In vivo elevation of extracellular potassium in the rat amygdala increases extracellular glutamate and aspartate and damages neurons

Pergamon

PII: S0306-4522(96)00171-6

Neuroscience Vol. 74, No. 3, pp. 695–706, 1996 IBRO Published by Elsevier Science Ltd Printed in Great Britain

IN VIVO ELEVATION OF EXTRACELLULAR POTASSIUM IN THE RAT AMYGDALA INCREASES EXTRACELLULAR GLUTAMATE AND ASPARTATE AND DAMAGES NEURONS D. G. FUJIKAWA,*†‡ J. S. KIM,* A. H. DANIELS,* A. F. ALCARAZ* and T. B. SOHN* *Experimental Neurology Laboratory, Sepulveda VA Medical Center, Sepulveda, CA 91343, U.S.A. †Department of Neurology and Brain Research Institute, UCLA School of Medicine, Los Angeles, CA 90024, U.S.A. Abstract––It is well known that high potassium (K+) solutions introduced by microdialysis into normal brain increase the extracellular concentration of the excitatory amino acid glutamate, and in vitro studies suggest that a high exogenously applied glutamate concentration can produce excitotoxic neuronal death. However, only recently were in vivo studies undertaken to determine whether high-K+ exposure damages neurons. We implanted microdialysis probes into rat amygdalae bilaterally, and after a 2-h baseline period exposed one side to a modified Krebs–Ringer–bicarbonate solution containing 100 mmol/l KCl for 30, 50 and 70 min, followed by a 2-h recovery period, and 70 min and 3 h without a recovery period. Of 100.9&2.0 mmol/l KCl, 12.0&1.0% was extracted by amygdalar tissue in vivo. Elevation of the extracellular K+ concentration in the amygdala for 70 min or longer without a recovery period produced extensive neuronal damage and edematous-appearing neuropil in the tissue dialysed, as well as loss of normal neurons. Histological evidence of edema subsided in the groups with a 2-h recovery period. Although the number of damaged neurons was not significantly higher in the group with a 70 min high-K+ exposure and 2-h recovery period, the number of normal neurons was reduced, suggesting cell loss. During 70-min high-K+ exposure, the extracellular glutamate concentration increased to 242–377% of baseline during the first 60 min, and extracellular aspartate rose to 162–213% during the first 50 min; extracellular taurine rose even higher, to 316–567% of baseline, and glutamine fell to 14–27% of baseline. Extracellular serine was decreased at 20, 50 and 70 min of high-K+ exposure; extracellular glycine was unchanged. The elevated extracellular glutamate and aspartate concentrations suggest that exposure of the amygdala to high extracellular K+ may produce cell death through an excitotoxic process, and point the way to future studies to define the specific mechanisms involved. Key words: high potassium, cell death, neuronal depolarization, glutamine, taurine.

Introducing high-K+ concentrations in vivo through a microdialysis probe35,41,66 or in vitro into the medium bathing brain slices16 or dissociated neurons in tissue culture14,59 is a routine method of evoking neuronal depolarization and the release of amino acids such as glutamate and aspartate. In vivo perfusion of rat hippocampus with a high-K+ solution for 30 min depolarizes neurons and evokes the sustained release of glutamate into the extracellular space.66 It is well known that exogenously administered glutamate damages neurons in vitro15,59,60 and in vivo.64 Despite the evidence that introducing a high-K+ solution into the brain by microdialysis produces a substantial ‡To whom correspondence should be addressed. Abbreviations: ANOVA, analysis of variance; EEG, electroencephalograpic; EVF, extracellular volume fraction; HPLC, high-performance liquid chromatography; ID, ischemic depolarization; [K+]o, extracellular potassium concentration; KRB, Kreb’s–Ringer–bicarbonate; NMDA, N-methyl--aspartate; SD, spreading depression; SE, status epilepticus.

increase in the endogenous extracellular glutamate concentration (GLUo), until recently there were no data on whether this increase is associated with neuronal damage. In vitro evidence is inconsistent regarding whether exposure of neurons to high extracellular potassium concentration ([K+]o) is neurotoxic and whether the calcium influx triggered by high-K+ exposure is primarily through N-methyl-aspartate (NMDA)-receptor-operated cation channels33,57,62,67 or -type voltage-dependent Ca2+ channels.27,68 We undertook the current study to answer the following questions. (1) Does high-K+ exposure kill neurons in vivo? (2) What is the shortest exposure time required? (3) What percentage of 100 mmol/l KCl in the perfusate is extracted by tissue? (4) What happens to extracellular amino acid concentrations during high-K+ exposure? Previous microdialysis studies have used 100 mmol/l KCl to depolarize neurons30,53,66 and elevate extracellular glutamate;53,66 we used the same concentration for this

695

696

D. G. Fujikawa et al.

reason. Only one of these studies was concerned with the question of whether neuronal injury occurs,30 and none pursued the question of how much of the perfused 100 mmol/l KCl was actually taken up by the brain, both of which are addressed in this study. Our results have been published in part in abstract form.19,21 EXPERIMENTAL PROCEDURES

Materials Wistar rats were obtained from Charles River Laboratories (Wilmington, MA, U.S.A.). Carnegie-Medecin CMA/11 microdialysis probes (2 mm membrane length, 240 ìm outer diameter, 20,000 mol. wt cut-off) were used in all microdialysis experiments and were obtained from Bioanalytical Systems (West Lafayette, IN, U.S.A.). Potassium-sensitive electrodes and reference electrodes were purchased from Microelectrodes (Londonderry, NH, U.S.A.). Amino acid standards (25 nmol/ml) were purchased from Sigma Chemical Co. (St Louis, MO, U.S.A.); o-phthaldialdehyde was obtained from Pierce (Rockford, IL, U.S.A.); atropine (400 ìg/ml) was obtained from Elkins-Sinn (Cherry Hill, NJ, U.S.A.), ketamine HCl (Ketalar>, 50 mg/ml) from Parke-Davis (Morris Plains, NJ, U.S.A.) and pentobarbital sodium (Nembutal>, 50 mg/ml) from Abbott Laboratories (North Chicago, IL, U.S.A.). Bouin’s solution, used for brain fixation, was comprised of picric acid (Aldrich Chemical Company, Milwaukee, WI, U.S.A.), formaldehyde (37% solution; Sigma) and glacial acetic acid (VWR Scientific, Cerritos, CA, U.S.A.), and was prepared as previously described.44 Microdialysis experiments Male Wistar rats (245–450 g) were anesthetized with ketamine (25 mg/kg) and pentobarbital (40 mg/kg) i.p. for placement of skull screws for cortical electroencephalographic (EEG) recording and bilateral guide cannulae. The guide cannulae were positioned to place concentric CMA/11 microdialysis probes (2 mm membrane length) within the basolateral nuclei of the amygdalae (guide cannula stereotaxic coordinates: AP "3.0 mm, ML 4.7 mm, DV 7.0 mm55). In the last 18 of the 27 rats used in this study, bipolar electrodes, the tips of which extended 1 mm beyond the ends of the cannulae, were glued to the cannulae to permit local EEG recording 200 ìm from the probe surfaces. The next day the rats had abdomens and hindlimbs restrained, the probes were inserted bilaterally, and EEG and rectal temperature monitoring were begun (rectal temperatures remained at 37.6&1.0)C during these experiments). Krebs–Ringer–bicarbonate (KRB) solution (NaCl 122 mmol/l, KCl 3.4 mmol/l, CaCl2 1.2 mmol/l, MgSO4 1.2 mmol/l, KH2PO4 0.4 mmol/l and NaHCO3 25 mmol/l, at pH 7.4) was pumped through the probes at a rate of 4.0 ìl/min, and samples were obtained every 10 min. After a 2-h baseline period, the perfusate was randomly switched on one side to 100 mmol/l KCl in KRB, in which Na+ was appropriately reduced to maintain iso-osmolarity, for 30 min (n=6), 50 min (n=7), 70 min (n=10) or 3 h (n=4), followed by a switch back to KRB for 2 h in the 30- and 50-min groups. The 70-min group was subdivided into those with (n=5) and those without (n=5) a 2-h recovery period. The switch back to KRB solution for 2 h was to ensure enough time for neuronal cytoplasmic acidophilia and nuclear pyknosis, light-microscopic signs of neuronal necrosis,31 to appear. The dead space of the probes and the outlet tubing from the rats was 4.5 ìl; since the flow rate was 4 ìl/min, there was a 72-s lag-time from the brain to the dialysate collection

vial. In addition, there was 5.0 ìl of dead space from the microdialysis pump syringes to the probes, so that there was a 75-s delay once one side was switched to or from the high-K+ solution. The total time lag because of dead space was therefore 147 s, or approximately 2.5 min. Thus, the first (0–10 min) high-K+ sample represented dialysate fluid exposed to KRB solution for the first 2.5 min and to the high-K+ solution for 7.5 min, the second (10–20 min) sample, dialysate fluid exposed to high-K+ from 7.5 to 17.5 min, etc. In order to simplify the reporting of results, the first sample is called the 10-min sample, the second the 20-min sample, etc. Similarly, the first (0–10 min) recovery sample contained dialysate fluid exposed to the high-K+ solution for the first 2.5 min and to KRB solution for 7.5 min, up to the 120–130-min, or ‘‘2-h’’ sample, which represented dialysate fluid exposed to KRB solution for 117.5–127.5 min. At the end of 70 min or 3 h of high-K+ perfusion or 2 h following 30, 50 or 70 min of high-K+ perfusion, rats were deeply anesthetized with 400 mg/kg pentobarbital i.p. and underwent transcardiac brain perfusion–fixation with heparinized Ringer’s lactate solution for several minutes and then 200 ml of Bouin’s solution, at a perfusion pressure of 120 mmHg. Brains were removed 1–3 h later, put in Bouin’s solution overnight, then transferred to 70% ethanol until further processing. All efforts were made to minimize animal suffering, to keep the number of rats used to a minimum and to use alternatives to in vivo methods. The protocol was approved by the Sepulveda VA Medical Center’s Animal Studies Subcommittee of the Research and Development Committee, and the guidelines published in the NIH Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 86-23, Revised 1985) were followed. Histological processing Brains were processed to obtain 6-ìm-thick hematoxylinand eosin-stained coronal sections, as previously described.20 In brief, brains were cut coronally into 2–3-mm-thick blocks, then dehydrated in graded ethanol solutions and cleared with xylene. The blocks were then embedded in paraffin, sectioned at 6 ìm thickness, deparaffinized and hydrated to water, then stained with hematoxylin and eosin. Histological procedures Coronal sections were used which showed, if possible, the entire probe track; in all cases a section with the longest probe track was used. A 20-power objective and a square eyepiece grid divided into 10 54-ìm boxes at that power were used to count neurons within 270#540-ìm areas immediately medial and lateral to the edges of each probe track (270 ìm perpendicular and 540 ìm parallel to each edge). Only neurons lying entirely within each box or touching its right and bottom edges were counted. The neuronal counts on each side of each track were summed. Separate counts were obtained for acidophilic, irreversibly damaged neurons, normal-appearing neurons and the sum of the two, or total number of neurons. In addition, the distances that edema of the neuropil and acidophilic neurons extended from the edges of the probe tracks were recorded (in ìm) on both the high-K+ and KRB sides. The severity of the edema was estimated visually according to the following grading scale: 0=no edema, 0.5=slight edema, 1.0=mild edema, 2.0=moderate edema, and 3.0=severe edema. The above histological procedures were performed by one observer (D.G.F.), who was blinded with respect to which amygdala received the high-K+ solution. In vitro and in vivo measurements of probe dialysate extraction fractions for potassium The dialysate extraction fraction (E) is the relative extraction from the probe dialysate of a substance when the probe

High extracellular potassium damages neurons

697

Table 1. Percentage of 100 mmol/l KCl extracted by rat amygdala perfused with a CMA/11 microdialysis probe at a flow rate of 4 ìl/min

[K+]p in modified KRB [K+]d [K+]p " [K+]d Extraction fraction of K+ (%) Number of samples

In vitro

In vivo

101.2&0.5 mmol/l 90.6&0.4 mmol/l 10.6&0.4 mmol/l 10.9&0.4% 24

100.9&2.0 mmol/l 89.2&2.1 mmol/l 11.7&0.9 mmol/l 12.0&1.0% 42

Results are expressed as mean & S.E.M. [K+]p is the K+ concentration in the perfusate and [K+]d is the K+ concentration of dialysate samples. In the in vitro experiments, 100 mmol/l KCl in modified isotonic KRB solution was perfused through microdialysis probes used in the in vivo experiments while the probes were immersed in small tubes filled with standard KRB solution in a 37)C water bath. In both the in vitro and in vivo experiments, the K+ concentrations in the high KCl solution perfusing the probes and in the dialysate samples from the probes were measured with K+-sensitive electrodes. The values for the extraction fraction of 100 mmol/l KCl (expressed as a percentage) were calculated as described in Experimental Procedures. is used as an infusion device or the relative extraction from the tissue when the probe is used as a sampling device.13,48 In other words, it is the per cent efficiency (E#100) of a microdialysis probe in either delivering a substance to the outside environment (probe delivery) or recovering a substance from that environment (probe recovery). The efficiencies of some of the microdialysis probes used in 70-min high-K+ experiments in delivering 100 mmol/l KCl to a KRB solution in vitro and in delivering 100 mmol/l K+ to amygdala in vivo were measured with K+-sensitive microelectrodes at the perfusion rate of 4 ìl/min. The high-K+ solution used to deliver K+ to tissue consisted of a modified isotonic KRB solution containing 100 mmol/l KCl (KCl 100 mmol/l, NaCl 25 mmol/l, CaCl2 1.2 mmol/l, MgSO4 1.2 mmol/l, KH2PO4 0.4 mmol/l and NaHCO3 25 mmol/l, at pH 7.4). The in vitro probe delivery experiments were performed at 37)C to approximate body temperature more closely, since probe recoveries are temperature dependent.4 Potassium concentrations were determined by measuring the DC potential changes (mV) of isotonic KRB solutions containing 60–120 mmol/l KCl, plotting mV against the log of KCl concentrations, obtaining the log of K+ and calculating the K+ concentration from it. The DC potential changes of the 100 mmol/l KCl solutions used and of dialysate samples (collected in triplicate in vitro and at each 10-min time-point up to 70 min in vivo) were then measured, and the K+ concentrations of the dialysate samples were calculated. For the purpose of calculating the extraction fraction (E) of 100 mmol/l K+ in the perfusate (i.e. the fraction delivered), [K+]o at a distance far enough from the probe not to be affected by the dialysis was 3.4 mmol/l in the in vitro experiments and was assumed to be equal to this concentration in the in vivo experiments. Thus, E=(Cd "Cp)/(Co "Cp),

serine concentrations were analysed in order to control for non-specific increases in extracellular amino acid concentrations from high-K+ perfusion. Since the shortest period of high-K+ perfusion which produced neuronal damage was 70 min, the dialysate samples for rats in the 70-min high-K+ subgroups with and without a recovery period were analysed (four of the five rats without and six rats with 2-h recovery periods). Twenty-five microliters of 1 ìmol/l asparagine, which was used as an internal standard, was added to a 10-ìl aliquot of dialysate. A Shimadzu autosampler/controller was programmed to add 42 ìl of o-phthaldialdehyde to the 35-ìl dialysate/asparagine mixture, so that the final asparagine concentration was 0.325 ìmol/l, and 50 ìl of the total volume of 77 ìl was injected on to a ternary-gradient Shimadzu L600 chromatograph fitted with a Shimadzu fluorescence detector (340 nm excitation, 455 nm emission) and an Alltech 3-ìm 15-cm C18 analytical column. The o-phthaldialdehyde derivatives were separated by gradient elution with acetonitrile (10–23%) and 60–70 mmol/l sodium acetate buffer, pH 5.7 (77–90%). Chromatographic peaks were recorded and integrated with a Shimadzu integrator; the amino acid concentrations of the dialysate samples were obtained by comparing the integrated sample peaks to those of varying concentrations of amino acid standards. The calculated amino acid concentrations from the last three 10-min baseline dialysate samples were averaged for each rat (Table 2), and the amino acid concentrations of subsequent samples were expressed as the percentage of baseline for both the KRB and high-K+ sides (Fig. 1).

Table 2. Dialysate concentrations of amino acids

13,48

where Cd is the measured [K+] in the dialysate, Cp is the measured [K+] in the perfusate and Co is assumed to be 3.4 mmol/l. The probe delivery of 100 mmol/l K+, expressed as a percentage, is 100 # E. Results are summarized in Table 1. Determination of amino acid concentrations in the dialysate Precolumn derivatization with o-phthaldialdehyde, reversed-phase high-performance liquid chromatography (HPLC) and fluorescence detection were used to analyse the dialysate concentrations of the amino acids glutamate, aspartate, serine, glutamine, glycine and taurine, in the 70-min high-K+ group (n=10), using a modification of the method described by Jarrett and co-workers.32 Glycine and

Baseline concentrations (ìmol/l) High-K+ side KRB side Aspartate Glutamate Serine Glutamine Glycine Taurine

0.27&0.06 0.37&0.07 1.16&0.16 5.81&1.26 0.88&0.17 0.67&0.14

0.25&0.05 0.32&0.05 1.15&0.15 5.36&1.04 0.94&0.14 0.57&0.11

Values represent mean & S.E.M. (n=10). The baseline dialysate concentrations for these amino acids are comparable to those obtained in the rat hippocampus9 and neocortex18 with 2-mm-long CMA/10 probes, and in human hippocampus and frontal and temporal neocortex with 6- or 10-mm-long CMA/10 probes.58

698

D. G. Fujikawa et al.

Fig. 1. Changes in the dialysate concentrations of amino acids relative to baseline concentrations during 70 min of high-K+ (filled bars) or control KRB perfusion (open bars), and during the last 30 min of a 2-h recovery period (Rec). Results are expressed as mean & S.E.M.; n=10 except for the 2-h recovery sample (n=6). *P¦0.05, **P¦0.01, ***P¦0.001, using one-way repeated-measures ANOVA and post hoc t-tests with pooled S.D. Although the pooled S.D. was used in post hoc t-tests, the individual S.D. for each time-point is shown here. The dashed lines represent the baseline concentrations of each amino acid on both the high-K+ and KRB sides.

Statistical analysis A one-way repeated-measures analysis of variance (ANOVA) was used to compare each of the six amino acids to its baseline concentration on both the high-K+- and the KRB-perfused sides at 10-min intervals during the 70 min of high-K+ perfusion. In the group with a 2-h recovery period, the amino acid concentrations during the last 30 min of the recovery period were also compared to their baseline concentrations. Post hoc paired t-tests were performed with pooled S.D., with á=0.05. One-way ANOVA and post hoc paired t-tests were used to compare the number of necrotic neurons, normal neurons and total number of neurons on the high-K+ and KRB sides after 30, 50, 70 min and 3 h of high-K+ perfusion, with á=0.05. One-way ANOVA and post hoc paired t-tests were also used to compare the distances that edema and acidophilic neurons extended

from the probe tracks. To determine whether the edema scores were normally distributed so that parametric statistical testing could be carried out, ‘‘residuals’’ (individual scores less group means) were plotted for both the high-K+ and KRB sides. Normal curves were approximated, so paired t-tests were also used to compare edema scores of the high-K+ and KRB sides in each of the five groups. RESULTS

In vitro and in vivo probe recoveries for potassium We tested CMA/11 2-mm probes used in the in vivo microdialysis experiments and found that about 11– 12% of a 100 mmol/l KCl modified KRB solution perfused through the probes was taken up by both

High extracellular potassium damages neurons Table 3. In vitro dialysis probe recoveries (%)

Aspartate Glutamate Serine Glutamine Glycine Taurine

High-K+ side

KRB side

2.84&0.29 2.87&0.27 4.26&0.56 3.03&0.57 4.56&0.83 4.00&0.71

3.10&0.45 3.19&0.47 4.52&0.67 3.11&0.51 4.25&0.68 3.86&0.69

Values represent mean & S.E.M. (n=10). These in vitro recovery values were obtained by placing 2-mm-long CMA/11 microdialysis probes in a 25 ìmol/l standard amino acid solution, perfusing the probes with KRB solution and measuring the concentrations of amino acids recovered in the dialysate. These recovery values underestimate the in vivo recoveries of amino acids from the extracellular space of the rat amygdala because of characteristics of the tissue, the most comprehensive quantitative analysis of which is presented in Bungay et al.13 and Morrison et al.48

amygdala in vivo and standard KRB solution in vitro at a flow rate of 4.0 ìl/min (Table 1). The fact that in vivo and in vitro probe recoveries for K+ are essentially the same was also shown previously with an artificial cerebrospinal fluid solution containing 3.0 mmol/l K+.7 For many compounds, including the amino acids that we measured, in vitro and in vivo probe recoveries differ, with the in vivo recoveries tending to be higher because of the complicating factors in the extracellular space, which include the extracellular volume fraction (EVF, or the fraction of total tissue volume which is extracellular), a ‘‘tortuosity’’ factor, metabolism, intracellular–extracellular and extracellular–microvascular exchange.7,13,48 However, this is not the case for K+, which rapidly crosses cell membranes, a consequence of which is that its diffusion characteristics are similar in KRB and brain tissue.7 Extracellular amino acid high-potassium exposure

concentrations

during

The baseline dialysate concentrations of the six amino acids that we measured are listed in Table 2; Table 3 gives the in vitro probe recoveries for each amino acid. The extracellular concentrations of glutamate rose to 242–377% of baseline, and aspartate (ASPo) to 162–213% during the first 60 min and 50 min, respectively, of a 70-min exposure to highK+. On the high-K+ side there were dramatic increases in extracellular taurine (TAUo) (to 316– 567% of baseline) and reductions in extracellular glutamine (GLNo) (to 14–27% of baseline) throughout the 70-min exposure time (Fig. 1). The mean concentrations of serine on the high-K+ side were in general lower than their baseline concentrations and were significantly decreased to 67%, 71% and 79% of baseline after 20, 50 and 70 min of high-K+ perfusion, respectively (Fig. 1); in addition, serine was also decreased to 78% of baseline in the control

699

KRB-perfused amygdala 60 min after the start of high-K+ perfusion. High-K+ perfusion produced no significant changes in glycine, except at 20 min on the KRB side, at which time glycine rose to 164% of baseline, an unexpected and probably fortuitous result (Fig. 1). High-potassium-induced neuronal damage in the dialysed tissue We counted neurons in 270#540-ìm areas medial and lateral to the probes in hematoxylin and eosin coronal sections showing the maximal probe track length. Immediately after 70 min and 3 h of high-K+ perfusion there were 1.8–2.4 times as many acidophilic neurons on the high-K+ side as on the KRB side, and the normal and total numbers of neurons were reduced after 70 min and markedly reduced after 3 h of high-K+ perfusion (Table 4 and Fig. 2). Although no differences in the numbers of acidophilic neurons were found between the high-K+- and control KRB-perfused sides in the groups with a 2-h recovery period, the number of normal neurons was substantially lower and the total number of neurons was also lower on the high-K+ side after 70 but not 30 nor 50 min of high-K+ exposure (Table 4). There was also pronounced edema of the neuropil on the high-K+ side immediately after 70 min and 3 h of high-K+ exposure, but not after 30, 50 or 70 min of high-K+ exposure and a 2-h recovery period, although edema was seen 1.7 times as far away from the probe tracks 2 h after 70 min of high-K+ exposure compared to the control side (Table 5 and Fig. 2). Coincident with the more pronounced edema immediately after 70 min and 3 h of high-K+ exposure, the maximal distance away from the probes that edema was seen was 4.0 and 4.9 times that of the respective control sides. Also, the mean distance away from the probe tracks that acidophilic neurons were seen immediately after 70 min of high-K+ perfusion was 2.3-fold greater on the high-K+ side than the KRB side; 2 h after 30, 50 and 70 min of high-K+ perfusion and immediately after 3 h of high-K+ perfusion the difference was not statistically significant (Table 5). Seizure activity during high-potassium perfusion Our experiments occasionally triggered behavioral limbic seizures and frontal EEG spike discharges. This occurred in four out of 31 (13%); these rats were excluded from the study. Of the 27 rats included in this study, alternating current bipolar EEG recordings within 200 ìm of the probe surface in the last 18 consistently showed normal EEG activity on both the high-K+- and KRB-perfused sides. These bipolar electrodes recorded EEG activity with amplitudes in the 20–200 ìV range and frequencies in the 50– 500 ms range. Although these electrodes can detect seizure discharges, they cannot measure high-K+induced, spreading-depression (SD)-like neuronal

700

D. G. Fujikawa et al. Table 4. Relationship between the duration of high-K+ exposure and neuronal damage 30 min 2-hour Recovery High-K+ KRB side side

Number of neurons Acidophilic 31&5 Normal 120&8 Total 151&7

36&4 112&7 147&9

50 min 2-hour Recovery High-K+ KRB side side

70 min 2-hour Recovery High-K+ KRB side side

70 min No Recovery High-K+ KRB side side

26&3 142&10 167&9

23&4 118&14 141&17

35&7 103&11 138&9

17&4 143&6 160&8

17&5 145&9** 162&13*

3h No Recovery + High-K KRB side side

19&6** 149&10** 168&6**

40&1 46&8 85&9

17&5** 121&14* 138&12*

Results are mean & S.E.M. *P¦0.05, **P¦0.01, using one-way ANOVA and post hoc two-tailed paired t-tests to compare the amygdala given isotonic high-K+ solution to its contralateral control (KRB) side at each time point. The numbers of rats used for 30-, 50- and 70-min high-K+ perfusions with a 2-h recovery period were 6, 7 and 5, respectively; the numbers used for 70-min and 3-h high-K+ perfusions without a recovery period were 5 and 4, respectively. The number of neurons in 270#540-ìm areas medial and lateral to the surfaces of microdialysis probes were counted and summed for each rat in representative coronal brain sections. The decrease in the numbers of normal neurons in the 70-min high-K+ perfusion groups with and without a 2-h recovery period and the even more pronounced decrease in the 3-h high-K+ perfusion group are probably due to an ongoing loss of neurons. The reason for the high number of acidophilic neurons on the control KRB side with a 30-min high-K+ exposure is unclear; in the other four groups the KRB side showed a remarkably consistent lower number of acidophilic neurons. Table 5. Extent of edema and neuronal necrosis 30 min 2-hour Recovery High-K+ KRB side side Edema score 1.0&0.2 Edema (ìm) 86&11 Acidophilic neurons (ìm) 117&17

50 min 2-hour Recovery High-K+ KRB side side

70 min 2-hour Recovery High-K+ KRB side side

70 min No Recovery High-K+ KRB side side

3h No Recovery + High-K KRB side side

1.0&0.1 72&6

1.3&0.2 85&9

0.8&0.1 69&6

1.1&0.4 150&46

0.8&0.4 89&36*

2.7&0.2 252&15

0.7&0.1** 63&6***

3.0&0.0 297&35

0.9&0.2** 61&7**

140&18

108&12

100&14

116&15

99&16

188&15

83&10**

179&34

112&28

Results are mean & S.E.M. *P¦0.05, **P¦0.01, ***P¦0.001, using one-way ANOVA and post hoc two-tailed paired t-tests to compare the amygdala given isotonic high-K+ solution for the specified time to its contralateral control side, which received KRB solution. The number of rats in each group is the same as in Table 4. The values for the edema score (0=no edema, 0.5=slight, 1=mild, 2=moderate and 3=severe) were compared with paired t-tests because ‘‘residuals’’ (individual scores less group means) approximated normal curves, thereby allowing parametric testing. The edema scores represent the most severe edema seen, the severity of which decreased with the distance from the probe tracks. The values for edema and acidophilic neurons in ìm are the maximal distances from the probe tracks at which each was seen.

depolarizations, which are "20 to "40 mV changes in the DC potential of the dialysed tissue lasting 1 min or longer.30,66 DISCUSSION

Our data provide direct histological evidence that a 70-min high-K+ perfusion, which elevates GLUo to 242–377% of baseline during the first 60 min and ASPo to 162–213% of baseline during the first 50 min, also produces irreversible neuronal damage in previously normal brain. Associated with the elevated GLUo was an even greater increase in TAUo, to 316–567% of baseline, and a dramatic drop in GLNo, to 14–27% of baseline, throughout the 70-min high-K+ exposure. Of the 100 mmol/l KCl in the perfusate, approximately 12% was extracted by amygdalar tissue, so that during each high-K+ perfusion [K+]o adjacent to the probes was elevated above baseline levels. Recurrent SD-like neuronal depolarizations and rapid increases in [K+]o to the 60 mmol/l range were produced by microdialysis of the rat hippocampus with 100 or 125 mmol/l KCl,30

and it is likely that similar depolarizations and increases in [K+]o occurred in our experiments. The appearance of cellular edema and a consequent decrease in the EVF also contributes to elevations in [K+]o (see below). High-potassium-induced changes amino acid concentrations

in

extracellular

The baseline dialysate concentrations of the amino acids that we measured (Table 2) are comparable not only to those obtained in rat hippocampus and neocortex with 2-mm-long, 500-ìm-diameter CMA/10 microdialysis probes9,18 but also to those obtained from human hippocampal and frontal and temporal neocortical tissue with 6- and 10-mm-long, 500-ìm-diameter CMA/10 microdialysis probes.58 Based upon our in vitro dialysate glutamate concentrations (Table 2) and probe recovery results (Table 3), the baseline in vivo tissue extracellular glutamate concentration would be approximately 11 ìmol/l (0.34 ìmol/l/0.0303, the mean dialysate concentration and probe recovery for glutamate from Table 2 and

High extracellular potassium damages neurons

Table 3). However, the in vivo probe recovery values are underestimated by in vitro measurements because of different diffusion characteristics of glutamate in situ and in an aqueous solution.7,13,48 Therefore the high-K+-induced 2.4–3.8-fold elevation of extracellular glutamate would in vivo be even greater than 27–42 ìmol/l. The in vitro ED50 concentrations of a 5-min exposure of glutamate to fetal murine cortical neurons in tissue culture which induced neuronal death were in the 50–100 ìmol/l range.15 Although of interest, particularly since our estimated in vivo glutamate concentrations approached the lethal in vitro concentrations reported by Choi et al.,15 a direct comparison between the two should be approached with caution because of the obvious differences in conditions. In addition, there is a substantial difference between the estimated GLUo in normal brain obtained from in vivo microdialysis data and that based on the stoichiometry of the glutamate uptake carrier. Assuming an intracellular glutamate concentration of 10 mmol/l,36 Bouvier et al.10 calculated a minimum GLUo of 0.6 ìM from the Nerst equation. However, Benveniste et al.9 found that the average dialysate concentration from the CA1 region of the rat hippocampus was 0.44 ìM, remarkably close to our average baseline values of 0.37 and 0.32 ìM for rat amygdalae (Table 2). Taking into account her measured in vitro probe recovery for glutamate of 5% and the different diffusion characteristics of glutamate in brain tissue compared to an aqueous solution,7,13,48 Benveniste estimated that GLUo is at least 180-fold higher than her observed average dialysate concentration.8 Although the disparity in GLUo calculated from microdialysis data and from the stoichiometry of the glutamate uptake carrier has not been resolved, it should be remembered that the value of 0.6 ìM calculated by Bouvier et al.10 is the minimum expected GLUo, and that concentric microdialysis probes in the order of 2 mm in length and 240– 520 ìm in diameter do not sample synaptic glutamate concentrations. A reduction in GLNo has been observed in status epilepticus (SE),12,22,38,47,70 transient global cerebral ischemia,5,53 hypoglycemia61 and high-K+-perfusion of hippocampus41,53 and striatum.35 Glutamine, which is converted to glutamate in presynaptic terminals, may be reduced because an increased rate of presynaptic glutamate release triggers an increased rate of presynaptic glutamine uptake, which may outpace its release from glial cells.51 The actions of taurine are complex, and the reasons for its K+-induced elevation may be multifold. It is well documented that TAUo increases when extracellular hypoosmolarity and cell swelling occur.39,65,69 Cell swelling may reduce the EVF by up to 50%, which is the extent to which it is reduced in SD and cerebral hypoxia–ischemia.25 Thus, up to a 200% increase in GLUo, ASPo and TAUo in these conditions may be due to a reduction in the EVF

701

alone. Cell swelling may be the reason that TAUo is increased during high-K+ perfusion and during SE,22,37,40,70 transient global cerebral ischemia1,5,24,53 and hypoglycemia.61 The reduction of SERo to about 70% of baseline during high-K+ perfusion was unexpected and of uncertain cause. A decrease in hippocampal SERo to about 80% of baseline was found during kainic-acidinduced SE,12 but no change was found in other studies of SE.22,47,70 Increases in SERo up to 125% and 153% of baseline occurred during cortical SD and anoxic depolarization, respectively,18 and there was either no change5 or an increase as high as 270% of baseline1 during transient global cerebral ischemia. Bruhn et al.12 suggest that since serine is a precursor in the biosynthesis of glycine,17 the reduction of SERo that they found during SE could be the result of increased intracellular glycine turnover triggering increased cellular uptake of serine. A similar mechanism may have been responsible for our results. High-potassium-induced neuronal damage in the dialysed tissue A comparison of the edema results of the 70-min high-K+ perfusion subgroups with and without a 2-h recovery period suggests that the edema of the neuropil was mild 2 h following 30, 50 and 70 min of high-K+ perfusion because more severe edema subsided during the 2-h recovery period (Table 5). A 2-h recovery period was used after shorter high-K+ exposure times in order to allow enough time for acidophilic neurons to appear. However, the 2-h recovery period complicated the interpretation of results in the 70-min high-K+ perfusion group. The difference between the high-K+ and KRB sides in the number of acidophilic neurons in the subgroup with 2 h of recovery just missed statistical significance (P=0.06; Table 4). The significant decrease in the number of normal neurons on the high-K+ compared to the KRB side in this subgroup strongly suggests that the lack of significant difference in the numbers of damaged neurons between the two sides was caused by lysis and disappearance of those neurons during the recovery period (Table 4). It is unlikely that loss of neurons occurred at 30 and 50 min of high-K+ perfusion, because the number of normal neurons was not reduced on the high-K+ side, as it was at 70 min of high-K+ perfusion. Severe edema cannot explain the reduced number of normal neurons and total number of neurons immediately after 70 min of high-K+ perfusion, because the number of normal neurons and total number of neurons in the group with a 2-h recovery period were also decreased to a comparable extent without the presence of significant edema (31% vs 19% reductions in normal neurons, respectively; P=0.44; Table 4 and Table 5). The increased number

702

D. G. Fujikawa et al.

Fig. 2.

High extracellular potassium damages neurons

of damaged neurons following 3 h of high-K+ perfusion occurred in spite of severe edema and 62% and 38% reductions in the number of normal neurons and total number of neurons, respectively (Table 4 and Table 5). The reduction in the total neuron count probably reflects ongoing lysis of dead neurons during the longer high-K+ exposure time. It is unlikely that the edematous neuropil itself was responsible, since (i) comparable edema was also present immediately after 70 min of high-K+ perfusion, at which time the number of normal neurons was two-fold higher, and (ii) there was a significant decrease in the number of normal neurons after 70 min of high-K+ perfusion and a 2-h recovery period, even though there was only mild edema (Table 4 and Table 5). Type of cell death produced by high-potassium exposure There are two major forms of cell death, necrosis and apoptosis.71 The light-microscopic and ultrastructural features of neuronal necrosis have been described in such pathological conditions as cerebral hypoxia–ischemia11 and SE.31,50 However, recent reports indicate that both global29,34,45,52,54,63 and focal cerebral ischemia42,43,46 and prolonged seizures induced by kainic acid56 can trigger apoptosis, which was previously thought to be restricted to normal developmental processes, cell turnover, tissue atrophy, cell-mediated immunity, neoplasms and miscellaneous stimuli such as radiation, hyperthermia and chemotherapeutic drugs.72 Acidophilic cells by light microscopy have been correlated with electronmicroscopic evidence of cellular necrosis in the brain2,31 and in peripheral tissue.71 It is likely that the acute neuronal death which occurs during high-K+ exposure is necrosis because of the rapidity of its appearance and because the irreversibly damaged neurons are acidophilic. However, a more definitive determination of the type of cell death produced by high-K+ exposure awaits further study. How long should experiments be delayed after microdialysis probe insertion? Because of focal and generalized changes in cerebral blood flow and glucose utilization which were demonstrated during the first 2 h after probe insertion, changes which largely resolved 24 h later,6 it is

703

widely accepted that experiments should be delayed for up to 24 h. Our waiting period of 1.5 h after probe implantation before collecting baseline samples could be criticized on this basis, although many studies have employed a baseline interval of 1.5 h or less, including those by several groups investigating changes in extracellular amino acids by microdialysis during cerebral ischemia.3,9,23 In addition, the baseline GLUo in striatum was raised 10-fold and GLNo was reduced eight-fold at 24 h compared to 2 h after probe implantation, and high-K+ perfusion failed to increase GLUo at 24 h, whereas a four-fold increase was found at 2 h,53 comparable to the 2.4–3.8-fold increases we found in the amygdala. These data suggest that to demonstrate K+-invoked changes in extracellular amino acid concentrations, experiments should be performed within several hours after probe implantation. Finally, we found that 3–5 h after probe implantation (i.e. after 2-h baseline periods and (i) 70 min and 3 h of high-K+ perfusion or (ii) a 2-h recovery period following 30, 50 and 70 min of high-K+ perfusion), there was minimal tissue disruption adjacent to the probes on the control KRB side (Tables 4 and 5 and Fig. 2). In vivo studies relating high-potassium exposure to neuronal damage It has been known for years that intracerebral microdialysis of high-K+ solutions evokes the release of glutamate and other amino acids into the extracellular space,35,41,66 but the question of whether the high-K+-induced elevation of GLUo damages neurons has only recently been raised. In the only other in vivo study which addressed this question prior to the current investigation, the rat hippocampus was perfused with 100 or 125 mmol/l KCl through microdialysis probes.30 Direct current potential recordings showed that recurrent waves of a SD-type of neuronal depolarization were produced, and in about 50% of the rats studied, a prolonged, unstable SD occurred (i.e. rapid, repetitive depolarizations for at least 5 min). However, irreversible loss of evoked potentials occurred consistently only with combined irrigation of surgically exposed hippocampus plus microdialysis of 100 or 125 mmol/l KCl. Assuming similar dialysis membrane characteristics to ours, the in vivo probe recoveries in these

Fig. 2. Seventy minutes of high-K+ perfusion produces extensive neuronal damage and edema. Photomicrographs are shown of the amygdalae of a rat dialysed for 70 min with KRB solution on one side (A, C), and 100 mmol/l KCl in modified KRB solution on the other (B, D); about 12% of the 100 mmol/l KCl was taken up by tissue (see Results section and Table 1). (C, D) Higher power views of the areas outlined by the rectangles in A and B. The probe surface is at the right edge of the photograph in C and at the left edge in D. Scale bars=120 ìm (A, B); 30 ìm (C, D). The difference between the two sides is apparent even in the low-power views (A, B). In C there are a few dead neurons (arrows point to two) and many normal-appearing neuronal nuclei (the arrowhead points to one); aside from mild edema on the right side of the photomicrograph, the neuropil looks normal. In D severe edema of the neuropil and extensive neuronal damage are evident. The cell bodies of the acidophilic neurons (arrows point to some) are shrunken and surrounded by edematous-appearing neuropil. Only a few relatively normal-appearing neuronal nuclei can be seen (the arrowhead points to one).

704

D. G. Fujikawa et al.

experiments would have been 8.2% or 10.2%, comparable to but lower than our in vivo data. The minimal period of time required for electrophysiological evidence of high-K+-induced irreversible loss of the hippocampal CA1 population spike was 60 min, very close to the 70 min of high-K+ exposure needed to kill amygdala neurons in our study (Table 4). Relationship of high-potassium experiments spreading depression and ischemic depolarization

to

We found that of the 100 mmol/l KCl in the perfusate, 12.0&1.0% was taken up by dialysed amygdalae in vivo (Table 1). It has been shown that if [K+]o exceeds the ‘‘ceiling’’ of 10 mmol/l which occurs during seizures,28 SD and a rapid rise in [K+]o to the 60 mmol/l range occur.26 As a result of ischemic depolarization (ID), extracellular K+ also increases to the 60 mmol/l range during cerebral ischemia.26 Although we did not measure tissue DC potentials, it is likely that the elevated [K+]o triggered SD-like depolarizations in our model, as Herreras and Somjen30 found in the rat hippocampus. It is probably the short duration of the K+-induced neuronal depolarizations in SD which explains its lack of damaging effects.49 Neurons and glia can normally establish ionic equilibrium quickly between the intracellular and extracellular spaces, dropping [K+]o and re-establishing a normal or nearly normal resting potential. Continuous microdialysis of a high-K+ solution may hamper the ability of neurons and glia to re-establish ionic gradients continuously between the intracellular and extracellular compartments, and results in irreversible neuronal damage if the exposure to high K+ persists for 70 min under our conditions. Under hypoxic–ischemic conditions, where [K+]o remains elevated and neuronal depolarization persists because of ATP depletion and energy

failure, the time required for irreversible cell injury may be shortened to a few minutes. The effects of SD and ID on rat neocortical extracellular amino acid concentrations were reported recently.18 SD resulted in significant increases in extracellular concentrations of eight out of nine amino acids, including GLUo (163&9%; mean & S.E.M.), ASPo (160&17%), GLYo (158&21%) and TAUo (172&15%). Five minutes after decapitation (i.e. during ID), eight out of 10 amino acid concentrations were elevated, including GLUo (1696&546%), ASPo (3458&656%), GLYo (297& 37%) and TAUo (1721&98%). Interestingly, the elevations of GLUo (242–377%), ASPo (162–213%) and TAUo (316–567%) in the current study fall between the results obtained by Fabricius et al. for SD and ID. CONCLUSIONS

In the current study, the high-K+-induced elevations of GLUo and ASPo in the normal rat amygdala in vivo were associated with neuronal death and edematous-appearing neuropil 70 min after the onset of high-K+ exposure. Perfusion of an isotonic high-K+ solution into normal brain through a microdialysis probe is a well-known method of depolarizing neurons, and this study indicates that if the high-K+ perfusion persists for a sufficient period of time then neuronal death ensues. The elevated GLUo and ASPo suggest an excitotoxic process, but the specific mechanisms involved must be clarified by further investigation. Acknowledgements—This research was supported by the Research Service of the Department of Veterans Affairs. Dr Jeffrey Gornbein of the UCLA Department of Biomathematics provided expert assistance in statistical methodology.

REFERENCES

1. 2. 3. 4. 5.

6. 7. 8. 9.

10. 11.

Andine´ P., Orwar O., Jacobson I., Sandberg M. and Hagberg H. (1991) Changes in extracellular amino acids and spontaneous neuronal activity during ischemia and extended reflow in the CA1 of the rat hippocampus. J. Neurochem. 57, 222–229. Auer R. N., Kalimo H., Olsson Y. and Siesjo¨ B. K. (1985) The temporal evolution of hypoglycemic brain damage. I. Light- and electronmicroscopic findings in the rat cerebral cortex. Acta neuropath., Berlin 67, 13–24. Baker A. J., Zornov M. H., Scheller M. S., Yaksh T. L., Skilling S. R., Smullin D. H., Larson A. A. and Kuczenski R. (1991) Changes in extracellular concentrations of glutamate, aspartate, glycine, dopamine, serotonin and dopamine metabolites after transient global ischemia in the rabbit brain. J. Neurochem. 57, 1370–1379. Benveniste H. (1989) Brain microdialysis. J. Neurochem. 52, 1667–1679. Benveniste H., Drejer J., Schousboe A. and Diemer N. H. (1984) Elevation of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis. J. Neurochem. 43, 1369–1374. Benveniste H., Drejer J., Schousboe A. and Diemer N. H. (1987) Regional cerebral glucose phosphorylation and blood flow after insertion of a microdialysis fiber through the dorsal hippocampus in the rat. J. Neurochem. 49, 729–734. Benveniste H., Hansen A. J. and Ottosen N. S. (1989) Determination of brain interstitial concentrations by microdialysis. J. Neurochem. 52, 1741–1750. Benveniste H. and Hu¨ttemeier P. C. (1990) Microdialysis—theory and application. Prog. Neurobiol. 35, 195–215. Benveniste H., Jørgensen M. B., Sandberg M., Christensen T., Hagberg H. and Diemer N. H. (1989) Ischemic damage in hippocampal CA1 is dependent on glutamate release and intact innervation from CA3. J. cerebr. Blood Flow Metab. 9, 629–639. Bouvier M., Szatkowski M., Amato A. and Attwell D. (1992) The glial glutamate uptake carrier countertransports pH-changing anions. Nature 360, 471–474. Brown A. W. (1977) Structural abnormalities in neurones. J. clin. Path. 30 Suppl. 11, 155–169.

High extracellular potassium damages neurons 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

25. 26. 27. 28. 29.

30. 31. 32. 33. 34. 35. 36.

37.

38. 39. 40. 41. 42. 43. 44. 45.

705

Bruhn T., Cobo M., Berg M. and Diemer N. H. (1992) Limbic seizure-induced changes in extracellular amino acid levels in the hippocampal formation: a microdialysis study of freely moving rats. Acta neurol. scand. 86, 455–461. Bungay P. M., Morrison P. F. and Dedrick R. L. (1990) Steady-state theory for quantitative microdialysis of solutes and water in vivo and in vitro. Life Sci. 46, 105–119. Choi D. W. (1987) Ionic dependence of glutamate neurotoxicity. J. Neurosci. 7, 369–379. Choi D. W., Maulucci M. and Kriegstein A. R. (1987) Glutamate neurotoxicity in cortical cell culture. J. Neurosci. 7, 357–368. Crowder J. M., Croucher M. J., Bradford H. F. and Collins J. F. (1987) Excitatory amino acid receptors and depolarization-induced Ca2+ influx into hippocampal slices. J. Neurochem. 48, 1917–1924. Daly E. C. and Aprison M. H. (1983) Glycine. In Handbook of Neurochemistry (ed. Lajtha A.), Vol. 3, pp. 467–499. Plenum Press, New York. Fabricius M., Jensen L. H. and Lauritzen M. (1993) Microdialysis of interstitial amino acids during spreading depression and anoxic depolarization in rat neocortex. Brain Res. 612, 61–69. Fujikawa D. G., Daniels A. H. and Kim J. S. (1992) How long a period of depolarization induced by high extracellular potassium is needed to kill rat amygdala neurons in vivo? Soc. Neurosci. Abstr. 18, 1361. Fujikawa D. G., Daniels A. H. and Kim J. S. (1994) The competitive NMDA-receptor antagonist CGP 40116 protects against status epilepticus-induced neuronal damage. Epilepsy Res. 17, 207–219. Fujikawa D. G., Kim J. S., Daniels A. H. and Alcaraz A. F. (1993) In vivo elevation of extracellular potassium in the rat amygdala increases extracellular glutamate and damages neurons. J. cerebr. Blood Flow Metab. 13 Suppl. 1, S103. Fujikawa D. G., Kim J. S., Daniels A. H. and Alcaraz A. F. (1993) Lithium-pilocarpine-induced status epilepticus produces changes in extracellular amino acid concentrations in the rat amygdala. Soc. Neurosci. Abstr. 19, 1697. Globus M. Y.-T., Busto R., Martinez E., Valde´s I., Dietrich W. D. and Ginsberg M. D. (1991) Comparative effect of transient global ischemia on extracellular levels of glutamate, glycine, and ã-aminobutyric acid in vulnerable and nonvulnerable brain regions in the rat. J. Neurochem. 57, 470–478. Hagberg H., Lehmann A., Sandberg M., Nystro¨m B., Jacobson I. and Hamberger A. (1985) Ischemia-induced shift of inhibitory and excitatory amino acids from intra- to extracellular compartments. J. cerebr. Blood Flow Metab. 5, 413–419. Hansen A. J. and Olsen C. E. (1980) Brain extracellular space during spreading depression and ischemia. Acta physiol. scand. 108, 355–365. Hansen A. J. and Zeuthen T. (1981) Extracellular ion concentrations during spreading depression and ischemia in the rat brain cortex. Acta physiol. scand. 113, 437–445. Hartley D. M., Kurth M. C., Bjerkness L., Weiss J. H. and Choi D. W. (1993) Glutamate receptor-induced 45Ca2+ accumulation in cortical cell culture correlates with subsequent neuronal degeneration. J. Neurosci. 13, 1993–2000. Heinemann U. and Lux H. D. (1977) Ceiling of stimulus induced rises in extracellular potassium concentration in the cerebral cortex of cat. Brain Res. 120, 231–249. Heron A., Pollard H., Dessi F., Moreau J., Lasbennes F. and Ben-Ari Y. (1993) Regional variability in DNA fragmentation after global ischemia evidenced by combined histological and gel electrophoresis observations in the rat brain. J. Neurochem. 61, 1973–1976. Herreras O. and Somjen G. G. (1993) Effects of prolonged elevation of potassium on hippocampus of anesthetized rats. Brain Res. 617, 194–204. Ingvar M., Morgan P. F. and Auer R. N. (1988) The nature and timing of excitotoxic neuronal necrosis in the cerebral cortex, hippocampus and thalamus due to flurothyl-induced status epilepticus. Acta neuropath., Berlin 75, 362–369. Jarrett H. W., Cooksy K. D., Ellis B. and Anderson J. M. (1986) The separation of o-phthalaldehyde derivatives of amino acids by reversed-phase chromatography on octylsilica columns. Analyt. Biochem. 153, 189–198. Jing J., Aitken P. G. and Somjen G. G. (1991) Lasting neuron depression induced by high potassium and its prevention by low calcium and NMDA receptor blockade. Brain Res. 557, 177–183. Kihara S., Shiraishi T., Nakagawa S., Toda K. and Tabuchi K. (1994) Visualization of DNA double strand breaks in the gerbil hippocampal CA1 following transient ischemia. Neurosci. Lett. 175, 133–136. Korf J. and Venema K. (1985) Amino acids in rat striatal dialysates: methodological aspects and changes after electroshock. J. Neurochem. 45, 1341–1348. Kvamme E., Schousboe A., Hertz L., Torgner I. A. and Svenneby G. (1985) Developmental change of endogenous glutamate and gamma-glutamyl transferase in cultured cerebral cortical interneurons and cerebellar granule cells, and in mouse cerebral cortex and cerebellum in vivo. Neurochem. Res. 10, 993–1008. Lallement G., Carpentier P., Collet A., Pernot-Marino I., Baubichon D. and Blanchet G. (1991) Effects of soman-induced seizures on different extracellular amino acid levels and on glutamate uptake in rat hippocampus. Brain Res. 563, 234–240. Lehmann A. (1987) Alterations in hippocampal extracellular amino acids and purine catabolites during limbic seizures induced by folate injections into the rabbit amygdala. Neuroscience 22, 573–578. Lehmann A. (1989) Effects of microdialysis-perfusion with anisoosmotic media on extracellular amino acids in the rat hippocampus and skeletal muscle. J. Neurochem. 53, 525–535. Lehmann A., Hagberg H., Jacobson I. and Hamberger A. (1985) Effects of status epilepticus on extracellular amino acids in the hippocampus. Brain Res. 359, 147–151. Lehmann A., Isacsson H. and Hamberger A. (1983) Effects of in vivo administration of kainic acid on the extracellular amino acid pool in the rabbit hippocampus. J. Neurochem. 40, 1314–1320. Li Y., Chopp M., Jiang N. and Zaloga C. (1995) In situ detection of DNA fragmentation after focal cerebral ischemia in mice. Molec. Brain Res. 28, 164–168. Linnik M. D., Zobrist R. H. and Hatfield M. D. (1993) Evidence supporting a role for programmed cell death in focal cerebral ischemia in rats. Stroke 24, 2002–2008. Luna L. G. (1972) Manual of Histologic Staining Methods of the Armed Forces Institute of Pathology, p. 5. McGraw-Hill, New York. MacManus J. P., Buchan A. M., Hill I. E., Rasquinha I. and Preston E. (1993) Global ischemia can cause DNA fragmentation indicative of apoptosis in rat brain. Neurosci. Lett. 164, 89–92.

706 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

D. G. Fujikawa et al. MacManus J. P., Hill I. E., Huang Z. G., Rasquinha I., Xue D. and Buchan A. M. (1994) DNA damage consistent with apoptosis in transient focal ischaemic neocortex. Neuroreport 5, 493–496. Millan M. H., Chapman A. G. and Meldrum B. S. (1993) Extracellular amino acid levels in hippocampus during pilocarpine-induced seizures. Epilepsy Res. 14, 139–148. Morrison P. F., Bungay P. M., Hsiao J. K., Ball B. A., Mefford I. N. and Dedrick R. L. (1991) Quantitative microdialysis: analysis of transients and application to pharmacokinetics in brain. J. Neurochem. 57, 103–109. Nedergaard M. and Hansen A. J. (1988) Spreading depression is not associated with neuronal injury in the normal brain. Brain Res. 449, 395–398. Nevander G., Ingvar M., Auer R. and Siesjo¨ B. K. (1985) Status epilepticus in well-oxygenated rats causes neuronal necrosis. Ann. Neurol. 18, 281–290. Nicholls D. and Attwell D. (1990) The release and uptake of excitatory amino acids. Trends pharmac. Sci. 11, 462–468. Nitatori T., Sato N., Waguri S., Karasawa Y., Araki H., Shibanai K., Kominami E. and Uchiyama Y. (1995) Delayed neuronal death in the CA1 pyramidal cell layer of the gerbil hippocampus following transient ischemia is apoptosis. J. Neurosci. 15, 1001–1011. Obrenovitch T. P., Urenjak J., Richards D. A., Ueda Y., Curzon G. and Symon L. (1993) Extracellular neuroactive amino acids in the rat striatum during ischaemia: comparison between penumbral conditions and ischaemia with sustained anoxic depolarisation. J. Neurochem. 61, 178–186. Okamoto M., Matsumoto M., Ohtsuki T., Taguchi A., Mikoshiba K. and Yanagihara T. (1993) Internucleosomal DNA cleavage involved in ischemia-induced neuronal death. Biochem. biophys. Res. Commun. 196, 1356–1362. Paxinos G. and Watson C. (1986) The Rat Brain in Stereotaxic Coordinates. Academic Press, Sydney. Pollard H., Charriaut-Marlangue C., Cantagrel S., Represa A., Robain O., Moreau J. and Ben-Ari Y. (1994) Kainate-induced apoptotic cell death in hippocampal neurons. Neuroscience 63, 7–18. Ramnath R. R., Strange K. and Rosenberg P. A. (1993) Neuronal injury evoked by depolarizing agents in rat cortical cultures. Neuroscience 51, 931–939. Ronne-Engstro¨m E., Hillered L., Flink R., Spa¨nnare B., Ungerstedt U. and Carlson H. (1992) Intracerebral microdialysis of extracellular amino acids in the human epileptic fous. J. cerebr. Blood Flow Metab. 12, 873–876. Rothman S.M. (1985) The neurotoxicity of excitatory amino acids is produced by passive chloride flux. J. Neurosci. 5, 1483–1489. Rothman S. M., Thurston J. H. and Hauhart R. E. (1987) Delayed neurotoxicity of excitatory amino acids in vitro. Neuroscience 22, 471–480. Sandberg M., Butcher S. P. and Hagberg H. (1986) Extracellular overflow of neuroactive amino acids during severe insulin-induced hypoglycemia—in vivo dialysis of the rat hippocampus. J. Neurochem. 47, 178–184. Schramm M., Eimerl S. and Costa E. (1990) Serum and depolarizing agents cause acute neurotoxicity in cultured cerebellar granule cells: role of the glutamate receptor responsive to N-methyl--aspartate. Proc. natn. Acad. Sci. U.S.A. 87, 1193–1197. Sei Y., Von Lubitz K. J., Basile A. S., Borner M. M., Lin R. C., Skolnick P. and Fossom L. H. (1994) Internucleosomal DNA fragmentation in gerbil hippocampus following forebrain ischemia. Neurosci. Lett. 171, 179–182. Sloviter R. S. and Dempster D. D. (1985) ‘‘Epileptic’’ brain damage is replicated qualitatively in the rat hippocampus by central injection of glutamate or aspartate but not by GABA or acetylcholine. Brain Res. Bull. 15, 39–60. Solı´s J. M., Herranz A. S., Herreras O., Lerma J. and Martı´n del Rio R. (1988) Does taurine act as an osmoregulatory substance in the rat brain? Neurosci. Lett. 91, 53–58. Szerb J. C. (1991) Glutamate release and spreading depression in the fascia dentata in response to microdialysis with high K+: role of glia. Brain Res. 542, 259–265. Takahashi M., Liou S.-Y. and Kunihara M. (1995) Ca2+- and Cl"-dependent, NMDA receptor-mediated neuronal death induced by depolarization in rat hippocampal organotypic cultures. Brain Res. 675, 249–256. Tymianski M., Charlton M. P., Carlen P. L. and Tator C. H. (1993) Source specificity of early calcium neurotoxicity in cultured embryonic spinal neurons. J. Neurosci. 13, 2085–2104. Wade J. V., Olson J. P., Samson F. E., Nelson S. R. and Pazdernik T. L. (1988) A possible role for taurine in osmoregulation within the brain. J. Neurochem. 51, 740–745. Wade J. V., Samson F. E., Nelson S. R. and Pazdernik T. L. (1987) Changes in extracellular amino acids during soman- and kainic acid-induced seizures. J. Neurochem. 49, 645–650. Wyllie A. H. (1981) Cell death: a new classification separating apoptosis from necrosis. In Cell Death in Biology and Pathology (eds Bowen I. D. and Lockshin R. A.), pp. 9–34. Chapman & Hall, London. Wyllie A. H., Kerr J. F. R. and Currie A. R. (1980) Cell death: the significance of apoptosis. Int. Rev. Cytol. 68, 251–306. (Accepted 11 March 1996)