Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Serotonin or 5-hydroxytryptamine (5-HT) is a monoaminergic neurotransmitter modulating numerous neuronal functions. The 5-HT₃ receptor subtype is a serotonin-gated ion channel found in the peripheral and central nervous system (CNS) (Maricq et al., 1991; Tecott et al., 1993). This receptor is implicated in many brain functions. For example, antagonists of this receptor are important antiemetic drugs (Aapro, 1991). These antagonists potentially have other clinical applications (Costall et al., 1990; Gerlach, 1991; White et al., 1991), because they may act on other receptors. This is possible, because 5-HT₃, -aminobutyric acid (GABA_A), glycine, and nicotinic acetylcholine (nACh) receptors all belong to a superfamily of ligand-gated ion channels. The subunits of these receptors exhibit extensive amino acid sequence homology (Betz, 1990). Because of these similarities, many drugs that act on one type of receptor often act on other receptors in this group. For example, some 5-HT₃ receptor antagonists act on the GABA_A receptor complex in addition to their effects on 5-HT₃ receptors (Klein et al., 1994). Using patch clamp technique, we recently found that ondansetron, an antagonist of the 5-HT₃ receptor, suppresses GABA_A current of rat CNS neurons (Ye et al., 1997).

The glycine receptor/Cl channel (GlyR), like the GABA_A receptor complex, is a major inhibitory receptor in the adult mammalian CNS (Betz, 1991). Activation of GlyRs increases the postsynaptic membrane Cl conductance and inhibits neuronal excitation. Modulation of GlyR function would be expected to alter neuronal excitability. The fact that the potent convulsant agent strychnine selectively antagonizes the GlyR clearly demonstrates the importance of the GlyR in the CNS. We undertook the present experiments to determine whether ondansetron modulates the glycine response of hippocampal neurons. Our results demonstrate that ondansetron suppresses glycine-induced responses, which may contribute to in vivo effects of ondansetron.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of Neurons and Electrophysiological Recording. Hippocampal CA1 pyramidal neurons were prepared as described previously (Ye et al., 1997). Briefly, 5- to 11-day old (neonatal group) and 24- to 30-day old (mature group) Sprague-Dawley rats were decapitated. Their brains were quickly excised and placed into iced standard external solution containing: 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 10 mM glucose, and 10 mM HEPES; pH was adjusted to 7.4 with Tris base and osmolarity to 320 mM/kg with sucrose. The brain was then glued to the chilled stage of a vibratome (Campden Instrument, LTD, Cambridge, England) and sliced to a thickness of 400 µm. Slices were transferred to standard solution containing 1 mg pronase/6 ml and incubated (31°C) for 20 min. After an additional 20-min incubation in 1 mg thermolysine/6 ml, micropunches of the hippocampal CA1 region were isolated and transferred to a 35-mm culture dish. Mild trituration of these tissue punches through heat-polished pipettes of progressively smaller tip diameter served to dissociate single neurons. Within 20 min of trituration, isolated neurons had attached to the bottom of the culture dish and were ready for electrophysiological experiments. The experimental protocol was approved by the Institutional Animal Care and Use Committee.

Chemical Application. Solutions of glycine (Sigma Chemical Co.) and ondansetron hydrochloride (Glaxo Wellcome Inc., Hertfordshire, England) were prepared on the day of experiments. These solutions were applied to a dissociated neuron with a superfusion system via a multibarreled pipette as described previously (Ye et al., 1997). The tip of the superfusion pipette was normally placed 50 to 100 µm away from the cell, a position that allowed rapid as well as uniform drug application and preserved the mechanical stability of the cell. By keeping the dead volume small and the flow rate high, solution exchange could be accomplished within 15 ms. Throughout all experimental procedures the bath was continuously perfused with the standard external solution.

Data Analysis. Data were statistically compared using Student's t test or ANOVA, as noted. Statistical analyses of concentration-response data were performed using the nonlinear curve-fitting program (Sigma Plot). For all experiments, average values are expressed as mean ± S.E.M.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Ondansetron Inhibits Glycine Current (I_Gly). The effects of ondansetron on I_Gly were tested with gramicidin perforated patch technique (Abe et al., 1994). Gramicidin forms pores permeant only to monovalent cations. Therefore, the concentration of intracellular Cl and proteins is virtually undisturbed. As for hypothalamic neurons (Abe et al., 1994), the reversal potential for I_Gly of neonatal rat hippocampal neurons was between 50 and 10 mV with an average value of 25 ± 10 mV (n = 7). At a holding potential negative to the reversal potential, glycine induced inward current. As expected, I_Gly was sensitive to strychnine. For example, 20 nM strychnine depressed I_Gly to 50% (data not shown). As illustrated in Fig. 1, 10 to 54 µM ondansetron suppressed the peak current induced by 30 µM glycine. I_Gly completely recovered after ondansetron washout (Fig. 1A, e). On average, 13.5, 27, and 54 µM ondansetron decreased the peak amplitude of current induced by 30 µM glycine by 41 ± 3% (n = 6), 48 ± 1% (n = 9), and 54 ± 6% (n = 7), respectively. Ondansetron inhibition of current activated by 30 µM glycine was observed in all neurons tested (n = 50), and exhibited a clear concentration dependence (Fig. 1B). Fit of the Hill equation to the concentration-response data of Fig. 1B indicated that the IC₅₀ for ondansetron reduction of peak current induced by 30 µM glycine was 25 µM. Interestingly, there is a remarkable difference in the effects of ondansetron on the peak versus steady state I_Gly. Figure 1A (also Figs. 2 and 3) shows that ondansetron has essentially no effect on the current at the end of the pulse.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   A, records of current induced by 30 µM glycine in the absence (A, a-e) and presence of 13.5 (A, b), 27 (A, c), and 54 µM (A, d) ondansetron, respectively. Records are sequential current traces (from left to right) obtained from a single neuron. Bars above each current trace indicate the duration of drug application. Hatched bar indicates that ondansetron was present before and during glycine; that is referred to as the ++ protocol. Holding potential was 50 mV. B, concentration-response relation for ondansetron-induced decrease of IGly. The normalized peak IGly in response to 30 µM glycine plus varying concentrations of ondansetron is plotted as a function of the ondansetron concentration; peak IGly was first normalized to the peak IGly in response to 30 µM glycine alone. Each point is the mean of four to eight neurons. Vertical bars show ± S.E.M. IGly was recorded at a holding potential of 50 mV. For estimation of the dissociation constant (Kd) and the Hill coefficient (n) of the concentration-response curve, the Hill equation, I/IGly =1/(1+(C/Kd)n) was fit to the data; I is the peak current with ondansetron, IGly is the control peak IGly, C is the concentration of ondansetron. Note, data in Figs. 1-8 were obtained from neonatal rats whereas that in Fig. 8, D and E, were from a 26-day-old rat.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Dependence of ondansetron inhibition on glycine concentration. A, representative current traces in response to glycine at the concentrations indicated. B, ondansetron (54 µM) was coapplied with glycine with preincubation (++ protocol): note peak currents, especially in response to 10 and 30 µM glycine, were more sensitive to ondansetron. Holding potential was 50 mV. C, amplitude of the peak IGly was first normalized to the peak current value in response to 30 µM glycine and plotted as a function of glycine concentration. Each data point is the mean (±S.E.M.) for four to six neurons exposed to glycine alone () or glycine plus 54 µM ondansetron (). Solid lines are least square fit of the following form of the Michaelis-Menten equation to the experimental data: I = (IMax * Cn)/(Cn + Kdn) where, I, IMax, C, Kd and n are IGly, maximal IGly, glycine concentration, the concentration at which IGly is 50% of maximum and the Hill coefficient, respectively. D, double reciprocal plot of ondansetron suppression of IGly, suggesting a competitive inhibition.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Ondansetron has a stronger effect when applied before glycine. IGly was recorded in response to 30 µM glycine (open bars) alone or with 54 µM ondansetron (filled bars). A, a and B, a are currents in response to 30 µM glycine alone. A, b, 5-s prepulse of 54 µM ondansetron before the coapplication of ondansetron and glycine (++) suppressed peak IGly. A, c, coapplication of ondansetron and glycine without prepulse (+) suppressed both the peak and the steady state current. A, d, superimposition of traces a, b, and c. B, b, 5-s prepulse of 54 µM ondansetron did not affect IGly induced by subsequently applied glycine alone. B, c, superimposition of traces a and b. Current traces of A and B were from two different neurons. Holding potential was 50 mV in both cases.

Ondansetron Inhibited Only Current Induced by Subsaturating Concentration of Glycine. To test the hypothesis that ondansetron inhibits I_Gly by a competitive mechanism, the effects of ondansetron were tested on current induced by 10 to 1000 µM glycine. Figure 2 shows typical I_Gly records obtained without (A) and with (B) ondansetron. Ondansetron (54 µM) depressed I_Gly induced by 10 and 30 µM glycine. However, ondansetron had no effect on the current induced by 1 mM glycine. On average, 54 µM ondansetron decreased the amplitude of peak current activated by 10, 30, and 1000 µM glycine by 85 ± 5% (n = 6), 59 ± 5 (n = 7), and 6 ± 4% (n = 5), respectively. Figure 2C summarizes the concentration-response relationships for glycine (10-1000 µM) in control and in the presence of 54 µM ondansetron. The EC₅₀ and Hill coefficient of glycine was 40 µM and 1.9 in the absence of ondansetron and 73 µM and 1.8 in the presence of 54 µM ondansetron, respectively. The Lineweaver-Burke plot of Fig. 2D indicates that ondansetron competes with glycine for binding to the GlyR.

Ondansetron Has a Stronger Effect on I_Gly When Applied with a Prepulse. To determine the mechanism of ondansetron action on I_Gly, we compared ondansetron's effect on I_Gly when it was preapplied to neurons for 5 s before coapplication with glycine (++ paradigm, Fig. 3A, b) to its effect without this prepulse (+ paradigm, Fig. 3A, c). Superimposition of records a, b, and c produced Fig. 3A, d.This composite record reveals that ondansetron has a stronger effect on the peak I_Gly with the ++ paradigm. To determine the proper preincubation time, we evaluated ondansetron's effect with preincubation of 1 to 50 s. Although the inhibitory action of ondansetron on I_Gly significantly increased with preincubation time within 3 to 5 s, there was no significant change in the range of 5 to 50 s (data not shown). This suggests that ondansetron is at equilibrium with the receptor with 5-s preincubation time. To examine the possibility that ondansetron affects the receptors when they are not activated, we applied ondansetron alone for 5 s before the application of glycine alone (+ paradigm), as illustrated in Fig. 3B, b. Superimposition of records a and b produced Fig. 3B, c. This composite record reveals that ondansetron applied in a + paradigm has no effect on I_Gly.

Ondansetron's Action on I_Gly Has a Slower Onset than Offset. To explore the kinetics of the ondansetron action, we applied a brief pulse of 54 µM ondansetron during a longer lasting pulse of 30 µM glycine. Ondansetron decreased I_Gly immediately and with cessation of ondansetron, I_Gly recovered (Fig. 4A, a). As the continuous lines of Fig. 4A, d show a single exponential could fit both the onset and offset of ondansetron's effect. The time constant (n = 5) for onset and offset of the ondansetron effect was 1760 ± 30 ms and 385 ± 20 ms, respectively.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   The kinetics of ondansetron's effect. A, a, brief pulse of ondansetron (54 µM) reduced IGly in response to 30 µM glycine. A, b, control IGly in response to two consecutive pulses of glycine. A, c, control IGly in response to a continuously applied glycine pulse. A, d, both the onset and offset of ondansetron's effect could be fit by a single exponential, as shown by the continuous lines. A, e, superimposition of a, b, and c. Note the onset and offset of ondansetron's action is slower compared with the onset and offset of glycine alone. B, repeated pulses of ondansetron (54 µM) were applied during a continuous application of 30 µM glycine. Holding potential was 50 mV.

Ondansetron Does Not Increase Receptor Desensitization. Ondansetron inhibition of I_Gly could result from ondansetron enhancement of receptor desensitization. To test this hypothesis, we studied the desensitization of I_Gly in the absence and presence of ondansetron. As shown in Fig. 5A, the decay rate of current activated by 30 µM glycine was decreased, rather than increased, by application of 27 µM ondansetron. For six neurons, 27 µM ondansetron significantly increased the time constant of desensitization (Student's t test, p < .05).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effects of ondansetron on desensitization of IGly. A, current traces illustrating the time constants of desensitization (d) of current activated by 30 µM glycine in the absence and presence of 27 µM ondansetron. d values in A are those obtained for the records shown. Average values of d of current activated by 30 µM glycine were 3.9 ± 1.5 s and 7.3 ± 1.1 s in the absence and presence of 27 µM ondansetron, respectively; these values are significantly different (Student's t test, p < .01; n = 5). B, current traces illustrating the d of currents activated by 1 mM glycine, a saturating concentration, in the absence and presence of 54 µM ondansetron. d values in B are those obtained for the records shown. Average values of d of current activated by 1 mM glycine in the absence and presence of 54 µM ondansetron were not significantly different (2.9 ± 0.3 s versus 3.2 ±0.2 s, respectively, Student's t test, p > .5; n = 5). Desensitization of the IGly in A and B was well fit using a single-exponential equation. Holding potential was 50 mV.

Ondansetron Decreases Receptor Activation Rate without Affecting Deactivation Rate. The preceding data suggest that ondansetron may inhibit I_Gly by competing with the agonist for the same binding site on the GlyR. To further examine this hypothesis, we determined the activation and deactivation time constants of I_Gly in the absence and presence of ondansetron. To allow accurate measurement of time constants within the limits of the fast perfusion system (time constant of 9 ± 1.3 ms for solution changes in patch-clamped cells; Fig. 6A, inset), we used glycine concentrations of 30 µM or lower. As shown in Fig. 6A, in the presence of 0, 108, and 218 µM ondansetron, the activation time constant _on was 284, 860, and 1003 ms, respectively. The activation time constants (Fig. 6B) were highly dependent on both glycine concentration (ANOVA p < .01; n = 6) and ondansetron concentration (ANOVA p < .01, n = 6). In contrast, the deactivation time constants (Fig. 6C) were dependent on glycine concentration (ANOVA p < .05, n = 5) but independent of ondansetron concentration (ANOVA p > .25; n = 5).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Effect of ondansetron on activation and deactivation time constants of glycine-activated current. A, records of currents activated by 30 µM glycine in the absence and presence of 54 and 109 µM ondansetron. Both activation and deactivation of the current were well fitted using single-exponential equations (continuous curves). Horizontal bars indicate that we applied ondansetron for 5 s before and after application of glycine to eliminate the possible effect of the rates of onset and offset of ondansetron action on the measurement of the rates of activation and deactivation of the receptor by glycine. Note that both 54 and 109 µM ondansetron had relatively little effect on deactivation time constant (off), but greatly increased the activation time constant (on). Inset shows the time constant for solution change in a patch-clamped cell. The external solution containing 100 µM kainate was abruptly changed from 140 mM Na+ to 10 mM Na+ plus 130 mM N-methyl-D-glucamine (NMDG). B, graph plotting averaged on values as a function of glycine concentration in the absence (), and presence of 54 (), and 109 µM () ondansetron. The average on of glycine induced current was highly dependent upon ondansetron concentration (ANOVA, p < .01). Holding potential was 50 mV. C, graph plotting average off values as a function of glycine concentration in the absence (), and presence of 54 (), and 109 µM () ondansetron. Average off of IGly depends on glycine concentration (ANOVA, p < .05; n = 5), but was independent of ondansetron concentration (ANOVA, p > .25; n = 5).

Ondansetron Depression of I_Gly Is Independent of Voltage. To explore the voltage dependence of ondansetron's action, we studied the current-voltage relationships (I-V) of I_Gly in the absence and presence of ondansetron with a voltage ramp protocol. As illustrated in Fig. 7, over the voltage range of +20 to 70 mV, the I-V curves are linear. Ondansetron depressed I_Gly to a similar extent at all voltages tested. Thus, the effect on I_Gly is not voltage dependent. Furthermore, the glycine-activated channel remained selectively permeable to Cl because the reversal potential of I_Gly was approximately 31 mV with or without ondansetron.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Ondansetron depression of IGly is independent of voltage. Effect of ondansetron on the current-voltage relationship of IGly was studied with a voltage ramp protocol. A pair of voltage ramps ranging from + 20 to 70 mV was applied to the neuron at a rate of 1 mV/10 ms. Drugs were applied to the cell and covered the second ramp in each pair. Traces obtained from the first voltage ramp measured background or leakage current. The current-voltage curve was produced by subtracting the trace obtained in the first ramp from that in the second ramp. A, typical IGly recorded from a neuron exposed to 30 µM glycine alone and in combination with 27 µM ondansetron. B, current-voltage curve derived from A shows that ondansetron depressed IGly at all potentials without changing the apparent reversal potential of this current. Similar data were obtained from three other cells. C, to determine the voltage dependence, currents recorded in control and in the presence of ondansetron were first normalized to the values obtained at +20 mV. Normalized current-voltage relations from the same experiment as B showing that ondansetron depression is not voltage dependent.

Ondansetron Suppresses Glycine-Induced Membrane Potential Changes. The ondansetron effect on the glycine response was also studied under current clamp conditions. As stated earlier, the average reversal potential of I_Gly was 25 mV in the neonatal group. Because resting membrane potentials (68 ± 2.5 mV, n = 5) are more negative than the reversal potential of glycine at this age, GlyR activation would produce membrane depolarization. As illustrated in Fig. 8A, glycine induced membrane depolarization. On average, 30 µM glycine induced 20 ± 5 mV (n = 6) depolarization. This effect of glycine was also concentration dependent, with an EC₅₀ of 40 µM, close to the EC₅₀ for I_Gly. Ondansetron suppressed the depolarizing response induced by glycine (Fig. 8B). Figure 8C illustrates the concentration-response relationships of glycine in the absence and presence of ondansetron. Ondansetron shifted the curve to the right without affecting the maximal voltage change induced by glycine. Similar experiments were repeated in hippocampal neurons dissociated from 26- to 30-day old rats. In accord with other studies (Ito and Cherubini, 1991), glycine induced hyperpolarization (Fig. 8D). Figure 8E shows that 27 µM ondansetron reduced glycine-induced hyperpolarization to 46%; similar results were obtained from two other neurons. These data indicate that the glycine-induced response of mature neurons has a sensitivity to ondansetron similar to that of neonatal neurons.

View larger version (13K): [in this window] [in a new window]   Fig. 8.   Ondansetron suppressed glycine-induced changes of membrane potential. A, glycine (30 µM) induced depolarization of a hippocampal neuron of a neonatal rat is reduced by ondansetron (B). C, amplitude of the peak glycine induced depolarization (VGly) was first normalized to the peak VGly value in response to 30 µM glycine and plotted as a function of glycine concentration. Each data point is the mean (±S.E.M.) for 4 to 10 neurons exposed to glycine alone () or glycine plus 54 µM ondansetron (). Solid lines are least-squares fit of the following form of the Michaelis-Menten equation to the experimental data: V = (VMax * Cn)/(Cn + Kdn) where, V, VMax, C, Kd, and n are VGly, maximal VGly, glycine concentration, the concentration at which VGly is 50% of maximum, and the Hill coefficient, respectively. D, glycine (30 µM) induced hyperpolarization of a hippocampal neuron of a mature (26-day-old) rat is reduced by ondansetron (27 µM, E). Similar results were observed in three other neurons. Calibration, vertical bar, 25 mV for A and B, 13 mV for D and E.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

5-HT₃ receptor antagonists inhibit GABA actions because they act as inverse agonists at the benzodiazepine site on GABA_A receptors (Klein et al., 1994). Recently, we reported that ondansetron inhibits GABA current of rat central neurons (Ye et al., 1997). Squires and Saederup (1999) demonstrated that ondansetron reversed the inhibitory effect of 1 µM GABA on [³⁵S]t-butylbicyclophosphorothionate binding to whole rat forebrain membranes. In this study, we describe the depressant effects of ondansetron on the glycine-induced response of rat hippocampal pyramidal neurons.

The data indicate that ondansetron inhibits the glycine receptor by shifting the agonist concentration-response curve to the right in a parallel manner without affecting the maximal response to glycine. This effect could result either from competitive inhibition by ondansetron, or by interaction with an allosteric site on the receptor channel resulting in a decreased affinity of the receptor for glycine. The latter mechanism has been demonstrated for inhibition of GABA_A receptors by benzodiazepine site inverse agonists (Kemp et al., 1987). Competitive antagonists will decrease the activation rate of the receptor channel without changing its deactivation rate (Clements and Westbrook, 1991). Allosteric antagonists that decrease the affinity of the receptor for agonist will increase the deactivation rate without changing its activation rate (Li et al., 1997). In the present study, ondansetron decreased the activation rate without affecting the deactivation rate of I_Gly. This observation suggests that ondansetron may compete with the agonist for binding to the GlyR. The observation that I_Gly recovered after a brief pulse of ondansetron suggests that during the pulse, GlyRs were occupied by ondansetron instead of glycine.

There are several explanations for ondansetron's weaker effect on steady state than peak I_Gly. For example, ondansetron may reduce desensitization of the GlyRs. That is, in the early part of the pulse, ondansetron occupied an agonist binding site of the GlyRs and thus prevented activation of the GlyR. Because desensitization of GlyRs is proportional to activation of GlyRs, there will be less desensitization. Consequently, the steady state I_Gly has a greater amplitude. This hypothesis is supported by the observation that ondansetron reduced desensitization of GlyRs (Fig. 5). Alternatively, there may be on rate competition between ondansetron and glycine. Presumably, in the ++ paradigm, GlyRs were occupied by ondansetron. Therefore, less GlyRs were available to bind glycine when it arrived. However, as the pulse continued, glycine might bind by competing with ondansetron, as suggested by the gradual increase of I_Gly (Figs. 2B, b and 3A, b). If the resultant increment of I_Gly offsets the ondansetron depression of I_Gly, ondansetron would not affect the steady state current. The competitive nature of ondansetron's effect supports this hypothesis.

In addition to competitive, noncompetitive block is also a major mechanism underlying receptor inhibition. Open-channel block and increase of desensitization are two common mechanisms underlying noncompetitive inhibition. Because ondansetron has a pK_a of 7.4 (Glaxo-Wellcome), it is 50% charged in the external solution used. Because ondansetron inhibited I_Gly only in the presence of agonist, ondansetron may act as an open channel blocker. However, two lines of evidence do not support this hypothesis. First, open-channel block by charged molecules is usually voltage dependent (Hille, 1992); ondansetron inhibition of I_Gly was independent of voltage. Conceivably, ondansetron could bind to a site within the ion channel beyond the influence of the membrane electrical field. In this case, ondansetron could be an open channel blocker but independent of voltage. This mechanism is unlikely, because ondansetron inhibition can be overcome by increasing the concentration of glycine. Analysis of the current-voltage relationships reveals that ondansetron did not change the ion selectivity because the reversal potential of I_Gly did not alter in the presence of ondansetron. Secondly, use dependence is a feature often associated with an open channel blocker. However, repeated application of ondansetron suppressed I_Gly to a similar extent. Although ondansetron could depress I_Gly by increasing desensitization of the receptor, this is unlikely because ondansetron actually decreased the rate of desensitization of current activated by a submaximal concentration of glycine. Finally, ondansetron did not alter the desensitization rate of current activated by a saturating concentration (1 mM) of glycine.

The suppression of peak I_Gly by the ++ paradigm is much larger than the + paradigm. Because ondansetron has a molecular weight of 365.9, much greater than glycine (75.1), it will take much longer for ondansetron to reach its site of action than glycine. In the + paradigm, ondansetron has not been allowed to pre-equilibrate with the GlyR before the presentation of the agonist. The onset time constant of 1790 ms supports this hypothesis, explaining why ondansetron suppressed I_Gly less when applied with a short pulse as shown in Fig. 4A, a (see below for details).

The fact that both the onset and offset of ondansetron action could be fit by a single exponential suggests that the ondansetron-glycine receptor interaction can be modeled as a simple bimolecular reaction and expressed as:
where R and D are glycine receptor and ondansetron, respectively; k₁ and k₁ are the forward and backward rate constants, respectively. The time constant for the onset and offset could be expressed as 1/(k₁[D] + k₁) and 1/k₁, respectively, where [D] is the concentration of ondansetron. The experiments of Fig. 4A, d give time constants for onset and offset of 1790 ms and 397 ms in the presence of 54 µM ondansetron. These values result in k₁ = 18.0 M¹ s¹ and k₁ = 2.52 s¹. Thus, the apparent dissociation constant K_D is 69 µM. This value is much higher than 25 µM ondansetron required to inhibit 50% I_Gly activated by 30 µM glycine. This difference is consistent with the observation that ondansetron had a stronger effect when applied with pretreatment (++ paradigm) and the peak I_Gly was more sensitive to ondansetron than the steady state I_Gly. An alternative interpretation is that, because glycine has a considerably faster "on" rate (_on = 284 ms) than ondansetron (1790 ms), ondansetron applied in the middle of a longer lasting pulse of glycine would only partially inhibit I_Gly, thus increasing the IC₅₀ value for ondansetron. Because the forward rate constant k₁ (18.0 M¹ s¹) is much slower than free diffusion in solution, the binding site for ondansetron is not freely accessible.

There are several differences between ondansetron effects on I_Gly and current induced by GABA (I_GABA). First, peak I_Gly is more sensitive to ondansetron than the steady-state current. This is just the opposite of I_GABA, where the steady-state, rather than peak current was more sensitive to ondansetron. Secondly, although ondansetron inhibition of steady-state I_GABA is noncompetitive, the effect on peak I_Gly is competitive. Furthermore, in contrast to the observation that ondansetron had an effect on the resting state of the GABA_A receptor, ondansetron had no effect on the GlyR in its resting state. Finally, ondansetron had a faster onset than offset of effect on I_GABA; the converse was observed for I_Gly.

Glycine and GABA are the principal inhibitory neurotransmitters in the adult mammalian CNS. Several studies show that glycine receptor activation depolarizes embryonic and neonatal neurons but hyperpolarizes mature neurons (Abe et al., 1994; Chen et al., 1996; Ito and Cherubini, 1991). In this study, we demonstrated that glycine depolarized and hyperpolarized neurons of neonatal and mature rats, respectively. Ondansetron suppressed both effects of glycine. Thus, it is possible that ondansetron inhibition of the glycine response can have different effects on the CNS of neonates and adults. That is, by suppressing the glycine and GABA response, ondansetron decreases and increases the excitability of the CNS of neonates and adults, respectively.

The clinical significance of ondansetron's effect on I_Gly is obvious. Despite the fact that the effects of ondansetron observed here were obtained with concentrations that are relatively high compared with those expected to arise from clinically safe doses, ondansetron could be much more potent when applied in vivo. Studies have shown that in the intact brain, whenever the potency of GABAergic and/or glycinergic inhibition is diminished, epileptiform activity appears (Krnjevi, 1983). It is possible that in vivo synaptic inhibition is reduced at ondansetron concentrations much lower than the IC₅₀ for in vitro suppresses GABA or glycine responses. The loss of inhibitory restraint thereby permits unopposed excitatory drive, leading to hyperexcitation and convulsions. In view of the critical importance of the glycinergic system in the central nervous system, the antiglycine action of ondansetron may result in behavioral changes after repeated use. Thus, in addition to producing an antiemetic effect by blocking 5-HT₃ receptors, ondansetron can also produce CNS disinhibition by blocking glycine receptors. The later action is likely to play a significant role in the ondansetron-induced seizures observed in vivo.