INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Central nervous system processing of gustatory information first occurs in the rostral nucleus of the solitary tract (rNST). Investigations of this processing have revealed that while excitation is an important component of the synaptic activity in the rNST, the inhibitory neurotransmitter -aminobutyric acid (GABA) has also been shown to play a significant role. Many rNST neurons contain GABA (Lasiter and Kachele 1988) and about half of the synaptic terminals in the rNST are GABAergic (Leonard et al. 1999). In addition, both in vitro and in vivo application of GABA inhibits rNST neurons, and GABA is also involved in a corticofugal tonic inhibition of rNST neurons (Smith and Li 1998, 2000; Wang and Bradley 1993).

In a recent series of studies, we have examined the plasticity of rNST GABAergic inhibition (Grabauskas and Bradley 1996, 1999). Tetanic stimuli mimicking afferent patterns of gustatory input to the rNST can enhance inhibition by increasing the length and shape of the decay time course of the inhibitory postsynaptic potentials (IPSP) (Grabauskas and Bradley 1998, 1999, 2001). Thus the normally fast synaptic inhibition is significantly prolonged, which could influence the extent of the temporal and spatial summation of synaptic activity (Jonas 2000). Lengthening of the IPSP decay time is especially apparent in newborn animals. In addition, at tetanic frequencies between 20 and 50 Hz, the exponential decay time course has an S shape in which the amplitude of the IPSP is sustained even after termination of the stimulus (Grabauskas and Bradley 2001). This type of synaptic response, which disappears after the first two post-natal weeks, has been suggested to be the result of afferent taste-initiated modifications of rNST synapses that results in sharpening of the tuning of rNST taste-responsive neurons (Grabauskas and Bradley 2001).

To date little is known about the mechanisms that contribute to the decay time of GABAergic synapses in rNST. We therefore studied inhibitory synaptic transmission in the rNST of newborn animals to learn about processes that are involved in tetanic stimulus evoked plasticity of GABAergic synapses. We have analyzed single stimulus and tetanic stimulus evoked inhibitory postsynaptic potentials or currents (IPSP/Cs) to demonstrate that both receptor kinetics and diffusion of neurotransmitter from the synaptic cleft contribute to the decay phase of single-stimulus-evoked IPSP/Cs, and only the rate of clearance of neurotransmitter from the synaptic cleft defines the time course of tetanic-stimulus-evoked IPSP/Cs during the first postnatal week in the rNST.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Brain slice preparation

Brain stem slices were prepared from 0- to 7-day-old Sprague-Dawley rats. The preparation of horizontal rNST brain slices has already been described in detail (Bradley and Sweazey 1992; Grabauskas and Bradley 1996, 2001). Rats were decapitated, and the whole brain, including the brain stem, was rapidly removed and placed in ice-cold physiological saline containing (in mM) 124 NaCl, 2.5 KCl, 2.5 CaCl₂, 1.3 MgSO₄, 26 NaHCO₃, 1.25 KH₂PO₄, and 10 glucose, gassed with 95% O₂-5% CO₂ to give a pH of 7.3. The brain was transected at the level of the pons and just below the obex and the cerebellum was removed. Horizontal 300-µm slices containing the whole NST were cut on a Vibratome and placed in a holding chamber. Following 1-6 h recovery, the slice containing the NST was transferred to the recording chamber (volume of ~1 ml, Warner Instrument, Hamden, CT), where it was submerged and held in place by a net and continuously superfused (1-2 ml/min) with physiological saline at 22-37°C.

Electrophysiological recordings

Electrodes were positioned in the rNST using coordinates established in previous anatomical and electrophysiological investigations of the developing rNST (Bao et al. 1995; Grabauskas and Bradley 2001; Lasiter 1992; Lasiter et al. 1989). Whole cell patch-clamp recordings were performed on 92 rNST neurons. Patch pipettes, pulled in two stages from 1.5 mm OD borosilicate filament glass, were filled with a solution containing (in mM) 130 K-gluconate, 10 HEPES, 10 EGTA, 1.0 MgCl₂, 2.5 CaCl₂, 2.0 ATP, 0.2 GTP. Pipette solutions were adjusted to a pH 7.2-7.3 with KOH and had an osmolarity of 275-292 mosM. Electrode resistance was between 5 and 8 M.

Inhibitory postsynaptic potentials or currents were evoked by delivering a stimulus shock (0.1-ms duration) via a bipolar stimulating electrode consisting of a pair of a tightly twisted teflon-insulated platinum/iridium wires (~200 µm overall diameter) placed under visual control in the most rostral portion of the NST. Stimulus intensity was adjusted to evoke IPSP/Cs and ranged from 0.1 to 3 mA. In experiments in which Ca²⁺ concentration of the physiological saline was manipulated, equimolar substitution of Na⁺ ions for Ca²⁺ ions was made.

Intracellular labeling with biocytin

Because it was difficult to visualize the NST in immature brain slices, neurons were intracellularly labeled with 0.2-0.5% biocytin to confirm that the recorded neurons were in the rNST. Biocytin (Sigma) was diluted in the pipette-filling solution and placed in the tip of the recording pipette (Horikawa and Armstrong 1988). The neurons were filled with biocytin by diffusion. Following the experiment the slices were removed from the recording chamber, placed on a piece of filter paper and fixed in 4% neutral-buffered formalin for 24 hr.

After fixation, slices were rinsed in phosphate buffer for 30 min, embedded in agar (4% in distilled water), and cut into 50 to 60-µm-thick sections on a Vibratome. The sections were incubated for 2 hr in avidin-horseradish peroxidase (avidin-HRP) at room temperature. The avidin-HRP was diluted 1:200 in phosphate-buffered saline containing 0.3% Triton-X. The sections were then rinsed three times and reacted with 0.025% diaminobenzidine and 0.01% H₂O₂ for 6-10 min. After rinsing, the sections were mounted on gelatin-coated slides, dried overnight, and then counter stained with cresyl violet (0.05%). The slides were then coverslipped and examined using a light microscope. Because the cresyl violet counter stain labeled cell nuclei, it was possible to clearly identify the unstained solitary tract and therefore identify the extent of the solitary nucleus and confirm that the recorded neurons were located in the rNST. The filled neurons were photographed and examined to determine their morphology and then classified based on previous morphological studies of neuron types in the NST (King and Bradley 1994; Bao et al. 1995).

Drug application

To evoke pure IPSP/Cs, all experiments were performed in the presence of glutamate receptor antagonists (D)-2-amino-5-phosphonopentanoic acid (APV, 50 µM) and 6-cyano-7-nitroquinoxaline-2,3-dione disodium (CNQX, 20 µM, Sigma-RBI). Even though the volume of the slice chamber was small enough to allow for rapid exchange of the contents, 3-5 min were allowed to elapse before making further recordings to allow the concentration of drug and the cell to stabilize after the superfusing solutions were changed. Diazepam was purchased from Sigma-RBI (St. Louis, MO), thapsigargin and t-BuBHQ from Alomone laboratories (Jerusalem, Israel). Diazepam, a benzodiazepine, was used to modulate the GABA_A receptor, and thapsigargin and t-BuBHQ are ATP-dependent Ca²⁺ pump blockers used to modulate Ca²⁺ concentration. The concentrations used were those we have used in our previous investigations or derived from the literature (Fossier et al. 1999).

Data analysis

To analyze the decay kinetics of the IPSCs, exponential curve fitting using pCLAMP 8 software (Axon instruments, Foster City, CA) was used. Visual inspection of the fitting results indicated that for some of the IPSCs the decay kinetics could be fitted by a single exponential, but for others the decay time was best fitted with two exponentials. For a biphasic decay, a mean decay time constant (_m) was calculated where _m = A_fast × _fast + A_slow × _slow where _fast and _slow were the time constants and A_fast and A_slow were amplitude constants. Errors in all measured quantities are given as means ± SE. Statistical significance was determined using the paired Student's t-test. Intergroup comparisons were analyzed with one-way ANOVA followed by the Bonferoni test for individual post hoc comparisons. Groups were considered significantly different when the P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Location of labeled neurons

The results are based on recordings from 92 neurons. Because it is difficult to identify the rNST in slices from neonatal animals, neurons were filled with biocytin to confirm their location. Thirty-six of these neurons were successfully filled and recovered in cresyl violet counterstained slices. Subsequent analysis revealed that the neurons were located between the most rostral and intermediate parts of the NST corresponding to the area of the gustatory NST as defined in previous developmental studies (Fig. 1A) (Bao et al. 1995; Lasiter 1992; Lasiter et al. 1989). Recordings from the remaining neurons were confined to this same location based on the position of the labeled neurons relative to the IVth ventricle (Fig. 1A).

View larger version (104K): [in this window] [in a new window]   Fig. 1. A: diagram of the recording site (stippled) in the horizontal plane of the brain stem that includes the nucleus of the solitary tract (NST; left). The location of the solitary tract is represented as dashed lines. The box indicates the site of the low-magnification photomicrograph (right). The photomicrograph shows a neuron (arrow) that was recorded in whole cell mode and injected with 0.5% biocytin. The slice was counterstained with cresyl violet. B: photomicrographs of filled neurons. A multipolar neuron is illustrated on the left, an ovoid neuron in the middle and an elongate neuron on the right. Scale bar = 400 µm in A and 40 µm in B.

The filled neurons could be separated into three groups (elongate, multipolar, and ovoid) based on previously established morphological criteria (Davis and Jang 1988; King and Bradley 1994; King and Hill 1993; Lasiter and Kachele 1988). Most of the filled neurons were classified as ovoid neurons (42%), whereas 39% were identified as multipolar neurons and 19% as elongate neurons (Fig. 1B).

Single-stimulus shock-evoked IPSCs

Electrical stimulation of the rNST in the presence of glutamate receptor antagonists APV and CNQX evoked IPSCs in all tested neurons. At a holding potential of 60 mV, the mean IPSC amplitudes were 50 ± 7 (n = 36) pA. However, the peak amplitudes of the averaged IPSCs varied from stimulus to stimulus even though the inter-stimulus interval was 10 s and the stimulus strength remained constant. The average decay time of the IPSCs could be fitted by the sum of two exponentials for most of the neurons (30 of 36). The fast (_fast) and slow (_slow) decay phases of the IPSCs were fitted with exponentials having time constants of 38 ± 4 and 181 ± 35 ms, respectively. The decay time of the remaining six neurons was fitted with a single exponential with a mean time constant of 59 ± 8 ms. Comparison of the amplitudes and _m of IPSCs for different morphological types of neurons revealed no significant difference between groups.

Effect of paired-pulse stimulation on the amplitudes of the IPSCs

Paired-pulse stimulation was used to investigate activity-dependent synaptic plasticity in the rNST neurons. Depending on the length of the inter-stimulus interval, the IPSC evoked by the test stimulus could occur before the IPSC evoked by the conditioning stimulus had decayed to baseline. Thus when the inter-stimulus interval was <200 ms, the test-stimulus (I₂)-evoked IPSC summed with the tail current of the conditioning-stimulus-evoked IPSC (I₁; n = 33, Fig. 2A). For 82% of the neurons, the IPSC evoked by the test stimulus was of a greater amplitude than the conditioning-stimulus-evoked IPSC and the ratio of the current amplitudes evoked by conditioning and test stimuli was >1 (I₂/I₁ > 1, Fig. 2, A, C, and D). The increase of the amplitude of the test-stimulus-evoked IPSCs were significant different when the interstimulus intervals were 100 ms (30, 50, and 100 ms, Fig. 2D, P < 0.05). In the remaining neurons (18%), there was a reduction in the IPSC evoked by the test stimulus (I₂/I₁) < 1 (Fig. 2B). Algebraic subtraction of the IPSC tail current evoked by the conditioning stimulus from the IPSC evoked by the test stimulus (Fig. 2, C and D) revealed the contribution of the test stimulus current (I) to the summed IPSC. Thus even though I₂/I₁ > 1 for most neurons, the subtraction demonstrated that for 94% of the neurons the current contributed by test stimulus (I₂) is significantly less than that evoked by the conditioning stimulus (I/I₁ < 1, Fig. 2D, P < 0.05).

View larger version (22K): [in this window] [in a new window]   Fig. 2. A and B: paired-pulse experiments demonstrate that the amplitude of inhibitory postsynaptic currents (IPSCs) evoked by the test stimuli (I2) are the result of summation of the temporal profiles of IPSCs evoked by both conditioning (I1) and test stimuli. However, the result of summation may be greater (A) or less (B) than the amplitude of the IPSCs evoked by the conditioning stimuli. C: diagram to illustrate the derivation of . D: the relationship () between the ratio of the amplitudes of the conditioning and test stimuli evoked IPSCs (I2/I1) and the interstimulus interval reveals that the ratio is greater when test stimulus immediately follows conditioning stimulus and that summation of these 2 stimuli takes place when the interstimulus interval is <200 ms. The relationship () between the increase of the IPSC amplitude resulting from a test stimulus () and the interstimulus interval reveals that the test stimulus contributes less when it immediately follows the conditioning stimulus and almost equals the amplitude of IPSC evoked by conditioning stimulus (I1) when the interstimulus interval is >100 ms. *, significant differences from the normalized amplitude of the control stimulus evoked IPSC (I1 = 1, P < 0.05).

Effect of paired-pulse stimulation on the IPSC decay time

The mean decay time constant (_m) of the IPSC evoked by the test stimulus was longer than the decay time constant of the conditioning stimulus (Fig. 3, Aa and B). The significant increase of the test stimulus evoked IPSC time constant was sustained 100 ms (Fig. 3B, P < 0.05). Further analysis revealed, that the significantly longer mean decay time (_m) of the test-stimulus-evoked IPSC resulted from an increase in the amplitude of the slow component (A_slow/(A_fast + A_slow)) to the total amplitude of the IPSC (Fig. 3C, P < 0.05). Superimposition of the normalized decay times of the IPSCs evoked by the conditioning and the test stimuli demonstrates the increased contribution of the slow component, whereas the time constant of the fast exponential (_fast) remained constant (40 ± 6 ms, n = 23, P = 0.35, Fig. 3Ab).

View larger version (28K): [in this window] [in a new window]   Fig. 3. The effect of paired-pulse stimulation on the decay phase of the IPSCs. A, a: paired-pulse stimulation results in the amplitude summation of IPSCs evoked by conditioning and test stimuli in the presence of 2.5 mM [Ca2+]o. b: superimposition of the decay phase of the IPSCs evoked by both stimuli reveals that the decay of the IPSC evoked by test stimulus (2nd) is slowed compared to the decay time course evoked by the conditioning stimulus (1st). c and d: a similar result is observed in the presence of elevated (10 mM) [Ca2+]o. e: superimposition of the decay times of the IPSC evoked in 2.5 and 10 mM [Ca2+]o reveals that elevation of [Ca2+]o slows down both the decay time of conditioning- and test-stimulus-evoked IPSCs. Notice that the decay time course of the IPSC following the test stimulus in control (2.5 mM) [Ca2+]o is identical to the decay time course of the IPSC following the control stimulus in the presence of elevated (10 mM) [Ca2+]o. B: relationship between the decay time m = (fast + slow) and the interstimulus interval reveals that both the duration of the interstimulus interval and the [Ca2+]o contribute to the decay constant of the IPSCs. #, significant difference between the IPSC decay time constants evoked in control and high [Ca2+]o; *, significant differences between the IPSC decay time constant evoked by a conditioning and test stimulus in control and high [Ca2+]o. C: relationship between the proportional "weight" of the Aslow in the total amplitude of the IPSC (Afast + Aslow) and the inter-stimulus interval reveals the fraction of the amplitude of the "slow" component. Aslow depends on both the inter-stimulus interval and on [Ca2+]o. #, significant differences between the "weight" of the slow component of the amplitude of the IPSC evoked by conditioning stimuli in control and high [Ca2+]o. *, significant differences between the "weight" of the slow component of the amplitude of the IPSC evoked by a conditioning and test stimulus in control and high [Ca2+]o.

The increased amplitude of the test-stimulus-evoked IPSCs may result from either the facilitation of GABA release from presynaptic terminals or an increase in the postsynaptic response due to activation of additional postsynaptic receptors. Alternatively both of these mechanisms may be involved. However, the fact that the amplitude of the IPSC evoked by a test stimulus (I₂) increases when compared to the amplitude of the IPSC evoked by the conditioning stimulus (I₁), leads to the conclusion that the conditioning stimulus activates only a fraction of the available postsynaptic receptors.

Effect of elevation of [Ca²⁺]_o on the paired-pulse evoked IPSCs

The results of the paired-pulse experiments suggest that a conditioning stimulus shock might not activate all the available postsynaptic receptors, i.e. the postsynaptic receptors are nonsaturated. The amplitude of the postsynaptic responses at nonsaturated synapses depends on the neurotransmitter concentration released from the presynaptic terminal, and the amount of neurotransmitter released from presynaptic terminals depends on the extracellular calcium concentration (Creager et al. 1980; Walmsley et al. 1998). Therefore changing the external calcium concentration should result in changes in the amplitude of the IPSCs. This was tested in 23 neurons. Elevation of external calcium concentration 10 mM increased both the amplitude (142 ± 22%, P < 0.05) and the mean decay time constant (_m = 75 ± 7 ms, n = 12; P < 0.05) of the conditioning-stimulus shock-evoked IPSCs (Fig. 3, A, c and d, B, and C). The data indicate that in the presence of a high external calcium concentration, the conditioning and test IPSCs have slower decay times when compared to IPSCs decay times in control external calcium levels (Fig. 3, A-C). When the normalized IPSCs evoked by the conditioning and test stimuli at control and high calcium concentrations are superimposed the decay time courses are identical, indicating that calcium elevation and the test stimulus have the same effect on the decay time course (Fig. 3Ae). This observation suggests that slowing of the decay kinetics of the IPSCs by paired-pulse stimulation or elevation of extracellular [Ca²⁺]_o may be due to a similar cellular mechanism (Fig. 3Ae).

Prolongation of the decay time resulting from both paired-pulse stimulation and elevation of extracellular calcium concentration suggests that accumulation of neurotransmitter in the synaptic cleft might be responsible for this phenomenon. Repetitive stimulation should therefore slow the decay time even more than paired-pulse stimulation. Increasing the number of stimulus shocks (at 30 Hz) in the presence of high external calcium demonstrated that the duration of the IPSC decay time depends on the number of stimulus shocks (n = 5, Fig. 4A). The superimposed and normalized decay times evoked by the last stimulus in the series indicates that the decay time is slowed by increasing the number of stimuli (Fig. 4B). The means of five experiments graphed in Fig. 4C illustrates that increasing the number of stimuli results in an increase in the length of the decay time constant. These results support the hypothesis, that the mechanism responsible for slowing the decay time of the IPSCs results from accumulation of neurotransmitter in the synaptic cleft.

View larger version (20K): [in this window] [in a new window]   Fig. 4. A: effect of varying the number of stimuli on the amplitude and decay time of IPSCs in the presence of 10 mM [Ca2+]o. B: superimposed decay times of the IPSCs reveal that the decay time of the IPSCs slows down with each consecutive stimulus. C: relationship between the decay time constant (m) and the number of stimuli.

Tetanic-stimulus-evoked IPSP/Cs

The paired-pulse and multiple stimulation experiments demonstrate that inhibitory synapses in rNST are changed as a result of accumulation of neurotransmitter in the synaptic cleft. Because neural discharge patterns in gustatory afferent fibers consists of trains or bursts of action potentials (Frank et al. 1988; Ogawa et al. 1968, 1974), it is possible that the in vivo input to the rNST results in similar changes in synaptic activity. In adult animals, we have already shown that tetanic stimulation results in summation of IPSP/Cs amplitudes (Grabauskas and Bradley 1998, 1999). Summation also occurred in postnatal animals. The amplitude of the tetanic-stimulus-elicited IPSPs reached a maximal level after two to eight stimuli. Further stimulation resulted in a relatively sustained IPSP amplitude (Fig. 5A). After termination of the tetanic stimulus, the IPSP/Cs decayed back to the resting membrane potential with a prolonged time courses (Fig. 5B). In addition, the decay time of the tetanic-stimulus-evoked IPSP/C was more complex than that evoked by a single stimulus shock. In 72% of the neurons, a 1-s tetanic stimulus at 30 Hz resulted in IPSP/Cs with a decay time course that could be fitted with two or three exponentials (30-70, 150-400, and >2,000 ms). The remaining neurons (18%) had an IPSP/C decay time course that was initially sustained after termination of the stimulus before decaying back to the resting level with an S-shaped time course (Fig. 5B, arrow).

View larger version (12K): [in this window] [in a new window]   Fig. 5. A: tetanic stimulation at 30 Hz results in summation of the amplitudes of individual inhibitory postsynaptic potentials (IPSPs) evoked by the 1st 3 stimuli. Later stimulation does not result in an increase of the amplitude of the IPSP. However, the decay time of the amplitude decreases with each consecutive stimulus. B: tetanic stimulation at 50 Hz (bar) results in sustained hyperpolarization. After termination of the tetanic stimulus, the membrane potential decays back to the resting membrane potential. The decay time course has an S shape (arrow), where the amplitude of the IPSP is sustained after termination of the tetanic stimulus before it starts to decay exponentially to the resting membrane potential. C: tetanic stimulation at 20 Hz results in sustained hyperpolarization. The IPSP decays exponentially back to the resting membrane potential after termination of the tetanic stimulus. Tetanic stimulation of the same neuron at 50 Hz results in initial hyperpolarization of the postsynaptic neuron that later responds with an increasing hyperpolarizing amplitude (arrow). After termination of the tetanic stimulus the decay time course has an S shape.

Two neurons responded to tetanic stimulation with an increase in the IPSP amplitude. At low tetanic frequencies (10-30 Hz), these two neurons responded with sustained IPSP amplitude, but at higher tetanic frequencies (30-50 Hz) responded with an initial sustained hyperpolarization followed with an increasing hyperpolarizing amplitude (Fig. 5C).

Tetanic stimulation resulted in IPSP/Cs with sustained amplitudes that were independent of stimulus frequency except at relatively low tetanic stimulus frequencies (10 Hz) when the amplitudes of the IPSPs failed to reach a sustained level (Fig. 6, A and Ba, n = 12, P < 0.05). In contrast, the duration of the decay time was dependent on stimulus frequency. By measuring the time required for the amplitude of the IPSP to decay to half amplitude (T_50%), it was possible to analyze the influence of stimulus frequency on the IPSP time course. T_50% significantly increases with increasing tetanic stimulus frequency (Fig. 6, A and Bb, P < 0.05). Interestingly, at high stimulus frequencies (50 Hz), the IPSP amplitude is sustained briefly after termination of the tetanic stimulus before it starts to decay back to the resting membrane potential (Fig. 6A). These results suggest that tetanic stimulation results in an increase in neurotransmitter concentration in the synaptic cleft that ultimately reaches a saturation level. At relatively low tetanic stimulus frequencies (5-20 Hz), the GABA concentration builds up to a saturating concentration resulting in an IPSP with an exponential decay time. At high frequencies, the neurotransmitter concentration in the synaptic cleft probably exceeds the saturation level and as a result the IPSC amplitude is sustained even after termination of the tetanic stimulus (Figs. 5B and 6A).

View larger version (18K): [in this window] [in a new window]   Fig. 6. A: tetanic stimulation (bar) at different frequencies (10, 30, and 50 Hz) results in hyperpolarizing IPSPs with different amplitudes. The decay time course of the IPSPs is different depending on the frequency of the tetanic stimulus. Notice that tetanic stimulation at 10 Hz did not produce sustained hyperpolarization of the postsynaptic neuron. Ba: relationship between amplitude of the tetanic-stimulus-evoked IPSPs and frequency of the tetanic stimulus. *, significant difference between the amplitude of a tetanic stimulus evoked at 5 and 30 Hz when the IPSP/C is at a sustained level. Bb: relationship between the half time of the decay of the IPSP amplitude and the frequency of the tetanic stimulus. C: voltage-clamp recording of IPSCs evoked by single stimulus shock (arrow head) and tetanic-stimulation (bar)-evoked outward current in control conditions and in the presence of the GABAB receptor antagonist 2-hydroxysaclofen. Superimposed traces indicate that addition of the GABAB receptor antagonist had no effect on the amplitudes and decay time courses of both the IPSC and the outward current evoked by the tetanic stimulus (bottom).

Previously we reported the presence of postsynaptic GABA_B receptors within the rNST (Grabauskas and Bradley 2001). Activation of "slow" metabotropic GABA_B receptors can prolong decay times (for review, see Kerr and Ong 1995). However, application of the GABA_B receptor antagonist 2-hydroxysaclofen (200 µM) had no significant effect on the amplitudes or decay times of either single- or tetanic-stimulus shock-evoked IPSP/Cs (n = 7; Fig. 6C). Thus GABA_B receptors are not involved in slowing of the IPSP/Cs decay time.

Goda and Stevens (1994) suggested that accumulation of intracellular Ca²⁺ might extend neurotransmitter release. We tested whether Ca²⁺ accumulation in the presynaptic terminal might be responsible for the prolongation of the decay time during repetitive stimulation by using the ATP-dependent Ca²⁺ pump blockers tBuBUQ and thapsigargin (Fossier et al. 1999). A 1- to 3-h incubation of the slices in 10 µM tBuBUQ or 10 µM thapsagargin did not abolish activity-dependent prolongation of the decay time of the IPSP/Cs (n = 5 and 6 respectively, data not shown), indicating that buffering of Ca²⁺ in pre-synaptic stores was not involved in prolongation of the decay time.

Effect of elevation of [Ca²⁺]_o on tetanic-stimulus-evoked IPSCs

The results support the hypothesis that inhibitory synapses in the rNST are not saturated by a single stimulus shock and only high-frequency stimulation results in sufficient accumulation of neurotransmitter in the synaptic cleft to activate all available postsynaptic receptors. The variability of the _slow also suggests that neurotransmitter diffusion but not binding-unbinding kinetics defines the slow phase of IPSP/Cs. To further test this possibility, we manipulated [Ca²⁺]_o to facilitate neurotransmitter release from the presynaptic terminal. Elevation of [Ca²⁺]_o 10 mM had no significant effect on the amplitudes of tetanic-stimulus-evoked IPSCs (104 ± 6% of the control amplitudes; P = 0.4). However, elevation of [Ca²⁺]_o slowed the decay time of the tetanic stimulus evoked IPSP/Cs. Moreover, in 5 of 29 neurons tested, the exponential decay time course of the tetanic-stimulus-evoked IPSP/C was converted to an S shape (note difference in decay shape indicated by arrowheads in Fig. 7, A and B). The effect of [Ca²⁺]_o on the decay time course was reversible (Fig. 7C). This result suggests that transients of neurotransmitter in the synaptic cleft but not neurotransmitter binding-unbinding defines the decay time of the tetanic stimulus evoked IPSP/Cs.

View larger version (11K): [in this window] [in a new window]   Fig. 7. A: single-stimulus shock- and tetanic-stimulus (30 Hz)-evoked IPSPs recorded in control conditions (2.5 mM [Ca2+]o). After termination of the tetanic stimulus, the neuron membrane potential decays back with an exponential time course (). B: elevation of [Ca2+]o to 10 mM increases the amplitude and duration of decay time of the single-stimulus-evoked IPSP and slows the decay time course of the tetanic-stimulus-evoked IPSP, which has an S-shape decay (). However, elevation of [Ca2+]o has no effect on the amplitude of the tetanic-stimulus-evoked IPSP. C: the effect of "high" [Ca2+]o is reversible.

Relationship between paired-pulse and tetanic-stimulus-evoked IPSC amplitudes

The data indicate that single-stimulus shock-evoked IPSCs do not saturate the postsynaptic receptors and that repetitive stimulation results in accumulation of neurotransmitter in the synaptic cleft. We reasoned that variability in the ratio of the IPSC amplitudes evoked by test and conditioning stimuli (I₂/I₁) might be due to differences in saturation of the postsynaptic receptors by the amount of neurotransmitter released during stimulus shock. For synapses in which the conditioning stimulus results in high saturation of the postsynaptic receptors, the remaining fraction of available receptors would be small, resulting in little increase in the amplitude produced by the test stimulus. On the other hand, if the conditioning stimulus activates a small fraction of the available receptors, a test stimulus would produce a significant increase in IPSC amplitude. Tetanic stimulation at high frequency saturates all available receptors and the ratio between the amplitude of the IPSC evoked by a single stimulus shock and the amplitude of the IPSC evoked by tetanic stimulus shock (I₁/I_tet) is an indicator of a the degree of postsynaptic saturation produced by a single stimulus shock. The relationship between the facilitation evoked by paired-pulse (I₂/I₁) and the degree of synaptic saturation resulting from tetanic stimulation (I₁/I_tet) reveals that synapses with high initial release probability (value of I₁/I_tet ~ 1) have a I₂/I₁ ratio that is close to 1. In contrast those with a low initial release probability (value of I₁/I_tet  1) demonstrate facilitation (I₂/I₁  1, Fig. 8). Also when the external calcium concentration is raised the degree of postsynaptic saturation is increased and the probability of facilitation by paired-pulse stimulation is lowered (Fig. 8).

View larger version (22K): [in this window] [in a new window]   Fig. 8. Comparison of the amplitudes of the paired-pulse- and tetanic-stimulus-evoked IPSCs. Top left: a paired-pulse-evoked IPSC is facilitated with a ratio of I2/I1 < 1. Right: a tetanic-stimulus-evoked IPSC illustrates that the amplitude of the IPSC (Itet) is greater than the IPSC evoked by the paired-pulse stimulation. When these results are plotted (bottom) for a number of different neurons in control and increased external Ca2+, the extent of facilitation (I2/I1) of the IPSC amplitude evoked by the test stimulus (I2) is inversely related to the degree of saturation of the postsynaptic receptors by the conditioning stimulus (I1).

Effect of temperature and diazepam on IPSCs

To test the hypothesis that neurotransmitter transients in the synaptic cleft and not receptor opening and closing kinetics define the decay phase of the IPSP/Cs, we used a benzodiazepine to modulate the GABA_A receptor. Benzodiazepines increase the frequency of channel opening, which in turn increases the amplitude and length of the decay time of both spontaneous and evoked IPSCs (MacDonald and Olsen 1994; Poncer et al. 1996). Thus application of the benzodiazepine, diazepam, should increase the decay time constant if receptor kinetics define the decay time of the IPSCs but will not produce an effect if neurotransmitter concentration defines the decay time.

Bath application of 1 and 4 µM diazepam did not increase the decay constant in 66% (10/15) of single-stimulus shock- and 100% (10/10) of tetanic-stimulus-evoked IPSCs when measured at room temperature. However, when measured at 32°C, the decay time of the single-shock-evoked IPSCs became shorter in 11 of 15 neurons (Fig. 9A). Moreover, the effect was concentration dependent. Superfusion of 1 and 4 µM diazepam at 32°C prolonged the decay time of the IPSC from 38 ± 4 to 46 ± 4 ms (P = 0.1) and 49 ± 4.5 ms (P < 0.05), respectively (Fig. 9B). However, no effect of diazepam was observed on the decay kinetics of tetanic stimulus evoked IPSPs (Fig. 9C, a and b, P = 0.32) indicating that decay kinetics of the single-stimulus-evoked IPSCs are defined by the opening-closing kinetics of the GABA_A receptors (Fig. 9, B and Cc), whereas the decay time of the tetanic-stimulus-evoked IPSCs is defined by the rate of the neurotransmitter clearance from synaptic cleft (Fig. 9Cd).

View larger version (22K): [in this window] [in a new window]   Fig. 9. Superimposed IPSCs to show that the temperature dependent effect of diazepam. A: Diazepam has no effect on the decay time of the IPSC at room temperature (22°C, top); however, at 32°C, diazepam increases the decay time of the IPSC (bottom). B: relationships between the decay time constant of the single-stimulus shock-evoked IPSC (m) and the concentration of the extracellular diazepam at 22°C () and 32°C (). *, significant difference compare to control. C, a and b: diazepam has a different effect on the decay times of the single stimulus shock () and tetanic stimulus () evoked IPSC at 32°C. Superimposed traces of the single-stimulus shock-evoked IPSCs and tetanic-stimulus-evoked outward currents show that diazepam prolongs the decay time of the single stimulus shock (c); however, it has no influence on the decay time course of a tetanic stimulus evoked current (d).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we have demonstrated that the synaptic strength of inhibitory synapses in the rNST are modified by prior activity. Specifically, the amplitude and decay time of IPSP/Cs depend on previous synaptic activity resulting in accumulation of neurotransmitter in the synaptic cleft. In addition, the transient concentration of GABA in the synaptic cleft defines the decay phase of the tetanic stimulus evoked IPSCs while neurotransmitter binding-unbinding kinetics from postsynaptic receptors defines the fast decay phase of single stimulus shock evoked IPSCs.

Activity-dependent increase of synaptic strength

Stimulus-evoked IPSP/Cs result from vesicular release of neurotransmitter from one or several presynaptic sites to bind receptors on the surface of the postsynaptic neuron. The number of postsynaptic receptors activated defines the amplitude of IPSCs. However, the concentration and the time course of GABA in the synaptic cleft define the number of GABA_A receptors that open on the postsynaptic neuron. For synapses where neurotransmitter concentration is high enough to activate all available receptors, the IPSC amplitude is determined by the number and intrinsic properties of postsynaptic receptors. On the other hand, when all the receptors are not activated, the IPSC amplitude is determined by both the amount of GABA released from the presynaptic terminal and the intrinsic properties of the postsynaptic GABA_A receptors (for review, Walmsley et al. 1998). Data from the present study demonstrate that a single stimulus shock does not saturate all the available postsynaptic receptors at the GABAergic synapse in the rNST. However, high-frequency stimulation does saturate the postsynaptic receptors. These conclusions are supported by the evidence that high-frequency stimulation produced IPSCs that have sustained (i.e. saturated) amplitudes that were unaffected by [Ca²⁺]_o manipulation, whereas single-stimulus shock-evoked IPSP/Cs amplitudes were affected by [Ca²⁺]_o manipulation. During tetanic stimulation, each consecutive stimulus produced an IPSC that summed with the previous IPSC until the amplitude of IPSPCs reached a sustained or saturated level. Analysis of the relationship (see Fig. 8) between the coefficient of facilitation (I₂/I₁) and the coefficient of saturation (I₁/I_tet) indicates that synapses that show evidence of facilitation (I₂ > I₁) have small coefficients of saturation (I₁/I_tet  1), suggesting that at this type of synapse a single-stimulus shock activates a small fraction of postsynaptic GABA receptors. Thus the neurotransmitter released during a test stimulus can activate an additional number of postsynaptic receptors. For synapses in which the control (I₁) stimulus activated all available receptors (I₁ = I_tet), the neurotransmitter released by the test stimulus (I₂) only activated receptors that remained unbound with the neurotransmitter released from the previous stimulus (i.e. I₁ = I₂). Thus synapses that are facilitated (I₂/I₁ > 1) during paired-pulse stimulation are unsaturated. Similar results have been reported by Charpier et al. (1995) analyzing activity-dependent changes in goldfish Mauthner cell in which facilitation was more frequent in weak rather than in "normal" inhibitory connections. The weak connections were more sensitive to drugs that enhance synaptic release.

Biexponential decay of IPSC

For 30 of 36 neurons tested, single-stimulus shock evoked IPSCs with a biexponential decay time course, whereas the decay time of the other six neurons could be fitted with a single exponential. However, using either paired-pulse stimulation or alteration of [Ca²⁺]_o, all the neurons had a biexponential decay time course. It is possible that all the neurons in fact had a biexponential decay time course but for a small subset _fast and _slow were similar and could not easily be separated. Other investigators also had difficulty separating the two exponentials reporting that when the ratio _slow and _fast is <5, it becomes increasingly difficult to resolve the exponential components (Dempster 1993; Roepstorff and Lambert 1994).

Edwards et al. (1990) and Pearce (1993) suggested that the multiple decay time constants of IPSCs result from the contribution of receptors with different kinetic and pharmacological properties. According to these authors, the fast component of the decay phase is mediated by rapidly unbinding receptors, whereas the slow component is mediated by a slowly unbinding receptor subtype. An alternative explanation was proposed by Jones and Westbrook (1995) investigating channel gating in outside-out patches of cultured hippocampal neurons. They suggested that long-channel closed states before reopening contributed to the slow component of the IPSC and that the fast decay component is due to rapid desensitization of the postsynaptic receptors. However, none of these explanation account for the activity-dependent increase of the decay time of the IPSP/C in the present study because they would predict changes in the amplitudes of the decay time constants but not the values of the time constant itself.

Most earlier electrophysiological studies of synaptic transmission in the CNS are based on the assumption that neurotransmitter release is a single process. However, there is evidence that at least two distinct components are involved: a fast, synchronous component and slower asynchronous component, which indicate the existence of two Ca²⁺-buffering systems at the presynaptic terminal (Goda and Stevens 1994). According to this explanation, repetitive stimulation results in accumulation of Ca²⁺ in a high-affinity buffering system at the pre-synaptic terminal. After termination of the stimulus, Ca²⁺ clearance from this system accounts for the second component of release. However, the lack of an effect of the ATP-dependent Ca²⁺ pump blockers tBuBHQ or thapsigargin on the decay phase of the IPSP/Cs indicates that internal stores of Ca²⁺ are not responsible for the activity-dependent prolongation of the decay phase of the IPSP/Cs.

Rossi and Hamann (1998) suggested that spillover from neighboring but not directly connected axon terminals is responsible for the slow inhibitory transmission at Golgi to granule cell synapses. They demonstrated that high-affinity GABA receptors containing ₆ subunit located extra-synaptically might mediate the slow component of the IPSP/Cs. Our data also indicate that GABA spillover might take place in rNST. Two neurons in control conditions and four in the presence of elevated [Ca²⁺]_o responded to tetanic stimulation with an increase in the IPSP amplitude. These increases in amplitudes of the hyperpolarizing potentials can be best explained by the spread of GABA to adjacent postsynaptic sites (see Fig. 5C). Repetitive activation of presynaptic GABA terminals might create a "cloud" of GABA in the extra-synaptic space. This "cloud" then activates extra-synaptic receptors and also decreases the diffusion gradient of neurotransmitter allowing buildup of GABA in the synaptic cleft. The delay in removal of GABA from the synaptic cleft then increases the decay time of the postsynaptic GABA receptors (Roepstorff and Lambert 1994). The buildup of neurotransmitter in the synaptic cleft during tetanic stimulation is also demonstrated by the increase in the amplitude, the increase in the length of the sustained phase, and the prolongation of the decay time of the IPSP/Cs.

Effect of diazepam on the IPSC kinetics

GABA_A receptors are pentameric structures made up of combination of different subunit types. So far six isoforms (, , , , , ) and 15 subtypes have been discovered. The subunit combination of a particular GABA_A receptor determines its pharmacological properties. Molecular analysis has revealed that the benzodiazapine binding site is located on the  subunit; however, the  subunit is required also (Sieghart 1995).

When benzodiazepines bind to the GABA_A receptor, they increase the affinity of the receptor for GABA; this in turn results in an increase in the frequency of opening of the receptor without changing their mean open time and burst duration (Rogers and Twyman 1994). Benzodiazepine binding to the GABA_A receptor leads to a leftward shift in the GABA dose-response relationship resulting in potentiation of the GABA response.

The predominant effect of benzodiazepines on GABAergic IPSCs is prolongation of the time course rather than an increase in current amplitudes. Because the transmitter concentration in the synaptic cleft falls rapidly to a basal level, benzodiazepine receptor agonists prolong the decay time of both evoked and spontaneous miniature IPSCs in the dentate gyrus (Otis and Mody 1992) and CA3 region of the hippocampus (Poncer et al. 1996) and in the visual cortex (Perrais and Ropert 1999). Our observation that diazepam increases the duration of single-stimulus shock-evoked IPSCs indicates that rNST neurons also express benzodiazepine receptors. However, the sensitivity to diazepam is absent at room temperature for both single shock- and tetanic-stimulus-evoked IPSCs. Temperature is known to affect the kinetics of IPSCs and is believed to reflect changes in the gating kinetics of the GABA receptors as has been reported in various CNS locations (Collingridge et al. 1984; Nusser et al. 1997; Otis and Mody 1992; Poncer et al. 1996; Strecker et al. 1999). Our results support the hypothesis that the decay kinetics of single-stimulus shock-evoked IPSCs are defined by the GABA unbinding kinetics from the GABA_A receptor, whereas the absence of diazepam sensitivity on tetanic-stimulus-evoked IPSCs indicates that neurotransmitter clearance or diffusion from the synaptic cleft defines the decay time characteristics.

Functional significance

The results of this study extend our previous work in which we reported short- and long-term changes at the GABAergic synapse in the rNST of mature animals (Bradley and Grabauskas 1998; Grabauskas and Bradley 1998, 1999). In addition, we have reported profound changes of kinetics and pharmacological properties at inhibitory synapse in the rNST during maturation (Grabauskas and Bradley 2001). All of these results indicate that GABAergic synapses could play an active role in shaping gustatory responses in the rNST. Data from the present study indicate that the characteristics of short-term plasticity of rNST GABAergic synapse are similar in both newborn and mature animals. However, there are some differences, the most important of which is that high-frequency tetanic-stimulus-evoked IPSP/Cs have sustained amplitudes after termination of the tetanus in the younger animals.

In other CNS areas short- and long-term plasticity are described as memory-related processes (Fisher et al. 1997; Malenka and Nicoll 1999) so that it is conceivable that GABAergic synapse plasticity is involved in taste-related learning phenomenon. Short-term plasticity could provide a flexible and adaptive mechanism that might contribute to processing of temporal information. The activity-dependent plasticity at GABAergic synapses in the rNST might also be important in the encoding of submodalities of taste responses and also be involved in comparison processes between different kind of stimuli. Moreover, the dynamic range of stimulus frequencies that result in synaptic plasticity (0-70 Hz) corresponds to the breadth of frequencies that travels via afferent gustatory nerve fibers in response to taste stimuli (Ogawa et al. 1974). Thus the frequency-dependent changes demonstrated in vitro are likely to occur in solitary nucleus circuits processing gustatory afferent information.