Loss of Stat3 in Osteoblasts Impaired the Bone Remodeling in In ammatory Microenvironment

Jingyi Feng Sun Yat-Sen University https://orcid.org/0000-0001-5064-7484 Zijing Huang Sun Yat-Sen University Laiting Chan Sun Yat-Sen University Jiarui Lu Sun Yat-Sen University Xin Feng Sun Yat-Sen University Lizhen Lei Sun Yat-Sen University Zhuwei Huang Sun Yat-Sen University Lichieh Lin Sun Yat-Sen University Yichen Yao Sun Yat-Sen University Xiaolei Zhang (  zhangxl35@mail.sysu.edu.cn ) Sun Yat-sen University https://orcid.org/0000-0003-2410-179X


Introduction
The in ammatory oral bone diseases, including periapical periodontitis, periodontitis and peri-implantitis, are characterized by in ammation, resorption of alveolar bone and connective tissues [1,2]. Accordingly, how to control the in ammatory destruction of bone tissue and to promote bone regeneration is an important clinic challenge and bench topic.
Osteoblasts and osteoclasts are regulated, in part, by many local factors including cytokines in the bone microenvironment. Janus kinase (JAK)-signal transducer and activator of transcription (Stat) pathway have been also found in bone cells [3]. The Stats consisting of Stat1, Stat2, Stat3, Stat4, Stat5α, Stat5β and Stat6, are an important family of intracellular transcription factors that mediate cell signal transduction with dual functions of transmitting cytoplasmic signals and initiating gene transcription in the nucleus [4]. Stat3 mediated by numerous growth factors, cytokines and oncoproteins, plays an important role in cell proliferation, survival, apoptosis, differentiation, transformation, cellular immunity and so on [5].
Among the seven Stats, Stat3 was reported to be the most relevant to bone disorders [6]. In human, the Job syndrome (also named as the high IgE syndrome) is caused by Stat3 mutation, which is an autosomal dominant multisystem disorder characterized by recurrent skin and lung infections, high serum IgE levels, and obvious craniofacial and skeletal features [7,8]. Fractures, joint hyperextension and scoliosis were observed on Job syndrome patients [9]. Similarly, our previous study showed that the selective inactivation of mesenchymal progenitors-derived Stat3 in mice led to shortened limbs, multiple fractures and reduced bone mass due to inhibition of bone formation, suggesting an important role of Stat3 during bone development [10]. What's more, we also found that overexpression of Stat3 rescued the decreased osteoblast differentiation of Ror2-knockdown mBMSCs, indicating Stat3 could be the downstream molecule of Ror2 [11]. Stat3 can be found in osteoblasts, osteocytes, chondrocytes and osteoclasts [6]. Studies have shown that the depletion of Stat3 from osteoclasts caused a marked increase of trabecular bone mass in adult mice, in which the number of osteoclast was decreased without in uencing the activity of osteoblasts [12]. Meanwhile, Stat3 was reported to play a vital role in the innate immune response, anti-in ammatory and repair of damaged tissues [13][14][15]. Although the bone formation was known to be affected by Stat3, the effect of osteoblast-derived Stat3 on bone remodeling in in ammatory lesions is yet to study.
Hence this study aimed to investigate the effect of Stat3 on the osteoblast-mediated bone remodeling in the in ammatory lesion. The conditional knockout of Stat3 in osteoblasts (Stat3 CKO) was generated via the Cre-loxP recombination system using Osterix-Cre (Osx-Cre) transgenic mice [16][17][18]. Furthermore, the lipopolysaccharide (LPS) -induced calvarial bone lesions were established on both Stat3 CKO and wildtype mice. With the depletion of Stat3 from osteoblasts, the bone remodeling in in ammatory microenvironments was assessed in vivo and in vitro.

Mouse lines
The mice were purchased from the Jackson Laboratory (Osx-GFP-Cre, stock no. 006361 and Stat3 f/f , stock no. 016923) [19,20]. To generate the mice of Stat3 conditional knockout in osteoblasts (Stat3 CKO), the Stat3 f/f mice were rst crossed with the Osx-GFP-Cre transgenic mice to obtain the Osx-GFP-Cre; Stat3 f/+ mice. Then, the Osx-GFP-Cre; Stat3 f/+ mice were crossed with the Stat3 f/f mice to get the Osx-GFP-Cre; Stat3 f/f (i.e. Stat3 CKO) mice. Genotyping was conducted by polymerase chain reaction (PCR) of tail lysate. Primer sequences were provided in Table 1.
For temporal deletion of Stat3 in osteoblasts, the doxycycline administration in drinking water (0.2 g doxycycline per liter water; Sigma Aldrich, St Louis, MO) was applied on the Stat3 CKO mice (i.e. to tet-off the function of conditional knockout by Osx-GFP-Cre) until the age of 3-weeks (2 weeks before the establishment of in ammatory calvarial lesions). With the least in uence of Stat3-de ciency on the development of animal skeleton, the in vivo knockout of Stat3 in osteoblasts was implemented during the period of in ammatory lesion establishment and sample harvest. For in vitro culture of osteoblasts, the Stat3 CKO mice were raised without doxycycline administration. The animals were used according to guidelines with approval by the Animal Ethical and Welfare Committee of Sun Yat-sen University (approval number SYSU-IACUC-2020-000069).

Establishment of in ammatory calvarial osteolytic lesions
To study the effects of osteoblastic Stat3 on bone remodeling in in ammatory microenvironments, the LPS-induced calvaria osteolytic model was established on both Stat3 CKO and the littermate (the control) mice as described in literature [21,22]. In brief, the LPS (Sigma-Aldrich, St. Louis, MO, USA, L2880, 25 mg/kg) or PBS (sham group) were injected subcutaneously into the tissue pocket surrounding the calvaria and near the midline of the skull between the ears and eyes in 5-week mice. After the injection for 3, 7 and 14 days, the animals (n=3 per group) were sacri ced respectively to collect the calvaria. The timeline of doxycycline administration, in ammatory calvaria osteolysis establishment and sample harvest was displayed in Figure 1.

Micro-CT scanning and analysis
The calvaria were xed in 4 % para-formaldehyde at 4℃ overnight, stored in 70 % ethanol for 24 h and analyzed using a high-resolution micro-CT (Scanco, Brüttisellen, Switzerland) instrument. The scanning protocol was set at an isometric resolution of 15mm, with X-ray energy settings of 70kV and 114µA. After reconstruction, a square region of interest (ROI) of the calvaria was selected for further quantitative analysis, including bone volume fraction (BV/TV), trabecular thickness (Tb.Th.), trabecular number (Tb.N.) and trabecular separation (Tb.Sp.). The calvaria were reconstructed into 3-dimensional images using Mimics software (Medical Version 17.0; Materialise, Leuven, Belgium).

Histological staining
After xed in 4 % para-formaldehyde at 4℃ overnight, the calvaria were decalci ed in 10% ethylenediaminetetraacetic acid for 2 weeks, and dehydrated through a series of alcohols and were embedded in para n. Para n-embedded specimens were sectioned through the coronal plane at a thickness of 5 µm. Histological analysis was performed as reported previously [23,24]. The sections were prepared for hematoxylin and eosin (HE) staining and Masson staining, then analyzed under a microscope. To observe the osteoclast activity, the tartrate-resistant acid phosphatase (TRAP) staining was performed using a TRAP kit (kit no. 387A; Sigma-Aldrich, St Louis, MO) according to the manufacturer's instructions. Osteoclasts were identi ed morphologically as the large multinucleated cells within the resorption lacunae. For quantitative analysis, the number of TRAP-positive multinucleated osteoclasts of bone surface area were counted. Immunohistochemistry staining Immunohistochemistry (IHC) of Stat3, runt related transcription factor 2 (Runx2), osteopontin (OPN), alpha-1 type I collagen (COL1A1) were performed to visualize the spatial expression of these proteins in para n-embedded sections from calvaria samples. All sections were stained using the HRP-DAB staining kit (CTS017; R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. The primary antibodies were listed in Table 2. The counterstaining was performed with hematoxylin. The negative controls were treated with phosphate-buffered saline rather than the primary antibodies.
Alizarin red staining and calcium deposition determination After mineralization induction for 21 days, the cells were gently washed three times with PBS, xed in 4% paraformaldehyde for 20 min, and stained with alizarin red S solution for 15 min. Images of extracellular matrix mineralization nodules were obtained using an inverted microscope with a digital camera. For quantitative analysis, the mineralization nodules were extracted using 10% cetylpyridinium chloride. The absorbance was measured at a wavelength of 540 nm on a spectrophotometer (Thermo Fisher Scienti c).

Cell proliferation assay
A 5-ethynyl-20-deoxyuridine (EdU) assay kit (Beyotime Biotechnology, Shanghai, China) was applied to assess the proliferation of cells. Cells were seeded into a 6-well plate with density of 2 × 10 5 cells per well. When the cells reached 60-70% con uence, they were incubated with EdU buffer at 37℃ for 2 h, xed with 4% formaldehyde for 0.5 h and permeabilized with 0.1% Triton X-100 for 20 min. EdU staining solution was added into the plate followed by the staining of nuclei with Hoechst. Fluorescence microscopy was performed in ve randomly selected elds with an inverted microscope (Axio Observer Z1, Zeiss, Oberkochen, Germany) to analyze the proliferation rates.

Transwell migration assay
Transwell plates (8µm pore size; Costar, Corning, NY, USA) were used to perform the cell migration assay according to the manufacture's protocol. Brie y, a total of 100µl of a serum-free suspension 2.5×10 5 cells/ml was seeded in the upper membrane of inserts in the presence of LPS (0 and 1 µg/ml) while 600µl of medium with 10% serum was added in the lower chamber. After incubation for 12h at 37℃, the cells remaining at the upper surface of the membrane were removed with cotton swabs, and the cells on the lower surface of the membrane were considered as the migrated cells. Migrated cells were xed in 4% paraformaldehyde for 20 min and stained with crystal violet for 30 min at room temperature. The cells in ve random microscopy elds were counted using an inverted microscope (Axio Observer Z1, Zeiss, Oberkochen, Germany).
RNA isolation, reverse transcription, and qRT-PCR analysis

Statistical analysis
All animal studies were performed using gender-and age-matched animals, with the wild-type littermates as controls (CON). Data were derived from multiple independent experiments from distinct mice or cell culture plates. The data were presented as means ± standard deviation (SD). Student's t-tests were used to compare the difference between Stat3 CKO and CON. The GraphPad Prism 7.0 software was applied for statistical analysis. The signi cance level was set at P < 0.05.

Characterization of Stat3 CKO mice
To achieve the conditional knockout of Stat3 in osteoblast, the transgenic mice of Osx-Cre was used to cross with Stat3 f/f mice. In Fig. 2A, the mouse genotypes, including Osx-Cre; Stat3 f/f (Stat3 CKO) and Cre negative littermate control, were displayed. The calvaria isolated from 5-week-old Stat3 CKO mice showed positive GFP signal due to Osx-GFP-Cre, whereas it was absent in CON mice (Fig. 2B). To verify the elimination of Stat3 in osteoblasts, the quantitative real-time polymerase chain reaction (qRT-PCR) and western blot were performed to assess the expression of Stat3. Stat3 were found in the osteoblasts in the calvaria of the 5-week-old Con mice, but were hardly expressed in the calvaria of the Stat3 CKO mice (Fig. 2C, D) (**P<0.01).

Depletion of Stat3 in osteoblasts led to a decrease of bone mass in calvarial osteolytic lesions
To explore the impact of Stat3-de ciency on bone regeneration in in ammatory microenvironment, the LPS-stimulated calvaria lesion was induced on both Stat3 CKO and control mice. According to the morphological observation of calvaria from day 3 to day 14 (Fig. 3A), the knockout of Stat3 in osteoblasts led to an increase of LPS-induced osteolysis, which was characterized by visible osteolysis damage and bone surface erosion, especially at day 3 and day 7. Micro-CT 3D reconstruction observations indicated an obvious bone loss in the Stat3 CKO mice on calvaria at day 7 (Fig. 3B). Meanwhile, no signi cant difference of microstructural indices between CON and Stat3 CKO mice was observed at day 3 and 14. On the other hand, a marked reduction of trabecular bone mass in Stat3 CKO mice was found at day 7, assessed by the CKO vs. CON groups of BV/TV, Tb.Th., Tb.N., and Tb.Sp. (Fig. 3C) (*P<0.05, **P<0.01).

Stat3 de ciency in osteoblasts caused an increased in ammatory osteolysis
Considering that the obvious lesion difference between CON and Stat3 CKO mice was found at day 7 after LPS-stimulation, these samples were collected for para n section preparation. According to HE staining, the in ammatory cell in ltration along with signi cant bone erosion was observed in the calvaria of Stat3 CKO mice (Fig. 4A). The decrease of mineralized bone matrix in Stat3 CKO mice was con rmed by Masson staining (Fig. 4B) (**P<0.01). Consistent with the increased bone erosion, the osteolytic lesion with abundant TRAP-positive osteoclasts was found on Stat3 CKO mice, whereas the CON group was presented with fewer TRAP-positive osteoclasts (Fig. 4C) (**P<0.01). Moreover, the increased mRNA expression of RANKL in osteolytic lesion was found on Stat3 CKO group (Fig. 4D) (vs. CON, *P<0.05).
4. In the bone destructive lesion, the osteogenic markers were decreased in Stat3 CKO mice The IHC staining observations revealed that the number of Stat3 positive osteoblasts in Stat3 CKO group was dramatically dropped compared with CON group (***P<0.001) (Fig. 5A). Meanwhile, the numbers of Runx2, OPN and COL1A1 positive osteoblasts per hpf in osteolytic lesion were signi cantly decreased in Stats3 CKO mice (**P<0.01) (Fig. 5A). A considerably reduction of the mRNA and protein expressions of osteogenic markers Runx2, OPN, and COL1A1 were detected in the calvaria of Stat3 CKO mice, compared with CON group (*P<0.05, **P<0.01) (Fig. 5B-D).

Knockdown of Stat3 in osteoblasts led to a reduced mineralization in the in ammatory microenvironment in vitro
To gain insight into the Stat3-mediated bone matrix mineralization in osteoblasts, osteoblasts from primary cell culture of passage 2-3 were seeded into plates followed by mineralization induction for 21 days. Less formation of calcium deposit was observed in the Stat3-di cient cells when compared with the Con, which was visualized by the Alizarin Red staining (***P<0.001) (Fig. 6A).
The expression of total Stat3 and nuclear phosphorylated Stat3 (p-Stat3) was signi cantly decreased in Stat3 CKO cells in both normal and in ammatory microenvironment, con rming that Stat3 gene was effectively knocked out in osteoblasts of Stat3 CKO mice (*P<0.05, ***P<0.001) (Fig. 6B-C). The lack of Stat3 signi cantly suppressed the mRNA expression of Runx2, COL1A1, and OPN in osteoblasts (*P<0.05, **P<0.01, ***P<0.001) (Fig. 6 B-C)). The mRNA expression of RANKL in osteoblasts was increased while difference was not found in the expression of OPG, which was consistent with the in vivo data (*P<0.05) ( Fig. 6D-E). The protein levels of osteogenic markers were also down-regulated in both normal and in ammatory microenvironment, compared with the control (*P<0.05, **P<0.01) (Fig. 5F-G).

The proliferation and migration of Stat3-di cient osteoblasts were inhibited in the in ammatory microenvironment
According to the EdU assay, fewer EdU positive cells were observed in Stat3 CKO group in both normal and in ammatory microenvironment (**P<0.01) (Fig. 7A, B). The number of migrated osteoblasts in the Stat3 CKO group was signi cantly decreased in normal microenvironment and further less in in ammatory microenvironment, compared with the control (***P<0.001) (Fig. 7C, D).

Discussion
Bone remodeling in in ammatory bone destructive lesion has always been an important clinical research topic in stomatology. To explore how Stat3 regulates osteoblast-mediated bone remodeling in the in ammatory lesion, we cultivated osteoblast-speci c Stat3 knockout transgenic mice (Stat3 CKO) and established LPS-induced in ammatory bone destructive lesion on both Stat3 CKO and wildtype mice.
Osteoblasts from the calvaria of neonatal mice were isolated and cultured in vitro to investigate the effect of Stat3 de ciency on the ossi cation, signals transduction in osteoclast activation, proliferation and migration of osteoblasts in in ammatory microenvironments. Our ndings indicated that Stat3 plays a promoting role in bone regeneration and remodeling mediated by osteoblasts in in ammatory lesion.
Germline Stat3 depletion-induced early embryonic lethality was observed in 1997 [28]. The impaired contractility and metabolic homeostasis by Stat3 de ciency was found in cardiomyocyte-targeted Stat3 knockout (KO) mice [29]. During the forced swim, tail suspension and open-eld tests, antidepressive-like behavior in microglia-speci c Stat3 KO mice proved that microglial Stat3 regulates depression-related behaviors via neuronal M-CSF-mediated synaptic activity [30]. Collectively, Stat3 is widely expressed in different tissues and cells, participates in the regulation of many physiological functions including proliferation, differentiation, growth, apoptosis and in ammation [31]. In this study, we used Tet-Off system in Osx-Cre mice to control the expression of Stat3 in osteoblasts in time for doxycycline can reversibly inhibit the expression of Cre-GFP fusion protein and prevent the activation of CKO. The purpose of withdrawing doxycycline before 3 weeks of age is to ensure the bone development was not affected and bone mass between two groups was similar. So it could avoid errors caused by their own development defects and body size in further research. After next two weeks of metabolism, the original osteoblast-derived Stat3 would be consumed by apoptosis, then Stat3 would be conditionally knocked out in the newborn osteoblasts.
LPS from the cell membrane of gram-negative bacteria is recognized as the main pathogenic factor positively related to infectious bone destruction. A recent literature has shown that LPS inhibited bone remodeling and promoted oxidative stress accumulation in alveolar bone lesions [32]. The hypothesis that factors within the in amed joint impede bone formation at arthritic sites by inhibiting osteoblast maturation was supported by less mineralized bone formed at bone surfaces adjacent to in ammation, compared to surfaces adjacent to normal bone marrow [33]. These studies raise the interesting question as to how in ammation impairs osteoblast maturation and function. Bandow K et al. subsequently demonstrated that the inhibitory effect of LPS on osteoblast differentiation is Myd88-dependent: LPS suppresses the mRNA expression of Runx2, osterix (Osx), and activating transcription factor 4 (ATF4) during osteoblast differentiation in the wild-type, but not in the myd88 ox/ ox osteoblasts [34]. Pretreatment with JNK inhibitor not only attenuated both of phosphorylation-JNK and ERK1/2 enhanced by LPS in MC3T3-E1 cells, but also reversed the down-regulated expressions of osteoblast-speci c genes including ALP and BSP induced by LPS, reaching to the conclusion that LPS could inhibit osteoblast differentiation via activation of JNK pathway [35]. It has also been reported that LPS can suppress osteoblastic differentiation by affecting multiple pathways, such as BMP/Smad signaling, Wnt/β-catenin signaling, and Notch signaling [36-38]. As was shown in this study, the depletion of Stat3 in osteoblasts resulted in a decline of regenerated bone mass at the bone-repair site in accordance with micro-CT analysis. Our data from Masson staining also indicated that bone fraction was reduced in the lesion of Stat3 CKO mice. Given the ndings that the lack of Stat3 led to less regenerated bone, we also found that Stat3 could affect bone matrix mineralization function of osteoblasts in vitro. Moreover, the protein and mRNA expressions of osteogenic markers Runx2, OPN and COL1A1 were diminished in Stat3 CKO group, further con rming the role of Stat3 as a positive regulator of bone regeneration in osteoblasts. Taken together, our ndings revealed that loss of Stat3 delays bone remodeling in mice with impaired osteoblast bone formation activity.
LPS-induced overproduction of in ammatory such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β) and RANKL [39][40][41][42] contributes to most in ammatory response that are involved in the pathogenesis of in ammation-stimulated oral bone diseases [43]. The role of in ammation has been closely associated with bone loss and osteoclastogenesis [44]. Among infection-induced in ammation, the components of invading bacteria (e.g. LPS) will activate macrophages. Factors secreted from activated macrophages increase sphingosine kinase 1 (SPHK1) activity in bone marrow stromal cells (BMSCs), resulting in the overproduction of sphingosine-1-phosphate (S1P). Then S1P acts on its receptor sphingosine-1-phosphate receptor 1 (S1PR1) in BMSCs, which results in the overproduction of RANKL and eventually induces bone resorption via the activation of RANK in osteoclast precursors [45]. During the active phase of rheumatoid arthritis, the in amed synovium comprising T-cells, synovial broblasts, and macrophages, produces in ammatory cytokines and RANKL which can activate osteoclasts and lead to bone destruction [46]. RANKL/RANK/OPG is an important pathway modulated by osteoblasts which affect the bone mass density via the regulation of osteoblasts and osteoclasts functions [47]. A closer look at the osteoclast activity via TRAP staining showed that osteoclast number of Stat3 CKO mice was signi cantly greater in our study. Besides, Stat3 de ciency in osteoblasts promoted LPS-induced osteoclastogenesis mediated by RANKL expression was found both in vivo and in vitro. A possible explanation might be that, in the in ammatory environment, Stat3 not only plays a role in ossi cation of osteoblasts, but also enhances osteoclast activity indirectly by regulating osteoclast-related signal molecules secreted by osteoblasts. But its underlying mechanism still needs further investigation.
Existing studies have suggested Stat3 can induce the expression of proliferative, anti-apoptotic, angiogenic and metastatic genes that promote tumor development and growth, such as B-cell lymphomaextra large (Bcl-xL), B-cell lymphoma-2,Bcl-2 (Bcl-2), Cyclin D1, c-Myc, vascular endothelial growth factor (VEGF) and HIF1α [48][49][50][51]. Sim and other scholars have shown that gp130-Stat1/3 signaling pathway plays a key role in stimulating chondrocyte proliferation [52]; under hypoxia, interleukin-8 (IL-8) enhances the proliferation of bone marrow mesenchymal stem cells (BMSCs) through Akt-Stat3 pathway and reduces the apoptosis rate of cells [53]. Recent work from Yu X et al. reported that hypoxia can enhance BMSCs migration through Stat3 signal transduction, and up-regulated gene expression related to cell migration, such as matrix metalloproteinase 7 (MMP7), matrix metalloproteinase 9 (MMP9) and C-X-C chemokine receptor type 4 (CXCR4) [54]. In the present study, the notion that the proliferation and migration ability of osteoblasts are inhibited owing to Stat3 de ciency and in ammation is supported by decreased migration cells number and EdU positive cells number of Stat3 CKO group in vitro. This may explain the underlying mechanism of in ammation and Stat3 de ciency leading to changes in osteoblast mineralization. In brief, osteoblast-derived Stat3 is associated with pro-proliferative, anti-in ammatory and osteogenic growth factor activities during bone remodeling.
Taken together, our study indicated Stat3 plays an important role in osteoblast-mediated bone remodeling in in ammatory microenvironment. Loss of Stat3 in osteoblasts caused decreased bone regeneration in in ammatory microenvironment, which might be resulted from the effect of Stat3 on the ossi cation, signals transduction in osteoclast activation, proliferation and migration of osteoblasts. Based on these results, Stat3 could serve as a promising treatment target for in ammatory bone diseases.

Availability of data and materials
All datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
Mice were used according to federal guidelines and as approved by the Animal Ethical and Welfare Committee of Sun Yat-sen University (approval number SYSU-IACUC-2020-000069).

Consent for publication
Not applicable. Figure 1 The process of establishing in ammatory bone destruction model in murine calvaria. For temporal deletion of Stat3 in osteoblasts, the doxycycline administration in drinking water was applied on the Stat3 CKO mice until the age of 3-weeks (2 weeks before the establishment of in ammatory calvarial lesions). The LPS or PBS (sham group) were injected subcutaneously into the tissue pocket surrounding the calvaria and near the midline of the skull between the ears and eyes in 5-week-old Stat3 CKO and the littermate (the control) mice. After the injection for 3, 7 and 14 days, the animals (n=3 per group) were sacri ced respectively to collect the calvaria.    Stat3 CKO mice were signi cantly decreased compared with CON mice. n=3, **P<0.01 vs. CON. C. Western blot results of the osteogenic proteins in calvaria from CON mice and Stat3 CKO mice were shown. n=4, **P<0.01 v.s. CON.

Figure 6
De ciency of Stat3 in osteoblasts led to reduced bone matrix mineralization in in ammatory microenvironment. A. Representative alizarin red staining images were shown after mineralization