Human bone marrow-derived mesenchymal stem cells (MSCs) were isolated from bone marrow according to the previous studies (Zhang et al., 2011; Feng et al., 2018). Briefly, bone marrow was aspirated from a healthy 38 year-old male donor with formal consent and approval by the local ethics committee. The isolated cells were cultured in a-minimum essential medium (α-MEM, Invitrogen, Carlsbad, CA, United States), supplemented with 10% fetal bovine serum (FBS, Gibco, United States) and 1% penicillin-streptomycin (Gibco) in a humidified atmosphere containing 5% CO2 at 37°C. The osteogenic differentiation of MSCs was induced by the classical inducers including 10 nM dexamethasone (Sigma-Aldrich, St. Louis, MO, United States), 50 μg/ml ascorbic acid (Sigma-Aldrich), and 10 mM glycerol 2-phosphate (Sigma-Aldrich) as described previously (Sun et al., 2015). The differentiation medium was replaced every 3 days. TRX was purchased from Aladdin (Shanghai, China) and dissolved in 0.1% DMSO for usage 0.1% DMSO was used as control.

The human MSCs were seeded in 96-well plate at density 1.5 × 10⁴ per well for 24 h, then treated with different concentrations of TRX (from 0 to 200 µM). After incubated for 24, 48, and 72 h, the methylthiazolyl tetrazolium (MTT) assays were performed. The MTT solution (0.5 mg/ml) was added, and after 4 h, the deposition was dissolved with 100 µL DMSO. The absorbance was measured at the wavelength of 570 nm using Spectramax Gemini dual-scanning microplate reader (Molecular Devices, United States).

The human MSCs were treated with TRX in the osteogenic induction medium (OIM) for 7 days, then the cells were collected, qualitative, and quantitative examination of ALP activity were carried out according to the protocol of BCIP/NBT Alkaline phosphatase Color Development Kit (Beyotime, Shanghai, China). For the alizarin red S staining, the TRX-treated MSCs were incubated in OIM for 14 days, and washed with PBS twice and fixed with ice-cold 75% ethanol for 15 min. The MSCs were then stained with 2% Alizarin Red S staining solution (pH 4.2, Leagene, Beijing, China) for 30 min. The stained calcified nodules were dissolved with 10% Hexadecylpyridinium chloride monohydrate (Sigma-Aldrich) at room temperature and then the absorbance was detected at the wavelength of 562 nm.

Total RNA was extracted by the Animals Total RNA Isolation Kit (FOREGENE, Chengdu, China) according to the manufacture’s instruction. After the concentration and purity of total RNA were measured by Nanodrop (Thermo Fisher Scientific), cDNA was reversely transcribed from RNA samples by PrimeScript™ RT Reagent Kit (TaKaRa, Japan). The PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific) was applied for the qPCR examination using the LightCycler 480 system (Roche, Basel, Switzerland). The relative fold changes of candidate genes were analyzed by using the 2^−ΔΔCt method. The house-keeping gene GAPDH was served as internal control. Primer sequences used in qPCR examination were displayed in Table 1.

The hMSCs were seeded into six well at the density of 2 × 10⁴ for 24 h. Cells grown to 70–80% confluence were transiently transfected with TOPFlash and PGMLR-TK by using the Lipofactamine™ 3,000 Reagent (Invitrogen). About 24 h later, the cell culture medium that contains TRX (100 μM, 200 μM) was added. After 2 days incubation with TRX, the cells were harvested and lysed for measuring the luciferase activity with Bright-GloTM luciferase Assay System (Promega) following the manufacturer’s instruction.

Total protein was extracted by RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific) supplemented with the protease inhibitor cocktail (Roche). And the nuclear and cytoplasmic protein were isolated by the Nuclear and Cytoplasmic Protein Extraction Kit (KeyGEN, Nanjing, China) according to the manufacturer’s instructions. The protein was qualified by the BCA assay kit (Thermo Fisher Scientific). Then the soluble protein was separated by SDS-PAGE (10%) and transferred to PVDF membranes. The membranes were then blocked with 5% skimmed milk and probed with the following antibody: β-catenin (1:1,000; Cell Signaling Technology, United States) or GADPH (1:2,000; Cell Signaling Technology, United States). After incubation with the appropriate secondary antibody conjugated with HRP (1:1,000 dilution), the chemiluminescence (ECL, Hangzhou, China) was applied to visualize the bands.

Twenty-four Sprague-Dawley rats (male, 12-week-old) were purchased from the Laboratory Animal Research Centre, Southern Medical University. Animal ethics was approved by the Institutional Animal Care and Use Committee (IACUC) of Southern Medical University (Guangzhou, China). The modified rat closed transverse femoral fracture model was applied in this study (Nyman et al., 2009; Liang et al., 2019). Briefly, the surgery was carried under general anesthesia (40 mg/kg ketamine and 4 mg/kg xylazine) and sterile condition. A 2-cm incision was made in the lateral aspect of right thigh, the osteotomy was made with an electric saw at the middle site of the femur, and then the femur was fixed with Kirschner’s needle penetrated into medullary cavity. Animals were randomly divided into four groups (n = 6 per group): surgery without any treatment (Blank); femur fracture with 0.1% DMSO treatment group (Control); femur fracture with a low dose injection of TRX (100 µM); and femur fracture with a high dose of TRX (200 µM). TRX was administered intramuscularly every other day after surgery. X-rays radiography was taken to determine the status of fracture healing at week 4 and 6. At week 4 and 6 after surgery, animals were randomly selected to sacrifice and the femurs were collected for further analyses.

The fractured femur was examined using a high-resolution micro-CT instrument (SkyScan1172, Bruker, German). The samples were scanned by the procedure with a source voltage of 70 keV, current of 80 µA, and 9 µm isotropic resolution. The region of interest (ROI) for scanning was set at 2.2 mm up and below of the fracture plate. Two-dimensional data obtained from the scanned femur were applied for the three-dimensional reconstruction and several morphometric parameters including trabecular bone volume (BV), tissue volume (TV), Bone surface (BS), BV/TV and BS/TV were calculated using the CT-VOX software (CtAN, Bruker, German).

All femurs were initially fixed with 10% formalin for 72 h, followed by decalcification in 10% EDTA solution for 2 weeks. The samples were cut into 5 µm thick sections and stained with hematoxylin and eosin (H and E) according to the previous study (Molvik and Khan, 2015). The femurs were sliced along the long axis in the coronal plane for the bone-fracture study. IHC staining was performed as previously reported (Sun et al., 2015). Secretions were incubated with primary antibodies of osteocalcin (OCN, 1:100, abcam13418), osterix (OSX, 1:100, ab22552) and β-catenin (1:100; Cell Signaling Technology) overnight at 4°C. The horse-radish peroxidase-streptavidin system (Dako, United States) was applied for IHC signal detection, followed by counterstaining with hematoxylin. Finally, the images were captured under the Positive Fluorescence Microscope (Olympus BX53F, Japan). The analysis of the positive-stained cell area was performed using NIH ImageJ software.

At least triplicates were taken in each experiment and the data were recorded as the mean ± SD. Student’s t test and One-way ANOVA were used for statistically analysis between inter-groups analyses, and a p-value of less than 0.05 was considered statistically significant.

In this study, we firstly examined the inhibitory effects of TRX on cell proliferation of human MSCs. By using MTT examination, TRX, its chemical structure showed in Figure 1A, exhibited no significant inhibitory effects on cell viability of human MSCs, even with the concentration of 200 µM (Figures 1B–D), suggesting it has low cytotoxicity to MSCs.

To investigate the effects of TRX on osteogenesis, TRX-treated MSCs were induced to differentiate into osteoblast. ALP activity, an early marker of osteogenic differentiation, was monitored at day 7. As shown in Figure 2A (up-panel) and 2B, TRX treatment enhanced the ALP activity in a dose-dependent manner from 0 to 200 µM. The ARS staining examination also confirmed that TRX promoted calcium nodule formation at day 14, consistent with ALP activity (Figure 2A down-panel and 2C). Based on the results of ALP activity and ARS staining, TRX with the concentration of 100 and 200 µM were selected for further investigation. We next examined the expression of osteogenic markers, including OSX, Runx2, and OPN by qRT-PCR examination; and the results showed that they were significantly up-regulated by 100 and 200 µM TRX, especially by 200 µM TRX (Figure 2D).

Considering that Wnt/β-catenin signaling plays a pivotal role in regulating osteoblast differentiation and bone formation, we wondered whether this signaling was involved in TRX-mediated osteogenesis. As shown in Figure 3A, the luciferase activities of the Wnt/β-catenin signaling reporter TOPflash were significantly promoted by TRX. And the expression of total β-catenin, the key regulator of Wnt signaling, was promoted by TRX treatment at protein level (Figure 3B). It is well known that β-catenin accumulates in the nucleus and stimulates Wnt/β-catenin signaling and regulates gene transcription. We therefore examined the expression of intranuclear β-catenin and intracytoplasmic β-catenin, and the results showed that intranuclear β-catenin were significantly increased in TRX treated MSC cells (Figures 3B,C). Furthermore, several downstream target genes of Wnt/β-catenin signaling such as Cmyc, CD44, and Survivin were examined, and they were obviously up-regulated by TRX (Figure 2D). All of these data suggest that TRX induced the activation of Wnt/β-catenin signaling during the osteogenic differentiation via stimulating the translocation of β-catenin from the cytoplasm to the nucleus.

To further evaluate whether API could enhance fracture healing in vivo, a rat fracture model was established and TRX were locally injected into the fracture sites every other day (Figure 4A). The broken bones were monitored by X-rays examination and the results showed that the gaps between the fracture sites almost disappeared in TRX treated groups while it also remained clearly visible in control groups, especially in the 200 µM TRX treated groups (Figure 4B). We also statistically evaluated the callus and the results showed that the thickness of newly formed callus was significantly improved by TRX administration at week 4 and 6 (Figures 4C,D). These data indicated that the fracture healing process was accelerated in TRX treated animals.

We next investigated the bone formation during the fracture sites using micro-CT scanning, and images were reconstructed by three-dimensional modeling system. As shown in Figures 5A,B, more mineralized calluses were observed in the TRX-treated groups when compared with their respective control groups at week 4 and 6. In addition, the bone volume (BV), tissue volume (TV), and bone surface (BS) were recorded, and the data showed that they were significantly increased by TRX treatment at week 4 and 6 (Figures 5C,D). The ratios of BV/TV and BS/TV also exhibited a significant increase in the TRX-treated groups at week 4 and 6 (Figures 5C,D), indicating more newly formed bone with TRX treatment. We further evaluated the newly formed bone tissues by histological examination, the H and E staining showed that more osteoblasts were converged at the fracture sites of TRX treatment groups, suggesting more active bone regeneration in TRX groups (Figures 6A,B). Moreover, the OCN and OSX expression in the animal tissues was evaluated by immunohistochemistry staining. Representative micrograph images showed that positive staining of OCN and OSX was increased in TRX-treated groups (Figures 7A,B, Supplementary Figure S1), indicating the promoting effect of TRX on bone formation. We also investigated the β-catenin expression around the fracture sites, and the results showed that it was promoted by TRX in animal specimens (Supplementary Figure S2).