GBE was provided by Wanbangde Pharmaceutical Group Co., Ltd. Chemicals such as L-glutamate, MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), DCFH-DA and DMSO were obtained from Sigma-Aldrich; Merck KGaA. Antibody were obtained from Santa Cruz Biotechnology, Inc. or Abcam: p-SRC (ab40660, Abcam), SRC (ab109381, Abcam), p-VAV2 (ab86695, Abcam), Vav2 (ab52640, Abcam), p-p66Shc (ab68166, Abcam), p66Shc (ab33770, Abcam), cytochrome c (ab133504, Abcam), β-actin (sc-58673, Santa Cruz Biotechnology, Inc.), prohibitin (sc-377037, Santa Cruz Biotechnology, Inc.), Goat Anti-Rabbit IgG H&L (HRP) (ab7090, Abcam), Goat Anti-Mouse IgG H&L (HRP) (ab205719, Abcam). Rac1 activity assay kit was obtained from Cell Biolabs. Amplex Red hydrogen peroxide/peroxidase assay kit was obtained from Thermo Fisher Scientific, Inc. Caspase-3 Activity assay kit was obtained from Abcam. Other chemicals and reagents used in this study were obtained from Beyotime.

Human neuroblastoma SH-SY5Y cells were purchased from the American Type Culture Collection (ATCC). Cells were maintained in culture medium of DMEM (Gibco; Thermo Fisher Scientific, Inc.) containing 10% fetal bovine serum (FBS, Gibco; Thermo Fisher Scientific, Inc.) with addition of antibiotics (100 U/ml penicillin and 100 µg/ml streptomycin, Gibco; Thermo Fisher Scientific, Inc.) at 37˚C under 5% CO₂. Cells were treated with compounds (Glutamate: 50 mM or GBE: 25, 50 and 100 µg/ml) or transfected with the gene-specific siRNAs (p66Shc: Sense :5'-AUGAGUCUCUGUCAUCGCUTT-3', antisense: 5'-AGCGAUGACAGAGACUCAUTC-3') using Lipofectamine 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the previous study (16).

MTT assay was used to determine cell viability. Briefly, cells with or without treatment were incubated with 0.5 mg/ml MTT solution (100 µl) for 3 h at 37˚C. The culture medium was then replaced with DMSO (150 µl). The absorbance was recorded at a wavelength of 490 nm using the microstrip reader (Bio-Rad Laboratories).

DCFH-DA staining was adopted to assess intracellular ROS generation. Briefly, cells with or without treatment were incubated with DCFH-DA (10 µM) for 15 min at 37˚C. ROS concentration was analyzed using the fluorescence microscope (Leica Microsystems). Intracellular H₂O₂ level was measured by Amplex Red assay as previously reported (17,18).

NOX activity was determined by luminescence assay using lucigenin as the electron acceptor generated by the NADPH oxidase complex according to the previous study (19). Cells with or without treatment were sonicated in PBS buffer containing 1 mM MgCl₂, 1 mM EGTA and protease inhibitors. The lysates (250 µg/ml of protein) were then incubated with 20 µM lucigenin and 100 µM NADPH. The emitted luminescence was detected by the luminometer (FluoStar Optima, BMG Labtech). Typically, the readings of the luminescence increased linearly within 5 min and the slope of the trend line was defined as the relative NOX activity.

The cytosol and mitochondrial fractions were prepared for cytochrome c release analysis according to the previous study (20). After treatment, mitochondrial and cytosolic fractions were extracted from the cells using Cell Mitochondria Isolation Kit (Beyotime) according to the manufacturer's instructions. The levels of mitochondrial cytochrome c (Mito Cyto c) and cytosolic cytochrome c (Cyto Cyto c) were determined by western blot analysis. β-actin was used as loading control for cytosolic fraction and prohibitin was used as loading control for mitochondrial fraction.

Total protein extract (20 µg) from cultured cells was separated by SDS-PAGE and electrophoretically transferred to polyvinylidene fluoride (PVDF) membrane (Beyotime). Target protein was probed by primary antibody (1:1,000) followed with horseradish peroxidase-conjugated secondary antibody (1:500). The protein signals were detected with chemiluminescence (ECL) detection reagent (Beyotime) and exposed in a dark room. Image J was used to quantify band densities with normalization to that of β-actin or prohibitin.

Caspase activity was assessed using fluorometric assay (Beyotime). Cells with or without treatment were lysed and protein lysate was incubated with caspase-3 substrate DEVD-AFC at 37˚C for 30 min. The caspase activity was assessed with the spectrofluorometer (Molecular Devices) (excitation wavelength at 400 nm, emission wavelength at 505 nm).

Statistical analysis was performed using SPSS 16.0 software package. All data were done in triplicates and the results were from three independent studies. Results of multiple experiments were expressed as means ± SD. Statistical comparisons were made by Student's t-test between two groups and one-way ANOVA followed by Tukey's post hoc test among multiple groups. P<0.05 was accepted as statistically significant.

The SH-SY5Y cells were pre-treatment with or without GBE at a range of concentrations (0, 25, 50, 100 µg/ml) for 24 h, and followed with glutamate incubation (50 mM, 6 h). Cell viability was evaluated using MTT assay. Upon GBE pre-treatment (up to 100 µg/ml), cell viability was not significantly affected, which suggested that GBE is non-toxic to the cells (Fig. 1A). Comparing the groups treated with glutamate, GBE dose-dependently attenuated the glutamate-induced cytotoxicity (Fig. 1B). Such findings demonstrated the protective effect of GBE against glutamate-induced toxicity in SH-SY5Y cells.

The effect of GBE against glutamate-induced oxidative injury in SH-SY5Y cells was further evaluated with the measurement of ROS and H₂O₂. As shown in Fig. 2, exposure of cells to glutamate (50 mM) for 2 h resulted in the significantly upregulated levels of intracellular ROS and H₂O₂, potentiating the oxidative toxicity of glutamate. However, cells pre-treated with GBE (0, 25, 50, 100 µg/ml, 24 h) prior to glutamate incubation resulted in a dose-dependent attenuation of ROS and H₂O₂ generation (Fig. 2). These results indicated that GBE exerted its protective effect against glutamate-induced toxicity by attenuating oxidative stress in SH-SY5Y cells.

To examine the molecular mechanism responsible for the neuroprotective effect of GBE, redoxosome activity was measured. As shown in Fig. 3, glutamate treatment (50 mM, 2 h) induced the upregulated phosphorylation of Src tyrosine and Vav2 tyrosine, which in turn induced the increased expression of active Rac1-GTP, leading to NADPH oxidase (NOX) activation in the Rac1-GTP-dependent manner. However, the incubation of Rac inhibitor (NSC23766, 80 µM, 6 h) or NOX inhibitor (VAS2870, 10 µM, 6 h) potently protected the cells from the cytotoxic effect of glutamate (Fig. 4). Furthermore, the pre-treatment of GBE (0, 25, 50, 100 µg/ml, 24 h) prior to glutamate incubation (50 mM, 6 h) resulted in a dose-dependent attenuation of redoxosome activation.

To investigate whether glutamate-induced redoxosome activation impacts on p66Shc activity, the expressions of p-p66Shc and distribution of p66Shc were measured with or without GBE pre-treatment. As shown in the Fig. 5, glutamate treatment (50 mM, 2 h) induced p66Shc serine36 phosphorylation and mitochondrial translocation of p66Shc. However, Rac inhibitor (NSC23766, 80 µM, 6 h) or NOX inhibitor (VAS2870, 10 µM, 6 h) incubation blocked the effect of glutamate on p66Shc activation. Moreover, cells pre-treated with GBE (0, 25, 50, 100 µg/ml, 24 h) prior to glutamate incubation (50 mM, 2 h) led to a dose-dependent attenuation of p66Shc activation (Fig. 6).

Mitochondria translocation of p66Shc induces mitochondrial dysfunction. Thus, we also assessed the mitochondrial function in cells with or without GBE pre-treatment. As shown in the Fig. 7, glutamate treatment (50 mM, 6 h) resulted in the reduced mitochondrial membrane potential (MMP), increased cytochrome c release and activation of caspase-3. However, p66Shc inhibition blocked glutamate-induced mitochondrial dysfunction. Furthermore, cells pre-treated with GBE (0, 25, 50, 100 µg/ml, 24 h) prior to glutamate incubation (50 mM, 6 h) showed the decreased mitochondrial dysfunction dose-dependently (Fig. 8).