The U2OS human osteosarcoma cell line was obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). The malignant transformed hFOB1.19 cell line (MTH) was established by treatment with an initiating agent, N-methyl-N′-nitro-N-nitrosoguanidine (MNNG, 1.0 µg/ml) and a promoter, 12-O-tetradecanoylphorbol-13-acetate (TPA, 200 ng/ml), as previously reported (26). Human umbilical vein endothelial cells (HUVECs) were obtained from Southwest Hospital, Third Military Medical University (TMMU) (Chongqing, China). The MTH and U2OS cells were cultured in Dulbecco's modified Eagle's medium (DMEM; HyClone; GE Healthcare Life Sciences, Logan, UT, USA); HUVECs were cultured in RPMI-1640 medium (HyClone; GE Healthcare Life Sciences) in 5% CO₂ and 95% humidified air at 37°C. Fetal bovine serum (FBS, 10%; Biological Industries, Beit Haemek Ltd., Beit HaEmek, Israel) and 1% penicillin-streptomycin (HyClone; GE Healthcare Life Sciences) were used to supplement the cultures.

To establish the anoikis-resistant model, we cultured the MTH and U2OS cells in ultra-low attachment 6-well plates (Corning Inc., Corning, NY, USA) for 3 days, transferred to normal plates with attachment culture until they formed monolayers, cultured again in ultra-low attachment 6-well plates for 14 days, and finally transferred to normal plates; the re-adherent cells were considered anoikis-resistant cells, and were designated MTHar and U2OSar, respectively (21). An inverted light microscope (Olympus Corp., Tokyo, Japan) was used to observe the cell morphology.

We cultured the MTH, U2OS, MTHar and U2OSar cells in ultra-low attachment 6-well plates for 3 days before detection. Then, 1×10⁶ cells were harvested and an Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) Apoptosis Detection Kit (BD Biosciences, Franklin Lakes, NJ, USA) was used to detect the cells. We analyzed the apoptotic cell percentage using a flow cytometer (BD Biosciences) after staining and repeated the experiment independently 3 times.

Complete medium (DMEM, 10% FBS, 1% penicillin-streptomycin) was used to culture the MTH, U2OS, MTHar and U2OSar cells, or the cells were pretreated with 10 µmol/l Src inhibitor PP2 (Selleckchem, Houston, TX, USA) for 24 h. Phosphate-buffered saline (PBS) was used to wash the cells three times and the cells were transferred to serum-free medium when the cell density was ~70–80%. Then, the CM was collected 24 h after we had changed the medium; we stored the CM at −80°C until use.

We seeded the HUVECs in 6-well plates and cultured in complete medium until they reached approximately 90% confluence. A 10-µl pipette tip was used to create wounds, and the HUVECs were treated with CM. We monitored cell migration to the wound at 0 and 20 h using an inverted light microscope, and quantified it by measuring the distance between the start point of the scratch wound to the migrated point, and the experiment was repeated thrice.

We examined the cell proliferative potential using Cell Counting Kit-8 (CCK-8; Dojindo Laboratories, Kumamoto, Japan). We seeded 1×10³ HUVECs in 96-well plates; the cells were treated with CM. Following 1 to 5 days of culture, we added 10 µl/well CCK-8 reagent to each plate and cultured the cells for 1 h. A microplate reader (Bio-Rad Laboratories, Hercules, CA, USA) was used to measure the absorbance at 450 nm.

We dissolved Matrigel (BD Biosciences) at 4°C overnight. We coated a pre-chilled 96-well plate with thawed Matrigel (50 µl), and incubated it at 37°C for 30 min. After the gel had solidified, 1×10⁴ HUVECs/well were suspended in culture medium comprised of 50% complete medium and 50% CM and seeded in the wells. After 4–12 h of incubation at 37°C, we photographed tube formation at 6 h using an inverted microscope. A capillary-like tube formation assay is often used to determine angiogenesis in vitro. Tube branches and total tube length were measured with ImageJ software (http://rsb.info.nih.gov/ij/). Their number or total length represents the capacity of angiogenesis (27). In this study, the number of tube branches was used for statistical analysis.

We cultured the MTH, U2OS, MTHar, and U2OSar cells in 6-well plates. We changed confluent cells to serum-free medium or pretreated them with PP2 (10 µmol/l). After 24 h, we detected supernatant VEGF-A levels using a human VEGF-A ELISA kit (Neobioscience, Shenzhen, China).

We extracted total RNA from the osteosarcoma cells with TRIzol (Takara Bio Inc., Shiga, Japan). We reverse-transcribed RNA into complementary DNA (cDNA) using a QuantScript kit for reverse transcription (RT) (Takara Bio Inc.). We performed RT-PCR with SYBR Green using primers based on human VEGF-A and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) sequences. The primer sequences were: VEGF-A forward, 5′-CTTCGCTTACTCTCACCTGCTTCTG-3′ and reverse, 5′-GCTGCTTCTTCCAACAATGTGTCTC-3′; GAPDH forward, 5′-CTTTGGTATCGTGGAAGGACTC-3′ and reverse, 5′-GTAGAGGCAGGGATGATGTTCT-3′. The cycling conditions were polymerase activation at 95°C for 30 sec, and then 40 cycles at 95°C for 15 sec and 60°C for 60 sec. We performed triplicate quantitative PCR (qPCR) with StepOnePlus (Applied Biosystems, Foster City, CA, USA). The levels of the experiment group mRNA were calculated with the 2^−ΔΔCq method (28).

We performed western blotting as previously described (21). Proteins were extracted using RIPA buffer (cat. no. 89900; Thermo Fisher Scientific, Inc.) and their concentrations were measured by BCA Protein Assay kit (cat. no. P0010S; Beyotime Institute of Biotechnology, Haimen, China). Proteins (40 µg/lane) were separated by 10% SDS-PAGE (cat. no. P0012A; Beyotime Institute of Biotechnology) and transferred to polyvinylidene difluoride (PVDF) membranes (cat. no. 3010040001; Roche, Shanghai, China). Skim milk (5%) was applied to block the membranes for 1 h at room temperature. Then, the membranes were incubated with primary antibodies overnight at 4°C. After 3 washes with PBST (0.05% Tween-20 in PBS), the blots were incubated with the corresponding secondary antibodies for 1 h at room temperature. We used the following primary antibodies: Rabbit anti-phosphorylated (p)-Src (dilution 1:1,000; cat. no. 2105) and Src (dilution 1:1,000; cat. no. 2109; both from Cell Signaling Technology, Inc., Danvers, MA, USA); mouse anti-VEGF (dilution 1:200; cat. no. sc-7269; Santa Cruz Biotechnology, Santa Cruz, CA, USA); rabbit anti-JNK (dilution 1:1,000; cat. no. AJ518), rabbit anti-ERK (dilution 1:1,000; cat. no. AM076), rabbit anti-p-ERK (dilution 1:1,000; cat. no. AF1891), rabbit anti-p-JNK (dilution 1:1,000; cat. no. AF1762), mouse anti-p38 (dilution 1:1,000; cat. no. AM065) and mouse anti-p-p38 (dilution 1:1,000; cat. no. AM063; all from Beyotime Institute of Biotechnology). The secondary antibodies were horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulin (IgG) (dilution 1:5,000; cat. no. BA1056) and HRP-conjugated goat anti-mouse IgG (dilution 1:5,000; cat. no. BA1050; both from Biodragon Immunotech, Beijing, China). GAPDH (dilution 1:1,000; cat. no. 5174; Cell Signaling Technology, Inc.) was used as a loading control. Finally, the blots were visualized using an ECL kit (cat. no. P0018FFT; BeyoECL Moon; Beyotime Institute of Biotechnology) and quantitative data were obtained using ImageJ software (http://rsb.info.nih.gov/ij/).

We collected specimens from 26 patients (13 males and 13 females, aged from 8 to 49 years old, average age, 19.31±9.30 years) who had been diagnosed with osteosarcoma before radiation therapy or chemotherapy from the Southwest Hospital, TMMU from January 2011 to April 2014 (Table I). We had obtained informed consent previously from the patients or their guardians according to the standards set by the Declaration of Helsinki. The TMMU Institutional Ethical Committee approved the present study.

p-Src, VEGF, CD34 and CD31 expression were determined with a kit (ZSGB-BIO, Beijing, China) as previously described (21). Two pathologists blinded to the pathological and clinical data examined the specimens under a light microscope. We used primary rabbit anti-p-Src (dilution 1:200; cat. no. 2105; Cell Signaling Technology, Inc.), mouse anti-VEGF (dilution 1:50; cat. no. sc-7269; Santa Cruz Biotechnology, Inc.) and mouse anti-CD31 (dilution 1:100; cat. no. ZM-0044; ZSGB-BIO, Beijing, China) for human osteosarcoma specimens. For mouse xenograft specimens, primary rabbit anti-p-Src (dilution 1:200; cat. no. 2105; Cell Signaling Technology, Inc.), mouse anti-VEGF (dilution 1:50; cat. no. sc-7269; Santa Cruz Biotechnology, Inc.) and rabbit anti-CD34 (dilution 1:100; cat. no. Ab81289; Abcam, Cambridge, MA, USA) were used. Semi-quantitative estimation was calculated to interpret the IHC results based on the stained cells per 100 cells (percentage) in >5 randomly selected fields under high-power microscopy (magnification, ×400). We calculated the IHC scores by multiplying the intensity of staining (0, no staining; 1, stained yellow; 2, stained brown; and 3, stained dark brown) and staining extent scores, i.e., percentage of positive tumor cells (0, 0–5%; 1, 6–25%; 2, 26–50%; 3, 51–75%; 4, >75%). Protein expression was categorized using the final IHC scores as follows: 0 (−); 1–3 (+); 4–7 (++); and 8–12 (+++); positive expression was denoted by scores ≥4. We assessed microvessel density (MVD) using a CD31 antibody. We scanned the immunostained sections at a low magnification (×40) to detect hotspots, i.e., areas with the greatest vascular intensity. Then, the number of microvessels were counted in 5 randomly-selected fields per hotspot under a high-power field (magnification, ×200). The average number of microvessels in each sample was determined and was defined as the MVD.

We performed the animal experiments adhering to the Institutional Animal Care and Use Committee-approved protocols at Xinqiao Hospital, TMMU, according to the Directive 2010/63/EU. Eight female 4-week-old nude mice were obtained from the Laboratory Animal Centre of Xinqiao Hospital, TMMU (Chongqing, China). We randomly divided mice into two groups (n=4 mice in each group) and the average weight was ~15 g. Housing conditions included temperature (22±2°C), humidity (40–60%), 12-h light/dark cycle and fed ad libitum. We resuspended MTH/MTHar cells (1×10⁶ for each mouse) in serum-free medium, and the cells were subcutaneously injected into each mouse. The mice were sacrificed by cervical dislocation after 3 weeks and the harvested tumors were fixed with formalin. The maximal volumes of xenografts were 1,058 mm³ (MTH) and 1,573 mm³ (MTHar). p-Src, VEGF-A and CD34 expression were determined by IHC.

We performed all statistical analyses using SPSS 19.0 (version 19.0; IBM Corp., Armonk, NY, USA). We reported the data as the mean ± standard deviation (SD). We analyzed between-group differences with Student's t-test of variance, Spearman's rank correlation coefficients or Pearson's χ² test. P<0.05 was considered to indicate a statistically significant difference.

The anoikis-resistant human osteosarcoma cell model was established by cell suspension culture according to a previous study (21). The osteosarcoma cells were round and gathered into clusters gradually in the suspension culture. The clusters grew bigger and tighter over time (Fig. 1A). The regular polygonal osteosarcoma cells became irregular, elongated and spindle-shaped when they were transferred into normal plates and these cells (the re-adherent cells) were considered anoikis-resistant cells (Fig. 1B). Flow cytometry demonstrated significantly less apoptosis in the anoikis-resistant osteosarcoma cells under suspension conditions compared with the parental osteosarcoma cells (Fig. 1C).

Tumor metastasis requires angiogenesis, a critical factor in controlling cancer progression (27,29). Anoikis-resistant cells play a key role in metastasis; however, the effects of resistance to anoikis in human osteosarcoma cells on angiogenesis remain largely unknown. Angiogenesis includes HUVEC migration, proliferation, and tube formation for the formation of new blood vessels (30). Therefore, we investigated the effects of resistance to anoikis in osteosarcoma cells from humans on HUVEC angiogenesis using in vitro migration, tube formation and proliferation assays. CM from anoikis-resistant osteosarcoma cells promoted HUVEC migration, tube formation and proliferation (Fig. 2A-C). RT-PCR, western blotting and ELISA revealed increased mRNA and protein expression of VEGF-A in the anoikis-resistant osteosarcoma cells (Fig. 2D-F). These data demonstrated that osteosarcoma cells that were resistant to anoikis had increased expression of VEGF-A and angiogenesis.

Src kinase activationis frequently detected in a variety of anoikis-resistant tumor cells as demonstrated in Fig. 3G. Recent research has focused on kinases directly modulating the apoptosis machinery in anoikis resistance. However, the relationship between p-Src and VEGF-A has been poorly explored in anoikis-resistant osteosarcoma cells in humans. To verify whether the expression of VEGF-A involved Src kinase activation in anoikis-resistant osteosarcoma cells, we pretreated anoikis-resistant cells with an Src inhibitor (PP2) for 24 h. PP2 inhibited the expression of p-Src and VEGF-A in the osteosarcoma cells resistant to anoikis (Fig. 3D-G). Compared with the control group, PP2 also reduced HUVEC proliferation, migration and tube formation (Fig. 3A-C). In various cancers, the MAPK (JNK/ERK/p38) signaling pathway is a downstream molecule of Src kinase (24). Accordingly, we investigated whether MAPK (JNK/ERK/P38) signaling pathway molecules were involved in the expression of VEGF-A induced by p-Src in anoikis-resistant osteosarcoma cells. We investigated JNK and ERK phosphorylation in anoikis-resistant cells and p-p38 was not altered; PP2 decreased JNK and ERK phosphorylation in anoikis-resistant cells (Fig. 3G). These results revealed that JNK/ERK may be an Src target downstream in controlling VEGF-A expression. Altogether, our findings indicated that Src kinase activation induced the expression of VEGF-A and angiogenesis via the JNK/ERK pathway activation in anoikis-resistant osteosarcoma cells.

We quantified the expression of P-Src and VEGF-A in osteosarcoma tissues from humans using IHC staining, and the areas of positive expression were mainly localized in the (Fig. 4). To elucidate their clinical significance, we assessed the correlation between P-Src and VEGF-A expression and the patients' available clinicopathological parameters (Table I). Both P-Src and VEGF-A expression was significantly associated with lung metastasis, and there was an association between P-Src expression and Enneking stage, but not with age, sex, tumor size, pathological type and recurrence. The positive rate of P-Src and VEGF-A expression was 50% (13/26) and 61.5% (16/26), respectively, in the osteosarcoma samples (Table I). The relationship between P-Src and VEGF-A expression was calculated and showed that high P-Src expression correlated with expression of VEGF-A (r=0.474; P=0.014; Table II). To assess the P-Src-angiogenesis association, we detected the MVD via CD31 IHC staining in osteosarcoma tissues (Fig. 4). MVD was significantly higher in tumors with positive expression of P-Src (r=0.545; P=0.004; Table II). Taken together, these data demonstrate that P-Src and VEGF-A may be important clinical markers in human osteosarcoma with lung metastasis, and Src kinase may promote angiogenesis via VEGF-A expression in osteosarcoma development.

We found that VEGF-A expression and HUVEC proliferation, migration, and tube formation were enhanced in anoikis-resistant osteosarcoma cells via Src kinase activation in vitro. To verify our experimental results, a xenograft model in mice was used to assess the association between anoikis resistance and angiogenesis. p-Src, VEGF-A and CD34 expression in the xenograft tumors formed in the nude mice were detected using IHC (Fig. 5A). p-Src and VEGF-A expression was higher in anoikis-resistant cells, as was MVD (Fig. 5B). Collectively, these data indicated that anoikis-resistant osteosarcoma cells enhanced osteosarcoma angiogenesis via p-Src and VEGF-A overexpression.