Control of interneurone firing pattern by axonal autoreceptors in the juvenile rat cerebellum
1 Laboratoire de Physiologie Cérébrale, CNRS UMR 8118, 45 rue des Saints Pères, 75006 Paris, France
Abstract
Recent work has shown that certain neurones have axonal GABAA receptors, whose tonic activation modifies their firing properties and neurotransmitter release capability. In addition, results obtained in interneurones of the molecular layer of the cerebellum indicate that action potential-released GABA binds back to the axon that released it, generating an autoreceptor current. In the present paper, we show that at physiological Cl–i concentration (15 mM) and at 34–36°C, the autoreceptor current generates a large amplitude (up to 21 mV) afterdepolarization that lasts for about 150 ms, and that occasionally leads to double firing. Furthermore we show that elimination of the afterdepolarization, by either blocking GABAA receptors, or eliminating the autoreceptor currents through prolonged whole-cell recording, decreases burst firing. Ih (a hyperpolarization-activated current) was previously found to be prominent in interneurone axons. We show that blocking Ih leads to an increase in the amplitude of the autoreceptor current as well as of the associated afterdepolarization, suggesting a shunting effect of Ih on autoreceptor-mediated afterdepolarization. Conversely, blocking Ih accentuates burst firing. The effects of autoreceptor-mediated afterdepolarization on firing are prominent during a period of development when interneurone synapses are stabilized and vanish by postnatal day 17 (PN 17), together with the expression of the autoreceptor current. Altogether, this work reveals a new role for autoreceptors in the regulation of cell excitability and firing pattern, which may contribute to the development and stabilization of the cerebellar network.
, 百拇医药
Introduction
Molecular layer interneurones (MLIs) form a network of interconnected cells in the cerebellar cortex. Like many other GABAergic neurones throughout the brain, MLIs display spontaneous activity both in vitro (Midtgaard, 1992a; Llano & Gerschenfeld, 1993a; Vincent & Marty, 1996; Husser & Clark, 1997) and in vivo (Jorntell & Ekerot, 2003). MLIs make powerful GABAergic synapses onto Purkinje cells and other MLIs, in both cases modulating the firing of their postsynaptic cells (Midtgaard, 1992b; Vincent et al. 1992; Callaway et al. 1995; Husser & Clark, 1997; Kondo & Marty, 1998; Chavas & Marty, 2003).
, 百拇医药
MLI firing can be periodic (Midtgaard, 1992a; Husser & Clark, 1997; Carter & Regehr, 2002), irregular (Jorntell & Ekerot, 2003) or bursting (Mann-Metzer & Yarom, 2002), but the mechanisms responsible for switching from one mode to another are poorly understood. MLIs exhibit several intrinsic conductances around the resting membrane potential, including Ih (Midtgaard, 1992a; Southan et al. 2000), and a non-inactivating Na+ current (Mann-Metzer & Yarom, 2002) that contributes to determining the duration of the train of action potentials induced by an excitatory current. Following a train of spikes, MLIs exhibit a biphasic hyperpolarizing afterpotential (Midtgaard, 1992a), which presumably reflects, at least in part, the activation of axonal Ca2+-dependent channels (Tan & Llano, 1999) by a biphasic calcium concentration decay (Collin et al. 2005).
, http://www.100md.com
Recently a novel tail current was described in voltage-clamped MLIs following depolarization-induced axonal firing (Pouzat & Marty, 1999). This current was abolished when blocking either GABAA receptors (GABAARs) or neurotransmitter release (either internally, using a calcium buffer, or externally, by blocking voltage-dependent calcium entry). It appeared to reflect the activation of axonal GABAARs following action potential-mediated release of GABA from the cell axon, and was therefore called autoreceptor current.
, 百拇医药
Several recent reports show that the activation of presynaptic ionotropic receptors for glycine and GABA can modulate neurotransmitter release and consequently synaptic transmission either in a facilitatory or inhibitory way (Jackson & Zhang, 1995; Turecek & Trussell, 2001, 2002; Belenky et al. 2003; Kawa, 2003; Axmacher & Draguhn, 2004; Ye et al. 2004). Likewise, sustained activation of these receptors has been shown to change the excitability of the presynaptic cell in the spinal cord (Rudomin et al. 1981), in the pituitary gland (Jackson & Zhang, 1995) and in hippocampal mossy fibres (Ruiz et al. 2003). However, the potential effects of phasic activation of axonal GABAA autoreceptors on cell excitability have not been explored. Since the activation of autoreceptor currents after each action potential might induce changes in the probability of firing, we studied the effects of the activation of autoreceptor currents on the spontaneous firing pattern of MLIs.
, 百拇医药
Our results show that the autoreceptor current does contribute to shaping the firing pattern of MLIs, thus revealing a new mechanism capable of modulating the excitability of central neurones and of inducing switches between various modes of firing.
Methods
Preparation
We used Sprague Dawley rats aged from 8 to 17 days. We obtained parasagittal, 180–200 μm thick cerebellar slices as previously described (Llano & Gerschenfeld, 1993a). Animals were decapitated after either cervical dislocation or deep anaesthesia with Metofane (Janssen; 0.5 ml in 2 l administered for 1–2 min). The extracellular solution used both for slicing and recording contained (mM): 125 NaCl, 2.5 KCl, 2 CaCl2, 1 MgCl2, 1.25 NaH2PO4, 26 NaHCO3, and 10 glucose. The pH was 7.4 when equilibrated with a 5% CO2 and 95% O2 mixture. Slices were maintained at 34°C for 1–5 h and transferred to the recording chamber as needed for the experiments. In the recording chamber, slices were continuously perfused at a rate of 1–1.5 ml min–1 with extracellular solution pre-equilibrated with a 5% CO2 and 95% O2 mixture. During recording, the chamber was heated in order to maintain a temperature of 34–36°C.
, http://www.100md.com
Both types of MLIs were included in the study. Following the criteria used by Llano & Gerschenfeld (1993a), cells located on the lower third and upper two thirds of the molecular layer were, respectively, classified as basket and stellate.
Drugs
(–)-Bicuculline methochloride (10 μM, Tocris), SR 95531 hydrobromide (Gabazine, 20 μM, Tocris) and picrotoxin (200 μM, Sigma) were used as pharmacological blockers of GABAARs. 4-Ethylphenylamino-1,2-dimethyl-6-methylaminopyridinium chloride (ZD7288, 10 μM, Tocris) was used as a blocker of Ih.
, http://www.100md.com
A total number of 77 experiments were performed to test the effects of various manipulations (application of GABAAR blockers, of ZD7288, and prolonged whole-cell recording) on the bursting behaviour of MLIs. Of these, 11 experiments (between 1 and 3 for each protocol) were done in the presence of 2,3-dihydroxy-6-nitro-7-sulphonyl-benzo[f]quinoxaline (2 μM, NBQX, Tocris), in order to check for the possible dependence of the phenomena studied on ionotropic glutamate receptor activity. Since these were indistinguishable from those performed without NBQX, the results in the presence and absence of this blocker were pooled together. All drugs were bath applied.
, http://www.100md.com
Electrophysiology
Whole-cell and cell-attached patch clamp recordings were performed using the following intracellular solution (mM): 10 GABA, 139 potassium gluconate, 6 KCl, 4.6 MgCl2, 0.1 EGTA, 10 Hepes, 0.4 Na2GTP, 4 Na2ATP; pH 7.4. The total Cl– concentration in this solution was 15 mM, which corresponds to the physiological value (Chavas & Marty, 2003). GABA was included since preliminary experiments showed that addition of 10 mM GABA to the pipette solution significantly slowed down the rundown of autoreceptor currents (J. Chavas, unpublished experiments; see also Smith & Jahr (2002) for the description of a similar effect in granule cells of the olfactory bulb). Pipette resistances were around 5–7 M for whole-cell and around 8–11 M for cell-attached experiments. In whole-cell mode, series resistance values during recording ranged from 10 to 35 M and were not compensated. Experiments were discontinued if the series resistance of a cell increased above 35 M. Cells were rejected if, at any moment during the experiment, holding currents in excess of 40 pA were required to maintain a potential of –70 mV.
, 百拇医药
Autoreceptor current. To test for the activation of autoreceptor currents we used a modification of the protocol described by Pouzat & Marty (1999). Briefly, in whole-cell configuration, cells were voltage clamped to –90 mV. A 1–2 ms long, 90 mV large depolarizing pulse was applied to induce the activation of a fast inward Na+ current, reflecting the generation of an action potential in the axon. The autoreceptor current was manifested as a slow inward tail current following the Na+ spike. This test was performed for each cell included in this study. Autoreceptor current amplitudes were determined 7 ms after the peak Na+ current. A high amplitude variability as well as a high rate of failures are characteristic features of autaptic currents (Pouzat & Marty, 1998; see the discussion on the distinction between autaptic currents and autoreceptor currents in Pouzat & Marty, 1999). Therefore, in each experiment, tail currents were examined for failures, and the coefficient of variation (CV) value for the amplitudes of the tail current was measured. In most cases, no failure was observed, and the CV had values between 0.12 and 0.24, indicating an absence of autaptic currents. In a single case, however, failures were observed, and the CV value was 0.49, indicating the presence of an autapse; this cell was removed from the analysis. The low incidence of autapses in PN 8–17 preparations which is apparent from these results is consistent with our previous study (Pouzat & Marty, 1999).
, 百拇医药
Afterpotential. In current clamp mode a holding current up to –40 pA was applied to keep the cells at a stable membrane potential between –60 and –75 mV, within the range of resting membrane potentials of MLIs (as described by Chavas & Marty, 2003). In these conditions, 20–60 ms depolarizing steps were applied to induce either a single action potential or up to four action potentials. The duration and amplitude of the steps were adjusted such that the down stroke of the last action potential coincided with the end of the step. In this way confounding effects of aborted spikes on the afterpotential were avoided. Ten to twenty traces were averaged. The afterpotential was measured as the integrated area of the membrane potential profile over a 200-ms-long time period starting at the end of the pulse, taking as a baseline the cell resting membrane potential before stimulation. In order to test for afterpotential dependence on voltage we did additional experiments in which the holding current was adjusted in order to vary the membrane potential in a range between –59 and –84 mV.
, 百拇医药
Ih. Ih was induced by injection of hyperpolarizing voltage steps (10 mV increments, 200 ms duration) from –80 to –120 mV in cells clamped at –70 mV. The amplitude of the inward current generated was determined as the average of the current during the last 50 ms from each step (when the current was already stationary) with respect to the initial current.
Spontaneous firing pattern. Spontaneous firing activity was recorded for periods of 1 min both in whole-cell current clamp mode and in cell-attached configurations. Action potentials were automatically localized, and interspike intervals (ISIs) were calculated using an Igor (Wavemetrics, Lake Oswego, OR, USA) routine written in the laboratory by C. Pouzat. In order to determine the effects of pharmacological treatments as well as of prolonged periods in whole-cell recording, we calculated the CV for the ISI distributions and compared them before and after experimental manipulations using a paired Student's t test.
, 百拇医药
Statistical treatment
All results are expressed as mean ±S.E.M. Experiments involving a control and a test period were analysed using a paired Student's t test. In other cases an ANOVA comparison was used. Differences were considered significant at P < 0.05.
Results
Autoreceptor current
In voltage-clamped MLIs, application of a short depolarizing voltage pulse induces a sequence of a classical fast inward Na+ current followed by a slow inward current that is sensitive to bicuculline or gabazine (Pouzat & Marty, 1998, 1999). Previous work indicates that the latter current is generated by GABA which is released from the axon of the MLI under study, and which activates either somatodendritic receptors (autaptic current: Pouzat & Marty, 1998) or axonal receptors (autoreceptor current: Pouzat & Marty, 1999). It was shown that autaptic and autoreceptor currents can be unambiguously distinguished, and that the two forms of currents have opposite developmental profiles. The present results were obtained in an age range where autoreceptor currents largely predominate (PN 11–14). Furthermore, occasional cells with autapses were identified and removed from the analysis (see Methods). The tail currents in the remaining cells are autoreceptor currents. These currents were characterized earlier under symmetrical Cl– conditions at room temperature (Pouzat & Marty, 1999). In the present work we used a physiological intracellular Cl– concentration (15 mM, see Chavas & Marty, 2003) and a recording temperature of 34–36°C. Under these conditions, an inward autoreceptor current was clearly measurable at –90 mV, but it was considerably smaller than that previously reported, and it did not always display a clear temporal separation from the early inward Na+ current. The total duration of this current was about 50 ms. Its amplitude was measured after a 7 ms delay following the peak of the Na+ current surge, giving an average of 34 ± 5 pA. This value significantly decreased to 7 ± 2 pA (n= 7) (paired t test, P < 0.01) after bath application of either 10 μM bicuculline or 20 μM gabazine (not shown).
, http://www.100md.com
Autoreceptor-driven afterdepolarization
To study the effects of the autoreceptor current under current clamp conditions, the firing rate of MLIs was reduced by hyperpolarizing them slightly to a range between –60 and –80 mV. Action potentials were generated by applying depolarizing current steps (Fig. 1). The action potentials were followed by a depolarizing afterpotential (Fig. 1A), which occasionally induced firing (3 cells out of 31; example in Fig. 1B). The afterdepolarization was variable from cell to cell, both in amplitude and in time course. For the largest amplitude values, a clear peak was observed around 25–35 ms after the peak of the last spike, but in other cases, no clear peak could be defined. Under the conditions of our experiments, GABAergic synaptic currents decay with a mean weighted time constant of 6 ms (authors' unpublished results), shorter than the interval between spikes in the experiments of Fig. 1 (10–20 ms). Therefore, little overlap was expected between GABA-induced conductances produced by successive action potentials. Nonetheless, we studied the dependence of the afterdepolarization on the number of action potentials induced by the stimulation pulse. As shown in Fig. 1B, we found no significant correlation between the number of action potentials and the integral of the afterdepolarization (integrated over a 200 ms period following the end of the stimulus) (ANOVA F= 0.25; P= 0.86, compared for 1–4 action potentials; n= 27 cells). MLI–Purkinje cell synapses show little or no potentiation under resting conditions in juvenile preparations (Pouzat & Hestrin, 1997; Caillard et al. 2000), and MLI terminals exhibit little accumulation of presynaptic calcium for short trains (Collin et al. 2005). Therefore, little potentiation of GABA release was expected. Overall, the lack of summation of afterdepolarizations is consistent with previous results on this preparation.
, http://www.100md.com
A, an example of depolarizing afterpotential elicited by one action potential. A subthreshold response (dashed trace) is shown for comparison. B, the depolarizing afterpotential had a size which was independent of the number of action potentials during the stimulation pulse, and occasionally elicited delayed firing. Left: superimposed current responses to depolarizing steps that were adjusted to produce one (black trace) or 4 (grey traces) action potentials. Afterdepolarization (ADP) size was quantified as the area under the voltage profile during a 200 ms long time window starting at the end of the stimulation pulse, taking the initial membrane potential as a baseline. Individual traces show no apparent change in afterpotential amplitude when inducing either one or four action potentials. As it can be seen, the afterdepolarization sometimes induced firing. Right: mean (±S.E.M.) afterdepolarization integral for current pulses inducing 1–4 action potentials (ANOVA F= 0.2516; P= 0.86, 27 cells). C, the afterdepolarization was sensitive to GABAAR blockers. Left: current clamp recordings from a cell held at –80 mV in control and after applying 10 μM bicuculline for 3 min in the bath solution. Right: afterdepolarization integral for 10 cells recorded in control and after 3 min in GABAAR blockers, showing a significant reduction (paired t test, P < 0.01) after the pharmacological treatment. D, the afterdepolarization was sensitive to ‘rundown’. Left: current clamp recordings from a cell held at –70 mV, 5 min after establishment of whole-cell recording (Control) and after > 20 min of whole-cell recording (Rundown). Right: mean afterdepolarization integral was significantly diminished after periods of between 20 and 38 min in whole-cell recording compared with the first 5 min (n= 7 cells).
, 百拇医药
To test whether the afterdepolarization recorded in current clamp was related to the activity of the autoreceptor currents, we studied the effects of experimental manipulations blocking these currents. For these comparisons, we pooled together results obtained with one to four action potentials, since the number of action potentials did not appear to significantly affect the size of the afterdepolarization. As shown in Fig. 1C, the amplitude of the afterdepolarization was always drastically reduced after 3 min application of a GABAAR blocker (either 10 μM bicuculline (n= 5), 20 μM gabazine (n= 4) or 200 μM picrotoxin (n= 1)). Pooling the results from these 10 cells together, the afterdepolarization integral fell from 438 ± 82 mV x ms to 91 ± 23 mV x ms (paired t test, P < 0.01). These results show that the major part of the afterdepolarization arises from the activation of GABAARs.
, 百拇医药
The size of the afterdepolarization was surprisingly large. In the above series, the afterdepolarization amplitude ranged from 3 to 21 mV, with a mean value of 10.6 ± 1.6 mV. In 6/10 cells, the peak depolarization was less negative than ECl (–58 mV), reaching up to –45 mV. The difference cannot be explained on the basis of the contribution of bicarbonate ions, which do not alter appreciably the reversal potential value of GABAARs in our whole-cell recording conditions (no bicarbonate inside: see Chavas & Marty, 2003). Therefore, the results suggest that, even though an activation of GABAARs underlies the afterdepolarization, other mechanisms are also involved. In light of the earlier results by Mann-Metzer & Yarom (2002) it seems likely that a non-inactivating Na+ current amplifies the afterdepolarization and, in about half of the cells, brings it above the value of ECl.
, http://www.100md.com
In whole-cell recording, autoreceptor currents undergo an irreversible decrease which is presumably due to the diffusion of one or several substances out of the cell (‘rundown’: Pouzat & Marty, 1999). We included GABA in the recording pipette to slow down this process, but even under these conditions, rundown of autoreceptor currents (as observed in voltage clamp) was severe or complete after > 20 min of whole-cell recording. Likewise, afterdepolarization diminished from 292 ± 89 mV x ms, when measured in the first 5 min after establishment of whole-cell recording, to –85 ± 100 mV x ms when measured after more than 20 min in whole-cell recording (paired t test, P < 0.02; 7 cells) (Fig. 1D). Therefore, rundown was observed in the same time domain for the afterdepolarization and for the autoreceptor current. Taken together, the results led us to conclude that the afterdepolarization was primarily due to the activation of GABAA autoreceptors following GABA release from the axon.
, 百拇医药
Lack of sensitivity of afterdepolarization on membrane potential
We next investigated the sensitivity of the afterdepolarization to the cell membrane potential. In isotonic intracellular Cl– condition (ECl= 0 mV), the peak amplitude of autoreceptor currents increases linearly with potential, as predicted for a conductance which is primarily carried by Cl– ions (Pouzat & Marty, 1999). We expected that with 15 mM intracellular Cl– concentration, the afterpotential amplitude would be modulated in the same way when the initial membrane potential would be moved away from ECl (–58 mV). Contrary to this expectation we found no statistical difference between afterdepolarization integrals measured at membrane potentials ranging from –59 to –84 mV (binned in 5 mV groups; ANOVA F= 1.405, P= 0.25; Fig. 2A). Nonetheless, since the afterpotential is mainly driven by the autoreceptor-dependent conductance, we expected it to be absent or reversed when the cell potential was near or above ECl. We therefore analysed single action potentials taken from segments of spontaneous activity recordings in which the membrane potential before the spike was near or above ECl. In this case, as shown in Fig. 2B, we found negative values for the afterpotential, averaging –753 ± 148 mV x ms (cell potential 200 ms before the spike: –47 to –50 mV; 6 cells). These results are consistent with the notion that the afterpotential is mediated primarily by a Cl–-selective conductance.
, http://www.100md.com
A, afterdepolarization (ADP) is not strongly voltage dependent in the hyperpolarizing range. Pooled results showing the afterdepolarization integral as a function of membrane potential, from a total of 15 cells. Afterdepolarization was measured at membrane potentials from –59 to –84 mV; results were binned in groups of 5 mV for analysis. Note the non-significant reduction of the integral for the most depolarized as well as the most hyperpolarized potentials (ANOVA F= 1.26; P= 0.31). B, lack of afterdepolarization in cells held above ECl. Example of time-averaged spontaneous action potentials, in a depolarized cell (membrane potential: –48 mV).
, http://www.100md.com
Interaction between autoreceptor current and Ih
A likely explanation for the anomalous evolution of the afterdepolarization with hyperpolarization was that the increase in driving force was counteracted by a shunt brought about by a hyperpolarization-driven conductance. Since MLIs exhibit a large Ih current which turns on around resting membrane potential (Midtgaard, 1992a), we decided to study the interaction of autoreceptor currents with Ih. In our voltage clamp recordings Ih was seen as an inwardly rectifying, slowly activating current, which was readily apparent for hyperpolarizing steps. In the presence of 10 μM ZD7288, Ih relaxation amplitudes in response to steps to –80 and –120 mV diminished from 10.7 ± 2.1 to 4.6 ± 1.0 pA and from 76.1 ± 5.1 to 23.3 ± 4.4 pA (n= 5; Fig. 3A). To test whether Ih contributes to the cell conductance near resting potential we compared cell resistance as measured at –70 mV before and with ZD7288, and found no difference (1.55 ± 0.40 and 1.73 ± 0.37 G, respectively).
, 百拇医药
A, consecutive 10 mV hyperpolarizing pulses (test potential: –80 to –120 mV) were applied in a MLI voltage clamped at –70 mV (first trace without hyperpolarizing pulse). In control conditions, hyperpolarization induced the activation of an inward Ih current characterized by its slow onset kinetics. This current was blocked by 10 μM ZD7288. B, in voltage clamp, Ih blockage significantly increased the size of autoreceptor currents. Upper panel: individual traces in a cell voltage clamped at –90 mV and shortly depolarized to 0 mV in control conditions (black trace) and in ZD7288 (grey trace). Lower panel: autoreceptor currents in control and in ZD7288 (n= 7) (paired t test, P < 0.02). C, left: individual current clamp traces in a cell held at –70 mV, showing an increase in afterdepolarization size after Ih blockade with ZD7288; this increase was accompanied by an increased proportion of cells spiking during afterdepolarization. Both traces were recorded from the same cell in exactly the same conditions. Note the increased cell excitability during the depolarizing pulse, which had to be reduced in ZD7288 compared with its control amplitude in order to limit firing to one spike. Right: data from 6 cells, showing a significant increase in the afterdepolarization integral upon ZD7288 application (paired t test, P < 0.05; black and grey traces as in B here and in following figures).
, http://www.100md.com
We next tested the effects of 10 μM ZD7288 on the amplitude of autoreceptor currents as measured in voltage clamp, as well as on that of the afterdepolarization associated with autoreceptor currents. We found that autoreceptor currents increased by 45%, from 31 ± 11 to 45 ± 10 pA (paired t test, P < 0.05; n= 7) (Fig. 3B). The most likely interpretation for this finding is that suppression of Ih increases the space constant of voltage control in the axon, leading to a recording of the autoreceptor current on a larger area in ZD7288 than in control conditions. Such a change in the axon space constant would not necessarily be apparent as a change in resistance as recorded in the somatic compartment, for two reasons. First, the increase in the space constant counteracts the effects of a reduction in specific membrane conductance on the input resistance. Second, the resting conductance of the somatodendritic domain may be little affected by ZD7288; since this conductance is in parallel with that of the axon, its contribution could occlude any change in axon input resistance. The hypothesis of a change in the axonal space constant is supported by the fact that Ih has been preferentially found in the axonal domain of MLIs (Santoro et al. 1997; Southan et al. 2000). In current clamp, ZD7288 likewise increased the afterdepolarization produced by a train of action potentials, tested for membrane potentials between –70 and –72 mV, from 345 ± 127 mV x ms to 1097 ± 303 mV x ms (paired t test, P < 0.05) (Fig. 3C, ZD7288; n= 6). Moreover the blockade of Ih increased the proportion of cells in which firing was observed during the afterdepolarization (5/7 cells; Fig. 3C) compared with control conditions where only 3/31 cells were spiking during the afterdepolarization. Finally, it was noted in these experiments that the amplitude of the injected current needed to be reduced in ZD7288 compared with the control situation in order to match the number of spikes during the depolarizing step, as exemplified in Fig. 3C. This effect, like those on the autoreceptor current and its associated depolarization, may be explained by an increase in the length constant of the axon in the presence of ZD7288, which would account for a higher efficacy of somatic depolarization in triggering spikes at their site of initiation. Taken together, the results reveal a negative interaction between autoreceptor current and Ih.
, 百拇医药
Influence of autoreceptors on firing pattern
In order to determine whether autoreceptor currents could influence the firing pattern of MLIs, we compared ISIs in control conditions and after experimental manipulations affecting this current. We analysed the spontaneous firing activity both in whole-cell and in cell-attached recordings. In whole-cell mode, the membrane potential was now set at –65 ± 5 mV, close to the normal resting values (Chavas & Marty, 2003). The average spontaneous firing frequency in this condition was 3.38 ± 0.85 Hz (n= 6). In our cell-attached experiments at physiological temperature the average frequency was 2.46 ± 0.52 Hz (n= 5), similar to the frequency in our whole-cell experiments and within the range reported in cell-attached conditions at room temperature in this age range (Llano & Gerschenfeld, 1993a; Vincent & Marty, 1996; Chavas & Marty, 2003).
, 百拇医药
In control conditions, action potential firing appeared highly irregular, with isolated spikes and short bursts of up to five spikes alternating with silent periods in the order of one or a few seconds in duration. ISI histograms displayed a peak near 0.1–0.2 s, followed by a tail extending to the seconds range. As shown in Fig. 4A and B, application of GABAAR blockers shifted the firing pattern to a more regular one. ISI histograms became more compact; their peak was shifted to higher values, and the slow tail became less conspicuous. To quantify the change of shape of ISI histograms, we analysed the CV values of ISIs, reasoning that the more compact histogram shape associated with the suppression of autoreceptor currents should decrease this parameter. Indeed we found that CV values were significantly diminished, from 2.56 ± 0.57 in control conditions to 0.97 ± 0.08 after 3–5 min application of GABAAR blockers in whole-cell recording configuration (paired t test, P < 0.02, n= 6) (Fig. 4A1), and from 1.30 ± 0.12 to 0.94 ± 0.04 in cell-attached configuration (paired t test, P < 0.05, n= 5) (Fig. 4B1). On the other hand, GABAAR blockers did not change the mean ISI significantly (whole-cell recording experiments: 0.82 ± 0.47 s in blockers versus 0.35 ± 0.08 s in control; paired t test: P= 0.39; n= 6; cell-attached experiments: 0.42 ± 0.13 s in blockers versus 0.51 ± 0.10 s in control; paired t test: P= 0.20; n= 5). Thus, the change in CV cannot be ascribed to a general increase or decrease in spiking.
, 百拇医药
A, upper traces: whole-cell recordings of spontaneous activity in current clamp mode in a cell held around –70 mV before and after bath application of bicuculline. The cell fired groups of 1–5 action potentials in control conditions, but displayed a more regular discharge after blocking GABAARs. Lower panels: this change was manifested in the distribution of the ISI histogram, which became more homogeneous, while its peak was shifted to the right. A1, CV values of ISIs diminished significantly in GABAAR blockers compared with control values (paired t test, P < 0.02, n= 6). A2, cumulative normalized ISI histograms for times shorter than 400 ms (same experiment as traces in A) show that the main changes occurred in this period as expected considering the duration of the afterdepolarization. B, spontaneous activity recorded in cell-attached mode before and after application of gabazine (upper traces). As in whole-cell recording (A), blocking GABAARs changed the firing pattern to a more regular one. Again, the ISI histogram became more homogeneous and its peak was displaced to the right (lower panels). B1, changes in the distribution of ISIs were reflected in a significant decrease in the CV values of ISIs (paired t test, P < 0.05, n= 5). B2, cumulative histograms for the first ISI bins showing a diminution in the number of short ISIs. C, spontaneous activity in current clamp under control conditions and after 10 μM ZD7288. Blockade of Ih favoured a ‘bursty-like’ pattern of discharge. In the example shown, the peak of the distribution became more prominent and was shifted to the left, while the number of ISIs larger than 1 s increased. C1, CV values of ISIs increased significantly after 10 μM ZD7288 (paired t test, P < 0.05, n= 6). C2, cumulative histograms showing an increase in the number of ISIs shorter than 400 ms after ZD7288. D, spontaneous activity of a MLI held near –70 mV, recorded in current clamp mode during the first 5 min of whole-cell recording (Control; left), and after more than 20 min of whole-cell recording (Rundown; right). The firing exhibited short bursts initially, and changed to a more regular discharge later. Accordingly, the ISI histograms became more homogeneously distributed, and the position of its peak shifted to the right (lower panels). D1, CV values for ISIs are significantly diminished after > 20 min of whole-cell recording (paired t test, P < 0.05, n= 6). D2, initial portion of the normalized cumulative ISI histograms from the same experiment, highlighting the initial surge in control (continuous line) contrasting with the delayed onset after rundown (dotted line).
, http://www.100md.com
According to our results, the GABAAR-driven afterdepolarization lasts roughly 150 ms after the spike (see Figs 1 and 3). On this basis the most striking effect of autoreceptors on the ISI histogram is expected during this time period (see Fig. 4A2 and B2; normalized cumulative histograms for the first bins in the distribution). Therefore we evaluated the percentage of spikes falling in the first three bins of our histograms (corresponding to 150 ms after spiking) before and after blocking the GABAARs. The percentage values diminished from 73% to 27% and from 35% to 24% for whole-cell (n= 6) and cell-attached (n= 5) experiments, respectively.
, http://www.100md.com
Conversely, since Ih blockade increases the amplitude of the autoreceptor current and of its associated afterdepolarization, we asked whether pharmacological blockade of Ih would favour the bursting of MLIs. In this experimental group, the mean ISI value was 0.64 ± 0.12 s in control conditions, and remained unchanged after 10 μM ZD7288 (0.49 ± 0.06 s, paired t test, P= 0.09, n= 6). Meanwhile, contrary to the results of previous manipulations, the blockage of Ih changed the firing pattern to a more ‘bursty’ one (Fig. 4C). This was not accompanied by any systematic change in the position of the ISI histogram peak (paired t test, P= 0.58). Reflecting the change in the shape of ISI distribution, CV values increased significantly after ZD7288 application, from 1.73 ± 0.16 to 2.61 ± 0.54 (paired t test, P < 0.05; Fig. 4C1). Thus, these results, like those illustrated in Fig. 3, indicate an antagonistic influence of Ih over autoreceptor current. They are also consistent with a previous finding that Cs+, a blocker of Ih, favours bursting of MLIs (Mann-Metzer & Yarom, 2002). Finally, ZD7288 was found to enhance the first bins of the cumulative ISI histogram (Fig. 4C2), contrary to the decrease observed by blocking the autoreceptor current.
, 百拇医药
Even though an implication of autoreceptors is an attractive explanation for the effects illustrated in Fig. 4A–C, other interpretations need also to be considered. While autoreceptors are axonal, GABAARs of the somatodendritic compartment are engaged in MLI–MLI synapses (Kondo & Marty, 1998) and in autapses (Pouzat & Marty, 1998). In addition it has been suggested that, in slices from adult rats, ambient GABA activates extrasynaptic receptors located in the somatodendritic domain of MLIs (Husser & Clark, 1997). A priori GABAAR blockers could modify the firing pattern by acting on any of these four groups of GABAARs, or on a combination of them. Autapses can be dismissed because their occurrence at the age considered here (PN 11–14) is very low (Pouzat & Marty, 1999) and because in whole-cell recording experiments, occasional cells containing an autapse were unambiguously identified (see Methods) and rejected. A contribution of extrasynaptic somatodendritic GABAARs is also unlikely since we found no change in mean firing frequency upon application of bicuculline or gabazine. This leaves us with classical synapses and with autoreceptors. To distinguish between these possibilities we made use of the fact that autoreceptor currents run down in whole-cell recording (Pouzat & Marty, 1999), whereas currents due to MLI–MLI synapses are stable for > 30 min of whole-cell recording (Llano & Gerschenfeld, 1993a). In our recording conditions we found that the rundown of autoreceptor currents was nearly complete in 20 min; therefore, we compared the spontaneous firing pattern at the beginning (first 5 min) and after at least 20 min of whole-cell recording (Fig. 4D). In these experiments we observed a shift in the peak of the ISI distribution from values under 0.1 s to values around 0.3 s (histograms in Fig. 4D) without significant changes in the mean ISI value (1.07 ± 0.36 and 1.10 ± 0.29, paired t test, P= 0.92, n= 6). As observed after application of GABAAR blockers, the percentage of ISIs smaller than 150 ms was decreased from 54 to 16% after periods longer than 20 min in whole-cell recordings (Fig. 4D2). Finally, and most importantly, statistical comparison of the CV of ISIs showed a significant decrease after rundown of autoreceptor currents from a value of 2.96 ± 0.49 to 1.07 ± 0.22 (paired t test, P < 0.02, n= 6) (Fig. 4D1).
, http://www.100md.com
In summary, the results from Fig. 4A–C indicate that activation of GABAARs favours burst firing. In addition, the experiments shown in Fig. 4D led us to conclude that axonal autoreceptors, rather than somatodendritic receptors, are primarily responsible for the modulation of MLI firing following application of GABAAR blockers.
Autoreceptor-driven afterdepolarization and bursting behaviour are developmentally regulated
All results presented this far have been obtained from animals aged 11–14 days. However, autoreceptor currents are developmentally regulated (Pouzat & Marty, 1999), thus raising the issue of how much the conclusions depend on age. Within the PN 11–14 age group, we found a 2-fold decrease in the mean afterdepolarization amplitude. However, since this decrease was not significant (ANOVA P= 0.24), the results for this age group have been pooled in the presentation of the results so far. This procedure seems justified because the age distribution of the cells used for each series of experiments was similar, and because in addition, our conclusions rest on the comparison between a control and a test period, using paired t tests.
, http://www.100md.com
To explore further the age dependence, we examined the afterdepolarization for two additional age groups: PN 8–10 and PN 17. We found that the afterdepolarization integral decreased from 1269 ± 316 mV x ms (n= 5) at PN 8–10 to 467 ± 89 mV x ms in the PN 11–14 group (n= 29) and vanished completely at PN 17 to a value of –47 ± 64 mV x ms (n= 8) (ANOVA P= 1.53 x 10–5) (Fig. 5A). We also examined the firing pattern in the three groups. In the PN 8–10 group, the application of gabazine induced a diminution in CV (from 1.70 ± 0.12 to 1.29 ± 0.12; paired t test, P < 0.05; n= 5), similar to the one obtained for PN 11–14. This diminution was not accompanied by a change in the mean ISI value in the absence and the presence of the drug (0.88 ± 0.17 and 1.27 ± 0.53 s, respectively; paired t test, P= 0.42; n= 5).
, 百拇医药
A, autoreceptor-driven afterdepolarization as a function of postnatal age. The afterdepolarization decreased in amplitude from the PN 8–10 group to the PN 11–14 group, and completely disappeared at PN 17. B, spontaneous activity recorded in current clamp mode in control conditions was unaltered after bath application of 20 μM gabazine in a PN 17 MLI. ISI histograms (lower panels) showed no change upon GABAAR blockade. C, initial portion of the cumulative ISI histogram, showing no difference between control and gabazine data for the cell in B. D, the CV values for ISIs was insensitive to GABAAR blockers (paired t test, P= 0.11; n= 6).
, 百拇医药
Likewise, in view of the absence of GABAAR-driven afterdepolarization at PN 17, we investigated at this age the effects of GABAAR blockers on the spontaneous firing pattern. In control conditions, the cells showed a more regular firing pattern compared with 11- to 14-day-old animals (Fig. 5B), which was reflected in a significant diminution of the CV value in the older group (P < 0.001), when compared with PN 11–14. This change in CV value was not accompanied by any modification of the average spiking frequency: the mean ISI value was 0.43 ± 0.09 s in control conditions, a value not significantly different from that obtained in the 11- to 14-day-old group (P= 0.15). The application of GABAAR blockers at PN 17 did not produce any significant change in the CV value, with an average of 1.32 ± 0.12 versus 1.06 ± 0.07 in control (paired t test, P= 0.11; n= 5) (Fig. 5D). Likewise the mean ISI value was similar in the presence of GABAAR blockers (0.33 ± 0.05 s) to the control value (see above; paired t test, P= 0.12; n= 6). The average position of the ISI histogram peak was smaller than 0.15 s both in control and in the presence of GABAAR blockers (Fig. 5B). Taken together, our results indicate that the presence of an autoreceptor-mediated afterdepolarization is responsible for the differences in firing pattern between PN 11–14 and PN 17. If this is correct, blocking the afterdepolarization at PN 11–14 should make the firing pattern similar to the one at PN 17. This was confirmed by the lack of difference between CV values at PN 17 in control conditions, compared with those obtained at PN 11–14 during application of GABAAR blockers or after prolonged periods of whole-cell recording (ANOVA P= 0.77).
, http://www.100md.com
Discussion
Our study indicates that a GABAAR-driven, developmentally regulated autoreceptor current shapes the firing pattern of MLIs during the second postnatal week. We also show that an Ih current, which is prominent in these cells (Midtgaard, 1992a), reduces the autoreceptor current, and counteracts its effects on MLI firing.
There are several possibilities to explain that GABAAR block can induce a shift from a ‘bursty-like’ firing pattern to a regular one. Of these, (1) the modulation of a constitutively active, GABAAR-mediated extrasynaptic conductance (Husser & Clark, 1997) is unlikely since the alteration in firing pattern induced by GABAAR blockers did not involve a change in the mean firing frequency, as would be expected from a change in tonic GABAA conductance. (2) MLI–MLI synapses could favour bursting, for instance in the case of two MLIs which would be reciprocally connected via excitatory synapses. However MLI–MLI synaptic currents are not affected by prolonged whole-cell recording (Llano & Gerschenfeld, 1993a), whereas the bursting pattern was sensitive to rundown. Furthermore at PN 17, whereas MLI–MLI synaptic currents are strong (Kondo & Marty, 1998), bursting was not apparent before the application of GABAAR blockers, and application of the blockers failed to decrease the CV of the ISIs. (3) The modulation of an autaptic rather than autoreceptor GABAAR-mediated current can be excluded since the proportion of cells presenting autaptic currents is very low at the age under study (about 5%; Pouzat & Marty, 1998, 1999), and since in the present work, the single cell which exhibited characteristic features of autaptic currents, i.e. high amplitude variability and failures, was rejected from the analysis.
, 百拇医药
On the other hand, the proportion of cells showing autoreceptor current at PN 11–14 is about 92% (Pouzat & Marty, 1999), and this current is sensitive to both rundown and GABAAR blockers. Moreover it is developmentally regulated and we found a diminution of both autoreceptor currents and its associated depolarization at PN 17, coincident with the appearance of a regular firing pattern (Fig. 5). The latter result is in accordance with previous results obtained in older animals (Husser & Clark, 1997; Carter & Regehr, 2002). Overall, the results strongly suggest a tight link between bursting and expression of autoreceptors.
, 百拇医药
Previous studies have addressed the role of axonal or presynaptic GABAARs in regulating cell firing. In primary spinal cord afferents (Rudomin et al. 1981; Verdier et al. 2003), in presynaptic terminals of the pituitary gland (Zhang & Jackson, 1995), and in hippocampal mossy fibres (Ruiz et al. 2003), activation of axonal GABAARs depolarizes the axonal membrane, yet inhibits action potential firing. The proposed explanation for this paradoxical inhibition is a depolarization-driven inactivation of Na+ channels (Zhang & Jackson, 1995). In our experiments, by contrast, we found that the autoreceptor current induces a depolarizing afterpotential and produces a net increase in firing probability during the first 150 ms afterspike (as suggested by the high density of ISIs within this range in control conditions), thus shaping firing in a ‘bursty-like’ manner. Consistent with an increase in firing probability, we found a decrease in the current amplitude required to produce a spike during the afterdepolarization (J. Chavas & A. Marty, unpublished observations). The key difference between our work and the previous publications presumably lies in the mode of activation of the GABAARs, which is phasic in the present work, and was tonic in the previous studies. As the rate of Na+ channel inactivation is slower than that of activation, a transient depolarization is more likely to be excitatory than a sustained one.
, 百拇医药
Understanding the exact reason why activation of the autoreceptor current changes firing probability after spikes would require a detailed knowledge of voltage-dependent conductances in MLI axons, which is presently lacking. Previous work has shown that the reversal potential of GABA-sensitive currents, EGABA, is rather depolarized in MLIs (–58 mV), and as a result of this situation, activation of MLI–MLI synapses at membrane potentials more negative than –58 mV leads to postsynaptic depolarization and to an increase in the firing probability (Chavas & Marty, 2003). The depolarized position of EGABA likewise accounts for the excitatory effect of the autoreceptors as revealed in our whole-cell experiments when cells were held at resting membrane potentials more negative than –58 mV. In contrast with our results, earlier work on GABAAR-mediated autapses in the hippocampus and neocortex revealed inhibitory effects, presumably because the values of EGABA in the concerned interneurone subtypes were more hyperpolarized than in the present preparation (Bacci et al. 2003; Pawelzik et al. 2003).
, 百拇医药
In spite of the relatively high level of EGABA in MLIs, Chavas & Marty (2003) showed in cell-attached experiments different populations of neurones in which the stimulation of a presynaptic cell can induce either increases or decreases on the firing rate of MLIs. This contrasts with the results of the present work, which indicate a consistent excitatory effect of GABAAR activation on cell firing. To explain the discrepancy, two explanations can be offered. One would be that EGABA is more depolarized in the axon than in the somatodendritic domain. The other would be that the axonal GABAARs responsible for the autoreceptor currents are associated more directly with depolarization-activated conductances such as non-inactivating Na+ channels (Mann-Metzer & Yarom, 2002). Both hypotheses are compatible with the finding that, in about half of the whole-cell recording experiments, the peak of the autoreceptor-associated depolarization was more depolarized than EGABA. Further work will be needed to decide between these two possibilities.
, 百拇医药
In addition to non-inactivating Na+ channels, other voltage-dependent conductances are likely to contribute together with the autoreceptor current to shape the firing pattern of MLIs. We observed a negative interaction between autoreceptor currents and Ih, possibly mediated by a change in the axonal space constant induced by the blockage of Ih channels, which are primarily localized in the axon (Santoro et al. 1997; Southan et al. 2000). Pharmacological blockade of Ih increased the size of autoreceptor-mediated afterspike depolarization and the probability of afterspike firing. Moreover, Ih removal by ZD7288 induced a change in the spontaneous firing pattern whose direction was contrary to the one induced by the blockage of autoreceptor currents, namely a shift to a more ‘bursty’ pattern. These results are in agreement with the ones by Chan et al. (2004) in GABAergic globus pallidus neurones, who reported a diminution in the regularity of spontaneous cell firing after pharmacological blockade of Ih. All of the above results fit with a scheme where the effects of the autoreceptors are negatively tuned by Ih, possibly through a change in axonal excitability. This raises the possibility that a functional modulation of Ih could alter the autoreceptor current and the bursting behaviour of MLIs. Accordingly, Chan et al. (2004) suggested a regulation of the activity of GABAergic neurones in the globus pallidus by modulation of an Ih conductance. Our results with Ih current could alternatively be explained through an effect on neurotransmitter release as the one reported previously by Beaumont & Zucker (2000) and Chevaleyre & Castillo (2002). However, in those systems Ih was promoting an increase in neurotransmitter release whereas in our preparation Ih diminished the autoreceptor currents. Furthermore, effects of Ih blockade on transmitter release develop after a delay of tens of minutes (Chevaleyre & Castillo, 2002), whereas those of ZD7288 in the present work do not show such a delay. Finally, a recent report (Saitow et al. 2005) showed that noradrenalin effects on MLI neurotransmitter release (Llano & Gerschenfeld, 1993b) are independent of Ih.
, 百拇医药
Finally, it is worth stressing that according to the present results, autoreceptors shape the firing pattern precisely during the period of development where MLI synaptic connections are under maturation (Pouzat & Hestrin, 1997). A similar transient bursting behaviour has been observed in many CNS regions, including developing retinal ganglion and hippocampal pyramidal cells (Costa et al. 1991; Liets et al. 2003; Chen et al. 2005). In the case of pyramidal cells, bursting behaviour was promoted by an afterspike depolarization, which was, however, not mediated by GABAARs. Spontaneous bursting activity is generally assumed to promote neurite elongation and synapse formation (e.g. in the visual system: Shatz, 1996; and in the spinal cord: Ren & Greer, 2003). How this could come about at GABAergic synapses is still unclear. In the present system, it was recently shown that repetitive activation of MLI synapses induces transient increases in Ca2+ concentration in the postsynaptic cells (Chavas et al. 2004). Such postsynaptic signals could modulate the amplitude of GABAergic events (as shown in developing hippocampal synapses by Woodin et al. 2003) and reinforce the strength of synaptic contacts.
, 百拇医药
References
Axmacher N & Draguhn A (2004). Inhibition of GABA release by presynaptic ionotropic GABA receptors in hippocampal CA3. Neuroreport 15, 329–334.
Bacci A, Huguenard JR & Prince DA (2003). Functional autaptic neurotransmission in fast-spiking interneurons: a novel form of feedback inhibition in the neocortex. J Neurosci 23, 859–866.
Beaumont V & Zucker RS (2000). Enhancement of synaptic transmission by cyclic AMP modulation of presynaptic Ih channels. Nat Neurosci 3, 133–141.
, http://www.100md.com
Belenky MA, Sagiv N, Fritschy JM & Yarom Y (2003). Presynaptic and postsynaptic GABAA receptors in rat suprachiasmatic nucleus. Neuroscience 118, 909–923.
Caillard O, Moreno H, Schwaller B, Llano I, Celio MR & Marty A (2000). Role of the calcium-binding protein parvalbumin in short-term synaptic plasticity. Proc Natl Acad Sci U S A 97, 13372–13377.
Callaway JC, Lasser-Ross N & Ross WN (1995). IPSPs strongly inhibit climbing fiber-activated [Ca2+]i increases in the dendrites of cerebellar Purkinje neurons. J Neurosci 15, 2777–2787.
, http://www.100md.com
Carter AG & Regehr WG (2002). Quantal events shape cerebellar interneuron firing. Nat Neurosci 5, 1309–1318.
Chan CS, Shigemoto R, Mercer JN & Surmeier DJ (2004). HCN2 and HCN1 channels govern the regularity of autonomous pacemaking and synaptic resetting in globus pallidus neurons. J Neurosci 24, 9921–9932.
Chavas J, Forero ME, Collin T, Llano I & Marty A (2004). Osmotic tension as a possible link between GABAA receptor activation and intracellular calcium elevation. Neuron 44, 701–713.
, 百拇医药
Chavas J & Marty A (2003). Coexistence of excitatory and inhibitory GABA synapses in the cerebellar interneuron network. J Neurosci 23, 2019–2031.
Chen S, Yue C & Yaari Y (2005). A transitional period of Ca2+-dependent bursting triggered by spike backpropagation into apical dendrites in developing rat CA1 neurons. J Physiol 567, 79–93.
Chevaleyre V & Castillo PE (2002). Assessing the role of Ih channels in synaptic transmission and mossy fiber LTP. Proc Natl Acad Sci U S A 99, 9538–9543.
, http://www.100md.com
Collin T, Chat M, Lucas MG, Moreno H, Racay P, Schwaller B, Marty A & Llano I (2005). Developmental changes in parvalbumin regulate presynaptic Ca2+ signaling. J Neurosci 25, 96–107.
Costa PF, Ribeiro MA & Santos AI (1991). Afterpotential characteristics and firing patterns in maturing rat hippocampal CA1 neurones in vitro slices. Brain Res Dev Brain Res 62, 263–272.
Husser M & Clark BA (1997). Tonic synaptic inhibition modulates neuronal output pattern and spatiotemporal synaptic integration. Neuron 19, 665–678.
, 百拇医药
Jackson MB & Zhang SJ (1995). Action potential propagation and propagation block by GABA in rat posterior pituitary nerve terminals. J Physiol 483, 597–611.
Jorntell H & Ekerot CF (2003). Receptive field plasticity profoundly alters the cutaneous parallel fiber synaptic input to cerebellar interneurons in vivo. J Neurosci 23, 9620–9631.
Kawa K (2003). Glycine facilitates transmitter release at developing synapses: a patch clamp study from Purkinje neurons of the newborn rat. Brain Res Dev Brain Res 144, 57–71.
, 百拇医药
Kondo S & Marty A (1998). Synaptic currents at individual connections among stellate cells in rat cerebellar slices. J Physiol 509, 221–232.
Liets LC, Olshausen BA, Wang GY & Chalupa LM (2003). Spontaneous activity of morphologically identified ganglion cells in the developing ferret retina. J Neurosci 23, 7343–7350.
Llano I & Gerschenfeld HM (1993a). Inhibitory synaptic currents in stellate cells of rat cerebellar slices. J Physiol 468, 177–200.
, 百拇医药
Llano I & Gerschenfeld HM (1993b). Beta-adrenergic enhancement of inhibitory synaptic activity in rat cerebellar stellate and Purkinje cells. J Physiol 468, 201–224.
Mann-Metzer P & Yarom Y (2002). Jittery trains induced by synaptic-like currents in cerebellar inhibitory interneurons. J Neurophysiol 87, 149–156.
Midtgaard J (1992a). Stellate cell inhibition of Purkinje cells in the turtle cerebellum in vitro. J Physiol 457, 355–367.
, 百拇医药
Midtgaard J (1992b). Membrane properties and synaptic responses of Golgi cells and stellate cells in the turtle cerebellum in vitro. J Physiol 457, 329–354.
Pawelzik H, Hughes DI & Thomson AM (2003). Modulation of inhibitory autapses and synapses on rat CA1 interneurones by GABAA receptor ligands. J Physiol 546, 701–716.
Pouzat C & Hestrin S (1997). Developmental regulation of basket/stellate cellPurkinje cell synapses in the cerebellum. J Neurosci 17, 9104–9112.
, 百拇医药
Pouzat C & Marty A (1998). Autaptic inhibitory currents recorded from interneurones in rat cerebellar slices. J Physiol 509, 777–783.
Pouzat C & Marty A (1999). Somatic recording of GABAergic autoreceptor current in cerebellar stellate and basket cells. J Neurosci 19, 1675–1690.
Ren J & Greer JJ (2003). Ontogeny of rhythmic motor patterns generated in the embryonic rat spinal cord. J Neurophysiol 89, 1187–1195.
, 百拇医药
Rudomin P, Engberg I & Jimenez I (1981). Mechanisms involved in presynaptic depolarization of group I and rubrospinal fibers in cat spinal cord. J Neurophysiol 46, 532–548.
Ruiz A, Fabian-Fine R, Scott R, Walker MC, Rusakov DA & Kullmann DM (2003). GABAA receptors at hippocampal mossy fibers. Neuron 39, 961–973.
Saitow F, Suzuki H & Konishi S (2005). -Adrenoceptor-mediated long-term up-regulation of the release machinery at rat cerebellar GABAergic synapses. J Physiol 565, 487–502.
, 百拇医药
Santoro B, Grant SG, Bartsch D & Kandel ER (1997). Interactive cloning with the SH3 domain of N-src identifies a new brain specific ion channel protein, with homology to eag and cyclic nucleotide-gated channels. Proc Natl Acad Sci U S A 94, 14815–14820.
Shatz CJ (1996). Emergence of order in visual system development. Proc Natl Acad Sci U S A 93, 62–608.
Smith TC & Jahr CE (2002). Self-inhibition of olfactory bulb neurons. Nat Neurosci 5, 760–766.
, 百拇医药
Southan AP, Morris NP, Stephens GJ & Robertson B (2000). Hyperpolarization-activated currents in presynaptic terminals of mouse cerebellar basket cells. J Physiol 526, 91–97.
Tan YP & Llano I (1999). Modulation by K+ channels of action potential-evoked intracellular Ca2+ concentration rises in rat cerebellar basket cell axons. J Physiol 520, 65–78.
Turecek R & Trussell LO (2001). Presynaptic glycine receptors enhance transmitter release at a mammalian central synapse. Nature 411, 587–590.
, http://www.100md.com
Turecek R & Trussell LO (2002). Reciprocal developmental regulation of presynaptic ionotropic receptors. Proc Natl Acad Sci U S A 99, 13884–13889.
Verdier D, Lund JP & Kolta A (2003). GABAergic control of action potential propagation along axonal branches of mammalian sensory neurons. J Neurosci 23, 2002–2007.
Vincent P, Armstrong CM & Marty A (1992). Inhibitory synaptic currents in rat cerebellar Purkinje cells: modulation by postsynaptic depolarization. J Physiol 456, 453–471.
, 百拇医药
Vincent P & Marty A (1996). Fluctuations of inhibitory postsynaptic currents in Purkinje cells from rat cerebellar slices. J Physiol 494, 183–199.
Woodin MA, Ganguly K & Poo MM (2003). Coincident pre- and postsynaptic activity modifies GABAergic synapses by postsynaptic changes in Cl– transporter activity. Neuron 39, 807–820.
Ye JH, Wang F, Krnjevic K, Wang W, Xiong ZG & Zhang J (2004). Presynaptic glycine receptors on GABAergic terminals facilitate discharge of dopaminergic neurons in ventral tegmental area. J Neurosci 24, 8961–8974.
Zhang SJ & Jackson MB (1995). GABAA receptor activation and the excitability of nerve terminals in the rat posterior pituitary. J Physiol 483, 583–595., 百拇医药(Sheyla Mejia-Gervacio and Alain Marty)
Abstract
Recent work has shown that certain neurones have axonal GABAA receptors, whose tonic activation modifies their firing properties and neurotransmitter release capability. In addition, results obtained in interneurones of the molecular layer of the cerebellum indicate that action potential-released GABA binds back to the axon that released it, generating an autoreceptor current. In the present paper, we show that at physiological Cl–i concentration (15 mM) and at 34–36°C, the autoreceptor current generates a large amplitude (up to 21 mV) afterdepolarization that lasts for about 150 ms, and that occasionally leads to double firing. Furthermore we show that elimination of the afterdepolarization, by either blocking GABAA receptors, or eliminating the autoreceptor currents through prolonged whole-cell recording, decreases burst firing. Ih (a hyperpolarization-activated current) was previously found to be prominent in interneurone axons. We show that blocking Ih leads to an increase in the amplitude of the autoreceptor current as well as of the associated afterdepolarization, suggesting a shunting effect of Ih on autoreceptor-mediated afterdepolarization. Conversely, blocking Ih accentuates burst firing. The effects of autoreceptor-mediated afterdepolarization on firing are prominent during a period of development when interneurone synapses are stabilized and vanish by postnatal day 17 (PN 17), together with the expression of the autoreceptor current. Altogether, this work reveals a new role for autoreceptors in the regulation of cell excitability and firing pattern, which may contribute to the development and stabilization of the cerebellar network.
, 百拇医药
Introduction
Molecular layer interneurones (MLIs) form a network of interconnected cells in the cerebellar cortex. Like many other GABAergic neurones throughout the brain, MLIs display spontaneous activity both in vitro (Midtgaard, 1992a; Llano & Gerschenfeld, 1993a; Vincent & Marty, 1996; Husser & Clark, 1997) and in vivo (Jorntell & Ekerot, 2003). MLIs make powerful GABAergic synapses onto Purkinje cells and other MLIs, in both cases modulating the firing of their postsynaptic cells (Midtgaard, 1992b; Vincent et al. 1992; Callaway et al. 1995; Husser & Clark, 1997; Kondo & Marty, 1998; Chavas & Marty, 2003).
, 百拇医药
MLI firing can be periodic (Midtgaard, 1992a; Husser & Clark, 1997; Carter & Regehr, 2002), irregular (Jorntell & Ekerot, 2003) or bursting (Mann-Metzer & Yarom, 2002), but the mechanisms responsible for switching from one mode to another are poorly understood. MLIs exhibit several intrinsic conductances around the resting membrane potential, including Ih (Midtgaard, 1992a; Southan et al. 2000), and a non-inactivating Na+ current (Mann-Metzer & Yarom, 2002) that contributes to determining the duration of the train of action potentials induced by an excitatory current. Following a train of spikes, MLIs exhibit a biphasic hyperpolarizing afterpotential (Midtgaard, 1992a), which presumably reflects, at least in part, the activation of axonal Ca2+-dependent channels (Tan & Llano, 1999) by a biphasic calcium concentration decay (Collin et al. 2005).
, http://www.100md.com
Recently a novel tail current was described in voltage-clamped MLIs following depolarization-induced axonal firing (Pouzat & Marty, 1999). This current was abolished when blocking either GABAA receptors (GABAARs) or neurotransmitter release (either internally, using a calcium buffer, or externally, by blocking voltage-dependent calcium entry). It appeared to reflect the activation of axonal GABAARs following action potential-mediated release of GABA from the cell axon, and was therefore called autoreceptor current.
, 百拇医药
Several recent reports show that the activation of presynaptic ionotropic receptors for glycine and GABA can modulate neurotransmitter release and consequently synaptic transmission either in a facilitatory or inhibitory way (Jackson & Zhang, 1995; Turecek & Trussell, 2001, 2002; Belenky et al. 2003; Kawa, 2003; Axmacher & Draguhn, 2004; Ye et al. 2004). Likewise, sustained activation of these receptors has been shown to change the excitability of the presynaptic cell in the spinal cord (Rudomin et al. 1981), in the pituitary gland (Jackson & Zhang, 1995) and in hippocampal mossy fibres (Ruiz et al. 2003). However, the potential effects of phasic activation of axonal GABAA autoreceptors on cell excitability have not been explored. Since the activation of autoreceptor currents after each action potential might induce changes in the probability of firing, we studied the effects of the activation of autoreceptor currents on the spontaneous firing pattern of MLIs.
, 百拇医药
Our results show that the autoreceptor current does contribute to shaping the firing pattern of MLIs, thus revealing a new mechanism capable of modulating the excitability of central neurones and of inducing switches between various modes of firing.
Methods
Preparation
We used Sprague Dawley rats aged from 8 to 17 days. We obtained parasagittal, 180–200 μm thick cerebellar slices as previously described (Llano & Gerschenfeld, 1993a). Animals were decapitated after either cervical dislocation or deep anaesthesia with Metofane (Janssen; 0.5 ml in 2 l administered for 1–2 min). The extracellular solution used both for slicing and recording contained (mM): 125 NaCl, 2.5 KCl, 2 CaCl2, 1 MgCl2, 1.25 NaH2PO4, 26 NaHCO3, and 10 glucose. The pH was 7.4 when equilibrated with a 5% CO2 and 95% O2 mixture. Slices were maintained at 34°C for 1–5 h and transferred to the recording chamber as needed for the experiments. In the recording chamber, slices were continuously perfused at a rate of 1–1.5 ml min–1 with extracellular solution pre-equilibrated with a 5% CO2 and 95% O2 mixture. During recording, the chamber was heated in order to maintain a temperature of 34–36°C.
, http://www.100md.com
Both types of MLIs were included in the study. Following the criteria used by Llano & Gerschenfeld (1993a), cells located on the lower third and upper two thirds of the molecular layer were, respectively, classified as basket and stellate.
Drugs
(–)-Bicuculline methochloride (10 μM, Tocris), SR 95531 hydrobromide (Gabazine, 20 μM, Tocris) and picrotoxin (200 μM, Sigma) were used as pharmacological blockers of GABAARs. 4-Ethylphenylamino-1,2-dimethyl-6-methylaminopyridinium chloride (ZD7288, 10 μM, Tocris) was used as a blocker of Ih.
, http://www.100md.com
A total number of 77 experiments were performed to test the effects of various manipulations (application of GABAAR blockers, of ZD7288, and prolonged whole-cell recording) on the bursting behaviour of MLIs. Of these, 11 experiments (between 1 and 3 for each protocol) were done in the presence of 2,3-dihydroxy-6-nitro-7-sulphonyl-benzo[f]quinoxaline (2 μM, NBQX, Tocris), in order to check for the possible dependence of the phenomena studied on ionotropic glutamate receptor activity. Since these were indistinguishable from those performed without NBQX, the results in the presence and absence of this blocker were pooled together. All drugs were bath applied.
, http://www.100md.com
Electrophysiology
Whole-cell and cell-attached patch clamp recordings were performed using the following intracellular solution (mM): 10 GABA, 139 potassium gluconate, 6 KCl, 4.6 MgCl2, 0.1 EGTA, 10 Hepes, 0.4 Na2GTP, 4 Na2ATP; pH 7.4. The total Cl– concentration in this solution was 15 mM, which corresponds to the physiological value (Chavas & Marty, 2003). GABA was included since preliminary experiments showed that addition of 10 mM GABA to the pipette solution significantly slowed down the rundown of autoreceptor currents (J. Chavas, unpublished experiments; see also Smith & Jahr (2002) for the description of a similar effect in granule cells of the olfactory bulb). Pipette resistances were around 5–7 M for whole-cell and around 8–11 M for cell-attached experiments. In whole-cell mode, series resistance values during recording ranged from 10 to 35 M and were not compensated. Experiments were discontinued if the series resistance of a cell increased above 35 M. Cells were rejected if, at any moment during the experiment, holding currents in excess of 40 pA were required to maintain a potential of –70 mV.
, 百拇医药
Autoreceptor current. To test for the activation of autoreceptor currents we used a modification of the protocol described by Pouzat & Marty (1999). Briefly, in whole-cell configuration, cells were voltage clamped to –90 mV. A 1–2 ms long, 90 mV large depolarizing pulse was applied to induce the activation of a fast inward Na+ current, reflecting the generation of an action potential in the axon. The autoreceptor current was manifested as a slow inward tail current following the Na+ spike. This test was performed for each cell included in this study. Autoreceptor current amplitudes were determined 7 ms after the peak Na+ current. A high amplitude variability as well as a high rate of failures are characteristic features of autaptic currents (Pouzat & Marty, 1998; see the discussion on the distinction between autaptic currents and autoreceptor currents in Pouzat & Marty, 1999). Therefore, in each experiment, tail currents were examined for failures, and the coefficient of variation (CV) value for the amplitudes of the tail current was measured. In most cases, no failure was observed, and the CV had values between 0.12 and 0.24, indicating an absence of autaptic currents. In a single case, however, failures were observed, and the CV value was 0.49, indicating the presence of an autapse; this cell was removed from the analysis. The low incidence of autapses in PN 8–17 preparations which is apparent from these results is consistent with our previous study (Pouzat & Marty, 1999).
, 百拇医药
Afterpotential. In current clamp mode a holding current up to –40 pA was applied to keep the cells at a stable membrane potential between –60 and –75 mV, within the range of resting membrane potentials of MLIs (as described by Chavas & Marty, 2003). In these conditions, 20–60 ms depolarizing steps were applied to induce either a single action potential or up to four action potentials. The duration and amplitude of the steps were adjusted such that the down stroke of the last action potential coincided with the end of the step. In this way confounding effects of aborted spikes on the afterpotential were avoided. Ten to twenty traces were averaged. The afterpotential was measured as the integrated area of the membrane potential profile over a 200-ms-long time period starting at the end of the pulse, taking as a baseline the cell resting membrane potential before stimulation. In order to test for afterpotential dependence on voltage we did additional experiments in which the holding current was adjusted in order to vary the membrane potential in a range between –59 and –84 mV.
, 百拇医药
Ih. Ih was induced by injection of hyperpolarizing voltage steps (10 mV increments, 200 ms duration) from –80 to –120 mV in cells clamped at –70 mV. The amplitude of the inward current generated was determined as the average of the current during the last 50 ms from each step (when the current was already stationary) with respect to the initial current.
Spontaneous firing pattern. Spontaneous firing activity was recorded for periods of 1 min both in whole-cell current clamp mode and in cell-attached configurations. Action potentials were automatically localized, and interspike intervals (ISIs) were calculated using an Igor (Wavemetrics, Lake Oswego, OR, USA) routine written in the laboratory by C. Pouzat. In order to determine the effects of pharmacological treatments as well as of prolonged periods in whole-cell recording, we calculated the CV for the ISI distributions and compared them before and after experimental manipulations using a paired Student's t test.
, 百拇医药
Statistical treatment
All results are expressed as mean ±S.E.M. Experiments involving a control and a test period were analysed using a paired Student's t test. In other cases an ANOVA comparison was used. Differences were considered significant at P < 0.05.
Results
Autoreceptor current
In voltage-clamped MLIs, application of a short depolarizing voltage pulse induces a sequence of a classical fast inward Na+ current followed by a slow inward current that is sensitive to bicuculline or gabazine (Pouzat & Marty, 1998, 1999). Previous work indicates that the latter current is generated by GABA which is released from the axon of the MLI under study, and which activates either somatodendritic receptors (autaptic current: Pouzat & Marty, 1998) or axonal receptors (autoreceptor current: Pouzat & Marty, 1999). It was shown that autaptic and autoreceptor currents can be unambiguously distinguished, and that the two forms of currents have opposite developmental profiles. The present results were obtained in an age range where autoreceptor currents largely predominate (PN 11–14). Furthermore, occasional cells with autapses were identified and removed from the analysis (see Methods). The tail currents in the remaining cells are autoreceptor currents. These currents were characterized earlier under symmetrical Cl– conditions at room temperature (Pouzat & Marty, 1999). In the present work we used a physiological intracellular Cl– concentration (15 mM, see Chavas & Marty, 2003) and a recording temperature of 34–36°C. Under these conditions, an inward autoreceptor current was clearly measurable at –90 mV, but it was considerably smaller than that previously reported, and it did not always display a clear temporal separation from the early inward Na+ current. The total duration of this current was about 50 ms. Its amplitude was measured after a 7 ms delay following the peak of the Na+ current surge, giving an average of 34 ± 5 pA. This value significantly decreased to 7 ± 2 pA (n= 7) (paired t test, P < 0.01) after bath application of either 10 μM bicuculline or 20 μM gabazine (not shown).
, http://www.100md.com
Autoreceptor-driven afterdepolarization
To study the effects of the autoreceptor current under current clamp conditions, the firing rate of MLIs was reduced by hyperpolarizing them slightly to a range between –60 and –80 mV. Action potentials were generated by applying depolarizing current steps (Fig. 1). The action potentials were followed by a depolarizing afterpotential (Fig. 1A), which occasionally induced firing (3 cells out of 31; example in Fig. 1B). The afterdepolarization was variable from cell to cell, both in amplitude and in time course. For the largest amplitude values, a clear peak was observed around 25–35 ms after the peak of the last spike, but in other cases, no clear peak could be defined. Under the conditions of our experiments, GABAergic synaptic currents decay with a mean weighted time constant of 6 ms (authors' unpublished results), shorter than the interval between spikes in the experiments of Fig. 1 (10–20 ms). Therefore, little overlap was expected between GABA-induced conductances produced by successive action potentials. Nonetheless, we studied the dependence of the afterdepolarization on the number of action potentials induced by the stimulation pulse. As shown in Fig. 1B, we found no significant correlation between the number of action potentials and the integral of the afterdepolarization (integrated over a 200 ms period following the end of the stimulus) (ANOVA F= 0.25; P= 0.86, compared for 1–4 action potentials; n= 27 cells). MLI–Purkinje cell synapses show little or no potentiation under resting conditions in juvenile preparations (Pouzat & Hestrin, 1997; Caillard et al. 2000), and MLI terminals exhibit little accumulation of presynaptic calcium for short trains (Collin et al. 2005). Therefore, little potentiation of GABA release was expected. Overall, the lack of summation of afterdepolarizations is consistent with previous results on this preparation.
, http://www.100md.com
A, an example of depolarizing afterpotential elicited by one action potential. A subthreshold response (dashed trace) is shown for comparison. B, the depolarizing afterpotential had a size which was independent of the number of action potentials during the stimulation pulse, and occasionally elicited delayed firing. Left: superimposed current responses to depolarizing steps that were adjusted to produce one (black trace) or 4 (grey traces) action potentials. Afterdepolarization (ADP) size was quantified as the area under the voltage profile during a 200 ms long time window starting at the end of the stimulation pulse, taking the initial membrane potential as a baseline. Individual traces show no apparent change in afterpotential amplitude when inducing either one or four action potentials. As it can be seen, the afterdepolarization sometimes induced firing. Right: mean (±S.E.M.) afterdepolarization integral for current pulses inducing 1–4 action potentials (ANOVA F= 0.2516; P= 0.86, 27 cells). C, the afterdepolarization was sensitive to GABAAR blockers. Left: current clamp recordings from a cell held at –80 mV in control and after applying 10 μM bicuculline for 3 min in the bath solution. Right: afterdepolarization integral for 10 cells recorded in control and after 3 min in GABAAR blockers, showing a significant reduction (paired t test, P < 0.01) after the pharmacological treatment. D, the afterdepolarization was sensitive to ‘rundown’. Left: current clamp recordings from a cell held at –70 mV, 5 min after establishment of whole-cell recording (Control) and after > 20 min of whole-cell recording (Rundown). Right: mean afterdepolarization integral was significantly diminished after periods of between 20 and 38 min in whole-cell recording compared with the first 5 min (n= 7 cells).
, 百拇医药
To test whether the afterdepolarization recorded in current clamp was related to the activity of the autoreceptor currents, we studied the effects of experimental manipulations blocking these currents. For these comparisons, we pooled together results obtained with one to four action potentials, since the number of action potentials did not appear to significantly affect the size of the afterdepolarization. As shown in Fig. 1C, the amplitude of the afterdepolarization was always drastically reduced after 3 min application of a GABAAR blocker (either 10 μM bicuculline (n= 5), 20 μM gabazine (n= 4) or 200 μM picrotoxin (n= 1)). Pooling the results from these 10 cells together, the afterdepolarization integral fell from 438 ± 82 mV x ms to 91 ± 23 mV x ms (paired t test, P < 0.01). These results show that the major part of the afterdepolarization arises from the activation of GABAARs.
, 百拇医药
The size of the afterdepolarization was surprisingly large. In the above series, the afterdepolarization amplitude ranged from 3 to 21 mV, with a mean value of 10.6 ± 1.6 mV. In 6/10 cells, the peak depolarization was less negative than ECl (–58 mV), reaching up to –45 mV. The difference cannot be explained on the basis of the contribution of bicarbonate ions, which do not alter appreciably the reversal potential value of GABAARs in our whole-cell recording conditions (no bicarbonate inside: see Chavas & Marty, 2003). Therefore, the results suggest that, even though an activation of GABAARs underlies the afterdepolarization, other mechanisms are also involved. In light of the earlier results by Mann-Metzer & Yarom (2002) it seems likely that a non-inactivating Na+ current amplifies the afterdepolarization and, in about half of the cells, brings it above the value of ECl.
, http://www.100md.com
In whole-cell recording, autoreceptor currents undergo an irreversible decrease which is presumably due to the diffusion of one or several substances out of the cell (‘rundown’: Pouzat & Marty, 1999). We included GABA in the recording pipette to slow down this process, but even under these conditions, rundown of autoreceptor currents (as observed in voltage clamp) was severe or complete after > 20 min of whole-cell recording. Likewise, afterdepolarization diminished from 292 ± 89 mV x ms, when measured in the first 5 min after establishment of whole-cell recording, to –85 ± 100 mV x ms when measured after more than 20 min in whole-cell recording (paired t test, P < 0.02; 7 cells) (Fig. 1D). Therefore, rundown was observed in the same time domain for the afterdepolarization and for the autoreceptor current. Taken together, the results led us to conclude that the afterdepolarization was primarily due to the activation of GABAA autoreceptors following GABA release from the axon.
, 百拇医药
Lack of sensitivity of afterdepolarization on membrane potential
We next investigated the sensitivity of the afterdepolarization to the cell membrane potential. In isotonic intracellular Cl– condition (ECl= 0 mV), the peak amplitude of autoreceptor currents increases linearly with potential, as predicted for a conductance which is primarily carried by Cl– ions (Pouzat & Marty, 1999). We expected that with 15 mM intracellular Cl– concentration, the afterpotential amplitude would be modulated in the same way when the initial membrane potential would be moved away from ECl (–58 mV). Contrary to this expectation we found no statistical difference between afterdepolarization integrals measured at membrane potentials ranging from –59 to –84 mV (binned in 5 mV groups; ANOVA F= 1.405, P= 0.25; Fig. 2A). Nonetheless, since the afterpotential is mainly driven by the autoreceptor-dependent conductance, we expected it to be absent or reversed when the cell potential was near or above ECl. We therefore analysed single action potentials taken from segments of spontaneous activity recordings in which the membrane potential before the spike was near or above ECl. In this case, as shown in Fig. 2B, we found negative values for the afterpotential, averaging –753 ± 148 mV x ms (cell potential 200 ms before the spike: –47 to –50 mV; 6 cells). These results are consistent with the notion that the afterpotential is mediated primarily by a Cl–-selective conductance.
, http://www.100md.com
A, afterdepolarization (ADP) is not strongly voltage dependent in the hyperpolarizing range. Pooled results showing the afterdepolarization integral as a function of membrane potential, from a total of 15 cells. Afterdepolarization was measured at membrane potentials from –59 to –84 mV; results were binned in groups of 5 mV for analysis. Note the non-significant reduction of the integral for the most depolarized as well as the most hyperpolarized potentials (ANOVA F= 1.26; P= 0.31). B, lack of afterdepolarization in cells held above ECl. Example of time-averaged spontaneous action potentials, in a depolarized cell (membrane potential: –48 mV).
, http://www.100md.com
Interaction between autoreceptor current and Ih
A likely explanation for the anomalous evolution of the afterdepolarization with hyperpolarization was that the increase in driving force was counteracted by a shunt brought about by a hyperpolarization-driven conductance. Since MLIs exhibit a large Ih current which turns on around resting membrane potential (Midtgaard, 1992a), we decided to study the interaction of autoreceptor currents with Ih. In our voltage clamp recordings Ih was seen as an inwardly rectifying, slowly activating current, which was readily apparent for hyperpolarizing steps. In the presence of 10 μM ZD7288, Ih relaxation amplitudes in response to steps to –80 and –120 mV diminished from 10.7 ± 2.1 to 4.6 ± 1.0 pA and from 76.1 ± 5.1 to 23.3 ± 4.4 pA (n= 5; Fig. 3A). To test whether Ih contributes to the cell conductance near resting potential we compared cell resistance as measured at –70 mV before and with ZD7288, and found no difference (1.55 ± 0.40 and 1.73 ± 0.37 G, respectively).
, 百拇医药
A, consecutive 10 mV hyperpolarizing pulses (test potential: –80 to –120 mV) were applied in a MLI voltage clamped at –70 mV (first trace without hyperpolarizing pulse). In control conditions, hyperpolarization induced the activation of an inward Ih current characterized by its slow onset kinetics. This current was blocked by 10 μM ZD7288. B, in voltage clamp, Ih blockage significantly increased the size of autoreceptor currents. Upper panel: individual traces in a cell voltage clamped at –90 mV and shortly depolarized to 0 mV in control conditions (black trace) and in ZD7288 (grey trace). Lower panel: autoreceptor currents in control and in ZD7288 (n= 7) (paired t test, P < 0.02). C, left: individual current clamp traces in a cell held at –70 mV, showing an increase in afterdepolarization size after Ih blockade with ZD7288; this increase was accompanied by an increased proportion of cells spiking during afterdepolarization. Both traces were recorded from the same cell in exactly the same conditions. Note the increased cell excitability during the depolarizing pulse, which had to be reduced in ZD7288 compared with its control amplitude in order to limit firing to one spike. Right: data from 6 cells, showing a significant increase in the afterdepolarization integral upon ZD7288 application (paired t test, P < 0.05; black and grey traces as in B here and in following figures).
, http://www.100md.com
We next tested the effects of 10 μM ZD7288 on the amplitude of autoreceptor currents as measured in voltage clamp, as well as on that of the afterdepolarization associated with autoreceptor currents. We found that autoreceptor currents increased by 45%, from 31 ± 11 to 45 ± 10 pA (paired t test, P < 0.05; n= 7) (Fig. 3B). The most likely interpretation for this finding is that suppression of Ih increases the space constant of voltage control in the axon, leading to a recording of the autoreceptor current on a larger area in ZD7288 than in control conditions. Such a change in the axon space constant would not necessarily be apparent as a change in resistance as recorded in the somatic compartment, for two reasons. First, the increase in the space constant counteracts the effects of a reduction in specific membrane conductance on the input resistance. Second, the resting conductance of the somatodendritic domain may be little affected by ZD7288; since this conductance is in parallel with that of the axon, its contribution could occlude any change in axon input resistance. The hypothesis of a change in the axonal space constant is supported by the fact that Ih has been preferentially found in the axonal domain of MLIs (Santoro et al. 1997; Southan et al. 2000). In current clamp, ZD7288 likewise increased the afterdepolarization produced by a train of action potentials, tested for membrane potentials between –70 and –72 mV, from 345 ± 127 mV x ms to 1097 ± 303 mV x ms (paired t test, P < 0.05) (Fig. 3C, ZD7288; n= 6). Moreover the blockade of Ih increased the proportion of cells in which firing was observed during the afterdepolarization (5/7 cells; Fig. 3C) compared with control conditions where only 3/31 cells were spiking during the afterdepolarization. Finally, it was noted in these experiments that the amplitude of the injected current needed to be reduced in ZD7288 compared with the control situation in order to match the number of spikes during the depolarizing step, as exemplified in Fig. 3C. This effect, like those on the autoreceptor current and its associated depolarization, may be explained by an increase in the length constant of the axon in the presence of ZD7288, which would account for a higher efficacy of somatic depolarization in triggering spikes at their site of initiation. Taken together, the results reveal a negative interaction between autoreceptor current and Ih.
, 百拇医药
Influence of autoreceptors on firing pattern
In order to determine whether autoreceptor currents could influence the firing pattern of MLIs, we compared ISIs in control conditions and after experimental manipulations affecting this current. We analysed the spontaneous firing activity both in whole-cell and in cell-attached recordings. In whole-cell mode, the membrane potential was now set at –65 ± 5 mV, close to the normal resting values (Chavas & Marty, 2003). The average spontaneous firing frequency in this condition was 3.38 ± 0.85 Hz (n= 6). In our cell-attached experiments at physiological temperature the average frequency was 2.46 ± 0.52 Hz (n= 5), similar to the frequency in our whole-cell experiments and within the range reported in cell-attached conditions at room temperature in this age range (Llano & Gerschenfeld, 1993a; Vincent & Marty, 1996; Chavas & Marty, 2003).
, 百拇医药
In control conditions, action potential firing appeared highly irregular, with isolated spikes and short bursts of up to five spikes alternating with silent periods in the order of one or a few seconds in duration. ISI histograms displayed a peak near 0.1–0.2 s, followed by a tail extending to the seconds range. As shown in Fig. 4A and B, application of GABAAR blockers shifted the firing pattern to a more regular one. ISI histograms became more compact; their peak was shifted to higher values, and the slow tail became less conspicuous. To quantify the change of shape of ISI histograms, we analysed the CV values of ISIs, reasoning that the more compact histogram shape associated with the suppression of autoreceptor currents should decrease this parameter. Indeed we found that CV values were significantly diminished, from 2.56 ± 0.57 in control conditions to 0.97 ± 0.08 after 3–5 min application of GABAAR blockers in whole-cell recording configuration (paired t test, P < 0.02, n= 6) (Fig. 4A1), and from 1.30 ± 0.12 to 0.94 ± 0.04 in cell-attached configuration (paired t test, P < 0.05, n= 5) (Fig. 4B1). On the other hand, GABAAR blockers did not change the mean ISI significantly (whole-cell recording experiments: 0.82 ± 0.47 s in blockers versus 0.35 ± 0.08 s in control; paired t test: P= 0.39; n= 6; cell-attached experiments: 0.42 ± 0.13 s in blockers versus 0.51 ± 0.10 s in control; paired t test: P= 0.20; n= 5). Thus, the change in CV cannot be ascribed to a general increase or decrease in spiking.
, 百拇医药
A, upper traces: whole-cell recordings of spontaneous activity in current clamp mode in a cell held around –70 mV before and after bath application of bicuculline. The cell fired groups of 1–5 action potentials in control conditions, but displayed a more regular discharge after blocking GABAARs. Lower panels: this change was manifested in the distribution of the ISI histogram, which became more homogeneous, while its peak was shifted to the right. A1, CV values of ISIs diminished significantly in GABAAR blockers compared with control values (paired t test, P < 0.02, n= 6). A2, cumulative normalized ISI histograms for times shorter than 400 ms (same experiment as traces in A) show that the main changes occurred in this period as expected considering the duration of the afterdepolarization. B, spontaneous activity recorded in cell-attached mode before and after application of gabazine (upper traces). As in whole-cell recording (A), blocking GABAARs changed the firing pattern to a more regular one. Again, the ISI histogram became more homogeneous and its peak was displaced to the right (lower panels). B1, changes in the distribution of ISIs were reflected in a significant decrease in the CV values of ISIs (paired t test, P < 0.05, n= 5). B2, cumulative histograms for the first ISI bins showing a diminution in the number of short ISIs. C, spontaneous activity in current clamp under control conditions and after 10 μM ZD7288. Blockade of Ih favoured a ‘bursty-like’ pattern of discharge. In the example shown, the peak of the distribution became more prominent and was shifted to the left, while the number of ISIs larger than 1 s increased. C1, CV values of ISIs increased significantly after 10 μM ZD7288 (paired t test, P < 0.05, n= 6). C2, cumulative histograms showing an increase in the number of ISIs shorter than 400 ms after ZD7288. D, spontaneous activity of a MLI held near –70 mV, recorded in current clamp mode during the first 5 min of whole-cell recording (Control; left), and after more than 20 min of whole-cell recording (Rundown; right). The firing exhibited short bursts initially, and changed to a more regular discharge later. Accordingly, the ISI histograms became more homogeneously distributed, and the position of its peak shifted to the right (lower panels). D1, CV values for ISIs are significantly diminished after > 20 min of whole-cell recording (paired t test, P < 0.05, n= 6). D2, initial portion of the normalized cumulative ISI histograms from the same experiment, highlighting the initial surge in control (continuous line) contrasting with the delayed onset after rundown (dotted line).
, http://www.100md.com
According to our results, the GABAAR-driven afterdepolarization lasts roughly 150 ms after the spike (see Figs 1 and 3). On this basis the most striking effect of autoreceptors on the ISI histogram is expected during this time period (see Fig. 4A2 and B2; normalized cumulative histograms for the first bins in the distribution). Therefore we evaluated the percentage of spikes falling in the first three bins of our histograms (corresponding to 150 ms after spiking) before and after blocking the GABAARs. The percentage values diminished from 73% to 27% and from 35% to 24% for whole-cell (n= 6) and cell-attached (n= 5) experiments, respectively.
, http://www.100md.com
Conversely, since Ih blockade increases the amplitude of the autoreceptor current and of its associated afterdepolarization, we asked whether pharmacological blockade of Ih would favour the bursting of MLIs. In this experimental group, the mean ISI value was 0.64 ± 0.12 s in control conditions, and remained unchanged after 10 μM ZD7288 (0.49 ± 0.06 s, paired t test, P= 0.09, n= 6). Meanwhile, contrary to the results of previous manipulations, the blockage of Ih changed the firing pattern to a more ‘bursty’ one (Fig. 4C). This was not accompanied by any systematic change in the position of the ISI histogram peak (paired t test, P= 0.58). Reflecting the change in the shape of ISI distribution, CV values increased significantly after ZD7288 application, from 1.73 ± 0.16 to 2.61 ± 0.54 (paired t test, P < 0.05; Fig. 4C1). Thus, these results, like those illustrated in Fig. 3, indicate an antagonistic influence of Ih over autoreceptor current. They are also consistent with a previous finding that Cs+, a blocker of Ih, favours bursting of MLIs (Mann-Metzer & Yarom, 2002). Finally, ZD7288 was found to enhance the first bins of the cumulative ISI histogram (Fig. 4C2), contrary to the decrease observed by blocking the autoreceptor current.
, 百拇医药
Even though an implication of autoreceptors is an attractive explanation for the effects illustrated in Fig. 4A–C, other interpretations need also to be considered. While autoreceptors are axonal, GABAARs of the somatodendritic compartment are engaged in MLI–MLI synapses (Kondo & Marty, 1998) and in autapses (Pouzat & Marty, 1998). In addition it has been suggested that, in slices from adult rats, ambient GABA activates extrasynaptic receptors located in the somatodendritic domain of MLIs (Husser & Clark, 1997). A priori GABAAR blockers could modify the firing pattern by acting on any of these four groups of GABAARs, or on a combination of them. Autapses can be dismissed because their occurrence at the age considered here (PN 11–14) is very low (Pouzat & Marty, 1999) and because in whole-cell recording experiments, occasional cells containing an autapse were unambiguously identified (see Methods) and rejected. A contribution of extrasynaptic somatodendritic GABAARs is also unlikely since we found no change in mean firing frequency upon application of bicuculline or gabazine. This leaves us with classical synapses and with autoreceptors. To distinguish between these possibilities we made use of the fact that autoreceptor currents run down in whole-cell recording (Pouzat & Marty, 1999), whereas currents due to MLI–MLI synapses are stable for > 30 min of whole-cell recording (Llano & Gerschenfeld, 1993a). In our recording conditions we found that the rundown of autoreceptor currents was nearly complete in 20 min; therefore, we compared the spontaneous firing pattern at the beginning (first 5 min) and after at least 20 min of whole-cell recording (Fig. 4D). In these experiments we observed a shift in the peak of the ISI distribution from values under 0.1 s to values around 0.3 s (histograms in Fig. 4D) without significant changes in the mean ISI value (1.07 ± 0.36 and 1.10 ± 0.29, paired t test, P= 0.92, n= 6). As observed after application of GABAAR blockers, the percentage of ISIs smaller than 150 ms was decreased from 54 to 16% after periods longer than 20 min in whole-cell recordings (Fig. 4D2). Finally, and most importantly, statistical comparison of the CV of ISIs showed a significant decrease after rundown of autoreceptor currents from a value of 2.96 ± 0.49 to 1.07 ± 0.22 (paired t test, P < 0.02, n= 6) (Fig. 4D1).
, http://www.100md.com
In summary, the results from Fig. 4A–C indicate that activation of GABAARs favours burst firing. In addition, the experiments shown in Fig. 4D led us to conclude that axonal autoreceptors, rather than somatodendritic receptors, are primarily responsible for the modulation of MLI firing following application of GABAAR blockers.
Autoreceptor-driven afterdepolarization and bursting behaviour are developmentally regulated
All results presented this far have been obtained from animals aged 11–14 days. However, autoreceptor currents are developmentally regulated (Pouzat & Marty, 1999), thus raising the issue of how much the conclusions depend on age. Within the PN 11–14 age group, we found a 2-fold decrease in the mean afterdepolarization amplitude. However, since this decrease was not significant (ANOVA P= 0.24), the results for this age group have been pooled in the presentation of the results so far. This procedure seems justified because the age distribution of the cells used for each series of experiments was similar, and because in addition, our conclusions rest on the comparison between a control and a test period, using paired t tests.
, http://www.100md.com
To explore further the age dependence, we examined the afterdepolarization for two additional age groups: PN 8–10 and PN 17. We found that the afterdepolarization integral decreased from 1269 ± 316 mV x ms (n= 5) at PN 8–10 to 467 ± 89 mV x ms in the PN 11–14 group (n= 29) and vanished completely at PN 17 to a value of –47 ± 64 mV x ms (n= 8) (ANOVA P= 1.53 x 10–5) (Fig. 5A). We also examined the firing pattern in the three groups. In the PN 8–10 group, the application of gabazine induced a diminution in CV (from 1.70 ± 0.12 to 1.29 ± 0.12; paired t test, P < 0.05; n= 5), similar to the one obtained for PN 11–14. This diminution was not accompanied by a change in the mean ISI value in the absence and the presence of the drug (0.88 ± 0.17 and 1.27 ± 0.53 s, respectively; paired t test, P= 0.42; n= 5).
, 百拇医药
A, autoreceptor-driven afterdepolarization as a function of postnatal age. The afterdepolarization decreased in amplitude from the PN 8–10 group to the PN 11–14 group, and completely disappeared at PN 17. B, spontaneous activity recorded in current clamp mode in control conditions was unaltered after bath application of 20 μM gabazine in a PN 17 MLI. ISI histograms (lower panels) showed no change upon GABAAR blockade. C, initial portion of the cumulative ISI histogram, showing no difference between control and gabazine data for the cell in B. D, the CV values for ISIs was insensitive to GABAAR blockers (paired t test, P= 0.11; n= 6).
, 百拇医药
Likewise, in view of the absence of GABAAR-driven afterdepolarization at PN 17, we investigated at this age the effects of GABAAR blockers on the spontaneous firing pattern. In control conditions, the cells showed a more regular firing pattern compared with 11- to 14-day-old animals (Fig. 5B), which was reflected in a significant diminution of the CV value in the older group (P < 0.001), when compared with PN 11–14. This change in CV value was not accompanied by any modification of the average spiking frequency: the mean ISI value was 0.43 ± 0.09 s in control conditions, a value not significantly different from that obtained in the 11- to 14-day-old group (P= 0.15). The application of GABAAR blockers at PN 17 did not produce any significant change in the CV value, with an average of 1.32 ± 0.12 versus 1.06 ± 0.07 in control (paired t test, P= 0.11; n= 5) (Fig. 5D). Likewise the mean ISI value was similar in the presence of GABAAR blockers (0.33 ± 0.05 s) to the control value (see above; paired t test, P= 0.12; n= 6). The average position of the ISI histogram peak was smaller than 0.15 s both in control and in the presence of GABAAR blockers (Fig. 5B). Taken together, our results indicate that the presence of an autoreceptor-mediated afterdepolarization is responsible for the differences in firing pattern between PN 11–14 and PN 17. If this is correct, blocking the afterdepolarization at PN 11–14 should make the firing pattern similar to the one at PN 17. This was confirmed by the lack of difference between CV values at PN 17 in control conditions, compared with those obtained at PN 11–14 during application of GABAAR blockers or after prolonged periods of whole-cell recording (ANOVA P= 0.77).
, http://www.100md.com
Discussion
Our study indicates that a GABAAR-driven, developmentally regulated autoreceptor current shapes the firing pattern of MLIs during the second postnatal week. We also show that an Ih current, which is prominent in these cells (Midtgaard, 1992a), reduces the autoreceptor current, and counteracts its effects on MLI firing.
There are several possibilities to explain that GABAAR block can induce a shift from a ‘bursty-like’ firing pattern to a regular one. Of these, (1) the modulation of a constitutively active, GABAAR-mediated extrasynaptic conductance (Husser & Clark, 1997) is unlikely since the alteration in firing pattern induced by GABAAR blockers did not involve a change in the mean firing frequency, as would be expected from a change in tonic GABAA conductance. (2) MLI–MLI synapses could favour bursting, for instance in the case of two MLIs which would be reciprocally connected via excitatory synapses. However MLI–MLI synaptic currents are not affected by prolonged whole-cell recording (Llano & Gerschenfeld, 1993a), whereas the bursting pattern was sensitive to rundown. Furthermore at PN 17, whereas MLI–MLI synaptic currents are strong (Kondo & Marty, 1998), bursting was not apparent before the application of GABAAR blockers, and application of the blockers failed to decrease the CV of the ISIs. (3) The modulation of an autaptic rather than autoreceptor GABAAR-mediated current can be excluded since the proportion of cells presenting autaptic currents is very low at the age under study (about 5%; Pouzat & Marty, 1998, 1999), and since in the present work, the single cell which exhibited characteristic features of autaptic currents, i.e. high amplitude variability and failures, was rejected from the analysis.
, 百拇医药
On the other hand, the proportion of cells showing autoreceptor current at PN 11–14 is about 92% (Pouzat & Marty, 1999), and this current is sensitive to both rundown and GABAAR blockers. Moreover it is developmentally regulated and we found a diminution of both autoreceptor currents and its associated depolarization at PN 17, coincident with the appearance of a regular firing pattern (Fig. 5). The latter result is in accordance with previous results obtained in older animals (Husser & Clark, 1997; Carter & Regehr, 2002). Overall, the results strongly suggest a tight link between bursting and expression of autoreceptors.
, 百拇医药
Previous studies have addressed the role of axonal or presynaptic GABAARs in regulating cell firing. In primary spinal cord afferents (Rudomin et al. 1981; Verdier et al. 2003), in presynaptic terminals of the pituitary gland (Zhang & Jackson, 1995), and in hippocampal mossy fibres (Ruiz et al. 2003), activation of axonal GABAARs depolarizes the axonal membrane, yet inhibits action potential firing. The proposed explanation for this paradoxical inhibition is a depolarization-driven inactivation of Na+ channels (Zhang & Jackson, 1995). In our experiments, by contrast, we found that the autoreceptor current induces a depolarizing afterpotential and produces a net increase in firing probability during the first 150 ms afterspike (as suggested by the high density of ISIs within this range in control conditions), thus shaping firing in a ‘bursty-like’ manner. Consistent with an increase in firing probability, we found a decrease in the current amplitude required to produce a spike during the afterdepolarization (J. Chavas & A. Marty, unpublished observations). The key difference between our work and the previous publications presumably lies in the mode of activation of the GABAARs, which is phasic in the present work, and was tonic in the previous studies. As the rate of Na+ channel inactivation is slower than that of activation, a transient depolarization is more likely to be excitatory than a sustained one.
, 百拇医药
Understanding the exact reason why activation of the autoreceptor current changes firing probability after spikes would require a detailed knowledge of voltage-dependent conductances in MLI axons, which is presently lacking. Previous work has shown that the reversal potential of GABA-sensitive currents, EGABA, is rather depolarized in MLIs (–58 mV), and as a result of this situation, activation of MLI–MLI synapses at membrane potentials more negative than –58 mV leads to postsynaptic depolarization and to an increase in the firing probability (Chavas & Marty, 2003). The depolarized position of EGABA likewise accounts for the excitatory effect of the autoreceptors as revealed in our whole-cell experiments when cells were held at resting membrane potentials more negative than –58 mV. In contrast with our results, earlier work on GABAAR-mediated autapses in the hippocampus and neocortex revealed inhibitory effects, presumably because the values of EGABA in the concerned interneurone subtypes were more hyperpolarized than in the present preparation (Bacci et al. 2003; Pawelzik et al. 2003).
, 百拇医药
In spite of the relatively high level of EGABA in MLIs, Chavas & Marty (2003) showed in cell-attached experiments different populations of neurones in which the stimulation of a presynaptic cell can induce either increases or decreases on the firing rate of MLIs. This contrasts with the results of the present work, which indicate a consistent excitatory effect of GABAAR activation on cell firing. To explain the discrepancy, two explanations can be offered. One would be that EGABA is more depolarized in the axon than in the somatodendritic domain. The other would be that the axonal GABAARs responsible for the autoreceptor currents are associated more directly with depolarization-activated conductances such as non-inactivating Na+ channels (Mann-Metzer & Yarom, 2002). Both hypotheses are compatible with the finding that, in about half of the whole-cell recording experiments, the peak of the autoreceptor-associated depolarization was more depolarized than EGABA. Further work will be needed to decide between these two possibilities.
, 百拇医药
In addition to non-inactivating Na+ channels, other voltage-dependent conductances are likely to contribute together with the autoreceptor current to shape the firing pattern of MLIs. We observed a negative interaction between autoreceptor currents and Ih, possibly mediated by a change in the axonal space constant induced by the blockage of Ih channels, which are primarily localized in the axon (Santoro et al. 1997; Southan et al. 2000). Pharmacological blockade of Ih increased the size of autoreceptor-mediated afterspike depolarization and the probability of afterspike firing. Moreover, Ih removal by ZD7288 induced a change in the spontaneous firing pattern whose direction was contrary to the one induced by the blockage of autoreceptor currents, namely a shift to a more ‘bursty’ pattern. These results are in agreement with the ones by Chan et al. (2004) in GABAergic globus pallidus neurones, who reported a diminution in the regularity of spontaneous cell firing after pharmacological blockade of Ih. All of the above results fit with a scheme where the effects of the autoreceptors are negatively tuned by Ih, possibly through a change in axonal excitability. This raises the possibility that a functional modulation of Ih could alter the autoreceptor current and the bursting behaviour of MLIs. Accordingly, Chan et al. (2004) suggested a regulation of the activity of GABAergic neurones in the globus pallidus by modulation of an Ih conductance. Our results with Ih current could alternatively be explained through an effect on neurotransmitter release as the one reported previously by Beaumont & Zucker (2000) and Chevaleyre & Castillo (2002). However, in those systems Ih was promoting an increase in neurotransmitter release whereas in our preparation Ih diminished the autoreceptor currents. Furthermore, effects of Ih blockade on transmitter release develop after a delay of tens of minutes (Chevaleyre & Castillo, 2002), whereas those of ZD7288 in the present work do not show such a delay. Finally, a recent report (Saitow et al. 2005) showed that noradrenalin effects on MLI neurotransmitter release (Llano & Gerschenfeld, 1993b) are independent of Ih.
, 百拇医药
Finally, it is worth stressing that according to the present results, autoreceptors shape the firing pattern precisely during the period of development where MLI synaptic connections are under maturation (Pouzat & Hestrin, 1997). A similar transient bursting behaviour has been observed in many CNS regions, including developing retinal ganglion and hippocampal pyramidal cells (Costa et al. 1991; Liets et al. 2003; Chen et al. 2005). In the case of pyramidal cells, bursting behaviour was promoted by an afterspike depolarization, which was, however, not mediated by GABAARs. Spontaneous bursting activity is generally assumed to promote neurite elongation and synapse formation (e.g. in the visual system: Shatz, 1996; and in the spinal cord: Ren & Greer, 2003). How this could come about at GABAergic synapses is still unclear. In the present system, it was recently shown that repetitive activation of MLI synapses induces transient increases in Ca2+ concentration in the postsynaptic cells (Chavas et al. 2004). Such postsynaptic signals could modulate the amplitude of GABAergic events (as shown in developing hippocampal synapses by Woodin et al. 2003) and reinforce the strength of synaptic contacts.
, 百拇医药
References
Axmacher N & Draguhn A (2004). Inhibition of GABA release by presynaptic ionotropic GABA receptors in hippocampal CA3. Neuroreport 15, 329–334.
Bacci A, Huguenard JR & Prince DA (2003). Functional autaptic neurotransmission in fast-spiking interneurons: a novel form of feedback inhibition in the neocortex. J Neurosci 23, 859–866.
Beaumont V & Zucker RS (2000). Enhancement of synaptic transmission by cyclic AMP modulation of presynaptic Ih channels. Nat Neurosci 3, 133–141.
, http://www.100md.com
Belenky MA, Sagiv N, Fritschy JM & Yarom Y (2003). Presynaptic and postsynaptic GABAA receptors in rat suprachiasmatic nucleus. Neuroscience 118, 909–923.
Caillard O, Moreno H, Schwaller B, Llano I, Celio MR & Marty A (2000). Role of the calcium-binding protein parvalbumin in short-term synaptic plasticity. Proc Natl Acad Sci U S A 97, 13372–13377.
Callaway JC, Lasser-Ross N & Ross WN (1995). IPSPs strongly inhibit climbing fiber-activated [Ca2+]i increases in the dendrites of cerebellar Purkinje neurons. J Neurosci 15, 2777–2787.
, http://www.100md.com
Carter AG & Regehr WG (2002). Quantal events shape cerebellar interneuron firing. Nat Neurosci 5, 1309–1318.
Chan CS, Shigemoto R, Mercer JN & Surmeier DJ (2004). HCN2 and HCN1 channels govern the regularity of autonomous pacemaking and synaptic resetting in globus pallidus neurons. J Neurosci 24, 9921–9932.
Chavas J, Forero ME, Collin T, Llano I & Marty A (2004). Osmotic tension as a possible link between GABAA receptor activation and intracellular calcium elevation. Neuron 44, 701–713.
, 百拇医药
Chavas J & Marty A (2003). Coexistence of excitatory and inhibitory GABA synapses in the cerebellar interneuron network. J Neurosci 23, 2019–2031.
Chen S, Yue C & Yaari Y (2005). A transitional period of Ca2+-dependent bursting triggered by spike backpropagation into apical dendrites in developing rat CA1 neurons. J Physiol 567, 79–93.
Chevaleyre V & Castillo PE (2002). Assessing the role of Ih channels in synaptic transmission and mossy fiber LTP. Proc Natl Acad Sci U S A 99, 9538–9543.
, http://www.100md.com
Collin T, Chat M, Lucas MG, Moreno H, Racay P, Schwaller B, Marty A & Llano I (2005). Developmental changes in parvalbumin regulate presynaptic Ca2+ signaling. J Neurosci 25, 96–107.
Costa PF, Ribeiro MA & Santos AI (1991). Afterpotential characteristics and firing patterns in maturing rat hippocampal CA1 neurones in vitro slices. Brain Res Dev Brain Res 62, 263–272.
Husser M & Clark BA (1997). Tonic synaptic inhibition modulates neuronal output pattern and spatiotemporal synaptic integration. Neuron 19, 665–678.
, 百拇医药
Jackson MB & Zhang SJ (1995). Action potential propagation and propagation block by GABA in rat posterior pituitary nerve terminals. J Physiol 483, 597–611.
Jorntell H & Ekerot CF (2003). Receptive field plasticity profoundly alters the cutaneous parallel fiber synaptic input to cerebellar interneurons in vivo. J Neurosci 23, 9620–9631.
Kawa K (2003). Glycine facilitates transmitter release at developing synapses: a patch clamp study from Purkinje neurons of the newborn rat. Brain Res Dev Brain Res 144, 57–71.
, 百拇医药
Kondo S & Marty A (1998). Synaptic currents at individual connections among stellate cells in rat cerebellar slices. J Physiol 509, 221–232.
Liets LC, Olshausen BA, Wang GY & Chalupa LM (2003). Spontaneous activity of morphologically identified ganglion cells in the developing ferret retina. J Neurosci 23, 7343–7350.
Llano I & Gerschenfeld HM (1993a). Inhibitory synaptic currents in stellate cells of rat cerebellar slices. J Physiol 468, 177–200.
, 百拇医药
Llano I & Gerschenfeld HM (1993b). Beta-adrenergic enhancement of inhibitory synaptic activity in rat cerebellar stellate and Purkinje cells. J Physiol 468, 201–224.
Mann-Metzer P & Yarom Y (2002). Jittery trains induced by synaptic-like currents in cerebellar inhibitory interneurons. J Neurophysiol 87, 149–156.
Midtgaard J (1992a). Stellate cell inhibition of Purkinje cells in the turtle cerebellum in vitro. J Physiol 457, 355–367.
, 百拇医药
Midtgaard J (1992b). Membrane properties and synaptic responses of Golgi cells and stellate cells in the turtle cerebellum in vitro. J Physiol 457, 329–354.
Pawelzik H, Hughes DI & Thomson AM (2003). Modulation of inhibitory autapses and synapses on rat CA1 interneurones by GABAA receptor ligands. J Physiol 546, 701–716.
Pouzat C & Hestrin S (1997). Developmental regulation of basket/stellate cellPurkinje cell synapses in the cerebellum. J Neurosci 17, 9104–9112.
, 百拇医药
Pouzat C & Marty A (1998). Autaptic inhibitory currents recorded from interneurones in rat cerebellar slices. J Physiol 509, 777–783.
Pouzat C & Marty A (1999). Somatic recording of GABAergic autoreceptor current in cerebellar stellate and basket cells. J Neurosci 19, 1675–1690.
Ren J & Greer JJ (2003). Ontogeny of rhythmic motor patterns generated in the embryonic rat spinal cord. J Neurophysiol 89, 1187–1195.
, 百拇医药
Rudomin P, Engberg I & Jimenez I (1981). Mechanisms involved in presynaptic depolarization of group I and rubrospinal fibers in cat spinal cord. J Neurophysiol 46, 532–548.
Ruiz A, Fabian-Fine R, Scott R, Walker MC, Rusakov DA & Kullmann DM (2003). GABAA receptors at hippocampal mossy fibers. Neuron 39, 961–973.
Saitow F, Suzuki H & Konishi S (2005). -Adrenoceptor-mediated long-term up-regulation of the release machinery at rat cerebellar GABAergic synapses. J Physiol 565, 487–502.
, 百拇医药
Santoro B, Grant SG, Bartsch D & Kandel ER (1997). Interactive cloning with the SH3 domain of N-src identifies a new brain specific ion channel protein, with homology to eag and cyclic nucleotide-gated channels. Proc Natl Acad Sci U S A 94, 14815–14820.
Shatz CJ (1996). Emergence of order in visual system development. Proc Natl Acad Sci U S A 93, 62–608.
Smith TC & Jahr CE (2002). Self-inhibition of olfactory bulb neurons. Nat Neurosci 5, 760–766.
, 百拇医药
Southan AP, Morris NP, Stephens GJ & Robertson B (2000). Hyperpolarization-activated currents in presynaptic terminals of mouse cerebellar basket cells. J Physiol 526, 91–97.
Tan YP & Llano I (1999). Modulation by K+ channels of action potential-evoked intracellular Ca2+ concentration rises in rat cerebellar basket cell axons. J Physiol 520, 65–78.
Turecek R & Trussell LO (2001). Presynaptic glycine receptors enhance transmitter release at a mammalian central synapse. Nature 411, 587–590.
, http://www.100md.com
Turecek R & Trussell LO (2002). Reciprocal developmental regulation of presynaptic ionotropic receptors. Proc Natl Acad Sci U S A 99, 13884–13889.
Verdier D, Lund JP & Kolta A (2003). GABAergic control of action potential propagation along axonal branches of mammalian sensory neurons. J Neurosci 23, 2002–2007.
Vincent P, Armstrong CM & Marty A (1992). Inhibitory synaptic currents in rat cerebellar Purkinje cells: modulation by postsynaptic depolarization. J Physiol 456, 453–471.
, 百拇医药
Vincent P & Marty A (1996). Fluctuations of inhibitory postsynaptic currents in Purkinje cells from rat cerebellar slices. J Physiol 494, 183–199.
Woodin MA, Ganguly K & Poo MM (2003). Coincident pre- and postsynaptic activity modifies GABAergic synapses by postsynaptic changes in Cl– transporter activity. Neuron 39, 807–820.
Ye JH, Wang F, Krnjevic K, Wang W, Xiong ZG & Zhang J (2004). Presynaptic glycine receptors on GABAergic terminals facilitate discharge of dopaminergic neurons in ventral tegmental area. J Neurosci 24, 8961–8974.
Zhang SJ & Jackson MB (1995). GABAA receptor activation and the excitability of nerve terminals in the rat posterior pituitary. J Physiol 483, 583–595., 百拇医药(Sheyla Mejia-Gervacio and Alain Marty)