INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The mossy fiber synapse has a number of unusual features, including large-sized terminals (3-8 µm), multiple release sites (up to 37), and a proximal termination zone along the apical dendrites of CA3 pyramidal neurons (Chicurel and Harris 1992; Claiborne et al. 1986). The large size and well-defined terminal zone for mossy fiber synapses make them ideal for studying presynaptic Ca²⁺ responses. The initial characterization of Ca²⁺ signals at single mossy fiber terminals was accomplished by Regehr and Tank through localized perfusion of the AM ester form of the fluorescent calcium indicator fura-2 to the mossy fiber tracts (Regehr and Tank 1991b). This method for selective labeling and measuring changes in [Ca²⁺]_i of presynaptic terminals has now been applied to many other synapses in hippocampus and cerebellum (Regehr and Atluri 1995; Regehr and Tank 1991a,b; Sabatini and Regehr 1998; Wu and Saggau 1994). Recently, direct electrical recordings have also been made from single presynaptic mossy fiber boutons (Geiger and Jonas 2000).

It is known that both endoplasmic reticulum (ER) and mitochondria are widely distributed within neurons, being present in dendrites and dendritic spines, axons and presynaptic nerve terminals, and in growth cones (Cheng and Reese 1985; Chicurel and Harris 1992; Dailey and Bridgman 1989; Deitch and Banker 1993; Kanaseki et al. 1998; Levesque et al. 1999), and Ca²⁺ pumps and internal ligand-gated Ca²⁺ channels, including ryanodine receptors (RyRs) and inositol (1,4,5)-trisphosphate receptors (IP₃Rs), participate in Ca²⁺ signaling under certain conditions (Berridge 1998). The spatiotemporal distribution of intracellular free Ca²⁺ levels is critically involved in higher brain activities including learning and memory (Alkon et al. 1998; Bliss and Collingridge 1993; Finkbeiner and Greenberg 1997; Teyler et al. 1994). During neuronal activation, elevation of cytoplasmic [Ca²⁺]_i is initially generated by Ca²⁺ influx through voltage- or ligand-gated Ca²⁺ channels. Subsequently, activation of IP₃Rs and RyRs on the endoplasmic reticulum membrane leads to Ca²⁺ release from intracellular stores (calcium-induced calcium release, CICR), which further amplifies and/or prolongs Ca²⁺ signals in specific subcellular compartments (Pozzan et al. 1994; Sorrentino and Volpe 1993; Sutko and Airey 1996). Electron microscope studies have provided suggestive but limited evidence showing the presence of mitochondria and ER in mossy fiber terminals (Chicurel and Harris 1992). Immunohistochemical staining experiments have shown clear differences in the localization of IP₃R and RyR. In the hippocampus, IP₃R is most concentrated in pyramidal cells of CA1. By contrast, RyR staining is much greater in area CA3 than CA1 and is particularly prominent in the granule cell layer of the dentate gyrus (Sharp et al. 1993). Further study suggested that of the three isoforms of RyR (RyR1, RyR2, and RyR3), RyR2 mRNA was shown to be widely expressed in the rat CNS at high concentrations, and high levels of RyR2 mRNA signal were detected in the dentate gyrus and area CA3 (Zhao et al. 2000).

Little is known about the possible contribution of internal calcium release to the rise in presynaptic calcium at the mossy fiber synapse. There have been, however, previous studies in area CA1 suggesting that intracellular calcium release may have an effect on synaptic transmission. Depleting the stores or blocking CICR reduces the transients and also reduces a form of short-term synaptic plasticity, paired-pulse facilitation of excitatory postsynaptic potentials (EPSPs) (Emptage et al. 2001). Thapsigargin, which depletes intracellular calcium stores, has been shown to block long-term potentiation elicited by weak stimuli, but not that induced with strong stimuli suggesting that calcium release may, under some conditions, play a role in induction of synaptic plasticity (Behnisch and Reymann 1995). Anoxia causes an increase in the frequency of spontaneous miniature excitatory postsynaptic currents measured in CA1 neurons; this increased frequency of release could be blocked by agents that decrease the release of intracellular calcium (Katchman and Hershkowitz 1993). Reyes and Stanton found, also in area CA1, that presynaptic blockade of ryanodine receptors, or postsynaptic blockade of inositol triphosphate, triggered release of internal calcium could block the induction of long-term depression (Reyes and Stanton 1996).

In the present study, we applied optical recording techniques to characterize the presynaptic Ca²⁺ signal at single mossy fiber terminals in rat hippocampal slices by loading synapses with fura-2AM or other permeant indicators. We found that in response to a brief train of 20 stimuli at 100 Hz, presynaptic [Ca²⁺] increased and decayed rapidly, and the decay phase of [Ca²⁺] was often accompanied by a delayed peak. The existence of the second peak was variable, often apparent early in an individual recording, and then fading with time. In other experiments, the second peak was not apparent unless the number of stimuli in the train was increased (although we report here only results with a 20-stimuli train). This indicated to us that this component of the Ca²⁺ signal was under some level of physiological control and worthy of further study. Our data suggest that both the initial and the secondary peaks of Ca²⁺ reflect contributions of Ca²⁺ release from internal stores and that Ca²⁺ release from this ryanodine-sensitive store contributes significantly to the rise in preterminal [Ca²⁺].

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Preparation of hippocampal slices

Hippocampal slices were prepared from Sprague-Dawley rats (5-7 wk) using standard procedures (Magee and Johnston 1995). All experimental procedures were approved by the Animal Research Committee of Baylor College of Medicine. An anesthetic consisting of a mixture of ketamine (42.8 mg/ml), xylazine (8.6 mg/ml), and aceromazine (1.4 mg/ml) was injected intraperitoneally, and rats were perfused with a cold (~2°C) oxygenated cutting solution described below. Slices were cut on a Vibratome at a thickness of 350 µm, incubated in a holding chamber heated to 35°C for 20 min, and then stored at room temperature (~22°C). The holding chamber was continuously bubbled with 95% O₂-5% CO₂ and contained the bathing solution described below.

Drugs and solution

The cutting solution contained (in mM) 110 choline chloride, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 0.5 CaCl₂, 7 MgCl₂, 1.3 ascorbate, 3 pyruvate, and 7 dextrose, bubbled with 95% O₂-5% CO₂ during whole cutting process. The bathing solution contained (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2 CaCl₂, 2 MgCl₂, and 10 dextrose, bubbled with 95% O₂-5% CO₂ continuously. Where specified, one of the following drugs was included in the media: L-CCG-I (Tocris) or D,L-2-amino-5-phosphonovaleric acid (D,L-APV; Tocris) were prepared from a stock solution dissolved in 1 molar equivalent (1 eq.) NaOH; ryanodine (Sigma), cyclopiazonic acid (Sigma), thapsigargin (Alomone labs), 6,7-Dinitroquinoxaline-2,3-dione (DNQX; Sigma), or CGP 55845 (Tocris), prepared from a stock solution dissolved in 20, 30, 10, 20, or 10 mM DMSO, respectively; DCG IV (Tocris) or ruthenium red (Sigma), prepared from a concentrated stock solution in water. The glass stimulating electrode solution contained (in mM) 149 NaCl, 5 KCl, 10 HEPES, 2 CaCl₂, 1 MgCl₂, and 10 dextrose (pH 7.3).

Fluorescence imaging

In our experiments, the loading of mossy fiber terminals in hippocampal slices with cell-permeant fura-2AM took advantage of two previous findings (Gray et al. 1996; Regehr and Tank 1991b). First, CA3 neurons in the hippocampal slice do not load well with fura-2AM; second, presynaptic terminals tend to load well with fluorophores. We were therefore able to identify and locate mossy fiber terminals (MFTs) using fluorescence illumination.

For loading of slices with dye, 50 µg fura-2 acetoxymethyl ester (fura-2AM) (Molecular Probes) was added to 40 µl of 20% Pluronic F-127 (Molecular Probes) in DMSO (Sigma) to make 1.25 mM stock solution as described (Gray et al. 1996; Saggau et al. 1999). Individual slices were removed from the holding chamber and placed in a 35-mm Petri dish containing artificial cerebrospinal fluid (ACSF) and 8.3 µM cell-permeant fura-2AM for 10-15 min at 30°C. After loading, individual slices were then rinsed gently at least 10 times in ACSF and transferred to a chamber on the stage of the microscope. The recording chamber was continuously perfused with bathing solution at 30-32°C. A Zeiss Axioskop, fitted with a ×40 Zeiss water-immersion objective (N.A. 0.75) and differential interference contrast (DIC) optics, was used to view slices. MFTs were located using fluorescence illumination (380 nm), and single terminals were isolated by closing a diaphragm in the light path to a spot diameter of 10-15 µm (Fig. 1A). Light emission of 510 nm was measured with a photodiode (Hamamatsu S1336-18BK) over the terminal through a 1-µm pinhole to reduce collection of scattered light. Relative changes in [Ca²⁺]_i were quantified as changes in F/F, where F is fluorescence intensity before stimulation (after subtracting autofluorescence), and F is the change from this value during neuronal activity (corrected for bleaching during optical recording). The bleaching correction was determined by measuring fluorescence in the MFT without stimulation. Tissue autofluorescence and background fluorescence were subtracted by measuring fluorescence at a parallel location in the slice that was away from the loaded MFT. The terminal was stimulated with a glass microelectrode (2-4 µm diam) filled with HEPES-buffered internal solution and placed in stratum lucidum approximately 50 µm from the selected terminal (Fig. 1A) by giving a brief train of 20 stimuli at 100 Hz, which was the minimum stimulus to reliably produce a delayed peak in the Ca²⁺ signal.

View larger version (44K): [in this window] [in a new window]   Fig. 1. Detection of Ca2+ signals in single mossy fiber terminals. A: schematic of hippocampal slice showing stimulating (stim) and recording site (green circle). Bottom image shows several mossy fiber terminals (MFTs) loaded with fura-2AM (bright green spots). B: presynaptic Ca2+ influx induced through voltage-dependent Ca2+ channels by stimulating the mossy fibers can be completely blocked by 400 µM Cd2+. The bottom traces show frequency of stimuli and time course of a typical recording procedure. First, bleaching traces (no stimulation and normal saline) were recorded for 20 min (10 traces, 5-s trace acquired every 2 min), then responses from 20 stimuli for 20 min (in normal saline) were recorded (10 traces), followed by bath application of the drug. After waiting 10 min, a new sequence of bleaching traces (10 traces) were collected, then the Ca2+ responses were measured in the presence of the drug for 30 min (15 traces). C and D: effects of bath-applied group II mGluR agonist L-CCG-I (10 µM) or DCG IV (10 µM) to MFTs after blocking presynaptic GABAB receptors with 1 µM CGP 55845. Both L-CCG-I and DCG IV decreased presynaptic calcium responses significantly.

Statistics

All numerical values are represented as means ± SE, and the number of experiments (n) refers to the number of slices investigated. The differences between the experimental groups were evaluated using Student's paired t-test. In all cases a probability value of <0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The rapid rise in presynaptic [Ca²⁺] in response to MF stimulation appeared to result from activation of voltage-dependent Ca²⁺ channels because it was completely blocked by 400 µM Cd²⁺ (Fig. 1B). To test whether the recorded terminals were from the mossy fibers of the dentate gyrus and the evoked calcium signals were from mossy fiber presynaptic terminals, we sequentially co-stained slices with calcium green and the Zn²⁺ fluorescent dye TFLZn. While there was a high background fluorescence with TFLZn, significant co-staining was evident (data not shown). In addition, we applied the group II mGluR agonist L-CCG-I (10 µM) or DCG IV (10 µM) to MFTs after blocking presynaptic GABA_B receptors with 1 µM CGP 55845. Both L-CCG-I and DCG IV decreased presynaptic calcium influx to 66% (n = 6) or 55% (n = 3) of control, respectively (Fig. 1, C and D). The group II mGluR agonists have become useful tools in distinguishing MF synapses because they selectively inhibit MF synapses and do not affect C/A synapses (Kamiya et al. 1996; Yeckel et al. 1999). Under the same experimental conditions, we investigated the contribution of intracellular Ca²⁺ release to the total Ca²⁺ signal at mossy fiber presynaptic terminals by stimulating mossy fibers and bath applying agents known to interfere with intracellular Ca²⁺ signaling.

Ryanodine (20 µM), which blocks ryanodine receptors in the endoplasmic reticulum and prevents Ca²⁺ release in high concentration (Mattson et al. 2000; Sutko et al. 1997), decreased the amplitude of the Ca²⁺ signals. After bath application for 30 min, the fluorescent signal (F/F) induced by stimulating mossy fiber was decreased from 3.22 ± 0.64 to 1.73 ± 0.34 (mean ± SE, n = 6; P < 0.01; Fig. 2A). Cyclopiazonic acid (CPA 30 µM), a specific inhibitor of Ca²⁺-ATPase in the ER that blocks Ca²⁺ uptake and depletes Ca²⁺ stores without increasing intracellular IP₃ levels (Markram et al. 1995; Plenge-Tellechea et al. 1997; Seidler et al. 1989), decreased the amplitude of the Ca²⁺ signals from terminals. After washing in CPA for 30 min, the fluorescent signal (F/F) induced by stimulating mossy fiber was decreased from 2.93 ± 0.31 to 1.45 ± 0.31 (n = 5; P < 0.01; Fig. 2B). Thapsigargin (10 µM) can deplete ER Ca²⁺ by specific inhibition of ER Ca²⁺-ATPase (Markram et al. 1995; Mattson et al. 2000; Thastrup et al. 1990; Treiman et al. 1998). After bath application for 60 min, the Ca²⁺ signal amplitude decreased from 4.56 ± 0.2 to 2.06 ± 0.19 (n = 3; P < 0.01; Fig. 2C). Ruthenium red (100 µM), a Ca²⁺ antagonist that inhibits the ryanodine receptor and the mitochondrial Ca²⁺ uniporter (Chen and MacLennan 1994; Ma 1993; Moore 1971), decreased the amplitude of the Ca²⁺ signals significantly. After perfusing ruthenium red for 30 min, the fluorescent signal (F/F) was changed from 2.58 ± 0.18 to 0.76 ± 0.04 (n = 5; P < 0.01; Fig. 2D). We also measured the half decay time for both control and with blockers. We found that there were no statistical differences before and after blockers. Surprisingly, drugs that affect Ca²⁺ release also decreased the initial peak of the terminal Ca²⁺ signal, suggesting that release from intracellular stores is rapid and participates in the formation of the initial peak.

View larger version (26K): [in this window] [in a new window]   Fig. 2. Effects of bath-applied ryanodine (20 µM; A), CPA (30 µM; B), thapsigargin (10 µM; C), and ruthenium red (100 µM; D) on Ca2+ responses after giving a brief train of 20 stimuli. The Ca2+ signals from mossy fiber terminals were decreased significantly after wash-in one of the above agents for 30 min (60 min for thapsigargin). But under control condition (E) and in the presence of D,L-APV (50 µM; F) and DNQX (20 µM; G), the Ca2+ signals did not change. These results exclude the action of N-methyl-D-aspartate (NMDA) and non-NMDA receptors on the evoked Ca2+ signals and indicate the stability of the recordings.

Continuously perfusing normal external solution or applying DMSO (0.1%) after normal external solution in the same time period as drug application, the F/F signals were stable during recording (Fig. 2E). In addition, we applied D, L-APV (50 µM) and DNQX (20 µM), the selective N-methyl-D-aspartate (NMDA) and non-NMDA receptor antagonist, to exclude the action of NMDA and non-NMDA receptors on the evoked presynaptic Ca²⁺ signals (Fig. 2, F and G).

Alternative indicators with lower affinities for Ca²⁺ were used to address the concern that saturation of fura-2 might confound these results. Both calcium green-2AM (K_d = 0.55 µM; Fig. 3A) and fura-2FF-AM (K_d = 6 µM; Fig. 3B) showed changes similar to those measured with fura-2AM (K_d = approximately 0.22 µM; Fig. 2A) in response to application of ryanodine. 4-Aminopyridine (4-AP, 1 mM), which blocks transient K⁺ channels, increased the presynaptic signal by 107 ± 0.1% (n = 17, P < 0.01; Fig. 3C), further suggesting that saturation of indicator was not occurring in these experiments. Figure 3D summarized the effects of all tested agents on calcium responses at single MFTs.

View larger version (29K): [in this window] [in a new window]   Fig. 3. To test for saturation of fura-2, we compared the effects of ryanodine (20 µM) on Ca2+ release at mossy fiber terminals loaded with indicators having different affinities for Ca2+. A: calcium green-2AM (Kd = 0.55 µM). B: fura-2FF-AM (Kd = 6 µM). C: after focal application of 4-aminopyridine (4-AP; 1 mM), the transient K+ channel blocker, Ca2+ signals increased. D: summary of the effects of all tested agents on calcium responses at single MFTs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There are two principal results of this study. 1) Using optical methods, we recorded a Ca²⁺ response (F/F) at single synaptic boutons in stratum lucidum of the CA3 region in the hippocampus and stimulated the mossy fiber pathway. Because the evoked Ca²⁺ signals were reduced by group II mGluR agonists and the terminals co-stained for Zn²⁺, the results strongly support the hypothesis that the recorded terminals were from the mossy fibers of the dentate gyrus. This entire Ca²⁺ response could be completely blocked by the Ca²⁺ channel blocker Cd²⁺. 2) Several pharmacological agents thought to disrupt regulation of intracellular Ca²⁺ stores (resulting in depletion of internal Ca²⁺) decreased the early and the delayed increase in the Ca²⁺ signals. Local application of 4-AP and indicators with lower affinities for Ca²⁺ were used to control for possible saturation of fura-2.

Our experiments suggest the following sequence of events. Stimulation of the mossy fiber pathway generated action potentials that propagated down the axon to the terminals. Action potentials directly caused [Ca²⁺]_i increases by opening voltage-gated Ca²⁺ channels located on mossy fiber terminals. This transiently increased Ca²⁺ concentration rapidly triggered more Ca²⁺ release from ER through ryanodine receptors. The triggered release could be inhibited by either ryanodine receptors blockers (ryanodine and ruthenium red) or specific inhibition of ER Ca²⁺-ATPase (CPA and thapsigargin), which deplete the ER Ca²⁺ store. These results were in agreement with observations on Schaffer collateral boutons in cultured hippocampal slices (Emptage et al. 2001), in which they reported that action potentials reliably triggered large Ca²⁺ transients in boutons, due both to Ca²⁺ influx and to CICR from internal stores, depleting the stores (by CPA or thapsigargin) or blocking CICR (by ryanodine) reduced the evoked Ca²⁺ transients. Our proposed presynaptic mechanism is different from other reports on postsynaptic [Ca²⁺] changes (Finch and Augustine 1998; Kapur et al. 2001; Nakamura et al. 1999; Takechi et al. 1998). First, these reports all studied postsynaptic [Ca²⁺] changes, either in hippocampal pyramidal neurons or Purkinje cells in the cerebellum, all of which prominently express metabotropic glutamate receptor I (mGluRI) and inositol-1,4,5-trisphosphate (IP₃) receptors (Conn and Pin 1997; Walton et al. 1991); second, the results showed that the synaptically activated Ca²⁺ transients in these cells were mediated by activation of postsynaptic mGluRI and required IP₃-mediated Ca²⁺ release from internal stores. We investigated Ca²⁺ release at presynaptic mossy fiber synapses on CA3 pyramidal neurons in the hippocampus, where the levels of RyRs are particularly high (Padua et al. 1992; Sharp et al. 1993). The Ca²⁺ channel blocker, Cd²⁺ (400 µM) could completely inhibit the rise of presynaptic [Ca²⁺]. This is evidence that extracellular Ca²⁺ through voltage-gated Ca²⁺ channels on MFTs participated in the initial [Ca²⁺]_i increase. In addition, we applied D,L-APV and DNQX, the selective NMDA and non-NMDA receptor antagonists, to exclude the effects of NMDA and non-NMDA receptors on the evoked presynaptic Ca²⁺ signals.

Mossy fiber boutons terminate on the proximal portion of the apical dendrites of CA3 pyramidal neurons. There are multiple active zones at each bouton resulting in multiple release sites for neurotransmitter (Chicurel and Harris 1992; Claiborne et al. 1986). Two particularly noteworthy features of mossy fiber synaptic transmission are large facilitation and minimal depression during long-duration, high-frequency stimulation (Yeckel et al. 1999). One possibility is that the Ca²⁺ release described here contributes to the high capacity for neurotransmitter release at this synapse.

Although different mechanisms mediated the presynaptic and postsynaptic Ca²⁺ release from intracellular stores, our studies suggest that at mossy fiber presynaptic terminals, Ca²⁺ release from internal stores is an important component of the Ca²⁺ signal and contributes to the rise of presynaptic [Ca²⁺]_i during evoked activity. A recent report has also characterized mossy fiber-evoked Ca²⁺ release in postsynaptic CA3 neurons (Kapur et al. 2001).