CK was provided by Zhejiang Hongguan Bio-pharma Co., Ltd., and dissolved in DMSO, and diluted in PBS. Modified Eagle’s Medium of Alpha (α-MEM), Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12), fetal bovine serum (FBS), and penicillin/streptomycin were purchased from Gibco (United States). Beta-glycerolphosphate, dexamethasone, ascorbic acid phosphate, Safranine O, and Fast Green were purchased from Sigma (United States). Alizarin Red S was purchased from Solarbio (Beijing, China). Cell Counting Kit-8 (CCK-8) was purchased from Beyotime (Beijing, China). NucleoZOL reagent, Reverse Transcription Kit and SYBR-Green Master Mix were supplied by Takara (Japan). Hematoxylin-eosin (H&E) was purchased from biosharp (China). Primary antibodies against CD31, β-catenin, and DAPI were supplied by Santa Cruz Biotechnology (United States); Primary antibodies anti-CTSK, anti-ALP, anti-OPG, anti-RANKL, anti-Runx2, and anti-OPN were purchased from Bioss (China); Anti-GAPDH, and DAPI were obtained from Abcam (United States). Secondary antibodies HRP-conjugated Goat Anti-Rabbit IgG, Goat anti-Mouse IgG (H + L), Rabbit Anti-Rabbit IgM/Cy3 and Rabbit Anti-Mouse IgM/FITC were obtained from Bioss (China). Dual-Luciferase Reporter Assay System was supplied by Promega Company (United States). Matrigel was purchased from Becton Dickinson (United States).

After inducing differentiation for 3 days in 12-well plates with osteogenic induction medium, the total RNA was extracted using NucleoZOL, and cDNA was obtained from total RNA using a Reverse Transcription Kit. Next, qRT-PCR was performed using SYBR Green qPCR Master Mix. The relative gene expression was calculated by the 2^–ΔCT method, and GAPDH was used as a reference for normalization. The primers were purchased from Invitrogen (United States) and primer sequences are shown in Table 1.

The method of rat bone marrow mesenchymal stem cells (BMSCs) isolation and cultivation has been described previously (Shen et al., 2018). BMSCs were isolated from Sprague-Dawley rats (male, 2 weeks old, 30–40 g) in a sterile environment. Rats were euthanized and the bone marrow of the bilateral femoral was flushed out with serum-free α-MEM to obtain a single-cell suspension. Flushed bone marrow cells was centrifuged at 1200 rpm for 6 min, the supernatant liquid was removed, and the cell pellets were resuspended in α-MEM supplemented with 10% fetal bovine serum, 1% penicillin and streptomycin. The medium was discarded after 72 h of primary culture and then changed once every 3 days. Upon 80—90% confluence, the adherent cells were further expanded with trypsin. The medium was changed several times to obtain pure BMSCs. Cells from passages 3 to 5 were used in the study.

The cell viability of CK on BMSCs was assessed using a Cell Counting Kit-8 (CCK-8) kit (C0037, Beyotime Biotechnology Co., Ltd., Shanghai, China). BMSCs or HUVECs were seeded in a 96-well plate with 5000 cells in each well and cultured without or with CK (0, 2.5,5, 10, 20, 30 and 40 μM) for 48 and 72 h. Subsequently, CCK-8 reagent was added to each well and the plates were incubated for 1 h at 37°C. The Optical density (OD) of the samples was measured at 450 nm with a spectrophotometric microplate reader (Xianke Instruments, Shanghai, China). The experiment was independently repeated 3 times.

8 week-old male Sprague Dawley (SD) rats (220 ± 10 g, n = 18) were purchased from the Guangzhou Medicine Laboratory Animal Center. The animals were housed in the First Affiliated Hospital of Guangzhou University of Chinese medicine animal center. All experimental methods were approved by the Animal Care and Use Committee of Guangzhou University of Traditional Chinese Medicine. All rats were fed with standard chow and free access to water with a 12 h light-dark cycle (24 ± 1°C). An open femoral fracture model with internal fixation was established. In brief, the procedure was performed under general anesthesia (pentobarbital sodium 100 mg/kg, intraperitoneally) and aseptic conditions. The right femur was exposed, and transverse osteotomy was performed with a hand saw to create a gap size of 2 mm. A K-wire (diameter: 1.2 mm, Stryker Ltd., United States) was inserted into the right femoral bone marrow cavity to fix the fracture. The rats were randomly assigned to the following two groups: fracture + PBS (n = 9, 100 μl/day), fracture + CK (n = 9, CK = 500 μM, 100 μl every other day). 5 days after the surgery, CK or PBS was locally injected at the fracture sites every other day for 4 weeks. Then the animals were sacrificed and the right femurs were collected for further analysis.

The femurs were harvested, fixed in 10% neutral formalin for 24 h, decalcified in 10% Ethylene Diamine Tetraacetic Acid (EDTA) for 21 days, dehydrated and then embedded in paraffin. After cutting into 5-μm-thick sections and dewaxing in xylene and rehydration in a decreasing alcohol gradient and distilled water, samples were processed for hematoxylin-eosin (H&E), and Safranine O-Fast Green (SO-FG) staining. For immunohistochemical staining, the sections were incubated in 0.3% hydrogen peroxide for 20 min and antigen retrieval in 0.01 M citrate buffer at 60°C for 30 min, and blocking with 5% g BSA in PBS for 1 h and then incubated overnight at 4°C with a primary antibody. The sections were then incubated with the secondary antibodies for 1 h at 37°C, counterstained with hematoxylin, and visualized using an HRP-streptavidin system. The primary antibodies used in this study included anti-OPG, anti-RANKL, anti-OPN and anti-ALP antibodies. For immunofluorescence staining, the slides were incubated in antigen retrieval in 0.01 M citrate buffer, and blocking with 5% g BSA in PBS for 1 h and then incubated overnight at 4°C with a primary antibody. The sections were then incubated with the fluorescent secondary antibodies for 1 h at 37°C. The primary antibodies used in this study included anti-Runx2, anti-Endomucin, anti-CD31 and anti-β-catenin antibodies. DAPI staining was carried out to stain the nuclei. Images were acquired with the fluorescence microscope (Olympus, IX73 L, United States).

Microcomputer tomography (micro-CT) examination was applied for the fractured femurs. Samples were scanned by Skyscan 1176 micro-CT scanner (Bruker micro-CT, Kontich, Belgium), with a source voltage of 80 kV, current of 114 μA, Al 0.5 mm filter and 10.5 μm isotropic resolution. The fractured callus sites were defined as the volume of interest. The bone volume/tissue volume (BV/TV), mean volumetric bone mineral density (BMD), and Callus Volume were measured. Three dimensional images were generated using CTvol software (Bruker micro-CT, Kontich, Belgium).

Tube formation assay was performed as previously reported (Lin et al., 2019). The wells of the 12-well plate were coated with Matrigel and incubated for 30 min hBMSCs (5000 cells/well) were treated with or without CK (10 μM), and co-cultured with HUVECs (10⁵ cells/well). DMEM/F12 basal medium containing 2% FBS (Thermo Fisher Scientific) was used. Plates were incubated at 37°C, 5% CO₂ for 8 h. Then the tube formation was observed using a microscope. Tubes were then assessed through an inverted fluorescent microscope at 10×(Olympus). Image J with the Angiogenesis Analyzer plugin were used to quantify the tube length and branch points of tube networks. Images taken at 5× magnification.

Experiments were performed as described previously (Ren et al., 2019). Briefly, 293FT cells were seeded on 24-well plates and allowed to grow to 80% confluence. Cells were then transfected with TOPflash (500 ng) and Renilla reporter plasmid pRL-CMV (100 ng) using Lipofectamine 8000. 24 h after transfection, cells were treated with CK (10 μM) for 24 h. The luciferase activity was measured using a GloMax™ 20/20 single-tube luminometer (Promega, Madison, WI, United States).

Three-point bending biomechanical testing was performed as previously reported (Lin et al., 2019). Fractured femurs were tested to failure with a constant displacement rate of 4 mm/min by a 3-point bending device (H25KS; Tinius Olsen, United Kingdom). The fractured femur was loaded in the front and back directions, and the span of the two support points was set as 10 mm. The force loading point was set at the fracture site. After testing, ultimate load to failure and energy absorbed to failure were recorded and analyzed by the QMAT software.

We applied GraphPad Prism 5 for comparison. Quantitative data were expressed as mean ± standard deviation (SD). Statistics were analyzed by t-test for two-group comparison, and One-way or two-way analysis of variance (ANOVA) for multi-comparison between groups. We used Tukey’s post hoc multiple comparisons test as the posttest method for ANOVA. P values <0.05 were considered statistically significant.

The chemical structure depiction of CK was shown in Figure 1A. BMSCs were cultured without or with CK (0, 2.5,5, 10, 20, 30 and 40 μM) for 48 and 72 h. The CCK-8 assay results showed that CK exhibited no obvious cytotoxicity to BMSCs even at the high concentration of 40µM, and CK at 10–40 µM increased the cell viability of BMSCs (Figure 1B). To investigate the effect of CK on osteogenic differentiation of BMSCs, BMSCs were cultured in the osteogenic induction medium (OIM) with or without CK at various concentrations (0, 1, and 10 μM) for 3 and 14 days. At the third day of osteogenic induction, the expression of osteogenesis-related genes was detected and it was found that the RNA levels of osteopontin (OPN), alkaline phosphatase (ALP), osteocalcin (OCN) and osterix (OSX) were significantly up-regulated by CK (Figures 1C–F). Furthermore, Alizarin Red S staining showed that CK significantly increased the formation of calcium deposits after 14 days of induction (Figures 1G,H). As 10 μM CK showed the best effect, it was used for the following in vitro experiments.

HUVECs were cultured without or with CK (0, 2.5,5, 10, 20, 30 and 40 μM) for 48 and 72 h. The CCK-8 assay showed that CK exhibited no obvious cytotoxicity to HUVECs (Figure 1I). Some studies have reported MSCs can stimulate migration and angiogenesis of HUVECs (Chiang et al., 2018; Liu et al., 2019). To investigate the effect of CK on angiogenesis in the presence of MSCs, the HUVEC cells were co-cultured with BMSCs treated with or without CK (10 µM), and the tube length and branch points were evaluated (Figures 1J–L). We could observe that the tube formation capacity of HUVEC was increased by CK. The result demonstrated that CK enhanced angiogenesis in vitro compared with the control group.

It is well known that the Wnt/β-catenin signaling has been shown as an important regulatory pathway in the osteogenic differentiation of mesenchymal stem cells (Kim et al., 2013). To verify whether the Wnt/β-catenin signaling pathway is activated upon CK treatment in BMSCs, we performed immunofluorescence staining to detect the level of β-catenin and its co-location with Runx2 in rBMSCs. Immunofluorescence analysis revealed more nuclear translocation of β-catenin and an increased expression of Runx2 in the CK (10 μM) group, compared with the OIM group (Figures 2A,B). Additionally, the TOP flash assay was used to evaluate the effect of CK (10 μM) on the activation of the Wnt/β-catenin signaling pathway. After 24 h of stimulation, the luciferase activity was significantly increased by CK (Figure 2C).

We further conducted an open femoral fracture model to evaluate whether CK could accelerate fracture healing in rats. The time points of animal modeling and sample collection were shown in Figure 3A. The 3-dimensional images of the femurs obtained by micro-CT analysis showed that the fracture gap was almost filled by new bone in the CK-treated rats at 4w post-fracture, compared with that of the PBS group (Figure 3B). The BMD, BV/TV and callus volume at femoral callus sites of CK-treated rats were significantly higher than those in the PBS group (Figures 3C–E). In addition, results of biomechanical testing confirmed a much stronger biomechanical property in the femoral bones of CK- treated group than those of the PBS group (Figures 3F,G).

Furthermore, H&E and SO-FG staining of the callus showed varying amounts of newly formed trabecular bone, cartilage tissue and fibrous-like tissue. Callus of CK-treated rats exhibited enhanced bone regeneration after 4 weeks, in comparison with the PBS group, which was evidenced by more neo-formed trabecular bone and less cartilaginous and fibrous-like tissue in the CK group (Figures 4A,B). Additionally, higher expression of OPG, OCN, ALP and lower expression of RANKL within the callus areas of the CK group was confirmed using immunohistochemical analysis (Figures 4C–F).

In addition, the H-type vessel was observed in vivo in sections of femur. The CD31 and Emcn double immunofluorescent staining revealed a greater population of CD31^hiEmcn^hi cells within the callus of the CK group, compared with the PBS group (Figures 5A,B).

To detect the expression of β-catenin in fracture callus treated with CK, we performed immunofluorescent staining to detect the level of β-catenin and Runx2 in fracture callus. Immunofluorescence analysis revealed a higher expression of β-catenin and Runx2 within the callus of the CK group, compared with PBS group (Figures 6A,B), which is consistent with the in vitro experiment.