INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Changes in the concentration of intracellular Ca²⁺ ([Ca²⁺]_i) has many important cellular functions; for example; neuronal excitation, exocytosis, muscle contraction, gene expression, and second-messenger signaling (Berridge 1998; Dolmetsch et al. 1998; Li et al. 2000). The [Ca²⁺]_i in nerve terminals is also critical in regulating the probability of neurotransmitter release and synaptic plasticity (Kamiya and Zucker 1994; Schneggenburger and Neher 2000; Zucker 1996).

Increases in free [Ca²⁺]_i can arise from two compartments: from the extracellular fluid or from release from intracellular stores. The influx of Ca²⁺ from the extracellular fluid can be mediated via voltage-dependent Ca²⁺ channels (VDCCs) (Meir et al. 1999; Rhee et al. 1999) and/or via some ligand-gated cation channels (Chittajallu et al. 1996; Rhee et al. 2000). Mechanisms for release from intracellular stores includes IP₃-induced Ca²⁺ release and Ca²⁺-induced Ca²⁺ release systems (Berridge 1998; Finch and Augustine 1998). On activation of one of these mechanisms, the normally low [Ca²⁺]_i (~10⁷ M) can be raised to 10⁵~10⁴ M (Zucker 1996). The increased [Ca²⁺]_i is removed from the cytoplasm by various Ca²⁺ pumps, such as the ATP driven sarco-endoplasmic reticulum Ca²⁺ pump (Hartter et al. 1987) and the plasma membrane Ca²⁺ pump (DiPolo and Beauge 1983). Another mechanism for Ca²⁺ homeostasis is the Na⁺/Ca²⁺ exchanger (NCX), a transporter that has been identified in many kinds of tissues including neuronal cells (Blaustein and Lederer 1999; Bobbin et al. 1991; Kobayashi and Tachibana 1995; Reuter and Porzig 1995). NCX is located in the plasma membrane where it exchanges Na⁺ and Ca²⁺, the direction of net transport being determined by both the concentration gradients of Na⁺ and Ca²⁺ across the plasma membrane and by the membrane potential (Blaustein and Lederer 1999). Under normal physiological conditions, NCX transports Na⁺ into the cell and Ca²⁺ out of the cell (forward mode NCX). However, under conditions in which the relevant gradients are reversed, the NCX can operate in the reverse mode, providing a net influx of Ca²⁺ (Mulkey and Zucker 1992). The presence of NCX has been demonstrated in presynaptic nerve terminals of cultured hippocampal neurons (Juhaszova et al. 2000; Reuter and Porzig 1995), in the guinea pig cochlea (Bobbin et al. 1991), and in goldfish retinal bipolar cells (Kobayashi and Tachibana 1995). However, the functional and pharmacological properties of NCX have been rarely investigated by using high temporal resolution electrophysiological techniques and in GABAergic presynaptic nerve terminals in native mammalian CNS neurons.

It is technically difficult to examine channels and transporters in typical vertebrate presynaptic nerve terminals because of their small size. In the present study, we used the "synaptic bouton preparation" in which Meynert neurons were mechanically dissociated with the presynaptic nerve endings attached to the isolated neurons. This preparation allows the observation of presynaptic neurotransmitter release by electrophysiological recordings from the soma of a single isolated neuron (Rhee et al. 1999, 2000). This preparation also enables the solution bathing these single cells to be accurately controlled and rapidly exchanged. For rapid solution exchange and drug application, we used the "Y-tube method" (Murase et al. 1990). Using these techniques, we demonstrate the existence of NCX in GABAergic presynaptic nerve terminals projecting to Meynert neurons.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Preparation

Wistar rats (17-21 days old) were decapitated under pentobarbital anesthesia (50 mg/kg ip). The brain was quickly removed and transversely sliced at a thickness of 400 µm (DTK-1000, Dosaka, Kyoto, Japan). Slices were kept in incubation medium (see following text) saturated with 95% O₂-5% CO₂ at room temperature (21-24°C) for 1 h before being transferred into 35-mm culture dishes (Primaria 3801, Becton Dickinson). The Meynert nuclei were identified using a binocular microscope (SMZ-1, Nikon, Tokyo). For mechanical dissociation, we used a custom-built vibration device that has an arm equipped with a fire-polished glass pipette. The pipette tip was lightly placed on the surface of the Meynert nuclei using a micromanipulator and was vibrated at 3-5 Hz over a distance of ~0.2 mm. Within ~5 min, the dissociation was complete, and the remaining slice was removed. The mechanically dissociated neurons were left for ~20 min during which time the cells adhered to the bottom of the dish (Rhee et al. 1999). Single neurons that retained some of their original morphological features such as their proximal dendrites were used for recordings. All experiments conformed to the guiding principles for the care and use of animals approved by The Council of The Physiological Society of Japan, and all efforts were made to minimize the number of animals used and their suffering.

Electrical measurements

All electrical measurements were performed using the whole cell patch-clamp recording mode. Cells were voltage-clamped at a holding potential (V_H) of 60 mV, and currents were recorded with a patch-clamp amplifier (EPC-7, List Electric). Patch pipettes were made from borosilicate capillary glass (1.5 mm OD, 0.9 mm ID; G-1.5, Narishige, Tokyo) pulled in two stages on a vertical puller (RB-7, Narishige). The resistance of these electrodes was 5-7 M. Neurons were viewed under phase contrast on an inverted microscope (Diaphot, Nikon, Tokyo). Currents and voltage were continuously monitored on an oscilloscope (VC6025, Hitachi, Tokyo) and recorded on a chart recorder (Recti-Horiz 8K, Nippondenki San-ei, Tokyo) and on a DAT recorder (RD-130TE TEAC, Tokyo). Membrane currents were filtered at 1 kHz (E-3201A Decade Filter, NF Electronic Instruments, Tokyo). All experiments were performed at room temperature (21-24°C).

Solutions

The ionic composition of the incubation medium was (in mM) 124 NaCl, 5 KCl, 1.2 KH₂PO₄, 24 NaHCO₃, 2.4 CaCl₂, 1.3 MgSO₄, and 10 glucose. The pH was maintained at 7.4 by bubbling the medium with 95% O₂ and 5% CO₂ at room temperature. Standard external solution (standard solution) contained (in mM) 150 NaCl, 5 KCl, 2 CaCl_2, 1 MgCl₂, 10 glucose, and 10 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES). N-methyl-D-glucamine-Cl (NMG-Cl) external solution (Na⁺-free solution) contained (in mM) 150 NMG-Cl, 5 KCl_, 2 CaCl_2, 1 MgCl₂, 10 glucose, and 10 HEPES. The Ca²⁺-free external solution (Ca²⁺-free solution) contained (in mM) 140 NaCl, 5 KCl_, 6 MgCl₂, 2 ethylene glycol-bis (-aminoethyl ethel)-N,N,N',N'-tetraacetic acid (EGTA), 10 glucose, and 10 HEPES. The pH of these external solutions was adjusted to 7.4 with Tris-OH. To isolate GABAergic mIPSC, these solutions routinely contained 3 × 10⁷ M tetrodotoxin (TTX), a selective voltage-dependent Na⁺ channel blocker, and 10⁶ M 6-cyano-7-phosphovaleric acid (CNQX) and 10⁵ M DL-2-amino-5-phosphovaleric acid (APV) to suppress glutamatergic currents. The ionic composition of Cs⁺ internal solution (in mM) was: 100 CsCl, 50 Cs methanesulfonate, 5 TEA-Cl, 2 EGTA, 4 ATP-Mg, 10 HEPES. The pH was adjusted to 7.2 with Tris-OH.

Data analysis

Miniature inhibitory postsynaptic currents (mIPSCs) were detected and analyzed using DETECTiVENT (Ankri et al. 1994) and IGOR PRO software (Wavemetrics, Lake Oswego, OR). All traces were visually inspected before being accepted for further analysis to minimize artificial events (i.e., noise) that otherwise could have been detected by the automated analysis. mIPSCs amplitude was calculated by subtracting the baseline current value from the peak current value. When two or more events were overlaid, the baseline current for the latter events was estimated by extrapolating the decay phase of the preceding event. This procedure minimized the influence of changes in baseline current on the event amplitude. For mIPSCs analysis, all data were normalized to the control values. To investigate the time course of changes in mIPSCs frequency, the total number of events observed in bins of 10 s was plotted. The averaged frequency of mIPSCs events during the control period (cont. in Figs. 2-8; 5 min) was defined as 1.0, the control mIPSCs frequency. The amplitude of mIPSCs averaged over the control period was also normalized to 1.0, the control mIPSCs amplitude. The effects of the various experimental conditions on mIPSCs frequency and amplitude are expressed relative to the normalized control values. Differences in these mIPSCs parameters were examined using the Wilcoxon signed-ranks test for comparison between two groups. Differences in the amplitude of the exogenous GABA-induced inward currents under different experimental conditions were examined using Student t-test. Statistical analyses were conducted using SPSS 7.5J software (SPSS Japan, Tokyo). Differences were judged to be statistically significant at P < 0.05. Numerical values are provided as means ± SE.

Drugs

Drugs used in the present study were APV, CNQX, EGTA, ouabain, cyclopiazonic acid (CPA; all from Sigma, St. Louis, MO), and TTX (Wako Pure Chemicals, Tokyo). 2-[2-[4-(4-nitrobenzyloxy)- phenyl]ethyl] isothiourea methanesulfonate (KB-R7943) was kindly provided by Kanebo (Tokyo). Drugs and solutions were rapidly applied using the Y-tube method (Murase et al. 1990).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1A shows a typical example of spontaneous synaptic currents recorded at a holding potential (V_H) of 60 mV. These current transients were completely suppressed by 3 × 10⁵ M bicuculline. Figure 1B shows typical spontaneous synaptic currents recorded at various holding potentials. The currents reversed polarity at a V_H of 10.5 ± 0.4 mV, which was almost equal to the Cl equilibrium potential (E_Cl) of 10.7 mV, calculated from the Cl concentrations in the patch pipette (internal) and normal external solution. These results clearly identify these currents as GABAergic mIPSCs.

View larger version (26K): [in this window] [in a new window]   Fig. 1. GABAergic miniature inhibitory postsynaptic currents (mIPSCs). A: in the presence of TTX (3 × 107 M), 6-cyano-7-nitroquinoxalene-2,3-dione (CNQX, 3 × 106 M) and DL-2-amino-5-phosphovaleric acid (APV, 105 M), bicuculline (3 × 105 M) completely and reversibly blocked mIPSCs. Bottom: segments of the top trace (indicated by a and b) at greater resolution. B, left: representative current traces showing mIPSCs at various VHs. Right: mean mIPSCs amplitude current-voltage (I-V) relationship. All responses were normalized to the mean amplitude of mIPSCs at a VH of 40 mV (*). Each point, and the corresponding vertical error bars, show means ± SE from 5 neurons.

Effect of Na+ driving force on mIPSCs

The elevation of [Ca²⁺]_i in GABAergic presynaptic nerve endings would be expected to cause an increase in GABA release probability and result in an increase in mIPSCs frequency. Mechanisms that could result in an influx of Ca²⁺ from the extracellular solution include the activation of VDCCs, the activation of ligand-gated cation channels, and reverse mode operation of the Na⁺/Ca²⁺ exchanger (reverse mode NCX) (Blaustein and Lederer 1999; Meir et al. 1999). To determine whether the NCX exists in these presynaptic nerve endings, we manipulated the experimental conditions in an attempt to activate the reverse mode NCX, which should increase [Ca²⁺]_i and hence increase mIPSCs frequency.

The removal of Na⁺ from the external solution activates reverse mode NCX in neuronal somata and increases [Ca²⁺]_i (Blaustein and Lederer 1999; Meir et al. 1999). If NCX is present in these presynaptic nerve terminals, a similar procedure could also increase [Ca²⁺]_i by reverse mode NCX resulting in an increase in GABA release probability. When the external Na⁺ was replaced with the larger cation, N-methyl-D-glucamine (Na⁺-free solution), there was a small outward shift in the postsynaptic holding, as previously reported (Munakata et al. 1998; Shimura et al. 1999). However, there was no significant effect on both mIPSCs frequency (0.91 ± 0.11 of control, P = 0.063, n = 7) or amplitude (0.99 ± 0.10 of control, P = 0.43; Fig. 2).

View larger version (45K): [in this window] [in a new window]   Fig. 2. Effects of Na+-free solution on mIPSCs. A, top: current recordings before, during, and following perfusion with Na+-free, N-methyl-D-glucamine (NMG+)-containing solution as indicated (). Sections of this trace, as indicated (- - -), are shown at higher resolution. Bottom: time course of mIPSCs frequency before, during and after application of Na+-free solution. mIPSCs frequency was calculated and plotted for 10-s epochs and normalized to the mean control mIPSCs frequency (=1). Each point represents the mean relative frequency from 7 experiments, error bars show SE. B: the effects of Na+-free solution on the relative mIPSCs frequency (left) and amplitude (right) distributions. C: mean values of mIPSCs frequency and amplitude, normalized to those in standard solution. Vertical bars indicate ±SE.

Ouabain inhibits Na⁺/K⁺ ATPase (O'Brien et al. 1994), resulting in an elevation of the intracellular Na⁺ concentration ([Na⁺]_i (Blaustein 1993; Rose and Ransom 1997). Thus we next investigated the effects of removing external Na⁺ in the presence of ouabain. The application of 10⁴ M ouabain in standard solution caused little change in mIPSCs frequency or amplitude (cont. in Fig. 3, A and B). However, as shown in Fig. 3, when we removed external Na⁺ in the presence of ouabain, there was a marked but transient increase in mIPSCs frequency (period a in Fig. 3, A and B: 2.05 ± 0.29, P < 0.05, n = 7). mIPSCs frequency peaked ~30 s after application of Na⁺-free solution before declining back to control levels over the next 1-2 min (period b in Fig. 3, A and B: 1.14 ± 0.25, P = 0.74). When the external Na⁺ was reapplied, there was no further change in the mIPSC frequency (recov. period in Fig. 3, A and B). The averaged amplitude of mIPSCs changed little throughout all these experimental conditions (Fig. 3, B and C; a: 1.08 ± 0.06; b: 1.04 ± 0.03). These results indicate that the subsequent removal of external Na⁺ in the presence of ouabain causes a transient increase in mIPSCs frequency.

View larger version (45K): [in this window] [in a new window]   Fig. 3. Effect of Na+-free solution with ouabain. A, top: typical raw data current trace illustrating the effects of Na+-free solution in the presence of 104 M ouabain. Ouabain was first added to the standard solution (as indicated by the upper bar) followed by addition of Na+-free solution (black bar). Sections of this trace (as indicated by the dashed lines) are expanded for clarity. Bottom: time course of normalized mIPSCs frequency in standard solution and then in ouabain; before (cont.), during (a and b), and after (recov.) application of Na+-free solution. Each point represents the mean ± SE from 7 experiments. B: relative mIPSCs frequency (left) and amplitude (right) distributions during the times indicated by the arrows. C: mean mIPSCs frequency (left) and amplitude (right) expressed relative to the control values.

The preceding results suggest that reverse mode NCX, induced by reversal of the Na⁺ driving force, may be responsible for the presumed increase in [Ca²⁺]_i. Ouabain, however, may not only inhibit the Na⁺/K⁺ ATPase but may also cause membrane depolarization (e.g., Munakata et al. 1998). If the presynaptic nerve terminals are depolarized by the application of ouabain, one may predict a subsequent increase in [Ca²⁺]_i. To clarify this, we reversed the order of application of Na⁺-free and ouabain-containing solution. In the absence of external Na⁺, the application of 10⁵ M ouabain had no effect on mIPSCs frequency (Fig. 4; 1.10 ± 0.08, P = 0.34, n = 5). However, under the same experimental conditions, the application of 10 mM K⁺ solution markedly increased mIPSCs frequency (Fig. 4; 5.42 ± 0.43, P < 0.05, n = 5). The increase of mIPSCs frequency during the application of high-K⁺ solution presumably results from presynaptic depolarization and a subsequent activation of VDCCs in the presynaptic terminals. These results indicate the reason why the presence of ouabain enables the Na⁺-free solution to transiently increase mIPSC frequency; it is the ouabain-induced accumulation of Na⁺ in the presynaptic terminals rather than being due to any direct ouabain-induced depolarization.

View larger version (41K): [in this window] [in a new window]   Fig. 4. Comparison of the effects of ouabain and high-K+ solution on mIPSCs in Na+-free solution. A: typical current trace of mIPSCs in Na+-free solution illustrating the effects of brief applications of ouabain () and 10 mM K+ (). Sections of the control, ouabain, and high-K+ traces, as indicated (- - -), are shown at greater resolution. B: mean values of mIPSCs frequency (left) and amplitude (right) during ouabain and high-K+ perfusion. Data are expressed relative to the control values, error bars represent SE (n = 5; *, P < 0.05).

Involvement of VDCCs and Ca²+ stores in the transient increase of mIPSCs frequency

The preceding results (Fig. 4) suggested that Na⁺-free solution in the presence of ouabain did not cause a significant membrane depolarization. In fact, the perfusion of Na⁺-free solution may result in a hyperpolarization of the presynaptic membrane as observed for the postsynaptic membrane (e.g., Fig. 1). Nevertheless because the nerve terminals were not voltage-clamped, we wished to further investigate the possible participation of Ca²⁺ influxes into the terminals via VDCCs in the observed increase in mIPSC frequency (Fig. 3). Thus we repeated the experiments in the continued presence of Cd²⁺, a nonselective blocker of VDCCs. The addition of 10⁴ M Cd²⁺ to standard solution containing ouabain itself decreased mIPSCs frequency (Fig. 5A). However, mIPSCs frequency was still transiently increased by the subsequent removal of external Na⁺ (period a in Fig. 5, A-C: 3.07 ± 0.40, P < 0.05). Moreover, the increase in mIPSCs frequency was significantly elevated throughout the entire period of Na⁺-free solution (period b in Fig. 5, A-C: 1.61 ± 0.19, P < 0.05).

View larger version (45K): [in this window] [in a new window]   Fig. 5. Effects of blocking VDCCs on the facilitation of mIPSCs frequency by altered [Na+]i. A, top: typical raw data current trace showing the effect of Na+-free solution in the continued presence of 104 M Cd2+ and 104 M ouabain. Bottom: time course of mIPSCs frequency before, during, and after application of Na+-free solution in the continued presence of both ouabain and Cd2+. Each point represents mean ± SE, grouped into 10-s epochs and expressed relative to the control frequency (n = 8). B: relative mIPSCs frequency (left) and amplitude (right) distributions. C: mean values of mIPSCs frequency (left) and amplitude (right) expressed relative to those obtained in the control solution (*P < 0.05). D: current responses induced by exogenous application of 104 M GABA in the presence of ouabain and Cd2+ and in the presence of ouabain, Cd2+, and Na+-free solution. Top: typical raw data traces; bottom: mean data ± SE from 6 experiments.

In contrast to the control conditions (Fig. 3), the application of Na⁺-free solution in the presence of ouabain and Cd²⁺ caused a transient but significant increase in mIPSCs amplitude (Fig. 5, B and C; period a: 1.38 ± 0.11, P < 0.05; period b: 1.01 ± 0.01). Such an increase could be caused by an increase in the postsynaptic GABA_A receptor sensitivity, and therefore the postsynaptic responses induced by exogenous application of GABA were examined. When Na⁺ in the ouabain- and Cd²⁺-containing solution was replaced with N-methyl-D-glucamine, 10⁴ M GABA-induced currents were 1.80 ± 0.20 nA at 30 s (corresponding to period a) and 1.69 ± 0.20 nA at 3.5 min (corresponding to period b). These values were not significantly different from those recorded under control conditions (1.60 ± 0.18 nA; Fig. 5D; P = 0.69 at 30 s and P = 0.76 at 3.5 min, n = 6). These results indicate that the transient increase of mIPSCs amplitude in the Na⁺-free solution containing both Cd²⁺ and ouabain was not due to an increase in postsynaptic GABA_A receptor sensitivity. The increase more likely arises from unresolved overlapping mIPSCs caused by a rapid elevation of [Ca²⁺]_i in the presynaptic terminals and a similar rapid increase in GABA release. Thus the Ca²⁺ influx seen under these conditions of reversed Na⁺ gradients is not due to activation of VDCCs but more likely reflects reverse mode NCX.

The elevation of [Ca²⁺]_i in the presynaptic terminals may also be caused by Ca²⁺ release from presynaptic Ca²⁺ stores (Berridge 1998). [Ca²⁺]_i increase mediated by reverse mode NCX (a consequence of the reversed Na⁺ driving force in the presynaptic terminals) could activate Ca²⁺-induced Ca²⁺ release that would further amplify the [Ca²⁺]_i elevation (Litwin et al. 1996). Cyclopiazonic acid (CPA), a sarco-endoplasmic reticulum Ca²⁺ pump inhibitor, was used to explore the role of Ca²⁺-induced Ca²⁺ release in the mIPSCs frequency facilitation. The addition of 10⁵ M CPA to standard solution with ouabain and Cd²⁺ by itself caused a small increase in mIPSCs frequency (Fig. 6A). Subsequent removal of external Na⁺ again induced both a brief transient and a smaller sustained increase in mIPSCs frequency (Fig. 6, A-C; periods a and b, respectively). mIPSCs frequency during the first 2 min (period a) and during the subsequent minutes (period b) were both significantly higher than those in control. (Fig. 6C; period a: 3.96 ± 0.71, P < 0.05; period b: 1.99 ± 0.25, P < 0.05). Again there was an apparent increase in mIPSCs amplitude (Fig. 6, B and C; a: 1.20 ± 0.04, P < 0.05; b: 1.03 ± 0.01, n = 9), and thus the response to exogenous GABA application under these conditions was again examined. GABA-induced currents corresponding with period a (1.75 ± 0.18 nA) and period b (1.66 ± 0.17 nA) were not significantly different from those of control (1.56 ± 0.18 nA; Fig. 6D; P = 0.73, n = 7). The data indicate that the transient increase in GABA release is mainly caused by Ca²⁺ influx induced by reversing the Na⁺ driving force rather than by [Ca²⁺]_i amplified by Ca²⁺-induced Ca²⁺ release.

View larger version (43K): [in this window] [in a new window]   Fig. 6. Role of sarco-endoplasmic reticulum Ca2+ pump in mIPSCs frequency facilitation. A, top: typical raw data current trace showing the effect of Na+-free solution in the continued presence of cyclopiazonic acid (CPA), 104 M Cd2+, and 104 M ouabain. The top and bottom traces are continuous. Bottom: time course of mIPSCs frequency before, during, and after application of Na+-free solution in the continued presence of CPA, ouabain, and Cd2+. Each point represents means ± SE, grouped into 10-s epochs and expressed relative to the control frequency (n = 9). B: relative mIPSCs frequency (left) and amplitude (right) distributions during the times as indicated (). C: mean values of mIPSCs frequency (left) and amplitude (right) in Na+-free solution, expressed relative to those values obtained in the control solution (*P < 0.05). D: current responses induced by exogenous application of 104 M GABA in the presence of CPA, ouabain and Cd2+ and in the presence of CPA, ouabain, Cd2+, and Na+-free solution. Top: typical raw data traces; bottom: mean data ± SE from 7 experiments.

Evidence for reverse mode NCX in the presynaptic nerve terminal

As shown in Fig. 3, Na⁺-free solution increased mIPSCs frequency in the presence of ouabain. It seems that Ca²⁺ influx via reverse mode NCX induced by reversing the Na⁺ driving force mediates this response. If this hypothesis is correct, a transient increase of mIPSC frequency would not be induced by Na⁺-free solution in conditions in which reverse mode NCX is inhibited. Accordingly, the following two methods to inhibit the reverse mode NCX were examined: the removal of external Ca²⁺ that is required for the reverse mode NCX and the use of an inhibitor of reverse mode NCX, KB-R7943 (Iwamoto et al. 1996).

When external Ca²⁺ was removed in the presence of ouabain and CPA, both the frequency and amplitude of mIPSC were reduced (Fig. 7A). When Na⁺ was subsequently removed from the external solution, there was now no change in either the frequency or amplitude of mIPSCs (Fig. 7B). These results show that the mechanism mediating the transient increase of mIPSCs frequency requires external Ca²⁺.

View larger version (39K): [in this window] [in a new window]   Fig. 7. Inhibition of Na+/Ca2+ exchanger in Ca2+-free solution. A: typical raw current trace showing the lack of effects of Na+-free solution on mIPSCs when applied in the absence of external Ca2+ (CPA and ouabain present throughout). B: relative mIPSCs frequency (left) and amplitude (right) distributions, using the data shown in A.

We next examined the modulatory effect of KB-R7943 on the GABAergic mIPSCs. In the standard control solution, the application of KB-R7943 at concentrations less than 3 × 10⁶ M had little effect on either mIPSCs frequency (1.19 ± 0.08 at 3 × 10⁶ M KB-R7943) or amplitude (0.94 ± 0.04 at 3 × 10⁶ M KB-R7943). Higher doses (10⁵ M) of KB-R7943 increased mIPSCs frequency (1.94 ± 0.21, P < 0.05) and decreased mIPSCs amplitude (0.66 ± 0.04, P < 0.05; Fig. 8A). KB-R7943, at concentrations >10⁵ M, also reversibly suppressed the response to exogenous (10⁴ M) GABA application (Fig. 8B). The results indicate that the decrease in mIPSCs amplitude by 10⁵ M KB-R7943 (Fig. 8A) is due to the direct depression of the GABA_A receptor response. As the IC₅₀ of KB-R7943 for inhibition of reverse mode NCX is ~10⁶ M (Iwamoto et al. 1996), 3 × 10⁶ M KB-R7943 was used in the following experiments as a compromise between the inhibitory effect on the postsynaptic membrane and the inhibitory effect on the reverse mode NCX.

View larger version (33K): [in this window] [in a new window]   Fig. 8. Effect of KB-R7943 on mIPSCs and exogenous GABA response. A, a and b: typical raw current trace showing the effects of a low dose (a) and higher dose (b) of KB-R7943 on mIPSCs. Note the expanded sections of b. c: mean relative mIPSCs frequency and amplitude in the presence of various concentrations of KB-R 7943 (*P < 0.05). B: effects various concentrations of KB-R7943 on the current response to applied GABA.

The addition of 3 × 10⁶ M KB-R7943 to standard solution containing ouabain, Cd²⁺, and CPA induced little change in mIPSC frequency or amplitude of mIPSCs (Fig. 9A). However, the transient increase in mIPSCs frequency induced by the Na⁺-free solution was eliminated in the presence of 3 × 10⁶ M KB-R7943. The more sustained and smaller increase in mIPSC frequency (seen in period b in Figs. 4 and 5) was still observed in the presence of 3 × 10⁶ M KB-R7943 (Fig. 9; 1.26 ± 0.05, P < 0.05). The significant and transient increase in mIPSCs amplitude seen in these conditions on removal of Na⁺ (Figs. 5 and 6) was also abolished by KB-R7943 (Fig. 9, B and C; 0.98 ± 0.03, P = 0.40). The results suggest that a KB-R7943-sensitive mechanism, which is responsible for the observed transient increase in mIPSCs frequency, is present in these presynaptic terminals.

View larger version (63K): [in this window] [in a new window]   Fig. 9. Inhibitory effect of KB-R7943 on the transient increase of mIPSCs frequency. A, top: typical current recording showing the effects of application of Na+-free solution in the continued presence of ouabain, Cd2+, CPA, and KB-R7943. Bottom: time course of relative mIPSCs frequency before, during, and after application of a Na+-free solution in the continued presence of ouabain, Cd2+, CPA and KB-R7943. Each point represents the mean ± SE from 7 experiments. B: relative mIPSCs frequency (left) and amplitude (right) distributions of the data shown in A. C: mean mIPSCs frequency (left) and amplitude (right) expressed relative to the control values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reverse mode Na+/Ca²⁺ exchanger in nerve endings

It has been proposed that the forward mode NCX, working together with the plasma membrane Ca²⁺ and sarco-endoplasmic reticulum Ca²⁺ pumps, is responsible for the removal of elevated [Ca²⁺]_i in presynaptic nerve terminals (Blaustein and Lederer 1999; Reuter and Porzig 1995). Attempts to isolate a particular role for the forward mode NCX are complicated by these other parallel pathways. Furthermore because [Ca²⁺]_i is usually maintained at very low levels in resting cells, it would be difficult to demonstrate the existence of forward mode NCX using a decease in the low basal rate of mIPSCs frequency as an indicator of a decrease in [Ca²⁺]_i. We attempted to measure forward mode NCX by comparing the decay phase of the transient increase in mIPSCs frequency induced by depolarization (high-K⁺ solution) in the absence and presence of KB-R7943. However, differences between the experiments with and without KB-R7943 were not clearly apparent (data not shown). Perhaps the rapid intracellular buffering systems dominate in the regulation of presynaptic [Ca²⁺]_i under these conditions so that any contribution of forward mode NCX is difficult to identify (using mIPSCs frequency). Hence to elucidate the presence of NCX in the presynaptic terminals, we utilized the reverse mode NCX that results in increases in presynaptic [Ca²⁺]_i.

VDCCs, ligand-gated cation channels, reverse mode NCX, IP₃-induced Ca²⁺ release, and Ca²⁺-induced Ca²⁺ release are all mechanisms that could cause an increase in [Ca²⁺]_i. Of these systems, VDCCs, ligand-gated cation channels and reverse mode NCX all utilize external Ca²⁺. Thus when VDCCs and ligand-gated cation channels are both blocked, then reverse mode NCX is the only pathway capable of bringing external Ca²⁺ into the nerve endings. Moreover, Ca²⁺ release from internal storage sites can be suppressed pharmacologically. Therefore under the appropriate experimental conditions, one can use increases in mIPSCs frequency to directly reflect Ca²⁺ influx due solely to reverse mode NCX.

Application of the Na⁺-free solution was expected to activate reverse mode NCX, resulting in elevated presynaptic [Ca²⁺]_i and enhanced neurotransmitter release. However, a clear increase of mIPSCs frequency on removal of external Na⁺ was only observed in the continued presence of ouabain (Figs. 2 and 3). The following possibility is considered to explain these results. Because the Na⁺/K⁺ ATPase maintains [Na⁺]_i at a very low level (Rose and Ransom 1997), the Na⁺-free external solution by itself could not reverse the Na⁺ driving force enough to induce reverse mode NCX. Ouabain causes an increase in [Na⁺]_i in presynaptic terminals reducing the driving force for inward Na⁺ flux (Blaustein 1993; Mulkey and Zucker 1992; Munakata et al. 1998; Rose and Ransom 1997). In the presence of ouabain, Na⁺-free solution could now induce a transient increase in mIPSCs frequency (Fig. 3), suggesting that low [Na⁺]_i is maintained by Na⁺/K⁺ ATPase under normal conditions. Thus it was difficult to turn forward mode NCX into reverse mode NCX by the simple removal of external Na⁺ without a concomitant increase of the [Na⁺]_i.

External Na⁺ is required for the re-uptake of GABA into the presynaptic terminals, since the GABA transporters utilize the inward Na⁺ driving force (Lu and Hilgemann 1999). Under the absence of external Na⁺, the amount of GABA repackaged into the vesicles may decrease. In the present study, however, Na⁺-free solution by itself had no effect on the amplitude or frequency of the GABAergic mIPSCs (Fig. 1B). Perhaps under our experimental conditions GABA actions are terminated by diffusion and/or that the vesicle stores are not depleted or that GABA level in the terminal cytoplasm is sufficient to replenish recycled vesicles. Whatever role GABA transporters have on the mIPSCs in the present study, it is clear that any effect of the Na⁺-free solution on the GABA transporter does not contribute to the increase in mIPSCs frequency observed in this study.

CPA induces Ca²⁺ store-operated Ca²⁺ entry (Emptage et al. 2001) and in our conditions caused a small increase in basal mIPSCs frequency (Fig. 6A). In the continued presence of CPA, mIPSCs frequency was still markedly increased by the Na⁺-free solution, as shown in Fig. 6A, a. In addition, Hoth and Penner (1993) reported that replacing Na⁺ with N-methyl-D-glucamine reduced the amount of Ca²⁺ influx via the store-operated Ca²⁺ entry in rat mast cells. These results suggest that Ca²⁺ entry pathway activated by Na⁺-free solution is not the same as mediates store-operated Ca²⁺ entry.

NCX in the postsynaptic membrane can also be activated by the experimental protocol used to activate presynaptic reverse mode NCX activity (Danaceau and Lucero 2000). However, any change in postsynaptic NCX activity is of little consequence in the interpretation of the results from the present study because the whole cell patch-clamp recording technique used here provides accurate control of the postsynaptic [Na⁺]_i and [Ca²⁺]_i. Furthermore, as during N-methyl-D-glucamine perfusion, both the intra- and extracellular postsynaptic concentration of Na⁺ was nominally zero, the NCX at the soma could hardly function as either forward mode NCX or reverse mode NCX. Taken together with the results of the exogenous GABA-induced responses, it is most likely that the transient increase of mIPSCs amplitude induced by the Na⁺-free solution (Figs. 5 and 6) was a consequence of the high rates of GABA release so that multiple mIPSCs overlapped rather than due to an increased response of postsynaptic GABA_A receptors.

The application of Na⁺-free solution increased mIPSCs frequency more persistently in the presence of both ouabain and Cd²⁺ than in the presence of ouabain alone. We propose the following explanation. In the presence of Cd²⁺, the basal influx of Ca²⁺ through VDCCs is decreased and the only mechanisms allowing for Ca²⁺ influx is via reverse mode NCX, hence the presynaptic terminals will hardly be saturated with Ca²⁺. The lower absolute [Ca²⁺]_i will prolong the duration of Ca²⁺ concentration gradients across the presynaptic membrane and hence the reverse mode NCX and the relative facilitation of GABA release might continue longer. The reduction in basal mIPSCs frequency by Cd²⁺ may also enable a more persistent subsequent elevation of GABA release due to differences in the size of the pool of readily releasable vesicles. In some experiments (e.g., Fig. 5), Cd²⁺ also caused a slight decrease in mIPSCs amplitude. While this was not further investigated, it is likely to reflect the ability of Cd²⁺ to directly inhibit the GABA_A receptor complex (Nakagawa et al. 1991).

Inhibition of reverse mode NCX in the presynaptic terminals

We used two strategies to inhibit the reverse mode NCX activity observed in this study: i.e., removal of external Ca²⁺ (Fig. 7A) and use of KB-R7943 (Fig. 9). Figure 7A indicates that the transient increase in mIPSCs frequency induced by a reversed Na⁺ driving force requires extracellular Ca²⁺.

KB-R7943 has been reported as a specific inhibitor of reverse mode NCX (Iwamoto et al. 1996). According to this report, the KB-R7943 concentration of 3 µM employed in this present study should be enough to suppress most of reverse mode NCX. The results from Fig. 9 demonstrate that, in the presence of ouabain, removal of external Na⁺ failed to increase mIPSCs frequency in the presence of KB-R7943, clearly implicating reverse mode NCX in the presynaptic bouton.

The time course of changes in mIPSCs frequency during Na⁺-free solution presumably reflects the changes in Ca²⁺ influx mediated through reverse mode NCX. As reverse mode NCX is stimulated by the Na⁺-free external solution, Na⁺ is pumped out of the nerve terminal. Thus [Na⁺]_i, initially increased by ouabain pretreatment, may gradually decline during reverse mode NCX. As a consequence, reverse mode NCX may gradually decrease, and Ca²⁺ influx and the extent of mIPSCs frequency facilitation, would also decrease. This is consistent with the known dependence of NCX on the simultaneous presence of both Ca²⁺ and Na⁺ in the intra- and extracellular spaces (Blaustein and Lederer 1999).

Functional role of NCX in the presynaptic terminal

As also pointed out by Blaustein and Lederer (1999), the question of whether NCX actually functions in both forward and reverse modes under physiological conditions is difficult to answer. We presume that, in typical physiological conditions, NCX in nerve endings operates in the forward mode using the inward Na⁺ driving force to pump out elevated [Ca²⁺]_i during excitation. High doses of KB-R7943 (10⁵ M) were actually observed to increase mIPSCs frequency (Fig. 8A). While the specificity of KB-R7943 at these higher doses needs further investigation this action would be consistent with an inhibition of forward mode NCX. Furthermore, lower doses of KB-R7943 (3 × 10⁶ M) increased the paired pulse facilitation of evoked IPSCs in Meynert neurons recorded in brain slices at 30°C (data not shown). These data are also supportive of a role for forward mode NCX in presynaptic [Ca²⁺]_i homeostasis under these somewhat more physiological conditions. In combination with other mechanisms maintaining low [Ca²⁺]_i, NCX might work as a net [Ca²⁺]_i extruder and prevent an excess of [Ca²⁺]_i in the presynaptic terminals.

As seen in the experiments using ouabain, the increased [Na⁺]_i facilitated reverse mode NCX, or at least reduced the ability of forward mode NCX to extrude Ca²⁺. Repetitive electrical stimulation or a massive activation of ionotropic glutamate receptors in neurons that may occur during intense excitation or during ischemia, leads to a rapid and significant increase of [Na⁺]_i, which decreases forward mode NCX activity and Ca²⁺ extrusion, resulting in an increase in cytosolic [Ca²⁺]_i (Hirono et al. 1998; Hoyt et al. 1998). Both VDCCs and NMDA receptors have been postulated as major Ca²⁺ entry pathways during anoxia/ischemia (Tymianski and Tator 1996). In addition, it was recently reported that reverse mode NCX plays an important role in facilitating intracellular Ca²⁺ overload and the subsequent induction of irreversible damage after anoxic/ischemia and traumatic injury in spinal neurons (Li et al. 2000) and hippocampal neurons (Breder et al. 2000). In the present study, we could not confirm the functional reverse mode NCX in presynaptic nerve terminals under physiological conditions. However, there is a possibility that the reverse mode NCX might contribute to the presynaptic Ca²⁺ accumulation and the subsequent deleterious increase of transmitter release under these pathological circumstances.