The roots of C. lanceolata collected from Heongseong, Gangwon (Korea) was washed in tap water and shade-dried at 20?30°C for 2 days. Dried C. lanceolata was steamed using a steam device (Dechang Stainless, Seoul, Korea) at 90°C for 8 h and repeated five times.

Fermented microorganism, Bifidobacterium longum (KACC 20587), Lactobacillus acidophilus (KACC 12419), and Leuconostoc mesenteroides (KACC 12312) were provided from the Korean Agricultural Culture Collection (Suwon, Korea). The steamed C. lanceolata was aseptically inoculated with approximately 10⁶ CFU/g of Bifidobacterium longum, Lactobacillus acidophilus, and Leuconostoc mesenteroides (1:1:1) in distilled water 8 times the weight of the C. lanceolata and fermented for 48 h at 30°C.

Fermented C. lanceolata was reflux extracted in 70% ethanol with 5 times the weight of the C. lanceolata at 80°C for 24 h. Then C. lanceolata extraction was filtered and concentrated using rotary evaporator. After evaporation, the fermented C. lanceolata sample was obtained by freeze dry.

The mouse hippocampal HT22 cells, a sub-line derived from parent HT4 cells is used to study glutamate-induced cell death mechanisms (Tan et al., 1998b). Hippocampal HT22 cells derived from mouse were obtained from Seoul National University, Korea. HT22 cells in DMEM containing 10% (v/v) fetal bovine serum (FBS) with 1% penicillin/streptomycin, NaHCO₃ (2 mg/ml), and 15 mM HEPES were seeded at a density of 6.7×10⁴/well (48 wells plate), and incubated at 37°C at 5% CO₂. After incubation for 24 h, 10, 100 and 500 μg/ml of SFC and trolox (50 μM) were treated for 1 h before treatment of 2 mM glutamate. Cell viability was accessed using MTT assay as described in our previous study (Weon et al., 2013). After 24 h, cells were treated with MTT solution (1 mg/ml) to each well and then incubated for 3 h. Dimethyl sulfoxide (DMSO) was added in each well and the optical density (OD) was measured 570 nm. Results are expressed as percentages of relative cell viability (%) of non-treated control group. Trolox was used as a positive control.

ROS was determined using 2‘,7‘-dichlorofluorescein diacetate (H₂-DCF-DA). HT22 cells were treated with 2 mM glutamate and 10, 100 and 500 μg/ml of SFC for 8 h. After incubation, the cells were washed with PBS and stained with 10 μM DCF-DA in Hanks’ balanced salt solution in the dark room. After 30 min, cells washed with PBS and then suspend in 1% Triton X-100. Fluorescence was measured at an excitation wavelength of 490 nm and emission wavelength of 525 nm. Accumulation of fluorescent dye, rhodamine 123 (Rho123) was evaluated for the ΔΨ change indirectly by assessing ROS level. HT22 cells stained for 15 min at 37°C with Rho123 and then washed. The Rho123 concentration was measured by spectrofluorometry at excitation wavelength of 488 nm and emission wavelength of 520 nm.

Cytosolic Ca²⁺ concentration in cultured HT22 cells was measured with the Fura-2AM. After treatment of 10, 100 and 500 μg/ml of SFC, 2 μM Fura-AM and glutamate was added to each well. The cells were then washed twice with PBS and extracted with 1% Triton X-100 in PBS for 10 min at 37°C. Fluorescence was analyzed for excitation wavelength of 400 nm and emission wavelength of 535 nm.

The production of NO was determined as nitrite assay using the Griess reaction with 4.25% phosphoric acid, 9.7 μM d-naphtylenediamine and 0.14 μM sulfanillic acid. Optical density was measured at 550 nm.

Cultured cells were washed with 0.2 M phosphate buffer (pH 7.4) and lysed with sulfosalicylic acid. Then, cells were centrifuged at 3000 g for 30 min at 4°C and supernatant was collected. Total glutathione (GSH) in the HT22 cell was investigated based on the reaction of reduced GSH with DTNB (5’,5’-dithiobis (2-nitrobenzoic acid)). To supernatant, 5 unit/ml glutathione disulfide reductase, 0.3 mM NADPH and 0.5 mM DTNB were mixed together. The absorbance was measured at 412 nm within 15 min. Glutathione reductase (GR) was determined by the reduction of glutathione in the presence of NADPH. The reaction mixture contains supernatant, 1 mM glutathione, phosphate buffer containing 0.1 mM NADPH. The decrease in absorbance at 340 nm is immediately measured at an interval time of 15 seconds for a total time of 120 min.

After treatment with glutamate and SFC, HT22 cells were washed with PBS and lysed in lysis buffer (20 mM Tris-HCl,150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10 mM NaF, 2 mM Na₃VO₄ and a protease inhibitor cocktail, pH 7.5) and centrifuged at 13,000×g for 20 min at 4°C. The supernatants were collected and stored at ?80°C. The amount of protein was determined by Bradford method. 20 μg of sample protein was subjected on 13% SDS-PAGE gel. Then, the proteins were transferred to a PVDF membrane (Bio-Rad) and blocked in 5 % skim milk in PBS (0.01% Tween 20) for 1 h at room temperature. The membrane was incubated with diluted primary antibodies, Bax and β-actin at 4°C overnight. The protein bands were immunodetected using secondary antibodies conjugated to horse radish peroxidase. Signals were detected using western blotting detection reagents (Advansta, CA, USA).

To investigate antioxidant activity of SFC, DPPH (1,1-diphenyl-2-picrylhydrazyl) radical scavenging activity was measured. Different concentrations of SFC sample added to 150 μl of 0.4 mM of DPPH solution in 96 well plate. Absorbance of DPPH solution at 517 nm was determined using an ELISA reader and calculated.

We analyzed phenolic compounds, including gallic acid, 4-hydroxybenzoic acid, caffeic acid, vanillic acid, 4-coumaric acid, trans-ferulic acid, and caffeine, in SFC using HPLC. HPLC analysis was performed with an Agilent 1260 series (Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with a diode array detector (DAD). Separation was conducted using a ZORBAX Eclipse XDB-C₁₈ (250×4.60 mm i.d., 5 μm), maintained at 35°C. The mobile phase was composed of 10% acetonitrile with 0.1% formic acid (A) and 0.1% formic acid in 40% acetonitrile and 40% methanol (B). A gradient elution system of the mobile phase as follows: 95% A at 0?15 min; 60% A at 15?33 min; 0% A at 33?42 min; 95% A at 42?50 min. Flow rate was set of 1 ml/min. The UV wavelength was set at 280 nm and filtered sample injection volume was 20 μl.

All results were expressed as mean ± SD. Statistical significance between two groups was analyzed using the t-student test. A value of p<0.05, p<0.01, and p<0.001 were considered to be statistically significant.

Previously, we investigated the effect of steamed and fermented C. lanceolata (SFC) on glutamate-induced cytotoxicity of HT22 cells by using an MTT assay (Weon et al., 2013). Similarly, the treatment of cells with glutamate increased cell death and co-treatment with SFC (500 μg/ml) significantly reduced glutamate-induced cell death to 69.39 ± 3.23% of control (p<0.05) (Fig. 1A). Trolox is an antidoxidant derived from vitamine E, and used as a positive control. As expected, the treatment of glutamate-treated HT22 cells with Trolox significantly reduced cell death. When SFC alone was treated, there are no significant difference between control and 10, 100 and 500 μg/ml of SFC. It means that SFC not affected cell proliferation (Fig. 1B).

We investigated morphological character of HT22 cells. Glutamate induced morphological change as shrunken cells with rounded shape. In comparison, SFC inhibited morphological change and increased cell density (Fig. 1C).

These results indicate that glutamate-induced HT22 cell death may be mediate by oxidative stress and the treatment of cells with SFC showed a neuroprotective effect against glutamate-induced cell death.

Glutamate was known to induce neurotoxicity by increaing intracellular ROS generation in CNS (Choi, 1988). ROS production mediated programmed cell death through oxidative stress (Suzuki et al., 1997). ROS production was also increased in response to Ca²⁺ influx and mitochondrial dysfunction (Wang et al., 1994). Thus, we assessed the intracellular ROS level using the H₂-DCF-DA to check whether SFC can inhibit glutamate-induced ROS production in HT22 cells. The treatment of HT22 cells with glutamate significantly increased the ROS production as compared with HT22 cells treated with vehicle only. Our data show that pretreatment of SFC (500 μg/ml) significantly inhibited the glutamate-induced elevation of ROS level to 115.91 ± 2.49% of control (p<0.01) (Fig. 2A). This result has confirmed that SFC protects neuronal cells from glutamate-induced cell death by inhibition of ROS generation.

In addition, based on the inhibition of ROS generation by SFC, we presumed that SFC exerted protective effect on mitochondrial dysfunction caused by glutamate. Glutamate was known to cause a decrease of mitochondrial membrane potential. Mitochondria have an essential role in life and death decision of neuronal cells. Mitochondria dysfunction induced cell apoptosis and was represented by the loss of mitochondrial membrane potential (Ly et al., 2003). Although glutamate-induced Ca²⁺ influx decreased mitochondrial membrane potential (Duchen, 2000), SFC significantly recovered the mitochondrial membrane potential to 97.88 ± 9.60% of control (p<0.05) at a concentration of 500 μg/ml (Fig. 2B) as judged by Rho123 incorporarion. Thus, our results indirectly indicated that SFC effectively restored mitochondrial membrane potential decreased by glutamate.

It was also well-known that glutamate increased intracellular Ca²⁺ concentration via N-methyl-D-aspartate (NMDA) receptors and neuronal cell death is associated with elevation of Ca²⁺ concentration (Butterfield and Pocernich, 2003). We examined the effect of SFC in glutamate-induced Ca²⁺ influx in HT22 cells. As shown in Fig. 3, exposure to glutamate increased Ca²⁺ concentration to 111.39 ± 0.73% of control, whereas SFC dose-dependently reduced the increased Ca²⁺ concentration compared to glutamate-treated cell. 100 and 500 μg/ml of SFC significantly reduced Ca²⁺ concentration to 96.14 ± 7.07 (p<0.05) and 84.90 ± 1.01% (p<0.001) of control, respectively. This suggests that SFC might decrease Ca²⁺ dependent neuronal cell death induced by glutamate.

Overproduction of NO in brain is mainly induced by neuronal nitric oxide synthase (nNOS) and implicated oxidative cell death. Influx of Ca²⁺ is also contributed to activate NO synthase (Knowles et al., 1989). Glutamate mediated the enhancement of nNOS activity through NMDA receptors (Garthwaite, 1991). In this study, we investigated the effect of SFC on NO production in HT22 cell. The increased NO production (121.29 ± 4.69%) induced by glutamate was significantly decreased by the treatment of SFC. SFC was significantly reduced to 106.41 ± 5.43% (p<0.01), 104.65 ± 6.58% (p<0.05) and 109.19 ± 4.98% (p<0.05) of control by 10, 100 and 500 μg/ml (Fig. 4). These results demonstrated that SFC effectively decreased NO production induced by glutamate treatment.

We confirmed that glutamate-induced cell death in HT22 cells is related to oxidative stress and the treatment of cells with SFC recovered the oxidative stress condition. Since glutathione (GSH) is important antioxidant in CNS and GSH reductase (GR) is a critical enzyme for the production of GSH. High concentration of glutamate leads to deprivation of GSH by inhibiting of cysteine uptake into cells (Albrecht et al., 2010). Depletion of GSH or antioxidant enzyme, such as GR causes neuronal cell death (Kane et al., 1993). We investigated the effect of SFC associated with GSH and GR, and the results of GSH (Fig. 5A) and GR expression levels (Fig. 5B) showed similar pattern. The level of GSH and GR expression in glutamate-injured cells was decreased to 83.68 ± 3.18% and 71.32 ± 3.17%, respectively. However, SFC prevented glutamate-induced depletion of GSH (92.31 ± 9.77% (p<0.05) at 500 μg/ml) and GR (81.52 ± 1.36% (p<0.05) at 500 μg/ml). These results suggest that SFC exerted neuroprotective effect by antioxidant effect.

Overexpression of pro-apoptotic regulators such as Bax and activity of caspase-3 is associated with neuronal cell death by mitochondrial dysfunction (Zhang and Bhavnani, 2005).

To elucidate possible mechanism of neuroprotective effect, we investigated the expression of some proteins related to neuronal cell death, including Bax and caspase-3 on HT22 cells by Western blot analysis. The level of Bax increased to 1.40 ± 0.07-fold compared to control by glutamate, whereas that of SFC reduced to 0.90 ± 0.08-fold (p<0.05) at concentration of 500 μg/ml (Fig. 6A). In addition, glutamate increased caspase-3 levels by 1.17 ± 0.02-fold compared to control (Fig. 6B). 500 μg/ml of SFC (0.59 ± 0.10-fold (p<0.01)) inhibited the up-regulation caspase-3 activity induced by glutamate. These results indicated that SFC reversed the level of Bax and caspase-3 activity increased by glutamate.

DPPH radical scavenging activity was investigated to determine antioxidant activity of SFC. SFC showed DPPH radical scavenging activity in this study. DPPH radical scavenging activity of SFC showed IC₅₀ value as 7160.67 μg/ml.

Contents of 7 phenolic compounds, gallic acid, 4-hydroxybenzoic acid, caffeic acid, vanillic acid, 4-coumaric acid, transferulic acid, and caffeine in SFC were evaluated by HPLCDAD and HPLC chromatogram was shown in Fig. 7A. The result showed that the highest content 5702.21 μg/g of gallic acid existed in SFC. The contents of 4-hydroxybenzoic acid, caffeic acid, vanillic acid, 4-coumaric acid, trans-ferulic acid, and caffeine were 241.27 μg/g, 20.86 μg/g, 210 μg/g, 1.5 μg/g, 369.17 μg/g and 18.75 μg/g, respectively. The data demonstrated that gallic acid was main phenolic compounds in SFC. Among 7 phenolic compounds, contents of gallic acid, vanillic acid and trans-ferulic acid were significantly increased compared to common C. lanceolata extraction (Fig. 7B).