The present study was approved by the Ethics Committee of Fudan University (Shanghai, China), based on the recommendations established in the Guide for the National Science Council of the Republic of China. The protocols were approved by the Ethics Committee of Fudan University. The approval number from the IRB is ‘20150572A259’. This manuscript was written in accordance with the ARRIVE (Animal Research: Reporting in vivo Experiments) guidelines. Pregnant C57/B6 female mice were housed 1 per cage in a temperature-controlled room (22°C±1°C) under a 12-hour dark/light cycle with free access to water and food in the Animal House of Fudan University. Newborn mice (<24 h) were sacrificed by decapitation, and the brains were removed.

GB and GK were extracted and separated from ginkgo leaf by recrystallization and high-performance liquid chromatography (HPLC) separation and had a purity >98%. GB and GK were dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) before the experiment. Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin were obtained from Gibco (Thermo Fisher Scientific, Inc., Waltham, MA, USA). All other reagents were from Sigma-Aldrich (Merck KGaA) unless otherwise stated.

Primary cortical astrocytes were prepared from newborn mice at postnatal 24 h as previously described with minor modifications (25). Briefly, meninges-free cortices were cut into small cubes (<1 mm³) and digested with 0.25% trypsin at 37°C for 15 min. The suspension of tissue fragments was passed through a 40-µm cell strainer and allowed pre-adherence for 1 h to remove contamination from fibroblasts. The supernatant containing unattached cells was transferred to 25 or 75-cm² flasks (Corning, Inc., Corning, NY, USA) and cultured in an incubator with 5% CO₂ at 37°C. The culture medium was changed with complete DMEM every 3–4 days. When the cultures reached 80–90% confluence, cells were sub-cultured for another 7 days before they were used. The purity of astrocytes in cultures was determined by staining for astrocytic marker glial fibrillary acidic protein (GFAP). More than 95% cells showed GFAP immunoreactivity in the cultures.

For OGD, cells were exposed to 95% nitrogen and 5% CO₂ maintained by constant gas flow at 37°C in all experiments (Forma Anaerobic System; Thermo Fisher Scientific, Inc.). Oxygen tension was maintained at 1–2% during the duration of the test. The media used in these experiments consisted of sugar-free culture medium containing salts, 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin. Control groups were grown under standard culture conditions (5% CO₂ and 95% oxygen).

Astrocytes were exposed to OGD for 3 h followed by re-oxygenation for 24 h. Before re-oxygenation, cells were washed in PBS, and the medium was replaced with complete DMEM containing glucose and 10% fetal bovine serum. At the same time, GB and GK (30 µg/ml, dissolved in DMSO) were added to different cell cultures (ratio of GB/GK: medium=1:1,000). The same volume of DMSO was added as a control.

In the preliminary experiments, we used the 5H-SY5Y/A53T cells to observe the effect of GK with different concentrations (2, 10, 30 and 50 ug/ml). The results showed that only high concentrations (30 and 50 ug/ml) of GK can promote autophagy to degrade alpha-synuclein and induce the BDNF in the 5H-SY5Y/A53T cells. Therefore, in this study, we used 30 ug/ml of GK.

For astrocyte morphology observation, astrocytes were washed in PBS, fixed with 4% paraformaldehyde and stained with the anti-GFAP antibody. The cells were observed under a fluorescence microscope (BX60; Olympus Imaging America Inc., Center Valley, PA, USA).

The cell viability was measured using an 3-(4,5-dimethylthazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Briefly, a total of 10 µl MTT (5 mg/ml in PBS, BDH Chemicals) was added to each 96-well containing 200 µl medium before the conduction of incubation at 37°C for 4 h. The reaction was stopped by the addition of 100 µl DMSO. The optical density (OD) was measured at 570 nm by a Synergy H1 microplate reader (BioTek Instruments, Inc., Winooski, VT, USA), and the results were expressed as an OD value.

Cell cytotoxicity was quantitatively assessed by measuring the activity of LDH released from the damaged cells into the culture medium. At the end of culture, the supernatants that contained detached cells were centrifuged at 2000 rpm for 10 min. The supernatant was then used for the LDH activity assay. The enzyme was quantified by using an assay kit according to the manufacturer's protocol. The absorbance of the samples was read at 440 nm using a microplate reader (Synergy H1; BioTek Instruments, Inc.). The LDH release was expressed as a percentage of experimental vs. 100% damaged cells.

Enzyme-linked immunosorbent assay (ELISA). The quantity of interleukin (IL)-1β, IL-6, IL-10, TNF-α, and PAF released by astrocytes was measured by ELISA. Supernatants from different cultures were collected and centrifuged at 2000 rpm for 10 min to remove cell debris. The concentrations of IL-1β, IL-6, IL-10, TNF-α and PAF in supernatants were measured using commercial immunoassay kits (IL-1β, IL-6, IL-10, and TNF-α from R&D System and PAF; Cloud-Clone Corp., Houston, TX, USA) following the manufacturer's instructions and quantified by reference to standard curves.

Protein from astrocytes was extracted on ice with RIPA buffer plus protease inhibitor and phosphatase inhibitor cocktail (both Thermo Fisher Scientific Inc.). Protein concentration was determined by BCA Protein Assay (Thermo Fisher Scientific Inc.). Protein extracts (30 µg) were separated by SDS-PAGE and transferred onto nitrocellulose membranes (AmershamProtran 0.2 NC; GE Healthcare Life Sciences, Chicago, IL, USA). The membranes were then incubated with anti-p-NF-κB/p65 (Cell Signaling Technology, Inc., Danvers, MA, USA), anti-Nrf2, anti-heme oxygenase-1 (HO-1; both Abcam, Cambridge, MA, USA), anti-Nlrp3, anti-PI3K, anti-p-Akt (all Cell Signaling Technology, Inc.), anti-WNT-1 (Abgent Inc., San Diego, CA, USA), anti-Fzd1 (R&D Systems, Inc., Minneapolis, MN, USA), anti-β-catenin (Cell Signaling Technology, Inc.), anti-BDNF (Abcam), anti-GDNF (Santa Cruz Biotechnology, Inc., Dallas, TX, USA), and anti-β-actin (Cell Signaling Technology, Inc.) antibodies overnight at 4°C. Bands were visualized by HRP-conjugated secondary antibodies and chemiluminescence (ECL) kit (EMD Millipore, Billerica, MA, USA) under a ChemiDoc XRS+ system (Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Cells were fixed for 20 min with 4% paraformaldehyde, blocked with Triton X-100 containing blocking buffer, and then incubated with antibodies against GFAP (Abcam) at 4°C overnight. After the cells were washed with PBS, the secondary antibody Alexa Fluor 555 goat anti-rabbit IgG (Thermo Fisher Scientific Inc.) was added. Immunoreactive cells were observed under fluorescence microscopy (BX60; Olympus Imaging America Inc.) by two investigators.

Data were presented as the mean ± standard error of the mean. The raw data were analyzed by using one-way analysis of variance with Tukey's post hoc test, using GraphPad Prism 5.0 package (GraphPad Software, Inc, La Jolla, CA, USA). P<0.05 was considered to indicate a statistically significant difference.

Fig. 1A illustrates the structure of GB and GK. GB (molecular formula: C20H24O10) is one of the diterpenes that occurs naturally in the leaves of the Ginkgo biloba (25). GK (molecular formula: C20H22O9) is a derivative compound of GB. The morphology of astrocytes stained with GFAP was not abnormal after treatment with OGD, OGD+GB or OGD+GK; however, the number of GFAP-positive cells appeared to be reduced in astrocytes treated with OGD alone (Fig. 1B). The viability of astrocytes exposed to OGD for 3 h declined by 26.8% compared with the viability of astrocytes exposed to normoxic conditions (OD 0.77±0.05 vs. 1.039±0.03), while the viability of OGD astrocytes treated with GB and GK for 24 h was not significantly different from that of astrocytes exposed to OGD alone (Fig. 1C). The LDH release from OGD astrocytes with or without GB/GK treatment was similar to that from cells exposed to normoxia (Fig. 1C). The results demonstrate that the concentration of GB and GK used in this experimental did not influence the viability or death of cultured primary astrocytes.

As shown in Fig. 2, PAF was xulated in cultured primary astrocytes treated with OGD for 3 h (P<0.05). Both GB and GK, as PAF antagonists, effectively inhibited the level of PAF compared with OGD treatment alone (P<0.05 for both).

In this study, both GB and GK inhibited the expression of p-NF-κB/p65 (Fig. 3A) (P<0.001 for both), but neither changed Nlrp3 expression in OGD astrocytes (Fig. 3A) (P>0.05). Compared to OGD astrocytes and GB-treated OGD astrocytes, OGD astrocytes treated with GK released less IL-6 and TNF-α and produced more IL-10 (Fig. 3B) (P<0.001 and P<0.01 for IL-6, respectively, P<0.001 for TNF-α, and P<0.05 for IL-10). These results show that both GB and GK suppressed the expression of p-NF-κB/p65 and that GK more effectively inhibited the production of inflammatory IL-6 and TNF-α in OGD astrocytes. We measured antioxidative Nrf2 and HO-1 expression by western blot analysis. As shown in Fig. 4, although GB increased the expression of Nrf2, GK more efficiently induced the expression of Nrf2 (Fig. 4) (P<0.05 vs. OGD astrocytes treated with GB). Simultaneously, HO-1 expression was significantly elevated in OGD astrocytes treated with GK (Fig. 4) (P<0.05). These results indicate that GK efficiently induced the expression of antioxidative Nrf2 and HO-1 in OGD astrocytes.

We measured antioxidative Nrf2 and HO-1 expression by Western blot. As shown in Fig. 4, although GB increased the expression of Nrf2, GK more efficiently induced the expression of Nrf2 (Fig. 4) (P<0.05 vs. OGD astrocytes treated with GB). Simultaneously, HO-1 expression was significantly elevated in OGD astrocytes treated with GK (Fig. 4) (P<0.05). These results indicate that GK efficiently induced the expression of antioxidative Nrf2 and HO-1 in OGD astrocytes.

Phosphorylated PI3K/Akt pathway serves an anti-inflammatory role through suppressing the phosphorylation of downstream NF-κB (26). Besides, the activation of Nrf2 can be modulated via PI3K/Akt pathway (27). We compared the role of GB and GK on PI3K/Akt pathway of OGD-astrocytes. As shown in Fig. 5, after OGD-astrocytes cells were treated with GB, the expression of PI3K and p-Akt was inhibited compared with OGD-astrocytes alone (for PI3K P<0.05, and for p-AKT P=0.084), although the latter did not reach statistical significance. However, the treatment of GK on OGD-astrocytes showed higher expression of PI3K and p-AKT compared with GB (Fig. 5) (P<0.05, respectively).

The secreted Wnt1 ligand binds to its corresponding receptor, Fzd1, which activates the downstream WNT pathway that regulates neuronal proliferation and differentiation (28) and promotes angiogenesis (29). Compared with OGD treatment alone and OGD treatment with GB, GK treatment induced the expression of WNT-1 (Fig. 6) (P<0.05 and P<0.01, respectively), but did not enhance the expression of either Fzd1 or β-catenin in OGD astrocytes (Fig. 6).

The PI3K/Akt pathway mediates BDNF effects (30), while BDNF promotes the growth of neurons in vitro through crosstalk with the WNT/β-catenin pathway (31). As shown in Fig. 7, GK treatment induced the expression of BDNF (Fig. 6) (P<0.05, respectively) but not GDNF compared with OGD treatment alone and OGD treatment with GB.