Primary antibodies against p65, JNK, phospho-JNK, NFATc1, Phalloidin Cruz Fluor™ 594 conjugate, and c-Fos were obtained from Santa Cruz (CA, USA). Primary antibodies against phospho-p65 were obtained from Abcam (Cambridge, UK). Primary antibodies against phospho-p38, p38, and IĸB-α were purchased from ImmunoWay (Plano, TX, USA). Other antibodies were acquired from Cell Signaling Technology (Danvers, MA, USA). The RNeasy Mini Kit was obtained from QIAGEN (Frankfurt, German). The TRAP staining kit was purchased from Sigma-Aldrich (St. Louis, MO, USA). Fetal bovine serum (FBS) and α-MEM were obtained from Gibco (Rockford, IL, USA). ANE (purity > 98%; Figure 1A ) was purchased from Shanghai Pureone Biotechnology (Shanghai, China), and dissolved in dimethyl sulfoxide (DMSO) followed by dilution with α-MEM medium. RANKL and M-CSF were procured from PeproTech (Rocky Hill, NJ, USA).

Bone marrow cells were isolated from the femora and tibiae of 6-week-old male C57BL/6J mice, as previously described (Xu et al., 2016; Zhang et al., 2018). Extracted cells were cultured in α-MEM medium supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin in the presence of 40 ng/ml M-CSF at 37°C with 5% CO₂. Non-adherent cells were collected the next day and incubated with α-MEM complete medium containing 40 ng/ml M-CSF for 2 days. Then, the cells were collected as bone marrow-derived macrophages (BMMs).

BMMs (6 × 10⁴ cells/well) were treated with or without 100 ng/ml RANKL and ANE (0–10 μM) in a 48-well plate for 4 days in the presence of M-CSF (40 ng/ml). α-MEM medium contains 0.1% DMSO. Then, cells were fixed and stained to detect TRAP activity as the TRAP staining kit’s protocol. TRAP-positive multinucleated cells (≥3 nuclei) osteoclasts were counted as osteoclasts.

Briefly, BMMs (10⁴ cells/well) were added to 96-well plates in the absence or presence of 40 ng/ml M-CSF, 100 ng/ml RANKL and ANE (0-10 μM) for 48 h at 37°C in 5% CO₂. The CCK-8 assay (Beyotime Institute of Biotechnology, Shanghai, China) was performed according to the manufacturer’s instructions and read at 450 nm using a microplate reader (Tecan, San Jose, CA, USA).

BMMs (2 × 10⁵ cells/well) were added to 24-well Osteo Assay Surface Plates (Corning, Tewksbury, MA, USA) and treated with or without RANKL (100 ng/ml) and ANE (0-10 µM) in the presence of M-CSF (40 ng/ml) for 4 days. Next, the cells were removed using 5% (w/v) sodium hypochlorite for 10 min and then stained using the modified von Kossa method (Xu et al., 2016; Zhang et al., 2018). 5% (w/v) silver nitrate was incubated in the wells avoiding light for 1 h at room temperature. And then 5% (w/v) sodium carbonate prepared in 10% formaldehyde was added to the wells, treated for 5 min at room temperature. Resorption pit area per well was imaged using a light microscope and counted by ImageJ software.

BMMs (6 × 10⁴ cells/well) were added to a 48-well plate and cultured with or without RANKL (100 ng/ml) and ANE (0-10 μM) in the presence of M-CSF (40 ng/ml) for 4 days. Cells were fixed with 4% paraformaldehyde for 30 min and permeabilized with 0.1% Triton X-100 in PBS for 10 min. Next, the cells were blocked with 5% BSA in PBS for 1 h and incubated with phalloidin for 1 h at 37°C. After washing with PBS, the cells were stained with DAPI and imaged using an Olympus BX35 fluorescent microscope.

BMMs (4 × 10⁵ cells/well) were added to 12-well plates in complete α-MEM medium, and cultured with or without RANKL (100 ng/ml) and ANE (10 μM) in the presence of M-CSF (40 ng/ml) for 4 days. Total RNA was extracted from cultured cells using the RNeasy Mini Kit and cDNA was synthesized using the Transcriptor First Strand cDNA Synthesis Kit (Roche, Basel, German). Real-Time PCR was performed using an ABI 7500 Fast Real-Time PCR System (Foster City, CA, USA). The primers used for the tested genes are detailed in Table 1 ; GAPDH was employed as a reference gene.

Cultured cells were lysed in RIPA buffer with 1% protease and phosphatase cocktail inhibitor (Sigma-Aldrich). Proteins (30–50 μg) were quantified using a bicinchoninic acid (BCA) assay kit, separated by 10% SDS-PAGE, and transferred to PVDF membranes. The membranes were blocked in 5% BSA for 2 h and then incubated with the indicated primary antibodies for approximately 12 h at 4°C. Next, the membranes were incubated with the corresponding secondary antibody. The membranes were imaged using the MicroChemi Chemiluminescence system (DNR, Jerusalem, Israel). Densitometry analysis of protein bands was performed using ImageJ software.

Next, this study evaluated the therapeutic potential of ANE for inflammatory bone lysis disease in vivo. This project fully considered and protected the rights and interests of the study objects and was based on the principle of the Institutional Animal Ethics Committee of Jilin University. Healthy 6-week-old male C57BL/6J mice were separated into the following four groups (10 mice per group): sham group (saline), LPS group (only LPS and saline injected), low dose ANE group (LPS injected and 2 mg/kg ANE), high dose ANE group (LPS injected and 10 mg/kg ANE). All mice were subcutaneously intraperitoneally injected with saline and/or ANE was administered at one day prior to LPS (10 mg/kg) treatment, repeated every other day and continued for 10 days. The animals were then euthanized and the tibiae, femora, and blood were collected for further analysis. The fixed left femur sections were decalcified in 12% EDTA and embedded in paraffin for H&E and TRAP staining to examine bone erosion and osteoclast activity. The right femora were scanned through a high-resolution microcomputed tomography 35 µCT scanner (Scanco Medical AG, Bassersdorf, Switzerland) at an isotropic voxel size of 10 μm. The region of interest (ROI) is 1 mm from the growth plate. The bone parameters analyzed using Scanco Medical software included bone volume per tissue volume (BV/TV), trabecular separation (Tb.Sp), trabecular thickness (Tb.Th), and trabecular number (Tb.N.).

Data were expressed as mean ± SD of at least triplicate independent experiments. Statistical analyses were compared using a Student’s t-test or one way ANOVA using GraphPad Prism 7.0 (GraphPad Software, La Jolla, CA, USA). Results with p < 0.05 were considered statistically significant.

First, the effect of ANE on osteoclast formation was assessed. ANE treatment decreased the number and size of TRAP-positive osteoclasts in a dose-dependent manner ( Figure 1B, C ). Next, to exclude the possibility cytotoxic effect of ANE on RANKL-induced osteoclasts, the viability of cells treated with different concentrations of ANE (0–10 μM) for 48 h was measured using the CCK-8 assay. As shown in Figure 1D , ANE had no obvious effect on the cell viability of BMMs at the indicated concentration. These findings demonstrate that ANE treatment affect the osteoclast formation without any cytotoxicity.

Osteoclasts reorganize the actin cytoskeleton to adhere to mineralized bone surface and to resorb the bone (Garbe et al., 2012). Next, this study examined whether ANE could suppress F-actin ring formation and bone resorption function. As expected, ANE treatment markedly decreased osteoclast-mediated bone resorption pits in a dose-dependent manner ( Figures 2A, B ). Addition, ANE treatment dose-dependently decreased the quantity of F-actin rings and treated cells contained fewer nuclei ( Figures 2C, D ). Taken together, ANE treatment inhibited RANKL-induced osteoclast differentiation and decreased osteoclast size and fusion without obvious cytotoxicity. In addition, it reduced the bone-resorbing function of osteoclasts by impairing F-actin cytoskeleton formation and osteoclastic absorption ability. These results suggest that ANE decreases osteoclast-mediated bone formation and resorption.

Thereafter, this study investigated whether ANE regulates the mRNA and protein expression of NFATc1 and c-Fos. As shown in Figure 3A , ANE repressed the mRNA levels of NFATc1 and c-Fos in RANKL-treated BMMs. Similarly, ANE lowered RANKL-stimulated NFATc1 and c-Fos protein level expression ( Figure 3B ). These results suggest the anti-osteoclastogenic effect of ANE may contribute to the inhibition of NFATc1 signaling.

NFATc1 is a master transcription factor for osteoclastogenesis (Asagiri et al., 2005; Kim and Kim, 2014), which controls the transcription of osteoclastogenesis-essential genes, including ATPase H+-transporting V0 subunit d2 (Atp6v0d2), dendritic cell specific transmembrane protein (DC-STAMP) and Integrin αv. Thus, the study evaluated the expression of these genes by Real-Time PCR. As shown in Figure 4 , after 24, 48, 72, or 96 h treatment with ANE, the gene expression of DC-STAMP, Integrin αv, and Atp6v0d2 strikingly decreased. These genes are involved in precursor cell fusion and bone resorption (Lee et al., 2006; Chiu and Ritchlin, 2016). Atp6v0d2 and DC-STAMP have been shown to be crucial for cell-to-cell fusion, which is critical for reorganizing the actin cytoskeleton during osteoclast differentiation (Yagi et al., 2005; Kim et al., 2008). Integrin αvβ3 is a transmembrane integrin involved in osteoclast attachment on bone, mediating the downward signaling pathways and subsequent bone resorption (Engleman et al., 1997; Crotti et al., 2006). Taken together, this study provides evidence that ANE might exert its anti-osteoclastogenesis effects via inhibition of these NFATc1 responsive genes.

As noted in many studies, NF-κB signaling activation is important for the initial induction of NFATc1, ultimately resulting in RANKL-induced osteoclast differentiation (Ono and Nakashima, 2018). To investigate whether ANE affected NFATc1 induction through NF-κB signaling, this study next examined the influence of ANE on RANKL-induced NF-κB activation by western blotting analysis. BMMs were pretreated with different concentrations of ANE for 1 h and then incubated with RANKL for the indicated time. RANKL-mediated p65 phosphorylation was considerably decreased by ANE treatment at 20 min of RANKL stimulation, while IκB-α degradation and phosphorylation was relatively unaffected in the ANE group ( Figure 5A ). Additionally, ANE pretreatment strongly reduced p65 phosphorylation in a concentration-dependent manner at 20 min RANKL stimulation ( Figure 5B ). These findings demonstrate that ANE treatment reduced RANKL-induced NF-κB activation through IκB-α independent manner.

MAPK signaling which consists of three major members p38, ERK1/2, and JNK, plays a crucial role in osteoclast differentiation, have shown to be activated in RANKL-induced osteoclastogenesis (Nakashima et al., 2012; Choi et al., 2017). To investigate whether ANE affects the MAPK pathways, BMMs pretreated with the indicated concentration of ANE for 1 h and then stimulated with RANKL for different times (0 to 30 min). The expression of phosphorylated ERK1/2 was significantly reduced compared with total ERK1/2 by ANE treatment at 20 min of RANKL stimulation. However, the relative expression of phosphorylated p38 and JNK was relatively unaffected in the ANE group ( Figure 6 ). Previous studies also reported that MSK1, a downstream kinase of ERK1/2, positively regulates p65 signaling pathways (Schmitz et al., 2001; Vermeulen et al., 2003). To investigate whether MSK1 is involved in mediating the phosphorylation of p65 in the ANE-treated osteoclasts, this study checked the phosphorylation of MSK-1. As shown in Figure 6 , the phosphorylation of MSK-1 was significantly reduced compared with total MSK-1 at 20 min of ANE treatment. These data suggest that ANE inhibits RANKL-mediated activation of the ERK1/2 signaling pathway without changing JNK and p38 signaling pathways.

In addition, the study provides insights into the effect of ANE on the regulation of NFATc1 in osteoclasts. To investigate whether ANE can modulate NFTAc1 phosphorylation, this study examined the effect of ANE on the activation of GSK-3β and PLCγ2. As shown in Figure 7 , RANKL induced inactivation of GSK-3β and activation PLCγ2, which were not affected by ANE treatment. Taken together, these results suggest that ANE treatment might not affect the phosphorylation modification of NFATc1 mediated by GSK-3β and PLCγ2 signaling.

Previous studies have reported that Blimp1 (Prdm1), a transcriptional repressor, is necessary for osteoclastogenesis (Nishikawa et al., 2010). NFATc1 expression is downregulated by transcriptional repressors such as IRF-8 and Bcl-6; Blimp-1 negatively regulates IRF-8 and Bcl-6 (Zhao and Ivashkiv, 2011). Thus, this study further examined whether ANE could affect the expression of these regulators of NFATc1. As shown in Figure 8 , ANE treatment markedly decreased the expression of Blimp1, while increasing the expression of IRF-8, and Bcl-6. These results indicate that ANE might suppress NFTAc1 transcription via inhibition of Blimp1 followed by enhancement of Bcl-6 and IRF-8. Taken together, these findings indicate that ANE might inhibit NFATc1 activation by modulating NF-κB and ERK1/2 signaling and enhancing the expression of NFATc1 negative regulators Bcl-6 and IRF-8.

Owing to these promising cellular effects, the effects of ANE on LPS-treated bone loss in a mouse model were evaluated using micro-CT and histology. Compared with the LPS group, bone microfracture parameters, including trabecular BV/TV, Tb.N, Tb.Th, and Tb.Sp, were significantly improved in the ANE high-dose group ( Figures 9A, B ), while the ANE low-dose groups did not shown an obvious effect. H&E and TRAP staining showed that ANE treatment obviously reduced LPS-induced bone fracture destruction ( Figure 9C ). Briefly, morphometric and histomorphometric analyses showed that ANE treatment significantly improved BV/TV and trabecular bone parameters and improved bone destruction in LPS-induced mice. These results indicate that ANE may have the potential to protect against bone loss.