Protection from neuronal damage induced by combined oxygen and glucose deprivation in organotypic hippocampal cultures by glutamate receptor antagonists

BRAIN RESEARCH Brain Research 687 (1995) 167-174

ELSEVIER

Research report

Protection from neuronal damage induced by combined oxygen and glucose deprivation in organotypic hippocampal cultures by glutamate receptor antagonists Uta Strasser, Giinther Fischer * Preclinical Research, PRPN, F. Hoffmann-La Roche Ltd., CH-4002 Basle, Switzerland

Accepted 11 April 1995

Abstract Organotypic hippocampal cultures were exposed to defined periods (30 and 60 min) of combined oxygen and glucose deprivation, mimicking transient ischemic conditions. The involvement of different glutamate receptors in individual hippocampal subfields (CA1, CA3 and dentate gyms) was studied using antagonists of NMDA (dizocilpine) and AMPA/kainate receptors (CNQX and GYKI 52466). Staining with the fluorescent dye propidium iodide (PI) allowed detection of damaged cells. For quantitative determination of neuronal damage, fluorescence intensity was measured after a 22 h recovery period and was related to maximal fluorescence intensity measured after fixation and PI restaining of the cultures at the end of the experiment. Dizocilpine (10 /~M), CNQX (100 /zM) and GYKI 52466 (100 /zM) provided complete protection in CA1, CA3 and dentate gyrus following the moderate ischemic insult, when the antagonists were present permanently. This indicates that none of the ionotropic glutamate receptor subtypes dominated toxicity in the most sensitive subpopulation of neurons. When applied only during the recovery period protection with dizocilpine (10 /xM) or CNQX (100 /~M) was drastically reduced by about 60% in the most sensitive area (CA1), but only slightly by 15% in CA3. Therefore the onset of irreversible damage seems to occur earlier in CA1 than in CA3. Blockade of AMPA/kainate receptors by GYKI 52466 (100 /xM) offered no neuroprotection if the compound was applied only during the recovery period. Combined blockade of NMDA and AMPA/kainate receptors during the recovery period only, did not significantly enhance the protection compared to blockade of either receptor alone. By increasing the period of oxygen and glucose deprivation to 60 min, protection by permanent blockade of NMDA or AMPA/kainate receptors was lost completely in CA1 and was significant in CA3 and dentate gyrus only after NMDA receptor blockade by dizocilpine ( ~ 50% protection). When a combination of dizocilpine and CNQX or GYKI 52466 was applied permanently, significant protection could be shown in all hippocampal areas, being about 65% for dizocilpine plus CNQX and almost complete for dizocilpine plus GYKI 52466. These findings indicate that in vitro combined blockade of NMDA and AMPA/kainate receptors gains importance in protecting neurons from degeneration following prolonged oxygen and glucose deprivation periods. When the receptors are blocked only during recovery, no significant protection was obtained in any hippocampal area, indicating that irreversible neuronal damage, mediated by ionotropic glutamate receptors, is manifested already during the 60 min deprivation period. Keywords: Organotypic hippocampal culture; Oxygen deprivation; Glucose deprivation; Glutamate receptor antagonists; Neuroprotection; Propidium

iodide staining

1. Introduction Cerebral ischemia results in severe neuronal degeneration and as a consequence leads to loss of brain functions. Animal models for transient forebrain ischemia identified the hippocampus as the most susceptible brain area and within the hippocampal formation, pyramidal cells in the

* Corresponding author. Fax: (41) (61) 688-1720. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0006-8993(95)00519-6

CA1 region are most vulnerable to ischemic insults. Elevated concentrations of excitatory amino acids during ischemia [2], protection from neuronal damage by blockade of glutamate release [19] and protective effects of antagonists for different types of glutamate receptors in animal models [5,12,16,23,25,28,33,34,39,40,43] led to the conclusion that ischemic neuronal death is mediated predominantly via glutamate receptor activation (for review see [1]). As the NMDA-subtype of glutamate receptors has its highest density in the most vulnerable hippocampal area CA1 [30], the correlation between sensitivity to ischemia

168

U. Strasser, G. Fischer/Brain Research 687 (1995) 167-174

and NMDA receptor activation has been the subject of many investigations (reviewed in [28]). For the examination of molecular mechanisms underlying ischemic cell death, the complex system of animal models bears some problems in controlling critical parameters such as blood pressure and body temperature. Additionally, application of drugs to the appropriate brain area sometimes is difficult, due to poor permeability through the blood brain barrier, and so potential protective effects might be masked. As an alternative model, neuronal monolayer cell cultures are used to study different aspects of neuronal degeneration in a more simplified system [8,11,14,17,18,27,31,32]. Drugs can easily be applied to these cultures and neuronal damage can be quantified by staining and counting of dead cells. However, this simplicity also includes some disadvantages: selective vulnerability of different brain regions is difficult to compare and intrinsic synaptic connections are missing, which may play an important role in glutamate-mediated neurodegeneration. Organotypic slice cultures offer a good compromise between the complexity of animal models and the simplification of monolayer cultures as tissue specific cell connections and local neuronal circuits with the appropriate innervation patterns are well preserved in this culture system [6,7,9,13,41,45] and the cells are easily accessible for pharmacological treatments. Therefore, organotypic hippocampal cultures have already been used in a variety of studies on neurodegenerative processes, including excitotoxicity [49], hypoglycaemia [44], combined oxygen and glucose deprivation [35] and metabolic inhibition [50]. Besides hypoglycaemia, hypoxia and glutamate toxicity, combined deprivation of glucose and oxygen might mimic most closely ischemic conditions in vivo. In our recent paper, propidium iodide staining combined with fluorescence intensity measurements was used to quantify ischemia-induced cell damage [42]. Propidium iodide is a good marker for damaged cells, as the dye can only enter cells with leaky membranes [26]. Since, in this ischemia model only neurons are dying, the measured fluorescence intensity values can be taken to determine percentage neuronal damage [42]. In the present study we examined the involvement of NMDA and AMPA/kainate receptors in ischemia-induced neuronal degeneration in vitro. Furthermore we studied the relative sensitivity of individual hippocampal subfields to ischemia as well as the time at which irreversible damage occurs after onset of ischemia.

2. Material and methods 2.1. Organotypic hippocampal cultures

Cultures were prepared as described previously [41]. Hippocampi of 3 day old rats were rapidly removed under sterile conditions and cut into 3 5 0 / z m slices with a tissue chopper. Slices were separated and transposed to ice-cold

preparation buffer with a small spatula. Preparation buffer (pH 7.3) was composed of Minimum Essential Medium (MEM, Gibco, UK) with HEPES (25 mM) and L-glutamine (4 mM). Millicell-CM filters (Millipore, USA) in 24-well-plates were pre-equilibrated in 0.3 ml of culture medium in a moist 5% CO 2 atmosphere at 37°C for 1 h. Culture medium contained MEM (50%), HEPES (25 mM), L-glutamine (4 mM), NaHCO 3 (0.013%, Fluka, Switzerland), HBSS (25%, Gibco, UK) and heat-inactivated horse serum (25%, Sigma, USA) and was titrated to pH 7.3. Each slice was put on a single filter and cultured for two weeks with a total change of medium every second day. During the second week of culture, 10 mM MgCI 2 was added to the medium to reduce spontaneously occurring cell death. Culturing the slices at the interface between medium and gas atmosphere makes them flatten to about 3 cell layers during 2 weeks of culture. 2.2. Combined oxygen and glucose deprivation

After a culture period of 15 days, control cultures were once washed with a controlled salt solution (CSS), pH 7.2, containing NaC1 120 mM, KCI 5.4 mM, MgCI 2 x 6 H 2 0 0.8 mM, CaC12×2 H 2 0 1.8 mM, Tris 25 mM and glucose 15 mM (all from Fluka, Switzerland) and then incubated in CSS at 37°C for 30 or 60 min in a room air incubator. For ischemic conditions, slice cultures were once washed in CSS where glucose was replaced with sorbitol (glucose-free CSS) and 35 mM NaHCO 3 was added, to buffer the CO 2 content within the anaerobic system. The glucose-free CSS was bubbled with a nitrogen gas mixture (5% CO2, 95% N 2) to get it oxygen-free. The slices were then incubated in oxygen-and glucose-free CSS in an oxygen-free atmosphere (5% CO 2, 10% H 2 and 85% N 2) at 37°C (Anaerobic system 1029, Forma Scientific) for 30 or 60 min. After the deprivation period, the slices were maintained in normal fresh culture medium in the incubator for a recovery time of 22 h. 2.3. Propidium iodide staining

Propidium iodide (15 /xg/ml, Sigma USA) was added throughout the deprivation and the recovery period to stain damaged cells. It has been described that astrocytes in hippocampal slices, cultured on Millicell filters, redistributed into a thin layer at the bottom of the slice, forming a kind of barrier between the tissue and the medium [6]. To ensure sufficient diffusion of propidium iodide and added glutamate receptor antagonists, all solutions during the whole experiment were added below the Millicell-filter as well as on top of the culture, 0.2 ml each. After the recovery period of 22 h, the slices were fixed for 15 min in 4% paraformaldehyde (in CSS). The fixative was washed away with CSS and the cultures were restained with propidium iodide (15 /zg/ml in CSS) for 2 h.

U. Strasser, G. Fischer~Brain Research 687 (1995) 167-174 Table 1 Protection from 30 min of combined oxygen and glucose deprivation by permanent application of glutamate receptor antagonists Permanent application of

Conc. mM

Percent protection in CA1 CA3

Dizocilpine CNQX GYKI 52466

0.01 0.10 0.10

1 0 8 ± 4 *** (9) 96_+3 *** (9) 96+_4 ** * (9)

1 1 3 ± 6 * * (5) 108_+6 * * * (8) 1 0 8 ± 7 * * * (9)

Protection from neuronal damage following 30 min of combined oxygen and glucose deprivation by application of glutamate receptor antagonists during the deprivation and the recovery period. Total protection was achieved in CA1 and CA3 by dizocilpine as well as by the AMPA/kainate receptor antagonists CNQX and GYKI 52466. Values show average ± S.E.M., the number of slices, which were analysed in minimum of two independent experiments, is given in parentheses. Significance was determined by Student's t-test and significant protection is marked by , ( P < 0.05), , , ( P < 0.005) and , , , ( P < 0.0005).

2.4. Measurement of propidium iodide fluorescence intensity Fluorescence intensity was measured as described [42]. After 22 h of recovery, the propidium iodide fluorescence images of the whole slice cultures were recorded with an intensified CCD camera (Hamamatsu C 2400-87) at a 4 × magnification (Axiovert 405M, Zeiss, Germany). The orientation and relative position of the slices on the microscopic stage was kept constant. Measurement of fluorescence intensity was performed with a Hamamatsu image analysis system (ICMS). A camera set-up was chosen to get signals of low fluorescence intensity in control cultures, without reaching the saturation when measuring the maximal fluorescence in fixed cultures. This camera setting was kept constant during the image acquisition. The video images of the fluorescence staining were saved by the image analysis system for later quantification.

2.5. Quantification of fluorescence measurements Every experiment was performed with three slices. In every slice three rectangular measurement windows were placed in each hippocampal area (CA1, CA3 and dentate gyrus). For location of the measurement windows, the fluorescence picture of the fixed slice was used, where demarcations of the respective hippocampal areas were clearly visible. The size of the measurement windows was chosen in a way that the complete frame was located within the respective hippocampal area. The relative position of the measurement windows was kept constant and as the position of the cultures was always the same when images were recorded, this set of measurement windows could be used in the picture after 22 h of recovery as well as in the image of the fixed slice. The intensity value was obtained by integrating the amount of pixels over the area of the measurement window. Fluorescence intensities in the measurement windows in the fixed slices were set as 100% cell damage. Each value, measured after 22 h of

169

recovery, was calculated as percentage of the maximal fluorescence intensity in the respective window of the fixed slice, to obtain the percentage of cell damage. Average percentage values of the three measurement windows in each area were calculated for every slice. For standard±sat±on of protection, mean values of damage in unprotected ischemic cultures were taken as 0% and values of control cultures as 100% protection. These reference conditions were included in every experiment. Values given in the tables (Table 1,Table 2,Table 3,Table 4) show mean protection compared to ischemic conditions _ S.E.M. of slices analysed in at least two independent experiments. Significance was determined by using Student's t-test with *P < 0.05, * *P < 0.005 and * * *P < 0.0005.

2.6. Protection by glutamate receptor antagonists To yield concentration-response curves, permanent application of different concentrations of dizocilpine, CNQX and GYKI 52466 was tested. The percentage of protection from cell damage following 30 min of oxygen and glucose deprivation was determined for the most sensitive area CA1. For protection analysis glutamate receptor antagonists were added either during the deprivation and the recovery period or only during the recovery period in a concentration which gave maximal protection. Antagonists were added in the following concentrations: dizocilpine 10 /xM, CNQX 100 /zM and GYKI 52466 100 /xM.

3. Results

3.1. Neurodegeneration following oxygen and glucose deprivation Analysing the cell damage in different hippocampal areas following the 30 min deprivation period showed that Table 2 Protection from 30 min of combined oxygen and glucose deprivation by application of glutamate receptor antagonists during the recovery period Antagonists applied during recovery

Conc. mM

Percent protection in CA1 CA3

Dizocilpine CNQX GYKI 52466 Dizocilpine + CNQX Dizocilpine + GYK152466

0.01 0.10 0.10

42+_6 * * (6) 3 4 ± 6 ** (7) 36 _+ 13 (9) 33 ± 6 * (6) 51 + 12 (16)

8 7 ± 7 * (6) 8 4 ± 9 * (7) - 13 ± 26 (10) 79 ± 15 * (6) 59 ± 14 (14)

Protection from cell damage following 30 min of combined oxygen and glucose deprivation by application of glutamate receptor antagonists only during the recovery period. Compared to permanent drug application, the protective effect is significantly reduced in both areas. Combined application of dizocilpine and the AMPA/kainate receptor antagonists CNQX and GYKI 52466 did not improve the protection by single drug treatment. Values show average ± S.E.M., the number of slices, which were analysed in minimum of two independent experiments, is given in parentheses. Significance was determined by Student's t-test and significant protection was marked by , ( P < 0.05) and , , ( P < 0.005).

U. Strasser, G. Fischer~Brain Research 687 (1995) 167-174

170

Table 3 Protection from 60 min of combined oxygen and glucose deprivation by permanent application of glutamate receptor antagonists Permanent application of

Dizocilpine CNQX GYKI Dizocilpine + CNQX Dizocilpine + GYKI 52466

Conc. mM

Percent protection in

0.01 0.10 0.10

CA1

CA3

Dentate gyrus

12 ± 7 (5) 1 ± 7 (6) 15 ± 7 (6) 62 ± 23 * * (6) 92 ± 8 * * * (6)

46 ± 12 ' (6) 28 ± 15 (6) - 4 ± 7 (6) 67 ± 12 * ' (6) 98 ± 4 * * * (6)

56 ± 12 * * (6) 13 ± 14 (6) 0 ± 12 (6) 66 ± l l * * (6) 96 ± 1 * * * (6)

Protection from cell damage following 60 min of combined oxygen and glucose deprivation by application of glutamate receptor antagonists during the deprivation and the recovery period. With single drug treatment, only dizocilpine showed significant protection in CA3 and dentate gyrus. Combined treatment with CNQX or GYKI 52466 greatly increased protection in all hippocampal areas. Values show average + S.E.M., the number of slices, analysed in minimum of two independent experiments, is given in parantheses. Significance was determined by Student's t-test and significant protection was marked by , ( P < 0.05), , , ( P < 0.005) and , , , ( P < 0.0005).

the highest percentage of cells, namely 23%, died in CA1, followed by 13% in CA3 and only 7% in dentate gyrus. After the prolonged deprivation period of 60 min, 66% of cell damage was measured in CA1 and 51% in CA3 as well as in dentate gyrus, as shown recently [42]. Immunohistochemical staining for MAP-2 and GFAP on paraffin sections of organotypic hippocampal cultures revealed, that no neurons survive the 60 min deprivation of oxygen and glucose whereas astrocytes remained unaffected [42]. Therefore the percentage of cell damage following 60 min of combined oxygen and glucose deprivation can be regarded as 100% neuronal damage.

3.2. Concentration-dependent protection by glutamate receptor antagonists To determine the maximal protective concentration of the glutamate receptor antagonists used, concentration-response relationships were measured in the most sensitive CA1 area with permanent application of drugs during the 30 min oxygen and glucose deprivation and the subsequent 22 h recovery period. All three drugs, dizocilpine, CNQX and GYKI 52466 showed concentration-dependent protection (Fig. 1). Full protection was obtained with 1 /zM dizocilpine, 30 /.~M CNQX and about 100 /zM GYKI 52466. For further studies with single drug or combination

treatments, 10 /xM dizocilpine, 100 /xM CNQX and 100 /.tM Gyki 52466 were used to ensure maximal protection.

3.3. Short-term (30 min) oxygen and glucose deprivation Because only a small percentage of cells died in the dentate gyrus following 30 min of combined oxygen and glucose deprivation, the protective effects of glutamate receptor antagonists were only quantified in the hippocampal areas CA1 and CA3. Following the short deprivation period, complete protection was achieved in both areas with all three glutamate receptor antagonist when applied permanently, namely during the deprivation and the recovery period (Table 1). Combinations of either dizocilpine plus CNQX or dizocilpine plus GYKI 52466 were as protective as single drug treatments. Application of the antagonists only during the recovery markedly influenced their protective effect. GYKI 52466 lost its protective effect in both areas, whereas dizocilpine and CNQX were still highly effective in CA3 with only slightly reduced protection to about 85%. Their efficacy in CA1 was still significant but drastically reduced to only about 40% protection (Table 2). The combined treatment with dizocilpine and CNQX during recovery did not improve neuroprotection over single drug treatment. A combined application of dizocilpine and GYKI 52466 did not result

Table 4 Protection from 60 min of combined oxygen and glucose deprivation by application of glutamate receptor antagonists during the recovery period Application of antagonists during recovery

Dizociipine CNQX GYKI 52466 Dizocilpine + CNQX Dizocilpine + GYKI 52466

Conc. mM

0.01 0.10 0.10

Percent protection in CA1

CA3

Dentate gyrus

9 + 7 (6) 8 5:6 (6) 12 + 8 (3) - 6 5:6 (6) 12 5:8 (5)

36 + 13 (6) 17 + 12 (6) 23 + 5 (5) 25 + 14 (6) 4 + 20 (6)

14 5:9 (6) 6 + 11 (6) 24 5:4 (4) 1 + 11 (6) - 13 + 19 (3)

Protection from cell damage following 60 min of combined oxygen and glucose deprivation by application of glutamate receptor antagonists only during the deprivation period. No significant protection could be measured. Even combined application of dizocilpine with CNQX or GYKI 52466 did not result in significant protection in any hippocampal area. Values show average -t-S.E.M., the number of slices, which were analysed in minimum of two independent experiments, is given in parentheses. Significance was determined by Student's t-test.

U. Strasser, G. Fischer/Brain Research 687 (1995) 167-174 120 100

o .,=

80

o e~

60

Dizocilpine 9

T ~

6

9

I

I0

9

3

40 20 .0001 .001

.01

.I

I00

~M 120

CNQX

1 O0

o

6•6

80 60 40 20 0

•

' ' " " 1

"

"

' ' " ' 1

10

"

"

" ' ' • ' 1

1 O0

1000

laM 120 = o •

GYKI 52466

(Table 3). Combined treatment with dizocilpine and either CNQX or GYKI 52466, however, still revealed significant protection in all three areas. Whereas the combination of dizocilpine and CNQX resulted in 62%, 67% and 66% protection, the combination with GYKI 52466 gave 92%, 98% and 96% protection in CA1, CA3 and dentate gyrus, respectively (Table 3). In CA1 both combinations showed a significant increase in neuroprotection compared to single drug treatment. In CA3 and dentate gyrus, the difference between application of dizocilpine alone or in combination with CNQX was not statistically significant. Combination with GYKI 52466, however, increased the protective effect significantly. The protective effect of dizocilpine plus GYKI 52466 was significantly higher than that of dizocilpine plus CNQX in CA3 and dentate gyrus. In CA1 however, the difference between the two combinations was not statistically significant. When the glutamate receptor antagonists were applied only during the recovery period following 60 min of combined oxygen and glucose deprivation, no significant protection could be measured in all hippocampal areas (Table 4) and even the combination of dizocilpine with CNQX or GYKI 52466 did not result in significant protection.

4. Discussion

I O0

3

80 60

e~ ~.

171

40 20 0

•

• •

H•

.

10

100

•

•

,•n|

1000

gM Fig. 1. Concentration-responsecurvesfor dizocilpine,CNOXand GYKI 52466. Protectionin the CA1 area of organotypichippocampalculturesis shown 5:S.E.M., the numbersof evaluatedslice cultures is given below the symbols. in significant protection, due to a high degree of variance in these results. 3.4. Long-term (60 min) oxygen and glucose deprivation

To increase the severity of ischemic conditions, the period of oxygen and glucose deprivation was prolonged to 60 min. Permanent treatment with single drugs did not result in significant protection in the most sensitive CA1 area. However, dizocilpine still showed 46% and 56% protection in CA3 and dentate gyrus, respectively, whereas the AMPA/kainate receptor antagonists CNQX and GYKI 52466 did not protect neurons in any hippocampal area

During cerebral ischemia extracellular concentrations of excitatory amino acids increase [2] and it is commonly agreed that neurodegeneration induced by ischemic insults is mediated predominantly by glutamate receptor activation (for review see [1]). Hippocampal slice cultures were used recently to study neuronal degeneration in a descriptive way after various stimuli including glutamate toxicity [48], hypoglycaemia, [44], metabolic inhibition [50] and combined oxygen and glucose deprivation [35,47]. In our previous paper we described a method to analyse neuronal damage quantitatively by staining damaged cells with propidium iodide and measuring fluorescence intensity [42]. This method was used to investigate the role of ionotropic glutamate receptors in ischemia induced neuronal damage in organotypic hippocampal cultures. Neurodegeneration induced by transient oxygen and glucose deprivation in different areas of organotypic hippocampal cultures is triggered, at least in part, by the activation of NMDA and AMPA/kainate receptors. A detailed analysis revealed, however, that dependent on the duration of the insult and the onset of receptor blockade by different antagonists, neuroprotection varies between different hippocampal areas. Neuronal damage induced by 30 min of combined oxygen and glucose deprivation was found to be mediated by activation of AMPA/kainate receptors as well as NMDA receptors. As shown recently, only subpopulations of neurons in different hippocampal areas degenerate un-

172

U. Strasser, G. Fischer/Brain Research 687 (1995) 167-174

der these moderate transient ischemic conditions, resulting in about 35%, 25% and 14% neuronal damage in CA1, CA3 and dentate gyrus, respectively [42]. Because of the overall small percentage of neurodegeneration in dentate gyrus under these moderate ischemic conditions, protective effects of glutamate receptor antagonists were only analysed in CA1 and CA3. When given throughout the deprivation and the recovery period, the NMDA receptor antagonist dizocilpine completely protected these sensitive neuronal subpopulations in CA1 and CA3 as did the AMPA/kainate receptor antagonists CNQX and GYKI 52466 (Table 1). Maximal protection with CNQX was only obtained at high doses ( > 30 /~M), which besides AMPA/kainate receptor blockade partially block NMDA receptors [24]. As the selective AMPA/kainate receptor antagonist GYKI 52466 also showed complete protection, the additional NMDA receptor blockade by CNQX seems not to be necessary for its protective effect under these conditions. The protective effect of the NMDA receptor antagonist dizocilpine has already been described in studies mimicking ischemic conditions by oxygen and glucose deprivation [35] or metabolic inhibition in organotypic hippocampal cultures [50]. The influence of AMPA/kainate receptor blockade, however, has not been investigated in these cultures so far. Our results indicate that under conditions of a moderate ischemic insult, both types of ionotropic glutamate receptors contribute to, but neither of them dominates, toxicity in organotypic hippocampal cultures. This is in contrast to findings with monolayer cell cultures of either cortical or hippocampal neurons, where NMDA receptors have to be blocked to reveal complete protection against short term oxygen and glucose deprivation [21,22,29]. When the deprivation period was prolonged from 30 to 60 min, all neurons died in all hippocampal areas [42], and no protection could be achieved by permanent treatment with either CNQX or GYKI 52466. Dizocilpine was not protective in CA1 but still partially protective in CA3 (46%) and dentate gyrus (56%) indicating that NMDA receptor activation dominates cell damage only in subpopulations of neurons in CA3 and dentate gyrus. Protection was significantly increased in all areas by combined drug treatment reaching about 65% with dizocilpine plus CNQX and almost complete protection with dizocilpine plus GYKI 52466 (Table 3). Therefore it became mandatory to block both, NMDA and AMPA/kainate receptors to achieve protection in CA1 and for subpopulations of neurons in CA3 and dentate gyrus. The underlying mechanism of the higher efficacy of GYKI 52466 over CNQX in combination with dizocilpine is unclear. One explanation might be that CNQX is a competitive antagonist, whereas GYKI 52466 acts non-competitive at the AMPA/kainate receptor. High glutamate concentrations in the cultures during the ischemic insult might therefore reduce the protective effect of CNQX but not influence the effects of GYKI

52466. GYKI 52466 might also have a different spectrum of selectivity to subtypes of AMPA/kainate receptors or have additional, so far unknown properties. The improved protection obtained by a combination of NMDA and AMPA receptor blockers is in agreement with prolonged oxygen and glucose deprivation experiments in monolayer cell cultures where NMDA receptor blockade is mandatory but not sufficient in drug combination studies to reveal neuroprotection [21,22,29]. To address the time window for protective intervention with ionotropic glutamate receptor antagonists the drugs were only given throughout the recovery period. Following the 30 min deprivation period GYKI 52466 failed to show significant protection whereas dizocilpine and CNQX showed reduced but still significant protection in CA1 being about 42% and 34%, respectively (Table 2). In contrast, most of the neurons (about 85%) could be protected in CA3 with dizocilpine as well as with CNQX. The higher protective potential of dizocilpine and CNQX in CA3 compared to CA1 might indicate that the time course for irreversible manifestation of neurotoxicity is different in these hippocampal areas. Either the trigger signal develops later in CA3 or the trigger is weaker and thus needs more time to induce irreversible cell damage. The failure of GYKI 52466 in comparison to CNQX to show significant protection even in the less sensitive area CA3 might indicate that the additional blockade of NMDA receptors at the high concentration of CNQX becomes important for neuroprotection with delayed application. The reason for the higher variability in the experiments with GYKI 52466, which may mask some protective effects, is not clear. Combined drug treatment only during the recovery period after 30 min of oxygen and glucose deprivation could not improve protection indicating that mechanisms independent from activation of ionotropic glutamate receptors and/or downstream events gain importance with time for neuronal degeneration. Application of the glutamate receptor antagonists only during the recovery period following 60 min of combined oxygen and glucose deprivation could not rescue neurons from damage in any hippocampal area, even when a combined treatment was performed (Table 4). Obviously cell damage already becomes irreversible during the 60 min deprivation period, concerning blockade of ionotropic glutamate receptors. In agreement with our data, a similar decrease in protection with increasing time delay of NMDA receptor blockade after a moderate oxygen and glucose deprivation was described in monolayer cell cultures [18]. A comparison of these findings with results from mild to moderate transient global ischemia models in vivo reveals some marked differences. Although in these animal models as well as in the organotypic cultures selective neuronal damage occurs, the time course of neuronal degeneration is delayed only in the animal models. Neurodegeneration in different brain areas in in vivo models critically depends on the duration of the ischemic period,

U. Strasser, G. Fischer/Brain Research 687 (1995) 167-174

which is often selected in a w a y that the most susceptible pyramidal n e u r o n s in the CA1 area are preferentially affected by delayed n e u r o d e g e n e r a t i o n that occurs after 2 to 3 days [38]. In these a n i m a l m o d e l s of transient global forebrain ischemia, dizocilpine showed no protective effect in the CA1 area [3,10,34,39], whereas protection could be achieved by AMPA/kainate receptor antagonists [4,5,12,23,25,34] even with a 24 h delayed application [39], indicating a different time course and m e c h a n i s m of the d e v e l o p m e n t of the trigger signals leading to neurodegeneration. This is supported b y a similar long therapeutic time w i n d o w for a N-type calcium c h a n n e l blocker in transient global forebrain ischemia in rats [46]. A synergistic protective effect of a c o m b i n e d blockade of N M D A and A M P A / k a i n a t e receptors in transient global ischemia m o d e l s is under debate and m a y critically d e p e n d on the respective model. W h e r e a s in gerbils a c o m b i n a t i o n of dizocilpine and the less selective glutamate receptor antagonist k y n u r e n i c acid i m p r o v e d protection [16], only N B Q X but not a c o m b i n a t i o n of N B Q X with dizocilpine was found to be protective in a more severe transient global ischemia model (four vessel occlusion) or a transient focal ischemia model in rats [4,51]. In focal p e r m a n e n t or transient ischemia the time course of n e u r o d e g e n e r a t i o n is faster and the therapeutic time w i n d o w is considerably shorter, indicating a different m e c h a n i s m of neurodegeneration. This is supported by findings that N M D A as well as A M P A / k a i n a t e receptor antagonists are protective [3,15,20,36,37]. The therapeutic time w i n d o w is in the range of 1 to 2 h [20,36]. T a k e n together, the analysis of n e u r o n a l degeneration in the organotypic cultures in different hippocampal areas allowed a m u c h more detailed analysis of the relative contribution of ionotropic glutamate receptors to ischemic d a m a g e of n e u r o n s than the use of m o n o l a y e r cell cultures. Especially the important role of A M P A / k a i n a t e receptors b e c a m e obvious. They are not only a target of second choice for neuroprotective treatment but for n e u r o n a l subpopulations as important as N M D A receptors at least after moderate insults. O n l y after prolonged and thus more severe challenge N M D A receptor blockade b e c o m e s mandatory to achieve protection.

Acknowledgements W e gratefully a c k n o w l e d g e support by Prof. Dr. C. K6hler and his critical reading of the manuscript.

References [1] Benveniste, H., The excitotoxin hypothesis in relation to cerebral ischemia, Cerebrovasc. Brain Metab. Rev., 3 (1991) 213-245. [2] Benveniste, H., Drejer, J., Schousboe, A. and Diemer, N.H., Eleva-

173

tion of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis, J. Neurochem., 43 (1984) 1369-1374. [3] Buchan, A., Li, H. and Pulsinelli, W.A., The N-methyl-D-aspartate antagonist MK-801, fails to protect against neuronal damage caused by transient severe forebrain ischemia in adult rats, J. Neurosci., 1 I (1991) 1044-1056. [4] Buchan, A.M., Lesiuk, H., Barnes, K.A., Li, H., Huang, Z.-G., Smith, K.E. and Zue, D., AMPA antagonists: do they hold morc promise for clinical stroke trials than NMDA antagonists?, Stroke Suppl., 24 (1993) 1-148-I-154. [5] Buchan, A.M., Li, H., Cho, S. and Pulsinelli, W.A., Blockade of thc AMPA receptor prevents CA1 hippocampal injury following severe but transient forebrain ischemia in adult rats, Neurosci Lett., 132 (1991) 255-258. [6] Buchs, P.-A., Stoppini, L. and Muller, D., Structural modifications associated with synaptic development in area CA1 of rat hippocampal organotypic cultures, Dev. Brain Res., 71 (1993) 81-91. [7] Caeser, M. and Aertsen, A., Morpholigical organization of rat hippocampal slice cultures, J. Comp. Neurol., 307 (1991) 87-106. [8] Choi, D.W., Maulucci-Gedde, M. and Kriegstein, A.R., Glutamate neurotoxicity in cortical cell culture, J. Neurosci., 7 (1987) 357-368. [9] del Rio, J.A., Heimrich, B., Soriano, E., Schwegler, H. and Frotscher, M., Proliferation and differentiation of glial fibrillary acidic proteinimmunoreactive glial cells in organotypic slice cultures of rat hippocampus, Neuroscience, 43 (1991) 335-347. [10] Diemer, N.H., Jorgensen, M.B., Johansen, F.F., Sheardown, M. and Honore, T., Protection against ischemic hippocampal CA1 damage in the rat with a new non-NMDA antagonist, NBQX, Acta NeuroL Scand., 86 (1992) 45-49. [11] Frandsen, A., Drejer, J. and Schousboe, A., Direct evidence that excitotoxicity in cultured neurons is mediated via N-methyl-Daspartate (NMDA) as well as non-NMDA receptors, J. Neurochem., 53 (1989) 297-299. [12] Frank, L., Bruhn, T. and Diemer, N.H., The effect of an AMPA antagonist (NBQX) on postischemic neuron loss and protein synthesis in the rat brain, Exp. Brain Res., 95 (1993) 70-76. [13] G~ihwiler, B.H., Organotypic cultures of neural tissue, Trends Neurosci., 11 (1988) 484-489. [14] Giffard, R.G., Monyer, H., Christine, C.W. and Choi, D.W., Acidosis reduces NMDA receptor activation, glutamate ncurotoxicity, and oxygen-glucose deprivation induced neuronal injury in cortical cultures, Brain Res., 506 (1990) 339-342. [15] Gill, R., Nordholm, L. and Lodge, D., The neuroprotective actions of 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(F)quinoxaline (NBQX) in a rat focal ischaemia model, Brain Res., 580 (1992) 35-43. [16] Gill, R. and Woodruff, G.N., The neuroprotective actions of kynurenic acid and MK801 in gerbils are synergistic and not related to hypothermia, Eur. J. Pharmacol., 176 (1990) 143-149. [17] Goldberg, M.P. and Choi, D.W., Intracellular free calcium increases in cultured cortical neurons deprived of oxygen and glucose, Stroke SuppL, 21 (1990) I11-75-III-77. [18] Gotdberg, M.P. and Choi, D.W., Combined oxygen and glucose deprivation in cortical cell culture: calcium-dependent and calciumindependent mechanisms of neuronal injury, J. Neurosci., 13 (1993) 3510-3524. [19] Graham, S.H., Chen, J., Sharp, F.R. and Simon, R.P., Limiting ischemic injury by inhibition of excitatory amino acid release, J. Cereb. Blood Flow. Metab., 13 (1993) 88-97. [20] Hatfield, R.H., Gill, R. and Brazell, C., The dose-response relationship and therapeutic window for dizocilpine (MK-801) in a rat focal ischaemia model, Eur. J. Pharmacol., 216 (1992) 1-7. [21] Kaku, D.A., Giffard, R.G. and Choi, D.W., Neuroprotective effects of glutamate antagonists and extracellular acidity, Science, 260 (1993) 1516-1518. [22] Kaku, D.A., Goldberg, M.P. and Choi, D.W., Antagonism of nonNMDA receptors augments the neuroprotective effect of NMDA

174

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

[35]

[36]

[37]

U. Strasser, G. Fischer/Brain Research 687 (1995) 167-174

receptor blockade in cortical cultures subjected to prolonged deprivation of oxygen and glucose, Brain Res., 554 (1991) 344-347. Le Peillet, E., Arvin, B., Moncada, C. and Meldrum, B.S., The non-NMDA antagonists, NBQX and GYKI 52466, protect against cortical and striatal cell loss following transient global ischaemia in the rat, Brain Res., 571 (1992) 115-120. Lester, R.A., Quarum, M.L., Parker, J.D., Weber, E. and Jahr, C.E., Interaction of 6-cyano-7-nitroquinoxaline-2,3-dione with the Nmethyl-~aspartate receptor-associated glycine binding site, Mol. Pharmacol., 35 (1989) 565-570. Li, H. and Buchan, A.M., Treatment with an AMPA antagonist 12 hours following severe normothermic forebrain ischemia prevents CA1 neuronal injury, J. Cereb. Blood Flow Metab., 13 (1993) 933-939. Macklis, J.D. and Madison, R.D., Progressive incorporation of propidium iodide in cultured mouse neurons correlates with declining electrophysiological status: a fluorescence scale of membrane integrity, J. Neurosci. Methods, 31 (1990) 43-46. Mattson, M.P. and Kater, S.B., Development and selective neurodegeneration in cell cultures from different hippocampal regions, Brain Res., 490 (1989) 110-125. Meldrum, B., Protection against ischaemic neuronal damage by drugs acting on excitatory neurotransmission, Cereb. Brain Metab. Rev., 2 (1990) 27-57. M6ckel, V. and Fischer, G., Vulnerability to excitotoxic stimuli of cultured rat hippocampal neurons containing the calcium-binding proteins calretinin and calbindin Dzsk, Brain Res., 648 (1994) 109-120. Monaghan, D.T. and Cotman, C.W., Distribution of N-methyl-Daspartate-sensitive L-(3H) glutamate-binding sites in rat brain, J. Neurosci., 5 (1985) 2909-2919. Monyer, H., Giffard, R.G., Hartley, D.M., Dugan, L.L., Goldberg, M.P. and Choi, D.W., Oxygen or glucose deprivation-induced neuronal injury in cortical cell cultures is reduced by tetanus toxin, Neuron, 8 (1992) 967-973. Monyer, H., Goldberg, M.P. and Choi, D.W., Glucose deprivation neuronal injury in cortical culture, Brain Res., 483 (1989) 347-354. Nellgard, B. and Wieloch, T., Cerebral protection by AMPA-and NMDA-receptor antagonists administered after severe insulin-induced hypoglycemia, Exp. Brain Res., 92 (1992) 259-266. Nellgard, B. and Wieloch, T., Postischemic blockade of AMPA but not NMDA receptors mitigates neuronal damage in the rat brain following transient severe cerebral ischemia, Jr. Cereb. Blood Flow Metab., 12 (1992) 2-11. Newell, D.W., Malouf, A.T. and Franck, J.E., Glutamate-mediated selective vulnerability to ischemia is present in organotypic cultures of hippocampus, Neurosci. Lett., 116 (1990) 325-330. Park, C.K., Nehls, D.G., Graham, D.I., Teasdale, G.M. and McCulloch, J., Focal cerebral ischaemia in the cat: treatment with the glutamate antagonist MK-801 after induction of ischaemia, J. Cereb. Blood Flow Metab., 8 (1988) 757-762. Pulsinelli, W., Sarokin, A. and Buchan, A., Antagonism of the NMDA and non-NMDA receptors in global versus focal brain ischemia. In K. Kogure, K.-A. Hossmann and B.K. Siesj6 (Eds.),

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

Neurobiology of lschemic Brain Damage, Progress in Brain Research, Vol. 96, Elsevier, Amsterdam 1993, pp. 125-135. Pulsinelli, W.A., Brierley, J.B. and Plum, F., Temporal profile of neuronal damage in a model of transient forebrain ischemia, Ann. Neurol., 11 (1982) 491-498. Sheardown, M.J., Suzdak, P.D. and Nordholm, L., AMPA, but not NMDA, receptor antagonism is neuroprotective in gerbil global ischaemia, even when delayed 24 h, Eur. J. Pharmacol., 236 (1993) 347-353. Smith, S.E. and Meldrum, B.S., Cerebroprotective effect of a nonN-methyl-D-aspartate antagonist, GYKI 52466, after focal ischemia in the rat, Stroke, 23 (1992) 861-864. Stoppini, L., Buchs, P.-A. and Muller, D., A simple method for organotypic cultures of nervous tissue, J. Neurosci. Methods, 37 (1991) 173-182. Strasser, U. and Fischer, G., Quantitative measurement of neuronal degeneration in organotypic hippocampal cultures after combined oxygen/glucose deprivation, J. Neurosci. Methods, 57 (1995) 177186. Swan, J.H. and Meldrum, B.S., Protection by NMDA antagonists against selective cell loss following transient ischaemia, J. Cereb. Blood Flow Metab., 10 (1990) 343-351. Tasker, R.C., Coyle, J.T. and Vornov, J.J., The regional vulnerability to hypoglycemia-induced neurotoxicity in organotypic hippocampal culture: protection by early tetrodotoxin or delayed MK-801, J. Neurosci., 12 (1992) 4298-4308. Torp, R., Haug, F.M., Tonder, N., Zimmer, 1. and Ottersen, O.P., Neuroactive amino acids in organotypic slice cultures of the rat hippocampus: an immunocytochemical study of the distribution of GABA, Glutamate, Glutamine and Taurine, Neuroscience, 46 (1992) 807-823. Valentino, K., Newcomb, R., Gadbois, T., Singh, T., Bowersox, S., Bitner, S., Justice, A., Yamashiro, D., Hoffman, B.B., Ciaranello, R., Miljanich, G. and Ramachandran, J., A selective N-type calcium channel antagonist protects against neuronal loss after global cerebral ischemia, Proc. Natl. Acad. Sci. USA, 90 (1993) 7894-7897. Vornov, J.J. and Coyle, J.T., Enhancement of NMDA receptormediated neurotoxicity in the hippocampal slice by depolarisation and ischemia, Brain Res., 555 (1991) 99-106. Vornov, J.J. and Coyle, J.T., Glutamate neurotoxicity and the inhibition of protein synthesis in the hippocampal slice, J. Neurochem., 56 (1991) 996-1006. Vornov, J.J., Tasker, R.C. and Coyle, J.T., Direct observation of the agonist-specific regional vulnerability to glutamate, NMDA, and kainate neurotoxicity in organotypic hippocampal cultures, Exp. Neurol., 114 (1991) 11-22. Vornov, J.J., Tasker, R.C. and Coyle, J.T., Delayed protection by MK-801 and tetrodotoxin in a rat organotypic hippocampal culture model of ischemia, Stroke, 25 (1994) 457-464. Xue, D., Huang, Z.-G., Barnes, K., Lesiuk, H.J., Smith, K.E. and Buchan, A.M., Delayed treatment with AMPA, but not NMDA, antagonists reduces neocortical infarction, J. Cereb. Blood Flow Metab., 14 (1994) 251-261.