INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In rat hippocampal CA1 neurons obtained from tissue slices, intracellular recordings revealed that superfusion with oxygen- and glucose-deprived medium (in vitro ischemia) produces a rapid depolarization at 5-6 min after onset (Tanaka et al. 1997; Uchikado et al. 2000). The rapid depolarization corresponds to a rapid negative DC potential (the rapid negative potential) recorded extracellularly from the CA1 region (Rader and Lanthorn 1989; Uchikado et al. 2000). When oxygen and glucose are immediately reintroduced after the onset of the rapid depolarization, the membrane potential becomes persistently depolarized, reaching 0 mV after 5 min (Tanaka et al. 1997). On the other hand, the extracellular negative potential transiently diminished immediately after the reintroduction of oxygen and glucose was started; however, a second slow negative DC potential (slow negative potential) occurs within 30 s (Uchikado et al. 2000).

The slow negative potential is likely generated by nonneuronal cells, since a second application of in vitro ischemia to the same slice produces a similar slow negative potential (Uchikado et al. 2000). In addition, a prolonged application (>10 min) of ischemia-simulating medium or application of either oxygen-free and glucose-free medium after ischemic exposure decreases the amplitude and slope of the slow negative potential, thus suggesting that the slow negative potential is energy-dependent. The maximal slope of the slow negative potential is decreased by a reduction in the external Na⁺ or addition of Co²⁺; however, it is not altered by a reduction in external Cl or K⁺. Neither antagonists for ionotropic glutamate (Glu) receptors nor Glu transporter antagonists affect the maximal slope. Na⁺/Ca²⁺ exchanger blockers, Ni²⁺ and benzamil hydrochloride, reversibly suppress the maximal slope in a dose-dependent manner, thus suggesting that the slow negative potential is due to the forward-mode operation of the Na⁺/Ca²⁺ exchanger after the reintroduction of oxygen and glucose (Uchikado et al. 2000).

An elevation in the intracellular Ca²⁺ ([Ca²⁺]_i) concentration during and after ischemic exposure have been reported in in vivo (Morris et al. 1985; Silver and Erecinska 1990; Uematsu et al. 1988) and in vitro (Ebine et al. 1994; Tanaka et al. 1997) experiments on various brain regions. In in vivo ischemia, the elevated [Ca²⁺]_i in rat hippocampal region is recovered to the preexposure level after cerebral blood flow is restored (Silver and Erecinska 1990). Similarly, in vitro experiments have shown that [Ca²⁺]_i signal in rat hippocampal CA1 region rises rapidly during the rapid depolarization and falls again after the reintroduction of oxygen and glucose (Ebine et al. 1994; Tanaka et al. 1997). The rapid rise of [Ca²⁺]_i is due to an increase in the nonselective ion permeability during the rapid depolarization (Tanaka et al. 1997). However, the mechanisms underlying the recovery of [Ca²⁺]_i signal after ischemia either in vivo and in vitro are still unclear. The recovery of [Ca²⁺]_i signal after ischemia may be due to the extrusion of Ca²⁺ from glial cells, since the CA1 neurons do not show any potential recovery after the reintroduction of oxygen and glucose (Tanaka et al. 1997, 1999). At least three mechanisms would be possible for the elimination of Ca²⁺ from the cytosol of glial cells after ischemic exposure: the reactivation of either Ca²⁺-ATPase or Na⁺/Ca²⁺ exchanger and the uptake by both endoplasmic reticulum and mitochondria.

The present study has addressed the relationship between the recovery of the [Ca²⁺]_i signal and the slow negative potential after in vitro ischemia in slices obtained from the rat hippocampal CA1 region. We mainly examined the effects of various extracellular ion concentrations and the application of an antagonist of Na⁺/Ca²⁺ exchangers on the recovery of the [Ca²⁺]_i signal after ischemic exposure. The rising and falling phases of the [Ca²⁺]_i signal corresponded to the rapid depolarization and a depolarizing hump in the repolarizing phase of glial cells, respectively. When oxygen and glucose were reintroduced to slices in the Na⁺-free or ouabain- or Ni²⁺-containing medium, the falling phase was suppressed. The falling phase was significantly accelerated in Ca²⁺- and Mg²⁺-free with EGTA-containing medium. In contrast, the falling phase was significantly slower in the Ca²⁺-free with a high Mg²⁺- and EGTA-containing medium. The results suggest that the forward-mode operation of Na⁺/Ca²⁺ exchangers in glial cells induces intracellular Ca²⁺ extrusion, which thus restores the [Ca²⁺]_i signal to its preischemic exposure level.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The forebrains of adult Wistar rats (male 250-350 g) were removed quickly under ether anesthesia and placed in chilled (4-6°C) Krebs solution, which was aerated with 95% O₂-5% CO₂. The composition of the Krebs solution was as follows (in mM): 117 NaCl, 3.6 KCl, 2.5 CaCl₂, 1.2 MgCl₂, 1.2 NaH₂PO₄, 25 NaHCO₃, and 11 glucose. The hippocampus was dissected and then sliced with a vibratome (Oxford) at a thickness of 400 µm. A slice was placed on thin glass (thickness, 170 µm) at the bottom of a recording chamber (volume, 500 µl) and stabilized with a titanium grid placed on the upper surface of the section. The preparation was completely submerged in the superfusing solution. The temperature in the recording chamber was continuously monitored and maintained at 36.0 ± 0.5°C, and solutions were perfused at a rate of 4-6 ml/min.

Extracellular recordings from the proximal part of the stratum radiatum (~300 µm from the stratum pyramidale) in the CA1 region were made with glass micropipettes filled with Krebs solution and having resistances of 100-140 M, because the slow negative potential is most prominent in the proximal part of the stratum radiatum in the CA1 region (Uchikado et al. 2000). It took ~60 min to stabilize the DC potential level after the recording electrode was inserted into the tissue slice. For this reason, we allowed 60 min before starting the recordings. In some experiments, simultaneous recordings were made of the extracellular DC potentials and intracellular membrane potentials recorded from glial cells. Intracellular recordings were made with glass micropipettes filled with 2 M K acetate and having resistances of 90-120 M. Intracellular recordings from glial cells were identified based on the following criteria: 1) a very negative resting membrane potential (~ 90 mV), a low input resistance (1-7 M), and a short membrane time constant (<1 ms); 2) the absence of synaptic responses; and 3) no excitation by even very large depolarizing pulses (4 nA) (Leblond and Krnjevic 1989; Schwartzkroin and Prince 1979).

The slices were made "ischemic" by superfusing them with medium equilibrated with 95% N₂-5% CO₂ and deprived of glucose, which was replaced with NaCl isoosmotically (ischemia-simulating medium). The low Na⁺ or Na⁺-free medium and low Ca²⁺ medium were made by replacing NaCl with Tris-Cl and by replacing CaCl₂ with MgCl₂, respectively. For experiments using low Na⁺ and Na⁺-free media, HCO was omitted, and HEPES (10 mM) was used as the buffer (pH 7.4 with Tris), and solutions were bubbled with 100% O₂. Ca²⁺-free, EGTA (5 mM)-containing medium and Ca²⁺- and Mg²⁺-free, EGTA-containing medium were made by omitting Ca²⁺ and adding Mg²⁺ (10 mM), and by omitting both Ca²⁺ and Mg²⁺, respectively. When switching between superfusing media, there was a delay of 15-20 s before the new medium reached the chamber due to the volume of the connecting tubing. As a result, the chamber was filled with the test solution ~30 s after switching the solution.

[Ca²⁺]_i was measured by incubating tissue slices with the fluorescent Ca²⁺ indicator, 1-[6-amino-2-(5-carboxy-2-oxazolyl)-5-benzofuranyloxy]-2-(2-amino-5-methylphenoxy)-ethane-N,N,N',N'-tetraacetic acid penta-acetoxymethyl ester (Fura-2/AM). Fluoroprobe loading was performed by soaking the slice in a solution containing Fura-2/AM (10 µM) for 60-70 min and then washing the slice in the Krebs' solution for 10 min at 33-34°C. The slice was then placed in the recording chamber and mounted on an inverted epifluorescence microscope (Nikon TMD) equipped with a xenon lamp and band-pass filters of 340 ± 5 nm (wavelength that allows activation of a Ca²⁺-dependent increase in the signal) and 380 ± 5 nm (resulting in a Ca²⁺-dependent decrease in the signal). Changes in the fluorescence intensities in the proximal site of stratum radiatum (area of measurement was 50 × 250 µm) of the CA1 region was measured by microspectrofluorometry (CAM-220, Japan Spectroscopic) using an alternative excitation at 340 and 380 nm (Ebine et al. 1994; Kudo and Ogura 1986; Kudo et al. 1986, 1987). In the present study, changes in the ratio of fluorescence intensities induced by excitation at wavelengths of 340 and 380 nm (R340/380) were used to monitor any changes in [Ca²⁺]_i, because the dissociation constant of Ca²⁺-Fura-2 complex in the cytosol cannot be precisely determined in brain slice preparations and also due to the fact that the autofluorescence of the slices changes during hypoxia (Sick and Rosenthal 1989).

As demonstrated in Fig. 1B, bottom, the latency (L) for the elevation of [Ca²⁺]_i signal was measured from the onset of ischemia to the onset of the rising phase of the [Ca²⁺]_i signal. The recovery level (RL) of [Ca²⁺]_i signal after ischemic exposure was arbitrarily taken 5 min from the peak of the point. The peak amplitude of [Ca²⁺]_i signal was measured between the peak (P) and the preexposure baseline level (BL). The amplitude of the falling phase of [Ca²⁺]_i signal was measured between the peak and the RL of [Ca²⁺]_i signal. The recovery ratio of the [Ca²⁺]_i signal was calculated as the ratio of the amplitude of the falling phase to the peak amplitude. The maximal slopes of both the rising and the falling phases of the [Ca²⁺]_i levels (SR and SF, respectively) were also measured.

View larger version (20K): [in this window] [in a new window]   Fig. 1. Simultaneous recordings of the changes in the extracellular DC potential and either intracellular glial membrane potential or the [Ca2+]i signal in the hippocampal CA1 region following ischemic exposure. In this and subsequent figures, the ischemia-simulating medium was applied between the times indicated by the downward and upward arrows. In each trace, the dotted line indicates the preexposure level. A: simultaneous extracellular recordings (top) and intracellular (bottom) recordings from a glial cell. Note that the slow negative potential and the depolarizing hump occurred simultaneously. In the bottom trace, downward deflections are hyperpolarizing electrotonic potentials elicited by anodal current pulses (0.4 nA for 200 ms every 3 s). B: simultaneous extracellular recordings (top trace) and recording of the ratio of 1-[6-amino-2-(5-carboxy-2-oxazolyl)-5-benzofuranyloxy]-2-(2-amino-5-methylphenoxy)-ethane-N,N,N',N'-tetraacetic acid penta-acetoxymethyl ester (Fura-2) fluorescence intensities (R340/380) (bottom trace). Parameters measured are shown in bottom trace: latency (L), peak (P), and maximal slope (SR) of the rapid increase (the rising phase) in the [Ca2+]i signal, and maximal slope (SF) of the rapid decrease (the falling phase), the level of the [Ca2+]i signal 5 min after the onset of the decrease (RL), and the preexposure baseline level (BL). The recovery ratio of R340/380 was calculated as the ratio of the amplitude of the reduction in R340/380 (P  RL) to that of the peak of R340/380 (P  BL). C: the relationship between the maximal slope of a slow negative potential and that of the falling phase of R340/380 from similar experiments as those shown in B. Correlation coefficient is expressed as R.

The drugs used were benzamil hydrochloride, HEPES, EGTA, ouabain (all from Sigma Chemical), and Fura-2/AM (from Dojin).

Four to five hippocampal slices were taken from each rat. Two to three slices were used for the controls and two to three slices were used for the test solutions to compare the results in the same animal. We used one slice for one experiment since the responses to superfusion with ischemia-simulating medium could not be reproduced after the first ischemic exposure. Slices were pretreated with media containing test compounds or various ionic media for 10 min before ischemic exposure, unless specified otherwise. All quantitative results were expressed as the means ± SE. The number of slices examined is given in parentheses. An analysis of variance (ANOVA) with the Scheffé post-hoc tests was used to compare the data, with P < 0.05 considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Changes in DC potential and [Ca²⁺]_i induced by ischemia-simulating medium

Figure 1A shows the simultaneous recordings of the changes in the membrane potential recorded intracellularly from a glial cell and in the DC potential in the CA1 region following superfusion with oxygen- and glucose-free medium (ischemia-simulating medium). Ischemic exposure produced a rapid depolarization ~5 min later. The reintroduction of oxygen and glucose transiently repolarized, then depolarized the membrane (a depolarizing hump) within 1 min, and finally restored to the preexposure level. These potential changes in the glial cell corresponded to changes in the DC potential. A rapid negative-going DC potential (rapid negative potential) during ischemic exposure and a slow negative-going DC potential (slow negative potential) after exposure corresponded to the rapid depolarization and the depolarizing hump, respectively.

Figure 1B shows the simultaneous recordings of the changes in the DC potential and in the ratio of fluorescence intensities recorded from the stratum radiatum in the CA1 region following ischemic exposure of the same Fura-2-loaded slice preparation. The intracellular Ca²⁺ ([Ca²⁺]_i) signal during ischemic exposure consisted of an initial increased [Ca²⁺]_i signal and a subsequent rapid increased [Ca²⁺]_i signal (rising phase of [Ca²⁺]_i signal), which corresponded to an initial positive-going DC potential and a subsequent rapid negative-going DC potential, respectively (Fig. 1B, bottom trace). When both oxygen and glucose were reintroduced, the elevated [Ca²⁺]_i showed a further slow increase and reached a peak 30 s to 1 min after the reintroduction. Subsequently, the [Ca²⁺]_i signal began to fall rapidly (falling phase of [Ca²⁺]_i signal), which occurred simultaneously with the slow negative potential, but was never restored to the preexposure level. Similar results were obtained in 35 other slices.

Figure 1C shows the relationship between the maximal slope of the slow negative potential and that of the falling phase of [Ca²⁺]_i signal (indicated as SF in the figure) observed in 35 slices. The results showed a linear correlation between the maximal slope of the slow negative potential and the maximal slope of the falling phase of [Ca²⁺]_i signal (correlation coefficient, 0.79, P < 0.02). The maximal slope of the slow negative potential was also related linearly with the peak amplitude, but the peak amplitude varied in each slice preparation and the maximal slope was relatively constant (Uchikado et al. 2000). In the following section, we describe the effects of various superfusing media on the falling phase of [Ca²⁺]_i signal, since their effects on the maximal slope of the slow negative potential have been previously reported (Uchikado et al. 2000).

Dependence of reduction in [Ca²⁺]_i after ischemic exposure on energy supply

Figure 2 illustrates the effects on the falling phase of [Ca²⁺]_i signal of reintroduction of oxygen and glucose at various times after the rising phase of [Ca²⁺]_i signal. As reintroduction of oxygen and glucose was delayed, the falling phase became slower (Fig. 2A), and both the maximal slope of the falling phase and the recovery ratio of [Ca²⁺]_i signal were significantly reduced (Fig. 2, B and C). When oxygen and glucose were reintroduced 0.5, 4, and 8 min after generating the rising phase of the response, the maximal slope was 0.36 ± 0.06 min¹ (n = 7), 0.30 ± 0.06 min¹ (n = 10), and 0.12 ± 0.12 min¹ (n = 8, P < 0.01), respectively, and the recovery ratio was 52 ± 3% (n = 7), 53 ± 4% (n = 10), and 30 ± 2% (n = 8, P < 0.01), respectively.

View larger version (18K): [in this window] [in a new window]   Fig. 2. Changes in the [Ca2+]i signal produced by ischemic exposure for various periods in the hippocampal CA1 region. A: from top to bottom, oxygen and glucose were reintroduced 0.5, 4, and 8 min, respectively, after generating the rising phase of R340/380. B and C: mean maximal slopes of the falling phase of R340/380 and mean recovery ratios of R340/380 after the reintroduction of oxygen and glucose at various times (indicated by min). The vertical error bar on each column in B and C represents the standard error. Note that the maximal slope of the falling phase and the recovery ratio were significantly reduced when oxygen and glucose were reintroduced 8 min after generating the rising phase. * and **, significant differences with P < 0.05 and P < 0.01, respectively.

We next examined the effects of applying either oxygen or glucose on the falling phase of [Ca²⁺]_i signal. When an oxygen- or glucose-containing medium was applied after generating the rising phase, the falling phase became slower and the recovery ratio decreased in comparison with that observed after the reintroduction of both oxygen and glucose (Fig. 3A). Figure 3, C and D, shows the summary. The maximal slope was 0.36 ± 0.06 min¹ (n = 7) in control, 0.18 ± 0.01 min¹ (n = 6, P < 0.01) in glucose-free solution, and 0.18 ± 0.01 min¹ (n = 6, P < 0.01) in oxygen-free solution. The recovery ratio was 46 ± 4% (n = 7) in control, 10 ± 3% (n = 6, P < 0.01) in glucose-free solution, and 8 ± 5% (n = 6, P < 0.01) in oxygen-free solution. In addition, the maximal slope in oxygen-free solution was considerably reduced compared with that in the glucose-free solution (P < 0.05). The results suggest that the falling phase of [Ca²⁺]_i signal is energy-dependent.

View larger version (26K): [in this window] [in a new window]   Fig. 3. Energy-dependence of the rapid decrease in the [Ca2+]i signal in the hippocampal CA1 region. A: three superimposed recordings of R340/380 were obtained from different slice preparations before and after superfusion with ischemic-simulating medium. Immediately after generating the rising phase of R340/380, oxygen- and glucose-containing medium or either oxygen- or glucose-containing medium was applied. B and C: mean maximal slopes of the falling phase and mean recovery ratios of the [Ca2+]i signal, respectively, were obtained from similar experiments shown in A. Vertical error bar on each column in B and C represents standard errors. Note that the application of either oxygen or glucose after generation of the rising phase reduces the maximal slope of the falling phase and recovery ratio of the [Ca2+]i signal.

Involvement of reactivation of Na⁺, K⁺-ATPase in reduction of [Ca²⁺]_i after ischemic exposure

Uchikado et al. (2000) demonstrated that a prolonged application of ouabain (30 µM)-containing normoxic medium produces a similar rapid negative potential, but does not generate a slow negative potential. Figure 4A shows that a prolonged application of ouabain (30 µM) produced a similar rapid increase in [Ca²⁺]_i signal with that induced by in vitro ischemia (n = 6). After washing out ouabain, [Ca²⁺]_i signal was further increased and never decreased 30 min after (not shown in the figure). The latency for the generation of the rapid increase was 10.0 ± 1.2 min (n = 6), the peak amplitude was 1.95 ± 0.07 (n = 6), and the level of [Ca²⁺]_i 5 min after the onset of washout was 2.44 ± 0.09 (n = 6).

View larger version (21K): [in this window] [in a new window]   Fig. 4. Na+, K+-ATPase-dependence of the rapid increase and decrease in the [Ca2+]i signal during and after ischemic exposure, respectively, in the hippocampal CA1 region. A: effects of superfusion with ouabain (30 µM) in normoxic medium on R340/380. Note that ouabain produces a rapid increase in the [Ca2+]i signal during exposure, but does not generate a decrease in the [Ca2+]i signal after exposure. B: effects of the addition of ouabain (1 mM) in reperfusion medium with oxygen and glucose on the slow negative potential and the [Ca2+]i signal. Simultaneous extracellular DC recordings (top) and recordings of the [Ca2+]i signal (bottom) were made from the stratum radiatum. Note that the application of ouabain in the reperfusion medium blocks the generation of the slow negative potential and the decrease in the [Ca2+]i signal after ischemic exposure.

To examine the effects of ouabain applied in reperfusion medium with oxygen and glucose on the slow negative potential and the falling phase of [Ca²⁺]_i signal, simultaneous recordings of changes in the DC potential and in [Ca²⁺]_i signal following ischemic exposure were obtained, as shown in Fig. 4B. A high concentration (1 mM) of ouabain completely prevented the generation of both the slow potential and the falling phase of [Ca²⁺]_i signal. The [Ca²⁺]_i signal continuously increased after the reintroduction of oxygen and glucose. Similar results were also obtained in six other slices tested. The peak amplitude of the rising phase was 2.45 ± 0.02 (n = 7), and the level of [Ca²⁺]_i 5 min after the reintroduction of oxygen and glucose was 2.89 ± 0.03 (n = 7). The peak amplitude was not significantly changed; however, the level of [Ca²⁺]_i after 5 min was significantly higher (P < 0.0001) in comparison to that during and after ischemic exposure. These results suggest that the reactivation of Na⁺, K⁺-ATPase after reintroduction of oxygen and glucose is necessary for generating the falling phase of the [Ca²⁺]_i signal.

Effects of various ionic media and blocker of Na⁺/Ca²⁺ exchanger on changes in [Ca²⁺]_i during and after ischemic exposure

The present study showed a linear correlation between the maximal slope of the slow negative potential and that of the falling phase of the [Ca²⁺]_i signal. The slow negative potential has been reported to induce Na⁺/Ca²⁺ exchanger to operate in a forward mode (Uchikado et al. 2000). We were, however, unable to examine the effect of the Na⁺/Ca²⁺ exchanger blocker, benzamil hydrochloride, on the falling phase, since solutions containing 100 µM benzamil hydrochloride had a marked autofluorescence at wavelengths of 340 and 380 nm. In contrast, the medium containing 1-5 mM Ni²⁺, which blocks Na⁺/Ca²⁺ exchanger as well as Ca²⁺ channels, did not change the intensity of fluorescence induced by excitation at wavelengths of 340 and 380 nm.

Table 1 summarizes the values of the latency, peak amplitude, and maximal slopes of the rising and falling phases of [Ca²⁺]_i signal and of the recovery ratio in various ionic media or in Ni²⁺-containing medium. In low Ca²⁺ (0.25 mM)-, high Mg²⁺ (10 mM)-containing medium, the maximal slopes of the rising and falling phase, and the peak amplitude of the increase in [Ca²⁺]_i signal were significantly reduced in comparison with the controls. In low Na⁺ (28 mM)-containing medium, the maximal slopes of the rising and falling phase were significantly reduced but the peak amplitude was unaffected. In the presence of Ni²⁺ (1 mM), the latency was significantly prolonged and the peak amplitude was significantly reduced, as were the maximal slopes of the rising and falling phases. The recovery ratio of the [Ca²⁺]_i signal tended to decrease in low Ca²⁺-, a high Mg²⁺-containing medium, low Na⁺-containing medium, and Ni²⁺-containing medium, although there was no significant difference between control medium and these ionic media. The results suggest that the falling phase of the [Ca²⁺]_i signal depends on the extracellular Na⁺ ([Na⁺]_o) or Ca²⁺ ([Ca²⁺]_o) concentration and that the Na⁺/Ca²⁺ exchanger blocker, Ni²⁺, suppresses the falling phase.

Effects of various ions and a blocker of Na⁺/Ca²⁺ exchanger in reperfusion media on changes in [Ca²⁺]_i after ischemic exposure

We next also studied the effects on the falling phase of the [Ca²⁺]_i signal of various ions or Ni²⁺ applied in the reperfusion media with oxygen and glucose. After generating the increase in [Ca²⁺]_i, oxygen and glucose were reintroduced with Na⁺-free or Ni²⁺ (5 mM)-containing medium (Fig. 5A). The maximal slope of the falling phase was significantly reduced in either Na⁺-free or Ni²⁺ (5 mM)-containing medium compared with the controls (Fig. 5B). The recovery ratio of the [Ca²⁺]_i signal was significantly reduced in the Ni²⁺-containing medium (Fig. 5C). The maximal slope was 0.54 ± 0.06 min¹ (n = 15) in the control, 0.30 ± 0.06 min¹ (n = 8, P < 0.01) in Na⁺-free medium, and 0.24 ± 0.02 min¹ (n = 7, P < 0.01) in the Ni²⁺-containing medium. The recovery ratio was 56 ± 5% (n = 15) in the control, 48 ± 6% (n = 8) in Na⁺-free medium, and 43 ± 4% (n = 7, P < 0.01) in the Ni²⁺-containing medium.

View larger version (21K): [in this window] [in a new window]   Fig. 5. Either the removal of external Na+ or the addition of Ni2+ depresses the rapid decrease in the [Ca2+]i signal in the hippocampal CA1 region after ischemic exposure. A: after generating the rising phase of R340/380, oxygen and glucose were reintroduced in normal ionic medium, Na+-free medium, and Ni2+ (5 mM)-containing medium, respectively, from top to bottom trace. B and C: mean maximal slopes of the falling phase and mean recovery ratios of [Ca2+]i signal were obtained from similar experiments shown in A. Vertical error bar on each column in B and C represents standard errors. Note that the maximal slope of the falling phase was reduced in Na+-free or Ni2+-containing medium.

Figure 6A illustrates effects of the Ca²⁺-free and EGTA-containing media on the falling phase of [Ca²⁺]_i signal. After generating the rising phase of [Ca²⁺]_i signal, oxygen and glucose were reintroduced with the Ca²⁺-free and EGTA (5 mM)-containing media with or without Mg²⁺. The maximal slope of the falling phase was significantly reduced in the Ca²⁺-free, high Mg²⁺ (10 mM)- and EGTA-containing medium, but was significantly accelerated in Ca²⁺- and Mg²⁺-free and EGTA-containing medium (Fig. 6B). The recovery ratio of the [Ca²⁺]_i signal was not significantly changed in the Ca²⁺-free and EGTA-containing media, with or without Mg²⁺, compared with the control (Fig. 6C). Conversely, the recovery ratio significantly decreased in the Ca²⁺-free, high Mg²⁺- and EGTA-containing medium, compared with the Ca²⁺- and Mg²⁺-free and EGTA-containing medium (P < 0.05). The maximal slope was 0.42 ± 0.06 min¹ (n = 13) in the control, 0.18 ± 0.02 min¹ (n = 8, P < 0.05) in the Ca²⁺-free, high Mg²⁺- and EGTA-containing medium, and 0.84 ± 0.06 min¹ (n = 9, P < 0.01) in the Ca²⁺- and Mg²⁺-free and EGTA-containing medium. The recovery ratio was 53 ± 3% (n = 13) in the control, 34 ± 3% (n = 8) in the Ca²⁺-free, high Mg²⁺- and EGTA-containing medium, and 66 ± 9% (n = 9) in the Ca²⁺- and Mg²⁺-free and EGTA-containing medium. These results indicate that the falling phase of the [Ca²⁺]_i signal is inhibited by either a reduction in [Na⁺]_o or an addition of Ni²⁺ or Mg²⁺.

View larger version (19K): [in this window] [in a new window]   Fig. 6. The addition of external Mg2+ depresses and removal of external Mg2+ accelerates the rapid decrease in the [Ca2+]i signal in the hippocampal CA1 region after ischemic exposure. A: after generating the rising phase of R340/380, oxygen and glucose were reintroduced in the normal ionic medium, and the Ca2+-free and EGTA (5 mM)-containing medium with or without Mg2+ (10 mM), respectively, from top to bottom trace. B and C: mean maximal slopes of the falling phase of [Ca2+]i signal and mean recovery ratios of the [Ca2+]i signal were obtained from similar experiments shown in A. Vertical error bar on each column in B and C represents standard errors. Note that the maximal slope of the falling phase was accelerated the in Ca2+- and Mg2+-free EGTA-containing medium.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mechanism for a rapid reduction in [Ca²⁺]_i after ischemic exposure

In extracellular recordings from the CA1 region of the rat hippocampal slice preparation, superfusion with the ischemia-simulating medium produced a rapid negative DC potential (rapid negative potential) 5-6 min after exposure to the ischemia-simulating medium. The rapid negative potential coincided with a rapid elevation (the rising phase) of the [Ca²⁺]_i signal during ischemic exposure. The reintroduction of oxygen and glucose produced a slow negative DC potential (slow negative potential) within 1 min and subsequently restored the DC potential to the preexposure level 5-6 min thereafter. These events corresponded to the rapid reduction (the falling phase) in [Ca²⁺]_i signal after ischemic exposure. Furthermore, changes in the membrane potential in glial cells during and after ischemic exposure were very similar to the changes in [Ca²⁺]_i signal and also coincided with the changes in extracellularly recorded DC potentials (bottom traces in Fig. 1, A and B), respectively. Our previous study suggested that the slow negative potential is mediated by the operation of Na⁺/Ca²⁺ exchangers with a forward mode in nonneuronal cells (Uchikado et al. 2000). Together, our data indicate that the falling phase of [Ca²⁺]_i signal may be produced mainly by operation of Na⁺/Ca²⁺ exchangers in nonneuronal cells, including glial cells. However, although the DC potential completely recovered after ischemic exposure, the [Ca²⁺]_i did not. The residual elevation of [Ca²⁺]_i after ischemic exposure is probably due to the accumulation of [Ca²⁺]_i in the CA1 neurons but not in glial cells, since the neuronal membrane could not repolarize after the reintroduction of oxygen and glucose (Tanaka et al. 1997; Uchikado et al. 2000).

The falling phase of the [Ca²⁺]_i signal was significantly depressed by a prolonged application of the ischemia-simulating medium (e.g., 13 min application). Superfusion with oxygen- or glucose-free media after generating the rising phase also depressed the falling phase. Superfusion with ouabain (30 µM) produced a rapid elevation of [Ca²⁺]_i that did not reverse after the compound was washed out. Moreover, the addition of a high concentration (1 mM) of ouabain applied in a reperfusion medium with oxygen and glucose prevented the generation of both the slow negative potential and the falling phase. These results suggest that the falling phase of the [Ca²⁺]_i signal after ischemic exposure is due to a reactivation of Na⁺, K⁺-ATPase. However, the activation of Na⁺, K⁺-ATPase itself does not cause Ca²⁺ movement (Läuger 1991a). It is therefore most likely that the return of the ATP supply after reintroduction of oxygen and glucose is not directly involved in the falling phase; however, the reactivation of Na⁺, K⁺ATPase is essential for a reduction in the [Ca²⁺]_i signal.

The ouabain concentration we used (30 µM or 1 mM) seems to be considerably high. However, an isoform of the  subunit of Na⁺, K⁺-ATPase with a low affinity (1, K_i = 10-100 µM) for ouabain distributes both neurons and glial cells (Erdmann et al. 1985; Emanuel et al. 1988; McGrail et al. 1991; Juhaszova and Blaustein 1997; Sweadner 1989). In addition, other isoforms of the  subunit with high affinity (2 and 3, K_i = 0.02-1 µM) distribute either glial cells (2) or neurons (3) (Hara et al. 1988; Urayama and Sweadner 1988). It is therefore necessary to apply more than 100 µM of ouabain for a complete inhibition of Na⁺, K⁺-ATPase. Since the ouabain sensitivity of Na⁺, K⁺-ATPase decreases in the presence of elevated K⁺ concentration (Walz and Hertz 1982) and the extracellular K⁺ concentration increases during and after the rapid negative potential (Donnelly et al. 1992; Kawasaki et al. 1990; Lehmenkühler et al. 1988), the high concentration of ouabain (1 mM) is thus considered to be reasonable.

Pretreatment of the slice preparations with Ni²⁺-containing medium or low Na⁺-containing medium inhibited the falling phase. The application of Na⁺-free medium immediately after the onset of the rising phase also suppressed the falling phase. In contrast, Ca²⁺- and Mg²⁺-free with EGTA-containing medium accelerated the falling phase. These results support the suggestion that activation of Na⁺/Ca²⁺ exchangers in forward mode contributes to the falling phase, since the chemical gradient for inward transport of Na⁺ via Na⁺/Ca²⁺ exchangers should be reduced in low Na⁺-containing medium, and chemical gradient for extrusion of Ca²⁺ via Na⁺/Ca²⁺ exchangers should be larger in the Ca²⁺-free medium. The present results are comparable with previous studies in the squid giant axon, where the Ca²⁺ efflux resulting from the operation of Na⁺/Ca²⁺ exchangers is markedly reduced by low [Na⁺]_o (Baker et al. 1969). Similarly, in sheep ventricular muscle, a decrease in [Ca²⁺]_o results in a decreased [Ca²⁺]_i level and an increased [Na⁺]_i level while an increase in [Ca²⁺]_o results in an increased [Ca²⁺]_i level and a decreased [Na⁺]_i level via the Na⁺-dependent Ca²⁺ efflux (Sheu and Fozzard 1982).

Ca²⁺-ATPases are widely distributed in the plasma membranes as well as in the membranes of cellular organelles (Läuger 1991b). The activation of the Ca²⁺-ATPase in the plasma membrane induces an outward current or a hyperpolarization when Ca²⁺ ions are transported from cytosol to extracellular space; however, the activation of the Ca²⁺ATPase in the membrane of endoplasmic reticulum and mitochondria does not produce any transmembrane current. Therefore the contribution of Ca²⁺-ATPase to the rapid reduction in [Ca²⁺]_i during the generation of the slow negative potential is very small, if any at all.

Effects of Mg²⁺ on the rapid reduction in [Ca²⁺]_i

The present study showed that the falling phase was significantly suppressed in low Ca²⁺ (0.25 mM) and high Mg²⁺ (10 mM) medium, or in Ca²⁺-free solution containing Mg²⁺ (10 mM) and EGTA (5 mM). The free Mg²⁺ concentration in the latter medium was calculated by the following equation (modified from Tsuda et al. 1988)added [Mg] = (1 + K'([EGTA] + [Mg]f)/1 + [Mg]fK')[Mg]f where [Mg]f represents the desired free-Mg²⁺ concentration, [Mg] represents the dissolved Mg²⁺ concentration, [EGTA] represents the concentration of EGTA, and K' is an apparent association constant for Mg²⁺-EGTA (i.e., 10^5.2 at 25°C) (Schmid and Reilley 1957). We used 5 mM EGTA and 10 mM Mg²⁺, thus resulting in a 5 mM free-Mg²⁺ concentration of the medium. The present results are comparable with those of a previous report which described that Na⁺-induced Ca²⁺ efflux via Na⁺/Ca²⁺ exchangers is reduced by Mg²⁺ (5 mM) (Philipson and Nishimoto 1981). Moreover, the Na⁺-dependent Ca²⁺ flux is inhibited by divalent cations such as Cd²⁺, Mn²⁺, Co²⁺, and Mg²⁺ (Bers et al. 1980). Taken together, it is likely that in the presence of Mg²⁺, suppression of the falling phase of [Ca²⁺]_i is due to the inhibition of Na⁺/Ca²⁺ exchangers by extracellular Mg²⁺ at a relatively high concentration.

The stoichiometry for Na⁺/Ca²⁺ exchange is 3 Na⁺:1 Ca²⁺ (Carafoli 1987, also see review by Blaustein 1988). From the findings of stoichiometry, the forward-mode operation of Na⁺/Ca²⁺ exchangers induces an inward current while the reverse-mode operation induces an outward current. We therefore consider that the slow negative potential is mediated by the forward-mode operation of Na⁺/Ca²⁺ exchangers (Uchikado et al. 2000). In conclusion, we believe that the rapid reduction in [Ca²⁺]_i is mainly the result of an extrusion of intracellular Ca²⁺ via a forward-mode operation of Na⁺/Ca²⁺ exchangers in nonneuronal cells, including glial cells.