Muscone (purity > 98%) was (Figure 1A) provided by Yuanye Biotechnology (Shanghai, China). RAW264.7 and BMMs cells were provided by Professor Wang (Navy Medical University). Reagents and antibodies, such as penicillin-streptomycin, fetal bovine serum (FBS), RANK antibody, and p65 antibody, were purchased from Puhe Biotechnology (Wuxi, China).

BMMs were cultured for 48 h with muscone at different concentrations (0, 10, 20, 40, 80, 160, and 320 μM). MTT solution was then added for 2 h. Finally, the absorbance was measured by enzyme-linked immunosorbent assay plate reader at 490 nm.

For primary cell culture, BMMs were isolated from the femur of C57BL/6 mice (Thummuri et al., 2017) and were treated with α-MEM and 30 ng/ml M-CSF for 3 days. The cells (1 × 10⁴ cells/well) were seeded into 96-well plates and induced by 20 ng/mL M-CSF and 50 ng/mL RANKL. Various concentrations of muscone (0, 10, 20, and 40 μM) were added into the plates on the day 0. After 7 days, tartrate-resistant acid phosphatase (TRAP) staining (Sigma, MO, USA) was performed, and TRAP-positive cells with more than 3 nuclei were counted as osteoclasts.

The actin-ring formation assay was performed as described previously (Koide et al., 2009). Cells, induced by M-CSF and RANKL for 7 days, were fixed and washed three times by PBS. Treated with fluorescein isothiocyanate (FITC)-phalloidin for 1 h, cells were then stained.

The pit formation assay was done as previously described (Chen et al., 2017). The induced mature osteoclasts were processed by collagenase and seeded onto FBS-coated dentin slices in a 96-well plate with M-CSF and RANKL. Muscone was added into the plate. After 2 days, the slices were treated and stained with 0.5% toluidine blue for 1 min.

With osteogenic differentiation medium, bone marrow-derived mesenchymal stem cells (BMSCs) were planted onto 24-well plates in the presence 0, 10, 20, and 40 μM MUS. Cells were washed 3 times with PBS, and were fixed with 4% paraformaldehyde for 15 min. It was 14 days and 21 days after the osteogenic induction that ALP staining (Sigma‐Aldrich) and ARS staining (Nanjing Jiancheng Chemical Industrial Co., Nanjing, China) were conducted, respectively.

Furthermore, to induce adipogenesis, BMSCs were cultured with 10% FBS α‐MEM supplied with 10 μg/mL insulin, 200 μmol/L indomethacin, 1 μmol/L dexamethasone, and 0.5 mmol/L 3‐isobutyl‐1‐methylxanthine (IBMX) (Cyagen Biosciences). Differentiated cells were then marked with Oil Red O staining (Sigma‐Aldrich).

Staining cells were observed under a microscope (Leica, Germany), and the mineralized area was analyzed by ImageJ software (National Institutes of Health, USA).

The effects of muscone (40 μM) on the p65 nuclear translocation, RANK, and TRAF6 in RAW264.7 cells were examined by immunofluorescence described previously (Lee et al., 2010). Cells were fixed and washed. Antibodies of targeted factors were added with anti-mouse IgG antibody and fluorescein-conjugated streptavidin sequentially for further observing and imaging.

RAW264.7 cells were divided into 3 groups. All groups were treated with M-CSF, and two groups were treated with only RANKL or RANKL + MUS (40 μM). Related factors in RANK-induced NF-κB and MAPK pathways were examined as described previously (Chen et al., 2018).

To assess whether muscone influences the binding of RANK and TRAF6 induced by RANKL, coimmunoprecipitation experiments were performed with or without 40 μM muscone, as described previously (Hu et al., 2019). And then, Western blotting was performed and analyzed.

Female C57BL/6 mice aged 8-week-old, weight 25 ± 1 g were obtained from Slack (Shanghai, China), and were maintained with free access to food and water. A total of 50 mice were divided into five groups (N=10 for each group) according to the method of random number table: control group, sham group, ovariectomy (OVX) group, muscone (MUS) treating group, and zoledronic acid (ZOL) treating group. The OVX model was performed in the specific pathogen free animal laboratory as described previously (Leng et al., 2005), in accordance with guidelines for reporting experiments involving animals. Muscone was injected intraperitoneally with 2 mg/kg every day and zoledronic acid (0.1mg/kg) (Hu et al., 2019) was injected subcutaneously once a week. 6 weeks later, mice were sacrificed by cervical dislocation, and then femurs were collected. And arterial blood samples were also reserved for further investigation. Ethical committee of Shanghai Changhai Hospital approved the treatment on animals.

H&E staining and TRAP staining were used for histologic analysis after femur bone samples were prepared and cut into 4-μm sections. Trabecular bone area and the number of osteoclasts were calculated by the image J software (National Institutes of Health, Bethesda, MD).

For each femur, 100 section planes from the growth plate were scanned (Skyscan1172, Antwerp, Belgium). We used built-in software to analyze index, including trabecular bone, bone mineral density (BMD), bone volume/total volume (BV/TV), bone surface area/total volume (BS/TV), and trabecular number (Tb.N), within the selected metaphyseal region. In addition, bone structure images in two-dimensional and three-dimensional version were reconstructed.

Cellular RNA was extracted using TRIzol (Invitrogen) as described previously (Li et al., 2011). PCR primers were shown in Supplemental Table 1.

All statistical analysis was performed using SAS 9.1 software. Results were repeated for five times and presented as means ± SEM. Results were compared with one way ANOVA followed by SNK test for multiple comparisons. P < 0.05 was regarded statistically significant.

Cell viability is examined by MTT assay. It shows that muscone has no obvious cytotoxic effects below 40 μM (Figure 1B). To examine the effects of muscone on osteoclastogenesis, BMMs and RAW264.7 cells are used and treated with 10, 20, and 40 μM muscone. As the results showed in Figures 1C, D, in the RANKL group, the number of TRAP-positive cells is significantly increased after RANKL induction within 5 to 7 days for BMMs and within 3 to 5 days for RAW264.7 cells, respectively. By contrast, muscone significantly declines the number of TRAP-positive cells in a dose-dependent manner. Thus, muscone suppresses osteoclastogenesis both in BMMs and RAW264.7 cells.

To explore the role of muscone in the osteoclast function and the cytoskeleton formation, actin-ring formation, the most important and apparent process of osteoclastogenesis, is investigated by FITC-phalloidin staining (Figure 2A). With the induction of RANKL, BMMs differentiate into mature osteoclasts and forms actin-ring cytoskeleton. However, when cells are incubated with 10, 20, and 40 μM muscone, the size and the number of actin-ring assemblies show a significantly decrease in a dose-dependent manner, which suggests that muscone inhibits the actin-ring formation by mature osteoclasts.

Furthermore, in bone resorption assays, induced with RANKL/M-CSF, RAW264.7 cells form apparent pits on the bone biomimetic synthetic surface, indicating that cells differentiate into mature osteoclasts. However, muscone significantly shrinks the resorbed range, which suggests that muscone can overturn resorption functions of osteoclasts (Figure 2B).

Besides osteoclasts, osteoblasts differentiated from BMSCs also modulates bone remodeling (Hamam et al., 2014). Therefore, we evaluate the osteogenesis effects of muscone by ALP staining and alizarin red staining. It finds that muscone (40 μM) has little influence on the materialization of calcium nodules (Figure 3). In addition, for adipogenesis of BMSCs, the oil red O staining is used, and it shows that muscone has little consequence on the formation of fat particles in vitro (Figure 3). As a result, muscone presents limited effect on BMSC differentiation and osteogenesis, which indicates that muscone prevents bone loss primarily by affecting osteoclastogenesis.

To investigate at which stage muscone inhibits osteoclastogenesis since it is a multistep processes, BMMs and RAW264.7 cells are divided into groups, according to the different days that 40 μM muscone is added after RANKL stimulation (set as day 0). Besides the RANKL group, BMMs are divided into three groups: day 1-7 group, day 3-7 group, and day 5-7 group, since muscone is added on day 1,3, and 5, respectively (Figures 4A, C). And similarly, RAW264.7 cells are divided into four groups: day 0-7 group, day 1-7 group, day 2-7 group, and day 3-7 group, as muscone is added on days 0, 1, and 3, respectively (Figures 4B, D). On day 7 after RANKL stimulation, TRAP staining is performed since it is supposed to be the granted “mature” day, and downstream mRNA expression including MMP-9, Cathepsin K, CTR, TRAF6, TRAP, and NFATc1 in BMMs are detected by RT-PCR for each group (Figure 4E). The results indicate that osteoclast differentiation is remarkably inhibited when muscone is treated on the early stage after RANKL stimulation but is less effective on later stages, suggesting that muscone has primarily effect on the stage of the osteoclastogenesis that monocyte differentiating into osteoclast precursor, but has less effect on mature osteoclasts.

Induced by RANKL, p65 translocates into the nucleus. Immunofluorescence staining shows that RANKL increases p65 location in RAW264.7 cells, but muscone significantly blocks the nuclear translocation of p65 (Figures 5A, B). In addition, Western blot results indicate that IκB phosphorylation increases and achieves its peak at 30 min, and the p50 phosphorylation and the p65 phosphorylation intensify at 60 min. Thus, it indicates that the NF-κB pathway is activated after induction of M-CSF and RANKL. However, muscone significantly suppresses phosphorylation levels in RAW264.7 cells (Figure 5C). In conclusion, muscone reduces the RANKL-induced NF-κB pathway activation.

MAPK pathway is another vital pathway in osteoclastogenesis (Li et al., 2016). For the major subfamilies of the MAPK pathway, such as ERK, p38, and JNK, are examined by western blot. The phosphorylation levels of these factors are analyzed by semi-quantitative detection, and the levels significantly increase after incubating with M-CSF and RANKL, and all reach their peaks at 30 min, but muscone administration inhibits their phosphorylations in RAW264.7 cells (Figure 5D). However, quantification results show that levels of these phosphorylation factors are significantly inhibited by muscone (Figure 5D). Together, it indicates that the muscone may inhibit osteoclastogenesis via targeting the MAPK signaling pathways.

Since RANKL-induced RANK recruits TRAF6, we examined TRAF6 by RT-PCR and Western blot. It is found that muscone suppresses the expression of TRAF6 significantly after the induction of RANKL (Figures 6A, B). In a time-dependent manner described as the above, both of the mRNA levels of RANK and TRAF6 are reduced by muscone at an early stage (Figures 7A, B). To further assess whether muscone influences the binding of RANK and TRAF6, immunofluorescence staining (Figures 7C, D) and coimmunoprecipitation (Figure 7E) are also performed, and as a result, muscone significantly inhibits the binding of TRAF6 to RANK, and suppresses the expression of TRAF6 rather than RANK.

And for down-stream gene expression, NFATc1 is a central regulator in osteoclastogenesis (Wu et al., 2017). In our study, the level of NFATc1 showed by RT-PCR and Western blot in RAW264.7 cells is over-expressed when induced by RANKL. And muscone suppresses its expression. Thus, the expression of osteclastogenesis related genes, such as MMP-9, Cathepsin K, TRAP, and CTR, are significantly down-regulated added with muscone (Figure 6).

In postmenopausal osteoporosis, as the level of estrogen decreases, the genesis and activity of osteoclasts are enhanced. And it leads to an imbalance in bone hemostasis and inflammatory response (Sapir-Koren and Livshits, 2017; Chen et al., 2018). Therefore, we examined the effects of muscone on OVX mice. Mice are sacrificed 6 weeks after operation. Hematoxylin and eosin (H&E) staining and micro computed tomography (CT) show that the OVX group has a significant trabecular bone loss, and the muscone group has more trabecular area (Figures 8A, E). In addition, detailed results for HE staining and micro CT analysis are shown in Figures 8B, F. Furthermore, the number of TRAP staining positive osteoclasts significantly increases in the OVX group, and muscone significantly decreases the number of osteoclasts (Figures 8B, D).

For pro-inflammatory cytokines, we examined the serum levels of tumor necrosis factor (TNF)-α, and interleukin (IL)-6. And for osteoclastogenesis serum markers, we inspect C-telopeptide of type I collagen (CTX-1) and tartrate-resistant acid phosphatase (TRAcp5B). Compared with the control group, levels of the above serum factors are significantly higher in OVX group. However, treated with muscone, the serum levels significantly decrease. As a result, muscone suppresses pro-inflammatory cytokines secretion and osteoclast activity in OVX mice (Figure 8G).