The present preclinical study was performed according to the recommendations in the Guide for the Care and Use of Laboratory Animals of China (16). All experimental protocols and animals were approved by the Committee on the Ethics of Animal Experiments Defence Research of the Second Hospital of Tianjin Medical University (Tianjin, China). Clinical samples were collected from The Second Hospital of Tianjin Medical University (Tianjin, China). All patients were required to write informed consent with a signature. The study was approved by the Ethics Committee of The Second Hospital of Tianjin Medical University (Tianjin, China). A total of 60 patients with osteosarcoma (27 males, 33 females; age range, 22–61 years; median age, 49 years) between July 2013 and March 2017 were recruited into the present study.

Osteosarcoma cell lines SAOS2 and MG63 cells and human normal osteoblast hFOB1.19 cells were purchased from American Type Culture Collection (Manassas, VA, USA). SAOS2 and MG63 cells were cultured in Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS; Invitrogen; Thermo Fisher Scientific, Inc.). hFOB1.19 cells were cultured in RPMI-1640 (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) medium supplemented with 10% fetal calf serum (Gibco; Thermo Fisher Scientific, Inc.). All cells were cultured in a 37°C humidified atmosphere containing 5% CO₂.

Total RNA was extracted from SAOS2, MG63 and hFOB1.19 cells using a RNAeasy Mini kit (Qiagen Sciences, Inc., Gaithersburg, MD, USA). A total of 1 µg total RNA was used to transcribe into cDNA by using the PrimeScript RT Master Mix (Qiagen Sciences, Inc.) according to the manufacturer's protocol. The cDNA (10 ng) was subjected to RT-qPCR with SYBR Green Master Mix system (Bio-Rad Laboratories, Inc., Hercules, CA, USA). All the forward and reverse primers were synthesized by Invitrogen (Thermo Fisher Scientific, Inc.). For the PCR experiments, the following forward and reverse primers were used: Caspase-3, forward, 5′-TGGCAGCAGTGACAGCAGCA-3′ and reverse, 5′-TACGGAGGTGGAGTGGGTGT-3′; caspase-8, forward, 5′-AGCCGAGGAAGAACTATGAAC-3′ and reverse, 5′-ATTTGAGGGTGAGGAATGGG-3′; P16, forward, 5′-GAGGGCAGAATCATCACGAAGT-3′ and reverse, 5′-TGAGAGATCTGGTTCCCGAAAC-3′; P21, forward, 5′-AGGCACGAGTAACAAGCTCAC-3′ and reverse, 5′-ATGAGGACATAACCAGCCACC-3′; Bcl-2, 5′-GTGGACATCCGCAAAGAC-3′ and reverse, 5′-AAAGGGTGTAACGCAACTA-3′; BAMBI, forward, 5′-AGGCACGAGTAACAAGCTCAC-3′ and reverse, 5′-ATGAGGACATAACCAGCCACC-3′; Bcl-2, 5′-AAGGAATTTGTAACAAAGGT-3′ and reverse, 5′-AGACCTGTGAGATGACCTCC-3′; and reference gene GAP DH, forward 5′-GTGGGCGCCCAGGCACCA-3′ and reverse, 5′-CTCCTTAATGTCACGCACGATTT-3′. Amplification conditions consisted of 5 sec of denaturation at 94°C, 9 sec of annealing at 55–60°C and 9 sec of extension at 72°C, for 45 cycles for each step. Relative mRNA expression changes were calculated by 2^−ΔΔCq (17). The results are expressed as the relative expression compared with control.

SAOS2 and MG63 cells were treated with BAMBI (0, 5, 10 and 15 mg/ml, Abcam, Cambridge, UK) or PBS and cultured in 96-well plates for 48 h. A total of 20 µl MTT (5 mg/ml) in PBS solution was added to each well, and the plate was further incubated for 4 h at 37°C. The medium was removed and 100 µl dimethyl sulfoxide (Sigma-Aldrich; Merck KGaA) was added into the wells to solubilize the crystals. The OD was measured by an iMark microplate absorbance reader (Bio-Rad Laboratories, Inc.) at wavelength of 450 nm.

TGF-β gene was cloned into a PMD-18-T vector and sequenced to identify its sequence, according to a previous report (18). The TGF-β gene was then cloned into a eukaryotic expression vector pCMVp-NEO-BAN (pTGF-β; Takara Biotechnology Co., Ltd, Dalian, China) to generate TGF-β-overexpressed SAOS2 or MG63 cells as described previously (18). SAOS2 and MG63 cells were cultured in 6-well plates in RPMI-1640 medium containing 10% FBS at 37°C until 90% confluence and the media was then removed. SAOS2 and MG63 (5×10⁶) cells were transfected with pTGF-β (1.0 µg, Takara Biotechnology Co.) or pvector (1.0 µg) using Lipofectamine^® 2000 (Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. Stable TGF-β-overexpression SAOS2 or MG63 cells were selected using the dihydrofolate reductase/glutamine synthetase (Invitrogen; Thermo Fisher Scientific, Inc.) screening system (19).

For analysis of the apoptosis of osteosarcoma, a TUNEL assay (Beyotime Institute of Biotechnology, Haimen, China) was used to detect TUNEL-positive cells. Tumor sections isolated from xenografted mice were fixed with 4% paraformaldehyde solution for 60 min at 4°C. The tumor tissues were washed with PBS three times and then permeabilized by immersing cells slides in 0.2% Triton X-100 solution in PBS for 30 min at 4°C. Subsequently, tumor tissues were incubated with equilibration buffer (Beyotime Institute of Biotechnology) for 30 min at 4°C. The tumor tissues were then incubated with 50 µl reaction mixture (Beyotime Institute of Biotechnology) at 37°C for 60 min and washed 3 times with PBS. These sections were incubated in 5% bovine serum albumin (BSA) for 30 min and the fragmented DNA was labeled with the TUNEL reaction solution at 37°C for 1 h. Hoechst 33258 (5 mg/l; H-33258; Sigma-Aldrich; Merck KGaA) was used to stain the nuclei for 10 min at room temperature. Converter-peroxidase was added to the sections at 37°C for 30 min prior to the TUNEL-positive nuclei being visualized by adding the DAB staining solution. Finally, tumor tissues images were captured with a confocal microscope at 488 nm. TUNEL-positive cells were counted in 5 randomly selected fields per section. The apoptosis rate was expressed as the ratio of TUNEL-positive cardiomyocytes to the total number of cardiomyocytes.

SAOS2 and MG63 cells were cultured in RPMI-1640 medium containing 10% FBS at 37°C until 90% confluence and then treated with BAMBI (10 mg/ml). Cells were continually cultured for 48 h at 37°C and then trypsinized, collected and washed in cold PBS 3 times. Subsequently, the cells (1×10⁶ cells/ml) were mixed with PBS, labeled with Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (Annexin V-FITC kit; BD, San Diego, CA, USA) according to the manufacturers protocol and analyzed with a FACScan flow cytometer (BD Biosciences, San Jose, CA, USA) and Quantity One software (version 3.0; Bio-Rad Laboratories, Inc.).

SAOS2 and MG63 cells were treated with BAMBI (5 mg/ml) for 24 h at 37°C and used to analyze the cell invasion and migration. SAOS2 and MG63 cells were placed in a 24-well culture plate with chamber inserts (BD Biosciences). For migration assays, 5×10⁴/well SAOS2 and MG63 cells in RPMI-1640 medium were placed into the upper chamber with the non-coated membrane at 37°C for 24 h. For invasion assays, the cells (5×10⁴/well) were placed into the upper chamber with the Matrigel-coated membrane. In the invasion assay, cells were treated with BAMBI (5 mg/ml) for 24 h and subjected to the tops of BD BioCoat Invasion Chambers (BD Biosciences), according to the manufacturer's protocols. The medium and serum in the lower chamber was DMEM plus 20% FBS. The number of tumor cells that invaded and migrated through the membrane were by stained with 0.5% crystal violet at 37°C for 30 min and counted at least three randomly selected fields per membrane under a light microscope in five random visual fields (magnification, ×200).

SAOS2 and MG63 cells were treated with SB431542 (5 ng/ml, cat. no 93-1674-1; Biovision Inc., Milpitas, CA, USA), cisplatin (10 mg/ml, Takara Biotechnology Co., Ltd.) and BAMBI (10 mg/ml) and harvested by scraping and lysed in radioimmunoprecipitation assay buffer (Invitrogen; Thermo Fisher Scientific, Inc.) followed by homogenization at 4°C for 10 min. Protein concentration was calculated using a BCA protein assay kit (Thermo Fisher Scientific, Inc.). Proteins (10 µg) were analyzed via 10% SDS-PAGE assays, followed by transferring onto polyvinylidene fluoride membrane. Proteins were then blocked with 5% BSA (Sigma-Aldrich; Merck KGaA) for 2 h at 37°C and incubated for 1 h at room temperature with primary rabbit anti-mouse antibodies against: BAMBI (1:500; cat. no. AF2387; R&D Systems, Inc., Minneapolis, USA); P21 (1:1,000; cat. no. ab109199; Abcam); P16 (1:1,000; cat. no. ab51243; Abcam); B-cell lymphoma 2 (1:1,000; Bcl-2; cat. no. ab692; Abcam); TGF-β (1:1,000; cat. no. AF532; R&D Systems); epithelial (E)-cadherin (1:200; cat. no. NBP238856; Novus Biologicals LLC, Littleton, CO, USA); vimentin (1:500; cat. no. PAB24865; Abnova, Taipei, Taiwan); Twist (1:500; cat. no. DR1088100UG; EMD Millipore, Billerica, MA, USA); Smad2 (1:500; cat. no. ab53110; Abcam); Smad3 (1:500; cat. no. ab40854; Abcam); caspase-3 (1:500; cat. no. ab13847; Abcam); caspase-8 (1:500; cat. no. ab25901; Abcam); pSMAD2 (1:500; cat. no. ab53100; Abcam); pSMAD3 (1:500; cat. no. ab63403; Abcam) and β-actin (1:500; cat. no. ab8226; Abcam). Subsequently, proteins were inoculated with rabbit horseradish peroxidase (HRP)-labeled IgG (1:10,000; cat. no. ab6728; Abcam) for 12 h at 4°C. The proteins expression levels were visualized using a chemiluminescence detection system (Nikon Corporation, Tokyo, Japan). Expression levels were determined relative to β-actin. The density of the bands was analyzed using Quantity One software version 4.62 (Bio-Rad Laboratories, Inc.). Protein expression was analyzed using BandScan 5.0 software (Glyko, Inc., Novato, CA, USA). All experiments were repeated ≥3 times.

Tumors from Mg63-bearing xenograph mice were fixed using 10% formaldehyde followed with being embedded in paraffin wax and cut into serial sections of 4 µm thickness. Paraffin-embedded tissue sections 4 µm thick were prepared and epitope retrieval was performed for further analysis. The paraffin sections were incubated with hydrogen peroxide (3%) for 10–15 min at 37°C and were subsequently blocked with a regular blocking solution (normal goat serum) for 10–15 min at 37°C. Tumor sections were incubated with primary antibodies against: TGF-β; E-cadherin; vimentin; and Twist. Subsequently, proteins were inoculated with rabbit HRP-labeled IgG (1:5,000; cat. no. ab6728, Abcam) for 12 h at 4°C. Specimens were visualized. Images were captured using fluorescence video microscopy (BZ-9000; Keyence Corporation, Osaka, Japan) at ×400 magnification. A Benchmark automated staining system (Ventana Medical Systems, Inc, Tucson, AZ USA) was used for observation of integrin.

Specific pathogen-free female nude (six-eight weeks old, 25–32 g) C57BL/6 mice were purchased from Shanghai Slack Experimental Animals Co., Ltd. (Shanghai, China). All mice were housed at room temperature with a 12/12 h light/dark cycle and fed ad libitum. Mouse breeding and experiments were carried out under the Institutional Animal Care and Use Committee approved protocols of Ethics Committee of Library Animals (16). Mg63 tumor cells (1×10⁷) were subcutaneously implanted into the right flank of C57BL/6 mice (n=60). Mice bearing osteosarcoma were randomly divided into two groups (n=30 in each group) and received treatment with BAMBI (10 mg/kg) or PBS. The treatments for tumor-bearing mice were initiated when tumor diameters reached 5–7 mm on day 3 following tumor inoculation. The detail procedures were referenced according to previous report (20). The treatments were continued seven times at intervals of every two days. Tumor diameters were recorded every two days and tumor volume was calculated using the formula: 0.52× smallest diameter² × largest diameter (21). The experimental mice were euthanized when tumor diameter reached 10 mm with 1% pentobarbital (200 mg/kg) administered via intravenous injection. On day 25, 10 mice in each group were sacrificed for further analysis, including immunohistochemistry.

All data are presented as the mean ± standard error of the mean of triplicate experiments. Unpaired data was determined by Student's t-test and comparisons of data between multiple groups were analyzed by analysis of one-way analysis of variance followed by Fisher's Least Significant Difference post hoc test. Kaplan-Meier analysis was used to estimate the risk of relapse and re-treatment during the 100-day treatment. All data analysis was performed using SPSS software (version 20.0; IBM Corp., Armonk, NY, USA). P<0.05 was considered to indicate a statistically significant difference.

Expression levels of BAMBI in osteosarcoma cell lines and clinical tumor tissues were analyzed. It was observed that mRNA and protein expression levels of BAMBI were upregulated in SAOS2 and MG63 cells, compared with normal cell line hFOB1.19 (Fig. 1A and B). It was also determined that BAMBI expression levels were lower in osteosarcoma tissues, compared with normal adjacent tissues, determined by immunohistochemistry (Fig. 1C). Results in Fig. 1D indicated that BAMBI expression may be an independent prognostic marker in osteosarcoma, as demonstrated by multivariate analyses. Collectively, these results indicated that BAMBI downregulation may be associated with osteosarcoma progression.

In order to analyze inhibitory effects of BAMBI on osteosarcoma cell growth and aggressiveness of SAOS2 and MG63 cells, BAMBI was added into cultured osteosarcoma cells. As depicted in Fig. 2A and B, BAMBI treatment significantly inhibited SAOS2 and MG63 cells growth in a dose-dependent manner (P<0.05). BAMBI (10 mg/ml) treatment was observed to notably suppress the migration and invasion of SAOS2 and MG63 cells after 24 h incubation (Fig. 2C and D). Collectively, the data indicated that BAMBI treatment inhibited the growth and aggressiveness of osteosarcoma cells in vitro.

Apoptosis of osteosarcoma cell lines SAOS2 and MG63 was investigated following treatment with cisplatin (10 mg/ml). As depicted in Fig. 3A, BAMBI (10 mg/ml) promoted the apoptosis of SAOS2 and MG63 cells induced by cisplatin, compared with non-treated cells. RT-qPCR demonstrated that gene and protein expression levels of caspase-3 and caspase-8 were upregulated in BAMBI-treated SAOS2 and MG63 cells (Fig. 3B-E). It was demonstrated that anti-apoptosis gene and protein expression levels of P16, P21 and Bcl-2 were decreased in SAOS2 and MG63 following BAMBI treatment (10 mg/ml), compared with the control (Fig. 3F-I). Collectively, these results indicated that BAMBI promotes apoptosis of osteosarcoma cells induced by cisplatin.

In order to investigate the potential mechanism underlying BAMBI-mediated inhibition of growth and aggressiveness in osteosarcoma cell lines, the TGF-β-induced EMT signaling pathway in SAOS2 and MG63 cells was analyzed. As depicted in Fig. 4A, BAMBI treatment suppressed TGF-β protein expression in SAOS2 and MG63 cells. Additionally, BAMBI treatment increased E-cadherin and inhibited vimentin and Twist expression in SAOS2 and MG63 cells (Fig. 4B). Results demonstrated that expression and phosphorylation levels of Smad2 and Smad3 were decreased by BAMBI treatment in SAOS2 and MG63 cells (Fig. 4C). It was determined that the blocked TGF-β receptor using SB431542 also suppressed the expression and phosphorylation levels of Smad2 and Smad3 in BAMBI-treated SAOS2 and MG63 cells (Fig. 4D); however, pTGF-β antagonized the expression and phosphorylation of Smad2 and Smad3, as well as EMT markers in BAMBI-treated SAOS2 and MG63 cells (Fig. 4E and F). Notably, it was determined that pTGF-β inhibited BAMBI-inhibited growth and invasion of SAOS2 and MG63 cells (Fig. 4G and H). Collectively, the results indicated that BAMBI reconstitution inhibits osteosarcoma growth and invasion via inactivating the TGF-β-induced EMT signaling pathway.

To further identify the therapeutic efficacy of BAMBI for osteosarcoma growth, osteosarcoma xenograft mice model were established. The results demonstrated that the intratumor injection of BAMBI (10 mg/ml) significantly inhibited tumor growth and tumor weight, compared with PBS-treated group (Fig. 5A and B; P<0.05). The TUNEL assay demonstrated that BAMBI increased apoptotic cells and decreased TGF-β expression in the tumor sections, compared with control group (Fig. 5C). Immunohistochemistry indicated that EMT markers were decreased in BAMBI-treated tumor tissues, compared with control (Fig. 5D). Collectively, these results indicated that BAMBI may be a potential anticancer agent for osteosarcoma and improve the overall survival rate of xenografted mice compared with control group (Fig. 5E, P<0.05).