Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Brain ischemia induces a substantial increase in the extracellular accumulation of excitatory amino acids, glutamate and aspartate, in vivo. The high level of extracellular glutamate and aspartate is thought to activate glutamatergic synaptic transmission through glutamate receptors and the excessive activation of the receptors is suspected to cause neuronal cell damage. Although several possibilities have been proposed (see review by Lipton and Rosenberg, 1994; Martin et al., 1994), the mechanism of accumulation and the origin of glutamate and aspartate are not well understood. Neuronal injury mediated by excitatory amino acids can be induced in animals by ischemia and in cell culture by deprivation of oxygen and glucose (in vitro ischemia) from medium. Pharmacological treatment of the ischemic animals and cultured neurons in in vitro ischemia has facilitated understanding the process leading to the neuronal injury.

Selective N-methyl-D-aspartate antagonists can markedly attenuate the cortical neuronal death induced by hypoxia in vitro (Goldberg et al., 1987) and reduce focal ischemic brain injury in vivo (Buchan, 1990). In contrast, although non-NMDA antagonists markedly reduce global ischemic cell death and neuronal damage after severe forebrain ischemia in vivo (Sheardown et al., 1990; Buchan et al., 1991; Gaspary et al., 1994), the blockade of non-NMDA receptors has little or no protective effect in vitro against the destruction of neurons induced either by exposure to high levels of glutamate, or by hypoxia (Koh and Choi, 1991; Koretz et al., 1994). These antagonists at NMDA and non-NMDA receptors will exert their neuroprotective effects by the blockade of postsynaptic receptors. However, non-NMDA antagonists, GYKI 52466 and NBQX, have been reported to block the increase in extracellular glutamate induced by ischemia in in vivo microdialysis experiments (Arvin et al., 1992, 1994; Gaspary et al., 1994). In our study, we have examined the possibility that NMDA and non-NMDA receptor antagonists might protect cultured hippocampal neurons by inhibiting the glutamate efflux induced by oxygen/glucose deprivation in addition to the postsynaptic action.

Disturbance in intracellular ion homeostasis induced by ischemia is implicated in the initiation and progression of neuronal injury. Massive influx of calcium ions through postsynaptic ionotropic glutamate receptors significantly contributes to propagation of the injury. The neuroprotective effects of the sodium channel blocker TTX have been demonstrated both in vivo and in vitro (Prenen et al., 1988; Boening et al., 1989; Yamasaki et al., 1991; Weber and Taylor, 1994; Xie et al., 1994). However, the implication of voltage-dependent calcium channels in the accumulation of extracellular glutamate and aspartate during ischemia remains controversial. Although several authors have reported that neuroprotective effects were evoked by L-type (Grotta et al., 1986; Kuwaki et al., 1989; Takakura et al., 1992; Lobner and Lipton, 1993) and by N-type (Silverstein et al., 1986; Madden et al., 1990; Yamada et al., 1994; Ooboshi et al., 1992) calcium channel blockers in ischemia, others have been unable to show a beneficial effect of L-, N- or P-type voltage-gated calcium channel blockers in ischemia (Goldberg and Choi, 1993; Bickler and Hansen, 1994; Koretz et al., 1994). We also examined whether subtype-specific calcium channel blockers reduce the glutamate and aspartate efflux induced by oxygen/glucose deprivation and protect neurons from subsequent neuronal cell damage.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals and reagents. -Conotoxin GVIA and -Aga-TK (Kuwada et al., 1994) were purchased from Peptide Institute Inc. (Osaka, Japan). Fetal calf serum, heat-inactivated horse serum, trypsin solution, penicillin, streptomycin, DMEM and HEPES were purchased from Life Technologies Inc. (Grand Island, NY), insulin, sodium selenite, putrescine, Ara C, DNase I, AMPA, kainate, NMDA, glycine and nifedipine from Sigma Chemical Co. (St. Louis, MO), o-phthalaldehyde from Wako (Osaka, Japan), -mercaptopropionic acid from Dojin Laboratories (Kumamoto, Japan), MK801 from Research Biochemicals Inc. (Natick, MA), CNQX from Tocris Neuramin Ltd. (Bristol, UK) and TTX from Sankyo Co., Ltd. (Tokyo, Japan). CGS19755, CPP, NBQX and GYKI 52466 were synthesized in our laboratory. All other chemicals used were of reagent grade.

Hippocampal cell cultures. Hippocampal cell cultures were prepared from fetal rats of the Wistar strain (gestational age of 17 days). The hippocampus was dissected and kept in ice-cold Hanks' balanced salt solution, then incubated at 37°C for 15 min in Ca⁺⁺/Mg⁺⁺ -free Hanks' balanced salt solution containing 0.25% trypsin and 0.2 mg/ml DNase I. The hippocampal tissues were dissociated to single cells by gentle trituration using a glass pipette with a fire-polished tip. The cell suspension was mixed with DMEM supplemented with 10% fetal calf serum, 10% heat-inactivated horse serum, 5 µg/ml insulin, 30 nM sodium selenite, 100 µM putrecine, 20 nM progesterone, 15 nM biotin, 100 U/ml penicillin, 100 µg/ml streptomycin and 1 mM sodium pyruvate, as described (Scholtz et al., 1988). The cell suspension was centrifuged and the resulting pellets were resuspended in the medium described above. The hippocampal cells were then pelleted again by centrifugation, suspended in the medium and plated onto poly-L-lysine-coated coverslips. The cells were cultured in a CO₂ incubator (5% CO₂) at 37°C for 1 day and the coverslips were then transferred onto a confluent glial cell layer (cell-side-up) and cultured for 8 days in DMEM containing 2% fetal calf serum and the same supplements as described above. After 8 days, the hippocampal cells were cultured in DMEM containing 2% fetal calf serum and the same supplements, but without glutamine. The hippocampal cells were treated with 10 µM Ara C for 1 day (it was added to the culture medium 3 days after plating) to reduce the growth of contaminating non-neuronal cells. The culture medium was changed every 3 to 4 days. The glial cells used were obtained from postnatal day 1 rats of the Wistar strain. The cerebral cortex was dissected and triturated in DMEM supplemented with 10% fetal calf serum, 100 U/ml penicillin and 100 µg/ml streptomycin, and the glial cells were cultured in a CO₂ incubator (5% CO₂) at 37°C for 11 to 18 days before use.

Oxygen-glucose deprivation. The oxygen/glucose deprivation treatment was performed on 15- to 17-day-old cultures. Hippocampal neurons cultured on coverslips were transferred to new plates containing glucose-free KR buffer with the following ionic composition (in mM): KCl (5.36), CaCl₂ (1.26), KH₂PO₄ (0.44), MgCl₂ (0.49), MgSO₄ (0.41), NaCl (137), NaHCO₃ (4.17), Na₂HPO₄ (0.34), HEPES (10), pH 7.4, which was gassed with 95% N₂/5% CO₂ at 37°C. The plates were transferred to an anaerobic chamber containing the same gas mixture and incubated at 37°C for 40 to 50 min. Oxygen/glucose deprivation was terminated by replacing the exposure medium with DMEM containing 33.3 mM glucose and the cultures were returned to a normoxic incubator. The KR buffer was used for measurement of glutamate and aspartate. Control cultures were similarly incubated in a KR-buffer containing 10 mM glucose. Receptor antagonists and channel blockers were added to the cultures immediately before and after the oxygen/glucose deprivation.

LDH assay. The measurement of LDH efflux into the media as an indicator of neuronal injury was made 20 hr after the oxygen/glucose deprivation period (Koh and Choi, 1987).

Amino acid analysis. Glutamate and aspartate released into the KR buffer during the oxygen/glucose deprivation period were determined by high-performance liquid chromatography (HPLC) with a fluorometric monitor (excitation 330 nm, emission 450 nm) after derivatization with o-phthalaldehyde as described (Ikeda et al., 1989). Briefly, o-phthalaldehyde solution was prepared as follows: o-phthalaldehyde (50 mg) was dissolved in 2.5 ml methanol and then -mercaptopropionic acid (0.025 ml) was added to the solution. This solution was mixed with 5 ml of .15 M sodium borate buffer (pH 10.5). Samples (0.05 ml) were mixed with the o-phthalaldehyde solution (0.05 ml) and the reaction was terminated by adding 0.1 M KH₂PO₄/66% CH3CN (0.1 ml) after 3 min. This derivatized samples were applied to HPLC. The HPLC apparatus was equipped with an ion exchange column (Asahipak ES-502N, Showa Denko, Tokyo, Japan), and 0.145 M sodium citrate buffer (pH 5) containing 50% acetonitrile was used as the mobile phase at a flow rate of 2.5 ml/min. The detection limit of amino acids in the buffer was approximately 0.002 µM by this assay.

Glutamate release by glutamate agonists. Treatment with glutamate agonists was performed on 14-day-old cultures. Hippocampal neurons cultured on coverslips were transferred to a new plate containing KR buffer with the following ionic composition (in mM): KCl (5.36), CaCl₂ (1.26), KH₂PO₄ (0.44), MgCl₂ (0.49), MgSO₄ (0.41), NaCl (137), NaHCO₃ (4.17), Na₂HPO₄ (0.34), HEPES (10), glucose (33.3) pH 7.4. The plates were incubated with glutamate agonists in the presence or absence of glutamate antagonists at 37°C for 5 min. Glutamate release into the KR buffer was determined by HPLC with a fluorometric monitor after derivatization with o-phthalaldehyde.

Protein measurement. Hippocampal cell cultures were lysed with 1% sodium dodecyl sulfate and 0.5 N NaOH and the protein of the lysate was measured using a Micro BCA protein assay kit (Pierce, Rockford, IL).

Statistical analysis. Data were analyzed by means of analysis of variance and Fisher's PLSD.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

The incubation of hippocampal cell cultures under an oxygen/glucose-deprived condition in KR buffer solution produced widespread neuronal swelling within 60 min, followed by neuronal degeneration over the ensuing several hours after return to the normal medium. When cultured hippocampal neurons were exposed to combined deprivation of oxygen and glucose for at least 40 min or more, the extracellular glutamate and aspartate levels increased linearly with incubation time up to 80 min (fig. 1). Glutamate concentration in the KR buffer was initially at .370 µM (n = 5) and it increased to 2.05 µM at 60 min of oxygen/glucose deprivation. Similarly, initial aspartate concentration in the KR buffer was 0.002 µM and 1.30 µM at 60 min of oxygen/glucose deprivation. The LDH efflux into the medium was measured 20 hr after exposure to oxygen/glucose deprivation as an indicator of neuronal cell death. Deprivation of oxygen/glucose for 40 min or more induced an increase in the efflux of LDH into the incubation medium (fig. 1). However, the incubation of hippocampal neurons in the KR buffer containing 10 mM glucose or the glucose-free buffer gassed with air/5% CO₂ failed to increase the extracellular levels of glutamate and aspartate.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Accumulation of extracellular glutamate and aspartate and LDH efflux induced by oxygen/glucose deprivation of cultured hippocampal neurons. Cultured hippocampal neurons were incubated for a period as indicated up to 80 min in the absence of oxygen in KR buffer solution without glucose. The buffer was then replaced by normal medium and the hippocampal neurons were incubated in a normoxic incubator for 20 hr. The concentrations of glutamate (A) and aspartate (B) accumulated in the KR buffer solution deprived of oxygen and glucose that was sampled at 0 to 80 min and the activity of LDH (C) in the normal medium at 20 hr were measured. Each value represents the mean with SEM (n = 5).

The effects of glutamate receptor antagonists on the increase in extracellular glutamate/aspartate and the LDH efflux induced by oxygen/glucose deprivation. The incubation of hippocampal cell cultures in the oxygen/glucose-deprived KR buffer solution for 50 min evoked substantial increases in the extracellular levels of glutamate and aspartate and the LDH efflux. Both NMDA receptor antagonists, MK-801, CGS19755 and CPP, and non-NMDA receptor antagonists, CNQX, NBQX and GYKI 52466, dose-dependently attenuated the increase in the extracellular levels of glutamate and aspartate and the LDH efflux (table 1). The LDH efflux in the presence of high concentrations of MK-801, CGS19755 and CPP was lower than the control efflux observed when the hippocampal neurons were incubated in the KR buffer solution containing 10 mM glucose instead of glucose-free KR buffer solution. NMDA receptor antagonists, MK-801, CGS19755 and CPP, had greater inhibitory effects on LDH efflux than on the increase in extracellular glutamate and aspartate, although non-NMDA receptor antagonists, NBQX and GYKI 52466, inhibited the increase in the levels of extracellular glutamate and aspartate more potently than the LDH efflux (fig. 2). A non-NMDA receptor antagonist, CNQX, inhibited the increase in the extracellular level of glutamate and the LDH efflux to the same extent (table 1). It has been reported that high concentrations of CNQX antagonize NMDA receptor-mediated effects by blocking the strychnine-insensitive glycine binding site on the NMDA receptor (Ogita and Yoneda, 1990). CNQX, therefore, might have dual effects as both an NMDA receptor antagonist and a non-NMDA receptor antagonist on the glutamate receptors. Although NBQX did not have significant neuroprotective effects at 10 to 30 µM, NBQX at the same concentrations significantly inhibited the increase in the extracellular levels of glutamate and aspartate. A combination of 30 µM NBQX and 0.1 µM MK-801 produced marked neuroprotection, compared to that evoked by MK-801 alone (table 1).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Effects of glutamate receptor antagonists on the increase in the extracellular level of glutamate and the LDH efflux induced by combined deprivation of oxygen and glucose. Plot of data from table 1.

The effects of glutamate receptor agonists on the glutamate release. We examined the effect of glutamate receptor agonists on the glutamate release in hippocampal neurons (fig. 3). Kainic acid evoked substantial glutamate release, which was inhibited completely by non-NMDA receptor antagonists, NBQX (10 µM) and GYKI52466 (10 µM), although AMPA or NMDA failed to induce glutamate release.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Effect of glutamate receptor agonists on glutamate release from hippocampal neurons. Hippocampal neurons were incubated with glutamate agonists in the presence or absence of glutamate antagonists for 5 min in KR buffer at 37°C. Control, incubation of the hippocampal neurons without glutamate agonist; KA, incubation with 1 mM kainic acid; KA NBQX, incubation with 1 mM kainic acid and 10 µM NBQX; KA GYKI, incubation with 1 mM kainic acid and 10 µM GYKI 52466; AMPA, incubation with 1 mM AMPA; NMDA, incubation with 1 mM NMDA and 1 µM glycine in magnesium-free KR buffer. Each value represents the mean with SEM (n = 4). ***P < .001 vs. control. Data were analyzed by means of analysis of variance and Fisher's PLSD.

The effects of calcium and sodium channel blockers on the increase in extracellular glutamate and aspartate and the LDH efflux induced by oxygen/glucose deprivation. We examined the effect of calcium and sodium channel blockers on the increase in extracellular glutamate and aspartate and the LDH efflux induced by oxygen/glucose deprivation. A voltage-dependent L-type calcium channel blocker, nifedipine, an N-type blocker, -conotoxin GVIA and a P-type blocker, -Aga-TK, all attenuated the neuronal injury (table 2). However, these calcium channel blockers did not significantly affect the increase in the extracellular levels of glutamate and aspartate. Combinations of two subtypes of the calcium channel blockers inhibited the neuronal damage by about 30% and in particular, the combination of P-type and N-type calcium channel blockers reduced the increase in extracellular glutamate and aspartate, in addition to ameliorating the neuronal damage (table 2). The combination of all three subtypes of calcium channel blockers attenuated similarly both the increase of extracellular glutamate and aspartate and the neuronal injury (table 2). When the oxygen/glucose deprivation period was made longer, the protective effects became weaker (table 3). A sodium channel blocker, TTX, potently attenuated the increase in extracellular glutamate and aspartate and the neuronal injury, and it showed a greater effect on the extracellular glutamate and aspartate than on the LDH efflux (table 2).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In vitro ischemia and glutamate release. The elevation of extracellular excitatory amino acid concentrations under brain ischemia may be caused by several mechanisms and the origin of released glutamate/aspartate is not well understood. In our study, we examined the mechanisms of neuronal cell damage induced by ischemia by means of pharmacological modulation of the enhanced glutamate/aspartate efflux and the consequent cell loss. We have established a culture system of rat hippocampal neurons in which the extracellular levels of excitatory amino acids in buffer solution are greatly enhanced by combined deprivation of oxygen and glucose. Incubation of the cells in ischemic buffer solution for at least 40 min evoked increases in the extracellular levels of glutamate/aspartate and also induced massive neuronal cell loss on further incubation of the cells in normal medium for another 20 hr under normoxic conditions. Severe ischemia has been reported to induce the release of glutamate from both neurons and glia into the extracellular space and to cause an alteration in the uptake mechanisms of glutamate (Wahl et al., 1994). In our culture system, although cytosine arabinofuranoside was added to the culture medium of hippocampal neurons, there was some contamination with glial cells. However, it is very likely that glutamate and aspartate were derived from the neurons, since extracellular glutamate/aspartate was not increased when glial cultures were deprived of oxygen/glucose for 120 min (data not shown). In addition, cultured glial cells were much more resistant to damage, assessed in terms of efflux of LDH, after oxygen/glucose deprivation than neurons. In oxygen/glucose deprivation, not only NMDA antagonists but also non-NMDA antagonists and calcium and sodium channel blockers reduced the extracellular levels of glutamate and aspartate (tables 1 and 2). These results suggest that oxygen/glucose deprivation activates the glutamate receptors and voltage-dependent ion channels, and thereby shifts the equilibrium of glutamate/aspartate flux through membranes in favor of efflux to the extracellular space.

Mechanisms of pharmacological blockade of glutamate release. The inhibitory effects of non-NMDA antagonists on the increase of extracellular levels of glutamate/aspartate induced by in vitro ischemia have not been reported previously, although the effects of NMDA antagonists have recently been reported (Strijbos et al., 1996; Probert et al., 1997). Kainic acid has been suggested to induce glutamate release by activation of the presynaptic receptors on glutamatergic terminals (Ferkany et al., 1982). We have also confirmed that glutamate release is evoked by kainic acid, not by NMDA or AMPA, in cultured hippocampal neurons (fig. 3). It is, therefore, possible that non-NMDA receptor antagonists reduce the extracellular levels of glutamate/aspartate by inhibiting the release from presynaptic terminals through blocking the kainate-type receptors. Alternatively, another possibility is that the Na⁺-syntransporter system of glutamate might take part in the non-NMDA receptor-mediated glutamate efflux under ischemic conditions. During ischemia, the drop in ATP levels leads to an increase of intracellular Na⁺ concentration owing to the reduction of Na⁺/K⁺ -ATPase activity, and this might trigger the release of glutamate through the reversed mode of operation of the Na⁺-syntransporter system (Barbour et al., 1988; Nicholls and Attwell, 1990; Szatkowski et al., 1990; Madl and Burgessr, 1993; Friedman and Haddad, 1994). It is, therefore, likely that the non-NMDA receptor antagonists suppress the glutamate efflux as a result of their inhibition of non-NMDA receptor-mediated influx of sodium ions.

Relationship between the inhibition of glutamate release and neuroprotection. A question raised by the present study is whether the inhibitory effect of these antagonists and blockers on the increased levels of extracellular glutamate/aspartate is a cause of their neuroprotective effects. An NMDA receptor antagonist, MK-801, blocked the cell damage completely, while it only partially reduced the increase in extracellular glutamate/aspartate (table 1). If the glutamate was released from damaged cells as a result of cell injury, the NMDA receptor antagonist would completely inhibit the glutamate release. Thus, the enhanced glutamate efflux might be a primary event and the cell damage be the consequence. However, although the increase in the extracellular level of glutamate is likely to be the primary cause of neuronal cell damage, non-NMDA receptor antagonists and TTX did not afford complete neuroprotection despite reducing the glutamate efflux to a large extent, suggesting that the enhanced glutamate/aspartate efflux is not the sole cause of neuronal cell loss. Although ischemic incubation of hippocampal neurons for 40 min elicited the significant, but small increase in the extracellular level of glutamate, it induced massive cell loss comparable to that after prolonged incubation period of 60 to 80 min. This observation indicates that a large inhibition of glutamate release is needed to achieve the complete inhibition of excitotoxicity in cultured neurons. Reduction of neuronal ATP level and energy charge under ischemic conditions would facilitate the cellular injury by enhancing vulnerability of neurons to excitotoxicity due to deterioration of intracellular calcium homeostasis even at relatively mild increase in extracellular glutamate.