HPLC grade MB (Fig. 1) was purchased from AppliChem GmbH (Cat. No. A9007; St. Louis, MO, USA). A stock solution (1 mM) was prepared by dissolving MB in dimethyl sulfoxide (DMSO). Recombinant human VEGF165 (Cat. No. 293-VE) is a product of R&D Systems (Minneapolis, MN, USA). L-sulforaphane (SFP; Cat. No. S6317; Sigma-Aldrich, Beijing, China), a potent inducer of inhibition of angiogenesis, and batimastat (Bat; Cat. No. SC-203833; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), a broad spectrum MMP inhibitor, were used in the present study and served as positive controls.

Human umbilical vein endothelial cells (HUVECs) were isolated as previously described (17) from umbilical cords obtained from a parturient at the Department of Maternity at the First Affiliated Hospital of Nanchang University, who provided written informed consent. The study protocol was approved by the Institutional Review Board (IRB) of Nanchang University. HUVECs were routinely grown in M199 supplemented with 10% fetal bovine serum (FBS) (both from Gibco, Grand Island, NY, USA) and EC growth supplement (BD Biosciences, Bedford, MA, USA) at 37°C and 5% CO₂. HUVECs between P3 and P4 were used for all experiments. The human umbilical vein cell line, EA.hy 926, was purchased from the cell bank of Shanghai Institute for Biological Sciences, CAS. EA.hy 926 cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, 100 IU/ml penicillin and 100 μg/ml streptomycin (Invitrogen, Beijing, China) in a humidified incubator containing 5% CO₂ at 37°C. Prior to performing each assay, the ECs were serum starved for 4 h.

Five to six-week-old male Sprague Dawley (SD) rats were purchased from the Vital River Laboratories (Beijing, China) to obtain thoracic aorta in the rat aortic ring angiogenesis assay. All animal experiments were conducted in strict accordance with full ethical approval of the Animal Ethics Committees of Jiangxi University of Traditional Chinese Medicines. All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.

The tube formation assay was used to investigate the effects of MB on angiogenesis in vitro. Briefly, 80 μl of liquid growth factor-reduced Matrigel (BD Biosciences, San Jose, CA, USA) were added to each well of a 96-well plate. After 45 min incubation at 37°C, 3×10⁴ ECs/well in 100 μl complete culture medium containing vehicle, MB or SFP was seeded in each well. Then, 100 μl serum-free medium containing VEGF (final concentration 2 ng/ml) was added. After 24 h incubation at 37°C and 5% CO₂, the images of each well were recorded by an inverted microscope (Leica DMI3000B, Germany).

Cell proliferation was assayed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Briefly, log-phase cells were seeded in 96-well plates 24 h before initiation of treatment with MB in the presence of 2 ng/ml VEGF. The vehicle-treated cells served as controls. Three duplicate wells were set up in each sample. After incubation with a range of MB or vehicle for 24 h, cells were incubated with MTT (final concentration 0.5 mg/ml) for 4 h at 37°C (Sigma-Aldrich). The media was carefully removed from each well and 150 μl of DMSO was added. The plates were gently agitated and optical absorbance at OD 570 nm and OD 450 nm was determined using a microplate reader (ELx800; BioTek, Winooski, VT, USA). Vehicle only-treated cells served as the indicator of 100% cell viability. Each experiment was repeated three times.

In the present study, an ex vivo tube formation system was used as previously described (18) to evaluate the anti-angiogenic effect of MB. Male SD rats (5–6 weeks old) were euthanized and thoracic aortas were retrieved. After rinsing with 1% antibiotic/antimycotic cocktail in 1X PBS (100 U/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B; Invitrogen), the surrounding fibroadipose tissue of thoracic aorta was completely removed with fine microdissection scissors, and the thoracic aorta was cut into 1 mm rings with a scalpel blade. Then, all the rings were implanted on a Matrigel-coated 96-well microtiter plate. Matrigel was added again to embed and fix rings. After 30 min incubation in 5% CO₂ at 37°C, the aorta rings were incubated in human endothelial serum-free medium (Gibco, Carlsbad, CA, USA) supplemented with 2% FBS, 50 U/ml penicillin and 50 μg/ml streptomycin (Invitrogen) for 24 h and then treated with different doses of MB for 13 days. Finally, 8 μg/ml Calcein AM (BD Biosciences) was added to stain microvessels. Images of the microvessels were obtained using an inverted fluorescence microscope (Leica DMI3000B).

Effects of MB on invasion of ECs were measured by a 48-well microchemotaxis system (AP48; Neuro Probe, Gaithersburg, MD, USA). Briefly, 5 μg of fibronectin in a volume of 50 μl was applied on the rough (lower) surface of the polycarbonate membrane, and 5 μg/filter Matrigel was plated to the smooth (upper) surface. The lower chambers of the plates were then filled with 30 μl medium containing 0.1% BSA. Log-phase cells were harvested and re-suspended in culture medium with 0.1% BSA. Cell suspensions (100 μl containing 1×10⁵ cells) and 2 ng/ml VEGF were added to the upper compartment and incubated for 16 h at 37°C in a 5% CO₂ atmosphere. After incubation, the filters were fixed with methanol and stained with 0.5% crystal violet for 60 min. The cells on the upper surface of the filters were removed by wiping with cotton swabs. The cells invading to the lower surface of the filter through Matrigel and filter were quantified with the Image-Pro Plus software 5.0 (Media Cybernetics Inc., Silver Spring, MD, USA) and the most representative results are illustrated in the figures. Each assay was performed in triplicate.

MMP activity was analyzed with gelatin zymograms. Following 24 h treatment with MB, Bat or vehicle in the presence of 2 ng/ml VEGF, ECs were harvested and lysed. Cell lysates were diluted in 50 mM Tris-HCl, pH 7.4, and separated by electrophoresis in 7.5% sodium dodecyl sulfate polyacrylamide gels containing 2 mg/ml gelatin. After electrophoresis, the gels were washed in 2.5% Triton X-100 for 30 min and then incubated for 16 h at 37°C in 50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 10 mM CaCl₂. Gels were then stained with Coomassie brilliant blue R-250 and destained with 40% methanol-10% acetic acid until clear bands appeared. Images were captured with the UVP EC3 gel imaging system.

MMP activities were also measured by the SensoLyte^® 570 Generic MMP Assay kit (AnaSpec, Fremont, CA, USA). This kit uses a fluorescence resonance energy transfer (FRET)-based method to detect the activities of a variety of MMPs including MMP-1, -2, -3, -7, -8, -9, -10, -11, -12, -13 and -14. It uses 5-FAM (fluorophore) and QXL520™ (quencher) labeled FRET peptide substrates for continuous measurement of MMP activities. In an intact FRET peptide, the fluorescence of 5-FAM is quenched by SensoLyte^®. Upon cleavage of FRET peptide by MMPs, the fluorescence of 5-FAM is recovered. Analyses were performed according to the manufacturer’s instructions. Briefly, supernatants of ECs were collected after incubation with or without MB for 24 h. MMPs in supernatants were activated by incubation with 4-aminophenylmercuric acetate for 1 h at 37°C. Fifty microliter MMP-containing samples and 50 μl MMP substrate solution were added into a 96-well plate. The reagents were mixed by shaking the plate gently for 30 sec. After a 50-min incubation period at 37°C, the reaction was stopped by adding stop solution, and fluorescence intensity was measured by a multilabel counter (Victor3™; Perkin-Elmer, Waltham, MA, USA) at Ex/Em = 540/575 nm.

Cells were rinsed twice with PBS, and total proteins were extracted in 500 μl lysis buffer. Aliquots of whole cell lysates were separated by 10% SDS-PAGE and then transferred to Hybond nitro blotting membranes. The membranes were blocked with 3% BSA in Tris-buffered saline containing 0.5 ml/l Tween-20 and then incubated with primary antibodies against MMP-9 (SC-6840), RUNX2 (SC-10758) (both from Santa Cruz), and phospho-RUNX2 (AP3559a; Abgent), followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies. Immunoreactive proteins were detected using an enhanced chemiluminescence kit (Millipore). β-actin (SC-130301; Santa Cruz) served as an internal control.

RUNX2 activity in EC nuclear extracts was detected using the TransAM™ AML-3/RUNX2 kit (Active Motif North America, Carlsbad, CA, USA) following the manufacturer’s instructions. Cell extracts were prepared using the Nuclear Extract kit (Active Motif) with ECs that were treated with either MB or DMSO for 24 h. Then, 20 μl of extracts diluted in complete lysis buffer and containing 15 μg nuclear extract were added into a 96-well plate. This plate immobilizes oligonucleotides containing RUNX2 consensus binding sites. Saos-2 nuclear extract served as a positive control for RUNX2 activation, and 20 μl complete lysis buffer served as the blank. The wild-type consensus oligonucleotide was provided as a competitor for RUNX2 binding to monitor the specificity of the assay. After 1 h incubation at room temperature, the plate was washed three times with washing buffer. Diluted primary antibody (100 μl) was added into wells and incubated for 1 h at room temperature without agitation. After three washes, HRP-labeled secondary antibody was added and incubated for 1 h at room temperature. Then, 100 μl developing solution was added to initiate the color reaction. After 100 μl stop solution was added, the absorbance was measured within 5 min at 450 nm with a reference wavelength of 655 nm using an ELx800 microplate reader (BioTek).

Chromatin immunoprecipitation assay (ChIP) was performed using the agarose ChIP kit as described in the manufacturer’s instructions (Pierce™ Agarose ChIP kit; Thermo Scientific). Briefly, crosslinking was performed for 10 min using formaldehyde (final concentration 1%). Crosslinked cells then were lysed with lysis buffer containing protease inhibitors. To obtain DNA fragments with an average size of 0.3 kb, micrococcal nuclease was added and samples were incubated for 5 min. Protein-DNA complexes were immunoprecipitated using anti-RUNX2 antibody (SC-10758; Santa Cruz). Normal rabbit IgG served as a negative control and anti-RNA polymerase II antibody served as positive control. After DNA recovery, purified DNA was subjected to PCR amplification on the Mastercycler gradient (Eppendorf) using the Phusion^® High-Fidelity PCR kit. The primers designed to amplify one of the RUNX2 binding regions (−220 bp; TGGGGTC) in the human mmp-9 promoter (GeneBank no. NM_004994) were: forward, 5′-ACA GTT CCC ACA AGC TCT GC-3′ and reverse, 5′-CAG CAT GAG AAA GGG CTT ACA-3′. The PCR profile was: 15 min at 95°C, followed by 30 three step cycles of 15 sec at 95°C, 30 sec at 58°C and 30 sec at 72°C. PCR reactions were subjected to a final extension at 72°C for 10 min. PCR analysis was performed using the Mastercycler gradient. Aliquots of total DNA before immunoprecipitation were saved as input, and these input lysates were also processed as above. All ChIP assays were performed at least three times, and the most representative results are illustrated in the figures.

The data are presented as means ± SD and were analyzed with SPSS for Windows (13.0) software program (Chicago, IL, USA). Comparison among different groups was carried out by one-way analysis of variance (the one-way ANOVA). Differences between means were considered statistically significant at P<0.05.

We first used an in vitro tube formation assay to assess the effects of MB on tumor-induced angiogenesis. As shown in Fig. 2A, when EA.hy 926 cells were seeded on Matrigel, capillary-like structures with a lumen were formed. After exposure to MB solution at various concentrations for 24 h, these structures were destroyed in a dose-dependent manner. MB treatment of HUVECs, a primary EC line derived from the endothelium of umbilical vein, showed that normal capillary-like structures were also significantly impaired by MB (Fig. 2B). Fig. 2C and D shows the growth inhibition effects of MB on EA.hy 926 and HUVECs; the IC₅₀ values of MB were 92.35±13.47 and 100.99±27.71 μM and the inhibitory rates of 40 μM MB were 17.06 and 3.24%, respectively. These data indicate that the observed ability of MB to suppress tube formation was not due to growth inhibition.

Next, the rat aortic ring angiogenesis assay was employed to verify the anti-angiogenic effects of MB. New blood vessels, triggered by the injury of the dissection procedure and mediated by growth factors produced from the aortic ring, were demolished by MB in a dose-dependent manner (Fig. 3).

Invasion of ECs through the basement membrane to form sprouting vessels is essential to angiogenesis. Thus, we next examined the anti-invasive effects of MB on ECs. As shown in Fig. 4A, MB significantly blocked the transmembrane invasion of EA.hy 926 cells. The invasion rate across the reconstituted basement membrane was 79.43%, 64.60% and 38.49% when the cells were incubated with 10, 20 and 40 μM MB for 14 h, respectively, compared with controls. Similar results were obtained with HUVECs (Fig. 4B). Collectively, these results demonstrated that MB significantly inhibited the invasion through the basement membrane of ECs, and this effect may contribute to the anti-angiogenic properties of MB.

Breakdown of the extracellular matrix by MMPs in surrounding tissues is a fundamental step of EC invasion in tumor-induced angiogenesis. Therefore, we examined the effects of MB on MMPs in ECs. Fig. 5 shows the effects of MB on MMPs in EA.hy 926 and HUVECs. Data from the zymography analysis showed that MB significantly inhibited hydrolyzation of gelatin in both EC lines (Fig. 5A). Using a FRET-based analysis, we found that MB exhibited significant suppression of FRET substrate cleavage of MMPs in a dose-dependent manner (Fig. 5B). To further determine whether MB inhibited the functional activity or expression of MMPs in ECs, we analyzed the expression of MMP-9, one of the important MMPs, in EA.hy 926 and HUVECs treated with MB. As shown in Fig. 5C, MB significantly decreased MMP-9 expression in both EC lines in a dose-dependent manner. These data suggested that the anti-angiogenic properties of MB may be due to downregulation of expression of MMP-9 in ECs.

From current observations, we understood that Runx2 is a ‘master’ transcriptional factor of metastatic growth of breast cancer cells. Several genes required for the formation of metastatic foci, including mmp-9, bsp, opn and vegf, are targets of this transcriptional factor. Therefore, next we examined RUNX2 activities in ECs when cells were co-incubated with VEGF at 2 ng/ml. As shown in Fig. 6A, MB had no effect on total RUNX2 expression, but caused a significant decrease of phospho-RUNX2 expression. These results indicate that MB-induced downregulation of MMP-9 expression may be associated with the suppression of RUNX2 activation.

To confirm the western blot results, Runx2 transcription factor assay and ChIP assay were designed. First, an enzyme-linked immunoabsorbent assay (ELISA)-based kit for the Runx2 transcription factor was used to analyze the effects of MB on Runx2. EC nuclear extracts incubated with MB or vehicle were prepared and the binding activity between RUNX2 with its target sequence was determined. The results showed that MB decreased the binding activity of RUNX2 to its target sequences in a dose-dependent manner (Fig. 6B). Using ChIP assay, we found that binding of RUNX2 to one of the RUNX2 binding domains (−220 bp; TGGGGTC) in the mmp-9 promoter region was inhibited by MB (Fig. 6C). Collectively, these findings suggested that the inhibitory effect of MB on MMP-9 was caused by MB inhibition of RUNX2 binding to the mmp-9 promoter, which may be associated with suppression of RUNX2 phosphorylation.