AIDA, a Class I Metabotropic Glutamate-receptor Antagonist Limits Kainate-induced Hippocampal Dysfunction

Authors


Address correspondence and reprint requests to Dr. L. Carmant at Hôpital Ste-Justine, Centre de recherche, 3175 Côte Ste-Catherine, Montreal, Quebec, Canada, H3T 1C5. E-mail: lionel.carmant@umontreal.ca

Abstract

Summary:  Purpose: In the developing animal, intraperitoneal injections of kainic acid (KA) lead to a prolonged initial seizure followed by chronic recurrent seizures and long-term hippocampal dysfunction. We investigated whether the class I metabotropic glutamate receptor (mGluR) antagonist 1-aminoindan-1,5-dicarboxylic acid (AIDA) is neuroprotective in the KA model of epilepsy.

Methods: Immature rats aged postnatal day 20 (P20) and P30 were injected with fixed volumes of KA, KA + AIDA, AIDA, or saline. We monitored recurrent seizures. Thirty days later, we tested hippocampal function with the Morris water-maze test or prepared hippocampal slices to record extracellularly evoked and spontaneous potentials from the CA1 area. In a third group, we performed neuronal counts.

Results: In both age groups, acute seizures were similar in KA and KA + AIDA groups. Rare spontaneous recurrent seizures occurred only in KA-injected rats. The KA P20 group performed significantly worse than controls in the water-maze test. The KA + AIDA group showed impaired performance on day 1, but learning improved substantially, reaching control values in the remaining 3 days. The P30 KA rats performed worse than controls on all trial days, whereas the KA + AIDA rats improved by day 3, but did not reach control values. Electrophysiologic recordings showed small but consistent differences between KA and control animals, suggestive of an adaptive modification in the γ-aminobutyric acid (GABA)ergic system, reversed by AIDA. On histology, we observed a loss of CA1 interneurons in both ages. Cell loss was reversed by the use of AIDA.

Conclusions: Blockade of the class I mGluR during KA-induced seizures in the developing brain limits seizure-induced hippocampal dysfunction.

The syndrome of mesial temporal epilepsy associated with hippocampal sclerosis is the most frequent form of refractory epilepsy in humans (1). Its clinical presentation is often characterized by a prolonged seizure in early life followed by a latent period and then chronic recurrent seizures. Initially identified in adults, hippocampal sclerosis is more and more frequently recognized in children because of better imaging techniques and, in certain children, appears to be secondary to the initial prolonged seizure (2). Associated with the development of hippocampal sclerosis, the recurrent seizures lead to a progressive loss of hippocampal functions, as shown by neuropsychological testing (3). In the developing brain, these seizures also can impede the acquisition of basic skills and lead to specific learning disabilities (4).

The study of animal models of temporal lobe epilepsy, such as the kainate model, increased our knowledge of the pathophysiology of hippocampal sclerosis (5). In the adult brain, the pathology shows neuronal loss and gliosis involving the mesial temporal structures. This is a consequence of the release of excessive amounts of excitatory amino acids during seizures (6), which in turn causes an excessive intracellular calcium (Ca2+) accumulation, leading to neuronal damage (7). In the developing brain, neuropathologic findings have been less consistent, but hippocampal dysfunction and increased seizure susceptibility have all been well described. Investigators have therefore attempted to prevent seizure-induced brain damage by blocking the ionotropic glutamate receptors that are Ca2+ permeable (8). As Ca2+ influx occurs primarily through the N-methyl-d-aspartate (NMDA) receptors, they were believed to be choice targets for therapeutic interventions (9). However, clinical and experimental trials using NMDA-receptor antagonists have shown only limited benefit with significant adverse effects in the developing animal (10,11). Blockade of other ionotropic glutamate receptors has been attempted with success in animals (12). However, blocking fast synaptic transmission in humans is a cause of great concern, especially in the developing brain, because of its role in synaptic transmission (13).

In addition to Ca2+ influx, the release of Ca2+ from the intracellular stores also can lead to neuronal death (14). A subclass of glutamate receptors, the metabotropic glutamate receptors (mGluRs) participate in intracellular Ca2+ accumulation, via its release from nonmitochondrial stores (15). The mGluRs are a family of G protein–coupled receptors grouped into three classes based on their effects on the secondary messengers inositol triphosphate (IP3) and cyclic adenosine monophosphate (cAMP) and further subdivided into eight subtypes with amino acid sequence similarity. The class I (mGluR1 and mGluR5) receptors are particularly interesting because they activate the IP3 cascade, causing the release of Ca2+ from the intracellular nonmitochondrial stores. The mGluRs are a good target for neuroprotective agents, because they have been shown to be located at the periphery of the synapse and are activated only during periods of synaptic hyperactivity, as seen during an epileptic seizure (16). The evidence that class I mGluR antagonists show neuroprotective effects in ischemia models lends further support to our hypothesis (17).

Kainic acid (KA) is an analogue of the excitatory amino acid neurotransmitter glutamate and a potent convulsant and neurotoxin. When injected intraperitoneally (i.p.), KA induces a prolonged initial seizure followed by recurrent chronic limbic seizures that lead to hippocampal dysfunction reminiscent of the human condition (18). The pathology seen in epileptic rats after KA injections is similar to that seen in humans with temporal lobe epilepsy (5). For these reasons, KA-induced seizures in rats are considered a model for the study of mesial temporal epilepsy. It is believed that the i.p. injection of KA causes a massive release of glutamate, as seizures do in humans (6), and the class I mGluRs appear to be implicated in the pathology associated with the KA model via the release of intracellular Ca2+ stores (19).

The study of seizure-induced brain damage in the immature animals gives us a unique opportunity to reproduce what is observed in the clinical situation and to understand better the mechanisms involved in cellular vulnerability and plasticity during the formation of neuronal circuits (20,21). In immature animals, before P10, there does not appear to be any anatomic or physiologic long-term consequence to these prolonged seizures. If seizures are induced after P20, cognitive functions are altered when animals are tested as adults, and they have an increased seizure susceptibility (22).

The goal of this study is to evaluate the effects of an antagonist of the class I mGluR on seizures and hippocampal dysfunction induced by KA in the developing brain. We used 1-aminoindan-1,5-dicarboxylic acid (AIDA), a compound with a median inhibitory (IC50) of 214 μM, developed by Pellicciari and Costantino (23). It is devoid of any action on class II and III receptors and is much more potent than the nonspecific blocker MCPG, which has an IC50 of 700 μM(24). More potent class I antagonists have recently been developed, like CPCCOEt and MPEP, but these are subtype specific for mGluR1 and 5, respectively (25). We measured the effect of AIDA on the severity of acute KA-induced seizures, on the incidence of chronic spontaneous recurrent seizures (SRSs), and on the long-term hippocampal function based on the performance in a modified version of the Morris Water Maze (18). We complemented this study by performing histologic cell counts and electrophysiologic recordings on tissue obtained from test (KA, KA + AIDA) and control animals (AIDA, saline).

METHODS

Animals

Male Sprague–Dawley rats, P20 (n = 97) and P30 (n = 68), were obtained from Charles River Laboratories (St-Constant, Québec, Canada). They were maintained on a 12-h light/dark cycle and given free access to standard laboratory chow and water. All experiments were performed in accordance with protocols and guidelines of the Canadian Animal Care Committee (1993).

Initial drug treatment.

In both P20 and P30 rats, four protocols were performed. One third of the animals received a fixed volume of 8 mg/kg KA i.p. (Tocris Cookson) diluted in saline (pH 7.4), and one third received 1.8 mg/kg of AIDA (Tocris Cookson) with 8 mg/kg KA i.p. diluted in saline. For the remaining third, half of the rats received 1.8 mg/kg AIDA in saline, and the other half received an identical volume of saline only. The dosage of AIDA was determined based on previous studies showing a significant behavioral effect of the drug after i.p. injections on hippocampal-dependent processes (26). Even though younger animals are more sensitive to the convulsant effects of KA, a similar dose was used in both groups, because older rats are known to show more extensive cell loss despite less severe seizures in a dose-dependent fashion (22). The initial seizure activity was monitored 2 h after the injection. In each group, we measured the latency to seizure onset, seizure duration, and seizure severity based on the scale developed by Sperk et al. (27).

Spontaneous seizure monitoring

For 30 days after the KA injection, we monitored daily the animals by filming for 1 h in the morning and 1 h in the evening to observe spontaneous recurrent seizures. By using a wide-angle lens video camera, we were able to monitor up to nine rats simultaneously. The behavioral episodes consistent with an ictal behavior were counted for each animal, including repetitive clonic limb movements as well as rearing/falling seizures, and were expressed as the number of events per day. We also monitored wet-dog shakes (WDSs), an abnormal behavior known to correlate with chronic epilepsy in animal models (28).

Water-maze testing

After the 30-day observation period, the animals were tested on a modified version of the Morris water maze (22). A circular tank was filled with water (26 ± 1°C) to a depth of 25 cm. An 8 × 8-cm platform was positioned at the center of one quadrant 1 cm under the surface of the water so the rat could escape. The water was made opaque to prevent visualization of the platform. Room lights illuminated the bath, and visual cues from the room remained stable over the learning period. We designated four points on the perimeter of the bath as entry points north, south, east, and west. On day 0, the rats were placed in the bath for 60 s without a platform to habituate them to the experimental condition. On the following 4 days of testing, rats were trained to exit the bath onto the platform by using the room's visual cues. Six trials per day were performed, and we measured on each trial the time to exit the platform. If a rat failed to find the platform after 60 s, it was placed on the platform for a 30-s rest period. Rats received a 30-s rest period between successive trials.

Electrophysiological testing

In a second group of 17 rats injected at P20, we performed field recordings in area CA1 of the hippocampus obtained at P50. The electrophysiologic experiments were carried on a group of rats that did not perform the water maze (n = 17). We used three groups injected respectively with KA, KA + AIDA, or saline. We did not use an AIDA-only group, as they behaved similar to saline controls. Animals were decapitated under halothane anesthesia, the brains removed, and transverse hippocampal slices 450- to 500-μm thick prepared. Slices were placed in two independently perfused submersion chambers, where they were perfused with warm (32 ± 1°C) oxygenated (95% O2/5% CO2) artificial cerebrospinal fluid (aCSF) of the following composition (in mM): NaCl, 124; KCl, 2; KH2PO4, 1.25; CaCl2, 2; MgSO4, 2; NaHCO3, 26; glucose, 10; and pH adjusted to 7.4. The volume of each recording channel was 0.28 ml, and the perfusion rate was 0.6 ml/min.

Extracellular stimulating bipolar electrodes, made from Teflon-covered steel wire, were placed at the CA1 Schaffer collateral– commissural pathway, and orthodromic stimuli were delivered every 25–40 s. Extracellular recording electrodes were filled with 4 M NaCl (resistance, 10–20 MΩ). Evoked responses were displayed on a storage oscilloscope, where four to eight successive responses were averaged and printed, and spontaneous responses were displayed in a continuous chart recording throughout the test. The maximal amplitude of field excitatory postsynaptic potentials (fEPSPs, in mV) was measured as the difference between the positive peak of the signal minus the prestimulus baseline. Duration of the fEPSP was measured from the artefact of stimulation to the intercept of the trace with the prestimulus baseline. The amplitude of the population spikes (PSs) was measured from their negative peak to the half distance of their positive peaks in mV; the optical detection limit of a PS was 0.2 mV. The peak latency of the first PS was measured from the stimulation artefact to its negative peak (ms). The duration of the first PS was measured by subtracting the early positive PS peak from the late positive PS peak (ms). As a standard procedure recording continued for ≥10 min after the washout of substances, to verify the reversibility of the observed effects.

Comparative measures were obtained in the presence of the γ-aminobutyric acid subtype A (GABAA)-receptor antagonist bicuculline methiodide (BMI, 10 μM, Sigma). The drug was dissolved in aCSF of the same composition as described earlier.

Tissue processing and histology

Cellular loss was assessed 30 days after the drug injections in a third group of animals (n = 20 P20 + 20 P30) that did not perform the Morris Water Maze or undergo electrophysiologic recordings. Rats were anesthetized with the injection of a solution of ketamine hydrochloride (90 mg/kg i.p.; Ketaset, Ayerst) and xylazine hydrochloride (10 mg/kg i.p.; Rompun, Bayer) and perfused transcardially with 0.1 M phosphate-buffered saline (PBS) followed by 4% paraformaldehyde in PBS at room temperature (RT) and pH 7.4. After perfusion, brains were removed and postfixed in the same fixative for 24 h and washed in PBS for another 24 h. The brains were then dehydrated through a graded series of ethanol solutions, cleared in methyl benzoate, and embedded in paraffin. Serial sections (10 μm) were cut with a microtome along the anterior–posterior axis through the hippocampal formation, collected on silane-coated slides, and stained with cresyl violet.

The occurrence of apoptosis was assessed 24 h and 8 days after the drug injections in 24 additional rats of each age group by using a terminal deoxynucleotidyltransferase (TdT)-mediated dUTP nick-end labeling (TUNEL) protocol (ApopTag kit; Oncor) (29). In brief, sections were deparaffinized with xylene and rehydrated through graded alcohols and water. They were afterward incubated with 20 μg/ml of proteinase K (Sigma Chemical, St. Louis, MO, U.S.A.) in PBS at RT for 15 min for permeabilization and then washed with dH2O (3 times for 5 min). Endogenous peroxidase activity was subsequently quenched by incubating the sections with 3% H2O2 in PBS for 5 min at RT. Sections were afterward incubated with a solution of TdT in reaction buffer for 1 h at 37°C, washed in stop buffer for 10 min, and incubated in a humidified chamber with anti–digoxigenin peroxidase conjugate for 30 min at RT. They were finally washed in PBS (4 times for 2 min), treated with 0.05% 3-3′-diaminobenzidine tetrahydrochloride (DAB) to develop the reaction, and counterstained with 0.5% methyl green to reveal nuclei. We performed negative controls by omitting TdT. Paraffin-embedded sections of weaning mammary glands furnished with the ApopTag kit were used as positive control.

Cell counting

Pyramidal cells of the CA1 and CA3 area as well as the hilar cells and the interneurons of O/A and R/LM strata were counted on sections confined to the septal area where the lateral and medial blades of the dentate gyrus were approximately equal and sectioned 500 μm posterior to that point (30). The estimations of the numeric density of the CA3 and CA1 pyramidal cells were calculated by using the optical dissector method (31). In brief, the counts were performed by using a ×40 lens with a numeric aperture of 0.65 and depth of field of 3.04 μm. A counting grid of 50 × 50 μm divided by 100 5 × 5-μm squares was designed by using an image-analysis system (Image-Pro, Media Cybergenics) and superimposed over the sampled sections throughout the length of the hippocampus. Cells were counted in four squares. The numbers of cells of the hilus and of the O/A and R/LM strata were estimated by the fractionator method (32). In this method, the total number of cells was counted, and the estimation of the number was made by multiplying the number of cells counted in one section by the number of sections used.

Statistical analysis

We compared the groups for the latency of the initial seizure, the duration of the initial seizure, the frequency of spontaneous recurrent seizures, and the average time to escape onto the platform by using mean values ± SD. Histology and electrophysiology data are presented as mean ± SEM. Values between multiple groups were compared with an analysis of variance (ANOVA) test (followed by the Bonferroni test for multiple comparisons). Differences between the means of two groups were compared by using Student's t test; in all cases, the p < 0.05 level was used to determine statistical significance.

RESULTS

Initial seizure severity

The injection of 8 mg/kg KA with or without AIDA led to seizures in all animals with no mortality in both age groups. By contrast, injection trials of 10 and 12 mg/kg led to a mortality rate of 25%, judged unacceptable by our ethics committee.

In the P20 rats (n = 36), the provoked seizures were characterized initially by periods of immobility lasting 20 s, followed by WDSs and by focal clonic activity in all animals treated with KA and KA + AIDA. Twenty-one of 24 animals injected with KA and KA + AIDA went on to have generalized tonic–clonic seizures. No difference was seen in the seizure severity scale (Table 1). The onset latency in the KA + AIDA group was somewhat shorter (16.7 ± 1.0 min) than in the KA group (19.1 ± 0.8 min), a difference not statistically significant. In addition, there was no statistical difference between the KA + AIDA and KA groups in seizure duration (45.6 ± 11.8 vs. 42.1 ± 12.4 min). No seizure was observed in either the AIDA or saline groups.

Table 1.  Seizure characteristics according to treatment groups. The initial seizure severity rating is expressed on a scale of 0–5 described by Sperk
GroupInitial
seizure severity
Total number of SRS
 R + F    nclonic
WDS mean number ± SD
  1. SRS = Spontaneous recurrent seizures.

  2. R + F = Rearing and falling.

  3. WDS = Wet dog shakes.

P20 KA (n = 12)3.923512.2 ± 6.5
P20 KA/AIDA (n = 12)3.8004.3 ± 2.3*
P20 AIDA (n = 6)0002/6
P20 saline (n = 6)0004/6
P30 KA (n = 8)3.02014.7 ± 3.9
P30 KA/AIDA (n = 8)3.00014.9 ± 2.9
P30 AIDA (n = 4)0002/4
P30 saline (n = 4)0001/4

In the 24 P30 rats injected with KA, the initial seizures in the rats evolved from periods of immobility to WDSs and to clonic activity, without, however, reaching the stage of rearing and falling seizures. The seizure severity estimated by the scale was comparable in the KA versus KA + AIDA groups (Table 1). Mean onset latency was similar in both the KA + AIDA group with 25 ± 2.9 versus 26.7 ± 1.7 min in the KA group. The mean seizure duration was 35.2 ± 14.2 min in the KA + AIDA group and 36.3 ± 11.1 min in the KA group. Again, no seizure was observed in the AIDA or saline groups.

Spontaneous recurrent seizures

In the P20 group (n = 36), SRSs were observed six of 12 of the KA-treated animals and in none of the 12 KA + AIDA–injected animals. The first seizures were observed by day 24. Convulsive behaviors were characterized by focal clonic activity or less frequently by generalized rearing and falling seizures (Table 1). They exhibited these behaviors with a mean frequency of 1.3 ± 0.4 per rat per 30-day observation period (Table 1). None of the KA + AIDA animals had SRSs.

Abnormal behavior, characterized by WDS, also was observed in KA-treated animals with and without AIDA. The injection of KA + AIDA reduced significantly the frequency of these WDSs (p < 0.05, two tailed t test). The KA group experienced 12.2 ± 6.5 WDSs per day versus 4.3 ± 2.3 for the KA + AIDA animals (Table 1).

In the P30 animals, SRSs were witnessed in two of eight KA-treated animals, with one rearing and falling seizure each on days 27 and 29. WDSs were observed in the P30 KA-treated animals, but contrary to the P20 group, AIDA did not significantly reduce their number (KA + AIDA: 14.7 ± 3.9; KA: 14.9 ± 2.9; Table 1).

Water-maze testing

At the end of the 30-day observation period, a modified Morris water-maze test was performed to assess the rats' visuospatial learning abilities, a specific hippocampal function known to be impaired after KA-induced seizures (22).

For the animals injected at P20 and tested at P50 (n = 36), KA-treated animals took significantly more time to reach the platform than did control animals on all 4 days of testing. The difference was maximal on the last day, with KA-treated animals reaching the platform in an average time of 18.3 ± 4.1 s versus 10.0 ± 2.2 s for the controls (p < 0.05, two-tailed Student's t test). AIDA limited the deleterious effects of KA, except on day 1. On the last 2 days of testing, the KA + AIDA group performed similar to the control group (Fig. 1A).

Figure 1.

Figure 1.

Morris water maze results of animals injected at P20–P30 and tested at P50–P60.

A (upper side panel)- Results show an increased time to reach the platform on all 4 days for rats treated with KA at P20. This was reversed by the co-injection of AIDA, as this group only showed prolonged results on day 1 of testing. KA= kainic acid; AIDA= 1-aminoindan-1,5-dicarboxylic acid; (* p< 0.05, two tailed).

B (lower side panel)- With P30 injected animals we saw an increased time to reach the platform on all 4 days for rats treated with KA. This was partially reversed by AIDA, with abnormal delays to reach the platform on days 1-2. However on days 3–4 these animals performed as well as controls. (* p< 0.05, ** p= 0.06).

Figure 1.

Figure 1.

Morris water maze results of animals injected at P20–P30 and tested at P50–P60.

A (upper side panel)- Results show an increased time to reach the platform on all 4 days for rats treated with KA at P20. This was reversed by the co-injection of AIDA, as this group only showed prolonged results on day 1 of testing. KA= kainic acid; AIDA= 1-aminoindan-1,5-dicarboxylic acid; (* p< 0.05, two tailed).

B (lower side panel)- With P30 injected animals we saw an increased time to reach the platform on all 4 days for rats treated with KA. This was partially reversed by AIDA, with abnormal delays to reach the platform on days 1-2. However on days 3–4 these animals performed as well as controls. (* p< 0.05, ** p= 0.06).

In animals injected at P30 and tested at P60 (n = 24), KA rats (17.9 ± 2.4) took a significantly longer time to reach the platform than did controls on the last 3 days of testing. Although this also was true on day 1, variability in the results seen in controls rendered the difference nonsignificant (17.9 ± 2.4 vs. 13.0 ± 3.7). On days 2–4, KA rats took more time to reach the platform than did controls, and again this difference was maximal on the last day of testing (p < 0.05, Student's t test, two-tailed). In this age group, AIDA limited the effects of KA, but KA + AIDA rats performed significantly worse than controls on day 2 (p < 0.05, Student's t test), and then improved on days 3–4, but never reached the performance levels of controls (Fig. 1B).

Electrophysiologic recordings

CA1 was chosen because it is the outflow tract of the hippocampus and has been described as showing functional long-term changes induced by KA (33) despite the presence of no or limited cell loss (5,22). Maximal-intensity stimulation of the Schaffer collateral/commissural pathway in standard aCSF evoked a CA1 fEPSP with a PS in 29 slices of five P50 saline-injected animals (controls); (Fig. 2, upper left trace). Similar results were obtained after orthodromic synaptic stimulation of the CA1 area of 22 slices from six P50 animals treated with KA at P20 and of 30 slices from six P50 animals treated with KA + AIDA at P20 (Fig. 2, middle and lower left traces). Figure 3 illustrates graphically the mean values ± SEM of the first PS amplitude (mV, A), the latency to its negative peak (ms, C), its duration (ms, D), and the fEPSPs duration (ms, E) of the three groups in control aCSF. In all indexes, consistent decreases were observed in slices from KA-treated animals. Orthodromic maximal PSs obtained in KA slices had significantly smaller amplitude compared with those of the controls (p < 0.05, one-tailed Student's t test), an effect reversed partially in the KA + AIDA group (Fig. 4A, left set of bars). In addition, a small decrease compared with controls was observed in the first PS latency to peak (Fig. 3C), its duration (Fig. 3D), and the fEPSP duration (Fig. 3E). All changes were reversed in the KA + AIDA group. No synchronous spontaneous activity was recorded in standard aCSF in any of the slices tested, suggesting no increased excitability in the CA1 area of rats that were submitted to KA-induced seizures despite the SRS. On the contrary, these in vitro recordings suggest decreased excitability in the KA slices, which in turn indicates either excitatory cell damage or adaptive inhibitory changes in the KA-treated animals. Coadministration of AIDA appears to have reversed the effects of KA on all the indexes measured, with the exception of PS amplitude.

Figure 2.

Characteristic evoked field potentials recorded from the CA1 pyramidal layer of hippocampal slices from control (saline), KA-treated (KA) and KA+AIDA-treated (KA+AIDA) animals. The responses shown on the left were recorded in standard ACSF (controls) and on the right, 10min (minimum) after the addition of the GABAA receptor antagonist bicuculline methiodide (BMI, 10μM) in the ACSF. Each trace is the average of 4 consecutive responses. Note the smaller amplitude of the 1st and the presence of a 2nd population spike in the KA slice. However, when GABAA receptors were blocked by BMI, all excitatory CA1 field potentials (saline, KA, KA+AIDA) appeared comparable. Calibration bars 2mV, 10ms.

Figure 3.

Graphic representation of the amplitude of the CA1 field potential indexes from saline, KA and KA+AIDA slices, recorded in standard ACSF (ctrl ASCF) as well as in the presence of BMI. A. PS amplitude (1st PS when in BMI) in KA (n=17) slices was lower than in control (saline, n=29) or KA+AIDA (n=23) slices both in standard ACSF and in BMI. This difference was statistically significant (unpaired student's t-test) between the saline and the KA slices and in one case between the KA and KA+AIDA (3rd PS, PS3) slices as well. B. The percent increase in the number of PSs following the addition of BMI was smaller in KA (n=22) slices compared to either saline (n=29) or KA+AIDA (n=23). C. The latency to the negative peak of the1st PS (ms) was slightly lower in KA (n=18) slices compared to either saline (n=29) or KA+AIDA (n=24) slices, both in control ACSF and in BMI. D. The duration of the 1st PS, measured by the difference (ms) of the two positive peaks was slightly lower in KA (n=18) slices compared to either saline (n=29) or KA+AIDA (n=23) slices, both in control ACSF and in BMI. E. Similarly to the previous, the duration (ms) of the fEPSP was slightly lower in KA slices (n=22) compared to either saline (n=29) or KA+AIDA (n=23) slices, in control ACSF as well as in BMI. None of the small differences in the graphs B–E was statistically significant although all were reversible. The vertical lines on top of each bar represent the SEM of the mean value.

Figure 4.

Figure 4.

Light photomicrographs of Cresyl violet-stained sections obtained 30 days after treatment in P20 and P30 rats showing the distribution of the interneurons in the O/A layer. An extensive loss of interneurons can be seen in the P20–P30 kainate-treated rats (A and E) compared to controls (C–D and G–H). In the K+AIDA groups (B and F) the effect of kainate was partly blocked. Arrows show the interneurons. ori, stratum oriens; pyr, stratum pyramidale; rad, stratum radiatum. Scale bar = 30 μm.

Figure 4.

Figure 4.

Light photomicrographs of Cresyl violet-stained sections obtained 30 days after treatment in P20 and P30 rats showing the distribution of the interneurons in the O/A layer. An extensive loss of interneurons can be seen in the P20–P30 kainate-treated rats (A and E) compared to controls (C–D and G–H). In the K+AIDA groups (B and F) the effect of kainate was partly blocked. Arrows show the interneurons. ori, stratum oriens; pyr, stratum pyramidale; rad, stratum radiatum. Scale bar = 30 μm.

The same slices were perfused with the GABAA-receptor antagonist bicuculline methiodide (BMI, 10 μm) to block synaptic inhibition and to examine whether possible differences in the field excitatory potentials differed in the three groups. This manipulation permits us to estimate the relative degree of functional GABAA-mediated inhibition, by comparing the percentage (%) BMI-induced increase in the measured synaptic activity indexes (larger changes would suggest robust GABAA inhibition; small changes, the contrary).

In BMI, the amplitude of the first PS increased, and the delay to its peak decreased, repetitive PSs were induced, and the duration of fEPSP was increased in all (n = 29) control slices, and similar findings were obtained from the KA (n = 22) or KA + AIDA (n = 23) slices (Fig. 2, right traces). The averaged indexes plotted in Fig. 3 were measured 15–20 min after the onset of BMI perfusion. A comparison of the mean values revealed the same differences as in those observed in standard aCSF, that is, smaller-amplitude PSs in KA slices (Fig. 4A), smaller first PS latency to peak, smaller first PS duration, and fEPSP duration. Again, the observed changes were small, but were all reversed in the KA + AIDA group (Fig. 3C–E). In addition, the increase in the number of repetitive PSs after BMI exposure was smaller in KA slices compared with either controls or KA + AIDA slices (Fig. 3B). BMI also induced spontaneous field potentials in three (10%) of 29, four (20%) of 20 KA, and in two (7%) of 30 KA + AIDA slices with very low frequencies ranging from 0.02 to 0.48 Hz (not shown).

Neuronal loss

No cell loss was observed in the CA1 and CA3 strata pyramidale as well as in the CA1 stratum R/LM and hilus. However, in P20 rats, the KA group exhibited a significant loss (±50%) of CA1 O/A interneurons compared with that in control groups (KA, 1,920.6 ± 186.5, vs. AIDA, 3,561.80 ± 140.50, and saline, 3,570.4 ± 127.0; p < 0.05). Figure 4 shows that this loss was partially blocked by AIDA (2,692.8 ± 196.9, p < 0.05) in the KA + AIDA group. In P30 kainate-treated rats, a significant loss (±50%) of CA1 O/A interneurons also was observed compared with control groups (KA, 1,911.8 ± 111.9, vs. AIDA, 3,617.0 ± 630.9, and saline, 3,652 ± 385.0; p < 0.05); AIDA (2,567.4 ± 235.12) partially blocked this effect (p < 0.05; Figs. 5 and 6). TUNEL-positive nuclei were not observed in the kainate-treated rats of both age groups. Moreover, observations at the light microscope revealed no structural changes in cellular morphology that are normally viewed in apoptotic cells (i.e., nuclear condensation and fragmentation, cell-surface protrusions, and formation of membrane-bound apoptotic bodies), which overall suggests that apoptosis does not contribute to the loss of the O/A interneurons observed.

Figure 5.

Numerical density of the pyramidal cells of the CA1 and CA3 area and total number of cells of the hilus and of the interneurons of the O/A and R/LM strata obtained 30 days after the treatment in P20 rats.

Figure 6.

Numerical density of the pyramidal cells of the CA1 and CA3 area and total number of cells of the hilus and of the interneurons of the O/A and R/LM strata obtained 30 days after the treatment in P30 rats.

DISCUSSION

Our major finding is that the selective class I mGluR antagonist AIDA limits the loss of hippocampal functions, as tested by the Morris water maze, in the KA model. This functional improvement, which was more marked in the P20- than in the P30-injected animals, appears to be mediated by a neuroprotective effect on CA1 oriens/alveus interneurons.

The group I mGluRs are coupled to phosphoinositide (PI) hydrolysis, increased levels of IP3, and intracellular calcium release. Both the elevation in hippocampal IP3 and the release of intracellular calcium stores have been associated with seizure-induced brain damage (34). Our data support the hypothesis that activation of class I mGluRs participates in seizure-induced brain damage in the KA model.

Lack of an anticonvulsant effect

Prevention of the seizure-induced brain damage could be secondary to an anticonvulsant effect of AIDA. However, our results in both P20 and P30 groups show no difference in acute seizure-onset latency, seizure duration, or seizure severity. The difference in seizure-onset latency between the two age groups (P20 vs. P30) is most probably due to maturational changes, because younger animals have been shown to be more sensitive to the convulsant effects of KA (22).

Proposed antiepileptogenic effect

An antiepileptogenic mechanism could also explain the less affected hippocampal functions in the KA + AIDA group. By this, we mean that blockade of class I mGluR could inhibit the cascade of events leading to the occurrence of SRSs. In both age groups of KA + AIDA–injected animals, we did not observe SRSs. These were observed only in the KA group, although not all the animals in this group had clinically detectable SRS. Our low observed seizure count could be due to the low dose of KA used to obtain an acceptable mortality rate. Similar findings are reported by other authors by using low-dose KA (27). We also monitored the presence of WDSs, an abnormal behavior observed in increased frequency in chronically epileptic rats (28). WDSs were significantly decreased in P20 rats who received KA + AIDA, but not in P30 animals, indicating that the antiepileptogenic effects of AIDA may be age related.

After a prolonged seizure in the developing brain, synaptic reorganization has been suggested as the underlying mechanism of SRSs (35). An interaction between some of the mGluRs and growth factors has been proposed to explain the observed neuroprotective effect (36). However, SRSs and evidence of synaptic reorganization are not required for the long-term deleterious effects of KA to be present (37). In our model, the differences in the occurrence of SRS could be explained by the preservation of oriens–alveus interneurons observed in the KA + AIDA group. Indeed, loss of interneurons was present in every tested animal and was present even in the P20 animals where no damage had been previously reported by using KA (22). CA1 is the outflow tract of the hippocampal formation, and the local interneuron circuit modulates the activity of the pyramidal cells. In area CA1, high and oscillating calcium responses have been observed after the application of glutamate in the oriens–alveus interneurons (38). These responses were shown to be mediated by mGluRs and to correlate with ictal-like discharges in the hippocampal slice model (38,39). Blockade of this excessive and prolonged accumulation of calcium by AIDA could protect these cells against seizure-induced cell death. A recent report also showed loss of CA1 interneurons in another animal model of temporal lobe epilepsy, the lithium–pilocarpine model, and this occurs during the latent period, leading to spontaneous recurrent seizures (40).

The electrophysiologic changes observed in slices were small but consistent compared with those in controls and were indicative of adaptive changes in the inhibitory system. These changes were partially reversed by the coadministration of KA + AIDA. In the absence of pyramidal cell loss, the smaller amplitude of CA1 PSs and fEPSP duration is consistent with a decrease in excitatory transmitter release, similar to what has been described after febrile seizures in the developing animal (41). The first PS latency and duration were shorter in KA slices, a change indicative of increased synchronization and in line with the inhibitory neuron loss. However, because these effects (on PS latency and duration) also were seen when the GABAA receptors had been blocked, they may be related to functional changes in other classes of GABA receptors and/or to increased excitatory connectivity (sprouting).

The fact that the observed changes were consistent in all measured indexes and that they reversed in the presence of AIDA suggests that they were not a chance finding. The fact that they were small may be due to the low-dose KA used in this protocol, to the lack of repeated exposure to KA, or even to the low sensitivity of the technique of extracellular recording. Conversely, the low sensitivity of the technique suggests also that such changes were widespread enough to have an impact on the synchronous activity of the slices recorded as field potentials.

Functional implications

On the Morris water maze, animals injected with KA at P20 or P30 took more time to reach the platform than did control animals on day 1 of testing and failed to improve significantly over the ensuing 4 days of testing. This effect was blocked by the injection of AIDA in both groups by day 3 of testing. This suggests that although abnormal, these animals have partially preserved hippocampal functions that enable them to learn the task. A previous report suggested that AIDA might have deleterious functional effects on hippocampal functions during normal synaptic activation (26). Another study using a nonspecific antagonist of mGluRs suggested that these receptors were required for visuospatial learning (42). In our study, a single dose of AIDA alone did not impair the long-term ability to learn the Morris water-maze task. However, because our animals were tested only 30 days after the injection, we have no data on the short-term effects of AIDA injections on water-maze performance. Our results argue therefore against any long-term detrimental effect of a single dose of this class I mGluR antagonist AIDA on hippocampal functions.

In conclusion, our findings suggest that blockade of class I mGluR receptors have a neuroprotective effect on hippocampal functions as well as a possible antiepileptogenic effect. This effect appears to be mediated by a direct neuroprotective effect on CA1 oriens–alveus interneurons and might be more specific to the developing brain, in line with the increased expression of class I mGluRs in the developing brain (43). Class I mGlu antagonists appear to hold promise in the prevention of the deleterious effect of prolonged seizures during development.

Acknowledgment: Preliminary results were presented at the American Epilepsy Society meeting in 2000. This study was funded by operating grants from Savoy Foundation for Epilepsy (L.C., C.P.), NSERC (C.P.), and studentships from Research Centre of Ste-Justine Hospital (J.R., S.M.).

Ancillary