LY900009 Regulates RANKL-Induced Osteoclast Formation and LPS-Induced Bone Resorption

To investigate the suppressive function of LY900009, a potent-secretase inhibitor, on RANKL-induced osteoclastogenesis. The cytotoxicity of LY900009 was evaluated. The suppressive effect and possible molecular mechanism of LY900009 on RANKL-induced osteoclastogenesis was evaluated both in vitro and in vivo. The IC50 of LY900009 was 2.93 mM. LY900009 treatment at different doses (100 nM, 200 nM, and 400 nM) effectively reduced osteoclast formation (number and arear) in a dose-dependent manner. The qPCR result shows that LY900009 attenuates RANKL-induced osteoclast formation and NFATc1 protein expression. The in vivo experiments demonstrated the inhibitory effect of LY900009 on LPS-induced bone resorption. LY900009 could potently inhibit osteoclastogenesis and bone resorption by down-regulating Notch/MAPK/Akt - mediated NFATc1 reduction in vitro. In accordance with the in vitro observations, we conrmed that LY900009 attenuated LPS-induced osteolysis in mice. In conclusion, our ndings indicate that Notch was a potential therapeutic target which could be used for osteolytic diseases treatment.


Introduction
Bone metabolism is a dynamic process that involves the simultaneous processes of bone resorption and formation 1 ; 2 . Thus, osteoblast and osteoclast disequilibrium can undermine bone integrity and normal functionality, promoting a series of pathological lytic bone disorders [3][4][5] . Thus, recent research has centered on the investigation and identi cation of novel agents (synthetic or natural) which may inhibit the pathological formation of osteoclasts and bone resorption 6 ; 7 .
Osteoclasts (OCs) are a type of multinucleated giant cell derived from macrophage precursors. OC differentiation is the rst stage of the bone resorptive process, which requires stimulation of receptor activator of NF-κB ligand (RANKL) and macrophage colony stimulating factor (M-CSF). Previous studies have shown that RANKL plays an essential role in OC multinucleated formation 8 ; 9 . After the interaction between RANKL and receptor activator of NF-κB (RANK), various downstream signaling pathways (e.g., MAPKs and PI3K/Akt) are activated. These speci c pathways are able to upregulate the key regulator of the osteoclast process: nuclear factor of activated T cells c1 (NFATc1) [10][11][12][13] .
In previous studies, Notch signaling has been shown to play an essential role in multiple developmental pathways 14 . The interaction between Notch receptors and Delta-like1/3/4 (Dlk 1/3/4) protein causes the release of tumor necrosis factor a-converting enzyme (TACE), and γ-secretase 15 . Former research con rmed the inhibition effect of γ-secretase down-regulation on Notch receptor cleavage 16 . Both enhancive and prohibitive effect of Notch on RANKL-induced bone resorption was reported in relative research 17 , suggesting that it may represent a promising targetable pathway for the development of antiresorptive agents 18 ; 19 . Moreover, LY900009, a γ-secretase inhibitor suppressing Notch cleavage, has not yet been con rmed its ability and possible signaling in regulation of osteoclast formation .
In this research, we assessed effect of LY900009 in suppressing RANKL-induced osteoclast formation and elucidated the potential molecular mechanisms associated with this process. We con rmed that LY900009 suppressed osteoclastogenesis and bone resorption by blocking the Notch/Hes1/MAPK(Erk and p38)/Akt pathway. This conclusion was also con rmed by the in vivo results.

BMMs and Culture System
BMMs were obtained and cultured according to former relative research [20][21][22] . Primary BMMs were isolated from mice ( ve-week-old C57/BL6 male) femurs and tiviae. All process were under the supervision of Shanghai General Hospital Animal Center Committee of Animal Care and Use

Osteoclast Formation and TRAP Staining Assay
The function of LY900009 on osteoclast formation was assessed. The BMMs were seeded into 96-well plates (1 × 10 4 cells per well) and stimulated by RANKL (100 ng/mL, R&D Systems, Inc. ), M-CSF (33.3 ng/mL) and different concentrations of LY900009 (0, 100, 200, and 400 nM). The culture medium was replaced every 48 h for 6 days. Afer the osteoclast was observed in control group, 4% paraformaldehyde was used for OC xation. TRAP ( Sigma-aldrich; Merck, Inc.) staining solution was added into each of the wells at 37 C for 1 h. TRAP + cells with at least three nuclei were considered to be osteoclasts. Image J software (National Institutes of Health) was used for cell enumeration.

Immuno uorescence analysis of the podosomal actin belt
OCs derived from BMMs were generated and processed. On days 5 -7, pancake-like osteoclasts were observed in the RANKL-treated control group, xed with 0.1% Triton X-100 (Sigma-Aldrich; Merck, Inc.), and permeabilized for 5 min. After blocking with 1% BSA-PBS for 1 h, the cytoskeletal actin structure was stained with rhodamine-conjugated phalloidin. The immuno uorescence images was obtained by using BioTek Cytation 3 Cell Imaging Reader. Image J software was used to analyze the size (diffusion area) and number of the dental actin bands.

Western Blotting Analysis
The total cellular proteins (TCPs) were extracted at different time points. Following treatment with 400 nM LY900009 for 2 h, RANKL (100 ng/mL) was used to stimulate the cells for different periods of time (short time course). The membrane was blocked in 1% TBST (tri-buffered saline, Tween 20)  Inc.). After incubating the samples overnight at 4°C, the secondary antibody was incubated at room temperature for 1 h using Odyssey V3 .0 image scan (Li-COR. Inc., Lincoln, NE, USA) to observe antibody reactivity.

Mouse Model of LPS-Induced Calvarial Bone resorption
All process were was carried out in compliance with the ARRIVE guidelines and under the supervision of Shanghai General Hospital Animal Center Committee of Animal Care and Use. All procedures were performed under the guidelines of the Ethical Conduct in the Care and Use of Nonhuman Animals in Research. C57BL/6 male mice, aged 7 to 8 weeks, were randomly assigned to four groups: LPS (control), and LPS with low(400nM) or high(800nM) concentrations of LY900009. 1% pentobarbital sodium (10 mg/kg) was intraperitoneally injected to anaesthetize the mice. Mice were injected with 100 µl LPS(10 mg/kg), and with or without LY900009, respectively, for two weeks. During the experiment period, no signi cant adverse effects and toxical effects were observed. The mice showed no signi cant weight loss and lack of spirit, and no mice died during the experiment period. After the operation, euthanasia was performed using pentobarbital sodium(100 mg/kg) 23 . The knee joints were resected and xed in 4% paraformaldehyde for 48h. After that, the specimens were prepared for micro-CT (Skyscan 1072; Skyscan, Inc.) and histological analyses.

Histological Staining and Histomorphometric Analysis
The 10% EDTA was used to decalci ed PFA-xed mice knees for two weeks and embedded in para n.
Next, the samples were sliced into speci c sections(4um thickness) and subjected to H&E and TRAP staining. Digital images were obtained using Axio ScopeA1 light microscope (ZEISS, Inc.). The osteoclasts amount were calculated by Image J software.

Statistical Analysis
All result were shown as the mean ± standard deviation (SD).The Student's t-test was used to assess the differences between control and therapeutic group. The results for the multiple group comparisons were analyzed using a Scheffe's test and one-way analysis of variance (ANOVA) with SPSS 22.0 software (SPSS, Inc.). Values were determined to be signi cant at *P < 0.05; **P < 0.01; and ***P < 0.001.

RANKL-Induced Osteoclast formation in vitro was down-regulated by LY900009
Cytotoxicity of LY900009 was rstly evaluated. The results showed that the IC50 of LY900009 in BMMs at 24 h, 48 h, and 96 h was 4.44 mM, 3.66 mM, and 2.93 mM respectively ( Figure 1A and B). Figure 1C was the chemical structure formula of LY900009 (provided by Selleck). Compared with control group, treatment of LY900009 effectively reduced the number and area of OCs in a dose-dependent manner ( Figure 1D, E and F). Accordingly, LY900009 could inhibit osteoclast formation in a speci c concentration under IC50.

Podosome actin belt formation and OC-mediated bone resorption activity was inhibited by LY900009
In order to assess the effect of LY900009 on OC-mediated bone resorption activity, resorption area on the hydroxyapatite-coated Osteo Assay plates was calculated (Figure 2A). In accrodance with the osteoclast reduction, the bone resorption area also reduced in a dose-dependent manner ( Figure 2B). The immuno uorescence results showed that the podosome actin belt formation was suppressed LY900009 in a dose-dependent manner ( Figure 2C). In summary, these part of data shows that LY900009 could effectivelly induce podosome actin belt formation( Figure 2D).

Osteoclastogenesis Relative Gene Expression was Depressed by LY900009
Quantitative PCR(qPCR) was used to evaluat the relative expression of osteoclastogenesis genes at the mRNA level. It was observed that relative Gene, including NFATc1, C-fos, Cath-K, and TRAP, was markedly downregulated by LY900009 in a dose-dependent and time-dependent manner. (Figure 3 A and B)

LY900009 mediate Osteoclastogenesis by inhibiting Notch, MAPK and Akt signalings in BMMs
To investigate the potential molecular mechanisms of LY900009 in osteoclastogenesis inhibition, the RANKL-induced NF-κB, MAPK, and PI3K/Akt pathways were examined using a Western blot analysis.
BMM treatment with LY900009 during RANKL-induced osteoclast formation induced cleaved -Notch1, Hes1 and NFATc1 protein expression (Figure 4 A and B). In addition, short-term stimulation of LY900009 signi cantly reduced the phosphorylation of p38, ERK1/2, p65, and Akt (Figure 4 C). Accordingly, our experiment suggests that LY900009 blocked osteoclastogenesis inhibited osteoclastogenesis by preventing Notch receptor cleavage and inhibiting Notch signaling, and perturbing the activation of ERK1/2, p38 and AKT signaling cascades.

LY900009 Administration Prevents LPS-Induced Bone Resorption in vivo
The in vitro experiments were used to investigate the effect of LY900009 on RANKL-induced osteoclastogenesis and bone resorption function by studying the phenotype and associated mechanism. We investigated the function of LY900009 in mice with LPS-induced bone resorption. According to the micro-CT three-dimensional recconstruction shown in Figure 5A, compared with control group, LY900009treated groups display fewer and smaller resorption pits. A signi cant reduction in BV/TV ,Tb.n and BMD was also detected in LPS group ( Figure 5B).
The histological analysis also reveal the inhibition of LY900009 on LPS-induced bone resorption in vivo.
In accordance with former result, extensive osteolysis was observed in the HE-stained images of the control group, whereas a lower osteolytic level was observed in the LY900009-treated groups ( Figure 6A). The TRAP staining results con rmed a decreased amount of OCs under LY900009 treatment ( Figure 6B and 6C). Accordingly, the in vivo results con rmed the potential therapeutic application of LY900009 in in ammation-relative osteolytic disease.

Discussion
A pathological bone homeostasis promotes the over-formation and activation of OCs, leading to massive bone disruption, which is a feature of several osteolytic diseases [3][4][5] . Medical therapies that suppress osteolysis are considered to be a potential treatment strategy for these diseases 24 ; 25 . However, a great number of these treatment have various side-effects 26 ; 27 . Accordingly, the development of medications that can inhibit osteoclastogenesis without side-effects are urgently required.
Our research reveals some potential signaling which may regulate RANKL-induced osteoclast formation.
Notch signaling is important in many function, such as morphogenesis and stem cell niche maintenance [28][29][30] . However, the function of Notch signalling on osteoclastogenesis is no clear, with evidence supporting both a stimulatory and inhibitory role. According to former research, Notch has been shown to work in conjunction with MAPK, NF-κB, and PI3K/Akt to possibly regulate NFATc1 expression during osteoclast differentiation 18 ; [31][32][33][34] . In addition, suppression of Notch1 in BMMs has been con rmed to promote osteoclastogenesis in vitro 35 , whereas suppression of Notch2 attenuates osteoclastogenesis 17 ; 36 .
Former research has con rmed the synergistic function between Notch and NF-kB signaling 37 . According to our research, NFATc1 expression was signi cantly decreased under the LY900009 treatment. LY900009, a potent γ-secretase inhibitor, doesn't change expression but blocked cleavage of Notch.
In accordance with previous studies 38 , we also observed a reduction of ERK, p38, Akt, and p65 phosphorylation under the treatment of LY900009. Herein, there is a signi cant crosstalk between Notch, ERK, and Akt signaling. Our data do suggest that the LY900009 can suppress bone loss and osteoclast formation by means of Notch mediated NFATc1 expression.
In this study, a widely-accepted LPS-induced osteolytic experimental model was established to assess the effect of LY900009. During the design of this research, Ovx, RANKL-induced bone resorption and LPSinduced bone resorption model were taken into consideration. Ovx model is inappropriate to be used in this research, because Ovx regulates not only osteoclast but also osteoblast formation. The reason of using RANKL in vitro is that RANKL is mainly produced by osteoblast and necessary for the formation of osteoclast. However, there is already has osteoblast in vivo. LPS-induced bone resorption, as a general approach, has been widely used for osteoclast relative in vivo study. [39][40][41][42][43] Accordingly, LPS-induced bone resorption was chosen as the in vivo model. Not only the micro-CT analysis but also the histological staining of TRAP con rmed a downregulated osteopenia in the LY900009-treated groups. However, there are still a series of limitations in present research. Our research didn't discuss the possibility of LY900009 regulating osteoclast via other signaling pathways. Secondly, the in uence of LY900009 on osteoblast and bone formation should be further assessed.
In conclusion, our data demonstrate that LY900009 can effectively attenuate RANKL-induced osteoclastogenesis in vitro via downregulating Notch/MAPK (ERK and p38)/Akt signaling. In accordance with the in vitro observations, prevention of LPS-induced osteolysis in vivo also con rmed the possibility of using LY900009 as a potential treatment strategy for osteolytic diseases in relation to excessive osteoclast-mediated bone resorption.