Male C57BL/6 mice (8-week-old, 20~25 g) were obtained from Daehan Biolink (Eumseong, Korea). The mice were housed in a controlled environment and provided with ad libitum food and water. Mice were divided into six groups as follows: i) Group 1: PBS-treated group (control group); ii) Group 2: 200 mg/kg silibinin-treated group; iii) Group 3: KA-treated group; iv) Group 4: KA+50 mg/kg silibinin-treated group; v) Group 5: KA+100 mg/kg silibinin-treated group; vi) Group 6: KA+200 mg/kg silibinin-treated group. All animal experiments were performed in accordance with approved animal protocols and guidelines established by the Animal Care Committee of Kyungpook National University (No. KNU 2016-42).

Injurious seizures were induced by unilateral intrahippocampal KA injection as previously described, with modifications [22]. Briefly, mice were anesthetized using chloral hydrate (360 mg/kg; Sigma, St. Louis, MO, USA), and placed in a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA). Mice received a unilateral injection of KA [0.2 µg in 2 µl phosphate-buffered saline (PBS); Sigma] into the hippocampus (AP: −2.0 mm; ML: −1.2 mm; DV: -1.5 mm, relative to the bregma; Franklin and Paxinos, 2004) using a 10-µl Hamilton syringe needle (30S) attached to a syringe pump (KD Scientific, NewHope, PA, USA). Mice injected with 2 µl PBS in the same manner were used as controls. After the injection, the needle was left in place for an additional 5 min to limit reflux along the injection track.

Following KA injection, mice were monitored for 6 hr to determine the onset time of seizures. Seizure stages was scored according to a modified scale as previously reported [6]: stage 1, facial movement; stage 2, head nodding and myoclonic twitching; stage 3, forelimb clonus with lordotic posture; stage 4, forelimb clonus with reared posture; and stage 5, tonic-clonic seizure without postural control. In the present study, the mice exhibiting at least stage 3 seizures were classified as positive for seizure onset. To examine the effect of silibinin on chronic seizures, mice were monitored to evaluate the frequency of spontaneous recurrent seizures (SRS) for 108 hr/wk (9 hr/d, 6 d/wk) starting from 3 wk after treatment with KA or silibinin and KA, by video recording as previously described, with some modification [19]. Only the mice exhibiting severe seizures (stages 3–5) were considered for evaluation of SRS. The frequency of SRS during the monitoring period was quantified.

Silibinin [(2R,3R)-3,5,7-trihydroxy-2-[(2R,3R)-3-(4-hydroxy-3-methoxyphenyl)-2-(hydroxymethyl)-2,3-dihydro-1,4-benzodioxin-6-yl]-2,3-dihydrochromen-4-one] is the major active component of silymarin, a standardized extract of milk thistle seeds, which is used as hepatoprotective and anti-cancer agents [20,21]. In order to investigate the effects of silibinin in preventing epileptic seizures, it was freshly prepared by suspension in 0.25% carboxymethylcellulose (Sigma-Aldrich) dissolved in 0.9% saline. Experiments were performed according to the scheme presented in Fig. 1A. For behavioral testing, mice were injected intraperitoneally with silibinin 1 day before KA injection, and then injected with silibinin for a further 2 (latency of seizure) or 35 days (SRS). To determine the effects of silibinin on GCD and neurotoxicity, silibinin treatment was initiated 1 day prior to KA injection, and then continued once daily for 7 days. To examine the expression profiles of protein involved in mTORC1 signaling, apoptosis, autophagy and neuroinflammation, silibinin was injected intraperitoneally daily for 3 days, as with the previous protocols, 1 day prior to KA injection.

Seven days after KA treatment, mice were anesthetized using chloral hydrate (360 mg/kg; Sigma) and perfused transcardially with 4% paraformaldehyde (PFA) in 0.1 M PBS. Brains were immediately removed and placed in the same fixative overnight at 4℃. After post-fixation, the brains ware cryoprotected with 30% sucrose in 0.1 M PBS for 24 hr at 4℃. Serial coronal sections of 30 µm thickness were cut on a freezing microtome (Microm International GmbH, Walldorf, Germany). For Nissl staining, the sections were mounted on gelatin-coated slides and stained with 0.5% cresyl violet (Sigma). The presence of GCD was determined by measuring the average width of the GCL in the mid and medial one-fourth portions of the upper blade of the DG [23]. GCD was quantified as the GCL width on the ipsilateral side and expressed as the percentage of the GCL width relative to the contralateral side.

Cryostat sections were used for immunocytochemistry and immunofluorescence staining using a free-floating procedure, as described previously [19]. Briefly, free-floating sections were washed with 0.1 M PBS and blocked with 0.5% BSA in 0.1 M PBS. Primary antibody incubation was performed two overnight at 4℃. The following primary antibodies were used: anti-neuronal nuclei (NeuN, 1:500, Millipore, Temecula, CA) and anti-ionized calcium-binding adapter molecule 1 (Iba1, 1:2000, Wako Pure Chemical Industries, Japan) and anti-phospho-4E-BP1 (p-4E-BP1, a phosphorylated form of the mTORC1 substrate 4E-BP1; 1:1000; Cell Signaling, Beverly, MA, USA). After repeated washes with PBS, the sections were incubated with biotin-conjugated secondary antibodies, followed by an avidin-biotin complex kit (Vector Laboratories, Burlingame, CA). The signal was detected by incubating the sections in 0.5 mg/ml 3,3′-diaminobenzidine (DAB, Sigma) in 0.1 M PBS containing 0.003% H₂O₂.

The sections used for immunofluorescence were incubated overnight with one pair of the following antibodies: anti-neuronal nuclei (NeuN, 1:500, Millipore, Temecula, CA), anti-cleaved caspase-3 (c-caspase-3, 1:400, Cell Signaling), anti-cleaved PARP (c-PARP1, 1:400, Cell Signaling), and anti-microtubule-associated protein 1A/1B-light chain 3 (LC3B, 1:200, Cell Signaling). The sections were then incubated with Texas Red- (1:400, Vector Laboratories, Burlingame, CA) and FITC-conjugated IgG (1:200, Jackson ImmunoResearch, West Grove, PA) and then mounted with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). The stained sections were examined under a microscope (Axio Imager; Carl Zeiss, Gottingen, Germany).

The number of hippocampal CA1 neurons was quantified by staining 30-µm-thick cryosections. Alternate sections were prepared from the coronal brain slice of each animal at 3.3, 3.6, 4.16, and 4.3 mm posterior to the bregma. To maintain consistency across animals, a rectangular box (1×0.05 mm) was centered over the CA1 cell layer beginning 1.5 mm lateral to the midline. Only neurons with visible nuclei were counted by light microscope (Carl Zeiss) at magnification of 400×. The mean number of CA1 neurons in the ipsilateral hippocampus was expressed as a percentage, compared to the contralateral control.

Brain tissues for western blotting were prepared as previously described [19]. DG and CA1 regions of the hippocampus were separated, and each tissue was homogenized, and lysed in buffer containing the protease inhibitor cocktail. Lysates were cleared by centrifugation, protein concentration was determined using a BCA kit (Bio-Rad Laboratories, Hercules, CA, USA). Proteins were separated using gel electrophoresis and transferred onto membranes using an electrophoretic transfer system (Bio-Rad Laboratories). Membranes were incubated overnight at 4℃ with the following antibodies: anti-4E-BP1 (1:1000, Cell Signaling), anti-p-4E-BP1 (1:1000, Cell Signaling), anti-p70 S6 kinase (p70S6K, 1:1000, Cell Signaling), anti-phospho-p70 S6 kinase (p-p70S6K, 1:1000, Cell signaling), anti-β-actin (1:4000, Santa Cruz), anti-LC3B (1:1000, Cell Signaling), anti-c-caspase-3 (1:1000, Cell Signaling), anti-caspase-3 (1:1000, Cell Signaling), anti-tumor necrosis factor alpha (TNF-α, 1:1000, Santa Cruz) and anti-Interleukin 1 beta (IL-1β, 1:1000, Santa Cruz). Membranes were incubated with suitable secondary anti-bodies (Amersham Biosciences, Piscataway, NJ, USA), and blots were developed using enhanced chemiluminescence (ECL) western blotting detection reagents (Amersham Biosciences). For semi-quantitative analyses, band densities were measured using a computerized imaging device and accompanying software (Fuji Film, Tokyo, Japan).

All values are expressed as mean±standard error of the mean (SEM). Multiple comparisons among the groups were performed by one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. All statistical analyses were performed using the SigmaPlot 12.0 software (Systat Software, San Leandro, CA).

Treatment with KA, an analog of the excitotoxic neurotransmitter glutamate, induces seizures which lead to neuronal death and biochemical changes similar to those observed in human patients with TLE [24]. To evaluate the effects of silibinin administration against KA-induced seizures, mice were injected intraperitoneally with silibinin at various concentrations, starting 1 day before KA injection and continued for 2 days. Silibinin treatment delayed the onset of seizure induced by KA treatment in a dose-dependent manner. Particularly, 200 mg/kg of silibinin significantly delayed seizure onset by 177 min, compared to KA alone (Fig. 1B; ^*p<0.01 vs. KA-treated mice). To further examine whether silibinin could suppress the occurrence of chronic spontaneous seizures after KA treatment, treatment with 200 mg/kg silibinin was initiated 1 day prior to KA injection and continued once daily for 35 days. As shown in Fig. 1C, silibinin treatment significantly decreased the frequency of chronic spontaneous seizure induced by KA treatment, compared to KA alone (^*p<0.05 vs. KA-treated mice). These results suggest that silibinin may have a considerably beneficial effect against epileptic seizures in vivo.

TLE can be characterized by morphological changes associated with broadening of the granule cell layer of the dentate gyrus (DG), termed granule cell dispersion (GCD) [2,3]. It is suggested that restriction of GCD may be useful for preventing and inhibiting epileptic seizures. To investigate whether silibinin attenuates the KA-induced morphological changes of GCD in the DG, brain sections from KA-induced epileptic mice, which are untreated or treated with silibinin, were stained with cresyl violet (Fig. 2A). As we reported previously [6,22], KA treatment induced significant morphological changes such as GCD in the DG compared to PBS-treated controls. However, treatment with various concentrations of silibinin resulted in a dose-dependent reduction of GCD (Fig. 2A and B). Quantitative analysis of GCD showed that KA treatment increased GCL width by 130%, compared to PBS-treated controls (Fig. 2B; ^*p<0.05 vs. PBS-treated mice). In particular, silibinin-treatment at a dose of 200 mg/kg dramatically decreased KA-induced GCD by 77%, compared to KA alone (Fig. 2B; ^#p<0.05 vs. KA-treated mice).

mTOR is a major regulator of cellular changes that mediate epileptogenesis [25], and the hyper-activation of mTOR complex 1 (mTORC1) signaling was observed in dispersed granule cells in a mouse model of TLE and in patients with sclerotic hippocampus [4,26]. In order to investigate whether silibinin administration diminishes KA-induced mTORC1 activation in the DG, we examined the effect of silibinin on phosphorylation of the downstream mTORC1 substrates 4E-BP1 (eukaryotic initiation factor 4E-binding protein 1) and p70S6K1 (p70 ribosomal protein S6 kinase 1). As shown in Fig. 2C, expression of p-4E-BP1 in the DG was increased by KA treatment as compared to PBS-treated controls, and this KA-induced mTORC1 activation was inhibited by silibinin treatment. Immunoblotting results were consistent with this, indicating that silibinin treatment caused a significant reduction of p-4E-BP1 and p-p70S6K levels in the DG which were increased upon KA treatment, compared to KA alone (Fig. 2D and E; ^#p<0.01 and ^##p<0.05 vs. KA-treated mice). Moreover, treatment with silibinin alone remarkably reduced the levels of p-4EBP1 in the DG compared to PBS-treated controls (Fig. 2D and E; ^**p<0.01 vs. PBS-treated mice), suggesting that silibinin might have a potent inhibitory effect against mTORC1 activation in the DG following KA administration. Taken together, our findings demonstrate that mTORC1 inhibition by silibinin in the DG may lead to the reduction of GCD.

We next investigated the protective effect of silibinin against KA-induced death of hippocampal neurons. Seven days after KA-treatment with daily intraperitoneal injection of various concentrations of silibinin, the surviving neurons in the CA1 region of the hippocampus were detected by NeuN staining (Fig. 3). As shown in Fig. 3A, NeuN staining indicated an apparent neuronal loss in the CA1 region of hippocampus of KA-treated mice, compared to that of PBS-treated mice. However, silibinin treatment prevented the KA-induced neuronal cell loss in the CA1 region of hippocampus. When quantified and expressed as a percentage of the CA1 neurons in the counting area of the ipsilateral hippocampus relative to the contralateral control, the percentage of surviving neurons in the hippocampal CA1 region was significantly lower in the KA-treated mice (45%, ^*p<0.01 vs. contralateral controls), whereas 91% of the hippocampal neurons were preserved in the silibinin (200 mg/kg)+KA-treated mice (Fig. 3B; ^#p<0.05 vs. KA-treated mice). Silibinin alone did not affect the number of hippocampal neurons compared to contralateral controls (Fig. 3B).

The excitotoxicity mediated by KA could induce neuronal death with apoptotic features in the hippocampus and striatum, resulting from induction of autophagic stress [13,27]. To investigate the protective effects of silibinin against KA-induced apoptosis, we examined the expression of various apoptosis markers, such as cleaved caspase-3 (c-caspase-3) and cleaved PARP-1 (c-PARP-1), in the hippocampus of KA-treated mice at post-lesion 2 days. As shown in Fig. 4A and B, the results of double-immunofluorescence staining demonstrated that numerous c-caspase-3 and c-PARP-1-positive cells were detected in the CA1 of the KA-treated mice, but less c-caspase-3 and c-PARP-1-positive cells were detected in the hippocampus of silibinin-treated mice compared with KA alone. Consistent with the results of double-immunofluorescence staining, western blot analysis revealed that KA treatment induced the significant increases in the levels of c-caspase-3 and c-PARP1 in the hippocampus compared with PBS-treated controls (Fig. 4C and D; ^*p<0.001 and ^**p<0.01 vs. PBS-treated mice), but silibinin administration could significantly inhibit the increases following KA treatment, compared to KA alone (Fig. 4C and D; ^#p<0.01 vs. KA-treated mice).

Microtubule-associated protein 1A/1B-light chain 3 (LC3) is an important mediator and the most reliable marker of autophagy [13]. LC3-II, required for the formation of autophagosome, is the cleaved and lipidated form of the cytosolic LC3-I [28,29]. To clarify the effects of silibinin against KA-induced autophagic stress, the expression levels of LC3-I and LC3-II were measured by double-immunofluorescence staining (Fig. 4E) and western blot analysis (Fig. 4F and G). Immunofluorescence results showed that KA-treatment could induce an increase in the level of LC3B expression in hippocampal CA1 neurons and its overexpression was inhibited by silibinin treatment (Fig. 4E). In addition, immunoblotting results showed that the levels of LC3-II were significantly increased in mice treated with KA compared to that of PBS-treated controls (Fig. 4F and G; ^*p<0.001 vs. PBS-treated mice), but the increased levels were attenuated by silibinin administration (Fig. 4F and G; ^#p<0.001 vs. KA-treated mice).

Recent studies showed that neuroinflammation may play an important role in epilepsy [30]. Moreover, our previous study had reported anti-inflammatory effects of silibinin through inhibition of microglial activation in the animal model of Parkinson's disease [21]. Thus, we further examined whether silibinin contributed to inhibition of microglial activation and reduction of proinflammatory cytokines, such as TNF-α and IL-1β, in the hippocampus of KA-treated mice. Our results showed that KA treatment could induce microglial activation in the hippocampus of mouse brain, and the activation of microglia following KA treatment was inhibited by silibinin treatment (Fig. 5A). Moreover, silibinin treatment significantly inhibited the increase in levels of TNF-α and IL-1β in KA-treated mice, compared to KA alone (Fig. 5B and C; ^#p<0.001 and ^##p<0.05 vs. KA-treated mice). Thus, these results suggest that silibinin can play a role as an important factor for protecting hippocampal neurons against KA-induced neurotoxicity in the adult brain.