Neobavaisoflavone inhibits osteoclastogenesis through blocking RANKL signalling‐mediated TRAF6 and c‐Src recruitment and NF‐κB, MAPK and Akt pathways

Abstract Psoralea corylifolia (P corylifolia) has been popularly applied in traditional Chinese medicine decoction for treating osteoporosis and promoting fracture healing since centuries ago. However, the bioactive natural components remain unknown. In this study, applying comprehensive two‐dimensional cell membrane chromatographic/C18 column/time‐of‐flight mass spectrometry (2D CMC/C18 column/TOFMS) system, neobavaisoflavone (NBIF), for the first time, was identified for the bioaffinity with RAW 264.7 cells membranes from the extracts of P corylifolia. Here, we revealed that NBIF inhibited RANKL‐mediated osteoclastogenesis in bone marrow monocytes (BMMCs) and RAW264.7 cells dose dependently at the early stage. Moreover, NBIF inhibited osteoclasts function demonstrated by actin ring formation assay and pit‐formation assay. With regard to the underlying molecular mechanism, co‐immunoprecipitation showed that both the interactions of RANK with TRAF6 and with c‐Src were disrupted. In addition, NBIF inhibited the phosphorylation of P50, P65, IκB in NF‐κB pathway, ERK, JNK, P38 in MAPKs pathway, AKT in Akt pathway, accompanied with a blockade of calcium oscillation and inactivation of nuclear translocation of nuclear factor of activated T cells cytoplasmic 1 (NFATc1). In vivo, NBIF inhibited osteoclastogenesis, promoted osteogenesis and ameliorated bone loss in ovariectomized mice. In summary, P corylifolia‐derived NBIF inhibited RANKL‐mediated osteoclastogenesis by suppressing the recruitment of TRAF6 and c‐Src to RANK, inactivating NF‐κB, MAPKs, and Akt signalling pathways and inhibiting calcium oscillation and NFATc1 translocation. NBIF might serve as a promising candidate for the treatment of osteoclast‐associated osteopenic diseases.


| INTRODUC TI ON
P corylifolia has been commonly used in herbal decoction based on the traditional Chinese medicine theory for its effectiveness in treating osteoporosis and promoting fracture healing. Abundant evidence has been emerging to support its efficacy for bone formation promotion and osteoporosis alleviation. 1,2 However, extracts from P corylifolia are bioactive mixtures and the specific effective monomers remain still by and large unknown. Figuring out potential bioactive products would pave way for the exploration of novel drugs for bone metabolism disorders.
Currently, a comprehensive two-dimensional cell membrane chromatographic/C18 column/time-of-flight mass spectrometry (2D CMC/C18 column/TOFMS) system analysis system has been developed for screening potential bioactive components from Chinese herbal medicines, based on the binding or bioaffinity of unidentified components to their membrane targets. 3,4 In our study, ethanol extracts from the seeds of P corylifolia were screened by 2D RAW 264.7 cells CMC/C18 column/TOFMS system, and neobavaisoflavone (NBIF) was identified as a potentially bioactive chemical component binding to the membrane of RAW 264.7 cells.
Bone metabolism homeostasis relies on the duel competitive role of osteoblasts and osteoclasts. 5 Disturbance of this delicate balance as a result of excessive bone resorption by overactivated osteoclasts contributes to the occurrence of various metabolic bone diseases, like post-menopausal osteoporosis (PMOP) and rheumatoid arthritis (RA). 6 Therefore, inhibiting overactivated osteoclastogenesis could be an effective strategy to find a cure for pathological bone loss in these diseases. [7][8][9][10] Osteoclasts originate from the haematopoietic cell line and differentiate from bone marrow monocytes that are stimulated by two indispensable cytokines, macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor-κB ligand (RANKL). 11,12 The M-CSF binding to c-Fms keeps the survival and proliferation of BMMCs and pre-osteoclasts and initiates BMMCs differentiation into osteoclast precursors, while the conjunct of RANKL and RANK leads to terminal differentiation into mature osteoclasts. 13 RAW 264.7 cells, a widely used mouse monocytic cell line, express RANK and have been shown to differentiate into functional osteoclasts upon recombinant RANKL stimulation. 14 In addition, the binding of RANKL to RANK recruits tumour necrosis factor receptor-associated factors (TRAFs), of which TRAF6 is the most important one. 15 It is worth noting that c-Src is also recruited by activated RANK to organize osteoclast's cytoskeleton and function. 16 As a result, several downstream pathways are subsequently tranduced by the activated RANK-TRAF6 complex and RANK-c-Src conjunct.
The well-recognized downstream signalling pathways involving in this cascade include NF-κB (IκB, P50, P52, Rel A, RelB, c-Rel) and MAPKs (ERK, JNK, P38) activated by TRAF6 recruitments, and Akt induced by c-Src recruitments. [17][18][19] RANKL-RANK interactions induced activation of Akt signalling pathway also triggers cytoplasmic calcium released by calcium oscillation, which ultimately increases the expression and translocation of NFATc1. [20][21][22] NFATc1 is the core transcriptional factor in the differentiation and maturation of osteoclasts, and it dominates the expression of multiple osteoclastogenesis-related genes, including tartrate-resistant acid phosphatase (TRAP), matrix metalloproteinase (MMP)-9, cathepsin K, calcitonin receptor (CTR), all of which are responsible for the terminal function of osteoclasts. [23][24][25][26] Recently, NBIF, an isoflavonoid originally isolated from the seeds of P corylifolia, has attracted much attention because of its anti-inflammation, anti-cancer and anti-oxidation properties. 27,28 Inflammation plays a pivotal role in both osteoclastogenesis and osteoporosis. 29 A recent study reported that NBIF inhibits inflammatory mediators in activated RAW264.7 macrophages. 30 Intriguingly, NBIF was previously reported to stimulate osteogenesis in vitro by P38 MAPK pathway. 31 However, the role of NBIF in osteoclastogenesis remains unclear. Here in our report, we demonstrated that NBIF might serve as an effective osteoclastogenesis inhibitor and an osteogenic promotor, ameliorate ovariectomy-induced bone loss in mice.

| Preparation of samples and RAW264.7-CMC column
Firstly, Psoralea corylifolia was smashed into powder by a pulverizer.
Then, the powder was mixed with 60% ethanol at 60-80°C water bath along with ultrasonic fragmentation for 2 hours. The ethanol extract was condensed by the rotary evaporator to 1 g/mL. After that, the ethanol extract was filtered by 0.2 μm filter membrane and stocked at 4°C for further use.
For RAW264.7 cell membrane preparation, 3.5 × 10 7 RAW264.7 cells were harvested and washed by PBS for 3 times and centrifuged at 110 × g for 10 minutes. Then, PBS was added to suspend cells, after which disrupted by an ultrasonic processor of 3 cycles (3 seconds for 400 W and 15 seconds for internal each cycle). The homogenate was then subjected to centrifugation at 1000 × g for 10 minutes. The supernatant was collected for further centrifugation at 12 000 × g for 20 minutes. The precipitation was collected and suspended in 5 mL PBS, which was regarded as cell membrane suspension.
For RAW264.7 cell membrane stationary phase (CMSP) preparation, 0.04 g activated silica under vacuum and agitation condition was mixed with RAW264.7 cell membrane suspension to absorb cell membrane. After incubation at 4°C overnight, the CMSP was washed by PBS for 3 times, centrifuging at 110 × g for 5 minutes.
The pellet was collected, suspended in PBS and put into a column.
The flow rate of packing was controlled by a linear program, and the equilibrated flow rate is 0.2 mL/min at 37°C. Then, the column was stored in PBS at 4°C.

| Identification of the bioactive compound from P Corylifolia
As previously described, 3

| Animal experimental designs
All animal experiments were conducted following the standards of Bioethics Committee in Changhai Hospital (SYXK 2015-0017).
Mice were randomly distributed to 3 groups (5 mice in each group): sham group, OVX group and NBIF group. Mice in the OVX group were treated with normal saline, while mice in the NBIF group were treated with NBIF. All mice were anaesthetized with 5% chloral hydrate. Ovaries were merely exposed from the surrounding adipose tissue in the sham group, and bilateral ovaries were removed in both OVX group and the NBIF group. After 1 day for post-operative recovery, OVX group and NBIF group were received intraperitoneal (i.p.) injection with normal saline and NBIF (30 mg/kg), respectively.
Six weeks later, all groups were killed by an overdose of chloral hydrate, after which bilateral femurs and blood were collected for further measurements.

| Cytotoxicity assays
The CCK-8 assay was executed according to the standard proto-

| Osteoclastogenesis assays
BMMCs were separated from the bilateral femurs of mice. RAW 264.7 cells and BMMCs were then seeded on 96-well plates
Differentiated cells were fixed and stained with alkaline phosphatase (ALP) staining at day 7 and stained with Alizarin Red staining at day 21.

| F-actin staining
For fibrous actin (F-actin) staining, RAW264.7 cells in 96-well plates (1.5 × 10 4 cells/well) induced by M-CSF (30 ng/mL) and RANKL (50 ng/mL) for 7 days were fixed with 3.75% formaldehyde in cold PBS for 15 minutes. After that, cells were subject to 0.5% Triton X-100 for 3-minute permeabilization followed by non-fat dry milk used for a blockade of non-specific binding. Then, permeabilized cells were incubated in rhodamine-conjugated dilution with 1% bovine serum albumin (BSA) for 20 minutes at room temperature.
Before being photographed, cells were counterstained by DAPI for 10 minutes. The number of F-actin rings and the number of osteoclasts were counted for further statistical analysis. Two days later, cell remains on bone biomimetic synthetic surface were cleaned up and the surface was rinsed by swirling in distilled water and subjected to air dry afterward. Pit-formation areas on the biomimetic plate surface were captured by Light Microscope (OLYMPUS-BX53) and measured by Image J software.
Micro-CT (Skyscan, Antwerp, Belgium) was used to scan each distal femoral metaphysis. The analysis conditions met the following parameters: the voltage was 80 kV, the electric current was 124 μA and the resolution was 8 μm. Then, scans were integrated into 2D and 3D images. Quantitative data of femur parameters were obtained as follows: bone surface/tissue volume (BS/TV), bone volume/tissue volume (BV/ TV), trabecular bone number (Tb.N) and bone mineral density (BMD).
These parameters were calculated using the built-in software.

| Histological analysis
Femurs were isolated, fixed in 4% PFA for 48 hours and decalcified for 2 weeks. Two weeks later, these femurs were embedded in paraffin, and then, 4 μm thick sections were cut into slices by the microtome and stained by haematoxylin and eosin (H&E), TRAP staining kit and stained for Osteocalcin (OCN). Light microscope (OLYMPUS-BX53) was used to observe and photograph the femur trabecular area. TRAP-positive multinucleated osteoclasts with 3 or more nuclei were calculated.

| Serum biochemistry analysis
Blood was sampled from mouse eyes. Sera were then collected after 1000 g centrifugation for 15 minutes. According to the general instructions, ELISA kit was employed for biochemical detection of serum carboxy-terminal telopeptide of type-l collagen (CTX-1), and Tartrate-resistant acid phosphatase 5b (TRAcp5B) (IDS plc, Boldon, UK), OCN (Immutopics, San Diego, CA, USA).

| Immunofluorescence assays
For RANK and c-Fms immunofluorescence staining, BMMCs cells were isolated and seeded on the 6-well plates at a density of

| Cytoplasmic Ca 2+ measurement
BMMCs cells were isolated and cultured on the 48-well plates

| Western blot
Western blot was used to detect the expression of osteoclastogenesis-associated markers genes (TRAP, CTR, MMP-9 and

| Statistical analysis
Total results were presented as mean ± SD. Two-tailed non-paired Student's t test statistical analysis was performed to compare 2 groups, and 3 or more groups were compared by one-way ANOVA.

| NBIF is a bioactive compound extracted from P corylifolia
Following a novel strategy to identify potential active components from interested traditional Chinese herbs, 3 we finally discovered that NBIF was a potential bioactive compound which could be extracted from P corylifolia.
As mentioned above in the Methods, using 2D RAW264.7 cells/ CMC/C18 column/TOFMS system ( Figure 1A-B), we discovered that only one component extracted from the seeds of P corylifolia displayed a fine affinity with RAW 264.7 cell membrane, represented by strong retention behaviour and the peak reaching time at about 5 minutes ( Figure 1C). We identified that the molecular formula of this compound was C 20 H 18 O 4 , and then, we searched and compared the components of P corylifolia on the Traditional Chinese Medicine Integrated Database. Finally, we confirmed that this bioactive ingredient was NBIF ( Figure 1D). Moreover, through literature research, we found no publication reporting the efficacy of NBIF related to osteoclastogenesis. Hence, we focused on NBIF in our following work.

| NBIF suppressed osteoclastogenesis in vitro
The

| NBIF inhibited the function of osteoclasts in vitro
Pit-formation assay and F-actin rings formation assay were carried out to detect whether osteoclasts' functions were impaired.
Pit-formation assay showed that bone resorption activity was dose-dependently attenuated by NBIF ( Figure 3A). Furthermore, F-actin rings were regarded as the characteristic structure of mature osteoclasts. To further determine the inhibitory effect of NBIF on osteoclast function, immunofluorescence staining of F-actin was performed in RAW264.7 cells which were exposed to NBIF treatments. Similar to the prior inhibitory effects, it showed that NBIF significantly inhibited the formation of F-actin ring structures, demonstrated by the decrement of F-actin numbers compared to total osteoclast numbers ( Figure 3B). In brief, the results showed that the function of osteoclasts could be notably impaired by NBIF.

| NBIF inhibited osteoclastogenesis at the early phase
To further identify at which specific phase NBIF displayed its inhibitory effect on osteoclastogenesis, RANKL-induced RAW264.7 cells were incubated with NBIF (8 μM) at different time-points and durations (day 1-3, day 3-5, day 5-6 and day 1-6). TRAP staining illustrated that osteoclast differentiation was more strongly inhibited by NBIF treatments in the first several days than that in the later days, implying that NBIF mainly suppressed osteoclast differentiation at the early phase ( Figure 3C).

| NBIF showed little effects on M-CSFstimulated BMMCs proliferation
Our CCK-8 assay results showed that BMMCs proliferation induced by M-CSF was not significantly influenced by NBIF (below 8 μM), indicating that NBIF exerted little influence on BMMCs proliferation ( Figure S1).

| NBIF showed little influence on RANK and c-Fms expressions during osteoclastogenesis
To assess whether NBIF influenced RANK (receptor of RANKL)  Figure S2A). Similar to the findings of RT-PCR, immunofluorescence assays confirmed that RANK and c-Fms fluorescent intensities were not significantly up-regulated or downregulated when exposed to NBIF ( Figure S2B). To sum up, these results implied that NBIF had no effect on RANK and c-Fms during osteoclast formation.

| NBIF exerted osteogenic effects on the osteoblastic differentiation of bone marrow mesenchymal stem cells (BMSCs)
To confirm the reported osteogenic effects of NBIF, we performed ALP, Alizarin Red staining after the osteogenic induction with or without NBIF (1 μM). Consistent with the published data, 31 ALP staining and Alizarin Red staining indicated that NBIF stimulated osteogenesis in BMSCs in vitro ( Figure S3). Thus, NBIF might serve as both as osteoclastogenesis inhibitor and osteogenic promotor at the same time.

| NBIF inhibited RANKL-induced NF-κB and MAPKs pathways activation and interrupted RANK-TRAF6 interaction in osteoclastogenesis
The activation of RANKL-stimulated NF-κB pathway is essential for cells. The results indicated that TRAF6 was recruited and bonded with RANK after RANKL stimulation, whereas NBIF notably blocked this process, suggesting that RANKL-induced recruitment of TRAF6 was impaired when exposed to NBIF ( Figure 4C). Taken together, the above findings suggested that NBIF RANKL-induced NF-κB, MAPKs pathways activation and abrogated the interaction of RANK and TRAF6 during osteoclastogenesis.

| NBIF inhibited RANKL-induced Akt pathway activation and blocked RANK-c-Src interaction
Activation of the Akt pathway stimulated by RANKL also plays a crucial role in the process of osteoclastogenesis. Western blot revealed that the phosphorylation of Akt was significantly elevated by the induction of RANKL, whereas remarkably down-regulated by NBIF treatments ( Figure 5A). Moreover, Co-IP was employed to determine whether NBIF disrupted the interaction between RANK and c-Src stimulated by RANKL incubation. Co-IP displayed that c-Src effectively bound to RANK after RANKL induction, while the recruitment was disrupted by the treatment of NBIF ( Figure 5B). Ca 2+ oscillation is also provoked by Akt signalling activation and serves as a vital messenger for signalling transduction of osteoclast differentiation.
Here, the fluctuation of the Ca 2+ level was detected to further identify the role of NBIF in calcium signalling. The results showed that RANKL induction stimulated cytoplasmic Ca 2+ release, while this effect was significantly diminished by NBIF intervention ( Figure 5C). immunofluorescence staining was applied to determine the nuclear translocation of NFATc1. As immunofluorescence staining displayed, the translocation of NFATc1 from cytoplasm to nucleus was remarkably increased by RANKL, while NBIF significantly attenuated this RANKL-mediated nuclear translocation of NFATc1 ( Figure 6A and B). To investigate the role of NBIF in the expression of these aforementioned marker genes, RAW264.7 cells induced by RANKL were exposed to increasing concentration of NBIF (0, 2, 4 and 8 μM). Cells were then harvested for Western blot analysis at day 7 after the induction. We found that the expression of all these marker genes, during the period of RANKL-induced osteoclastogenesis, was drastically decreased by NBIF treatments in a dose-dependent way ( Figure 6C). To sum up, the results implied that NBIF might exert inhibitory effects on osteoclastogenesis by suppressing osteoclastogenesis-related genes expression.

| NBIF attenuated bone loss by inhibiting osteoclast activation and promoting osteogenesis in ovariectomized mice
To further validate the efficacy of NBIF in vivo, ovariectomy was performed in mice to mimic the pathological bone loss after the withdrawal of oestrogen. Micro-CT showed that heavy trabecular bone loss was observed in the OVX group compared with that in the sham group, while NBIF treatment notably prevented the bone loss from ovariectomy as showing with the increased values of Tb.N, BMD, BS/TV and BV/TV in the 2-and 3-dimensional photographs ( Figure 7A). Compared with the OVX group, the NBIF group also illustrated higher bone mass maintenance presented by H&E staining images ( Figure 7B). Furthermore, we carried out TRAP staining to determine whether osteoclasts were inhibited by NBIF intervention. The results displayed that the distal femur sections from ovariectomized mice were spread with increased TRAP-positive multinucleated cells in the trabecular area, while fewer osteoclasts were found in the NBIF group than that in the OVX group ( Figure 7C). Besides, the level of the serum CTX-1 and TRAcp5B was measured to evaluate whether NBIF intervention ameliorated ovariectomy-associated bone loss via restraining osteoclast activity. In comparison with the OVX group, remarkably decreasing serum levels of these two markers were detected in the NBIF group, reflecting the inhibitory effect of NBIF on osteoclasts' function ( Figure 7D). Surprisingly, the number of OCNpositive osteoblasts and OCN serum level also saw a significant rise in the NBIF group in contrast to the OVX group, which indicated that NBIF also promoted osteogenesis in vivo ( Figure S4A and B).
Collectively, all these results suggested that NBIF attenuated OVXassociated pathological bone loss by disturbing osteoclast activity and promoting osteogenesis.

| D ISCUSS I ON
In this study, by the 2D RAW264.7/CMC/C18 column/TOFMS system, we discovered NBIF, a novel osteoclastogenesis-suppressive compound extracted from P corylifolia. As far as we know, it is for the first time that NBIF is reported for its inhibitory effects Disturbance to this delicate balance by highly activated osteoclasts, particularly, leads to excessive bone resorption and impaired bone microstructure, having been proved to be responsible for various osteopenic disorders, including PMOP. 35 As a result, inhibiting overactivated osteoclastogenesis could be an effective strategy to find a cure for pathological bone loss in these diseases.
In this study, we decided to explore the inhibitory effects of NBIF on osteoclastogenesis and its possible molecular mechanism.
In vitro, we found little cytotoxic effects on RAW264.7 cells and Application of P38 inhibitor abrogated the osteogenesis-stimulating effects induced by NBIF on the expression of osteogenic marker genes (Runx2, Osx). Consistent with this finding, our study discovered that NBIF also promoted osteogenesis in vivo, presented by more OCN-positive mature osteoblasts in the distal femurs and higher OCN level in the serum. It is well known that the bone remodelling process pathologically accelerates by ovariectomy-associated oestrogen deficiency. Under this circumstance, overactivated osteoclastogenesis is insufficient to couple equivalent bone formation by osteoblasts at a period of a single remodelling cycle, resulting in a net bone loss. 43 Intriguingly, our results discovered that NBIF rescued ovariectomy-induced bone loss and acted to correct the unbalanced bone coupling disorders by inhibiting osteoclastogenesis and promoting osteogenesis simultaneously. However, the molecular targets of NBIF with regard to osteoclastogenesis still need to be clarified in the future.
In summary, 2D RAW 264.7 cells CMC/C18 column/TOFMS system is useful in discovering potential natural products from traditional Chinese medicine. With the exploitation of this system, we explored NBIF and validated that NBIF might serve as an osteoclastogenesis inhibitor and also as an osteogenetic inducer at the same time, for the treatment of PMOP. As such, NBIF would be a promising and useful medical candidate for the treatment of osteoporosis and other osteolytic bone diseases.