Sec-O-glucosylhamaudol (purity ≥95%) was obtained from Nature Standard (Shanghai, China). Alpha modification of Eagle’s minimum essential medium (α-MEM), fetal bovine serum (FBS), and penicillin and streptomycin (PS) were purchased from Gibco (Grand Island, NY, USA). Recombinant murine M-CSF and RANKL were from Prospec (Rehovot, Israel). SB415286 was purchased from Selleckchem (Houston, TX, USA). LPS from Escherichia coli O55:B5, 2,5-diphenyltetrazolium bromide (MTT), zoledronic acid (ZOL), tartrate-resistant acid phosphatase (TRAP) staining kit, rhodamine phalloidin, 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI), and all other chemicals were obtained from Sigma-Aldrich (St Louis, MO, USA). Antibodies against phospho-P65, NFATc1, c-Fos, and PP2B-Aα were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). All other antibodies were from Cell Signaling Technology (Danvers, MA, USA).

Murine bone marrow cells were separated from the femur and tibia of 6–8-week-old mice (SLAC Laboratory Animal Co., Shanghai, China) and cultured in α-MEM supplemented with 10% FBS, 1% PS, and 30 ng/mL M-CSF for 24 h as previously described (14). Adherent cells were removed and floating cells were collected by centrifuging at 1000 × g for 5min, re-suspending, and culturing in 10-cm dishes in α-MEM complete medium with 30 ng/mL M-CSF at 37°C, 5% CO₂ for 2 days (d). Adherent cells were regarded as bone marrow macrophages (BMMs) and used as osteoclast precursors.

For osteoclast differentiation, BMMs (7 × 10⁴ cells/well) were cultured in 48-well plates with α-MEM complete medium supplemented with 30 ng/mL M-CSF and 100 ng/mL RANKL for 4 d. Various concentrations of SOG were added to cultures at indicated times. The TRAP staining assay was performed as described previously (14), and TRAP-positive cells containing three or more nuclei were regarded as osteoclasts.

BMMs (1.5 × 10⁴ cells/well) were cultured in 96-well plates with α-MEM complete medium containing 30 ng/mL M-CSF, 100 ng/mL RANKL, and SOG (0–200 μM) for 24 h or 48 h. After removing the medium, 100 μL of MTT (0.5 mg/mL) solution was added to each well. After 4 h, the supernatant was discarded and 150 μL dimethyl sulfoxide was added. The optical density at 570 nm was measured using a microplate reader.

BMMs (7 × 10⁴ cells/well) were cultured in 48-well plates with α-MEM complete medium containing 30 ng/mL M-CSF, 100 ng/mL RANKL, and SOG (0–200 μM) for 4 d. Cells were fixed with 4% paraformaldehyde (PFA) for 20 min, and permeabilized with 0.5% Triton X-100 for 5 min. After washing three times with phosphate-buffered saline (PBS), phalloidin was used to stain F-actin for 30 min and DAPI was used to stain nuclei for 5 min. Fluorescent images were acquired on a fluorescence microscope (RVL-110-G; ECHO, San Diego, CA, USA).

BMMs (1 × 10⁵ cell/well) were cultured in a Corning osteo assay surface 24-well plate with α-MEM complete medium containing 30 ng/mL M-CSF, 100 ng/mL RANKL, and SOG (0–200 μM) for 4 d. Cells were then removed with 5% sodium hypochlorite, and modified von Kossa staining was performed as previously described (15). The resorption area was observed under a light microscope and analyzed using ImageJ software (NIH, Bethesda, MD, USA). The results were expressed as a percentage of the total well area.

BMMs (1 × 10⁶ cells/well) were cultured in 6-well plates with α-MEM complete medium containing 30 ng/mL M-CSF, 100 ng/mL RANKL, and 200 μM SOG for the indicated periods. Total RNA was extracted using the RNA EASY MINI KIT (Qiagen, Hilden, Germany) according to the manufacturer’s instruction. The Transcriptor First Strand cDNA Synthesis Kit (Roche, Mannheim, Germany) was used to reverse-transcribe cDNA. mRNA expression levels of osteoclast-related genes were evaluated by real-time PCR using a SYBR Green kit (Roche), and glyceraldehyde 3-phosphate dehydrogenase as a control gene. PCR conditions were: 95°C for 10 min, followed by 40 cycles at 95°C for 10 s, 60°C for 15 s, and 72°C for 20 s. Primer sequences are shown in Table 1.

BMMs cultured in indicated plates at 37°C for 24 h were transfected with control siRNA or 5-LO siRNA at a final concentration of 50 nM using Ribo FECTTM CP Reagent (RiboBio, Guangzhou, China) according to the manufacturer’s protocol. Control siRNA was used as a negative control for the possible nonspecific effects of RNA interference. After removing transfection mixtures, cells were cultured with α-MEM complete medium containing 30 ng/mL M-CSF and 100 ng/mL RANKL for the indicated time. TRAP staining or western blotting assays were then performed.

BMMs were treated with 200 μM SOG in the presence or absence of 30 ng/mL M-CSF and 100 ng/mL RANKL for the indicated time. Cells were then collected and lysed using radioimmunoprecipitation assay lysis buffer on ice. Equal amounts of protein sample (30–50 μg) were separated by 10%–12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes. Membranes were blocked with 5% fat-free milk in Tris-buffered saline and Tween 20 (TBST) for 2 h and incubated with primary antibodies against target genes overnight at 4°C. After washing three times with TBST, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody for 2 h at room temperature. After washing five times with TBST, the membranes were detected using the MicroChemi Chemiluminescence system (DNR Bio Imaging System, Jerusalem, Israel), and band intensities were quantified using ImageJ software.

Male C57BL/6 mice (6-weeks-old) were obtained from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China). Animal experiments were performed strictly according to the Guide for the Humane Use and Care of Laboratory Animals (Animal protocol number: TJCAC-018-036). Mice were randomly assigned to five groups (n=10 per group): PBS control (Sham group), LPS injection (LPS group; 5 mg/kg body weight), SOG injection (only SOG group, 10 mg/kg body weight), LPS with 5 mg/kg body weight dose of zoledronic acid (ZOL group), and LPS with 10 mg/kg body weight dose of SOG (SOG group). SOG, ZOL, and LPS were injected intraperitoneally every other day for 8 d. A total of 2 h after SOG or ZOL injection, LPS or PBS were injected intraperitoneally. Mice were euthanized on d 8 after the first injection of LPS. Left and right femurs were collected for micro-computed tomography (CT) and histomorphometric analysis, respectively.

Left femurs were fixed with 4% PFA for 24 h, and scanned by micro-CT using a Quantum GX microCT system (PerkinElmer, Waltham, MA, USA). The bone mineral density (BMD), bone volume/total volume (BV/TV), and trabecular number (Tb.N), trabecular thickness (Tb.th) and decreased trabecular separation/spacing (Tb.Sp) were measured.

Right femurs were demineralized in 10% ethylenediaminetetraacetic acid for 2 weeks. They were then embedded in paraffin, cut into 5 μm-thick sections, and stained with H&E and TRAP. Staining was visualized using a high-quality microscope (RVL-110-G; ECHO, San Diego, CA, USA).

GraphPad Prism 7 software (San Diego, CA, USA) was used for data analysis. Data were expressed as means ± SD. Differences between groups were analyzed using one-way analysis of variance followed by the Student–Newman–Keuls test. P < 0.05 was considered statistically significant.

To determine whether SOG influences RANKL-mediated osteoclast differentiation, BMMs were cultured in the presence of M-CSF, RANKL, and indicated concentrations of SOG. SOG showed no obvious cytotoxicity at a concentration range of 50–200 μM (Figure 1B). As shown in Figure 1C, RANKL significantly induced the formation of TRAP-positive multinucleated cells, which was inhibited by SOG treatment in a concentration-dependent manner.

Next, the effects of SOG on RANKL-induced osteoclastogenesis during different stages of differentiation were investigated. SOG (200 μM) treatment significantly reduced the number and size of osteoclasts at all stages of cell differentiation, with stronger inhibitory effects on osteoclast formation seen during middle and late stages (d 3 and d 4) compared with early stages of cell differentiation (d 1 and d 2) (Figures 1D, E). Taken together, these results suggest that SOG suppresses RANKL-induced osteoclastogenesis, especially in the middle–late stages of osteoclast differentiation.

Previous studies reported that in osteoclasts plated on bone slices, dynamin was enriched in the actin-rich sealing zone that delineates the resorptive space (16). Similarly, when plated on serum-coated coverslips or regular plates, dynamin was highly concentrated at the periphery of the cell, where podosomes are localized as a belt, this belt is called F-actin ring. The formation of a well-defined F-actin ring is known to be essential for the bone resorption function of osteoclasts (4). Therefore, we examined the effects of SOG treatment on RANKL-induced osteoclast F-actin ring formation. Figure 2A shows that the number and size of F-actin rings were significantly reduced with increasing concentrations of SOG treatment of RANKL-stimulated BMMs. SOG treatment also significantly dose-dependently decreased the formation of resorption pits, as shown in Figure 2B. At a concentration of 200 μM, SOG treatment almost completely abolished the formation of F-actin rings and resorption pits in RANKL-stimulated BMMs. Meanwhile, 200 μM SOG was added to BMMs at day 1, 2, 3 and 4 during osteoclastogenesis for 24 h. SOG exhibited stronger inhibition on bone resorption pits formation at middle-late stage (day 3 and day 4 after RANKL treatment) compared to that at early stage (Figure S1). These results suggest that SOG inhibits RANKL-induced mineral resorption and formation of F-actin rings through inhibition of RANKL-induced osteoclast formation.

RANKL stimulation up-regulates osteoclast-related genes, which control the differentiation, maturation, and function of osteoclasts, while master transcription factor NFATc1 regulates the expression of osteoclast-specific genes (17). We therefore investigated the effects of SOG on the expression of NFATc1 and c-Fos, and several osteoclast marker genes including CTSK, TRAP, and DC-STAMP. SOG treatment significantly suppressed the induction of c-FOS and NFATc1 by RANKL at both the mRNA and protein level, in a time-dependent manner (Figures 3A, B). Figure 3C also shows that SOG treatment significantly suppressed mRNA expression levels of CTSK, TRAP, and DC-STAMP in a time-dependent manner in RANKL-stimulated BMMs. These results suggest that SOG inhibits osteoclastogenesis by suppressing the expression of c-Fos/NFATc1-mediated osteoclast-specific genes.

NF-κB, MAPK, and AKT-GSK3β pathways play a vital role in the early induction of NFATc1 (18–20). To further explore the molecular mechanism by which SOG modulates NFATc1 signaling, we evaluated its effects on the RANKL-initiated transient activation of NF-κB, MAPK, and AKT, and the inactivation of GSK3β using western blotting. As shown in Figure 4A, RANKL stimulation induced the transient phosphorylation of the NF-κB p65 subunit and MAPKs including p38, extracellular signal−regulated kinase (ERK)1/2, and c-Jun N-terminal kinase (JNK), which was unchanged by SOG treatment. The RANKL-induced transient phosphorylation of AKT and GSK3β was also unchanged by SOG treatment (Figure 4B). These results indicate that SOG does not alter the RANKL-induced transient activation of NF-κB, MAPK, and AKT signaling, implying that it might not influence the initial induction of NFATc1 during the early stages of osteoclast differentiation.

NFATc1 activation is regulated by calcineurin-mediated NFATc1 dephosphorylation and GSK3β-mediated NFATc1 phosphorylation during the middle–late stage of osteoclastogenesis (21). To investigate whether SOG affects NFATc1 activity at the middle–late stage of cell differentiation, the phosphorylation of AKT and GSK3β and protein levels of the calcineurin catalytic subunit PP2B-Aα were detected during different stages of osteoclastogenesis. RANKL-stimulated BMMs were treated with 200 µM SOG for 24 h at different periods of osteoclastogenesis. As shown in Figure 5A, SOG treatment had no effect on the protein levels of the calcineurin catalytic subunit PP2B-Aα in RANKL-stimulated BMMs. However, the phosphorylation of AKT and GSK3β was clearly inhibited by SOG in a time-dependent manner in RANKL-stimulated BMMs during the middle–late stage of osteoclast differentiation. These results suggest that SOG suppresses NFATc1 activation by inhibiting AKT-mediated GSK3β inactivation.

Further, the selective GSK3β inhibitor SB415286 was used to determine whether SOG suppresses RANKL-induced osteoclast differentiation through blocking AKT-mediated GSK3β inactivation. As shown in Figure 5B, the addition of 10 μM SB415286 weakened the inhibitory effect of SOG on osteoclast formation during the middle–late stage of cell differentiation. Taken together, these data suggest that SOG represses RANKL-induced osteoclast differentiation at least in part by inhibiting AKT-mediated GSK3β inactivation during the middle–late stage of osteoclastogenesis.

SOG was previously reported to inhibit 5-LO catalytic activity in vitro with an IC₅₀ value of 7.45 μM (13). To explore whether SOG blocks AKT-mediated GSK3β inactivation by inhibiting 5-LO, we silenced the expression of 5-LO in BMMs by 5-LO targeted siRNA transfection. As shown in Figure 6A, 5-LO knockdown significantly inhibited osteoclast formation compared with the control group. Additionally, NFATc1 induction by RANKL was significantly attenuated in 5-LO siRNA-transfected BMMs compared with controls (Figure 6B). However, 5-LO knockdown did not change the phosphorylation levels of AKT or GSK3β (Figure 6C). Taken together, these data indicate that 5-LO knockdown suppresses RANKL-induced osteoclastogenesis, but not through the AKT/GSK3β signaling pathway. SOG might also suppress RANKL-induced osteoclastogenesis by inhibiting both the independent 5-LO pathway and AKT-mediated GSK3β inactivation.

Given the inhibitory effects of SOG on osteoclastogenesis in vitro, we next evaluated the potential therapeutic effects of SOG in an LPS-induced mouse model of bone loss. Two- and three-dimensional reconstruction images of micro-CT revealed that the LPS group exhibited significant femur destruction compared with the sham group (Figure 7A). However, SOG treatment significantly improved bone destruction in LPS-treated mice. Similarly, zoledronic acid (ZOL), a known anti-bone resorptive drug used as a positive compound, markedly reduced LPS-induced bone loss. Quantitative analysis of bone microparameters showed that SOG treatment increased the values of bone mineral density (BMD), bone volume/total volume (BV/TV), and trabecular number (Tb.N), trabecular thickness (Tb.th) and decreased trabecular separation/spacing (Tb.Sp) compared with the LPS-treated group (Figure 7B). Hematoxylin and eosin (H&E) staining showed that SOG treatment significantly reduced LPS-induced bone destruction (Figure 7C). Moreover, TRAP staining revealed that SOG decreased the number of TRAP-positive osteoclasts compared with the LPS group (Figure 7C). These observations suggest that SOG attenuates LPS-induced bone loss in mice.