The Effect of Glutamate Receptor Blockade on Anoxic Depolarization and Cortical Spreading Depression

Transcription

1 Journal of Cerebral Blood Flow and Metabolism 2: The nternational Society of Cerebral Blood Flow and Metabolism Published by Raven Press, Ltd., New York The Effect of Glutamate Receptor Blockade on Anoxic Depolarization and Cortical Spreading Depression *tmartin Lauritzen and *:j:anker Jon Hansen *Department of General Physiology and Biophysics, Panum nstitute, University of Copenhagen, Copenhagen, tdepartment of Clinical Neurophysiology, Hvidovre Hospital, Hvidovre, and tcns Division, Novo Nordisk nc., S borg, Denmark Summary: We examined the effect of blockade of N methyl-d-aspartate (NMDA) and non-nmda subtype glutamate receptors on anoxic depolarization (AD) and cortical spreading depression (CSD). [K +le and the direct current (DC) potential were measured with microelectrodes in the cerebral cortex of barbiturate-anesthetized rats. NMDA blockade was achieved by injection of (+)- 5-methyl-to, -dihydro-5h-dibenzo[a, dlcyclohepten- 5,0-imine maleate [MK-80; 3 and to mg/kgl or amino- 7-phosphonoheptanoate (APH; 4.5 and to mg/kg). Non NMDA receptor blockade was achieved by injection of 2, 3-dihydroxy -6-ni tro-7 -sulfamoy benzo(f)quinoxaline (NBQX; 0 and 20 mg/kg). MK-80 and APH blocked CSD, while NBQX did not. n control rats, the latency from circulatory arrest to AD was 2. ± 0. min, while the amplitude of the DC shift was 2 ± m V, and [K + le increased to 50 ± 6 mm. All variables remained unchanged in animals treated with MK-80, APH, or NBQX. Finally, MK-80 (4 mg/kg) and NBQX (40 mg/ kg) were given in combination to examine the effect of total glutamate receptor blockade on AD. This combination slightly accelerated the onset of AD, probably owing to circulatory failure. n conclusion, AD was unaffected by glutamate receptor blockade. n contrast, NMDA receptors play a crucial role for CSD. Key Words: Cerebral ischemia-glutamate receptors-on homeostasis Migraine-Spreading depression. The aim of this study was to examine the mechanism of depolarization associated with cortical spreading depression (CSD) (Leao, 944) and cerebral ischemia (anoxic depolarization; AD) (Leao, 947). CSD and AD are two conditions in which brain function is perturbed, with a number of interesting similarities and differences. n both conditions spontaneous and evoked brain activity is depressed, the direct current (DC) potential under- Received February 22, 99; revised July 9, 99; accepted July 23, 99. Address correspondence and reprint requests to Dr. M. Lauritzen at Department of General Physiology and Biophysics, Panum nstitute, University of Copenhagen, Blegdamsvej 3c, 200-DK Copenhagen, Denmark. Abbreviations used: AD, anoxic depolarization; APH, amino- 7-phosphonoheptanoate; APV, amino-5-phosphonovalerate; CSD, cortical spreading depression; DC, direct current; MK- 80, (+ )-5-methyl-0, -dihydro-5h-dibenzo[a,dlcyclohepten- 5,O-imine maleate; NBQX, 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo(F)quinoxaline; NMDA, N-methyl-D-aspartate. Part of the data presented here have been published as preliminary communications (Hansen et al., 988; Lauritzen et al., 99). goes a predominantly negative change of 0-40 m V, and brain ion homeostasis fails owing to increased membrane permeability (Hansen and Zeuthen, 98 ; Nicholson and Kraig, 98; Hansen, 985). mportant differences are that in CSD the electrophysiological and metabolic changes revert spontaneously within a couple of minutes, while in cerebral ischemia blood flow is required to restore energy and ion metabolism to normal. Furthermore, in CSD the depolarization successively invades adjacent tissue (Leao, 944), while the AD occurs asynchronously in widespread brain regions (Leao, 947), without evidence of interaction between depolarized and nondepolarized regions (Hernandez Caceres et a., 987). The mechanisms of depolarization and membrane failure during CSD and AD are incompletely understood. The possibility that glutamate plays an active role in CSD and AD was originally suggested by Van Harreveld (959). Glutamate, aspartate, and agonists of the glutamate subtype receptors [Nmethyl-D-aspartate (NMDA), quisqualate, and 223

2 224 M. LAURTZEN AND A. J. HANSEN kainate] trigger CSD (Van Harreveld, 959; Curtis and Watkins, 963; Bures et a., 974; Lauritzen et a., 988). t is controversial whether glutamate, and to a lesser extent aspartate, is released during CSD (Van Harreveld and Kooiman, 965; Van Harreveld and Fifkova, 970; Gardino and do Carmo, 983; Scheller et a., 990). On the other hand, cerebral anoxia is accompanied by the release of glutamate, aspartate, and -y-aminobutyrate (Benveniste et a., 984). Artificial CSF with a low magnesium concentration triggers the opening of the NMDA receptorchannel complex and the production of CSD in the hippocampus and the cerebellum in vitro (Mody et a., 987; Lauritzen et a., 988). On some occasions the tissue remains depolarized much in the same manner as during AD (Lauritzen et a., 988). t is possible that NMDA receptor activation per se is sufficient to cause AD. NMDA receptor blockade inhibits CSD, both in vitro (Mody et a., 987; Lauritzen et a., 988) and in vivo (Gorelova et a., 987; Marrannes et a., 988). The role of NMDA receptor blockade in AD is controversial, since ketamine failed to influence AD in the neocortex (Hernandez-Caceres et a., 987; Amemori and Bures, 990) while amino-5-phosphonovalerate (APV) curtained hippocampal AD (Benveniste et a., 988). To elucidate the contribution of glutamatemediated neurotransmission further, we examined the effects on CSD and AD of (+ )-5-methyl-O,ldihydro-5H-dibenzo[a,d]cyciohepten-5, 0-imine maleate (MK-80 ) and amino-7 -phosphonoheptanoate (APH), both NMDA antagonists (Simon et a., 984; Wong et a., 986; Park et a., 988; Albers et a., 989), and 2,3-dihydroxy-6-nitro-7- sulfamoylbenzo(f)quinoxaline (NBQX), a newly developed non-nmda receptor antagonist with preference for a-amino-3-hydroxy-5-methyl-4- isoxazole propionic acid receptors and cerebroprotective properties for cerebral ischemia (Sheardown et a., 990). MATERALS AND METHODS Male Wistar rats ( g) were anesthetized with an initial dose of Nembutal (50 mg/kg) and supplementary doses as required later. Catheters were placed in one femoral artery and one femoral vein. The spontaneously breathing rats were placed in a headholder and craniotomies were performed ipsilaterally over the frontal and parietal cortices. The mean arterial blood pressure was monitored continuously while respiratory steady state was ensured by frequent sampling of arterial blood and measurements of Pa02, Pac02, and arterial ph. Rectal temperature was servo-controlled at 37 C with a heating pad. [K +]e and/or the DC potential were measured in the parietal cortex using single- or double-barreled glass mi- croelectrodes as described previously (Hansen and Zeuthen, 98). Spread of the CSD was monitored with a single-barreled microelectrode measuring the DC potential in the frontal cortex. An Ag/AgC wire located under the skin of the hind leg served as reference electrode. CSD was induced electrically in the parietal cortex by bipolar stimulation using two bared silver wires attached to the bottom of a cup placed on the dura (Fig. ). The recording electrode was positioned midway between the two stimulating wires. Stimulation consisted of a train of 5-ms pulses at 40 Hz lasting for 2 s at 8-0 V. The threshold stimulation for CSD elicitation was defined by varying the voltage, but the stimulus parameters varied very little between animals. The reliability of the threshold stimulus was tested by inducing CSD a couple of times before MK-80 at 3 or 0 mg/kg, APH at 4.5 or 0 mg/kg, or NBQX at 0 or 20 mg/kg was injected. Slow injection rates were used to avoid drop of arterial blood pressure. Stimulation was resumed after another few minutes, first at the threshold level. f CSD was blocked, stimulation was increased with respect to voltage (5-25 V) and stimulus duration (up to 4 s) so as to produce a local response of the DC potential andlor [K + le as during the preceding CSD under control conditions. The speed of the CSD spread was calculated from the latency difference of the negative DC shift at the electrodes in the parietal and frontal cortices and the distance between the electrodes. The amplitude of the DC potassium shift was measured from the baseline to the largest deflection. Animals were killed with an intravenous injection of concentrated KC. The AD latency was measured from the moment of circulatory arrest to the onset of the rapid part of the depolarization. Amplitudes were measured peak to peak from the positive shift occurring during the first phase of the AD to the negative peak terminating the AD. \ \ Stirn. o \ \ DC FG.. Schematic drawing of arrangement of stimulating and recording electrodes. Stimulating electrodes were two bared silver wires extending -2 mm out of the bottom of a cup placed at the dura over the parietal cortex. One recording electrode was placed between the two stimulating electrodes to monitor the triggering of cortical spreading depression (CSD) and the local response to electrical stimulation following injection of drugs. A second recording electrode was placed in the frontal cortex to monitor the spread of the CSD. DC, direct current.

3 GLUTAMATE N SPREADNG DEPRESSON AND SCHEMA 225 Values in the text are means ± SD. Student's t test (paired and unpaired) was used for statistics. RESULTS Under control conditions, the rate of the CSD spread was 3.S ± 0.9 mm/min while the amplitude of the DC shift was 23 ± 6 mv (n = 7 episodes). The latency to the onset of the AD was 2. ± 0. min, while the amplitude of the DC shift was 2 ± mv and [K +]e increased to 50 ± 6 mm (n = 4 rats). n eight rats treated with NBQX at 0 or 20 mg/ kg, the rate of CSD spread was 4. ±.2 mm/min before and 3.7 ± O.S mm/min after injection of the drug (n = 3; NS), while the change of the DC potential was 24 ± 7 m V before and 23 ± 5 m V after the drug (NS) (Fig. 2). The latency to the AD was.9 ± 0.3 min and the amplitude of the DC shift was 7 ± 4 mv (NS). Thus, CSD and AD are unaffected by NBQX. CSD was completely blocked in another 2 rats treated with MK-SO at 3 or 0 mg/kg CSD from a few minutes after the injection (Fig. 3). We tested the animals for up to 5 h after administration of MK-SOl, but found no evidence of adaptation. Threshold stimulation caused local changes of the DC potential and [K +]e' which were drastically reduced. ncreasing the stimulus strength and dura- tion produced local changes of DC potential and [K +]e as during the CSD recorded previously, but there was no spread to the electrode in the frontal cortex. n two rats a cottonball soaked in KC at M was applied to the cortex, but failed to produce a CSD. CSD was also inhibited when MK-SO at fa-m was applied directly to the cortex; however, the effect was irreversible and this type of experiment was pursued no further. Circulatory failure was instigated by intravenous injection of ml of M KC, typically at min after injection of MK-SO. NMDA receptors were still blocked at this time as evidenced by complete inhibition of CSD in response to electrical stimulation. The latency to the AD was 2.2 ± 0.2 min, while the DC potential change was 23 ± 3 m V, and [K +]e increased to 52 ± 6 mm, not different from controls (NS, n = S). Thus, MK-SO interferes with both the triggering and the spreading of CSD, but has no effect on AD. APR was tested in four rats. ntravenous injection of 4.5 mg/kg (n = ) induced incomplete blockade since suprathreshold stimulation induced a CSD on one occasion (Fig. 4). APR at 0 mg/kg (n = ) blocked CSD completely. APR at 0.5 mm applied to the brain surface at the site of electrical stimulation inhibited CSD completely, but CSD re- FG. 2. Shifts of the DC potential at electrodes in frontal cortex (upper trace) and parietal cortex (lower trace) of rats exposed to CSD and cerebral ischemia. CSD was elicited by local electrical stimulation of the parietal cortex. schemia was induced by intravenous injection of KCL After the second CSD episode, NBQX, a non-nmda antagonist, was given at 0 mg/ kg Lv. The drug had no effect on the triggering or spread of CSD, or the latency or amplitude of the AD. For abbreviations see the text. i-- ' ['0 mv JJ m - i n. 0 mv : -- ; - L : _-.J '\----" _--,- N_B_Q_X_lnjecti ' LJ m in mv i Circulatory arrest J Cereb Blood Flow Metab. Vol. 2, No.2, 992

4 226 M. LAURTZEN AND A.. HANSEN [K+Je mm l!:: 2 [K+Je E 0, JL 0 mm!20mv!, 3 0, [K+Je!, min -v-- ij mm!20mv:, min min R V ii..----l'-- (V AMi 8 0 V LJ -: '-.. _--- FG. 3. [K+le in parietal cortex (upper trace) and DC potential change in frontal cortex (lower trace) in rats exposed to CSD and cerebral ischemia. After the fourth CSD, MK-80, an NMDAantagonist, was given at 0 mg/ kg i. v. (arrow). The drug blocked CSD completely, but did not delay or curtail the [K+le changes during AD induced by intravenous injection of KC. For abbreviations see the text. appeared following prolonged drug outwash (n 2). Circulatory arrest was induced at a time when CSD remained inhibited owing to NMDA receptor blockade. The AD for all four rats followed with a latency of 2.3 ± 0.4 min and the amplitude of the DC shift was 2 ± my (NS). The variables for AD were the same whether the drug was applied locally or by intravenous injection. Finally, MK-80 at 4 mg/kg i.v. and NBQX at 40 mg/kg i. v. were given simultaneously to three rats over a 0-min period. The blood pressure-lowering effect of this rather high dose was dramatic and killed one of the rats, while the remaining two survived with a blood pressure at mm Hg for a few minutes. The latency from circulatory arrest to the AD was.6 ± 0.3 min, slightly faster than in controls (p < 0.05), while the amplitude of the DC shift in the parietal cortex was 6 ± 5 my, not significantly different from control rats. The shorter latency to AD is explained by the preischemic low mean arterial blood pressure, which induces brain hypoxia before cardiac arrest. DSCUSSON The main result of the present study is that AD in the cortex is unaffected by NBQX, MK-80, and APH alone and in combination. The results confirm the observations of Hernandez-Caceres et al. (987) and Amemori and Bures (990) that systematically administered NMDA antagonists in high doses are ineffective in blocking cortical AD. n the rat hippocampus, on the other hand, local injection of the competitive NMDA antagonist APY into the CAl region curtailed AD and the associated cellular influx of calcium (Benveniste et al., 988). AD and the calcium flux were also prevented by removal of the glutamatergic input to the CAl region (by killing

5 GLUTAMATE N SPREADNG DEPRESSON AND SCHEMA 227 FG. 4. Shifts of the potential at electrodes in the parietal cortex (upper trace) and frontal cortex (lower trace) of rats exposed to CSD and cerebral ischemia. After the third CSD, APH, an NMDA antagonist, was given at 4.5 mg/kg i.v. (arrow). APH blocked CSD rather effectively even in response to suprathreshold stimulation when the local response was of higher amplitude and longer duration than during control conditions. Numbers at arrows indicate the multiplication factor of the stimulus threshold voltage that was applied. The electrical stimulation induced CSD on one occasion following administration of the drug (bottom panel), but not during the following suprathreshold stimulations. For abbreviations see the text., mm r : : O. -+'-, "'" -+._,.. ' ' " i 2omvi=:l----n---'----.l=i L! : r i ",,! m " 2omv:--...u i_--,75hpc_= -=-- -_--- '-,t 'P-r-\:_-.-, ==-5 '.,...: P n ' ±= CA3 neurons with kainic acid) during a lo-min period of ischemia (Benveniste et a., 988). This may suggest that cortical and hippocampal ADs have different mechanisms. The latency from heart arrest to AD has previously been changed by varying the glucose stores of the brain or the cerebral metabolic rate (Bures et a., 974; Hansen, 978; Astrup et a., 980; Schaeffer and Lazdunski, 99 ). Antagonists of the ATPdependent K + channel are, however, without effect on the K + efflux during hippocampal ischemia (Schaeffer and Lazdunski, 99 ), despite the fact that the large ionic shifts are linked to a decline of energy charge (Siesjo and Bengtsson, 989). Energy charge remains constant during CSD, suggesting that energy failure is not a condition for membrane failure (Lauritzen et a., 990). MK-80 and APH blocked CSD, while NBQX did not. MK- 80 and APH interfered with both the triggering and the spreading mechanisms since a higher stimulation strength was required to obtain a local response of the same amplitude as in the untreated animal, and spread of the CSD occurred on only one occasion in one rat. The experimental setup with one recording electrode positioned between the two stimulating wires and the second electrode at a remote position appeared well suited for pharmacological studies of CSD since both the triggering and the spreading process were monitored. Our data strengthen the viewpoint that NMDA antagonists play a special role in CSD (Gorelova et a., 987; Mody et a., 987; Lauritzen et a., 988; Marrannes et a., 988; Sheardown, 989). Despite the similarities of CSD and AD with respect to ionic changes and DC shift, nerve cell depolarization, and electroencephalographic silence, the basic mechanisms appear entirely different. The different susceptibility to the NMDA antagonists may reflect quantitative aspects of transmitter release in the two conditions. Glutamate and aspartate are released in large quantities during ischemia (Benveniste et a., 984), while trace amounts are released during CSD (Van Harreveld and Kooiman, 965; Van Harreveld and Fifkova, 970; Gardino and do Carmo, 983; Scheller et a., 990). The dramatic increase of extracellular glutamate during cerebral ischemia occurs, however, much later than AD (Benveniste et a., 984) and has no immediate bearing on the ion fluxes associated with AD. Also the fact that the noncompetitive NMDA antagonist MK-80 was inefficient as a blocker of AD speaks against the importance of a larger efflux of glutamate during AD as compared to CSD. Studies of cultured hippocampal and cortical

6 228 M. LAURTZEN AND A.. HANSEN nerve cell cultures have demonstrated that acidosis inhibits the NMDA-induced Ca 2 + current (Giffard et a., 990), that mild acidosis protects the neurons from NMDA-mediated damage (Tombaugh and Sapolsky, 990), and that preservation of intracellular high-energy phosphates prevents the rundown of NMDA-gated channels (Mody et a., 988). This suggests that the NMDA receptor channel complex changes properties during ischemia. f so, NMDA receptor blockade by, e.g., MK-80 and APH may be neutralized or unimportant during the ischemic period when acidosis and energy depletion occur. During reperfusion, the "intrinsic" NMDA receptor blockade is expected to disappear concomitantly with the return to normal of ph and energy metabolism. The therapeutic time window of NMDA receptor blockers may be during this latter period. NBQX effectively inhibits the delayed neuronal death in the hippocampus following complete cerebral ischemia (Sheardown et a., 990), but has no effect on AD. Whether AD and the associated ion fluxes play any role in the development of neuronal cell death remains unknown. Finally, CSD may be a trigger mechanism of migraine attacks in humans (Lauritzen, 987). Considering the extreme sensitivity of CSD to NMDA antagonism, it is possible that NMDA receptor blockade could be a new therapeutic strategy for migraine. Acknowledgment: This work was supported by the Medical Research Council (Denmark), the Danish Migraine Society, the Foundation of 870, the Carlsberg Foundation, and the Foundation for Experimental Research in Neurology. REFERENCES Albers GW, Goldberg MP, Choi DW (989) N-Methyl-Daspartate antagonists: ready for clinical trial in brain ischemia? Ann Neurol 25: Amemori T, Bures J (990) Ketamine blockade of spreading depression: rapid development of tolerance. Brain Res 59: Astrup J, Rehncrona S, Siesjo BK (980) The increase in extracellular potassium concentration in the ischemic brain in relation to the preischemic functional activity and cerebral metabolic rate. Brain Res 99: 6-74 Benveniste H, Drejer J, Schousboe A, Diemer NH (984) Elevation of extracellular levels of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis. J Neurochem 43: Benveniste H, J rgensen MB, Diemer NH, Hansen AJ (988) Calcium accumulation by glutamate receptor activation is involved in hippocampal cell damage after ischemia. Acta Neurol Scand 78: Bures J, Buresova 0, Krivanek J (974) The Mechanism and Applications of Leao's Spreading Depression of Electroencephalographic Activity. New York, Academic Press Curtis DR, Watkins JC (963) Acidic amino acids with strong excitatory actions on mammalian neurons. J Physiol 66: - 4 Gardino FG, do Carmo RJ (983) Glutamate and spreading depression in chick retina. An Acad Bras Cienc 55: Giffard RG, Monyer H. Christine CW, Choi DW (990) Acidosis reduces receptor activation, glutamate neurotoxicity, and oxygen-glucose deprivation neuronal injury in cortical cultures. Brain Res 506: Gorelova NA. Koroleva V, Amemori T, Pavlik V, Bures J (987) Ketamine blockade of cortical spreading depression in rats. Electroencephalogr Clin Neurophysiol 66: Hansen AJ (978) The extracellular potassium concentration in brain cortex following ischemia in hypo- and hyperglycemic rats. Acta Physiol Scand 02: Hansen AJ (985) Effect of anoxia on ion distribution in the brain. Physiol Rev 66:0-48 Hansen AJ, Zeuthen T (98) Changes of brain extracellular ions during spreading depression and ischemia in rats. Acta Physiol Scand 3: Hansen AJ, Lauritzen M, Wieloch T (988) NMDA antagonists block cortical spreading depression but not anoxic depolarization. n: Frontiers in f.xcitatory Amino Acid Research (Cavalheiro EA. Lehmann J, Turski L, eds), New York, Alan R Liss, pp Hernandez-Caceres J, Macias-Gonzales R, Brozek G. Bures J (987) Systemic ketamine blocks cortical spreading depression, but does not delay the onset of terminal anoxic depolarization in rats. Brain Res 437: Lauritzen M (987) Cortical spreading depression as a putative migraine mechanism. Trends Neurosci 0:8-3 Lauritzen M, Rice ME, Okada YC Nicholson C (988) Quisqualate, kainate and NMDA can nitiate spreading depression in the turtle cerebellum. Brain Res 475: Lauritzen M, Hansen AJ, Kronborg D, Wieloch T (990) Cortical spreading depression is associated with arachidonic acid accumulation and preservation of energy charge. J Cereb Blood Flow Metab 0: 5-22 Lauritzen M. Sheardown M, Hansen AJ (99) The neuroprotective effect of the NMDA antagonist MK80 and non NMDA antagonist NBQX is not explained by an effect on anoxic depolarization, and only MK80 inhibits spreading depression. J Cereb Blood Flow Metab (suppl 2):S222 Leao AAP (944) Spreading depression of activity in the cerebral cortex. J Neurophysiol 7: Leao AAP (947) Further observations on the spreading depression of activity in the cerebral cortex. J Neurophysiol 0: Marrannes R, Willems R, De Prins E, Wauquier A (988) Evidence for a role of the N-methyl-D-aspartate (NMDA) receptor in cortical spreading depression in the rat. Brain Res 457: Mody, Lambert JDC, Heinemann U (987) Low extracellular magnesium induces epileptiform activity and spreading depression in rat hippocampal slices. J Neurophysiol 57: Mody, Salter MW, MacDonald JF (988) Requirement of NMDA receptor/channels for intracellular high-energy phosphates and the extent of intraneuronal calcium buffering in cultured mouse hippocampal neurons. Neurosci Lett 93:73-78 Nicholson C, Kraig RP (98) The behaviour of extracellular ions during spreading depression. n: The Application of on-selective Microelectrodes (Zeuthen T, ed), Amsterdam, Elsevier, pp Park CK, Nehls DG, Graham D, Teasdale GM, McCulloch J (988) The glutamate antagonist MK-80 reduces focal ischemic brain damage in the rat. Ann Neurol 24: Schaeffer P, Lazdunski M (99) K + effiux pathways and neurotransmitter release associated to hippocampal ischemia: effects of glucose and of K + channel blockers. Brain Res 539:55-58

7 GLUTAMATE N SPREADNG DEPRESSON AND SCHEMA 229 Scheller D, Heister U, Dengler K, Peters T (990) Do the excitatory amino acids aspartate and glutamate generate spreading depressions in vivo? n: Pharmacology of Cerebral schemia 990 (Krieglstein J, Oberpichler H, eds), Stuttgart, Wissentschaftliche Veriagsgesellschaft mbh, pp Sheardown MJ (989) The quinoxalinediones, a new series of potent and selective non-nmda receptor antagonists. Drugs Fut 4: Sheardown MJ, Nielsen E0, Hansen AJ, Jacobsen P, Honore T (990) 2,3-Dihydroxy-6-nitro-7 -sulfamoyl-benzo(f)quinoxaline: neuroprotectant for cerebral ischemia. Science 247: Siesjo BK, Bengtsson F (989) Calcium fluxes, calcium antagonists, and calcium-related pathology in brain ischemia, hypoglycemia, and spreading depression: a unifying hypothesis. J Cereb Blood Flow Metab 9:27-40 Simon RP, Swan JH, Griffiths T, Meldrum BS (984) Blockade of N-methyl-D-aspartate receptors may protect against ischemic damage in the brain. Science 226: Tombaugh GC, Sapolsky RM (990) Mild acidosis protects hippocampal neurons from injury induced by oxygen and glucose deprivation. Brain Res 506: Van Harreveld A (959) Compounds in brain extracts causing spreading depression of cerebral cortical activity and contraction of crustacean muscle. J Neurochem 3: Van Harreveld A, Fifkova E (970) Glutamate release from the retina during spreading depression. J Neurobiol 2: 3-29 Van Harreveld A, Kooiman M (965) Amino acid release from the cerebral cortex during spreading depression and asphyxiation. J Neurochem 2: Wong EHF, Kemp la, Priestley T, Knight AR, Woodruf GN, versen LL (986) The anticonvulsant MK80 is a potent N-methyl-D-aspartate antagonist. Proc Natl Acad Sci USA 83: J Cereb Blood Flow Me/ab, Vol. 2, No.2, 992