Kainate receptor-mediated presynaptic inhibition at the mouse hippocampal mossy fibre synapse

Abstract

1. The presynaptic action of kainate (KA) receptor activation at the mossy fibre-CA3 synapse was examined using fluorescence measurement of presynaptic Ca²⁺ influx as well as electrophysiological recordings in mouse hippocampal slices.

2. Bath application of a low concentration (0·2 μM) of KA reversibly increased the amplitude of presynaptic volley evoked by stimulation of mossy fibres to 146 ± 6 % of control (n = 6), whereas it reduced the field excitatory postsynaptic potential (EPSPs) to 30 ± 4 %.

3. The potentiating effect of KA on the presynaptic volleys was also observed in Ca²⁺-free solution, and was partly antagonized by (2S,4R)-4-methylglutamic acid (SYM 2081, 1 μM), which selectively desensitizes KA receptors.

4. The antidromic population spike of dentate granule cells evoked by stimulation of mossy fibres was increased by application of 0·2 μM KA to 160 ± 10 % of control (n = 6). Whole-cell current-clamp recordings revealed that the stimulus threshold for generating antidromic spikes recorded from a single granule cell was lowered by KA application.

5. Application of KA (0·2 μM) suppressed presynaptic Ca²⁺ influx to 78 ± 4 % of control (n = 6), whereas the amplitude of the presynaptic volley was increased.

6. KA at 0·2 μM reversibly suppressed excitatory postsynaptic currents (EPSCs) evoked by mossy fibre simulation to 38 ± 9 % of control (n = 5).

7. These results suggest that KA receptor activation enhances the excitability of mossy fibres, probably via axonal depolarization, and reduces action potential-induced Ca²⁺ influx, thereby inhibiting mossy fibre EPSCs presynaptically. This novel presynaptic inhibitory action of KA at the mossy fibre-CA3 synapse may regulate the excitability of highly interconnected CA3 networks.

Kainate-type ionotropic glutamate receptors (KA receptors) consist of various combinations of five subunits, that is, GluR5/6/7 and KA1/2 (for reviews, see Hollmann & Heinemann, 1994; Bettler & Mulle, 1995). Despite a widespread distribution within the central nervous system (CNS), the physiological function of the KA receptor remained unknown until recently, since the overlapping abilities of classical ligands (e.g. kainate and 6-cyano-7-nitroquinoxaline, CNQX) to interact with both AMPA and KA receptors had made it difficult to study the function of these receptors in isolated conditions. Recently, novel drugs which specifically act on either KA or AMPA receptors have been developed, and have made it feasible to separate the physiological responses of these receptors (Lerma et al. 1993; Paternain et al. 1995; Chittajallu et al. 1996; Clarke et al. 1997; Lerma et al. 1997). In the hippocampal CA3 region, which is one of the CNS areas expressing the highest density of kainate receptor subunits (Wisden & Seeburg, 1993; Petralia et al. 1994), synaptic release of glutamate was found to elicit slowly decaying synaptic currents by activating postsynaptic KA receptors (Castillo et al. 1997; Vignes & Collingridge, 1997).

On the other hand, earlier morphological studies have suggested the presynaptic localization of KA receptors at the mossy fibre-CA3 synapse. Represa et al. (1987) reported that selective lesion of dentate granule cells with intrahippocampal colchicine injection largely eliminated the KA binding sites in the stratum lucidum of the CA3 region, suggesting that a substantial proportion of the KA-binding sites (receptors) was associated with presynaptic elements of this synapse. Ultrastructural studies using specific antibodies raised against KA receptor subunits also suggested that GluR6/7 was localized on the unmyelinated axons in the CA3 region, possibly mossy fibres (Petralia et al. 1994). Despite accumulating evidence for the presynaptic localization of KA receptors at the mossy fibre-CA3 synapse, the physiological function of the receptors remains to be elucidated. Although Castillo et al. (1997) reported that KA application to the stratum lucidum failed to alter the frequency of miniature EPSCs recorded from CA3 neurons, this does not necessarily exclude other presynaptic modulating effects of KA. In fact, Vignes et al. (1998) have suggested the presynaptic inhibitory effect of KA receptor at this synapse, since a GluR5 selective agonist, (R,S)-2-amino-3-(3-hydroxy-5-tert-butylisoxazol-4-yl)-propionic acid (ATPA), suppressed EPSCs evoked by stimulation of mossy fibres without affecting holding current recorded from CA3 neurones. In this study, we examined the presynaptic action of KA receptors at the mossy fibre-CA3 synapse, using electrophysiological recording techniques, as well as fluorescence measurement of action potential-induced presynaptic Ca²⁺ influx. We found that activation of KA receptors enhanced the excitability of mossy fibres and reduced presynaptic Ca²⁺ influx, and thereby suppressed evoked transmitter release from the terminals. The physiological significance of the presynaptic inhibitory action of KA receptors at the mossy fibre-CA3 synapse is also discussed.

METHODS

Slice preparations and field potential recordings

The methods for preparing the slices and recording field excitatory postsynaptic potentials (EPSPs) induced by mossy fibre stimulation have been reported previously (Kamiya et al. 1996). All experiments were carried out according to the guidelines laid down by the Animal Care and Experimentation Committee of Gunma University, Showa Campus. Briefly, BALB/c mice (10-18 days old) were anaesthetized with ether and decapitated. The brain was quickly removed and immersed in an ice-cold oxygenated standard solution composed of (mM): NaCl, 127; KCl, 1.5; KH₂PO₄, 1.2; MgSO₄, 1.3; CaCl₂, 2.4; NaHCO₃, 26; and glucose, 10. The solution was saturated with 95 % O₂ and 5 % CO₂. Hippocampi were dissected out and transverse sections (0.3-0.4 mm thick) were prepared. They were incubated in standard Ringer solution at 32°C for at least 40 min, and then transferred into an observation chamber which was continuously superfused at a rate of approximately 2 ml min⁻¹ with the standard solution. In some experiments, Ca²⁺-free solution was used as indicated in the text. Electrical stimuli were delivered every 5 min (for simultaneous recordings of presynaptic Ca²⁺ transients and field EPSPs) or every 10 s (for field potential recordings alone) through a concentric bipolar stimulating electrode, and field potentials were recorded with a glass microelectrode of about 10 μm tip diameter filled with the extracellular solution. All recordings were made at 24–26°C.

Loading of Ca²⁺ indicator dye and fluorescence recordings

The relatively low-affinity Ca²⁺ indicator rhod-2 was used to load presynaptic structures of mossy fibre-CA3 synapses (Kamiya & Ozawa, 1999; see also Regehr & Tank, 1991). Axons and presynaptic boutons of the mossy fibre pathway were labelled by local pressure ejection of the membrane-permeable Ca²⁺ indicator rhod-2 AM into the extracellular space of the stratum lucidum, where the axon bundle is located. The labelling solution contained 0.1 mM rhod-2 AM dissolved in dimethyl sulphoxide (DMSO) and 2 % Pluronic F-127. The membrane-permeable rhod-2 AM is expected to enter the axons and diffuse or be transported to the presynaptic terminals after conversion into the membrane-impermeable form of rhod-2 by intracellular esterase. About 2 h after ejection, fluorescence (excited at 510–560 nm and monitored at above 580 nm) emerging from an area with a diameter of about 100 μm in the stratum lucidum approximately 500 μm distal from the ejection site was detected with a single photodiode. The output of the photodiode was I–V converted and filtered at 200 Hz with an eight-pole Bessel filter (FLA-01, Cygnus Technology, Delaware Water Gap, PA, USA) to improve the signal-to-noise ratio. The amplitude of the Ca²⁺ transient (ΔF) was measured as the difference between the resting fluorescence level (F) and the peak level after an electrical stimulation. The magnitude of the Ca²⁺ transient was expressed as relative fluorescence change (ΔF/F). Thus the signal size was corrected for dye bleaching and illumination irregularities, and represents the volume-averaged Ca²⁺ concentration change throughout all the presynaptic terminals loaded with rhod-2 in the recorded area. F was corrected for the background fluorescence of the tissue. The fluorescence signals and extracellular field potentials were digitized with a 12-bit A/D converter (Digidata 1200A, Axon Instruments, Foster City, CA, USA) and acquired at 10 kHz using pCLAMP software (Axon Instruments).

Whole-cell patch-clamp recordings

In separate experiments, recordings were made from visually identified dentate neurons in the stratum granulosum or CA3 neurones in the stratum pyramidale using a whole-cell patch-clamp technique. Membrane potentials (current-clamp experiments) or currents (voltage-clamp experiments) were recorded with an Axopatch-1D amplifier (Axon Instruments). In current-clamp experiments, patch pipettes were filled with an internal solution (pH 7.3) containing (mM): potassium gluconate, 140; KCl, 20; EGTA, 0.2; MgCl₂, 2.0; Hepes, 10; and Mg-ATP, 2.0. The internal soution used in voltage-clamp experiments contained (mM): caesium gluconate, 150; EGTA, 5.0; CaCl₂, 1.0; Hepes, 10; Mg-ATP, 2.0; and lidocain N-ethyl bromide quaternary salt (QX-314), 5.0 (pH 7.3). Electrode capacitance was compensated, but series resistance was not compensated in this study. The resistance of the pipette was 4–8 MΩ when filled with the internal solution. The access resistance was typically 15–30 MΩ immediately after obtaining whole-cell recordings, and was not allowed to vary by more than 10 % during the course of the experiment.

Chemicals

Drugs used in this study were: 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), (2S,1′R,2′R,3′R)-2-(2,3-dicarboxycyclo-propyl)glycine (DCG-IV), and (2S,4R)-4-methylglutamic acid (SYM 2081), all from Tocris Cookson, Bristol, UK; rhod-2 AM (Dojindo Laboratories, Kumamoto, Japan); Pluronic F-127 (Molecular Probes, Eugene, OR, USA); 1-(4-aminophenyl)-4-methyl-7,8-methyl-enedioxy-5H-2,3-benzodiazepine (GYKI 52466) hydrochloride (RBI, Natick, MA, USA); and QX-314 and DMSO from Sigma.

Statistical analysis

All values are expressed as means ±s.e.m. Statistical analysis was performed using the Mann-Whitney U test, and P < 0.05 was accepted for statistical significance.

RESULTS

Kainate-induced enhancement of the presynaptic volley at the mossy fibre-CA3 synapse

To examine the effect of KA receptor activation on synaptic transmission at the mossy fibre-CA3 synapse, we first studied the effect of a low concentration of KA (0.2 μM), which preferentially activates KA receptors while minimally affecting AMPA receptors, on the field potentials evoked by stimulating mossy fibres. The mossy fibres were stimulated at the stratum granulosum of the dentate gyrus, and the evoked field potentials were recorded through a glass microelectrode placed in the stratum lucidum of the CA3 region (Fig. 1A). As shown in Fig. 1A, the field potentials recorded under control conditions consisted of two components, i.e. a presynaptic volley component (initial biphasic deflection, •) and a subsequent field EPSP component (^). The field EPSPs were abolished almost completely by application of 1 μM DCG-IV, a group II-selective metabotropic glutamate receptor (mGluR) agonist, confirming that we selectively stimulated mossy fibres without stimulating the axon collaterals of CA3 neurons (Kamiya et al. 1996; Yokoi et al. 1996; see also Henze et al. 1997). Further application of 0.5 μM tetrodotoxin (TTX) abolished the initial biphasic deflection, suggesting that this component of field potentials could be presynaptic volley potentials evoked by stimulation of mossy fibres (Henze et al. 1997). Application of a low concentration of KA (0.2 μM, 5 min) reversibly increased the amplitude of presynaptic volley potentials (to 146 ± 6 % of control), whereas it reduced the amplitude of field EPSPs (to 30 ± 4 % of control, n = 6, Fig. 1B and C).

Figure 1. KA-induced enhancement of presynaptic volleys at the mossy fibre-CA3 synapse.

A, schematic diagram showing experimental arrangement. A stimulating electrode (Stim) was placed on the stratum granulosum of the dentate gyrus, and the field potentials were recorded (Rec, recording electrode) from the stratum lucidum of the CA3 region. Traces are representative records of the field potentials recorded before and during application of 1 μM DCG-IV, a group II-selective metabotropic glutamate receptor (mGluR) agonist, and 0.5 μM tetrodotoxin (TTX). Note that the initial biphasic component (•, presynaptic volley) was not affected by DCG-IV, whereas the following negative component (^, field EPSP) was completely abolished. B, effects of KA (0.2 μM) on field potentials evoked by mossy fibre stimulation. Traces are representative field potentials recorded under the control conditions (left, continuous trace) and during application of KA (centre, dotted trace). The two traces are superimposed for comparison in the right panel. Note that the presynaptic volley potential was enhanced, whereas the field EPSP was depressed by KA application. C, the relative amplitudes of the presynaptic volley (•) and field EPSP (^), with those before KA application as references, are plotted against time (means ±s.e.m., n = 6). KA (0.2 μM) and DCG-IV (1 μM) were applied during the period indicated.

It has been demonstrated that CA3 pyramidal neurons have extensive excitatory collaterals within CA3 region (Weisskopf & Nicoll, 1995; Henze et al. 1997). Therefore it seems to be possible that polysynaptic activation subsequent to orthodromic firing of CA3 neurons could be contaminated to the observed field EPSP component to some degree. To minimize the polysynaptic components in the field EPSPs, we used relatively low extracellular K⁺ solution (e.g. 2.7 mM, see Methods) to increase the threshold of CA3 neurons and tried to keep the stimulus intensity such that it did not evoke population spikes during the control period, as judged by the smooth decay phase of field EPSPs (see Fig. 1A and B). We also tested to see the effect of KA in the high Ca²⁺/Mg²⁺ solution (both at 4 mM), in which the threshold of CA3 neurons is expected to be elevated. As shown in Table 1, the relative changes in presynaptic volley amplitudes (to 135 ± 9 % of control) and field EPSPs (to 25 ± 3 %, n = 3) was not different from those observed in standard solution (P > 0.3, Mann-Whitney U test), suggesting that the relative contribution of polysynaptic activation was minimal in our experimental conditions.

Table 1.

Comparison of the effect of KA in different solutions

Solution	Standard	High Ca2+/Mg2+	High Ca2+/Mg2+ with picrotoxin
No. of experiments	6	3	3
Presynaptic volley (% of control)	146 ± 6	135 ± 9	142 ± 2
Field EPSP (% of control)	30 ± 4	25 ± 3	36 ± 10

The effects of application of 0.2 μM KA for 5 min are shown relative to control values (means ±s.e.m.). The standard solution contained 2.4 mM Ca²⁺ and 1.3 mM Mg²⁺. The concentrations of both Ca²⁺ and Mg²⁺ were raised to 4 mM for the high Ca²⁺/Mg²⁺ solution. The third solution was the high Ca²⁺/Mg²⁺ solution (both at 4 mM) containing 100 μM picrotoxin. The difference is not significant between any two groups (Mann-Whitney U test, P > 0.05).

We also examined the effect of 0.2 μM KA in the presence of 100 μM picrotoxin, to see the possible involvement of changes in IPSPs (inhibitory postsynaptic potentials) evoked disynaptically (Clarke et al. 1997; Cossart et al. 1998; Frerking et al. 1998). To avoid bursting discharges, the concentrations of both Ca²⁺ and Mg²⁺ were raised to 4 mM in these experiments. Again, relative changes in presynaptic volleys (to 142 ± 2 % of control) and field EPSPs (to 36 ± 10 %, n = 3) were not different from those in standard solution (P > 0.3, Mann-Whitney U test). These results suggest that application of 0.2 μM KA mainly affects the monosynaptic EPSPs evoked by stimulation of mossy fibres.

To isolate the presynaptic volley from the postsynaptic element, we next examined the effect of KA in Ca²⁺-free solution (Fig. 2A). KA at 0.2 μM also reversibly increased the amplitude of presynaptic volley potentials to almost the same degree as in Fig. 1 (to 140 ± 9 % of control, n = 6, Fig. 2B, •). The effect of KA on the presynaptic volley potential is mediated by activation of KA receptors, because the prior application of SYM 2081, which desensitizes KA receptors, antagonized the KA effect on the presynaptic volley. SYM 2081 (1 μM) alone did not affect the presynaptic volley (101 ± 1 % of control), whereas application of 0.2 μM KA in the presence of 1 μM SYM 2081 caused a smaller increase in presynaptic volley amplitude (to 110 ± 9 % of control, n = 6, Fig. 2B, ^). The difference in the degree of the KA-induced increase in the presynaptic volley in the presence and absence of SYM 2081 was statistically significant (P < 0.05, Mann-Whitney U test). In a separate set of experiments, we examined the effects of KA in the presence of the AMPA-selective antagonist GYKI 52466 (Donevan & Rogawski, 1993). Application of 0.2 μM KA in the presence of 100 μM GYKI 52466 increased the amplitude of presynaptic volley potentials to 145 ± 25 % of control (n = 4), which was almost equivalent to the value obtained in the absence of this drug. These results strongly suggest that the activation of KA receptor enhances the axonal excitability at the mossy fibre-CA3 synapse.

Figure 2. Effect of KA on the presynaptic volley potentials recorded in the Ca²⁺-free solution.

A, representative traces of presynaptic volleys recorded in the Ca²⁺-free solution before and during KA application, either in the absence or presence of SYM 2081 which selectively desensitizes KA receptors. B, the time course of the presynaptic volley amplitude when KA (0.2 μM) and TTX (0.5 μM) were applied during the periods indicated (means ±s.e.m., n = 6) is shown (•). In another set of experiments, 1 μM SYM 2081 was applied (^) during the period indicated by the open bar (n = 6).

We next examined the effect of higher doses of KA. As shown in Fig. 3A, 0.2 μM KA potentiated (to 154 ± 11 % of control) while 3 μM KA depressed the presynaptic volley potentials (to 67 ± 26 % of control, n = 4, Fig. 3C). It seems likely that 3 μM KA caused large depolarization of the axonal membrane which consequently blocked conduction of action potentials. Consistent with this notion, the presynaptic volley was invariably enhanced for the initial brief period of 3 μM KA application and during washout (Fig. 3B, n = 4), during which the concentration of KA in the bath was expected to be somewhat lower than 3 μM.

Figure 3. Effects of a higher concentration of KA.

A, representative traces of presynaptic volleys recorded before and during application of KA (0.2 μM or 3 μM). B, the graph shows the time course of a single representative experiment. KA (0.2 μM or 3 μM) and TTX (0.5 μM) were applied during the periods indicated. Note that 3 μM KA briefly potentiated (arrow) but soon depressed presynaptic volley potentials during KA application. C, the graph shows the averaged time course of four experiments (means ±s.e.m.). The brief increase is not evident in the average because of variations in its time of occurrence.

Enhancement of mossy fibre excitability by KA

In the experiments described above, a stimulating electrode was placed in the stratum granulosum of the dentate gyrus, and therefore we could not distinguish between whether KA increased the excitability of soma (granule cells) or that of axons (mossy fibres). To distinguish between these two possibilities, we examined the effects of KA on the presynaptic volleys evoked by stimulation of the stratum lucidum in Ca²⁺-free solution. In the arrangement shown in Fig. 4A, the mossy fibre was stimulated halfway between the soma of granule cells and the axon terminal. KA at 0.2 μM also reversibly increased the presynaptic volley (to 157 ± 25 %, n = 4, Fig. 4B and C) under these conditions. These values were not different from those of the presynaptic volleys evoked by stimulation of the stratum granulosum of the dentate gyrus (see Fig. 1C, P > 0.8, Mann-Whitney U test).

Figure 4. Effect of KA on presynaptic volley evoked by stimulation of the mossy fibre axons.

A, the stimulating electrode was placed on the stratum lucidum of the CA3c (hilar) region, to stimulate the mossy fibre pathway at the axon level. B, representative traces recorded before and during KA application. C, the graph shows the time course of the effect of KA on the amplitude of the presynaptic volley (means ±s.e.m., n = 4). KA (0.2 μM) and TTX (0.5 μM) were applied during the periods indicated.

Furthermore, we confirmed the increase in axonal excitability by measuring the amplitude of population antidromic spikes evoked by mossy fibre stimulation. The population antidromic spikes were evoked by stimulating the stratum lucidum of the CA3 region, and were recorded from the stratum granulosum of the dentate gyrus (Fig. 5A). KA at 0.2 μM reversibly increased the amplitude of the population antidromic spikes (to 160 ± 10 % of control, n = 6, Fig. 5B and C). All of these experiments indicate that KA enhances the exitability of mossy fibres.

Figure 5. Effect of KA on antidromic population spikes evoked by mossy fibre stimulation.

A, an electrical stimulus was delivered to the stratum lucidum of the CA3 region, and the evoked antidromic population spikes were recorded from the stratum granulosum of the dentate gyrus. B, specimen records of population spikes recorded before and during KA application. C, the graph represents the time course of the KA effect (means ±s.e.m., n = 6). KA (0.2 μM) and TTX (0.5 μM) were applied during the periods indicated.

To clarify the mechanism of the KA-induced increase in population antidromic spikes, we examined the KA effect on antidromic spikes recorded from a single granule cell. The membrane potential was recorded from a single granule cell by the whole-cell current-clamp method using K⁺-based pipette solution. Stimulation of mossy fibres induced typical antidromic spikes in granule cells (Fig. 6). When the stimulus intensity was set strong enough to evoke antidromic spikes in all trials, KA caused minimal changes in antidromic spikes recorded from single granule cells (Fig. 6B, n = 4). KA at 0.2 μM slightly depolarized the granule cells (by ∼5 mV), and slightly decreased the amplitude of antidromic spikes. However, when a weaker stimulus intensity that failed to evoke antidromic spikes in some trials under control conditions was used, KA application reversibly increased the probability of evoking antidromic spikes (Fig. 6A, n = 6). It is likely that KA depolarized the axonal membrane by activating presynaptic KA receptors, and thereby lowered the stimulus threshold for evoking the antidromic spike.

Figure 6. Effect of KA on antidromic spikes recorded from a single granule cell.

A, antidromic spikes recorded using the whole-cell current-clamp method. Ten consecutive traces are superimposed for each condition. When weak stimulus (stim) intensity that failed to evoke antidromic spikes in some trials was used, KA application increased the probability of evoking antidromic spikes. The graph shows changes in the failure ratio before, during, and after KA application (n = 6). The stimulus intensity was set to evoke antidromic spikes in 10–90 % of all trials under control conditions. B, when the stimulus intensity was set strong enough to evoke antidromic spikes in all trials, KA (0.2 μM) caused a small depolarization and the amplitude of the antidromic spikes was slightly reduced (n = 4). C, effect of KA on membrane potentials (V_m) of granule cells. V_m was recorded continuously for 15 min, and 0.2 μM KA was applied during the period indicated by the open bar (5 min). Antidromic spikes were evoked every 30 s. Note that there were no inter-stimulus spontaneous antidromic spikes, which would be expected to occur if KA-induced axonal depolarization caused ectopic spikes at mossy fibre axons, during KA application (n = 5).

It has been shown that primary afferent depolarization (PAD) in the spinal cord generates antidromic spikes if depolarization is large enough to evoke action potentials (Ménard et al. 1999; see also Cattaert et al. 1994; Pouzat & Marty, 1999). Such ectopic spikes would also occur in hippocampal neurons, since induction of electrographic seizure in hippocampal slices has been reported to accompany an increase in the number of baseline spikes in CA3 neurons (Stasheff & Wilson, 1990), which reflects invasion of antidromic spikes generated ectopically in axons (Stasheff et al. 1993). Therefore, we next examined the possibility that KA application causes a burst of antidromic spikes of ectopic (axonal) origin. For this purpose, we continuously monitored the membrane potential (V_m) of granule cells before and during KA application (Fig. 6C). Antidromic spikes were evoked every 30 s, to confirm that the axon of the test cell was located in the slice and functionally active. Under such conditions, KA application (0.2 μM, 5 min) caused a small depolarization, but no inter-stimulus antidromic spikes (n = 5). This result indicates that 0.2 μM KA does not depolarize the axon to an extent sufficient to generate antidromic spikes.

Suppression of presynaptic Ca²⁺ influx at the mossy fibre synapse by KA

How does the KA-induced enhancement of axonal excitability affect transmitter release from the mossy fibre terminals? To address this question, we examined the effect of KA on action potential-induced presynaptic Ca²⁺ influx, using the technique described previously (Kamiya & Ozawa, 1999) in which presynaptic structures of mossy fibre pathways were selectively loaded with the fluorescent Ca²⁺ indicator rhod-2 AM (Fig. 7A). Figure 7B shows a representative record of such experiments. Application of KA (0.2 μM) again increased the presynaptic volley (to 131 ± 11 % of control) and reduced the field EPSP (to 23 ± 5 % of control). Presynaptic Ca²⁺ influx (ΔF/F) was reduced, but to a lesser degree, by KA application (to 78 ± 4 % of control, n = 6, Fig. 7C). This change in presynaptic Ca²⁺ signal most likely reflects both an increase in the number of activated fibres by the given stimulus intensity and a decrease in the Ca²⁺ influx into the individual terminals. To estimate the changes in the amount of Ca²⁺ influx into each terminal, therefore, the ΔF/F value should be normalized by the presynaptic volley amplitude. This correction was also adopted by Dittman & Regehr (1996) who estimated a decrease in presynaptic Ca²⁺ influx induced by activation of either GABA_B or adenosine A₁ receptors at parallel fibre-Purkinje cell synapses in rat cerebellar slices. After this correction, the Ca²⁺ influx at individual mossy fibre terminals was estimated to be reduced to approximately 60 % of the control value during KA application. Since transmitter release at this synapse is steeply dependent on Ca²⁺ influx (Kamiya & Ozawa, 1999; see also Dodge & Rahamimoff, 1967; Landò & Zucker, 1994), this reduction in the Ca²⁺ influx would largely account for the prominent decrease in the field EPSP.

Figure 7. Suppression of presynaptic Ca²⁺ influx into mossy fibre terminals by KA.

A, schematic diagram showing experimental arrangement. Membrane-permeable Ca²⁺ indicator rhod-2 AM was pressure-ejected into the stratum lucidum, resulting in selective loading of the presynaptic terminals through the mossy fibre pathway. Fluorescence from the labelled region, which had a diameter of about 100 μm and was about 500 μm away from the ejection site, was measured with a single photodiode. B, representative records of presynaptic Ca²⁺ transients (ΔF/F) and field EPSPs observed before (smooth traces) and during application of 0.2 μM KA (dotted traces). C, time courses of the effects of KA on presynaptic Ca²⁺ transient, presynaptic volley and field EPSP. The relative amplitudes of presynaptic Ca²⁺ transients (ΔF/F, •), presynaptic volley (▵) and field EPSP (^) evoked every 5 min were plotted as a function of time (means ±s.e.m., n = 6). KA (0.2 μM) and a mixture of the AMPA receptor antagonist CNQX (10 μM) and the NMDA receptor antagonist D-AP5 (25 μM) were applied during the periods indicated.

Suppression of mossy fibre EPSCs by KA

Prolonged depolarization of CA3 neurons by activation of postsynaptic KA receptors might reduce the field EPSPs by decreasing the driving force for the generation of EPSPs. Therefore, we examined the effect of KA on EPSCs using whole-cell voltage-clamp recordings from visually identified CA3 pyramidal neurons. As shown in Fig. 8, 0.2 μM KA reversibly suppressed the evoked EPSCs (to 38 ± 9 % of control, n = 5) in the CA3 neurones held at -71 mV, while inducing small DC inward currents (93 ± 21 pA, n = 5) and reducing input resistance from 537 ± 161 to 312 ± 107 MΩ (n = 5). This indicates that the KA-induced suppression of the field EPSP is due mainly to reduced transmitter release from mossy fibre terminals.

Figure 8. Effect of KA on mossy fibre EPSCs.

A, representative EPSCs recorded in the control solution and during application of 0.2 μM KA. B, time course of the effect of KA on the EPSCs (means ±s.e.m., n = 5). KA (0.2 μM) and DCG-IV (1 μM) were applied during the period as indicated. These EPSCs were abolished almost completely by addition of the group II-selective mGluR agonist DCG-IV (1 μM).

DISCUSSION

In this study, we have demonstrated the KA receptor-mediated presynaptic inhibitory actions at hippocampal mossy fibre synapses. Application of KA reversibly enhanced the excitability of mossy fibres, possibly via depolarization of axons (mossy fibres) and reduced action potential-induced Ca²⁺ influx into the presynaptic terminals, and thereby inhibited synaptic transmission presynaptically. The involvement of KA receptors in the KA-induced increase in the excitability of mossy fibres has been corroborated pharmacologically by using SYM 2081, which specifically desensitizes KA receptors.

Presynaptic KA receptor at the mossy fibre-CA3 synapse

Earlier radioligand binding studies (Monaghan & Cotman, 1982; Miller et al. 1990) have shown that high- as well as low-affinity [³H]kainate binding sites are distributed throughout the whole CNS, and that some brain areas, including the stratum lucidum of the hippocampal CA3 region, show extremely intense labelling. Represa et al. (1987) suggested the presynaptic localization of the binding sites at the mossy fibre-CA3 synapse, since selective lesion of the dentate granule cells markedly reduced the high-affinity kainate binding in the stratum lucidum of the CA3 region. More recently, ultrastructural localization of KA receptor subunits has been examined using specific antibodies, and immunolabelling of GluR6/7 was detected in postsynaptic densities as well as in presumptive unmyelinated axons in the CA3 region (possibly mossy fibres; Petralia et al. 1994). These anatomical data support the notion that, in addition to the postsynaptic localization, KA receptors containing GluR6/7 are expressed selectively in hippocampal mossy fibres.

Recently, a biochemical study demonstrated the association of KA receptor subunits with members of the SAP90/PSD-95 family (Garcia et al. 1998). This study showed that GluR6 and KA2 both coimmunoprecipitate with SAP90(PSD-95) and SAP102, proteins localized on the postsynaptic densities (PSD). Interestingly, GluR6 also interacts with presynaptic protein SAP97. Since SAP97 is localized on the unmyelinated axons of the hippocampal CA3 region (Müller et al. 1995), it is possible that SAP97 anchors GluR6 on the mossy fibre axons.

KA receptor-mediated enhancement of mossy fibre excitability

The presynaptic volley potentials were reversibly enhanced by low concentrations of KA. We proposed a possible mechanism in which activation of KA receptors on the mossy fibres depolarizes the membranes of the axons, and therefore subthreshold fibres are recruited to generate action potentials by a given stimulus intensity. This notion was based on the following observations: (1) the amplitude of the presynaptic volley evoked by stimulation at the axon in the stratum lucidum increased to almost the same degree as the volleys evoked by stimulation of the stratum granulosum, (2) the amplitude of population antidromic spikes evoked by stimulation of mossy fibres increased during KA application, and (3) the stimulus threshold for generating antidromic spikes recorded from a single granule cell was lowered by KA application.

In this study, we paid much attention to stimulating mossy fibres selectively. It has been described previously that DCG-IV, a group II-selective mGluR agonist, suppressed the mossy fibre responses without affecting other inputs, e.g. commissural/associational input (Kamiya et al. 1996; see also Yokoi et al. 1996). As shown in Fig. 1A, DCG-IV abolished the field EPSP (^), suggesting that we selectively stimulated mossy fibres in this study. This notion is supported by the report of Henze et al. (1997), who analysed the nature of the presynaptic volleys evoked by stimulation of the dentate gyrus. They pointed out that the fibre volleys after blocking EPSP with kynurenate were apparently asynchronous and consisted of two (early and late) components. They further studied the origin of these two components, with a series of precise cuts in the slice, and concluded that the late component represented antidromic activation of CA3 collaterals in the hilus, whereas the early component was the fibre volley of mossy fibre origin. They also noticed that synchronous mossy fibre volleys without contamination of the late component can be recorded if the stimulating electrode was placed carefully on the stratum granulosum and the stimulus intensity was kept weak. Throughout this series of experiments, we adopted their criteria for selectively stimulating the mossy fibre alone. Therefore the presynaptic volley potential recorded in the presence of DCG-IV was invariably synchronous in this study, indicating that the DCG-IV-resistant initial biphasic deflection (Fig. 1A, •) reflects the presynaptic volley potential of mossy fibre origin.

Previous studies have shown that 1 μM KA does not affect the presynaptic volley potential at the Schaffer collateral/commissural synapse of the CA1 region (Collingridge et al. 1983; Kamiya & Ozawa, 1998). In the present study, a lower concentration of KA (0.2 μM) reversibly increased the amplitude of the presynaptic volley, suggesting that KA receptors depolarize mossy fibre axons specifically. KA-induced depolarization of axons has been reported for the primary afferent C fibres (Agrawal & Evans, 1986). In that case, higher concentrations of KA (5-10 μM) reduced the amplitude of C fibre volleys, possibly by depolarization block of action potentials. We also tested a higher concentration of KA (3 μM), and observed similar depression of presynaptic volleys evoked by mossy fibre stimulation. Since we used a relatively weak stimulus to evoke mossy fibre responses without stimulating the axon collateral of CA3 neurons (Weisskopf & Nicoll, 1995; Henze et al. 1997), only a certain fraction of mossy fibres could be stimulated under control conditions. The potentiating action of 0.2 μM KA would be caused by the recruitment of subthreshold fibres due to a slight depolarization of the axonal membrane. On the other hand, the higher concentration of KA (3 μM) would depolarize the membrane to a level which would cause depolarization block of action potential, and therefore depress presynaptic volley potentials. All results in this study are consistent with the notion that functional KA receptors exist on the mossy fibre axons.

Presynaptic inhibition by KA at the mossy fibre-CA3 synapse

In this study, we estimated that presynaptic Ca²⁺ influx induced by action potential at the individual terminals was suppressed to approximately 60 % of the control value. Since transmitter release at this synapse depends steeply on presynaptic Ca²⁺ influx (Kamiya & Ozawa, 1999), transmitter release would be reduced effectively by activation of presynaptic KA receptors. However, the mechanism responsible for the reduction of presynaptic Ca²⁺ influx is currently unclear. KA-induced depolarization of mossy fibres would cause: (1) reduction of action potential amplitude by inactivating axonal Na⁺ channels, (2) electrical shunting by a KA-induced increase in the membrane conductance in the presynaptic region, and (3) inactivating presynaptic Ca²⁺ channels (Forsythe et al. 1998). Although we have no information on the precise ionic mechanism in the mossy fibre terminals, we speculate that the combination of these mechanisms would contribute to the presynaptic inhibitory action of KA.

Involvement of KA receptors in the presynaptic action of KA at the mossy fibre-CA3 synapse

SYM 2081 is a newly developed glutamate analogue which selectively desensitizes KA receptor-mediated responses (Jones et al. 1997; Wilding & Huettner, 1997). Wilding & Huettner (1997) demonstrated that AMPA receptor-mediated currents in cultured hippocampal neurons were almost unaffected by 1 μM SYM 2081, whereas the KA receptor-mediated current was markedly inhibited. We found that SYM 2081 significantly suppressed the KA-induced enhancement of presynaptic volleys, whereas this drug itself does not affect presynaptic volley potentials. Furthermore, application of KA increased presynaptic volley amplitude in the presence of AMPA-selective antagonist GYKI 52466. Taken together these results suggest that KA receptors, rather than AMPA receptors, are involved in the presynaptic action of KA.

One striking difference between AMPA and KA receptors is their desensitization profiles to KA application. Continuous application of KA induces fast desensitizing current in the recombinant KA receptors (for review, see Bettler & Mulle, 1995). This may lead to suggestions that the bath application methods used in this study would barely activate KA receptors. However, the substantial non-desensitizing component in native KA receptor responses in postnatal hippocampal culture (∼30 % of the peak responses) has been demonstrated (Wilding & Huettner, 1997; see also Paternain et al. 1998). Furthermore, the time course of decay of the KA receptor-mediated EPSC at the mossy fibre-CA3 synapse was much slower than that of AMPA receptor-mediated EPSCs (Castillo et al. 1997; Vignes & Collingridge, 1997). An attractive hypothesis to explain the difference in desensitization profiles between recombinant and native KA receptors has recently been proposed by Garcia et al. (1998). They demonstrated that SAP90 (PSD-95) coexpressed with GluR6 or GluR6/KA2 receptors reduced desensitization to KA application, and suggested that the desensitization properties of native KA receptors were modified by association with SAP90/PSD-95 family proteins. Taken together, it is likely that the native KA receptor can be activated by the bath application of KA.

Functional implications

The physiological as well as pathophysiological functions of the presynaptic KA receptors in the hippocampal mossy fibre-CA3 synapse are currently unknown. The presence of multiple autoreceptors at this synapse (KA receptors, this study, see also Vignes et al. 1998; and group II mGluRs, Manzoni et al. 1995; Kamiya et al. 1996; Yokoi et al. 1996) suggests that their presynaptic inhibitory actions may limit excessive excitation of CA3 neurons and counteract the neurotoxic actions of massive glutamate release in some pathological conditions such as ischaemia (Choi & Rothman, 1990). Under physiological conditions, the presynaptic group II mGluRs are activated during high-frequency activation of mossy fibres (Scanzianni et al. 1997), either autaptically or heterosynaptically (Vogt & Nicoll, 1999). Since there are no axo-axonic synaptic contacts on the mossy fibre terminals, these studies suggest that glutamate spills out of the mossy fibre synaptic cleft and acts at a distance. It is possible that axonal KA receptors are also activated in some physiological conditions.

This study has shown that KA receptors on the mossy fibres inhibit the transmission at the mossy fibre-CA3 synapse by a presynaptic mechanism. This conclusion seems to be inconsistent with the previous findings that KA induces epileptiform activity (Westbrook & Lothman, 1983; Mulle et al. 1998) which usually originates from the CA3 region. KA-induced seizures might be generated by strong excitatory effects of ‘postsynaptic’ KA receptors at the mossy fibre- CA3 synapse (Robinson & Deadwyler, 1981; Castillo et al. 1997; Vignes & Collingridge, 1997; Yamamoto et al. 1998) and disinhibitory effects on the inhibitory transmission (Cossart et al. 1998; Frerking et al. 1998), which might overwhelm the inhibitory influence of ‘presynaptic’ KA receptors shown in this study. The mechanism controlling excitability of highly interconnected hippocampal CA3 networks through pre- and postsynaptic KA receptors may be clarified by using gene-targeting of KA receptor subunits (Mulle et al. 1998) as well as cell-targeting strategies (Watanabe et al. 1998) which disrupt specific cell types in the central neural circuits.