Magnoflorine [C₂₀H₂₄NO₄; MW, 342.41; purity, ≥98% (Meilun, Dalian, China)] was dissolved in dimethyl sulfoxide. Recombinant murine M-CSF and RANKL were from R&D Systems (Minneapolis, MN, USA). Fetal bovine serum and alpha-modified minimal essential medium (α-MEM) were supplied by Gibco-BRL (Sydney, Australia). The staining kit for tartrate-resistant acid phosphatase (TRAP) was obtained from Sigma-Aldrich (St. Louis, MO). The assay for cell viability, cell counting kit-8 (CCK-8), was from Dojindo Molecular Technology (Japan). Primary antibodies targeting extracellular signal-regulated kinase (ERK) and its phosphorylated form (phospho-ERK; Thr202/Tyr204), p38 and phospho-p38 (Thr180/Tyr182), TAK1 and phospho-TAK1, c-Jun N-terminal kinase (JNK) and phospho-JNK (Thr183/Tyr185), NF-κB (p65) and phospho-NF-κB, NF-κB inhibitor alpha (IκBα) and phospho-IκBα, c-Fos, β-actin and histone H3 (as internal controls), and appropriate secondaries conjugated to fluorescent dyes were from Cell Signaling Technology (Cambridge, MA). The primary antibody recognizing nuclear factor of activated T cells, cytoplasmic 1 (NFATc1) was supplied by Absin Bioscience Inc. (Shanghai, China). The Prime Script RT reagent kit and SYBR^® Premix Ex Taq™ II were from Takara Biotechnology (Otsu, Japan).

Commercial pure Ti particles (Alfa Aesar, Haverhill, MA, USA) were <20 µm in diameter. They were baked for 8 h at 180°C and then washed in 75% ethanol for 48 h for endotoxin removal before resuspension in sterile phosphate-buffered saline (PBS) at 0.3 g/ml and storage at 4°C. Ti particles were confirmed to be free of endotoxins via a Limulus amebocyte lysate assay.

The mouse model of calvarial osteolysis used in these studies was described previously (Tian et al., 2014; Wu et al., 2015). All experimental procedures were performed in strict accordance with the guidelines for Care and Use of Laboratory Animals and were approved by the Animal Care Committee of Shanghai Jiao Tong University. Twenty-four 6–8-week-old male C57BL/6 mice (Shanghai Slac Laboratory Animal Company) were housed under specific-pathogen-free conditions and placed into four groups of six mice each: those that underwent a sham operation and were treated only with PBS (sham group), those treated with Ti particles in PBS (vehicle group), those treated with Ti particles and 2 mg/kg magnoflorine (low‐dose magnoflorine group), and those treated with Ti particles and 5 mg/kg magnoflorine (high‐dose magnoflorine group). During the surgeries, the middle sutures of the calvarias received coatings of 30 mg sterilized Ti particles, which were applied under the periosteum. Beginning 2 days later, PBS or magnoflorine was injected under the scalp every other day for 14 days. Experimental procedures did not result in any adverse effects or mortality. After the last treatment, the mice were killed by an overdose of anesthetic, and the calvarias were harvested and fixed in 4% paraformaldehyde.

The degree of calvarial osteolysis was analyzed via high-resolution micro-CT (μCT-100; SCANCO, Brüttisellen, Switzerland) with a 10-µm isometric resolution, 300-ms exposure time, and 70 kV, 200 μA X-ray energy. Regions of interest (ROI) in reconstructed images were defined as a square around where the coronal and sagittal midline suture intersect. Bone mineral density, the bone-to-tissue volume ratio (BV/TV, %), the number of pores, and the percentage of total porosity of each sample were measured in the ROI as previously reported (Wedemeyer et al., 2007).

Paraformaldehyde-fixed calvarias were decalcified in 10% (w/v) EDTA (pH 7.4) for approximately 2 weeks before they were embedded in paraffin. Hematoxylin and eosin (H&E) and TRAP staining and immunohistochemistry for tumor necrosis factor alpha (TNF-α) and interleukin 1 beta (IL-1β) were performed and imaged by microscopy (Leica Microsystems, Wetzlar, Germany). The number of multinucleated TRAP-positive osteoclasts, the percentage of osteoclasts surface per bone surface (OcS/BS %) as well as amount of TNF-α- and IL-1β- positive cells were quantified with ImageJ software.

Primary BMMs were isolated from the femurs and tibiae of male C57/BL6 mice (6 weeks of age) as previously described (Tian et al., 2014; Zhai et al., 2014; Wu et al., 2015) and cultured in complete α‐MEM (containing 10% fetal bovine serum and 100 U/ml penicillin/streptomycin) supplemented with 30 ng/ml M‐CSF. The cells were maintained in a humidified environment of 5% CO₂ and 37°C for another 5–6 days to obtain BMMs.

For CCK-8 assays, 8 × 10³ BMMs cells were seeded in triplicate in a 96-well plate and cultured for 48, 72, or 96 h with complete α-MEM supplemented with M-CSF (30 ng/ml) and increasing concentrations of magnoflorine (0, 12.5, 25, 50, 100, 200, 400, and 800 μM). Then, 100 μl fresh medium containing 10 μl CCK-8 solution was added to each well. The plates were incubated at 37°C for another 2 h before the optical density at 450 nm wavelength was measured (with 630 nm serving as the reference wavelength) with a microplate reader. Cell viability is expressed as a percentage relative to vehicle-treated control cells (100%).

To assess osteoclast differentiation, 1.0×10⁴ BMMs were seeded in triplicates in 96-well plates and cultured in complete α-MEM supplemented with M-CSF (30 ng/ml), RANKL (50 ng/ml), and magnoflorine (0, 50, 100, and 200 μM). The media was changed every other day until multinucleated pancake-shaped osteoclasts were observed (day 5–7). After fixing for 20 min with 4% paraformaldehyde, the cells were incubated in TRAP staining solution for 1 h at 37°C. Multinucleated cells which were positive for TRAP staining and contained more than three nuclei were identified as osteoclasts. Digital images were captured with an optical light microscope (Olympus, Tokyo, Japan), and the numbers of osteoclasts and the areas of cell spread were then quantified using Image J software (NIH, Bethesda, MD, USA).

To observe immunofluorescence staining of podosome actin belt, BMMs were treated consistently with the TRAP staining assay as previously described. Briefly, after mature osteoclasts were formed, the cells were fixed for 20 min with 4% paraformaldehyde, and then permeabilized for 5 min with 0.2% Triton X in PBS. Cells were then incubated for 30 min in darkness with rhodamine-conjugated phalloidin diluted in PBS containing 1% bovine serum albumin (BSA). After extensively rinsing with PBS, cells were counterstained for 5 min with 4′, 6-diamidino-2-phenylindole (DAPI) to visualize cell nuclei. A fluorescence microscope (Leica Images) was used to acquire images, and the numbers and spread areas of podosomal actin belts were quantified with Image J software.

BMMs (1.5 × 10⁴ cells/well) were plated in triplicate in Osteo assay surface plates (Corning, NY) coated with calcium phosphate to mimic bone. The cells were incubated in complete α-MEM containing M-CSF (30 ng/ml), RANKL (50 ng/ml), and magnoflorine (0, 50, 100, and 200 μM). On day 7 when mature osteoclasts were observed, the plates were rinsed with a 5% sodium hypochlorite solution to remove adherent cells from the bottle of plates, and air dried. The resorption pits were imaged with a BioTek Cytation 3 Cell Imaging Reader (BioTek, Winooski, VT), and the resorption pit area was quantified with Image J software; the values are presented as ratios relative to the control (0 µM magnoflorine).

To assess Ti particle-mediated inflammatory response in vitro, quantitative real-time PCR and enzyme-linked immunosorbent assay (ELISA) were performed to detect the mRNA and protein levels of pro-inflammatory cytokines, respectively. 3 × 10⁵ cells/well BMMs were seeded into a 6-well plate and cultured in complete α-MEM supplemented with 30 ng/ml M-CSF. After adhering overnight, cells were stimulated for 24 h with 0.1 mg/ml Ti particles and magnoflorine (0, 50, 100, and 200 μM). Then, the supernatants were harvested for ELISA analysis, and total RNA was isolated from the adherent cells for real-time PCR analysis.

After culturing for 1 day, culture medium from Ti particle-exposed cells was harvested, centrifuged for 10 min at 2,500 rpm and then stored at -80°C until used for subsequent analyses. Murine ELISA kits were used to measure the concentrations of TNF-α, IL-1β, as well as IL-6 per manufacturer’s instructions. Absorbances at 450 nm were recorded with a microplate reader.

For osteoclast differentiation experiments, total RNA was extracted from BMM-derived osteoclasts cultivated for 5 days with 30 ng/ml M-CSF, 50 ng/ml RANKL in the absence or presence of different doses of magnoflorine (50, 100, and 200 μM) by using Axygen RNA Miniprep Kit (Axygen, Union City, CA, USA) per manufacturer’s directions. For inflammatory response evaluation, the total RNA was extracted as mentioned above. One milligram of total RNA was reversed transcribed with a PrimeScript RT reagent kit (TaKaRa Biotechnology), and the complementary DNAs (cDNAs) served as the template for qRT-PCR assay with the TB Green Premix Ex Taq kit (TaKaRa Biotechnology). All reactions were performed in triplicates and were normalized to the housekeeping gene Gapdh. The primer sequences (forward and reverse, 5′ to 3′) based on mouse sequences were as follows: GAPDH, GGTGAAGGTCGGTGTGAACG and CTCGCTCCTGGAAGATGGTG; NFATc1, TGCTCCTCCTCCTGCTGCTC and GCAGAAGGTGGAGGTGCAGC; cathepsin K (CTSK), GGGAGAAAAACCTGAAGC and ATTCTGGGGACTCAGAGC; V-ATPase d2, AAGCCTTTGTTTGACGCTGT and TTCGATGCCTCTGTGAGATG; TRAP, CTGGAGTGCACGATGCCAGCGACA and TCCGTGCTCGGCGATGGACCAGA; DC-STAMP, AAAACCCTTGGGCTGTTCTT and AATCATGGACGACTCCTTGG; c-Fos, CCAGTCAAGAGCATCAGCAA and AAGTAGTGCAGCCCGGAGTA; calcitonin receptor (CTR), TGCAGACAACTCTTGGTTGG and TCGGTTTCTTCTCCTCTGGA; TNF-α, CATCTTCTCAAAATTCGAGTGACA and TGGGAGTAGACAAGGTACAACCC; IL-1β, TGCCACCTTTTGACAGTGATG and TGATGTGCTGCTGCGAGATT; IL-6, TCCAGTTGCCTTCTTGGGAC and AGTCTCCTCTCCGGACTTGT.

In order to investigate the effects of magnoflorine on early RANKL‐mediated signaling pathway, total cellular proteins (TCPs) and nuclear proteins were extracted from BMMs pre-treated for 4 h with serum-free α-MEM with or without magnoflorine followed by RANKL (50 ng/ml) stimulation for 0, 5, 10, 20, 30, and 60 min. To assess c-Fos and NFATc1 expression, total cellular proteins were extracted from BMMs cultured with M-CSF (30 ng/ml) and RANKL (50 ng/ml) in the absence or presence of 200 μM magnoflorine for 0, 1, 3, and 5 days. The total proteins and nuclear proteins of these specific time points were harvested respectively with radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, Shanghai, China) supplemented with PMSF, protease and phosphatase inhibitor (Sigma-Aldrich) and with a nuclear protein extraction commercial kit (Beyotime, Shanghai, China). The lysate was centrifuged for 15 min at 12,000 × g, and the concentration of proteins in the supernatant was determined with a bicinchoninic acid assay (Biosharp Life Sciences, China) per manufacturer’s instructions. Equal amounts of the protein lysates (20 mg) were separated by electrophoresis on 10% SDS-PAGE and then transferred to PVDF membranes (Millipore, Bedford, MA). After blocking in 5% (w/v) skim milk in TBST (0.1% Tween 20 in TBS) at room temperature for 1 h, the membranes were incubated overnight at 4°C with primary antibodies diluted 1:1,000 including p38, p-p38, JNK, p-JNK, ERK, p-ERK, NF-κB-p65, p-NF-κB-p65, IκBα, p-IκBα, TAK1, p-TAK1, c-fos, NFATc1, β-actin, and histone H3. After incubating at room temperature for 1 h with appropriate secondary antibodies, antibody reactivity and protein bands were detected and quantified by exposure in an Odyssey V3.0 image scanner (LI-COR, Inc., Lincoln, NE).

All data were from at least three independent experiments and are presented as means ± standard deviation (SD). Statistical analyses were performed by combining one-way analysis of variance (ANOVA) and Student’s t tests using SPSS 22.0 software (SPSS Inc.). Differences were considered to be statistically significant at *P < 0.05, **P < 0.01 and ***P < 0.001.

To determine if magnoflorine is protective against pathological osteolysis, we first assessed calvarial bone loss in mice treated with Ti particles. Three-dimensional reconstructions of micro-CT images revealed considerable bone erosion, with large deep pits on the calvarial surface, occurred in the Ti group (Figure 1A). Treatment with magnoflorine on the other hand, dose-dependently alleviated the degree of bone resorption with less bone erosion in the high-dosage group than in the low-dosage group (Figure 1A), as demonstrated by quantitative analyses of bone morphometric parameters (Figures 1B–E). Notably, craniums treated with Ti particles exhibited marked decreases in BMD and BV/TV, with increases in the number of pores as well as percentage of porosity compared with those from the sham controls. These changes were mitigated by administration of magnoflorine. Results from H&E and TRAP staining similarly revealed profound calvarial bone loss induced by Ti particles as well as diffuse infiltration of inflammatory cells, such as macrophages and lymphocytes, and abundant TRAP-positive osteoclasts along the surface of the eroded bone (Figures 2A–D). The Ti particle-treated group had a higher number of osteoclasts and greater OcS/BS (Figures 2G, H). Consistent with the micro-CT findings, magnoflorine administration dose-dependently reduced in the area of bone erosion and the accumulation of osteoclasts (Figures 2A–D, G, H). Immunohistochemical staining revealed that the vehicle-treated group had more TNF-α- and IL-1β-positive cells in cranial tissue than the sham group (Figures 2E, F). Nonetheless, the number of TNF-α- and IL-1β- positive cells was considerably suppressed by magnoflorine treatment, indicating the alleviated release of inflammatory factors (Figures 2I, J). Together, the in vivo results demonstrated that magnoflorine is protective against Ti particle-mediated osteolysis by suppressing osteoclast activity and inflammation.

The chemical structure of magnoflorine has been shown in Figure 3A. The effect of magnoflorine on the RANKL‐stimulated differentiation of BMMs into osteoclasts was then investigated in vitro. As shown in Figure 3B, in the control group, BMMs differentiate into numerous TRAP-positive osteoclasts that are multinucleated, well spread, and “pancake”-shaped. However, the number and area of osteoclasts exposed to magnoflorine with different dose (50, 100, and 200 μM) were dose-dependently decreased (Figures 3B–D). The formation of mature osteoclasts was reduced approximately 20% with 50 μM magnoflorine and was almost completely inhibited with 200 µM magnoflorine (Figures 3B–D). Furthermore, to exclude the possibility that the reduced osteoclast differentiation was due to cytotoxicity, we assessed the viability of the BMMs with the CCK-8 assay. As demonstrated in Figure 3E, no cytotoxicity effect was detected on BMM precursors after treatment with magnoflorine at concentrations of up to 200 μM, the concentration at which magnoflorine effectively suppressed osteoclastogenesis. Collectively, these results suggest that magnoflorine inhibits osteoclastogenesis without inducing cytotoxicity.

A well‐formed cytoskeletal podosomal actin belt in the periphery of osteoclasts is an indicator of precursor cell fusion (Touaitahuata et al., 2016). Therefore, we examined whether this was influenced by magnoflorine. The immunofluorescence results showed osteoclasts exhibited a discernible and well-shaped podosomal actin belt in the RANKL‐only treated group (Figure 4A), whereas those treated with magnoflorine exhibited disrupted podosomal actin belts, with dose-dependent decreases in both size and number (Figures 4B, C). Notably, cells treated with higher concentrations of magnoflorine were primarily mononucleated, further indicating defects in precursor cell fusion (Figure 4D). Collectively, these findings indicated that magnoflorine treatment remarkably impaired podosomal actin belt formation in osteoclasts and inhibited osteoclast precursor cell fusion in vitro.

Considering that the formation of an intact podosomal actin belt is indispensable to the subsequent formation of well-polarized F‐actin ring which is also critical for osteoclastic bone resorption activity (Takito et al., 2018), we therefore investigated whether magnoflorine could inhibit osteoclastic bone resorption. RANKL‐only treated groups showed extensive resorption on the surfaces of hydroxyapatite-coated culture dishes. By contrast, the sizes of the resorption areas were markedly and dose-dependently decreased by magnoflorine (Figures 4E, F). The area of bone resorption was reduced to 80% and 65% by 50 and 100 μM magnoflorine, respectively, and groups treated with 200 μM magnoflorine exhibited only small resorption pits that were randomly scattered around the dish (Figures 4E, F).

RANKL-mediated osteoclast differentiation and function are associated with the upregulation on expression level of specific marker genes. To unveil the anti‐osteoclastic effect of magnoflorine, we analyzed and quantified the expression of osteoclast-related genes in the mRNA level by qRT-PCR. The expression of genes encoding the key transcription factors NFATc1 and c-Fos, which are responsible for initiating osteoclast differentiation (Gohda et al., 2005; Monje et al., 2005), was dose‐dependently decreased in cells treated with magnoflorine in comparison with that in cells treated only with RANKL (Figures 5A–G). Furthermore, other osteoclast-specific markers such as CTR, CTSK, TRAP, V-ATPase d2, and DC‐STAMP demonstrated similar expression trend as NFATc1 and c-fos, on account that the genes encoding these are transcriptionally co-regulated by NFATc1 and c-fos. These data provided further evidence that magnoflorine suppresses osteoclast differentiation and bone resorption.

We performed qRT-PCR and ELISA to examine whether magnoflorine influences the release of proinflammatory cytokines induced by Ti particles. As shown in Figures 6A–F, TNF-α, IL-1β, and IL-6 were increased at both the mRNA and protein levels in BMMs exposed to Ti particles. Notably, treatment with magnoflorine dose-dependently attenuated this proinflammatory effect. These findings were consistent with the in vivo results, showing that magnoflorine administration effectively alleviates inflammatory response in Ti particle-stimulated BMMs.

To further clarify the underlying molecular mechanism through which magnoflorine inhibits the formation and function of osteoclasts, we performed Western blotting to assess RANKL-induced signaling pathways, in which MAPK and NF-κB signalings are known to play pivotal roles in RANKL-mediated osteoclastogenesis (Ghosh and Karin, 2002; Ikeda et al., 2004). BMMs were stimulated with 50 ng/ml RANKL for various times, and the protein and phosphorylation levels of various members of these signaling pathways were measured. The phosphorylation of p38, ERK1/2, and JNK peaked within 5, 30, and 10 min, respectively, after RANKL stimulation (Figures 7A–D). However, a 4-h pretreatment with magnoflorine suppressed the activation of these signaling proteins. Moreover, RANKL stimulated the phosphorylation and translocation of NF-κB p65, and this activation was also suppressed by magnoflorine treatment (Figures 8A–C). NF-κB p65 activation and translocation are regulated by IκBα phosphorylation and degradation. RANKL stimulation resulted in transient phosphorylation of IκBα in control cells, which peaked at 30 min and decreased by 60 min after treatment. This effect was mitigated by pretreatment with magnoflorine (Figures 7E, G). The transient phosphorylation was accompanied by IκBα degradation 30 min after RANKL administration, which was also suppressed by magnoflorine, resulting in the activation of the NF-κB pathway (Figures 7E, F). The activation of both MAPK and NF-κB signalings were primarily due to TAK1 activation, as the phosphorylation of TAK1 was also suppressed in the presence of magnoflorine (Figures 7E, H). Figures 8D–F showed that the expression of NFATc1 and c-Fos increased gradually, consistent with activation of MAPK and NF-κB signalings, but were suppressed by administration of magnoflorine in a time-dependent manner.