Male BALB/c mice (25-30 g) were housed under diurnal lighting conditions and allowed food and tap water ad libitum. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The animal protocol used in this study has been reviewed by the INHA University-Institutional Animal Care and Use Committee (INHA-IACUC) on their ethical procedures and scientific care, and it has been approved (Approval Number INHA-110321-81). Animals were randomly assigned to a KA-treated, KA and GL-treated, or control group. At the start of the experiment, animals weighed 25-30 g and were 10 weeks old.

Intracerebroventricular (i.c.v.) injection of KA into brain was previously described (Cho et al., 2003). Briefly, male BALB/c mice (25-30 g) were anaesthetized by the intraperitoneal (i.p.) injection of a 2:1 mixture (3.5 µl/g body weight) of ketamine (50 mg/ml) and xylazine hydrochloride (23.3 mg/ml), and then placed on a stereotaxic apparatus (Stoelting Co, Wood Dale, IL). Four µl of saline solution containing KA (0.2 µg) was then injected into the right lateral ventricle (stereotaxic coordinates in mm with reference to the bregma were AP, -2.0; ML, -2.9; DV, -3.7) [17] of mice at 0.5 µl/min. After 5 min, the needle was removed over 3 min to minimize backflow. Mice were kept on a warm pad until awake.

Mice were observed for 150 min after the KA injection and scored using the following scale (Sperk et al., 1985) [18]: (+1) arrest of motion; (+2) myoclonic jerk of head and neck with brief twitching movements; (+3) unilateral clonic activity, frequent focal convulsions, salivation; (+4) bilateral forelimb tonic and clonic activity, frequent focal convulsions; and (+5) continuous generalized limbic seizures with loss of postural tone, and death within 2 hrs.

GL (Sigma, St. Louis, MO) was dissolved in 0.89% NaCl and administered intraperitoneal (i.p.). In total 300 µl of GL-containing solution was injected 30 min before KA (0.2 µg, i.c.v.) treatment.

BALB/c mice were sacrificed by cervical dislocation. Then the hippocampus CA1 and CA3 were removed quickly and placed in ice-cold RIPA buffer (50 mM Tris-HCl (pH7.4), 1% NP-40, 0.25% sodium-deoxycholate, 150 mM NaCl, and a complete mini protease inhibitor cocktail tablet (Roche, Basel, Switzerland)). After homogenization,lysates were centrifuged at 14,000 rpm at 4℃ for 15 min and supernatant liquors were frozen at -20℃.

Mice brains were fixed with 4% paraformaldehyde by transcardiac perfusion and post-fixed in the same solution overnight at 4℃. Fixed brain was incubated with 30% sucrose overnight and sections (30 µm) were prepared by using a vibratome. Immunological staining was performed following standard procedure. Primary antibodies were diluted 1:200 for anti-Neu N antibody (MAB377, Chemicon, Temecula, CA) and anti-ionized calcium binding adaptor molecule-1 (Iba-1) (Wako Pure Chemicals, Osaka, Japan), and 1:150 for anti-GFAP antibody (DB Bioscience, San Jose, CA). After washing with PBS containing 0.1% Triton X-100, sections were incubated with anti-mouse IgG (Vector Laboratories, Burlingame, CA) for anti-Neu N and anti-GFAP, or with anti-rabbit IgG (Vector Laboratories, Burlingame, CA) for anti-Iba-1 in PBS for 1hr at room temperature and visualized using the HRP/3,3'-diaminobenzidine (DAB) system. Numbers of NeuN-positive cells in 0.01 mm² (0.1×0.1 mm²) for CA1 or in 0.0225 mm² (0.15×0.15 mm²) for CA3 areas were obtained by counting 12 photographs, 3 photographs per experiment. The representative pictures were presented from three independent experiments.

Fifty µg of proteins were separated by 12% sodium dodecyl sulfate polyacrylamide gel. After blocking with 5% non-fat milk for 1 hr, membranes were incubated with primary antibodies diluted 1:1,000 for anti-Cox-2 (Santa Cruz Biotechnology, Inc, Delaware, CA), anti-iNOS (Abcam, Cambridge, MA), anti-IL-1β (Merck Millipore, Billerica, MA), and anti-α-tubulin (Cell Signaling, Denvers, MA) overnight at 4℃. The next day, membranes were detected using a chemiluminescence kit (Roche, Basel, Switzerland) using anti-rabbit HRP-conjugated secondary antibody (1:2,000, Santa Cruz Biotechnology).

Primary microglial cultures were prepared as previously described [19]. In brief, cells dissociated from the cerebral hemispheres of 2- to 13-day-old postnatal rat brains Sprague-Dawley strain) were seeded at a density of 1.2×10⁶ cells/ml in Dulbecco's modified Eagle's medium (DMEM; Gibco, Carlsbad, CA) containing 10% FBS (Hyclone, Logan, UT) and 1% penicillin-streptomycin (Gibco, Carsbad, CA). After 2 weeks, microglia were then detached from the flask by mild shaking and filtered through a nylon mesh to remove astrocytes. After centrifugation (1,000×g) for 5 min, the cells were resuspended in a fresh DMEM supplemented with 5% FBS, and plated at a final density of 1×10⁵ cells/well on a 24 multi-well culture plate. On the following day, cells were subjected to various treatments. BV2 cells were grown in DMEM supplemented with 1% penicillin, streptomycin and 5% FBS.

Primary microglial cultures plated on 24-well plates (1×10⁵ cells/well) were treated with lipopolysaccharide (LPS; 200 ng/ml) for 24 hrs. To measure the amount of NO produced by microglia, 100 µl of the conditioned medium was mixed with an equal volume of Griess reagent (0.5% sulfanilamide and 0.05% N-1-naphthylethylenediamine), and incubated for 10 min at room temperature. Absorbances were measured at 550 nm using a microplate reader.

Primary cortical cultures, including astrocytes and neurons, were prepared from embryonic day 15.5 mouse cortices and cultured as described previously by Kim et al. [20]. Dissociated cortical cells were plated at a density of approximately 4×10⁵ cells per well (five hemispheres per 24-well poly(d-lysine)- and laminin-coated plate). Cultures were maintained without antibiotics in MEM containing 5% horse serum, 5% fetal bovine serum, 2 mM glutamine, and 21 mM glucose. At day 7 in vitro (DIV 7), when astrocytes had reached confluence underneath neurons, cytosine arabinofuranoside was added to a final concentration of 10 µM, and cultures were maintained for 2 days to halt microglial growth. Fetal bovine serum and glutamine were not supplemented from day 7, and media were changed every other day after day 7. Cultures were used at DIV 12 to 14.

Primary cortical cells were treated with serum-free MEM or HEPES controlled salt solution (HCSS) containing 50 µM NMDA (Sigma, St. Louis, MO) for 10 min, 100 µM KA (Sigma, St. Louis, MO) for 12 hrs, or 100 µM glutamate (Sigma, St. Louis, MO) for 1 hr. The medium was then removed and replaced with fresh MEM, and cells were cultured for 24 hrs.

Twenty four hrs after treating cells with NMDA (50 µM, 10 min), KA (100 µM, 12 hrs), or glutamate (100 µM, 1 hr) 50 µl aliquots of media and 50 µl of LDH assay reagent (Roche, Mannheim, Germany) were mixed in a 96-well plate and incubated for 15 min. Optical densities were measured using a 96-well plate reader at 490 nm.

Statistical analysis was performed by analysis of variance (ANOVA) followed by the Newman-Keuls test. All data are presented as means±SEMs and statistical difference was accepted at the 5% level.

KA (0.2 µg, i.c.v.)-injected mice exhibited characteristic epileptic behavior as early as 5 min after treatment. Seizure progression from 0 to 150 min after KA administration was scored using Sperk's seizure scale [18]. These fixed epileptic behaviors intensified rapidly and the highest level of reaction was observed 90 min after KA administration (Fig. 1). To examine the neuroprotective effect of GL, it was administered at 30 min prior to KA (10 or 50 mg/kg, i.p.) and seizure behavior was assessed at the same time points as in KA-administered control animals. Although the severity and duration of seizure behavior were generally lower in the GL-administered group than in KA-administered controls, differences were not statistically significant (Fig. 1).

It has been known that KA administration leads to neuronal death in CA1 and CA3 regions of the hippocampus. To determine degrees of neuronal death, brain tissue slices containing hippocampus were stained with anti-NeuN antibody at 2 days after KA treatment. In saline-treated control animals, mean numbers of NeuN⁺ cells in the CA1 (0.1×0.1 mm²) and CA3 (0.15×0.15 mm²) regions (indicated as black box in Fig. 2A) of the hippocampus were 99.3±18.7 (n=12) and 178.7±9.1 (n=12), respectively (Fig. 2A, E, F) and in the KA-administered controls, mean numbers of NeuN⁺ cells were 42.9±6.9 (n=12) and 51.7±9.1 (n=12), respectively (Fig. 2B, E, F). When 10 mg/kg of GL was administered 30 min prior to KA, KA-induced neuronal death was significantly suppressed, and mean numbers of NeuN⁺ cells were 53.5±9.4 (n=12) and 97.2±9.7 (n=12), respectively (Fig. 2C, D, F). KA-induced neuronal death was further suppressed by the administration of 50 mg/kg of GL and corresponding mean numbers of NeuN⁺ cells were 63.0±8.9 (n=12) and 140.8±13.5 (n=12) (Fig. 2D, E, F). When GL (50 mg/kg) was administered 30 min after KA injection, KA-induced neuronal death was also significantly suppressed although the effects appeared to be slightly weaker than that observed in 30 min pre-administered animals (data not shown). These results indicated that GL suppressed KA-induced neuronal death in the mouse hippocampus.

We examined whether GL affected inflammatory processes in KA-injected mice. Anti-GFAP and anti-Iba-1 antibodies were used to access astrocyte and microglial activations, respectively. In saline-treated control animals, immunostaining with anti-GFAP antibody revealed that astrocyte cell bodies were small and their processes were long and thin (Fig. 3A). However, at 2 days after KA injection, astrocyte cell bodies were enlarged and their processes were shorter and thicker, that is, they possessed the characteristics of activating astrocytes (Fig. 3B). At 4 days after KA administration, astrocytes were activated further and exhibited the characteristic star-shape of activated astrocytes (Fig. 3C). However, when 10 mg/kg GL was injected 30 min prior to KA, astrocyte activation at 4 days after KA administration was remarkably suppressed (Fig. 3D) and it was further suppressed by 50 mg/kg GL, adopting morphology of the resting state (Fig. 3E). Staining with anti-Iba-1 antibody (a marker of cells of myeloid origin [21]) showed that in saline-treated control animals, Iba-1 -positive cells in the hippocampus were ramified (Fig. 3F). However, 2 days after KA injection, cell bodies of Iba-1-positive cells were enlarged and cytoplasmic processes were thicker and shorter in CA1 and CA3, displaying activated morphology of microglia (Fig. 3G). In contrast to astrocyte activation, staining with anti-Iba-1 antibody detected fewer number of Iba-1-positive cells exhibiting activated morphology at 4 days after KA administration (Fig. 3H). In 10 mg/kg GL-administered animals, number of Iba-1-positive cells exhibiting activated morphology was significantly suppressed at 2 days post-KA (Fig. 3I) and it was almost completely suppressed by 50 mg/kg GL, wherein cell bodies of Iba-1-positive cells were thin and their cytoplasmic processes were ramified (Fig. 3J). These results suggest that the activations of microglia/macrophage and astrocytes in the hippocampus of KA-administered animals were suppressed by GL.

Next, we examined whether the inhibition of astrocyte and microglia activation by GL is accompanied by the suppression of proinflammatory marker production in the hippocampus. Levels of proinflammatory markers in KA-injected animals were assessed by immunoblotting with anti-COX-2, anti-iNOS, and anti-IL-1β antibodies. Protein samples were prepared from both CA1 and CA3 hippocampal regions (indicated in Fig. 4A) one day after KA injection. COX-2 expression was notably increased both in CA1 and CA3 (Fig. 4B). After treatment with 10 mg/kg of GL, COX-2 expression was significantly suppressed and it was further suppressed by 50 mg/kg GL (Fig. 4B). iNOS and IL-1β expressions were also notably increased in KA-injected hippocampi and suppressed dose-dependently by GL (Fig. 4C, D). These results indicate that KA-induced proinflammatory marker productions were suppressed by GL.

To investigate the anti-inflammatory effect of GL, the GL-dependent suppression of microglial activation was examined in primary microglial cultures. Cells were stimulated with LPS (0.1 µg/ml) for 24 hrs with and without GL co-treatment, and nitrite production was measured. GL co-treatment reduced LPS-induced nitrite production dose-dependently; maximum inhibition was achieved at a GL concentration of 1 mM (Fig. 5A). In addition, GL dose-dependently repressed the inductions of proinflammatory markers (Cox-2, TNF-α, and iNOS) in LPS-treated primary microglial cultures (Fig. 5B-E). In contrast, α-tubulin levels were not changed at any GL concentration tested (Fig. 5B). These results indicate that GL suppresses LPS-induced microglial activation.

GL-mediated suppression of neuronal death in KA-administered animal prompted us to examine an anti-excitotoxic effect of GL in primary cortical cultures. Since glutamate-mediated excitotoxic cell death is the main neurotoxic mechanism in epilepsy [22], NMDA, glutamate, and KA were used to induce excitotoxic damage in primary cortical cultures and to examine the anti-excitotoxic effects of GL. Severe neuronal damage was observed 2 hrs after NMDA treatment (50 µM, 10 min) in primary cortical cultures (Fig. 6A). However, co-treatment with 0.5 mM of GL significantly reduced LDH release, and treatment with 1 mM of GL further reduced LDH release (Fig. 6A). Similar neuroprotective effects were also observed in KA (500 µM, 12 hrs)- or glutamate (100 µM, 1 hr)-treated primary cortical cultures (Figs. 6B, C), although, the protective efficacy of GL differed for the different treatments (Fig. 6). Together these results indicate that anti-excitotoxic effects in neurons contributes to the neuroprotective effects of GL in the postischemic brain.