Recent studies have emphasized that nonequilibrium conditions of postsynaptic GABAA receptor (GABAAR) activation is a key factor in shaping the time course of IPSCs (Puia et al., 1994; Jones and Westbrook, 1995). Such nonequilibrium, resulting from extremely fast agonist time course, may affect the interaction between pharmacological agents and postsynaptic GABAARs. In the present study we found that chlorpromazine (CPZ), a widely used antipsychotic drug known to interfere with several ligand and voltage-gated channels, reduces the amplitude and accelerates the decay of miniature IPSCs (mIPSCs). A good qualitative reproduction of the effects of CPZ on mIPSCs was obtained when mIPSCs were mimicked by responses to ultrafast GABA applications to excised patches. Our experimental data and model simulations indicate that CPZ affects mIPSCs by decreasing the binding (k on) and by increasing the unbinding (k off) rates of GABAARs. Because of reduction of k on by CPZ, the binding reaction becomes rate-limiting, and agonist exposure of GABAARs during mIPSC is too short to activate the receptors to the same extent as in control conditions. The increase in unbinding rate is implicated as the mechanism underlying the acceleration of mIPSC decaying phase. The effect of CPZ on GABAAR binding rate, resulting in slower onset of GABA-evoked currents, provides a tool to estimate the speed of synaptic clearance of GABA. Moreover, the onset kinetics of recorded responses allowed the estimate the peak synaptic GABA concentration.
- GABAA receptors
- miniature IPSCs
- binding and unbinding rate constants
Phenothiazines (PTZs) are a family of compounds commonly used in the treatment of psychiatric disorders. The mechanism whereby these drugs exert their therapeutic effects appears to be through blockade of dopamine receptors (Snyder et al., 1974; Seeman, 1980). However, these substances have been shown to affect also a number of other physiologically important sites. It is known, for instance, that PTZs compete for serotonin and α-adrenergic and histamine receptors (Peroutka and Snyder, 1980). More recently, electrophysiological studies have shown that PTZs interfere with a number of ligand- and voltage-activated channels (Gould et al., 1983;Sand et al., 1983; Changeux et al., 1986; Dinan et al., 1987; Zorumski and Yang, 1988; Ogata et al., 1989; Müller et al., 1991; Bolotina et al., 1992; Benoit and Changeux, 1993; Lidsky et al., 1997). In particular, it has been reported that PTZs block in a noncompetitive manner the responses evoked by exogenous application of GABA (Zorumski and Yang, 1988) and reduce the amplitude of IPSCs (Agopyan and Krnjevic, 1993), but the mechanism of these effects has not been elucidated. It is likely that the noncompetitive block of GABAA receptors by PTZs, reported by Zorumski and Yang (1988), is one of the processes underlying the reduction in amplitude of IPSCs observed by Agopyan and Krnjevic (1993). However, in the experiments of Zorumski and Yang (1988), GABA application system was too slow to mimic the time course of the agonist in the synapse, making it difficult to directly refer these results to the PTZs effects on synaptic currents.
A downregulation of synaptic inhibition may play an important role in the induction of epileptic activity. Thus, a further understanding of the effects of PTZ on inhibitory synaptic transmission appears to be particularly important because the use of PTZs may be associated (especially at high doses) with adverse effects, including seizures (Toone and Fenton, 1977; Itil and Soldatos, 1980).
The aim of the present work was to study the mechanisms underlying the effects of chlorpromazine (CPZ), a widely used PTZ, on miniature IPSCs (mIPSCs) in cultured hippocampal neurons. We report that CPZ reduces the amplitude of mIPSCs in a dose-dependent manner and accelerates their decay. The mechanism of CPZ action on GABAA receptors was studied using the ultrafast GABA applications to excised patches that fairly mimicked the synaptic currents. The results have been expressed in terms of a kinetic model, and it is suggested that the effects of CPZ are caused by a reduction in binding and increase in unbinding rates of GABAA receptor channel.
MATERIALS AND METHODS
Cell culture. Primary cell culture was prepared as described previously (Andjus et al., 1997). Briefly, hippocampi were dissected from 2- to 4-d-old rats, sliced and digested with trypsin, mechanically triturated, centrifuged twice at 40 × g, plated in the Petri dishes, and cultured for up to 12 d. Experiments were performed on cells between 5 and 12 d in culture.
Electrophysiological recordings. Currents were recorded in the whole-cell and outside-out configurations of the patch-clamp technique using the EPC-7 amplifier (List Medical, Darmstadt, Germany). In the case of whole-cell recordings of the synaptic and GABA-evoked currents, the series resistance (R s) was in the range of 4–8 MΩ, and 50–70% of R scompensation was accomplished. Both mIPSCs and GABA-elicited currents in the excised patch configuration were recorded at a holding potential (V h) of −70 mV. In the whole-cell configuration, GABA-evoked responses were often very large (>1 nA). To diminish a possible distortion caused byR s, the electrical driving force for chloride ions was reduced by setting the V h at −30 mV. The intrapipette solution contained (in mm) CsCl 137, CaCl2 1, MgCl2 2, 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid (BAPTA) 11, ATP 2, and HEPES 10, pH 7.2 with CsOH. The composition of the external solution was (in mm) NaCl 137, KCl 5, CaCl2 2, MgCl2 1, glucose 20, and HEPES 10, pH 7.4 with NaOH. mIPSCs were recorded in the presence of tetrodotoxin (TTX; 1 μm) and kynurenic acid (1 mm) to block fast sodium spikes and glutamatergic EPSCs. All the experiments were performed at room temperature 22–24°C.
The current signals were stored on a video recorder after pulse-code modulation. For the analysis requiring a high temporal resolution (e.g., rise time kinetics of synaptic or evoked currents), the signals were low-pass filtered at 10 kHz with a Butterworth filter and sampled at 50–100 kHz using the analog-to-digital converter CED 1401 (Cambridge Electronic Design Limited, Cambridge, UK) and stored on the computer hard disk. Otherwise, for analysis of slower events, the cutoff frequency of the filter and the sampling rate were lowered accordingly. The data and acquisition software were kindly given by Dr. J. Dempster (Strathclyde University, Glasgow, UK).
Two different perfusion systems for GABA applications were used: the multibarrel RSC-200 perfusion system (Bio-Logic, Grenoble, France) and the ultrafast system based on the use of a piezoelectric-driven theta glass application pipette (Jonas, 1995). The head of the multibarrel system was modified to improve the speed of drug application on cells adhering to the bottom of Petri dishes. This system was used either to evoke the whole-cell GABA-induced currents or to exchange the solution around the cell from which the synaptic activity was recorded. Judging from the onset of the liquid junction potentials, a complete exchange of the solution around the open-tip electrode occurred within 10–20 msec. A better indication of the exchange time around the cell was given by the rise time of whole-cell responses evoked by high concentrations of GABA (>1 mm). Because it is known that with those concentrations the rise time of GABA responses is less than or close to 1 msec (Maconochie et al., 1994; see also Results and Discussion), the observed rate of rise of the whole-cell current (15–30 msec) was mainly determined by the speed of the solution exchange. The piezoelectric translator used for ultrafast perfusion system was from Physik Instrumente (Waldbronn, Germany), and theta glass tubing was from Hilgenberg (Malsfeld, Germany). For this perfusion system, the open-tip recordings of the liquid junction potentials revealed that complete exchange of solution occurred within 80–250 μsec.
When studying the effect of CPZ on GABA-evoked currents, GABA was applied in the presence of CPZ, after pre-equilibration at the same CPZ concentration for at least 4 min.
Analysis. The decaying phase of the currents was fitted with a biexponential function in the form: Equation 1where A fast,A slow are the fraction of the fast and slow component, respectively, A s is the steady-state current, and τfast and τslow are the fast and the slow time constants. In the case of analysis of normalized currents, the fractions of kinetic components fulfilled the normalization condition: A fast +A slow + A s = 1. When fitting the IPSCs or deactivation currents, the steady-state componentA s was omitted.
To compare the time duration of the agonist presence in the case of different patterns of concentration time course, the effective exposureE was defined as following: Equation 2where c(t) is the concentration and t= time. In the case of the square-like (with amplitude A and time of application t appl) and exponential time course (A · exp(−t/τ)) the effective exposure would be: A ·t appl and A · τ, respectively.
The kinetic modelling was performed with the Bioq software kindly provided by Dr. H. Parnas from the Hebrew University, Jerusalem. The Bioq software converted the kinetic model (Fig. 8) into a set of differential equations and solved them numerically assuming, as the initial condition, that at t = 0, no bound or open receptors were present. Various experimental protocols were investigated by “clamping” the agonist concentration time course in the form of square-like pulses (ultrafast perfusion experiments) or of an exponentially decaying function (to model the synaptic clearance). The solution of such equations yielded the time courses of probabilities of all the states assumed in the model. The fit to the experimental data were performed by optimizing the values of rate constants for a given experimental protocol.
Data are expressed as mean ± SEM. Student’s unpairedt test was used for comparison of data.
CPZ reduces the amplitude of mIPSCs
mIPSCs were recorded in the whole-cell configuration of the patch-clamp technique, at a holding potential of −70 mV, in the presence of TTX (1 μm) and kynurenic acid (1 mm) in 9 of 15 cells. mIPSCs were GABAA-mediated because they were reversibly blocked by bicuculline (10 μm, data not shown). mIPSCs had a mean amplitude of −77.2 ± 15 pA (n = 9). CPZ applied in the bath at concentrations of 10–100 μm induced a dose-dependent decrease in mIPSC amplitude (Fig.1 A,B). As shown in the graph of Figure 1 E, in the presence of 10, 30, and 100 μm CPZ, the synaptic currents were decreased to 0.74 ± 0.1, 0.69 ± 0.12, and 0.09 ± 0.07 (n = 5) of the control value, respectively.
CPZ accelerates the decay of mIPSCs
The effectiveness of synaptic inhibition does not depend only on the amplitude but also on the time course of synaptic currents. Especially in the case of IPSC, the decay kinetics is particularly important for temporal integration of synaptic signals as the decaying phase of IPSCs are known to be long lasting, often exceeding hundreds of milliseconds (Edwards et al., 1990; Pearce, 1993; Puia et al., 1994;Jones and Westbrook, 1995).
Similar to previous reports (Edwards et al., 1990; Jones and Westbrook, 1995), the decay of mIPSCs recorded in our experiments could be well fitted with the sum of two exponentials (Fig. 1 A, Table 1). To compare the decay kinetics in control conditions and in the presence of CPZ, the synaptic events were averaged, and the traces were fitted with theoretical curves (Fig.1 A,B). When superimposing the normalized traces (Fig. 1 C), a clear acceleration of the decaying phase in the presence of 30 μm CPZ is evident. This effect was dose-dependent, being negligible at 10 μm, but gradually increased at higher CPZ concentrations (Table 1). The acceleration of mIPSC decay by CPZ was a consequence of a decrease in the area and in the time constant of the slower component (Table 1). Thus, a decreased amplitude and accelerated decay of mIPSCs are the two factors whereby CPZ reduces the charge transfer associated with a single mIPSC (charge transfer calculated as the integral of mIPSC). As shown in Figure 1 E, the CPZ-induced increase in the decay rate of mIPSCs contributes substantially to the decrease in the mIPSC area (charge transfer), especially at higher (≥30 μm) concentrations of CPZ.
We have also examined whether CPZ affects the rise time of mIPSCs. We found that in control conditions and in the presence of CPZ (10–100 μm) the 10–90% rise time was not significantly (p > 0.5) different (Fig. 1 D, Table 1).
CPZ affects whole-cell GABA-induced currents
To further explore the mechanisms underlying CPZ-induced decrease in the amplitude and acceleration of mIPSC decay kinetics, the effect of this drug was studied on the whole-cell currents evoked by applications of GABA from a multibarrel perfusion system (see Materials and Methods). These experiments were performed on the same neurons from which mIPSCs were recorded. The rise time of the whole-cell current induced by high doses of GABA (≥1 mm) was in the range of 15–30 msec. Figure 2 Ashows a typical example of the current evoked by GABA (1 mm, applied for 1.2 sec) at a holding potential of −30 mV. Application of the same GABA concentration in the presence of 100 μm CPZ led to a significant decrease in the amplitude and in the onset rate of GABA-evoked current (Fig.2 B,C). On average, the amplitude of GABA-induced currents in the presence of CPZ (100 μm) was 0.73 ± 0.03 of the control value (Fig. 2 D). This effect was independent of GABA concentration in the range from 5 to 1000 μm, suggesting a noncompetitive type of block. The most striking discrepancy between the effect of CPZ on mIPSCs and on the currents evoked by exogenous GABA application was the amount of their block by CPZ. As shown in Figure 2 D, whereas 100 μm CPZ reduced the mIPSCs amplitude by ∼90%, the same CPZ concentration induced only ∼30% depression of GABA-evoked responses. CPZ (100 μm) increased also the time-to-peak of GABA-evoked (1 mm) current from 21 ± 3.4 to 54 ± 7.4 msec (Fig. 3 C;n = 3). However, it is known that when GABA (1 mm) is applied with a fast perfusion system (Maconochie et al., 1994; Puia et al., 1994), the time-to-peak is close or <1 msec. This indicates that the rise time observed in our experiments reflects mainly the speed of drug application. The estimated rise time of the agonist concentration in the synaptic cleft (Clements, 1996) is at least two orders of magnitude faster than that obtained by our perfusion system. It is thus likely that the interaction of CPZ with mIPSCs occurs at a much faster time scale and for this reason cannot be reproduced by a conventional perfusion system. To test this possibility, a series of experiments was performed using an ultrafast perfusion system.
GABA applied by a ultrafast perfusion system mimics synaptic currents
The open-tip recordings of the liquid junction potentials showed that using the ultrafast perfusion system (see Materials and Methods), a complete exchange of solution occurred within 80–250 μsec. These values are similar to those reported by Jonas (1995) and Maconochie et al. (1994). Figure 3 A shows a typical current response to 1 mm GABA applied for 2 msec to an excised patch at a holding potential of −70 mV. Similar to what observed for mIPSCs, the currents were characterized by a biphasic decay described by two time constants (Table 2). The fast decay time constant (τfast) of mIPSCs and of GABA responses was not significantly different (p > 0.3; Tables 1, 2). However, the time constant of the slower component (τslow) was significantly longer (p < 0.05) in the case of currents evoked by fast GABA applications. The latter finding is similar to the observation of Jones and Westbrook (1995) who also found that the decay of GABA responses is slower than that of the synaptic currents. The rate of the rising phase was dependent on GABA concentration and similar to what was observed by Jones and Westbrook (1995), saturated at ∼3 mm GABA (Fig. 3 C). Thus, the above analysis shows that currents evoked by fast GABA applications reproduce qualitatively the major kinetic properties of mIPSCs. Taking this into account, such GABA responses were used as a tool to explore the mechanism of the effect of CPZ on the kinetics of mIPSCs.
The currents elicited by fast GABA applications (1 mm GABA for 2 msec) were also recorded in the presence of CPZ. Figure4 A shows typical responses to fast GABA applications in control conditions in the presence of 30 and 100 μm CPZ. On average, CPZ (30 and 100 μm) reduced the current amplitude to 0.71 ± 0.15 (n = 10) and 0.32 ± 0.09 (n= 17) of the control values (Fig. 4 C), respectively. As shown in Figure 4 C, the percentage of the amplitude reduction by 30 μm CPZ was very similar to that observed for mIPSCs but with 100 μm CPZ, the reduction of the current was smaller. However, with respect to the block of mIPSCs by CPZ, a much better agreement was obtained when mIPSCs were modelled by current responses evoked by fast GABA applications rather than by slower multibarrel perfusion (compare Figs. 2 D,4 D). This observation indicates that indeed the effect of CPZ on GABAA receptors strongly depends both on speed and time duration of agonist application.
CPZ affected also the decay of currents evoked by fast GABA pulses. At 10 and 30 μm CPZ, the effect was negligible, but at 100 μm a clear acceleration of the decay kinetics was observed (Fig. 4 B, Table 2) reproducing qualitatively the effect of CPZ on the mIPSC decaying phase.
What is the mechanism whereby CPZ affects so strongly mIPSCs and currents evoked by fast GABA applications? A hint to understand this problem came from the analysis of the effect of CPZ on rise time of currents evoked by brief pulses of GABA. As shown in Figure5 A, the onset of the current evoked by GABA (1 mm) in the presence of 100 μm CPZ was clearly slower than in control. On average, in control conditions, the 10–90% rise time was 1.12 ± 0.08 msec (n = 10), and in the presence of 100 μmCPZ the rise time was significantly slower (1.82 ± 0.06 msec;n = 10; p < 0.05). The effect of CPZ on the rise time of GABA-evoked currents was dose-dependent (Table 2). The observed decrease in the onset rate of GABA-evoked currents indicates that CPZ interferes with the activation kinetics of GABAA receptor. In terms of processes underlying receptor activation, this means that CPZ affects binding/unbinding rate constants and/or the transition rates between bound and open states. For instance, if CPZ reduces the binding rate, the presence of the agonist in the synaptic cleft could be too short to activate the receptors to the same extent as in control conditions. However, prolongation of the receptor exposure to agonist would increase the chance of forming bound receptors giving rise to larger currents. This hypothesis was tested by applying GABA pulses of different time duration using the fast perfusion system. As shown in Figure5 B, in the presence of 100 μm CPZ, the amplitude of GABA-evoked currents clearly increased with time of GABA application (six experiments). On the contrary, in control conditions, application of GABA for time ranging from 1 to 5 msec did not modify either the rise time or the current amplitude (Fig. 5 C; seven experiments). These results confirm that indeed CPZ interferes with the activation kinetics of GABAA receptors and that such effect is sufficient to decrease GABA-evoked current given that the agonist application is sufficiently short.
A similar effect was obtained by lowering the concentration of GABA from 1 to 0.3 mm (Fig. 5 D, four experiments). At such low concentration, binding of GABA (rate =k on · [GABA]; see also “Modelling” below) becomes rate limiting giving rise to a slower current rise time. Thus, when the presence of agonist is too short (e.g., 1 msec at 0.3 mm GABA), binding step is not completed, and, consequently, a smaller percentage of receptors can reach the open state.
As it will be discussed in details later, the rising phase depends basically both on binding and unbinding rates and on opening and closing rates of the bound receptor, and, a priori it is difficult to discriminate between these two mechanisms. However, the fact that the binding rate depends on the agonist concentration (k on · [GABA]), offers the opportunity to distinguish between these two possibilities. If CPZ lowers thek on rate constant, its effect on rise time would be equivalent to lowering the concentration of GABA (as in Fig.4 D). Moreover, the effect of CPZ on rise time should be compensated by applying a saturating dose of GABA (e.g., 10 mm).
Thus, to test whether CPZ affects the binding kinetics, first 1 mm and then 10 mm GABA pulses were applied in the presence of 100 μm CPZ. As shown in Figure6 A, application of 10 mm GABA strongly accelerated the rising phase with respect to 1 mm GABA. Moreover, the time-to-peak of currents evoked by 10 mm GABA in the presence of 100 μm CPZ (0.81 ± 0.06 msec; n = 6) and in control conditions (0.76 ± 0.05 msec; n = 5) were not significantly different (p > 0.3; Table 2; Fig.6 B,C). Therefore, the effect of CPZ on rise time of GABA-induced currents can be completely compensated by the increase in GABA concentration. In addition, at variance to 1 mm GABA responses (Fig. 5 B), when applying 10 mm GABA pulses of various time duration (1, 2, or 5 msec) in the presence of 100 μm CPZ, no differences either in amplitude or in time course were seen (Fig. 6 D, three experiments). Interestingly, on average, the amplitude of currents evoked by 10 mm GABA in the presence of 100 μm CPZ was only slightly smaller than that induced by the same GABA concentration in control conditions (Fig.6 E). This observation indicates that most of the observed CPZ-induced decrease in the amplitude of the GABA-evoked current was caused by a decrease in the binding rate.
In previous studies, Jones and Westbrook (1995, 1996) have demonstrated that the desensitization process of GABAAreceptors plays an important role in shaping inhibitory synaptic currents. These authors proposed that sojourns in the desensitized states preceding channel reopening effectively prolong the synaptic currents. Such single-channel openings separated by silent periods (sojourn in the desensitized state) were also observed in our experiments (Fig. 7 A). This finding indicates that also, in our case, desensitization participates in shaping mIPSCs and raises the possibility that CPZ could affect mIPSC decay kinetics by a modulation of the desensitization process. To test this hypothesis, long pulses of saturating doses of GABA (10 mm) were applied in the presence and in the absence of 100 μm CPZ. In these conditions, the binding reaction is not rate limiting, and the current time course is supposed to depend only on the opening and closing rates and desensitization kinetics. As shown in Figure 7, B and C, the currents evoked by long pulses of GABA (≥200 msec) could be well fitted by the sum of two exponentials (Eq. 1, Materials and Methods), indicating the presence of two desensitization components (in control: τfast = 8.7 ± 0.9 msec, A fast = 0.38 ± 0.03, τslow = 263 ± 56 msec, A slow= 0.35 ± 0.04, A s = 0.17 ± 0.05,n = 6; in 100 μm CPZ: τfast= 9.4 ± 0.9 msec, A fast = 0.59 ± 0.05, τslow = 143 ± 38 msec, Aslow = 0.21 ± 0.02, A s = 0.16 ± 0.04). The time constants of the fast components of the current decay were not statistically different (p > 0.5), indicating that the onset rate of the fast desensitization was not affected by CPZ. There was a difference in the percentage of the slow component of desensitization in controls and in the presence of CPZ (p < 0.05). However, because the time constant of slow desensitization is as slow as >150 msec (rate of onset < 0.007 msec−1), it is unlikely that this component would have any significant impact on the currents evoked by very brief GABA applications. Consequently, the lack of effect on CPZ on the kinetics of fast desensitization component indicates that the observed effect of CPZ on the deactivation is not caused by modification of the desensitization kinetics.
Kinetic modelling of currents evoked by brief GABA pulses
To provide a better quantitative description of the effect of CPZ on GABAA receptors, a kinetic model was investigated. The model (Fig. 8 A) proposed by Jones and Westbrook (1995) allows to predict all major properties of currents evoked by brief GABA applications. However, to adopt the model to our experimental data, the values of the rate constants were readjusted. The main difference between the prediction given by Jones and Westbrook’s parameters and our experimental data were the rise time of the responses evoked by brief GABA (1 mm) pulses (∼2 msec, Jones and Westbrook’s model; 1.12 ± 0.08 msec, our experimental data). Moreover, using their parameters, an increase in amplitude by ∼10% was predicted when increasing the time of GABA (1 mm) application from 1 to 5 msec. This finding is contrary to our experimental data, as shown in Figure 5 C, variation of application time within this interval did not lead to any significant change either in amplitude or in the time course of GABA-evoked currents. A good reproduction of the rising phase kinetics (Fig.9 A,B) was obtained by increasing the binding rate constant (k on) and the rate of transition from doubly bound to open state (β2) to the values indicated in Figure 8 B. The decaying phase was well reproduced by setting the unbinding rate k offand the rate constants governing the fast desensitization kinetics (d2, r2) as shown in Figure8 B. In the simulations of responses to fast GABA applications (≥1 mm; ≥1 msec) using the values of parameters as indicated in Figure 8 B, the occupancy of the singly bound open state (AR*) was very small (<2%).
Kinetic modelling of currents evoked by brief GABA pulses in the presence of CPZ
The model simulations were also used to explore the mechanisms underlying the observed effects of CPZ. The main differences between the responses recorded in the presence of CPZ and control ones were: (1) slower rise time, (2) faster deactivation kinetics, and (3) lower amplitude. As already mentioned, a decrease in the rate of onset of the GABA-evoked current could potentially involve modifications in:k on, k off, β2 and α2. As shown in Figure9 C,D, lowering k on is sufficient to predict a decrease both in amplitude (Fig. 9 C) and in the rate of onset (Fig. 9 D). Because β2and α2 are assumed not to depend on agonist concentration, the following observations argue against the possibility that the observed effect of CPZ on the current onset is related to modifications of these rate constants: (1) a decrease in the onset rate of current evoked by 1 mm GABA, observed in the presence of 100 μm CPZ (Fig. 5 A), can be reversed by increasing GABA concentration (Fig. 6 A), (2) the onset kinetics of currents evoked by a saturating GABA concentration (10 mm) in control conditions and in the presence of 100 μm CPZ are indistinguishable (Fig. 6 B), and (3) the currents evoked by 10 mm GABA in control conditions and in the presence of 100 μm CPZ have very similar amplitudes (Fig. 6 E). In addition, when modelling the effect of CPZ by decreasing β2 (leavingk on unchanged), it was impossible to reproduce the increase in current amplitude (Fig. 5 B) with duration of GABA application (simulation not shown). It is worth emphasizing that although the rate of binding reaction is determined by the reciprocal of (k on · [GABA] +k off), the kinetics of this process is dominated by the binding rate both in control conditions and in the presence of CPZ (k on · [GABA] ≫k off at [GABA] ≥ 1 mm, see also below for k off value). In our model, we mimicked the effect of 100 μm CPZ on the current onset by lowering the value of k on to 1–2 mm/msec (Fig. 8 B), leaving the values of β2 and α2 unchanged.
In our experiments we have also observed a very pronounced effect of CPZ on the decay kinetics of the evoked currents (Figs. 4 C,6 C). A decrease in k on, which accounts for the slower current onset and lower peak open probability, is insufficient to predict such strong acceleration of the current decay (Figs. 3, 4, 9 C,D). Although, as shown in Figure 9 D, a decrease in the binding rate tok on = 1.5 mm/msec gives rise to an acceleration of decay, but when increasing GABA concentration, a time course identical to control is restored. The last prediction results from the fact that the rate of binding reaction is determined by the product k on · [GABA] and, obviously, when multiplying [GABA] by the same factor (8 in Fig.9 C,D) by which k on was divided, the model resumes perfectly the control conditions. The acceleration of the decay kinetics obtained atk on = 1.5 mm/msec results from an increased probability of singly liganded open receptors (AR*, see legend of Fig. 9). Because the rate constant d 1is much smaller than d 2, the singly bound desensitized state (AD) cannot contribute to the decay kinetics to the same extent as A2D. However, in our experiments, the effect of CPZ on the decay kinetics was present also at saturating GABA concentration (10 mm, Fig. 6 C), indicating that other factors than k on must be also involved. In previous investigations, Jones and Westbrook (1995, 1997) pointed out that the decay kinetics of the deactivation currents strongly depends on the balance between the unbinding rate k offand the rate constants governing the fast desensitization (A2D). The reason for such relation is that the slower is the unbinding rate, the larger is the probability of “visiting” the desensitized state that, as already mentioned, leads to a prolongation of the deactivation currents. Thus, it is likely that the accelerated deactivation, observed in the presence of CPZ, involves a modification of the unbinding rate k off and/or of the fast desensitization kinetics. However, our data show that the onset kinetics of the fast desensitization is not affected by CPZ. In addition, as already mentioned, the slower desensitization phase is unlikely to affect the kinetics of the decay currents and for this reason this component was omitted in our model.
A good prediction of the effect of CPZ on the deactivation kinetics was obtained by increasing the value of the unbinding ratek off. As shown in Figure 9, E andF, an increase in the unbinding rate fromk off = 0.25 msec−1 to 0.5 msec−1, leads to a strong acceleration of the decay kinetics. When setting k off = 0.5 msec−1 (k on = 1.5 mm/msec), the faster deactivation kinetics is clearly present both in current evoked by 1 and 10 mm GABA (Fig.9 F). Moreover, as seen in Figure 9 F, the currents evoked by 10 mm GABA in control conditions and in the “presence of CPZ” (k on = 1.5 mm/msec; k off = 0.5 msec−1) have almost identical rising phases. These model predictions basically reproduce the effects of CPZ on current responses to fast GABA applications observed in our experiments (compare Figs. 5 A,B,6 B with Fig.9 E,F). Thus, the lack of effect of CPZ on the fast desensitization component and the above model simulation provide clear, although indirect, evidence that the mechanism underlying the acceleration of the deactivation kinetics is an increase in the unbinding rate k off.
In conclusion, the investigations described above show that the necessary requirement for a qualitative reproduction of the effect of CPZs on currents evoked by fast GABA application is a decrease in the binding and increase in the unbinding rate constantsk on and k off, respectively.
Kinetic modelling of mIPSCs in control conditions and in the presence of CPZ
In the previous sections we have pointed out that the currents evoked by brief pulses of GABA closely resembled mIPSCs and that the effects of CPZ on these currents were qualitatively similar. There were, however, some differences: (1) CPZ had no effect on the rising phase of the mIPSC but strongly affected the onset of the evoked currents and (2) at 100 μm concentration, CPZ decreased the amplitude of the synaptic currents to a larger extent than that of the evoked ones. It is possible that a source of these discrepancies is that the time course of the agonist applied using our perfusion system differs from that in the synaptic cleft. Although the time course of applied GABA concentration is expected to be of nearly rectangular shape, the time evolution of agonist in the synapse is different. The simulations as well as experiments (for review, seeClements, 1996) indicate that in the synapse, the onset of agonist concentration is very fast (time-to-peak ∼101μsec), and its clearance is characterized predominantly by a time constant in the range between 5 × 101 and 102 μsec (Clements, 1996; Silver et al., 1996). Because no perfusion system is able to apply the agonist with such kinetics, the model simulations offer a unique opportunity to explore the effects of CPZ on mIPSCs by extrapolating our experimental findings to the situation in the synapse. To simulate the synaptic response, the time course of the agonist was modelled by an exponential function:A · exp(−t/τ), where A is the maximum value of the agonist concentration and τ, the time constant. The time constant τ was chosen to be 100 μsec because, as it will be discussed later, a slower time constant of the agonist clearance (e.g., 200 or 300 μsec) would lead to a clear decrease in the “mIPSC” onset rate in the presence of CPZ (not seen in the experiments, Fig. 1 D). This value of agonist decay rate is in agreement with studies on the time course of agonist in the synapse (Clements, 1996; Silver et al., 1996). The peak concentrationA = 3 mm was chosen to assure the saturation of mIPSC. For instance, when choosingA = 1 mm, the maximum open probability was <0.4, and when increasing A, the maximumP open increased substantially, reaching saturation at A ≥ 3 mm(P open ∼0.78 at saturation). Thus, these simulations indicate that the peak synaptic concentration is at least 3 mm, as it is known that GABA released in the synapse is sufficient to saturate the amplitude of mIPSC (Mody et al., 1994). As shown in Figure 10, A andB, the synaptic currents modelled by “exponential agonist application” with parameters A = 3 mm and τ = 100 μsec had the rise time and decay kinetics almost indistinguishable from those obtained when simulating responses evoked by 1 msec pulses of 3 mm GABA. There was only a small difference in the maximum open probability (<5%, Fig.10 B).
As discussed above, the effect of CPZ was simulated by decreasing the binding rate k on and by increasing the unbinding rate k off (as indicated in Fig.8 B). As shown in Figure 10 C, the CPZ treatment results in a strong reduction in mIPSC amplitude, notably larger than in the case of effect of CPZ on the evoked currents. This prediction is in agreement with experimental observation that CPZ (100 μm) diminishes mIPSCs amplitudes to a larger extent than that of the evoked currents (Fig. 4 C). The simulation of the effect of CPZ on mIPSCs reproduced also the acceleration of the current decay kinetics (Fig. 10 D). Moreover, in agreement with the experimental data, in the presence of CPZ, the acceleration of deactivation is stronger for mIPSCs than for the currents evoked by brief pulses of GABA (Fig. 10 D). This difference is a consequence of a larger proportion of singly bound open states (AR*) in the case of mIPSCs (0.25 vs 0.11; Fig. 10, legend). Such larger percentage of singly bound states in mIPSCs results from a shorter exposure of receptors activated by “synaptic GABA application” (Fig.10 B,D,E,insets). Moreover, as shown in Figure 10, E andF, in the case of “synaptic” time course of the agonist, CPZ affected the rise time kinetics to a much smaller extent when compared with currents evoked by brief agonist application. However, when modelling the synaptic current, assuming slower time constant of agonist decay τ (e.g., 200 or 300 μsec), the model would predict a stronger effect of CPZ on the mIPSC rise time (similar to the effect of CPZ on the rise time of current evoked by 1 msec GABA application, data not shown) contrary to experimental evidence.
Thus, our analysis provides evidence that the differences between the effect of CPZ on mIPSCs and on the evoked currents are mostly caused by a different time course of agonist concentration in the two situations.
The major finding of the present work is that CPZ, a widely used antipsychotic drug, strongly affects the GABA-mediated inhibitory synaptic transmission by decreasing the amplitude and by accelerating the decay of mIPSCs. We provide evidence that these effects are caused by CPZ-induced changes in the binding and unbinding kinetics of GABAA receptors. Interestingly, our data show that the effects of CPZ on mIPSCs could be obtained at micromolar concentrations similar to those attained in the brain of psychotic patients (May and Van Putten, 1978).
A noncompetitive block of GABAA receptors by CPZ has been proposed by Zorumski and Yang (1988). The EC50 for CPZ reported by these authors (2.6 mm) was much larger than that obtained in our experiments even when using slower multibarrel perfusion system (30% of block by 100 μm CPZ indicates EC50 ∼230 μm). The source of this discrepancy is not clear. The limited speed of drug application used byZorumski et al. (1988) (as in the case of the multibarrel system) did not allow to elucidate the mechanisms of CPZ action. By using the fast perfusion system, we were able to observe the effects of CPZ with a much better time resolution, making it possible to get an insight into the mechanisms underlying the effects of CPZ. The modification of the kinetics of the mIPSCs was clearly of postsynaptic origin because this drug had a similar effect on the responses evoked by fast GABA applications to outside-out patches. The present data as well as the model simulations indicate that the decrease in the binding ratek on and the increase in the unbinding ratek off are sufficient to reproduce qualitatively the postsynaptic effects of CPZ on mIPSCs. In particular, the postulated decrease in k on by CPZ would slow agonist binding, making this process rate limiting for channel activation. Consequently, in conditions of very short exposure of GABAARs to GABA (especially in the case of mIPSC), a lower percentage of receptors would get bound and reach the open state. Strong acceleration of mIPSC decay kinetics is attributed mainly to CPZ-induced increase in the unbinding rate k off. According to the model (Fig. 8 A), the increase ink off leads to a decrease in the probability of visiting the desensitized states (not affected by CPZ), shortening thus the deactivation process (Jones and Westbrook, 1995, 1997). Thus, it seems that the mIPSC inhibition by CPZ is not a consequence of a direct block of channel pore by the drug but rather results from an “upsetting” of kinetics of the receptor, probably caused by an allosteric modulation by CPZ. An occlusion of channel pore by CPZ, if present, seems to be negligible because the current amplitudes elicited by saturating GABA concentration in control conditions and in the presence of CPZ are not significantly different (Fig.6 E). In addition, the observed mIPSC inhibition by CPZ does not fulfill the criteria of a classical competitive block. In the case of such antagonism, the receptor before it is activated, must dissociate the blocker molecule. We could suppose that the CPZ-induced decrease in the onset rate of GABA responses reflects a delay caused by dissociation of the CPZ molecule from the GABA-binding site. However, in classical mechanism of competitive block, antagonist dissociation rate does not depend on the agonist concentration. Thus, the fact that the effect of CPZ on rise time kinetics can be compensated by increasing dose of GABA (Fig. 6 A,B) argues against any significant contribution of competitive antagonism to the observed mIPSC inhibition by CPZ.
In our simulations we used the model (Fig. 8 A) previously proposed by Jones and Westbrook (1995). However, as described in Results, the values of the rate constants were changed to obtain a better fit to our experimental data. The most apparent was the difference between the rate constants determining the rising kinetics (k on, β2) reflecting faster current onset in our experiments. In the study of Jones and Westbrook (1995) (see Fig. 2 A), the rise time of the responses to 1 and 3 mm GABA were 3–4 msec and 1–2 msec, respectively, whereas in our experiments these values were 1.12 msec and 0.76 msec. This discrepancy may reflect different receptor kinetics in the two preparations and possibly a difference in speed of agonist application.
An interesting prediction of our model is that the effect of CPZ on mIPSCs and on currents evoked by brief GABA applications is associated with an increased proportion of monoliganded open states during current activation. This is a consequence of a slower activation kinetics while the agonist application remains short. As described in Results, the model simulations indicate that such increased proportion of monoliganded open states accelerates the deactivation kinetics because of a low probability of entering the singly bound desensitized state (AD). In our experiments, we have observed that CPZ affects the mIPSC decaying phase to a larger extent than that of the evoked currents [30 μm CPZ caused a strong acceleration of mIPSC decay (Fig.1), whereas its effect on GABA-evoked currents was negligible (Table2)]. A possible explanation of this difference could be a larger proportion of monoliganded open states during activation of mIPSC caused by much shorter receptor exposure to agonist in the case of the synaptic current.
CPZ effect reveals fast GABA clearance during mIPSC
The analysis of the effect of CPZ on mIPSCs and on evoked currents provides evidence that the agonist clearance is very fast (∼100 μsec). The key indication was the concomitant observation of a stronger CPZ effect on amplitude and smaller effect on the rising rate of mIPSC with respect to the currents evoked by brief GABA pulses. This indicates that during mIPSC, the presence of GABA is so short that in the case of slower receptor activation (caused by CPZ) less receptors activate, but still no effect is seen on the mIPSC onset. The above indications are nicely confirmed by the model simulations showing that for slower τ values of GABA clearance, the effect of CPZ on mIPSCs should be associated with a decrease in mIPSC rising rate (not observed in experiments). Thus, the effect of CPZ provides a tool to reveal kinetics of agonist clearance during GABAergic synaptic transmission.
Physiological implications from modelling mIPSCs
Although the fast perfusion system is an excellent and, so far, unique tool to mimic the synaptic events, the comparison of responses evoked using this technique to the synaptic ones should be done with caution. The main reason for possible discrepancies is that the synaptic agonist time course is still faster than that obtained with fast perfusion. For instance, we suggest that the differences in the effect of CPZ on mIPSCs and on evoked currents are consequences of faster agonist clearance during mIPSC.
Our data indicate that the peak GABA concentration in the synaptic cleft is at least 3 mm. This is supported by the observations that the rise time of currents evoked by fast GABA applications saturates at ∼3 mm and that this maximum rising rate is very similar to that of mIPSCs (Table 1, Fig. 3). However, this peak value of [GABA] differs substantially from those reported previously (Maconochie et al., 1994, 500–1000 μm; Jones and Wesbrook 1995, 527 μm). This discrepancy most likely reflects differences in the activation rates of GABA responses. Maconochie et al. (1994) have observed that the on rate of the evoked currents reaches saturation already at [GABA] = 1 mm, whereas in our experiments at [GABA] = 3 mm. Jones and Westbrook (1995), reported the IPSC rise time of ∼3 msec (rate of rise, 347 sec−1) whereas in our experiments the time-to-peak was 0.68 ± 0.09 msec. It needs to be emphasized, however, that also in a number of other studies a submillisecond rise time of IPSC was reported (Edwards et al., 1990;Maconochie et al., 1994; Bier et al., 1996; Mellor and Randall, 1998;Williams et al., 1998).
It seems interesting how sensitive the response of the postsynaptic receptors is to different patterns of agonist time course investigated in the present work. The effective exposure E (see Materials and Methods) for the simulation of synaptic current (A= 3 mm, τ = 100 μsec) is 10 times smaller than that for the square-like GABA pulse (3 mm for 1 msec) being 0.3 and 3 mm/msec, respectively. However, the charge transfer caused by the mIPSC is only slightly smaller (<5%) than that evoked by square GABA pulse. This suggests that receptors, bearing the properties predicted by our model, are particularly suitable to efficiently respond to agonist applications characterized by high peak concentration and very fast time course. Consequently, such receptors would assure a good performance of synaptic transmission. It is known that high peak agonist concentration is caused by a very small volume of the synaptic cleft, and that the fast time course is mainly caused by diffusion, that at such small distances is an extremely fast process (Hille, 1992). Thus, once released, the agonist time course must be very fast because its kinetics is determined by a spontaneous trend toward thermodynamic equilibrium. One can speculate that to assure an optimal performance of synaptic transmission, it was easier to “adopt” the receptor properties to the time course of agonist than vice versa. Consequently, such receptor properties may reflect a perfection of cooperation between agonist release and receptor activation kinetics in the synapse.
In conclusion, our results suggest that the agonist time course in the synaptic cleft is crucial not only for the kinetics of the synaptic currents but also for their modulation by pharmacological agents. A striking discrepancy between the effect of CPZ on mIPSCs and on the whole-cell GABA-evoked responses provides a clear example how large the impact of differences can be in nonequilibrium conditions underlying receptor activation. One of the practical implications of these findings is that the formalism based on the equilibrium Hill equation, widely used to describe the dose–response relationships, should be used with caution when describing the effects of various factors on synaptic currents. The reason for this is that this equation is based on the assumption that the system is in the equilibrium whereas, as already discussed, the synaptic transmission consists of a sequence of highly nonstationary processes.
This work was supported by a grant from Ministero dell’Università e Ricerca Scientifica e Tecnologica and from Consiglio Nazionale delle Ricerche. J.W.M. and K.M. were supported by Polish Committee for Scientific Research, research funds for Wroclaw University of Medicine. A.B. was supported by a fellowship from Novartis Pharmaceuticals. The Bioq software with which the kinetic modelling was performed was kindly provided by Dr. H. Parnas from the Hebrew University (Jerusalem).
Correspondence should be addressed to Enrico Cherubini, International School for Advanced Studies, via Beirut 2–4, 34 014 Trieste, Italy.