Roxadustat

HIF/Ca2+/NO/ROS is critical in roxadustat treating bone fracture by stimulating the proliferation and migration of BMSCs

Abstract

Aims: Fracture site is regionally hypoxic resulting from vasculature disruption. HIF-1αplays an essential role in fracture repair. This study aims to investigate the influence of FG4592 on the femur fracture of SD rats and the proliferation, migration of BMSCs.

Materials and methods: After the femoral fracture model was established, computed tomography imaging and histological analyses were used to quantify bone healing and the expression of CD90, HIF-1α, VEGF were observed by means of immunohistochemistry method on Day 10 and Day 20. In addition, CCK-8 assay, transwell, flow cytometric analysis, laser confocal microscopy assay, western blot and rT-PCR were performed to text the proliferation and migration of BMSCs using FG4592.

Key findings: In vivo, FG4592 facilitated the repair of bone fracture by increasing the number of BMSCs and cartilage formation. In vitro, FG4592 markedly improved the proliferation, migration of BMSCs via upregulation of intracellular Ca2+, NO and concomitant decrease of ROS. Gene silencing of HIF-1α resulted in the opposite phenomenon in BMSCs with the treatment of FG4592.

Significance: The transplantation of BMSCs is the most promising candidate for the treatment of fracture non- union. We illustrated that FG4592 promoted the proliferation, migration of BMSCs via the HIF/Ca2+/NO/ROS pathway and further accelerated fracture healing. These results provide a deeper understanding for the mechanism of HIF in promoting fracture healing.

1. Introduction

Delayed union or nonunion following fracture is a common problem in clinical management of bone fracture [1]. Cell transplantation as an advanced therapy deserves in-depth study [2–4]. BMSCs are the most appropriate cells for inducing bone repair. However, high mortality of BMSCs limits its application [2]. Great efforts have been made to induce MSC surviving using agents and the underlying mechanism promoting MSC proliferation are not noticeably clear.

Hypoxic signaling pathway plays a crucial role in fracture repair. HIF-1α is the central mediator of cellular response to hypoxia [5,6]. The activity of HIF-1α could be tuned by the hydrolysis of HIF-prolyl hy- droxylase [7]. In normal oxygen conditions, HIF-1α displays constitutive expression and quick degradation. Small molecule inhibitors of HIF- prolyl hydroxylase can block HIF-1α degradation and thus activate a broad array of genes [8]. HIF-1α can promote the secretion of VEGF in MSCs and play an important role in neovascularization and angiogenesis [9]. Angiogenesis helps support MSCs and osteoblasts, which are necessary for bone repair [10–12]. Increasing evidence suggests that HIF-1α is essential for progenitor cell migration, implying a role for HIF- 1α in hypoxia-induced MSC migration [13,14].

The increase of intracellular Ca2+ concentration augments the sta- bility of HIF-1α. CaMKII is of importance for Ca2+ signal transduction [15,16] and CaMKII inhibitors stunt the protein expression of HIF-1α [17]. However, the function of CaMKII in HIF-1α regulation remains controversial because a specific CaMKII inhibitor SMP-114 has no effect on HIF-1αexpression in rheumatoid synovial fibroblasts [18]. In addition to CaMKII, IP3R1 and RCN1 also affect Ca2+ level. As a huge Ca2+ channel, IP3R1 localized at the ER could control ER Ca2+ release to regulate physiological cellular processes [19,20]. RCN1 is a calcium- binding protein located in the lumen of the ER [21]. Therefore, we utilize prolyl hydroxylase inhibitors and investigate its possible effects on Ca2+ in BMSCs, as well as the underlying molecular mechanism.
NO also has emerged as a potent regulator for fracture healing, mechanically stimulated bone formation [22,23], and modulation of VEGF gene transcription in endothelial cells [24,25]. A growing body of evidence has suggested that NO is involved in regulating VEGF expres- sion via the activation of HIF [26]. NO has been found to mediate HIF-1α stabilization, nuclear translocation, and activation in mechanically stressed myocardiac cells [27]. So, we hypothesize that prolyl hydrox- ylase inhibitor treatment provokes an increase in NO concentration, which in turn boosts fracture healing.

FG4592, a potent oral inhibitor of HIF-1α prolyl hydroxylase, stabilized HIF-1α against degradation [28,29]. Transient inhibition of prolyl hydroxylase by FG4592 results in iron absorption, iron transport (transferrin), and heme synthesis [30]. Our previous investigation showed that activating HIF using prolyl hydroxylase inhibitors regulated MG63 cell activity and protected cells from apoptosis in low serum culture [31]. Herein, we put the focus on the role of FG4592 in fracture healing and exploit the strategy to promote BMSCs proliferation, migration. The fracture healing process was significantly accelerated. The proliferation and migration of BMSCs were also promoted by
FG4592 with the variation in Ca2+, NO and ROS.

2. Materials and methods

2.1. Rat femur fracture model

An open osteotomy model was conducted as described from litera- ture [32]. Eight-week-old male SD rats (200 ± 10 g) were anesthetized by chloral hydrate (10%, 0.3 mL/100 g). A kirschner wire (1.2 mm) was inserted into the femoral shaft retrograde and a midshaft femur fracture was created by manual three-point bending. In order to ensure the validity and repeatability of the experimental data, as well as consid- ering the possible death caused by the subsequent culture of animal models, we selected 6 rats in each group for follow-up experiments. The rats were randomly divided into model and FG4592 treatment groups. In the treatment group, the rats were given by gavage administration with 1 mL/100 g of FG4592 (1 mg/mL) once a day for 10 days and 20 days. Meanwhile, the model groups were treated with the same volume of corresponding solvent. Finally, the rats were sacrificed by cervical dislocation and femur samples were harvested for subsequent testing. Animal use was compliant with the Guide for the Care and Use of Lab- oratory Animals issued by the US National Institute of Health (NIH Publication NO. 85–23) and the Guide for the Care and Use of Labora- tory Animals issued by the Local Internal Evaluation Committee for Animal Welfare and Rights (Nanjing, China, 2018 DW-09-20). All pro- cedures were performed in compliance with relevant laws and institu- tional guidelines and that the appropriate institutional committee(s) had approved them.

2.2. Micro-CT analysis

Computed tomography was used to quantify bone healing using the mCT-80 system (energy/intensity: 70 kVp, 114 μA, 8 W) (Scanco Med- ical, Bassersdorf, Switzerland) and μCT 80 Micro-CT software. The scan range was set to be the full bone of the fracture sample. The resolution was set to 17.5 μm per voxel and 2048 × 2048 pixels. For fracture samples, the Kirschner wire was carefully removed, and femoral samples
were wrapped in wet gauze and fitted in the sample tube for micro-CT scanning. ROI was selected from 2D images with a standardized threshold (>165) as the mineralized tissue. 3D reconstruction of the mineralized tissue was performed with a low-pass Gaussian filter. Serial cross-sectional images of the osteotylus were collected to perform three- dimensional histomorphometric analysis, which included various bone parameters, such as bone volume (BV), total volume (TV), and bone mineral density (BMD) of TV, BMD of BV, and BV/TV.

2.3. Histological analyses

Isolated femurs were fixed, decalcified and then the tissues were embedded in paraffin and serial longitudinal sections (5 μm thickness) were cut for each sample. At least three or four consecutive sections obtained from the sagittal callus were stained with hematoxylin/eosin (HE, Beyotime) or safranine-fast green double dyeing (Servicebio, Wuhan, China) to assess femurs histology and morphology. The imag- ines were scanned using a digital slice canning system (Pannoramic 250, 3DHistech, Budapest, Hungary) with software CaseViewer 2.0 (3DHistech).

2.4. Immunohistochemistry

Paraffin sections (5 μm thickness) were routinely dewaxed to water and citrate buffer (pH = 6.0), which were used for antigen retrieval.
Then, the slides were immersed in 3% hydrogen peroxide for 25 min and blocked with 3% BSA at RT for 30 min. Sections were incubated with primary CD90 rabbit polyclonal antibodies (1:200, Bioss, Massachusetts, USA), VEGF and HIF-1α rabbit polyclonal antibodies (1:100, Santa Cruz, Texas, USA) at 4 ◦C overnight. The slides were washed with PBS and then incubated with HRP-labelled anti-rabbit secondary antibody
(1:500, Vector Lab. Inc., California, USA) at RT for 45 min. Then, the slides were washed with PBS and visualized with 3, 3′-diaminobenzidine (DAB, Sigma) for 5 min. The slides were counterstained with hematoxylin, dehydrated, sealed with neutral gum. The imagines were scanned using a digital slice canning system (Pannoramic 250, 3DHistech) with software CaseViewer 2.0 (3DHistech).

2.5. Determination of Ca2+, NO, H2O2 and Total Antioxidant Capacity in vivo

Blood was collected from SD rats on Day 0, Day 10, Day 20 by clipping their tails after administration of FG4592. Samples (n = 5) were centrifuged at 12,000 ×g for 10 min at 4 ◦C. Following centrifugation,the corresponding volume was added for subsequent measurement. The concentrations of Ca2+, NO, H2O2 and Total Antioxidant Capacity were determined by using the corresponding assay kit (Calcium Colorimetric Assay Kit, NO assay kit, Hydrogen Peroxide Assay Kit, Total Antioxidant Capacity Assay Kit with a Rapid ABTS method, Beyotime Institute of Biotechnology, Haimen, China) according to the manufacturer’s protocol. The absorbance at 575 nm (for Ca2+), 540 nm (for NO), 560 nm (for
H2O2) and 414 nm (for Total Antioxidant Capacity) were recorded by an automatic enzyme-linked immunosorbent assay plate reader (Thermo Scientific Varioskan Flash, Massachusetts, USA). The OD values of each well were measured to represent the concentrations.

2.6. Isolation and culture of BMSCs

SD rats were purchased from the Nantong University, which were kept at the SPF facility of Hospital in Jiangsu Province. BMSCs were harvested from the tibias/femurs of 4-week-old male SPF Spra- gue–Dawley (SD) rats. Primary cells were cultured in low-glucose Dul- becco’s modified Eagle’s medium (LG-DMEM) (Gibco, Invitrogen, Calif, USA) containing 10% fetal bovine serum (FBS, Gibco), 100 units/mL of penicillin (Sigma, St Louis, USA) and 100 mg/mL of streptomycin (Sigma) for 4–6 days by direct adherent culture methods. The medium was changed every 3 days. When cell confluency reached 90%, the attached BMSCs were released from the culture substratum by using 0.25% trypsin/0.02% EDTA (w/v, Gibco). Cells were then subcultured in new 75-cm2 flasks at 2 × 105 in 10 mL of medium at 37 ◦C in a hu-
midified atmosphere containing 5% CO2 with a change of culture me- dium every 2 days.

2.7. FACS analysis of BMSCs

The third generation of cells (1 × 106) were trypsinized and stained with anti-CD45-FITC [33], CD90-FITC [34], and CD11b-FITC [35]
(1:1000, Biolegend, California, USA) for 30 min, and then analyzed by BD LSRL Fortessa flow cytometer (Franklin Lakes, NJ, USA) with FlowJo
7.6 (Franklin Lakes, NJ, USA). IgG1-FITC and IgG2-FITC (1:1000, Bio- legend) were used as an isotype control.

2.8. Adipogenic and osteogenic differentiation

The third generation of cells was seeded into 6-well plates (4 × 105/ well). Adipogenic differentiation was assessed by incubation with an
adipogenic differentiation medium of BMSCs from SD rats (Cyagen, Guangzhou, China) for 3 weeks. Oil Red O staining was used to observe the fat droplets within the differentiated adipocytes from the BMSCs (Cyagen). Cell monolayers were fixed in 4% formaldehyde, washed with water, and stained with a 0.6% (w/v) Oil Red O solution (60% iso- propanol and 40% water) for 15 min at room temperature. After two intensive rinses with deionized water, the fat droplet staining was photographed under fluorescence microscope (X-Cite® 120PC Q, Lumen Dynamics, Ontario, Canada). Osteogenic differentiation was assessed by incubating the cells with the osteogenic differentiation medium of BMSCs from SD rats (Cyagen) for 3 weeks. To assess the mineralization, cultures were stained with an alizarin red dye solution (Cyagen), and the results were also analyzed by fluorescence microscopy (Lumen Dynamics).

2.9. Laser confocal microscopy assay

BMSCs were incubated with corresponding probes (Fluo-3 AM, DAF- AM and DCFH-DA probes of Ca2+, NO and ROS, respectively) for 35 min at 37 ◦C. The cells were then washed and visualized via confocal mi- croscopy (LSM 710, Carl Zeiss, Jena, Germany). The images were ac- quired using the green channel (λex = 488 nm, λem = 520 nm) to monitor the cytosolic fluorescence of Fluo-3 AM, DAF-AM and DCFH-DA probes, respectively. Fluorescence intensity measurements were per- formed using ZEN 2008 software. For each experiment, cells were monitored at 10-s intervals for a total of 300 s.

2.10. Determination of intracellular Ca2+, NO, ROS

The concentrations of Ca2+, NO and ROS were determined by measuring the conversion of cell permeable Fluo-3 AM to fluorescent Fluo-3 (λex = 488 nm, λem = 526 nm), DAF-AM diacetate to fluorescent DAF-AM (λex = 488 nm, λem = 515 nm) and 2′, 7′-DCF diacetate to fluorescent DCF (λex = 488 nm, λem = 525 nm), respectively (Beyo- time). BMSCs were stimulated with varying concentrations of FG4592 (1 μM, 10 μM) for 24 h and then incubated with PBS containing 10 μM of probes for 30 min at 37 ◦C and measured by BD LSRL Fortessa flow cytometer (Franklin Lakes) with FlowJo 7.6 (Franklin Lakes). The con- centrations of Ca2+, NO and ROS were expressed as a histogram of the fluorescence generated by 10,000 cells.

2.11. Plasmid DNA transfection

BMSCs were seeded into 6-well plates (4 × 105/well) and incubated with the plasmid DNA (pSIREN-RetroQ-ZsGreen vector) with Lipofect-
amine™ Stem Reagent (Invitrogen) for 12 h, and then used for subse- quent experiments. The HIF-1α siRNA duplex targeted nucleotides of the
HIF-1α mRNA sequence (NM001530) and were comprised of sense 5′- CUGAUGACCAGCAACUUGAdTdT-3′ and antisense 5′-UCAA- GUUGCUGGUCAUCAGdTdT-3′ [36]. The inverted HIF-1α control duplex did not target any gene and was comprised of sense 5′-AGUU- CAACGACCAGUAGUCdTdT-3′ and antisense 5′-GACUACUGGUCGUU- GAdTdT-3′. The expression of HIF-1α protein after transfection was evaluated by western blot.

2.12. Statistical analysis

Data were presented as mean ± standard deviation (SD) of three independent experiments. Student’s T. Test was used to compare two independent groups. For comparison between multiple groups were performed for ANOVA. Values of *p < 0.05, **p < 0.01 and ***p < 0.001 were considered to be statistically significant. Statistical analysis was performed with IBM SPSS Statistics 20. 3. Results 3.1. FG4592 enhanced fracture healing in SD rats To evaluate the effect of FG4592 on fracture healing, we established a rat bone fracture model. X-ray images revealed that FG4592 improved early callus formation in the fracture, and the osteotylus of the FG4592- treated group was larger than that of the model group (Supporting In- formation Fig. S1). μCT analyses showed the recognizable structure of the callus at Day 10, and the callus of the FG4592-treated group was significantly larger than that of the model (Fig. 1A, B, C). Additionally, fracture sites of the FG4592-treated group exhibited callus formation and nearly complete formation of external bridging bone collars at Day 20. In contrast, the model group failed to form external collars at the same time point (Fig. 1A). The statistical analysis of the Micro-CT pa- rameters TV, BV, BV/TV, BMD of TV and BMD of BV implied an increasing tendency from Day 10 to Day 20 (Fig. 1D). After the treat- ment of FG4592, TV, BV, BV/TV, BMD of TV and BMD of BV were significantly higher compared with the model group at both time points. 3.2. FG4592 induced new cartilage formation We further conducted the histological analyses in fracture site. HE staining (Fig. 2A) and Safranin O staining (Fig. 2B) showed that the fracture healing greatly accelerated after administration. The rats of the FG4592-treated group displayed a significantly high degree of newly formed cartilage (black arrow) and a meaningfully thick external callus at Day 10. Furthermore, At Day 20 post administration of FG4592, new bone (blue arrow) and bridging bone were observed, the femurs started remodeling and returning to normal geometries, but the percentage of newly formed cartilage area declined. 3.3. FG4592 up-regulated HIF-1α, CD90 and VEGF protein expression during fracture healing To determine the role of HIF pathway in fracture healing, the expression of HIF-1α, CD90 and VEGF proteins in fracture site were determined by immunohistochemistry. The results indicated that the expression levels of HIF-1α, CD90 and VEGF protein in the FG4592- treated group were significantly higher than that of controls at Day 10 (Fig. 3A, B, C). At Day 20 post administration of FG4592, the protein concentrations of HIF and CD90 revealed a significant decrease compared with that of Day 10, but still higher than that of the control group. The high expression of CD90, which was the surface marker of stem cells, represented an increase in the number of BMSCs. 3.4. FG4592 up-regulated Ca2+/NO/Total Antioxidant Capacity and down-regulated H2O2 level in SD rats The level of Ca2+, NO, H2O2 and Total Antioxidant Capacity in SD rats after administration of FG4592 were determined by an automatic enzyme-linked immunosorbent assay plate reader using the corresponding assay kits. The rats of the FG4592-treated group dis- played a significantly high degree of Ca2+, NO at Day 10 (Fig. 4A, B). At Day 20 post administration of FG4592, similar results were obtained. However, hydrogen peroxide levels dropped after administration, whether for Day 10 or Day 20 (Fig. 4C). Consistent with this, the Total Antioxidant Capacity increased (Fig. 4D). These results indicated that FG4592 up-regulated Ca2+/NO and removed ROS in SD rats. 3.5. FG4592 enhanced cell proliferation of BMSCs Flow cytometry demonstrated that adherent spindle cells from bone marrow were BMSCs (positive for CD90, but negative for CD11b and CD45). This was confirmed by the results of induction cells differenti- ation into osteoblasts and adipocytes (Supporting Information Fig. S2A). FG4592 boosted the survival of BMSCs significantly. 1 μM FG4592 increased cell viability by 50% (Fig. 5A). At low concentrations, the increase in BMSCs is more promoted, possibly because of the lower toxicity. The expression of HIF-1α protein promoted by FG4592 was about 2.5-fold higher than that of control group (Fig. 5B). And FG4592 restrained the cell death via inhibiting apoptosis (Supporting Informa- tion Fig. S2B) and induced a significant decline in the S ratio (Supporting Information Fig. S2C). Moreover, the ΔΨm was maintained in the BMSCs after FG4592 treatment (Supporting Information Fig. S2D). To demonstrate the effects of HIF-1α on the cell proliferative capacity, the BMSCs were transiently transfected with plasmid DNA HIF- 1α (si-HIF-1α) or si-NC (as a negative control). Western blot analysis showed that the expression of HIF-1α protein was approximately 83% knocked down after 12 h of transfection compared with the control group (Fig. 5C). FG4592 had insignificant effects on the expression of HIF-1α protein after transfection (Fig. 5C). When the expression of HIF was decreased, FG4592 no longer promoted the proliferation of BMSCs and the cell viability decreased by over 20% compared with the controls (Fig. 5A). These data confirmed that HIF-1α protein was required for the proliferation of BMSCs. 3.6. FG4592 induced cell migration of BMSCs In the absence of FG4592, BMSCs showed a limited ability to cross the filter. When FG4592 was added, the number of cells passing through the filter were significantly increased (Fig. 5D). The greater the con- centrations, the more the cells migrated. When HIF-1α was silenced, the number of cells passing through the filter decreased by 17%–30% (Fig. 5D). These data confirm that HIF-1α protein was also required for the cell migration of BMSCs. 3.7. FG4592 upregulated the concentration of intracellular Ca2+ via the HIF pathway The concentration of intracellular Ca2+ increased by 300% after dosing with FG4592 for 24 h (Fig. 6A), but decreased almost 20% when HIF-1α was silenced (Fig. 6A). Short-term results showed that intracel- lular Ca2+ elevated rapidly after administration, and then slowly decreased over time, but still maintained a higher level than that of the control group (Fig. 6B, C). 3.8. FG4592 upregulated the concentration of intracellular NO via the HIF pathway The concentration of NO was also increased 2.5-fold after dosing 24 h (Fig. 7A). When the HIF was knocked down, the intracellular NO decreased by over 30% (Fig. 7A). Short-term results also showed that intracellular NO enhanced rapidly after administration, and then slowly reduced over time, but still maintained a higher level than the control group (Fig. 7B, C). This was consistent with the changes in Ca2+. 3.9. FG4592 decreased the concentration of intracellular ROS via the HIF pathway The concentration of ROS in administration group significantly decreased by about 30% (Fig. 8A), which was consistent with the in- crease in proliferation and the inhibition of apoptosis. As the HIF decreased, the ROS level significantly increased by 20%–30% (Fig. 8A). Short-term results showed that intracellular ROS decreased rapidly to near zero after administration and still maintained in 300 s (Fig. 8B, C). 3.10. Regulation the expression of mRNA in the HIF-1α pathway The RT-PCR results (Fig. 9A, B) demonstrated that the expression of HIF-1α remained almost unchanged regardless of 1 μM or 10 μM FG4592. FG4592 induced the accumulation of the VEGF, NOS2, IP3R1 mRNAs in a dose-dependent manner and downregulated RCN1, CaMKII mRNAs compared with the control group. 3.11. Regulation the levels of protein expressions in the HIF-1α pathway FG4592 elevated iNOS and IP3R1 protein expressions of BMSCs in a concentration-dependent manner. However, RCN1 and CaMKII were sharply downregulated (Fig. 9C, D). In addition, an ELISA kit was used to measure the expression level of VEGF. FG4592 induced the accumula- tion of VEGF in 25–250% compared with the control group (Fig. 9E). 4. Discussion Regeneration and self-healing are the natural abilities of bone. Bone fracture recapitulates processes that operate during skeletal develop- ment and requires close temporal and spatial coordination of events involving resident bone cells, marrow stromal elements, and associated vascular structures [37,38]. The regenerative capacity of bone tissue resides in the population of BMSCs with multipotential capacity [39]. BMSCs can differentiate into osteoblasts, vascular endothelial cells, neurons, adipocytes and chondrocytes under specific conditions [40]. In recent years, as limitations of traditional therapies and the increasing number of bone grafting procedures, cell therapy has been proposed. BMSCs transplantation is an effective method to treat fracture non-union [2]. However, the low survival rate of cells after transplantation and shortage of safe, effective migration agents need to be solved. A prom- ising strategy to advance the proliferation and migration of BMSCs via HIF pathway was put forward. The present study has provided evidence both in vivo and in vitro for a contribution of HIF to the enhancement of fracture healing. A few points from our findings supported the results. First, FG4592 accelerated the healing of fractures with the number of BMSCs increased and endochondral ossification in a femoral fracture model. As aged rats have many self-factors and the conditions are relatively complex, adult rats were chosen to make fracture models. FG4592 displayed an increase in newly formed cartilage at Day 10 and the immunohistochemistry confirmed that this process was closely related to the expressions of HIF- 1α and CD90 protein. The expression of CD90 protein, surface marker of stem cell, was detected in fracture site. And the high expression of CD90 represented a growth in the number of BMSCs. In the early stage of fracture (at Day 10), the number of BMSCs rose rapidly with the increase of HIF protein expression after treatment with FG4592. At Day 20, cartilage formation accelerated while the number of BMSCs decreased compared with the controls, indicating that BMSCs differentiated into chondrocyte and boosted the fracture healing. Second, FG4592 induced a tremendously high proliferation rate response in BMSCs. Confocal imaging showed that ROS rose slowly in normal cultured cells. Once FG4592 was added, ROS was removed immediately. Further, the apoptosis of BMSCs was inhibited and the ΔΨm was maintained after treatment with FG4592. When HIF-1α gene was silenced, the concen- tration of intracellular ROS went up with the survival of BMSCs decreased. Combined with the results, we speculated that the FG4592 promoted cell proliferation mainly by eliminating ROS. Third, FG4592 markedly improved the migration activities of BMSCs and upregulated the concentration of intracellular Ca2+ and NO. In normal culture, the concentration of Ca2+ and NO remained constant. When the FG4592 was added, the concentration of Ca2+ and NO rose immediately. Once HIF-1α was silenced, the levels of Ca2+ and NO were declined with a reduction of cell migration. These results indicate that Ca2+ and NO, which are regulated by FG4592, are essential for the induction of BMSCs migration. To further explore the mechanism of FG4592 treatment, we detected the expression of the proteins based on the ITRAQ data in MG63 cells: the Ca2+-related proteins of CaMKII, IP3R1 and RCN1; the NO-related protein, iNOS; the vascular proliferation-related protein, VEGF, respectively. The intracellular free Ca2+ level is associated with the stability of HIF-1α and CaMKII [41]. CaMKII is not only necessary for Ca2+ homeostasis and reuptake in cardiomyocytes [42], but also plays a part in osteoclast differentiation and bone resorption [43]. CaMKII activation has been reported to participate in HIF-1α stabilization and the CaMKII regulation of HIF-1α is regarded as a cell type-dependent event [17,18]. Here, we reported that FG4592 decreased the expres- sion of the CaMKII protein and gene while maintaining the stability of the HIF-1α in BMSCs, which supported the cell type-dependent action of CaMKII. IP3R1 is a membrane glycoprotein complex that acts as a Ca2+ channel [44]. IP3R1 represents a dominant second messenger that leads to the release of Ca2+ from intracellular store sites. Strong evidence suggests that IP3R1 participates in the conversion of external stimuli to intracellular Ca2+ signals characterized by complex patterns relative to both space and time, such as Ca2+ waves and oscillations [45]. Our re- sults showed that the level of IP3R1 protein and gene rose significantly with the increase of HIF in BMSCs. Studies have shown RCN1 is expressed in endothelial cell (EC) lines, including bone marrow ECs. Herein we reported for the first time that RCN1 expressed in BMSCs, which could be regulated by FG4592. Together, we speculated that FG4592 mainly enhanced the level of intracellular Ca2+ by tuning the expression of CaMKII, IP3R1 and RCN1. Ca2+ released from α calcium sulfate led to increased levels of HIF-1α as well as improved expression of VEGF and osteoblast phenotype markers [46]. Our findings implied that FG4592 upregulated the HIF-1α expression with the elevation of intracellular Ca2+, in turn, the increase of Ca2+ mediate HIF-1α stabili- zation. This positive feedback mechanism plays a beneficial role in regulating disease treatment. In contrast to the critical calcium-dependent regulation of constitu- tive NOS enzymes, iNOS has been described as calcium-insensitive. Reher et al. have demonstrated that ultrasound increases in NO syn- thesis in osteoblasts, which are regulated by iNOS [47]. Our results are in agreement with this conclusion. After treatment with FG4592, both the gene and protein levels of iNOS grew with elevation of intracellular NO concentration. NO has been found to be involved in HIF-1α stabili- zation [27], and we also showed that the increasing protein concen- tration of HIF-1α mediate NO concentration in BMSCs. However, the effects of prolyl hydroxylase inhibitors on Ca2+ and NO in vivo need further exploration. Some studies indicate that the increase of Ca2+ and NO concentration can maintain the stability of HIF-1α, and our results showed that the up-regulation of HIF expression could also enhance the Ca2+ and NO level. Therefore, we can reasonably speculate that in the fracture model of SD rats, the increase of HIF protein concentration will be accompanied by the upregulation of Ca2+ and NO. It is well documented that VEGF has positive effect on angiogenesis and angiogenesis is an important step of fracture healing. Here, we demonstrated that FG4592 significantly upregulated VEGF production in rats and BMSCs. And the concentration of VEGF in FG4592-treated Day 20 group rose more significantly than that in Day 10. However, there was no significant difference in fracture healing between FG4592- treated groups of Day 10 and 20 (Fig. 5B). It can be concluded that the angiogenesis induced by VEGF and the proliferation, migration of BMSCs play a combined role in promoting fracture healing. In conclusion, a possible signaling pathway of FG4592 in bone fracture healing was established (Fig. 10). FG4592 promoted the accu- mulation of HIF, upregulated the intracellular Ca2+, NO and eliminated the ROS by regulating activity of a broad array of proteins (VEGF, iNOS, IP3R1, CaMKII and RCN1). And then these effects markedly improved the proliferation and migration of BMSCs. Furthermore, BMSCs migrated to fracture site, proliferated and differentiated into cartilage, promoting fracture repair. We are the first to report that Ca2+, NO and ROS play a critical role in the proliferation and migration of BMSCs. And for the first time, the relationship between HIF and IP3R1, RCN1 pro- teins were discovered in BMSCs. We proposed an effective strategy that enhanced BMSCs proliferation and migration. And a potential strategy to activate HIF pathway to facilitate fracture healing by inducing BMSCs proliferation and migration was provided.