RAW264.7, a murine monocytic cell line, was obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Bone marrow mononuclear cells (BMMCs) were isolated from the tibial and femoral bone marrow of 6–8 week C57BL/6 mice as we described previously (Guan et al., 2015). Both RAW264.7 and BMMCs were cultured in Dulbecco’s Modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum, streptomycin (100 μg/mL), and penicillin (100 U/mL) at 37°C in a humidified incubator with an atmosphere of 95% air plus 5% CO₂. The culture medium for BMMCs was supplemented with 25 ng/mL M-CSF (R&D Systems China). For the subsequent experiment, RAW264.7 and BMMCs were treated with RANKL (50 ng/ml, R&D Systems China) and various concentrations of DMY (12.5, 25, 50, or 75 μM) to induce osteoclast formation, and the culture medium was replaced every day.

Four days after treatment, tartrate resistant acid phosphatase (TRAP) staining was performed on cultured RAW264.7 cells and BMMCs with a TRAP staining kit (Sigma–Aldrich, Shanghai, China) according to the manufacturer’s protocol. TRAP-positive cells with three or more nuclei were counted as osteoclasts (Asagiri and Takayanagi, 2007). Cell images were taken using a digital camera attached to a Nikon ECLIPSE TE2000-S microscope (Nikon, Japan). TRAP enzyme activity was measured with a TRAP Assay Kit (Sigma–Aldrich, Shanghai, China) following the manufacturer’s instructions. Briefly, culture medium was collected from osteoclasts formed by BMMCs. TRAP enzyme activity was measured with a Synergy fluorescence plate reader at 405 nm on a colorimetric plate reader.

For cell viability assay, RAW264.7 cells and BMMCs were separately seeded in 96-well plates, the culture medium for BMMCs containing 25 ng/mL M-CSF. After 24 h, cells were treated with DMY as indicated concentrations, the culture medium containing DMY was replaced every day. At the end of the incubation, 100 μl medium with 5 mg/10 ml MTT solution (Sigma–Aldrich, Shanghai, China) were added and incubated at 37°C for an additional 4 h. Then, the medium was removed and 200 μl of DMSO was added to each well. Absorbance was measured at 570 nm with a microplate reader (ChemiDoc MP, Bio-Rad, United States).

Bone marrow mononuclear cells were treated with RANKL (50 ng/ml) and M-CSF (25 ng/mL) on 6-well collagen pre-coated plates for 4 days to form osteoclasts. Then mature osteoclasts were collected using 2.5 mg/mL collagenase in dissociation buffer (Life Technologies, Carlsbad, CA, United States) and seeded onto Corning Osteo Assay Surface (Corning Incorporated Life Science, Corning, NY, United States) in a multiple well plate in the presence of RANKL (50 ng/ml) and different concentrations of DMY supplemented with M-CSF (25 ng/mL) for 3 days. The disks were washed with 5% sodium hypochlorite for 5 min, and images were taken and resorption was quantified by image analysis (Bioquant Image Analysis, Nashville, TN, United States). The actin ring structure formation assay was examined by fluorescence microscopy as described previously (Soysa et al., 2009; Wilson et al., 2009).

Quantitative real-time polymerase chain reaction (qRT-PCR) was performed as described before (Guan et al., 2010, 2013). Briefly, RAW264.7 cells were treated with RANKL and 75 μM DMY for 3 days, total RNA was isolated using TRIzol reagents (Invitrogen Life Technologies, Carlsbad, CA, United States) and first-strand cDNA was synthesized with MMLV reverse transcriptase (Promega, Madison, WI, United States). Templates were amplified with QuantiTect SYBR Green PCR Kit (Qiagen, Valencia, CA, United States) on the iCycler real-time PCR instrument (BIO-RAD, Berkeley, CA, United States). Primers are listed in Table 1 (annealing temperature = 60°C).

Immunoblots were performed on RAW264.7 cells as described (Guan et al., 2010, 2013). RAW264.7 cells were treated with RANKL and 75 μM DMY for the indicated time, then protein was prepared. The following primary antibodies (1:1000 dilution) were obtained from Cell Signaling (Cell Signaling Technology, Boston, MA, United States): phospho-Akt (Ser-473), Akt, phospho-ERK1/2 (Thr202/Tyr204), ERK, phospho-JNK (Thr183/Tyr185), JNK, phospho-p38 (Thr180/Tyr182), p38. The antibodies (1:500 dilution) against GAPDH and β-Actin were obtained from BOSTER (BOSTER, Wuhan, China). Secondary antibodies (1:5000 dilution) used were goat anti-rabbit IgG-horseradish peroxidase (HRP; sc-2004, Santa Cruz, CA, United States). Signals were visualized with enhanced chemiluminescence and captured by a scanner (ChemiDoc MP, Bio-Rad, United States). Electrophoretic mobility shift assay (EMSA) was performed as described previously using a LightShift Chemiluminescent EMSA Kit (Thermo Fisher Scientific, China) (Guan et al., 2013). Briefly, RAW264.7 cells in 6-well plates were pretreated with DMY with a concentration of 75 μM for 2 h, then stimulated with RANKL (50 ng/ml) for 30 min. Nuclear extracts were prepared using Nuclear and Cytoplasmic Protein Extraction Kit according to the manufacturer’s instructions (Beyotime Institute of Biotechnology, Jiangsu, China) and quantified. An equal amount of nuclear extract was incubated with biotin end-labeled duplex DNA and electrophoresed on a 6% polyacrylamide native gel. The AP-1 and NF-κB probes (Beyotime Institute of Biotechnology, Jiangsu, China) used for EMSA, containing the consensus recognition site were as follows: AP-1, 5′-CGCTTGATGACTCAGCCGGAA-3′; NF-κB, 5′-AGTTGAGGGGACTTTCCCAGGC-3′.

RAW264.7 cells were cultured on 12-well plates, after treatment with DMY for 36 h, cells were incubated for 30 min in presence of RANKL (50 ng/ml). Subsequently, ROS production was measured by flow cytometry with an ROS assay kit (Beyotime Institute of Biotechnology, Jiangsu, China) as described previously (Guan et al., 2015).

This study was carried out in accordance with the re-commendations of Animal Experimentation Guidelines, the Ethics Committee on Animal Experimentation of Tongji Medical College, Huazhong University of Science and Technology (Wuhan, China). All animal procedures were approved by the Ethics Committee on Animal Experimentation of Tongji Medical College, Huazhong University of Science and Technology (Wuhan, China). C57/BL6 mice were purchased from the Experimental Animal Center of Tongji Medical College (Wuhan, China). All mice were kept in ventilated filter-top cages under standard laboratory conditions at a constant temperature of 25°C with a 12-h light/dark cycle, and were fed conventional rodent chow with water.

Four-month-old female C57BL/6 mice (21 ± 1 g) were divided randomly into three groups (n = 12 mice per group): sham-operated mice, bilateral ovariectomized (OVX) mice treated with VEH, and OVX mice treated with DMY. Sham operation was performed by identifying the bilateral ovaries and ovariectomy was performed by removing the bilateral ovaries, both through a dorsal approach. One day after operation, mice were injected intraperitoneally with VEH or DMY 50 mg/kg/d 5 days a week for 6 weeks. After 6 weeks, the mice were sacrificed for experiments and uterus wet weight was measured to validate the success of ovariectomy.

After removal of soft tissues, microcomputer tomography (μCT) (μ -CT50 ScancoMedical, Bassersdorf, Switzerland) was performed on the distal femur. Scans were taken with a source voltage of 80 kV and 80 μA source current with a voxel size of 10 μm. The bone structural parameters of bone mineral density (BMD), bone volume/tissue volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp) were quantitatively analyzed with the built-in software of the μCT. The three-dimensional bone structure image slices were reconstructed using the built-in software. Nomenclature and abbreviations of parameters follow the recommendations of the American Society of Bone and Mineral Research (Bouxsein et al., 2010).

For histological analysis, the femur samples were decalcified with 10% tetrasodium-EDTA aqueous solution at 4°C for 1 week. The samples were then embedded in paraffin. Paraffin-embedded bone sections (5 μm) of each femur were prepared for hematoxylin and eosin (H&E) staining and TRAP staining to observe the histology of the metaphysis below the primary spongiosa. Histological measurements and images were taken by a microscope equipped with a camera. Trabecular bone density was measured in sections with H&E staining, while numbers of osteoclasts of trabecular bone surface were counted in the sections with TRAP staining.

Blood from Sham+VEH, OVX+VEH, and OVX+DMY mice was collected by retro-orbital puncture before sacrifice. According to the manufacturer’s instructions, serum levels of TRAP, TNF-α, IL-1β, IL-6, RANKL, and OPG were measured by enzyme-linked immunosorbent assay (ELISA) kits: TRAP ELISA kit (BD Biosciences, San Jose, CA, United States), TNF-α, IL-1β, and IL-6 ELISA kits (eBioscience, San Diego, CA, United States), and RANKL and OPG ELISA kits (Boster, Wuhan, China).

All experiments were independently repeated three times with similar results. Data are expressed as mean ± standard deviation (SD). Student’s t-test was used for comparison between two groups, and analysis of variance (ANOVA) was used in multiple comparisons. All statistical analyses were carried out with SPSS13.0 software (SPSS, Chicago, IL, United States), statistical significance was considered as P < 0.05.

Treatment with DMY significantly reduced the numbers of the TRAP-positive multinuclear cells both in RAW264.7 cells (Figure 1A) and BMMCs (Figure 1B) in a dose-dependent manner, with the maximal effect at 75 μM of concentration. Likewise, DMY treatment decreased TRAP enzyme activity in BMMCs culture medium (Figure 1C). The potential toxicity of DMY was evaluated by a MTT assay; DMY up to 75 μM did not detectably inhibit the viability and proliferation of RAW264.7 cells and BMMCs (Figure 1D). The effect of DMY on the apoptosis of RAW264.7 cells was also assessed by Annexin V/PI double staining, DMY up to 100 μM did not affect the apoptosis of RAW264.7 cells (Supplementary Figure 1). Pit formation assays were performed to further assess osteoclast function. As shown by the diminished area of resorption pits formed by osteoclasts, osteoclast function was severely impaired by DMY treatment (Figures 2A,B). DMY markedly disrupted osteoclast actin ring formation (Figure 3), which is essential for the attachment and bone resorption of osteoclasts (Wilson et al., 2009).

In RAW264.7 cells, treatment with DMY strikingly inhibited Trap, Mmp-9, Cathepsin K, Nfatc1, Rank, and C-Fos mRNA expression (Figure 4A). Immunoblot analysis demonstrated the downregulation of these osteoclast-specific proteins by DMY (Figure 4B). Among the three major subfamilies of MAPK, DMY suppressed RANKL-induced phosphorylation of JNK and ERK (Figure 5A). Furthermore, DMY inhibited the levels of AKT phosphorylation (p-AKT), IkBα phosphorylation (p-IkBα), and p65 phosphorylation (p-p65) (Figure 5A). The RANK protein level was depressed in our immunoblot analysis (Figure 5B).

Transcription factors such as NF-κB and AP-1 play an essential role in osteoclastogenesis (Takayanagi, 2007a,b). By EMSA assays (Figure 5C), without RANKL stimulation, DMY had no significant influence on baseline NF-κB or AP-1 DNA-binding activity. Whereas, DMY remarkably inhibited the activation of NF-κB and AP-1 by RANKL.

The intracellular level of ROS is increased by RANKL stimulation, and ROS also activates osteoclast differentiation (Lee et al., 2005). Our results showed that the production of ROS was increased through RANKL stimulation, but was effectively decreased by DMY in a dose-dependent manner (Figure 5D).

We evaluated the effects of DMY on osteopenia using an OVX mouse model which mimics menopause-induced bone loss (Bouxsein et al., 2005). The weight of the uterus was used as an indicator of estrogen status to assess the success of ovariectomy and OVX resulted in a significant decrease in uterus weight (Ohlsson et al., 2014) (Supplementary Figure 2). As expected, 6 weeks after operation, treatment with DMY in OVX mice dramatically attenuated trabecular bone loss as shown by the μCT, compared to the sham-operated mice. OVX mice treated with VEH exhibited a significant loss of trabecular bone, as revealed by decreased BMD, BV/TV, Tb.N, and Tb.Th and by increased Tb.Sp (Figure 6). The results were further corroborated by H&E (Supplementary Figure 3A) and TRAP staining of decalcified bone sections. Compared with the Sham+VEH group, femoral sections from OVX mice treated with VEH demonstrated a paucity of cancellous bone both proximal and distal to the growth plate. DMY treatment in the OVX mice induced a marked increase in bone density and much fewer TRAP-positive multinucleated cells, which was also shown by the osteoclast numbers per bone surface (N.Oc/BS) (Figure 7). Moreover, we further confirmed that DMY (50 mg/kg) did not induce tissue damage in vivo, compared with the Sham+VEH group, as no histopathological changes in the liver and kidney tissues were observed in H&E staining after treating with DMY (Supplementary Figures 3B,C).

We next examined the serum levels of TRAP, TNF-α, IL-1β, IL-6, RANKL, and OPG. Serum TRAP level, a serologic marker of osteoclast function, was decreased by DMY (Figure 8). These observations suggested that DMY functioned as an inhibitor of osteoclastogenesis and resorption activity. DMY also effectively decreased serum levels of TNF-α, IL-1β, IL-6, and RANKL and increased the serum level of OPG.