Rhinacanthin C Inhibits Osteoclast Differentiation and Bone Resorption: Roles of TRAF6/TAK1/MAPKs/NF-κB/NFATc1 Signaling

Rhinacanthin C is a naphthoquinone ester with anti-inflammatory activity, found in Rhinacanthus nasutus (L) Kurz (Acanthaceae). We found that rhinacanthin C inhibited osteoclast differentiation stimulated by the receptor activator of nuclear factor-κB ligand (RANKL) in mouse bone marrow macrophage cultures, although the precise molecular mechanisms underlying this phenomenon are unclear. In this study, we investigated the inhibitory mechanisms of rhinacanthin C in osteoclastogenesis. Rhinacanthin C suppressed RANKL-induced nuclear factor of activated T cells c1 (NFATc1) expression. Phosphorylation of ERK, JNK, and NF-κB, but not p38, was inhibited by rhinacanthin C, which also inhibited RANKL-stimulated TRAF6-TAK1 complex formation. Thus, the anti-osteoclastogenic effect of rhinacanthin C is mediated by a cascade of inhibition of RANKL-induced TRAF6-TAK1 association followed by activation of MAPKs/NF-κB; this leads to suppression of c-Fos and NFATc1, which regulate transcription of genes associated with osteoclast differentiation. In vivo, rhinacanthin C also reduced RANKL-induced osteoclast formation and bone resorption in mouse calvaria. Rhinacanthin C also suppressed LPS-stimulated osteoclastogenesis and bone resorption in vitro and in vivo. Rhinacanthin C may provide a novel therapy for abnormal bone lysis that occurs during inflammatory bone resorption.


Introduction
Bone formation by osteoblasts and bone resorption by osteoclasts are balanced to maintain bone homeostasis. Enhanced bone resorption can lead to impaired bone structure and bone fracture in osteoporosis, rheumatoid arthritis, and bone metastatic disease.
Herbal extracts are widely used as traditional medicines. Screening for phytochemicals with anti-osteoclastogenic activity may lead to the discovery and development of novel drugs to protect against or treat osteopenia and osteoporosis with minimal side effects. The shrub Rhinacanthus nasutus (L) Kurz (Acanthaceae) is distributed in Southeast Asia and is used in traditional medicines for the treatment of pneumonia, diabetes, hypertension, and skin diseases. The root and aerial portions of R. nasutus contain naphthoquinone esters, such as rhinacanthins C, D, N, and Q, which exhibit anti-inflammatory, anti-allergic, anti-tumor, and antiviral activities [6][7][8][9][10]. Recently, we reported anti-osteoclastogenic activity in the ethyl acetatesoluble fraction of a methanol extract of the R. nasutus root [11]. We isolated five components (rhinacanthins C, G, N, and Q, and rhinacanthone) from the ethyl acetate-soluble fraction of this plant and found rhinacanthin C is a potent inhibitor of RANKL-stimulated osteoclastogenesis in mouse bone marrow macrophage (BMM) cultures [12]. In this study, we investigated the molecular mechanism by which rhinacanthin C suppresses RANKL-induced osteoclastogenesis of BMMs and validated its efficacy in vivo.

Macrophage formation and phagocytosis assay
In vitro macrophage colony formation in mouse bone marrow cells was measured according to manufacturer instructions (Stemcell Technologies, Vancouver, Canada). Briefly, bone marrow cells (BMCs: 1 × 10 4 cells/mL) were cultured for 7 days in MethoCult (methylcellulose) in the presence of macrophage colony-stimulating factor (M-CSF, 10 ng/mL; R & D Systems, Minneapolis, MN) alone or with rhinacanthin C. Macrophage colonies were counted by phasecontrast microscopy. For the macrophage phagocytosis assay, bone marrow macrophages (BMMs) were cultured in 96-well plates for 3 days in the presence of M-CSF with or without rhinacanthin C. Fluorescein-conjugated zymosan A BioParticles (5 μg) were added and incubated for 30 min. After washing, cell-incorporated zymosan particles were visualized and counted by fluorescence microscopy.

Osteoclast formation from bone marrow cells
Osteoclasts were differentiated from bone marrow macrophages as described [13]. Briefly, BMCs collected from the femurs and tibias of 8 week-old ddY mice were cultured in α-MEM containing 10% FBS and macrophage colony-stimulating factor (M-CSF, 10 ng/mL; R & D Systems, Minneapolis, MN) for 3 days. The bone marrow macrophages were cultured in M-CSF alone or M-CSF plus RANKL (10 ng/mL; R & D Systems) with or without rhinacanthin C for 3 days. In experiments testing the effects of rhinacanthin C on the LPS-induced osteoclastogenesis, BMMs were primed with RANKL(1 ng/mL) for 24 h before LPS treatment.

Osteoclast differentiation and bone resorption activity
Tartrate-resistant acid phosphatase (TRAP) activity, an osteoclast differentiation marker, was measured as described [13]. Histochemical staining of TRAP was performed using a leukocyte acid phosphatase kit (Sigma-Aldrich). Cultured cells were fixed in 100% methanol for 1 min at room temperature and air-dried. After TRAP staining, TRAP-positive multinucleated (>3 nuclei) cells (MNCs) were imaged by phase-contrast microscopy. Bone resorption activity was assessed by resorption pit formation on dentin slices as described [14]. Briefly, BMMs cultured on dentin slices with M-CSF plus RANKL in the presence or absence of rhinacanthin C for 6 days were scraped off and the resorption pits were stained with acid-hematoxylin solution (Sigma-Aldrich). The absorbed pit areas were quantified with NSI Element imaging software (Nikon Instech, Tokyo, Japan).

Cell viability
Viability was evaluated using a cell proliferation assay kit (Promega, Madison, WI). At the end of culture, a 1/10 volume of reagent was added and incubated for 2 h. Absorbance was measured at 492 nm.

RANKL and endotoxin-induced bone loss in vivo
Eight-week-old male mice (ddy) were divided into three groups (n = 3-4/group). Vehicle (30% DMSO) or RANKL (0.7 mg/kg body weight) with or without rhinacanthin C (2 mg/kg body weight) was injected daily into the subcutaneous tissue overlying the calvaria under anesthesia with pentobarbital (Somnopentyl, Kyoritsu Seiyaku; 40-50 mg/kg body weight). Eight-week-old normal (wild-type) male or osteoprotegerin (OPG) knockout mice (OPG-/-) (C57BL/6) were used for the LPS experiments. Vehicle (30% DMSO) or lipopolysaccharide (LPS, 3.3 mg/kg body weight; Sigma-Aldrich) with or without rhinacanthin C (2 mg/kg body weight) was injected. The mice were sacrificed on day 5. Calvariae were dissected, fixed in 4% paraformaldehyde, and stained for TRAP activity. Three-dimensional reconstruction images of calvarial bone were obtained by micro-focal computed tomography (μCT) scanning (MCT-CB130MF; Hitachi Medical, Tokyo, Japan) at the Research Institute of Bone Structure Analysis (Yokosuka, Kanagawa, Japan). The X-ray tube was operated at an acceleration voltage of 60 kV and current of 100 μA. The geometry of the setup and resulting constraints in resolution lead to a voxel size of 32 × 32 × 32 μm. The 3-D images were analyzed using Tri-3D Bon software (Ratoc System Engineering, Tokyo, Japan). Trabecular bone volume (BV; mm 3 ), total tissue volume (TV; mm 3 ), bone volume fraction (BV/TV; %) were calculated in 3D mode. Static bone morphometric parameters trabecular separation (Tb.Sp; μm), and trabecular bone pattern factor (TBPf; 1/mm) were also calculated.

Ethics Statement
This study was performed according to the Fundamental Guidelines of the Ministry of Education, Culture, Sports, Science and Technology of Japan for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions. The Animal Care and Use Committee of Meikai University approved all animal experiments.
All animal experiments were performed under anesthesia to limit pain. The mice were sacrificed by spinal cord dislocation under anesthesia with pentobarbital.

Statistical analysis
Data were analyzed by ANOVA and are presented as the means ± SD of at least three independent experiments. In all analyses, P < 0.05 was taken to indicate statistical significance.

Rhinacanthin C inhibits RANKL-mediated osteoclast formation from mouse BMMs without cytotoxicity
Rhinacanthin C is a naphthoquinone derivative containing an alkyl side chain ( Fig 1A). We previously reported that rhinacanthin C is a strong inhibitor of RANKL-stimulated TRAP-positive multinucleated cell formation from mouse BMMs [12]. In this study, we investigated the effects of rhinacanthin C on osteoclast differentiation and bone resorption pit formation. Rhinacanthin C produced dose-dependent inhibition of RANKL-induced TRAP-positive multinuclear osteoclast formation and TRAP activity in BMM culture (0.25-2.0 μM) (Fig 1B to 1D). In our previous study, rhinacanthin C exhibited non-apoptotic cytotoxicity against tumor cells [12]. To investigate whether osteoclastogenesis inhibition by rhinacanthin C is due to its cytotoxicity, we measured the viability of RANKL-induced osteoclasts exposed to rhinacanthin C. Fig 1E shows that rhinacanthin C did not suppress viability even at 2.0 μM, at which concentration it abolished RANKL-stimulated osteoclast formation. Thus, the inhibitory effects of rhinacanthin C on osteoclastogenesis seemed not to be due to cytotoxicity.
Osteoclasts differentiated from the monocyte-macrophage lineage. Therefore, we examined the effects of rhinacanthin C on macrophage colony formation with M-CSF from BMCs. The Thus, macrophage formation and phagocytic function stimulated by M-CSF are not inhibited by rhinacanthin C. We further characterized the effects of rhinacanthin C on osteoclastogenesis by performing two experiments. First, BMCs were cultured with rhinacanthin C for 3 days in the presence of M-CSF, followed by RANKL plus M-CSF to induce osteoclast formation. In the second experiment, BMCs were cultured with rhinacanthin C and then the BMMs were cultured with RANKL plus M-CSF, also in the presence of rhinacanthin C. In experiment 1, RANKL-stimulated osteoclast formation remained essentially the same, regardless of pretreatment with rhinacanthin C. However, after RANKL stimulation, rhinacanthin C provided dose-dependent inhibition of osteoclast formation, as shown in experiment 2 (S2D Fig). These results suggest the inhibitory role of rhinacanthin C in osteoclastogenesis does not occur at the stage macrophage formation and function as an osteoclast progenitor but at the stage of RANKL-stimulated osteoclast formation step.
Next, we sought to identify what stage of osteoclast differentiation is more strongly affected by rhinacanthin C. Rhinacanthin C was added at various stages of osteoclastogenesis over 3 days ( BMMs. The numbers of TRAP-positive multinucleated cells and TRAP activity were higher in cultures treated on day 1 (1RC+2) than in those treated on day 2 (1+1RC+1). These results suggest the inhibitory effects of rhinacanthin C is reversible, and differentiation potential of BMMs treated with rhinacanthin C on day 1 is recovered over 2 days culture without rhinacanthin C. Indeed, we observed greater numbers of TRAP-positive multinucleated cells in the 1+1RC+1 culture when normal culture conditions were continued for one more day (data not shown). Taken together, these results suggest rhinacanthin C reversibly suppresses the early to middle stage of osteoclast differentiation.
We next examined the effects of rhinacanthin C on bone resorption activity, which reflects the acquisition of differentiated osteoclast function. BMMs were cultured on dentin slices with M-CSF and RANKL for 6 days in the presence or absence of increasing concentrations of rhinacanthin C, and the osteoclastic bone resorption pits were examined. As shown in Fig  2, rhinacanthin C inhibited RANKL-stimulated bone resorption in a dose-dependent manner. These results were consistent with the inhibition of osteoclast differentiation by rhinacanthin C.

Rhinacanthin C suppresses induction of RANKL-induced proteins involved in osteoclast differentiation and function
NFATc1 is a master transcription factor in osteoclastogenesis. c-Fos, a major component of transcription factor AP-1, induces NFATc1 expression to regulate osteoclast differentiation [5,15]. RANKL significantly increased NFATc1 and c-Fos protein levels after 2 days, whereas rhinacanthin C suppressed the RANKL-stimulated induction of both transcription factors (Fig  3). Integrin β3 and c-Src are important for actin ring formation, which is essential for bone resorption [16]. RANKL increased c-Src and integrin β3 expression, while rhinacanthin C suppressed these effects (Fig 3).

Rhinacanthin C prevents RANKL-induced bone resorption in vivo
To determine whether rhinacanthin C prevents RANKL-induced bone loss in vivo, RANKL was administered once daily for 5 days with or without rhinacanthin C on the calvaria. We then stained for TRAP activity and performed morphometric analysis of calvarial bone by micro-CT. The TRAP-positive area increased in the RANKL-injected calvaria versus the vehicle-injected control (Fig 6A and 6B). Simultaneous administration of rhinacanthin C reduced the RANKL-induced TRAP-positive area. Representative micro-CT images of RANKL-treated calvaria showed that RANKL treatment increased excavated bone destruction, whereas rhinacanthin C-injected calvaria showed less bone destruction in comparison to RANKL-injection alone (Fig 6C). Micro-CT analysis of RANKL-treated calvaria showed decreased BV/TV, and increased trabecular separation (Tb.Sp) compared with vehicle-treated control calvaria (Fig 6D  and 6E). On the other hand, rhinacanthin C treatment significantly increased BV/TV and decreased trabecular separation, suggesting rhinacanthin C prevented RANKL-induced calvarial bone destruction. TBPf is the ratio of convex structures to concave structures on the trabecular surface, reflecting the connectivity of trabecular bone. TBPf was increased in the RANKL-treated group versus the control, suggesting RANKL destroys the microarchitecture of trabecular bone (Fig 6E). Rhinacanthin C reduced the RANKL-induced increase of TBPf. Thus, rhinacanthin C prevents osteoclastic bone resorption in vitro and in vivo.

Rhinacanthin C inhibits LPS-induced osteoclastogenesis and bone resorption
Bacterial infection is associated with bone destruction in periodontitis. LPS, a cell wall component of Gram-negative bacteria, also induces bone resorption [23][24][25]. We also demonstrated the effects of rhinacanthin C on LPS-induced osteoclastogenesis from BMMs (Fig 7) and bone resorption of mouse calvaria (Fig 8). RANKL priming is needed for LPS-induced osteoclastogenesis [26]. We treated BMMs with RANKL (1 ng/mL) for 24 h followed by LPS treatment with or without rhinacanthin C for 3 days (Fig 7). LPS significantly increased TRAP activity (Fig 7B) and the numbers of MNCs (Fig 7A and 7C) in RANKL-primed BMMs but not in non-primed BMMs. Rhinacanthin C provided dose-dependent inhibition of LPS-stimulated osteoclastogenesis from BMMs.
Finally, we tested the inhibitory effects of rhinacanthin C on LPS-stimulated calvarial bone erosion in vivo using normal and OPG-/-mice. Osteoprotegerin-deficient (OPG-/-) mice exhibit osteoporosis due to enhanced osteoclastogenesis caused by a lack of soluble decoy receptor for RANKL [27,28]. As shown in Fig 8, LPS significantly increased TRAP staining of whole calvariae in normal and OPG-/-mice (Fig 8A). Simultaneous administration of rhinacanthin C reduced LPS-induced osteoclast formation. μCT analysis (BV/TV, Tb.Sp, and TBPf) showed that LPS induced bone destruction in OPG-/-mice (Fig 8B). Rhinacanthin C ameliorated LPS-induced calvarial bone resorption of normal and OPG knockout mice. Thus, rhinacanthin C inhibits LPS-induced and RANKL-induced osteoclastogenesis.

Discussion
Identification and characterization of natural anti-osteoclastogenic compounds from herbal sources support the development of novel therapeutics for patients suffering from abnormal bone lysis and exploration of the regulatory mechanisms underlying osteoclast formation and activation. Previously, we found that rhinacanthin C is a potent inhibitor of RANKL-induced osteoclast formation from BMMs [12]. In this study, we demonstrated the inhibitory mechanisms of rhinacanthin C and its effects in vivo.
Rhinacanthin C suppressed the osteoclast-specific master transcription factor, NFATc1, which regulates osteoclast specific/related gene expression. Indeed, c-Src and integrin β, which are late osteoclastogenesis genes involved in bone resorption, were down-regulated by Fig 6. Protective effects of rhinacanthin C on RANKL-induced mouse calvarial osteolysis. Vehicle (30% DMSO) or rhinacanthin C (RC, 2 mg/kg body weight) with or without RANKL (0.7 mg/kg body weight) was daily injected into the subcutaneous tissue overlying the calvaria of 8-week-old ddy mice. The mice were sacrificed on day 5. A, TRAP staining of whole calvaria (upper) and high magnification of TRAP stain on calvaria (lower). Bar, 400 μm. B, TRAP + area was measured by densitometry using Image J.C, Three-dimensional micro-CT image of calvaria. D, Bone volume/total tissue volume ratio (BV/TV). E, Trabecular separation (Tb.Sp). F, Trabecular bone pattern factor (TBPf). *P < 0.05, **P < 0.01. doi:10.1371/journal.pone.0130174.g006 rhinacanthin C. Therefore, suppression of TRAP activity and pit formation by rhinacanthin C are attributed to reduced levels of osteoclast specific/related gene expression governed by NFATc1. c-Fos induces NFATc1 expression and is an essential regulator of osteoclastogenesis. c-Fos knockout mice develop osteopetrosis due to a lack of osteoclast formation [5,15]. Suppression of NFATc1 accumulation by rhinacanthin C may be caused by down-regulation of c-Fos. The NFATc1 promoter is initially activated by NF-κB and AP-1 in the presence of RANKL [29]. NFATc1 binds to its own promoter, followed by NFATc1-mediated auto-amplification of gene induction [30]. c-Jun and c-Fos (components of AP-1) are phosphorylated and activated by JNK and ERK, respectively [31]. Rhinacanthin C inhibited the RANKL-induced phosphorylation of ERK and JNK. Thus, rhinacanthin C may suppress AP-1 activation via inhibition of MAPKs. Rhinacanthin C also suppressed RANKL-induced phosphorylation of IκB and NF-κB/p-65, suggesting that the inhibitory effects of rhinacanthin C on osteoclast differentiation may also be caused by inhibition of NF-κB activity. Thus, rhinacanthin C suppresses the early step of osteoclast differentiation via inhibition of RANKL-stimulated NF-κB and MAPK activation.
As a consequence of RANKL binding, RANK activates TRAF6 to initiate complex formation with TAK1, which facilitates NF-κB and MAPK activation [22]. TRAF6 siRNA and  TRAF6 decoy peptides inhibit the formation of TRAP-positive multinucleated cell and bone resorption [32]. TRAF6 knockout mice exhibit severe osteopetrosis with defects in bone remodeling caused by impaired osteoclast function [22,33]. Our immunoprecipitation assay revealed that association of TRAF6 and TAK1 was increased by RANKL, while rhinacanthin C suppressed this effect. Rhinacanthin C may disrupt intracellular signal processing by targeting TRAF6-TAK1 complex formation, thus impairing osteoclast differentiation. The direct targets of rhinacanthin C remain unknown. Recently, it was reported that (+)-Vistin A and Norisoboldine inhibit osteoclast differentiation and prevent TRA6-TAK1 formation [34,35]. Whether these natural compounds and rhinacanthin C target the same molecule will be the topic of future research.
Consistent with its anti-osteoclastogenic effects in vitro, rhinacanthin C also inhibits RANKL-induced osteoclast formation and bone destruction in vivo. Furthermore, caldecrin inhibits LPS-stimulated osteoclast formation from BMMS and calvarial bone resorption. LPS is recognized by Toll-like receptor (TLR), which triggers recruitment of TRAF6 and TAK1, leading to activation of the NF-κB pathway [36,37]. TRAF6 knockout cells fail to respond to LPS stimulation [38]. Rhinacanthin C might inhibit LPS-induced osteoclastogenesis by suppressing TRAF6-TAK1-NF-κB signaling as it does in RANKL-induced osteoclastogenesis. These results and its reversible effects on osteoclast formation suggest the potential utility of rhinacanthin C in the treatment of conditions associated with abnormal bone lysis such as osteoporosis and inflammation-induced bone loss (rheumatoid arthritis and periodontal bone erosion).
Rhinacanthins are composed of naphthoquinone with variable side chains. Among the rhinacanthins (rhinacanthin C, D, G, N, and Q), rhinacanthin C has the most potent anti-osteoclastogenic activity [12] (data on rhinacanthin D are unpublished). Rhinacanthin C has a 2,6-dimethyl-2,6-octadienoic acid ester acid side chain. The relationship between side chain structure and anti-osteoclastogenic activity are a topic of future research.
In summary, we have demonstrated that rhinacanthin C suppresses RANKL-stimulated TRAF6-TAK1 complex formation and activation of the MAPK/NF-κB/NFATc1 pathways, thereby suppressing expression of osteoclast marker proteins, finally resulting in impaired osteoclast differentiation and activation of bone resorption in cultured BMMs. Rhinacanthin C also suppresses LPS-induced osteoclast formation and bone destruction in vivo. The direct targets of rhinacanthin C that mediate its anti-osteogenic activity warrants further study. A, Rhinacanthin C (RC, 1 μM) was added to BMM culture at various time points under RANKL stimulation. Black column indicates RC treatment period. Culture medium was exchanged daily. Culture conditions; 3, Continuous treatment without rhinacanthin C for 3 days; 3RC, Continuous treatment with rhinacanthin C for 3 days; 1+2RC, Rhinacanthin C treatment on days 2 to 3; 2+1RC, Rhinacanthin C on day 3; 1RC+2, Rhinacanthin C on day 1; 1+1RC+1, Rhinacanthin C on day 2; 2RC+1, Rhinacanthin C on days 1 to 2. B, TRAP staining of osteoclasts cultured by various conditions as described in (A). Bar, 100 μm. TRAP activity (C) and number of TRAP-positive multi-nuclear cells (D) of osteoclasts are shown. Ã P < 0.05, ÃÃ P < 0.01 (TIF)