Ginkgo biloba L. leaf extract injection (GBE) was purchased from Youcare Pharmaceutical Group (Beijing, China, Approval No. H20070226, Supporting Information 1). Its initial concentration is 3.5 mg/ml and prepared in aqueous solvent with excipients such as sorbitol, ethanol, sodium hydroxide. Lipopolysaccharide (LPS, #L2880) were obtained from sigma, and prepared in DMEM medium at concentration of 10 mg/ml. Primary antibodies for COX-2 (#4842, 1:1000), IκB-α (L35A5) (#4814, 1:1000), p-IκB-α (14D4) (#2859, 1:1000), NF-κB p65 (D14E12) (#8242, 1:1000) and p-p65 (93H1) (#3033, 1:1000), NF-κB p50 (D4P4D) (#13586, 1:1000) were obtained from Cell Signaling Technology, Inc. (Danvers, MA, United States). Primary antibodies for TNF-α (7B8A11) (#60291-1-Ig, 1:1000), IL-6 (#21865-1-AP, 1:1000), GAPDH (1E6D9) (#60004-1-Ig, 1:1000), β-actin (2D4H5) (#66009-1-Ig, 1:1000) and Lamin B1 (#12987-1-AP, 1:1000) were obtained from Proteintech Group Inc. (Rosemount, IL, United States). Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were obtained from Gibco Laboratories.

Immortalized mouse BV-2 microglial cell line and human U87 glioblastoma cells were cultured in DMEM medium with 10% FBS and 1% penicillin/streptomycin at 37 °C in a humidified atmosphere containing 95% air and 5% CO₂, and the medium was changed every 2 days. The cells were split with 0.25% trypsin when they reached 90% confluence and sub-cultured for further passages. Cells were pretreated with various concentrations of GBE for 1 h, and treated with or without LPS (0.1 µg/ml) for 24 h.

Cell viability was determined using CCK-8 (Cell Counting Kit-8) assay. In brief, 5 × 10³ BV-2 or U87 cells were inoculated on 96-well to adhere overnight. Then, the cells were incubated with a fresh medium containing GBE (5–100 µg/ml) for 1 h, followed by lipopolysaccharide (LPS) (0.1 µg/ml) treatment. After treatment for 24 h, CCK-8 solution (10 μl) was added, and the absorbance was measured at 450 nm using a Biotek Multimode Plate Reader. Cell viability in the vehicle control group was considered to be 100%. Each assay was performed in triplicates.

BV-2 or U87 cells were pre-incubated with various concentrations of GBE for 1 h, followed by LPS (0.1 µg/ml) treatment. After incubation for another 24 h, supernatants were collected and centrifuged at 1500 × g for 10 min at 4°C. TNF-α, IL-1β, and IL-6 productions were assessed in the supernatants using ELISA kits (Elabscience Biotechnology Co., Ltd., Wuhan, China), according to the manufacturer’s instructions.

Protein (30–50 μg) lysates were separated via SDS-PAGE using 10–12% gel, and then transferred to a PVDF membrane. After blocking with 5% nonfat milk, the PVDF membrane was washed and probed with specific primary and secondary antibodies. The protein bands were visualized by enhanced chemiluminescence. The experiments were performed at least thrice.

293T cells in a six-well plate were transfected with pNFκB-luc (Beyotime Biotechnology) plasmid, TRAF6 wild-type, or TRAF6 mutated 3′-UTR, with Lipofectamine 2000 Reagent (Invitrogen, Grand Island, NY, United States). After transfection, the cells were pre-treated with GBE (25 and 50 µg/ml) for 1 h, followed by LPS (0.1 μg/ml) treatment for 24 h. The luciferase activity was measure using a dual-luciferase reporter assay system (Promega, Madison, WI, United States), and normalized to the corresponding Renilla luciferase activity.

MiR-146b-5p mimic (miR-146b) and its control (miR-ctrl) and miR-146b-5p inhibitor (antimiR-146b) and its control (antimiR-ctrl) were obtained from General Biological System (Anhui, China). Transfection was performed using Lipofectamine 2000 when the confluence of BV-2 or U87 cells reached approximately 70%. After transfection for 24 h, the cells were used for further study.

A streptavidin-agarose pulldown assay was performed to detect the binding activity of NF-κB p65/p50 to the COX-2 core promoter probe (−20 nt to −498 nt), which was biotin-labeled. Briefly, 400 μg of nuclear extract proteins, 4 μg of biotinylated DNA probe, and 40 μl of 4% streptavidin-conjugated agarose beads were mixed with appropriate volume of PBSi buffer (PBS buffer with 1 mM EDTA, 1 mM DTT, and protease inhibitor cocktail complete) at room temperature (25°C) for 5 h on a rotating shaker, with a total volume of 400 μl. After washing with PBSi buffer twice, the beads were resuspended in 2X SDS-PAGE loading buffer and boiled for 10 min. The supernatant was analyzed by western blotting.

Immunofluorescence staining was performed in BV-2 cells that were cultured in chamber slides with the indicated GBE and/or LPS treatment. After fixture and permeabilization with 4% paraformaldehyde and 0.2% TritonX-100, the slides were probed with primary against p65 and p50 overnight at 4 °C. Following washing with PBS, samples were incubated with fluorescein isothiocyanate- (#SA00003-1, 1:50) and rhodamine-(# SA00007-2, 1:50) conjugated secondary antibodies for 1 h. Then, the slides were stained with DAPI to counterstain the cell nuclei. The samples were examined under a Leica DM14000B confocal microscope (Leica, Germany).

SPSS 17.0 (IBM Corp., Armonk, NY, United States) was used to perform all the statistical analyses. Unpaired Student’s t-test was used to assess differences between two groups, whereas analysis of variance (ANOVA) followed by the least significant difference test was used for comparisons of multiple groups. Differences were considered statistically significant at p < 0.05.

To investigate the cytotoxic effects of GBE on cell viability, a CCK-8 assay was performed on BV-2 microglial cells. LPS (0.1 μg/ml) had no obvious effects on the viability of BV-2 cells, and the combination with various GBE concentrations up to 100 µg/ml did not change cell viability compared to vehicle only (DMSO)-treated cells (Figure 1A). Similar results were observed for U87 cells (Figure 1B). Therefore, we used doses of 25 and 50 µg/ml GBE in all subsequent experiments.

LPS stimulation can activate microglial cells, accompanied by the release of various inflammatory cytokines, including TNF-α, IL-1β, and IL-6 (Mar et al., 2015; Guo et al., 2016). To evaluate the inhibitory effect of GBE on LPS-induced release of these cytokines, BV-2 cells were pre-incubated with GBE (0, 25, and 50 µg/ml) for 1 h and then stimulated with LPS (0.1 μg/ml) for 24 h. LPS stimulation increased the robust amounts of TNFα, IL-1β, and IL-6 compared to the vehicle control, and this increase was significantly inhibited by GBE in a dose-dependent manner in BV-2 cells (Figure 2A). Similar inhibitory effects were also observed in LPS-induced U87 cells (Figure 2B).

Additionally, quercetin was selected as a positive control drug (Tang et al., 2019), and its cytotoxicity and anti-inflammatory effects were also evaluated. The results showed that quercetin has no obvious effects on the cell viability of both BV-2 and U87 cells, however, it could effectively weaken LPS-induced release of TNFα, IL-1β or IL-6 (Figures 1, 2).

As inflammatory mediators play an important role in chronic inflammation, we examined the effects of GBE on the expression of TNF-α, IL-6, and COX-2 in LPS-stimulated BV-2 cells. The results revealed that GBE treatment significantly suppressed LPS-mediated expression of TNF-α, IL-6, and COX-2 (Figure 3A). Because the expression of these three major inflammatory cytokines is regulated by the transcription factor NF-κB (Zeinali et al., 2017), we examined the effect of GBE on NF-κB activity by a dual-luciferase reporter assay. As shown in Figure 3B, when treated with LPS, luciferase activity was immediately enhanced, and the high activity decreased in a dose-dependent manner with GBE treatment. Furthermore, as NF-κB activation depends on phosphorylation of IκB proteins, we evaluated the effects of GBE on phosphorylation of IκBα and p65. As expected, GBE clearly suppressed LPS-mediated phosphorylation of IκBα and p65 (Figure 3C).

We further assessed the effect of GBE on the binding activity of NF-κB to the COX-2 promoter using a streptavidin-agarose pulldown assay. As shown in Figure 4A, GBE treatment markedly attenuated the binding of NF-κB p50/p65 subunits to the COX-2 promoter DNA probe (Figure 4A) compared with LPS treatment. Moreover, we detected whether GBE suppressed the expression of NF-κB p50/p65 subunits. The results showed that GBE significantly decreased p65/p50 protein levels in nuclear lysates (Figure 4B), but had no obvious effects on their protein expression in LPS-stimulated BV-2 cells (Figure 4C). Based on the above results, we hypothesized that GBE markedly inhibits the translocation of NF-κB p65/p50 subunits from the cell cytoplasm to the nucleus. We next performed the immunofluorescence assay to verify this hypothesis. As expected, GBE markedly inhibited the translocation of NF-κB p65/p50 subunits in BV-2 cells (Figure 4D). The above results indicate that the anti-neuroinflammatory effects of GBE might be mediated by the inhibition of NF-κB translocation from the cytoplasm to the nucleus, which then further inhibits COX-2 expression.

To further investigate the mechanism of suppression of neuroinflammation progression by GBE, we focused on microRNAs that play a vital role in the development of neuroinflammation. After screening differentially expressed miRNAs for both GBE and AD in the gene expression omnibus (GEO) database (GSE44692 and GSE139252) (p < 0.05, log|FC| ≥ 2), we found that miR-146b-5p, miR-31-3p, and miR-495-3p were present in both GBE-treated and AD groups (Figure 5A). Quantitative PCR (qPCR) results showed that miR-146b-5p expression changed significantly in GBE-treated cells (Figure 5B). Therefore, hsa-miR-146b-5p was selected as a potential target gene of GBE for further studies. As shown in Figure 5C, miR-146b-5p overexpression significantly reduced the LPS-induced release of IL-1β and IL-6 but had a minor effect on endogenous expression. Furthermore, luciferase reporter assays also showed that miR-146b-5p overexpression reduced reporter activity driven by the NF-κB promoter (Figure 5D). These results suggest that miR-146b-5p suppresses the inflammatory response.

Next, we examined the effect of GBE on miR-146b-5p expression in LPS-stimulated U87 astrocytes using qPCR. From Figure 6A, the results showed that miR-146b-5p expression was downregulated in the LPS group compared to that in the negative control (NC) group. GBE treatment clearly reversed the expression of miR-146b-5p in LPS-simulated U87 astrocytes in a dose-dependent manner. The results indicate that miR-146b-5p may participate in the regulation of GBE-mediated anti-neuroinflammatory activity. To further examine the effect of miR-146b-5p in GBE-treated astrocytes, a miR-146b-5p inhibitor was employed to exogenously regulate miR-146b-5p expression in LPS-stimulated U87 astrocytes. We found that miR-146b-5p expression was clearly reduced by transfection with the miR-146b-5p inhibitor (antimiR-146-5p), compared to cells transfected with the NC inhibitor (antimiR-control) (Figure 6B). MiR-146b-5p suppression reversed the GBE-induced inflammatory cytokine reduction in LPS-stimulated U87 astrocytes (Figure 6C). Furthermore, the luciferase activity, inhibited by GBE, was also attenuated by miR-146b-5p suppression (Figure 6D). The above data indicate that GBE suppresses neuroinflammation by upregulating miR-146b-5p expression.

To identify the target genes that are regulated by miR-146b-5p, we screened the Starbase and TargetScanHuman databases. In total, 177 differentially expressed genes were identified as candidate targets of miR-146b-5p, as they were observed in both the databases (Figure 7A). After gene ontology analysis of the 177 candidate proteins, an inflammation-related functional cluster of “activation of NF-κB-inducing kinase activity” was identified, mainly including CARD10, IRAK1, and TRAF6 (Figure 7B). This finding was consistent with the experimental results mentioned above. qPCR analysis further revealed that TRAF6, an upstream regulator of NF-κB, showed the greatest change in expression with miR-146b-5p treatment (Figure 7C).

The complementary sites between miR-146b-5p and TRAF6 predicted by the TargetScanHuman database (Figure 8A) share high homology between mice and humans. Therefore, we investigated the association between miR-146b-5p expression and TRAF6 expression. A dual-luciferase reporter assay was performed to validate whether miR-146b-5p could bind to the 3′-UTR of TRAF6. As shown in Figure 8B, the luciferase activity of TRAF6-wild type (TRAF6-wt) was markedly decreased in the miR-146b-5p group and increased in the anti-miR-146b group compared with the miR-control group. Additionally, the inhibitory effect of miR-146b-5p on TRAF luciferase was abrogated in the mutation group compared to that in the wild-type group. Furthermore, miR-146b-5p overexpression reduced TRAF6 expression at both the mRNA and protein levels in LPS-induced U87 astrocytes compared to that in control cells (Figures 8C,D). In contrast, miR-146b-5p inhibitor stimulation increased the mRNA and protein levels of TRAF6 (Figures 8C,D), confirming that miR-146b-5p negatively regulates the expression of TRAF6. These results support that miR-146b-5p could directly bind to the 3′-UTR of TRAF6 and suppress its expression.

To further investigate the molecular mechanism of GBE function in TRAF6 expression, the effect of different concentrations of GBE on miR-146b-5p expression was evaluated. As shown in Figures 8E,F, GBE inhibited TRAF6 expression at both the mRNA and protein levels in LPS-stimulated U87 astrocytes. This inhibition was reversed by miR-146b-5p suppression, implying that GBE decreases TRAF6 levels in a miR-146b-5p-dependent manner.