Vigabatrin protects against hippocampal damage but is not antiepileptogenic in the lithium-pilocarpine model of temporal lobe epilepsy

Epilepsy Research 47 (2001) 99 – 117 www.elsevier.com/locate/epilepsyres

Vigabatrin protects against hippocampal damage but is not antiepileptogenic in the lithium-pilocarpine model of temporal lobe epilepsy Ve´ronique Andre´, Arielle Ferrandon, Christian Marescaux, Astrid Nehlig * INSERM U398, Faculty of Medicine, Uni6ersite´ Louis Pasteur, 11, rue Humann, 67085, Strasbourg, Cedex, France Received 1 March 2001; received in revised form 10 June 2001; accepted 22 July 2001

Abstract In temporal lobe epilepsy (TLE), the nature of the structures involved in the development of the epileptogenic circuit is still not clearly identified. In the lithium–pilocarpine model, neuronal damage occurs both in the structures belonging to the circuit of initiation and maintenance of the seizures (forebrain limbic system) as well as in the propagation areas (cortex and thalamus) and in the circuit of remote control of seizures (substantia nigra pars reticulata). In order to determine whether protection of some brain areas could prevent the epileptogenesis induced by status epilepticus (SE) and to identify the cerebral structures involved in the genesis of TLE, we studied the effects of the chronic exposure to Vigabatrin (gamma-vinyl-GABA, GVG) on neuronal damage and epileptogenesis induced by lithium-pilocarpine SE. The animals were subjected to SE and GVG treatment (250 mg/kg) was initiated at 10 min after pilocarpine injection and maintained daily for 45 days. These pilo– GVG rats were compared with rats subjected to SE followed by a daily saline treatment (pilo–saline) and to control rats not subjected to SE (saline– saline). GVG treatment induced a marked, almost total neuroprotection in CA3, an efficient protection in CA1 and a moderate one in the hilus of the dentate gyrus while damage in the entorhinal cortex was slightly worsened by the treatment. All pilo–GVG and pilo–saline rats became epileptic after the same latency. Glutamic acid decarboxylase (GAD67) immunoreactivity was restored in pilo–GVG rats compared with pilo– saline rats in all areas of the hippocampus, while it was increased over control levels in the optical layer of the superior colliculus and the substantia nigra pars reticulata. Thus, the present data indicate that neuroprotection of principal cells in the Ammon’s horn of the hippocampus is not sufficient to prevent epileptogenesis, suggesting that the hilus and extra-hippocampal structures, that were not protected in this study, may play a role in the genesis of spontaneous recurrent seizures in this model. Furthermore, the study performed in non-epileptic rats indicates that chronic treatment with a GABAmimetic drug upregulates the expression of the protein GAD67 in specific areas of the brain, independently from the seizures. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Temporal lobe epilepsy; Neuronal damage; GAD67; Interneurons

* Corresponding author. Tel.: + 33-390-24-3243; fax: + 33-390-24-3248. E-mail address: [email protected] (A. Nehlig). 0920-1211/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 1 ) 0 0 2 9 9 - 6

100

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

1. Introduction Temporal lobe epilepsy (TLE) is often associated with hippocampal lesions (Babb et al., 1984; Mathern et al., 1995a; Engel, 1996). This epilepsy is usually characterized by an initial precipitating injury occurring generally before the age of 4 years (French et al., 1993; Mathern et al., 1995b). Hilar GABAergic interneurons and mossy cells as well as CA1 and CA3 pyramidal neurons are selectively damaged in human TLE, and lesions eventually spread to extrahippocampal structures (de Lanerolle et al., 1989; Cendes et al., 1993; DeFelipe et al., 1993). This cell death pattern can be reproduced in animal models of TLE resulting from status epilepticus (SE) induced by kainate or pilocarpine (Sloviter, 1987; Turski et al., 1987; Houser, 1991). However, the relation between the extent of cell loss, the nature of the damaged structures and the genesis of epilepsy is still questioned. In this respect the model of epilepsy induced by lithium –pilocarpine may help to answer some of these questions. Indeed it is characterized by an acute phase of SE followed as in human TLE, by a silent phase during which lesions develop leading to the plastic reconstruction of a hyperexcitable circuit underlying the expression of spontaneous recurrent seizures (SRS) (Turski et al., 1983; Leite et al., 1990; Cavalheiro et al., 1991). In previous studies, we tried to prevent epileptogenesis and to protect the brain structures from lesions by using preconditioning with brief isolated seizures applied prior to lithium– pilocarpine SE. Our data, in accordance with many other morphological and functional studies, established the important role of the entorhinal and perirhinal cortices in the establishment of epilepsy (McIntyre and Kelly, 1990; Spiller and Racine, 1994; Du et al., 1995; Miettinen et al., 1998; Scharfman, 1996; Scharfman et al., 1998; Andre´ et al., 2000). In order to further clarify the relationship between the nature of the structures lesioned and the occurrence of epilepsy, in the present study, we explored the effect of a reinforcement of the GABAergic neurotransmission. We used the antiepileptic drug, Vigabatrin or gamma-vinyl-GABA (GVG), which is prescribed

for the treatment of refractory complex partial seizures (Gherpelli et al., 1997). Vigabatrin is an irreversible GABA transaminase (GABA-T) inhibitor, which prevents the degradation of GABA and elevates brain GABA concentrations (Grove et al., 1981; Preece and Cerdan, 1996; Qume and Fowler, 1996). GVG also decreases the brain concentrations of glutamate and aspartate in both rats (Halonen et al., 1991; Lo¨ scher and Horstermann, 1994) and epileptic patients (Petroff et al., 1995). In a model of ischemia, and models of SE induced by kainate or perforant path sustained stimulation, GVG exhibits neuroprotective effects in the rat hippocampus (Shuaib et al., 1992; Ylinen et al., 1991; Halonen et al., 1995; Jolkkonen et al., 1996). GVG-mediated neuroprotection appears to be selective to somatostatin positive interneurons, mainly in the septal part the of hilus (Ylinen et al., 1991; Pitka¨ nen et al., 1999). However, the relationship between the extent of neuroprotection induced by GVG and its potential consequences on the prevention of the plastic phenomena leading to epileptogenesis have not been explored. Therefore, we assessed both the neuroprotective and potential antiepileptogenic effects of GVG in the lithium–pilocarpine model of TLE. GVG was injected daily at a dose of 250 mg/kg, starting 10 min after pilocarpine and lasting for 45 days. We studied the effects of GVG treatment on the severity and duration of SE, on the electrographic characteristics of the silent and chronic periods, and on the latency to SRS. Neuronal damage was quantified using thionin staining and glutamic acid decarboxylase (GAD67) immunohistochemistry.

2. Material and methods

2.1. Lithium–pilocarpine and GVG treatments 2.1.1. Animals Male Wistar rats weighing 225–250 g, provided by Janvier Breeding Center (Le Genest-St-Isle, France) were housed under controlled standard conditions (light/dark cycle, 07:00–19:00 h lights

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

on), with food and water available ad libitum. All animal experimentation was performed in accordance with the rules of the European Communities Council Directive of November 24, 1986 (86/609/EEC), and the French Department of Agriculture (License Number 67-97). For electrode implantation, rats were anesthetized by an i.p. injection of 2.5 mg/kg diazepam (DZP, Valium, Roche, France) and 1 mg/kg ketamine chlorhydrate (Imalgene 1000, Rhone Merrieux, France). Four single-contact recording electrodes were placed on the skull, over the parietal cortex, two on each side, as previously described (Vergnes et al., 1982).

2.1.2. Status epilepticus induction, GVG treatment and occurrence of SRS All rats received lithium chloride (3 meq/kg, i.p., Sigma, St Louis, Mo, USA); about 20 h later, animals were placed into plexiglas boxes, in order to record baseline cortical EEG. Methyl scopolamine bromide (1 mg/kg, s.c., Sigma) was administered to limit the peripheral effects of the convulsant. SE was induced by injecting pilocarpine hydrochloride (25 mg/kg, s.c., Sigma) 30 min after methyl scopolamine. DZP (1– 2 mg/kg) was injected i.p. 2 h after SE induction, and every 4 h, 2 –3 times, to improve survival. The bilateral EEG cortical activity was recorded during the whole duration of SE and behavioral changes were noted. The effects of GVG treatment were studied on three groups of rats. The first group received 250 mg/kg GVG, i.p., 10 min after pilocarpine (pilo– GVG). Another group was injected the same volume of saline, 10 min after pilocarpine (pilo – saline). The control group received saline instead of pilocarpine and GVG (saline– saline). The pilo–GVG rats surviving SE were then injected daily with GVG, 100 mg/kg, s.c. during the 6 first days after SE and with 250 mg/kg from day 6 to 45. After SE, rats remained weak during several days. We noticed that rats injected with GVG, 250 mg/kg did not eat, presumably because of the drowsiness induced by GVG. Thus, we injected only 100 mg/kg of GVG until they recovered and were able to eat alone. Pilo– saline and saline –saline rats received daily an equivalent vol-

101

ume of saline. The effects of GVG on the EEG and on the latency to occurrence of SRS were investigated by the visual observation of the animals 5 days a week for 5–6 h and the recording of the electrographic activity twice a week for 8 h.

2.1.3. Quantification of cell densities Quantification of cell densities was performed at 6 days after SE on four pilo–saline, six pilo– GVG and four saline–saline rats. Animals were deeply anesthetized with 1.8 g/kg pentobarbital (Dolethal®, Ve´ toquinol, Lure, France), transcardially perfused with 100 ml ice-cold phosphatebuffer saline (PBS, pH 7.4) followed by 400 ml of ice-cold 4% paraformaldehyde (PFA) in 0.1 M PBS. Brains were then removed and postfixed with the perfusion solution during 2 h at 4 °C, cryoprotected with a 30% sucrose solution in 0.1. M PBS and frozen. Serial 20 mm slices were cut in a cryostat, air-dried during several days before thionin staining. Quantification of cell densities was performed with a 10× 10 boxes 1 cm2 microscopic grid on coronal sections according to the stereotaxic coordinates of the rat brain atlas (Paxinos and Watson, 1986). Cell counts were performed twice in a blind manner and were the average of at least three values from two adjacent sections in each individual animal. Counts involved only cells larger than 10 mm, smaller ones being considered as glial cells. Principal cells were quantified in the dorsal hippocampus in the CA1 stratum pyramidale and in the CA3 stratum pyramidale (15×10 − 3 mm2). Counts of non-pyramidal cells were performed in the hilus of the DG, the stratum oriens of CA1 the stratum radiatum of CA3. Values are expressed as mean9 S.D. (n= 4–6 animals). Surface areas were measured using a computerized image analysis software (Visioscan 2000, Biocom, Les Ulis, France) and were the average between both sides of 2–3 sections per animal (n= 4–6). 2.1.4. GAD67 immunohistochemistry A subset of animals exhibiting SRS (pilo–saline, n=5 and pilo–GVG, n=6) and saline–saline rats (n= 5) were deeply anesthetized at 45 days after SE with pentobarbital (Dolethal®, 1.8

102

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

g/kg) and perfused transcardially with PBS followed by ice-cold fixative (0.1 M PBS, 4% PFA, 1.5% sucrose, pH 7.4). Brains were removed immediately after perfusion and postfixed in the same fixative for 2 h at 4 °C. Transverse 50 mm sections were cut on a vibratome and stored at − 20 °C in an antifreeze solution (0.05 M PBS, 30% sucrose, 30% ethylene glycol, pH 7.2), as previously described (DiLullo and Martin, 1992). Sections were then sequentially incubated twice in PBS, once in 0.6% hydrogen peroxide in PBS and twice in PBS containing 0.4% normal goat serum (Vector Laboratories, Burlingame, CA, USA) and 0.25% Triton-X100. Immunoperoxidase staining was performed with a rabbit polyclonal antibody against GAD (67 kDa, Chemicon International, Inc, Temecula, CA, USA). Freefloating sections were incubated overnight at 4 °C with the diluted antibody (1:5000 in 0.1 M PBS, pH 7.4, 0.4% goat serum, 0.25% Triton X-100). Sections were then washed three times (PBS containing 0.4% goat serum) and incubated with the diluted secondary biotinylated goat anti-rabbit antibody (Biosys, Vector, dilution 1:400 in a Triton-X100/PBS mixture). Sections were rinsed twice in the latter medium and covered with the ABC reagent (Vectastain Kit, Vector) for 1 h at 20 °C. Sections were rinsed twice in PBS and incubated for 5– 8 min in a mixture of 0.02% diaminobenzidine, 0.5% nickel chloride and 0.05% H2O2 in PBS. Sections were then dehydrated in ethanol and coverslipped. Determination of the number of GAD67-stained cells was assessed as for thionin staining.

2.1.5. GVG treatment in control rats not subjected to lithium– pilocarpine Male Wistar non-epileptic rats from our breeding colony, weighing 225– 250 g were implanted with cortical electrodes as described above. Rats were injected daily with 250 mg/kg GVG, i.p. or saline during 45 days. They were divided in three groups. Group 1 received a daily saline injection (n= 7); groups 2 and 3 received a daily GVG injection during 45 days. In group 2, animals (n = 7) were sacrificed on the day following the end of the treatment while the rats of group 3 were only sacrificed at 15 days after the end of the

GVG treatment (n= 7). The electrographic activity of the rats was recorded during periods of 30 min, immediately after the injection (h0), at 1 h (h1), at 4 h (h4) and at 8 h (h8) post injection. Rats were recorded on day 1 (d1), d2, d8, d15, d30 and d45, the day of the last injection.

2.1.6. Data analysis For the comparison of the characteristics of SE in pilo–saline and pilo–GVG animals, a nonpaired Student’s t-test was used. The comparison between the number of rats seizing in both groups was performed by means of a  2 test. For neuronal damage and GAD immunohistochemistry staining, statistical analysis between groups was performed using analysis of variance (ANOVA) followed by a Fisher’s test for multiple comparisons using the Statview software.

3. Results

3.1. Effects of GVG treatment in rats with lithium–pilocarpine status epilepticus 3.1.1. Induction of status epilepticus Amongst rats which were injected with pilocarpine, 74% (17/23) of pilo–saline rats and 87% (20/23) of pilo–GVG rats developed SE. The difference was not statistically significant ( 2 = 0.5525, 0.15BPB 0.46). The same latencies to score 4–5 seizures were recorded for pilo–saline rats (23.49 5.5 min; mean9S.D.) and for pilo– GVG rats (24.095.6 min). The electrographic and behavioral characteristics of SE were identical in both pilo–saline and pilo–GVG groups. During SE, the same degree of mortality was noted in both groups: 53% (9/17) of pilo–saline rats and 50% (10/20) of pilo–GVG rats died ( 2 = 0.023, 3.84B PB5.41). 3.1.2. Silent and chronic periods All the rats studied until the chronic phase developed SRS with a similar latency. The latency for pilo–saline rats was 26.09 9.2 days (n= 7) and for pilo–GVG rats, 23.59 2.0 days (n= 7, P =0.65). None of the saline–saline rats (n= 5) developed SRS.

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

The EEG patterns during the silent and the chronic periods were different in pilo– saline and pilo – GVG rats, depending whether EEG was recorded before of after GVG administration. Two days after SE, the EEG was still disturbed in both groups, showing spikes and spike-andwaves (Fig. 1, silent period). In pilo–saline rats, at 3 days after SE, EEG returned to baseline, with occasional occurrence of low amplitude, fast activities, bursts of slow rhythms and atypical 4–5 Hz spikes-and-waves. In contrast, in pilo– GVG rats the EEG remained disturbed, before

103

and after GVG injection, and was punctuated by long lasting rhythmic spikes and spike-andwaves, with a low amplitude baseline (data not shown). Later during the silent period, at 6 days after SE, the EEG pattern returned to baseline in 4/7 pilo–GVG rats, punctuated by abnormalities similar to those seen in pilo– saline rats. However, in 3/7 pilo– GVG rats, the EEG remained abnormal either before or after GVG injection and was characterized by long lasting and rhythmic spikes and spikes-and-waves (Fig. 1, silent period).

Fig. 1. Cortical electrographic recordings in rats that underwent lithium – pilocarpine-induced SE, during the silent period (top) and the chronic period (bottom) in rats receiving daily saline (left) or 250 mg/kg GVG (right) and were recorded either just before or 4 h after GVG injection. At 2 days after SE, both groups of rats showed electrical abnormalities, spikes, spike-and-waves and burst of spikes. At 6 days, while the saline-treated rats showed a normal EEG punctuated by bursts of spikes, GVG-treated rats showed an abnormal EEG, with continuous post-ictal activities. During the chronic period, the EEG pattern returned to baseline in GVG-treated rats. The administration of GVG induced the occurrence of long lasting interictal activity 4 h after the injection. Sleep recordings were normal in pilo –saline while they were constantly disturbed, punctuated by spikes and bursts of spikes, in pilo – GVG rats whatever the time of recording after GVG injection.

104

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

During the chronic period, the EEG pattern recorded between 26 and 45 days in pilo– GVG rats changed depending on the delay between GVG injection and recording. Before the daily GVG administration, the EEG pattern of pilo– GVG rats was similar to the EEG pattern of pilo –saline rats, punctuated by occasional low amplitude, fast activities, slow rhythms and bursts of spikes, especially when the animals were quiet or sleeping. At 2 h after GVG, no change in EEG was observed, while at 4– 5 h, long bursts of spikes and spikes-and-waves appeared and were still present at 12 h after GVG administration (Fig. 1, chronic period). However, the EEG pattern of 3/7 pilo – GVG rats was continuously disturbed by long bursts of spikes and spike-and-waves, as those seen in the other four pilo–GVG rats, at 4–5 h after GVG administration. During the silent and chronic periods, pilo– GVG rats showed a longer duration of sleep and fast activities and frequent spikes were always present on the sleep recordings whatever the delay after GVG injection while these abnormalities were only rarely seen in pilo – saline rats (Fig. 1, sleep).

3.1.3. Cell densities and hippocampal surface areas In pilo–saline rats, the number of cells was massively decreased in the CA1 region of the hippocampus (80% cell loss in the pyramidal cell layer and 63% in the stratum oriens) while the CA3 region was less extensively damaged (48% cell loss in the pyramidal cell layer and 52% in the stratum radiatum, Figs. 2 and 3). In the dentate gyrus, the lithium–pilocarpine exposure induced extensive cell loss in the hilus (75%) while the granule cell layer did not show visible damage (Fig. 3). Similar damage was observed in the ventral hippocampus but cell counts were not performed in this region. In the piriform cortex, cell loss was too extensive to permit quantification, as layer II was not visible in pilo–saline rats (Fig. 3). In the entorhinal cortex, layer II underwent slight damage (32%) compared with layers III– IV (60%). No significant damage was seen in the perirhinal cortex (data not shown), the substantia nigra pars reticulata and the superior colliculus of pilo –saline rats (Fig. 2).

Fig. 2. Histograms of the number (mean 9 S.D.) of thioninstained cell profiles in various areas of control (saline – saline), pilo– saline and pilo – GVG rats at 6 days after SE. Abbreviations; CA1pyr, stratum pyramidale of the area CA1 of the hippocampus; CA1so, stratum oriens of the CA1; CA3pyr: stratum pyramidale of the area CA3; CA3srad, stratum radiatum of the CA3; ent II, layer II of the entorhinal cortex; ent III – IV; layers III-IV of the entorhinal cortex; PRh, perirhinal cortex; SNr, substantia nigra pars reticulata; SuG, superficial layer of the superior colliculus. °PB 0.05, statistically significant difference versus the saline – saline group. * PB 0.05: statistically significant difference versus the pilo – saline group

In pilo–GVG animals, cell loss was reduced in the hippocampus compared with pilo– saline rats. The CA1 pyramidal layer in which the cell loss reached 80% after lithium–pilocarpine, underwent a partial neuroprotection, and was limited to 40% in the pilo–GVG treated animals while the stratum oriens was not protected. In the CA3 pyramidal layer and stratum radiatum, the

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

105

Fig. 3. Photomicrographs of thionin-stained slices taken at the level of the hippocampus, in the CA1, CA3 and hilus (first, second and third rows) and the piriform cortex (bottom row) of a saline – saline rat (left column), a pilo – saline rat (middle column) and a pilo –GVG rat (right column) at 6 days after SE. The pilo – saline rat underwent damage in the pyramidal CA1 and CA3 regions while the pilo – GVG rat did not. In the hilus, the granule cells were preserved in the pilo – saline and pilo – GVG rats while hilar cells were lost in both groups. The piriform cortex of the pilo –saline and pilo – GVG rats underwent a similar extensive damage. Scale bar, 100 mm.

106

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

neuroprotection induced by daily GVG in lithium –pilocarpine exposed rats was total, as the number of cells in thionin-stained slices did not differ in the pilo–GVG group from the saline– saline rats. In the dentate gyrus, the cell loss in the hilus was slightly reduced, reaching 60% in pilo– GVG animals, compared with 75% in pilo– saline animals. This protection could be detected by cell counts, but was not visible (Figs. 2 and 3). The massive loss of cells observed in the piriform cortex of pilo–saline rats was identical in pilo– GVG rats and did not allow any counting in any of the lithium–pilocarpine treated groups (Fig. 3). In the entorhinal cortex, neuroprotection was not induced by GVG administration. Instead, we observed a slight worsening of cell loss in pilo– GVG rats compared with pilo– saline rats. This worsening was significant in layer II which underwent a loss of 44% in pilo–GVG compared with 22% in pilo – saline rats and non significant in layers III– IV (Fig. 2). No change was observed in the number of cells in the perirhinal cortex, substantia nigra reticulata and superior colliculus of pilo– GVG rats compared with pilo– saline and saline– saline rats (Fig. 2). The surface area of the hilus did not differ between the saline– saline group and the pilo– GVG group, while it was significantly smaller in the pilo– saline group versus the saline– saline group (208 vs. 272× 10 − 3 mm2). There was no significant difference in the hilar surface area of pilo and pilo–GVG rats. In the strata oriens and radiatum, the surface areas were similar in the three groups (data not shown).

3.1.4. GAD67 immunoreacti6ity GAD67 immunoreactivity was performed on rats which developed SRS and were sacrificed at 45 days after lithium– pilocarpine SE. In pilo– saline rats, in the CA1, GAD67-positive cells underwent a 42% decrease in the stratum oriens, while there was no change in the pyramidal cell layer. Similarly, the pyramidal cell layer of CA3 did not undergo GAD67-positive cell decrease while in the stratum radiatum the number of GAD67-positive cells was reduced by 30% (Figs. 4 and 5). In the dentate gyrus, rats underwent a 42% decrease in the number of GAD67-positive cells in the

hilus while the granule cell layer did not show any change. The superior colliculus and the substantia nigra reticulata that were not damaged by SE did not undergo changes in the number of GAD67positive cells either. In the entorhinal cortex that was damaged by lithium–pilocarpine SE, the number of GAD67-positive cells was not significantly decreased (Fig. 4). In pilo–GVG rats, the number of GAD67-positive cells was not decreased in any region of the brain. In contrast, it was significantly increased compared with pilo– saline rats in the strata radiatum and pyramidale of CA3, and in the hilus where it reached the number of cells of the saline–saline rats (Figs. 4 and 5). In more posterior regions, there was no decrease of GAD67-positive cells either. Surprisingly, some regions that did not undergo cell death, i.e. the optical layer of the superior colliculus and the substantia nigra pars reticulata showed a significantly higher number of GAD67-immunoreactive cells (respectively, + 100 and + 60%) compared with the saline–saline and pilo–saline rats (Figs. 4 and 5).

3.2. Effects of GVG treatment in control animals 3.2.1. EEG recordings The rats used for this part of the study were non-epileptic rats that we selected from our breeding colony and that did not exhibit absence seizures. This strain was selected for this part of the work to avoid any interference between the spike-and-wave pattern naturally occurring in most strains of rats (Jando et al., 1995) and the potential changes in EEG that could be induced by GVG. The baseline EEG confirmed that the animals did not exhibit spikes-and-waves. Up to 30 days, the daily injection of GVG did not induce any change in the EEG recordings. At that time, short bursts of 8–9 Hz spikes-and-waves appeared on the recordings in 3/8 rats. During the first days of GVG treatment, the rats lost some weight, compared with the rats receiving saline but recovered after a few days. 3.2.2. GAD immunohistochemistry In the hippocampus, the number of GAD67stained cells was similar in the hilus, the stratum

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

oriens of CA1 and stratum radiatum of CA3 of saline and GVG-exposed rats sacrificed at the end or 15 days after the end of the GVG treatment. The number of GAD67-immunoreactive cells was enhanced by 2–2.5-fold in the stratum pyramidale of CA1 in the groups of rats exposed to GVG whatever the time of sacrifice after the end of GVG exposure. In the hilus, the number of

107

GAD67-positive cells was not changed by the GVG treatment itself but increased by twofold at 15 days after the end of the treatment (Fig. 6). In the substantia nigra pars reticulata the number of GAD67-immunoreactive cells was enhanced by 8-fold in GVG-treated animals but returned to the control level at 15 days after the interruption of GVG treatment. Finally, in the optical layer of

Fig. 4. Histograms of the number (means 9 S.D.) of GAD67-stained cell profiles in various structures of saline – saline, pilo – saline and pilo – GVG rats sacrificed at 45 days after SE. Abbreviations, CA1pyr, stratum pyramidale of the area CA1 of the hippocampus; CA1so, stratum oriens of the CA1; CA3pyr, stratum pyramidale of the area CA3; CA3srad, stratum radiatum of CA3; EntII, layer II of the entorhinal cortex; GCL, granule cell layer; Op, optical layer of the superior colliculus; SNr, substantia nigra pars reticulata; SuG, superficial layer of the superior colliculus. °PB 0.05, statistically significant difference versus the saline – saline group. * P B0.05: statistically significant difference versus the pilo –saline group.

108

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

Fig. 5. Photomicrographs of GAD67 immunohistochemistry in the CA3 stratum radiatum (srad) (first row), the optical layer of the superior colliculus (Op) (middle row) and the substantia nigra pars reticulata (SNr) (last row) of a saline – saline rat (left column), a pilo– saline rat (middle column) and a pilo – GVG rat (right column) at 45 days after SE. In the CA3 srad, GAD67-positive cells were visible in the saline-saline and pilo – GVG rats while they were not detected in the pilo – saline rat. There was no difference in the number of GAD67-positive cells between the saline –saline and pilo – saline rats in Op and SNr. In contrast, the number of cells expressing GAD67 was increased in substantia nigra pars reticulata and in the optical layer of the superior colliculus of the pilo–GVG rat compared with the saline – saline rat.

the superior colliculus, the number of GAD67stained cells was increased by about 2-fold over control values by the GVG treatment at both times of sacrifice but the difference was not significant at 15 days after the end of the treatment because of the large interindividual variability of the data (Fig. 6).

4. Discussion The present study demonstrates that chronic treatment with an antiepileptic drug that increases levels of GABA, GVG could protect hippocampus from brain damage efficiently in the Ammon’s horn, and to a smaller extent in the hilus while it

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

had no effect on extra-hippocampal structures. Although we started the treatment early, 10 min after pilocarpine injection, and despite the partial neuroprotection of the hippocampus, GVGtreated rats were not protected from epilepsy. Finally, in epileptic rats as well as in non-epileptic rats, chronically administrated GVG induced specific increases in the expression of GAD67, the enzyme responsible for the synthesis of GABA.

4.1. Chronic administration of GVG induces hippocampal neuroprotection The extension and location of damage in rats subjected to lithium– pilocarpine SE are in accordance with previous studies (Honchar et al., 1983; Motte et al., 1998). As shown by cell counting and by GAD67 immunoreactivity, the chronic GVG treatment afforded total neuroprotection in the strata pyramidale and radiatum of CA3 while the stratum pyramidale of CA1 and the hilus were partly protected. These data are in accordance with the hippocampal protection linked to GVG in models of SE induced by kainate or perforant path sustained stimulation (Ylinen et al., 1991;

109

Halonen et al., 1995; Jolkkonen et al., 1996). Our data are in contradiction with a recent study reporting the lack of neuroprotection of GVG treatment in the temporal lobe of epileptic rats (Halonen et al., 2001). However, in this study, the treatment was started 2 days after SE. At that time, extensive neuronal damaged has already occurred; on the opposite, we started the treatment 10 min after SE. These findings suggest that immediate treatment with GVG induces protection of cells in the hippocampus while delayed treatment does not have this protective action. In other parts of the brain that were damaged following lithium– pilocarpine SE, cell loss was not reduced by GVG. The piriform and the entorhinal cortices were not protected. Instead, cell loss was slightly enhanced in the entorhinal cortex by chronic administration of GVG. The same worsening of damage was observed in another study (Halonen et al., 2001) and might indicate that GVG-mediated increases of GABA do not have the same effects depending on the regions of the brain as postulated by Lo¨ scher et al. (1989). In the present study, the number of GAD67positive neurons in the hilus of pilo–saline rats

Fig. 6. Histogram of the number (mean 9 S.D.) of GAD67-stained cell profiles in the various structures of saline-treated rats, and GVG-treated rats sacrificed either at the end of the 45 days GVG treatment (GVG) or at 15 days after the end of the GVG treatment (GVG +15d). Abbreviations, CA1pyr, stratum pyramidale of the area CA1 of the hippocampus; CA1so, stratum oriens of the CA1; CA3pyr, stratum pyramidale of the area CA3; CA3srad, stratum radiatum of CA3; SNr, substantia nigra pars reticulata; Op, optical layer of the superior colliculus. °PB 0.05: statistically significant difference versus the saline-treated group. *P B0.05: statistically significant difference versus the GVG group.

110

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

was decreased compared with saline– saline rats, suggesting either a loss of interneurons, or a decrease in the expression of the protein. The number of GAD67-positive neurons of pilo– GVG-treated rats was not different from control rats, suggesting either a total protection of this type of cells by GVG or an increase in the amount of GAD expression in the remaining hilar neurons. In the Ammon’s horn, GAD67 immunoreactivity was also decreased in pilo-saline rats in the CA3 stratum radiatum and in the CA1 stratum oriens. This decrease was paralleled by a loss of neurons stained by thionin, indicating that the loss of immunoreactivity corresponded to cell loss. In pilo– GVG-treated rats, GAD67 immunoreactivity was totally restored in the Ammon’s horn, indicating that the interneurons were totally protected by GVG. Likewise, in other studies, GVG induced neuroprotection of somatostatin-positive neurons in the hippocampus of rats submitted to SE (Jolkkonen et al., 1996; Pitka¨ nen et al., 1999). This neuroprotection might be mediated by GVG-induced glutamate decrease (Halonen et al., 1991). As glutamate is involved in cell death and in excitotoxic mechanisms, GVGinduced decrease of glutamate might prevent neuronal death (Pitka¨ nen et al., 1996; Jolkkonen et al., 1996). The lithium–pilocarpine treatment induces a sclerosis of the hippocampus. In order to determine if the decrease in the number of cells was induced by the loss of cells or by a decreased size of the cerebral regions we studied, we measured the surface areas on the counted slices of the hippocampus. Surface area was decreased by 24% in the hilus of pilo rats compared with saline– saline or pilo–GVG rats while the number of cells was reduced by 80% in pilo– GVG compared with saline– saline rats in that region. Therefore, the limited protection afforded by GVG to the hilus of pilo–GVG compared with pilo– saline rats might partly reflect the treatment-related differences in surface area of the structure. However, the surface areas were not different in the Ammon’s horn, also protected by the GVG treatment, suggesting that the cell loss in pilo-rats and the neuroprotection induced by GVG are effective. The administration of GVG, 10 min after

pilocarpine did not induce any change in the behavioral and EEG characteristics of SE. This is in agreement with previous studies reporting that the increase of intracerebral GABA and decrease in glutamate underlying the efficacy of GVG is measurable only several hours after injection (Lo¨ scher et al., 1989; Valdizan et al., 1999). Conversely to other neuroprotective anticonvulsants, GVG did not decrease the severity and the duration of the electrographic seizures during lithium– pilocarpine-induced SE (Morrisett et al., 1987; George and Kulkarni, 1996). Thus, all the neuroprotective effects recorded in the present study are likely to reflect the effects of the drug and not the consequence of a change in the severity of SE.

4.2. Chronic administration of GVG disturbs electrographic acti6ities during the silent and chronic periods In pilo–saline rats, the EEG returned to baseline at 3 days after SE, with the occurrence of only occasional abnormalities. In contrast, in pilo–GVG rats, the EEG remained disturbed for longer periods. In the three out of seven pilo– GVG rats, the EEG was continuously punctuated by long lasting rhythmic spikes and spike-andwaves that could be recorded during the whole silent phase whatever the time of the day and the delay after GVG injection. In the four other rats, the EEG returned to baseline 1 week after SE. After this time, interictal activities varied with the delay after GVG injection. Long lasting spikes, spike-and-waves and bursts of spikes started to occur 4 h after the daily GVG administration and lasted at least up to 8 h after the injection, while the pilo–saline rats exhibited occasional interictal activities, independently on the time between saline injection and recording. The onset occurrence of GVG-mediated EEG disturbances at 4 h after injection is in agreement with the kinetic properties of GVG leading to significant changes in GABA and glutamate starting at about 4 h after the injection (Lo¨ scher et al., 1989; Valdizan et al., 1999). Part of the changes recorded on the EEG of the pilo–GVG rats may directly relate to the increase in GABA levels, as the EEG of nonepileptic control rats shows some disturbances

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

during a chronic GVG treatment, as seen in the present study and as reported by Halonen et al. (1992). The appearance of abnormal EEG activities in pilo –GVG rats could reflect the decrease in GABAergic inhibition that was reported to occur in rat hippocampal slices in the presence of high concentrations of GVG. In these conditions, there is a decrease of chloride ions uptake through the GABAA receptor leading to the appearance of multiple population spikes (Suzuki et al., 1991; Jackson et al., 1994). Likewise, when cultured neurons are subjected to trauma (excess heat, osmotic imbalance) GABA depolarizes the neurons and induces Ca2 + entry (van den Pol et al., 1996). The authors postulate that the long-lasting capacity of GABA to raise Ca2 + may allow GABA to play a greater role during recovery from trauma. Thus, in the present study, the occurrence of the electrographic interictal activities in pilo–GVG rats could be triggered by an excitatory action of enhanced GABA concentration during several days and may allow, as hypothesized by van den Pol et al. (1996) better recovery after SE which is in accordance with the neuroprotective effects we have seen in pilo– GVG compared with pilo– saline rats. In the model of amygdala kindling, the effects of GVG treatment depend on the delay between GVG administration and stimulation, and may result either in aggravating or anticonvulsant effects of the drug (Shin et al., 1986; Lo¨ scher et al., 1989). These opposite effects were attributed to the possible differential effects of GVG on nerve terminal concentrations in the striatum or the amygdala of kindled rats (Lo¨ scher et al., 1989). Thus, the aggravating effects of GVG could be due partly to GABA increases in areas like the striatum and the substantia nigra pars reticulata in which there is a relationship between GABAergic activity and modulation of seizure activity. Increases of GABA in the striatum and the substantia nigra pars reticulata aggravate seizures, possibly by preventing disinhibition of the superior colliculus (Gale and Casu, 1981; Le Gal La Salle et al., 1983; Gale, 1985). In pilo– GVG rats, the number of GAD67-positive cells was enhanced in the substantia nigra. If the enzyme is

111

functional in these neurons, this would mean that the inhibitory projection towards the superior colliculus is enhanced and may contribute to the occurrence of spikes during the interictal state in these rats. This hypothesis is supported by the report that nigrostriatal dopaminergic neurons develop tolerance to the chronic effects of GABAergic drugs: in these structures, [3H]GABA binding to the GABAA receptor was enhanced when GABA concentrations were increased (Beart et al., 1985). Thus, GVG-induced occurrence of spikes and spike-and-waves on the EEG recorded in the present study in pilo–GVG rats might be induced by the elevation of GABA in selective regions involved in the remote control of seizures, as postulated by Lo¨ scher et al. (1989). This mechanism may be particularly relevant for the GVGinduced occurrence of spikes and spike-and-waves on the EEG recorded after the first few days following SE since those cannot be attributed to the initial insult because they have disappeared by that time in pilo–saline rats.

4.3. Chronic administration of GVG has no antiepileptogenic effect The pilo–GVG rats developed SRS at the same time as pilo–saline rats indicating no antiepileptogenic effect of the GVG treatment. Conversely, it was reported that GVG-treated rats require a higher number of stimulations to develop amygdala kindling (Shin et al., 1986). The lithium–pilocarpine is a lesional model and though GVG induced hippocampal neuroprotection, this protection was total only in the CA3 pyramidal region, mild in the CA1 region and weak in the dentate gyrus. All GAD67-positive cells were restored in the hilus of pilo–GVG rats, although the cell counts performed on thionin-stained slices indicate that the protection is not total. The hilus contains two different types of cells, the excitatory mossy cells and inhibitory interneurons, the latter expressing GAD67 and GAD65 proteins. Although, GAD67 is expressed at higher levels than GAD65, GAD65 can be preferentially expressed in some cells (Esclapez et al., 1993; Jongen-Reˆ lo et al., 1999; Stone et al., 1999). Therefore, in our study, it is likely that we did not stain all GAD-

112

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

positive cells in the hilus and although the number of GAD67-positive cells was not different in the pilo –GVG rats compared with the saline– saline rats, it is likely that GVG treatment did protect a subset of the GAD-positive cells only. Nevertheless, these data indicate that damage in the hilus might be responsible for epileptogenesis and confirm our previous observations using another neuroprotection strategy. We showed that, even though a pretreatment by amygdala kindling applied before lithium– pilocarpine SE afforded neuroprotection in almost all brain regions, except the hilus of the dentate gyrus, we could not prevent the occurrence of SRS, which in addition appeared after the same latency in prekindled and nonkindled rats subjected to SE (Andre´ et al., 2000). The dentate gyrus is a gate for activities arriving to the hippocampus and loss of mossy cells and GABAergic cells in the hilus is likely to affect this control and could lead to seizures (Sloviter, 1991; Heinemann et al., 1992; Buckmaster and Dudek, 1997). GVG might not induce a sufficient protection to restore the loss of hilar cells after SE and to prevent reorganizations. Supporting this hypothesis, another study reported that GVG is able to protect somatostatin interneurons against kainate-induced damage only in the septal part of the hippocampus but not in the temporal part, where cell loss is more severe (Pitka¨ nen et al., 1999). In this study, mossy fiber sprouting was not prevented by chronic administration of GVG and was even increased if GVG was administrated during SE. Mossy fiber sprouting has been hypothesized to promote epileptogenesis (Cavazos et al., 1991; Mathern et al., 1993; Moriwaki et al., 1996; Wuarin and Dudek, 1996; Buckmaster and Dudek, 1997). Although, the role of mossy fiber sprouting in epileptogenesis is controversial, several studies reported that the sprouting of the granule cell fibers could mediate hippocampal hyperactivity (Cronin et al., 1992; Buhl et al., 1996; Longo and Mello, 1997, 1998) Thus, the GVG-mediated increase of mossy fiber sprouting could be responsible for the enhanced interictal activities that we recorded in pilo– GVG compared with pilo– saline rats. Furthermore, no other region was protected by GVG against SE-induced cell damage. The role of

extra-hippocampal lesions in epileptogenesis remains to be clarified. In prekindled animals, SEinduced damage is persisting predominantly in the hilus and the deep layers of the entorhinal cortex (Andre´ et al., 2000) which might contribute to the hyperexcitability seen in epilepsy. The entorhino– hippocampal loop is a very excitable structure and the entorhinal cortex can generate seizures and participate in their propagation (Jones and Lambert, 1990; Heinemann et al., 1992). Likewise, the piriform cortex and the amygdala play a major role in propagation of seizures (McIntyre and Kelly, 1989, 1993). The chronic administration of GVG did not prevent the occurrence of seizures although in human epilepsy as in amygdala kindling, GVG has an anticonvulsant effect against complex partial and secondarily generalized seizures (Le Gal La Salle et al., 1983; Guberman, 1996). Spontaneous seizures in the pilocarpine model can be prevented by other classical anticonvulsants used to treat complex partial seizures as phenytoin, phenobarbital and carbamazepine but the seizures most often reoccur after several days, suggesting the development of tolerance (Leite and Cavalheiro, 1995). Likewise, the anticonvulsant effect of GVG against amygdala-kindled seizures is lost during the second week of treatment (Rundfeldt and Lo¨ scher, 1992). In our study, GVG was injected before the occurrence of seizures and during 45 days. It is thus not surprising that the drug did not prevent seizures since by the time the SRS occur (more than 3 weeks); tolerance is likely to have developed.

4.4. The GABAergic system is modulated by GVG In rats subjected to lithium-pilo SE, the GVG treatment was able either to restore the number of neurons lost in animals not treated by the antiepileptic drug or even to raise the number of GAD67-positive neurons to levels higher than in control rats not subjected to SE. The latter increase was induced only in some parts of the brain, and could be seen both in pilocarpinetreated rats and in control rats which received GVG.

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

In the hippocampus of pilo– GVG rats, the number of GAD67-positive neurons was increased compared with pilo– saline rats, but not compared with saline– saline rats. This effect could be due to either a neuroprotective effect of GVG on the GAD67-positive cells, or to an increased expression of GAD67 in the remaining GAD-expressing neurons. Several studies reported that in TLE, GAD67 levels were upregulated in interneurons as well as in principal cells (Babb et al., 1989; Schwarzer and Sperk, 1995; Sloviter et al., 1996; Esclapez and Houser, 1999). One explanation for the increased expression of GAD67 is that the neurons are part of a network in which neuronal activity is increased, presumably by reorganizations in the sclerotic hippocampus. However, we did not quantify GAD67 levels per cell, but the number of cells expressing GAD67. We observed an increased number of cells expressing GAD67 in the pilo– GVG rats only, and the counts performed on thionin-stained slices confirmed that in the pilo– saline rats, cells were lost in the hilus, in the strata oriens and radiatum, known to contain interneurons. This cell loss was less marked in pilo– GVG rats. Therefore, the increased number of GAD67positive cells in the hippocampus of pilo– GVG rats might be due in part to the neuroprotective effect of GVG. As GAD is sensitive to neuronal activation, it is also possible that GVG induced disinhibition of cells that did not express GAD67 and that this disinhibition induced expression of the protein. On the other hand, in structures that were not damaged after SE, i.e. the substantia nigra and optical layer of the superior colliculus, GAD67positive cells reached a number higher in pilo– GVG than in saline– saline rats. The number of GAD67-positive cells in pilo– GVG rats increased by 2-fold in the optical layer of the superior colliculus. In these structures that do not undergo cell loss after SE, it is, therefore, likely that the amount of GAD67 was increased, allowing the detection of more cells positive for GAD67. The optical layer of the superior colliculus is innervated by retinal axons (Berson, 1988) and the superior colliculus has been implicated in several visual localizing functions (Sahibzada et al., 1986;

113

Westby et al., 1990). Furthermore, rats with lesions of the superior colliculus fail to respond to distracting visual stimuli presented in the peripheral field when they are running towards a central stimulus (Overton et al., 1985). These data are in agreement with the visual field constriction induced by GVG in humans (Baulac et al., 1998; Hardus et al., 2000). It is possible that visual functions in rats with increased GAD67-positive cells in the superior colliculus are modified as in humans treated against seizures with GVG.

4.5. Is the GABAergic system modulated by GVG or by both epilepsy and GVG? To determine whether the enhanced GAD67 expression was mediated by GVG itself or by the combination of epilepsy and GVG, we assessed GAD67 immunoreactivity on brains from control rats receiving GVG, but not subjected to SE. In the substantia nigra, a 8-fold increase of GAD67positive cells occurred in control rats treated with GVG. This increase was reversible as the number of GAD-positive cells returned to baseline level at 15 days after the interruption of the GVG treatment. In the optical layer of the superior colliculus, a 2-fold enhancement in the number of GAD67-positive cells was observed after a 45 days GVG treatment. At 15 days after the discontinuation, the level of GAD67-positive cells was still as high as during the treatment, although it was no longer significantly different from the control level. This result is in line with the fact that the typical GVG-attributed visual defects in humans were still present at periods ranging from 1 to 4 years after cessation of the treatment (Harding et al., 2000). GVG also induced enhanced expression of GAD67 in the CA1 stratum pyramidale and the hilus. In the CA1, the effect of GVG was not reversed at 15 days after the cessation of treatment in the hilus, a further 2-fold elevation in the number of cells expressing GAD67 was observed in the rats in which the GVG treatment was stopped. In the present study, the number of cells expressing GAD67 was enhanced in GVG-treated rats, though GABA concentration is highly increased in the brain of these rats (Grove et al.,

114

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

1981; Halonen et al., 1991). Previous studies reported that the increase of brain GABA concentrations induced by GVG induces regional decreases of the activity of the enzyme (Neal and Shah, 1990; Sheikh and Martin, 1998). However, these increases did not occur in all brain regions; there was no change in hippocampus, hypothalamus and striatum and an increase in the olfactory bulb (Sheikh and Martin, 1998). Moreover, while the decreases in GAD activity were attributed to the expression of GAD67 protein (Rimvall and Martin, 1994), the measurement was performed on the whole brain and was probably not influenced by changes even as marked as those we recorded in the present study in discrete regions like the superior colliculus, the substantia nigra and subregions of the hippocampus. Thus, the data of the present study would tend to confirm that the increases of GABA have regionally different effects, as postulated previously by Lo¨ scher et al. (1989). The most striking changes occur in regions involved in the remote control of seizures (the substantia nigra), in the superior colliculus, the brainstem visual relay nucleus (the visual field is sensitive to Vigabatrin) and in subregions of the hippocampus, mainly the hilus of the dentate gyrus largely involved in the SRS and the hyperexcitability recorded in the lithium– pilocarpine model of epilepsy. Finally, it remains possible that the decrease of expression of GAD67 in some neurons triggers the expression of GAD67 in neurons that normally do not express the protein. This phenomenon was observed in the granule cells of the dentate gyrus that express GAD, neuropeptide Y (NPY) and somatostatin after SE (Sloviter et al., 1996; Ding et al., 1998). When we compare the effects of the chronic GVG treatment in pilo– GVG and saline– GVG rats, it appears that the changes in the number of GAD67 positive neurons occur in the same structures and are, therefore, the result of the treatment rather than the consequences of the epilepsy.

5. Conclusion Chronic administration of GVG after lithium– pilocarpine induced SE selectively protected the

hippocampus from brain damage while no protection was afforded in the entorhinal and piriform cortices. This study allowed us to determine that early treatment with GVG during SE can prevent partly hippocampal cell loss while another study reported that a GVG treatment with later onset, 2 days after SE, could not prevent hippocampal neuronal damage (Halonen et al., 2001). Clinically, it might be useful to treat patients as soon as possible in order to avoid brain damage. This partial neuroprotection did not prevent the occurrence of SRS. These results indicate that clinically, neuroprotection with GVG may not be the ultimate strategy to prevent epileptogenesis. However, hippocampal damage is directly related to memory impairment and a protection of the cells, even partial in patients with TLE might help to maintain memory processes. Finally, selective enhancement of the number of cells positive for GAD67 was induced by GVG but not by epilepsy in correlation with GVG, in the hilus of the dentate gyrus, substantia nigra and the optical layer of the superior colliculus. The enhanced number of GAD67 positive cells in the superior colliculus might relate to the loss of concentric visual fields in patients treated with GVG.

Acknowledgements This work was supported by grants from the Institut National de la Sante´ et de la Recherche Me´ dicale (U 398), the Fondation pour la Recherche Me´ dicale and Marion-Merrell Laboratories. Vigabatrin was a gift from Marion-Merrell Laboratories.

References Andre´ , V., Ferrandon, A., Marescaux, C., Nehlig, A., 2000. The lesional and epileptogenic consequences of lithium – pilocarpine induced status epilepticus are affected by previous exposure to isolated seizures: effects of amygdala kindling and maximal electroshocks. Neuroscience 99, 469 – 481. Babb, T.L., Brown, W.J., Pretorius, J., Davenport, C., Lieb, J.P., Crandall, P.H., 1984. Temporal lobe volumetric cell densities in temporal lobe epilepsy. Epilepsia 25, 729 – 740.

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117 Babb, T.L., Pretorius, J.K., Kupfer, W.R., Feldblum, S., 1989. Recovery of decreased glutamate decarboxylase immunoreactivity after rat hippocampal kindling. Epilepsy Res. 3, 18 – 30. Baulac, M., Nordmann, J.P., Lanoe, Y., 1998. Severe visualfield constriction and side-effects of GABA-mimetic antiepileptic agents. Lancet 352, 546. Beart, P.M., Scatton, B., Lloyd, K.G., 1985. Subchronic administration of GABAergic agonists elevates [3H]GABA binding and produces tolerance in striatal dopamine catabolism. Brain Res. 335, 169 –173. Berson, D.M., 1988. Retinal and cortical inputs to cat superior colliculus: composition, convergence and laminar specificity. Prog. Brain Res. 75, 17 –26. Buckmaster, P.S., Dudek, F.E., 1997. Neuron loss, granule cell axon reorganization, and functional changes in the dentate gyrus of epileptic kainate-treated rats. J. Comp. Neurol. 385, 385 – 404. Buhl, E.H., Otis, T.S., Mody, I., 1996. Zinc-induced collapse of augmented inhibition by GABA in a temporal lobe epilepsy model. Science 271, 369 –373. Cavalheiro, E.A., Leite, J.P., Bortolotto, Z.A., Turski, W.A., Ikonomidou, C., Turski, L., 1991. Long-term effects of pilocarpine in rats: structural damage of the brain triggers kindling and spontaneous recurrent seizures. Epilepsia 32, 778 – 782. Cavazos, J.E., Golarai, G., Sutula, T.P., 1991. Mossy fiber synaptic reorganization induced by kindling: time course of development, progression, and permanence. J. Neurosci. 11, 2795 – 2803. Cendes, F., Andermann, F., Gloor, P., Lopes, C.I., Andermann, E., Melanson, D., Jones, G.M., Robitaille, Y., Evans, A., Peters, T., 1993. Atrophy of mesial structures in patients with temporal lobe epilepsy: cause or consequence of repeated seizures? Ann. Neurol. 34, 795 – 801. Cronin, J., Obenaus, A., Houser, C.R., Dudek, F.E., 1992. Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers. Brain Res. 573, 305 – 310. DeFelipe, J., Garcia Sola, R., Marco, P., del Rio, M.R., Pulido, P., Ramon y Cajal, S, 1993. Selective changes in the microorganization of the human epileptogenic neocortex revealed by parvalbumin immunoreactivity. Cereb. Cortex 3, 39 – 48. de Lanerolle, N.C., Kim, J.H., Robbins, R.J., Spencer, D.D., 1989. Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy. Brain Res. 495, 387 –395. DiLullo, S.L., Martin, I.M., 1992. Evidence for presynaptic dopamine mechanisms underlying amphetamine- conditioned locomotion. Brain Res. 578, 161 –167. Ding, R.G., Asada, H., Obata, K., 1998. Changes in extracellular glutamate and GABA levels in the hippocampal CA(3) and CA(1) areas and the induction of glutamic acid decarboxylase-67 in dentate granule cells of rats treated with kainic acid. Brain Res. 800, 105 –113. Du, F., Eid, T., Lothman, E.W., Kohler, C., Schwarcz, R., 1995. Preferential neuronal loss in layer III of the medial

115

entorhinal cortex in rat models of temporal lobe epilepsy. J. Neurosci. 15, 6301 – 6313. Engel, J.J., 1996. Introduction to temporal lobe epilepsy. Epilepsy Res. 26, 141 – 150. Esclapez, M., Houser, C.R., 1999. Up-regulation of GAD65 and GAD67 in remaining hippocampal GABA neurons in a model of temporal lobe epilepsy. J. Comp. Neurol. 412, 488 – 505. Esclapez, M., Tillakaratne, N.J., Tobin, A.J., Houser, C.R., 1993. Comparative localization of mRNAs encoding two forms of glutamic acid decarboxylase with nonradioactive in situ hybridization methods. J. Comp. Neurol. 331 (3), 339 – 362. French, J.A., Williamson, P.D., Thadani, V.M., Darcey, T.M., Mattson, R.H., Spencer, S.S., Spencer, D.D., 1993. Characteristics of medial temporal lobe epilepsy: I. Results of history and physical examination. Ann. Neurol. 34, 774 – 780. Gale, K., 1985. Mechanisms of seizure control mediated by gamma-aminobutyric acid: role of the substantia nigra. Fed. Proc. 44, 2414 – 2424. Gale, K., Casu, M., 1981. Dynamic utilization of GABA in substantia nigra: regulation by dopamine and GABA in the striatum, and its clinical and behavioral implications. Mol. Cell. Biochem. 39, 369 – 405. George, B., Kulkarni, S.K., 1996. Protective effects of GABAergic drugs and other anticonvulsants in lithium-pilocarpine-induced status epilepticus. Method Find Exp. Clin. Pharmacol. 18, 335 – 340. Gherpelli, J.L., Guerreiro, M.M., da, C.J., Rotta, N.T., Manreza, M.L., Reed, U.C., Diament, A., Silva, E.A., Guerreiro, C.A., Nunes, M.L., Palmini, A., Vega, G.L., Vizioli, J., Pedroso, F., Chiste, M.A., 1997. Vigabatrin in refractory childhood epilepsy. The Brazilian Multicenter Study. Epilepsy Res. 29, 1 – 6. Grove, J., Schechter, P.J., Tell, G., Koch-Weser, J., Sjoerdsma, A., Warter, J.M., Marescaux, C., Rumbach, L., 1981. Increased gamma-aminobutyric acid (GABA), homocarnosine and beta-alanine in cerebrospinal fluid of patients treated with gamma-vinyl GABA (4-amino-hex-5enoic acid). Life Sci. 28, 2431 – 2439. Guberman, A., 1996. Vigabatrin. Can. J. Neurol. Sci. 23, S13 – S17. Halonen, T., Pitka¨ nen, A., Saano, V., Riekkinen, P.J., 1991. Effects of vigabatrin (gamma-vinyl GABA) on neurotransmission-related amino acids and on GABA and benzodiazepine receptor binding in rats. Epilepsia 32, 242 – 249. Halonen, T., Pitka¨ nen, A., Koivisto, E., Partanen, J., Riekkinen, P.J., 1992. Effects of vigabatrin on the electroencephalogram of rats. Epilepsia 33, 122 – 127. Halonen, T., Miettinen, R., Toppinen, A., Tuunanen, J., Kotti, T., Riekkinen, P.J.S., 1995. Vigabatrin protects against kainic acid-induced neuronal damage in the rat hippocampus. Neurosci. Lett. 195, 13 – 16. Halonen, T., Nissinen, J., Pitka¨ nen, A., 2001. Chronic elevation of brain GABA levels beginning two days after status epilepticus does not prevent epileptogenesis in rats. Neuropharmacology 40, 536 – 550.

116

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117

Harding, G.F.A., Wild, J.M., Robertson, K.A., Lawden, M.C., Betts, T.A., Barber, C., Barnes, P.M.F., 2000. Electro-oculography, electroretinography, visual evoked potentials, and multifocal electroretinography in patients with vigabatrin-attributed visual field constriction. Epilepsia 41, 1420 – 1431. Hardus, P., Verduin, W.M., Postma, G., Stilma, J.S., Berendschot, T.T., Van, V.C., 2000. Concentric contraction of the visual field in patients with temporal lobe epilepsy and its association with the use of vigabatrin medication. Epilepsia 41, 581 – 587. Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J., Zhang, C.L., 1992. The dentate gyrus as a regulated gate for the propagation of epileptiform activity. In: Ribak, C.E., Gall, C.M., Mody, I. (Eds.), The Dentate Gyrus and its Role in Seizures. Elsevier, Amsterdam, pp. 273 – 280. Honchar, M.P., Olney, J.W., Sherman, W.R., 1983. Systemic cholinergic agents induce seizures and brain damage in lithium- treated rats. Science 220, 323 –325. Houser, C.R., 1991. GABA neurons in seizure disorders: a review of immunocytochemical studies. Neurochem. Res. 16, 295 – 308. Jackson, M.F., Dennis, T., Esplin, B., Capek, R., 1994. Acute effects of gamma-vinyl GABA (vigabatrin) on hippocampal GABAergic inhibition in vitro. Brain Res. 651, 85 –91. Jando, G., Carpi, D., Kandel, A., Urioste, R., Horvath, Z., Pierre, E., Vadi, D., Vadasz, C., Buzsaki, G., 1995. Spikeand-wave epilepsy in rats: sex differences and inheritance of physiological traits. Neuroscience 64, 301 –317. Jolkkonen, J., Halonen, T., Jolkkonen, E., Nissinen, J., Pitka¨ nen, A., 1996. Seizure-induced damage to the hippocampus is prevented by modulation of the GABAergic system. Neuroreport 7, 2031 –2035. Jones, R.S., Lambert, J.D., 1990. Synchronous discharges in the rat entorhinal cortex in vitro: site of initiation and the role of excitatory amino acid receptors. Neuroscience 34, 657 – 670. Jongen-Reˆ lo, A.L., Pitka¨ nen, A., Amaral, D.G., 1999. Distribution of GABAergic cells and fibers in the hippocampal formation of the macaque monkey: an immunohistochemical and in situ hybridization study. J. Comp. Neurol. 408, 237 – 271. Le Gal La Salle, G., Kaijima, M., Feldblum, S., 1983. Abortive amygdaloid kindled seizures following microinjection of gamma-vinyl-GABA in the vicinity of substantia nigra in rats. Neurosci. Lett. 36, 69 – 74. Leite, J.P., Cavalheiro, E.A., 1995. Effects of conventional antiepileptic drugs in a model of spontaneous recurrent seizures in rats. Epilepsy Res. 20, 93 – 104. Leite, J.P., Bortolotto, Z.A., Cavalheiro, E.A., 1990. Spontaneous recurrent seizures in rats: an experimental model of partial epilepsy. Neurosci. Biobehav. Rev. 14, 511 – 517. Longo, B.M., Mello, L.E.A.M., 1997. Blockade of pilocarpine- or kainate-induced mossy fiber sprouting by cycloheximide does not prevent subsequent epileptogenesis in rats. Neurosci. Lett. 226, 163 –166.

Longo, B.M., Mello, L.E.A.M., 1998. Supragranular mossy fiber sprouting is not necessary for spontaneous seizures in the intrahippocampal kainate model of epilepsy in the rat. Epilepsy Res. 32, 172 – 182. Lo¨ scher, W., Horstermann, D., 1994. Differential effects of vigabatrin, gamma-acetylenic GABA, aminooxyacetic acid, and valproate on levels of various amino acids in rat brain regions and plasma. Naunyn-Schmiedeberg’s Arch. Pharmacol. 349, 270 – 278. Lo¨ scher, W., Jackel, R., Muller, F., 1989. Anticonvulsant and proconvulsant effects of inhibitors of GABA degradation in the amygdala-kindling model. Eur. J. Pharmacol. 163, 1 – 14. Mathern, G.W., Cifuentes, F., Leite, J.P., Pretorius, J.K., Babb, T.L., 1993. Hippocampal EEG excitability and chronic spontaneous seizures are associated with aberrant synaptic eorganization in the rat intrahippocampal kainate model. Electroencephalogr. Clin. Neurophysiol. 87, 326 – 339. Mathern, G.W., Babb, T.L., Pretorius, J.K., Melendez, M., Levesque, M.F., 1995a. The pathophysiologic relationships between lesion pathology, intracranial ictal EEG onsets, and hippocampal neuron losses in temporal lobe epilepsy. Epilepsy Res. 21, 133 – 147. Mathern, G.W., Pretorius, J.K., Babb, T.L., 1995b. Influence of the type of initial precipitating injury and at what age it occurs on course and outcome in patients with temporal lobe seizures. J. Neurosurg. 82, 220 – 227. McIntyre, D.C., Kelly, M.E., 1989. Pyriform cortex involvement in kindling. Neurosci. Biobehav. Rev. 13, 277 – 280. McIntyre, D.C., Kelly, M.E., 1990. Pyriform cortex involvement in limbic kindling. Soc. Neurosci. Abstr. 16, 1607. McIntyre, D.C., Kelly, M.E., 1993. Are differences in dorsal hippocampal kindling related to amygdala-piriform area excitability? Epilepsy Res. 14, 49 – 61. Miettinen, R., Kotti, T., Tuunanen, J., Toppinen, A., Riekkinen, P., Halonen, T., 1998. Hippocampal damage after injection of kainic acid into the rat entorhinal cortex. Brain Res. 813, 9. Moriwaki, A., Lu, Y.F., Hayashi, Y., Tomizawa, K., Tokuda, M., Itano, T., Hatase, O., Matsui, H., 1996. Immunosuppressant FK506 prevents mossy fiber sprouting induced by kindling stimulation. Neurosci. Res. 25, 191 – 194. Morrisett, R.A., Jope, R.S., Snead, O.C., 1987. Effects of drugs on the initiation and maintenance of status epilepticus induced by administration of pilocarpine to lithiumpretreated rats. Exp. Neurol. 97, 193 – 200. Motte, J., Fernandes, M.J.D., Baram, T.Z., Nehlig, A., 1998. Spatial and temporal evolution of neuronal activation, stress and injury in lithium-pilocarpine seizures in adult rats. Brain Res. 793, 61 – 72. Neal, M.J., Shah, M.A., 1990. Development of tolerance to the effects of vigabatrin (gamma-vinyl-GABA) on GABA release from rat cerebral cortex, spinal cord and retina. Br. J. Pharmacol. 100, 324 – 328. Overton, P., Dean, P., Redgrave, P., 1985. Detection of visual stimuli in far periphery by rats: possible role of superior colliculus. Exp. Brain Res. 59, 559 – 569.

V. Andre´ et al. / Epilepsy Research 47 (2001) 99–117 Paxinos, G., Watson, C., 1986. The Rat Brain in Stereotaxic Coordinates, second ed. Academic Press, San Diego. Petroff, O.A.C., Rothman, D.L., Behar, K.L., Mattson, R.H., 1995. Initial observations on effect of vigabatrin on in vivo 1 H spectroscopic measurements of g-aminobutyric acid, glutamate, and glutamine in human brain. Epilepsia 36, 457– 464. Pitka¨ nen, A., Tuunanen, J., Halone, T., 1996. Vigabatrin and Carbamazepine have different efficacies in the prevention of status epilepticus-induced neuronal damage in the hippocampus and amygdala. Epilepsy Res. 24, 29 – 45. Pitka¨ nen, A., Nissinen, J., Jolkkonen, E., Tuunanen, J., Halonen, T., 1999. Effects of vigabatrin treatment on status epilepticus-induced neuronal damage and mossy fiber sprouting in the rat hippocampus. Epilepsy Res. 33, 67 – 85. Preece, N.E., Cerdan, S., 1996. Metabolic precursors and compartmentation of cerebral GABA in vigabatrin-treated rats. J. Neurochem. 67, 1718 –1725. Qume, M., Fowler, L.J., 1996. Effects of chronic oral treatment with GABA-transaminase inhibitors on the GABA system in brain, liver, kidney, and plasma of the rat. Biochem. Pharmacol. 52, 1355 –1363. Rimvall, K., Martin, D.L., 1994. The level of GAD67 protein is highly sensitive to small increases in intraneuronal gammaaminobutyric acid levels. J. Neurochem. 62, 1375 – 1381. Rundfeldt, C., Lo¨ scher, W., 1992. Development of tolerance to the anticonvulsant effect of vigabatrin in amygdala-kindled rats. Eur. J. Pharmacol. 213, 351 –366. Sahibzada, N., Dean, P., Redgrave, P., 1986. Movements resembling orientation or avoidance elicited by electrical stimulation of the superior colliculus in rats. J. Neurosci. 6, 723 – 733. Scharfman, H.E., 1996. Hyperexcitability of entorhinal cortex and hippocampus after application of aminooxyacetic acid (AOAA) to layer III of the rat medial entorhinal cortex in vitro. Neurophysiology 76, 2986 –3001. Scharfman, H.E., Goodman, J.H., Du, F., Schwarcz, R., 1998. Chronic changes in synaptic responses of entorhinal and hippocampal neurons after amino-oxyacetic acid (AOAA)induced entorhinal cortical neuron loss. Neurophysiology 80, 3031 – 3046. Schwarzer, C., Sperk, G., 1995. Hippocampal granule cells express glutamic acid decarboxylase-67 after limbic seizures in the rat. Neuroscience 69, 705 –709. Sheikh, S.N., Martin, D.L., 1998. Elevation of brain GABA levels with vigabatrin (gamma-vinyl-GABA) differentially affects GAD65 and GAD67 expression in various regions of rat brain. Neurosci. Res. 52, 736 –741. Shin, C., Rigsbee, L.C., McNamara, J.O., 1986. Anti-seizure and anti-epileptogenic effect of gamma-vinyl gammaaminobutyric acid in amygdaloid kindling. Brain Res. 398, 370 – 374. Shuaib, A., Ijaz, S., Hasan, S., Kalra, J., 1992. Gamma-vinyl GABA prevents hippocampal and substantia nigra reticulata damage in repetitive transient forebrain ischemia. Brain Res. 590, 13 – 17. Sloviter, R.S., 1987. Decreased hippocampal inhibition and a

117

selective loss of interneurons in experimental epilepsy. Science 235, 73 – 76. Sloviter, R.S., 1991. Permanently altered hippocampal structure, excitability, and inhibition after experimental status epilepticus in the rat: the ‘dormant basket cell’ hypothesis and its possible relevance to temporal lobe epilepsy. Hippocampus 1, 41 – 66. Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E., Goodman, J.H., Sollas, A.L., Martin, D.L., 1996. Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus. J. Comp. Neurol. 373, 593 – 618. Spiller, A.E., Racine, R.J., 1994. Transfer kindling between sites in the entorhinal cortex-perforant path-dentate gyrus system. Brain Res. 635, 130 – 138. Stone, D.J., Walsh, J., Benes, F.M., 1999. Localization of cells preferentially expressing GAD(67) with negligible GAD(65) transcripts in the rat hippocampus. A double in situ hybridization study. Brain Res. Mol. Brain Res. 71, 201 – 209. Suzuki, Y., Mimaki, T., Arai, H., Okada, S., Kuriyama, K., 1991. Effect of gamma-vinyl gamma-aminobutyric acid on the gamma-aminobutyric acid receptor-coupled chloride ion channel in vesicles from the brain of the rat. Neuropharmacology 30, 423 – 427. Turski, W.A., Cavalheiro, E.A., Schwarz, M., Czuczwar, S.J., Kleinrok, Z., Turski, L., 1983. Limbic seizures produced by pilocarpine in rats: behavioural, electroencephalographic and neuropathological study. Behav. Brain Res. 9, 315 – 335. Turski, L., Cavalheiro, E.A., Czuczwar, S.J., Turski, W.A., Kleinrok, Z., 1987. The seizures induced by pilocarpine: behavioral, electroencephalographic and neuropathological studies in rodents. Pol. J. Pharmacol. Pharm. 39, 545 – 555. Valdizan, E.M., Garcia, A.P., Armijo, J.A., 1999. Time course of the GABAergic effects of vigabatrin: is the time course of brain GABA related to platelet GABA-transaminase inhibition? Epilepsia 40, 1062 – 1069. van den Pol, A., Obrietan, K., Chen, G., 1996. Excitatory actions of GABA after neuronal trauma. J. Neurosci. 16, 4283 – 4292. Vergnes, M., Marescaux, C., Micheletti, G., Reis, J., Depaulis, A., Rumbach, L., Warter, J.M., 1982. Spontaneous paroxysmal electroclinical patterns in rat: a model of generalized non-convulsive epilepsy. Neurosci. Lett. 33, 97 – 101. Westby, G.W., Keay, K.A., Redgrave, P., Dean, P., Bannister, M., 1990. Output pathways from the rat superior colliculus mediating approach and avoidance have different sensory properties. Exp. Brain Res. 81, 626 – 638. Wuarin, J.P., Dudek, F.E., 1996. Electrographic seizures and new recurrent excitatory circuits in the dentate gyrus of hippocampal slices from kainate-treated epileptic rats. J. Neurosci. 16, 4438 – 4448. Ylinen, A.M.A, Miettinen, R., Pitka¨ nen, A., Gulyas, A.I., Freund, T.F., Riekkinen, P.J., 1991. Enhanced GABAergic inhibition preserves hippocampal structure and function in a model of epilepsy. Proc. Natl. Acad. Sci. USA 88, 7650 – 7653.