HMJ-30 was designed and synthesized by Dr Mann-Jen Hour (School of Pharmacy, China Medical University, Taichung, Taiwan). Dimethyl sulfoxide (DMSO), potassium phosphate, trypan blue, propidium iodide (PI), Triton X-100, Tris-HCl, MTT and ribonuclease-A were obtained from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany). McCoy's 5a medium, L-glutamine, fetal bovine serum (FBS), trypsin-EDTA, penicillin G and streptomycin were obtained from Gibco; Thermo Fisher Scientific (Waltham, MA, USA). Antibodies against phosphorylated (p-) protein kinase B (AKT) (cat. no. 4060), AKT (cat. no. 4691), p-extracellular signal-regulated kinase (ERK) (cat. no. 4370), ERK (cat. no. 4695), p-c-Jun amino-terminal kinase (JNK) (cat. no. 4668), JNK (cat. no. 9252), p-p38 (cat. no. 4511), p38 (cat. no. 8690), p-IGF-1R β (Tyr1131)/insulin receptor β (IRβ; Tyr1146) (cat. no. 3021), and Ras (cat. no. 3965) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). Antibodies against β-actin (cat. no. sc-47778), α-tubulin (cat. no. sc-5286), IGF-1R (cat. no. sc-9038), IR (cat. no. sc-710), PI3K (p85) (cat. no. sc-1637), IKK (cat. no. sc-7607), nuclear factor-κB (NF-κB; p65) (cat. no. sc-372), p-IκB (cat. no. sc-8404), IκB (cat. no. sc-371), matrix metalloproteinase (MMP)-2 (cat. no. sc-13595), MMP-9 (cat. no. sc-21733), tissue inhibitor of metalloproteinase (TIMP)-1 (cat. no. sc-5538), TIMP-2 (cat. no. sc-5539), and all goat anti-mouse (cat. no. sc-2005) and anti-rabbit (cat. no. sc-2004) immunoglobulin (IgG)-horseradish peroxidase (HRP) secondary antibodies were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Epithelial cadherin (E-cadherin) (cat. no. 07-697), β-catenin (cat. no. 05-613), fibronectin (cat. no. CP70) and vimentin (cat. no. CBL202) primary antibodies were from EMD Millipore (Billerica, MA, USA). Materials and chemicals for electrophoresis were obtained from Bio-Rad Laboratories, Inc. (Hercules, CA, USA).

U-2 OS, a human osteogenic sarcoma cell line, was obtained from the Food Industry Research and Development Institute (Hsinchu, Taiwan). U-2 OS cells were cultured in 75 cm² tissue culture flasks with 90% McCoy's 5a medium containing 1.5 mM L-glutamine adjusted to 1.5 mg/l sodium bicarbonate and supplemented with 10% FBS and 1% penicillin-streptomycin (100 U/ml penicillin G and 100 μg/ml streptomycin) at 37°C under a humidified 5% CO₂ atmosphere, as previously described (33–36).

To evaluate the cytotoxicity of HMJ-30 in osteocarcinoma U-2 OS cells, an MTT assay was performed. Briefly, viable cells were counted on a Neubauer chamber with a microscope. U-2 OS cells were placed in a 96-well cell culture plate at an initial concentration of 1×10⁵ cells/ml, and treated with HMJ-30 at different concentrations (0, 5, 10, 20, and 40 μM) or with 0.1% DMSO for 24 h at 37°C. Following a 24 h incubation period, MTT solution (0.5 mg/ml) was added to each well for 4 h at 37°C. Subsequently, the culture medium was removed, and the formazan crystals formed by oxidation of the MTT solution were dissolved with DMSO in isopropanol and measured spectrophotometrically at 490 nm. The cell survival ratio was expressed as a percentage of the control, as previously described (37–39).

U-2 OS cells were cultured in 24-well plates at a density of 2.5×10⁵ cells/well/ml prior to treatment with 20 μM HMJ-30. Following treatment with HMJ-30 for 24 h at 37°C, morphological changes were determined using a phase-contrast microscope (×200 and ×400) as previously described (6).

To determine cell migration, U-2 OS cells were placed in a 6-well tissue culture plate for 24 h and grown to 80–90% confluence. Individual wells were scratched with a micropipette tip to create a denuded zone of constant width (1 mm). Cells were then cultured in serum-free McCoy's 5a medium and incubated with 0, 5, 10, 20, and 40 μM of HMJ-30 or with 0.1% DMSO for 24 h at 37°C. Cells were photographed under phase-contrast microscopy (×100) as previously described (35,40,41).

Cell invasion was measured using a Matrigel-coated invasion chamber. Initially, a 24-well Transwell insert with 8 μm porosity polycarbonate filters (EMD Millipore) recoated with 30 μg Englebreth-Holm-Swarm sarcoma tumor matrix (EHS Matrigel Basement Membrane Matrix) at 25°C for 1 h to form a genuine reconstituted basement membrane. Cells were then maintained for 24 h in serum-free McCoy's 5a medium. The serum-starved cells were then re-suspended in serum-free medium and placed in the upper chamber of the Transwell insert (5×10⁴ cells/well) and treated with different concentrations of HMJ-30 (0, 5, 10, 20 and 40 μM) or with 0.1% DMSO for 24 h at 37°C. The culture medium supplemented with 10% FBS was added to the lower chamber. The cells were then incubated at 37°C in a humidified atmosphere with 95% air and 5% CO₂ to allow sufficient time for invasion. Following incubation for 24 h, the cells were fixed with 4% formaldehyde for 15 min in phosphate-buffered-saline (PBS) and stained with 2% crystal violet for 5 min at room temperature. Finally, the non-invading cells in the top well were removed with a cotton swab, and the invading cells which penetrated through the Matrigel to the bottom wells were counted under a light microscope (×200) as previously described (6,36,42).

To determine the activity of MMP-2 and MMP-9, gelatin zymography was used. Briefly, U-2 OS cells were treated with HMJ-30 at different concentrations (0, 5, 10, 20 and 40 μM) or with 0.1% DMSO for 24 h at 37°C in the absence of serum. Cells were then collected and separated by dilution in a zymography sample buffer. Samples were mixed with loading buffer and electrophoresed on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel with 0.1% gelatin. Gels were washed twice in denaturing buffer (2.5% Triton X-100 in double-distilled H₂O), and incubated with development buffer (50 mM Tris, pH 7.5, 200 mM NaCl, 5 mM CaCl₂, 1 mM ZnCl₂, 0.02% Briji-35) at 37°C for 18 h, stained with 0.5% Coomassie blue G-250 for 1 h, and de-stained with de-staining solution. Non-staining bands indicating proteolytic activities were digitized using a scanning digital system and analyzed using NIH image software. MMP-2 (72-62 kDa) and MMP-9 (92-85 kDa) activity was expressed as the ratio to MMP-2 standard to avoid differences among gels (43).

The crystal structure of the IGF-1R kinase domain complexed with a benzimidazole (Protein Data Bank code = 2oj9) was retrieved from the Protein Data Bank (http://www.rcsb.org/pdb). Automated docking was then performed. LigandFit, within the software package Discovery Studio 2.5 (Accelrys, San Diego, CA, USA) was used to evaluate and predict the in silico binding free energy of the inhibitors within the macromolecules. First, the prepare protein protocol was used to prepare the 2oj9 protein structure, including standardizing atom names, inserting missing atoms in residues and removing alternate conformations, inserting missing loop regions based on SEQRES data, optimizing short and medium size loop regions with Looper Algorithm, minimizing the remaining loop regions, and calculating pK and protonate structure. A binding pocket of the native benzimidazole ligand was selected as the binding site for the present study. Following typing of the receptor model with the CHARMm forcefield, the binding site was identified by the LigandFit flood-filling algorithm. This docking protocol employed total ligand flexibility, whereby the final ligand conformations were determined by the Monte Carlo conformation search method set to a variable number of trial runs. The docked ligands were further refined using in situ ligand minimization with the Smart Minimizer algorithm. Each minimization was performed in two steps, first using the steepest descent minimization for 200 cycles and then using conjugate gradient minimization, until the average gradient fell below 0.01 kcal/mol. All atoms within 6.0 Å of the inhibitor were allowed to relax during the minimization, whereas those atoms beyond 6.0 Å were held rigid. Finally, the Dock score was used to estimate the binding free energies of the ligands. Among the docked conformations, the pose with the highest value of Dock score was selected for the calculation of binding free energy (∆Gb) and inhibition constant (Ki).

An assay to determine tyrosine kinase activity was performed according to the manufacturer's protocols (Tyrosine kinase assay kit; EMD Millipore). U-2 OS cells were plated in 12-well plates at an initial density of 5.0×10⁶ cells and incubated with 0, 20 or 40 μM of HMJ-30 or with 10 μM picopodophyllotoxin (PPP) (Sigma-Aldrich; Merck KGaA) for 6 h at 37°C. Cells were harvested and total proteins were collected under non-denaturing conditions. Samples were incubated with tyrosine kinase reaction buffer for 30 min at 30°C, followed by the addition of p-tyrosine (4G10) antibodies at a 1:5,000 dilution for 3 min at room temperature, and then detected using an ELISA reader (Anthos 2001) (Anthos Labtec Instruments GmbH, Salzburg, Austria) at a 450 nm wavelength, as previously described (44).

p-IGF-1R β (Tyr1131) (cat. no. 7302), p-p38 MAPK (Thr180/Tyr182) (cat. no. 7946), p-p44/42 MAPK (Thr202/Tyr204) (cat. no. 7177), and p-Akt (Thr308) (cat. no. 7252) assays were performed according to the manufacturer's protocols (PathScan Sandwich ELISA kits; Cell Signaling Technology, Inc.). In total, ~1×10⁷ U-2 OS cells/ml were plated on 24-well plates and treated with 0, 20 or 40 μM of HMJ-30 or with specific protein kinase inhibitors (10 μM PPP, U0126, SB203580 or wortmannin; Sigma-Aldrich; Merck KGaA) for 6 h at 37°C. These cells were harvested and total proteins were collected. Samples were incubated in appropriate antibody-coated microwells for 2 h at 37°C or overnight at 4°C. To each well, 100 μl of the appropriate antibody was added for 1 h at 37°C, and HRP-linked secondary antibody was added for 30 min at 37°C. Absorbance was measured with an ELISA reader (Anthos 2001) at a wavelength of 450 nm as previously described (45).

U-2 OS cells were seeded in a 10-cm dish at an initial concentration of 1.0×10⁷ cells and treated with 20 or 40 μM HMJ-30 or with 0.1% DMSO for 6 h. Cells were then harvested and total proteins were collected using PRO-PREP Protein Extraction Solution (iNtRON Biotechnology, Seongnam, Korea). Samples were incubated with IGF-1R or IR antibodies overnight at 4°C, followed by adding A/G-agarose beads with gentle rocking for 3 h at 4°C. Following centrifugation at 300 x g for 10 min at 4°C, the pellets were washed with 1X cell lysis buffer and re-suspended in 3X SDS sample buffer. These samples were subjected to 10% SDS gel electrophoresis followed by incubation with p-tyrosine (4G10) or p-IGF-1R β (Tyr1131)/IRβ (Tyr1146) antibodies at a 1:1,000 dilution overnight at 4°C to measure specific protein levels, as previously described (45).

U-2 OS cells (5×10⁴ cells/well) plated on a 4-well chamber slide were treated with 20 μM HMJ-30 or with 0.1% DMSO for 24 h. Cells were then fixed in 4% ice formaldehyde for 5 min at room temperature and placed in 1% bovine serum albumin (Sigma-Aldrich; Merck KGaA) containing 0.1% Triton X-100 at 37°C for 30 min. Subsequently, the cells were washed twice with PBS for 5 min and incubated with primary antibodies against E-cadherin (1:250 dilution), β-catenin (1:250 dilution), fibronectin (1:250 dilution), and vimentin (1:250 dilution) overnight at 4°C and then exposed to specific secondary antibodies [fluorescein isothiocyanate-conjugated goat anti-mouse (cat. no. 31569) or anti-rabbit (cat. no. 31635) IgG antibodies; Invitrogen; Thermo Fisher Scientific] at a 1:500 dilution for 1 h at room temperature, followed by DNA staining with PI. Photomicrographs were obtained using a Leica TCS SP2 confocal spectral microscope (Leica Microsystems GmbH, Wetzlar, Germany) as previously described (46,47).

To analyze the expression of proteins associated with invasiveness, western blot analysis was performed. Briefly, the U-2 OS cells were plated in 10 cm-dish at an initial concentration of 1.0×10⁷ cells and treated with HMJ-30 at different concentrations (0, 10, 20, and 40 μM) or with 0.1% DMSO for 6 or 24 h at 37°C. Cells were harvested and total proteins were collected using PRO-PREP Protein Extraction Solution (iNtRON Biotechnology). Protein samples were quantified using the Pierce bicinchoninic acid (BCA) Protein assay kit (Pierce; Thermo Fisher Scientific). Samples (40 μg) were electrophoresed by 10% SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose membranes (Invitrogen; Thermo Fisher Scientific). The membranes were then incubated in blocking solution (0.1% Tween-20 in PBS plus 5% powdered non-fat milk) for 1 h at room temperature, and incubated overnight at 4°C with each of the indicated primary antibodies [MMP-9, MMP-2, TIMP-1, TIMP-2, E-cadherin, β-catenin, vimentin, fibronectin, Ras, p-ERK, ERK, p-JNK, JNK, p-p38, p38, PI3K (p85), p-AKT, AKT, IKK, p-IκB, IκB, NF-κB (p65), β-actin and α-tubulin] (1:1,000 dilution) diluted in blocking solution. Subsequently, the membranes were washed with PBST three times for 10 min and incubated with the appropriate HRP-conjugated secondary IgG antibodies (1:10,000 dilution; HRP-conjugated goat anti-rabbit and goat anti-mouse IgG) for 1 h at room temperature and then washed three times in PBST. Bands were detected using enhanced chemiluminescence with ECL reagents (GE Healthcare Life Sciences, Little Chalfont, UK) and exposed to X-OMAT AR films (Kodak, Rochester, NY, USA). The auto-radiograms were scanned on a UMAX PowerLook Scanner (UMAX Technologies, Fremont, CA, USA) with Photoshop CS5 software (Adobe Systems, Inc., San Jose, CA, USA), as previously described (43,48).

Nuclear proteins were extracted from U-2 OS cells following treatment with 20 μM HMJ-30 or 10 μM ammonium pyrrolidinedithiocarbamate (PDTC; NF-κB inhibitor) (Sigma-Aldrich; Merck KGaA) for 12 h at 37°C using the NE-PER Nuclear and Cytoplasmic Extraction Reagents kit (Pierce; Thermo Fisher Scientific, Inc.). The protein concentrations were determined using the Pierce BCA Protein assay kit (Pierce; Thermo Fisher Scientific), and biotin end-labeled oligonucleotide sequences (5′-Biotin-GATCCAGGGGACTTTCCCTAGC-3′) corresponding to the consensus site of NF-κB were designed. Nuclear extract proteins (5 μg) were used for EMSA with a LightShift Chemiluminescent EMSA kit (Pierce; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Biotin end-labeled duplex DNA was incubated with a nuclear extract or purified factor and electrophoresed on a 6% non-denaturing polyacrylamide native gel. For competition experiments, a 100-fold excess of unlabeled double-stranded oligonucleotide was added to the reaction. Then, the DNA was rapidly transferred to a positively-charged nitrocellulose membrane, UV cross-linked and probed with streptavidin-HRP conjugate (1:300 dilution) with gentle agitation for 15 min at room temperature. Bands were detected by enhanced chemiluminescence with ECL reagents (GE Healthcare Life Sciences) and exposed to X-OMAT AR films (Kodak) as previously described (35,49).

All data were expressed as the mean ± standard deviation from three separate experiments. Statistical analysis was performed using one-way analysis of variance followed by Dunnett's test, with SPSS software version 16.0 (SPSS, Inc., Chicago, IL, USA). ^*P<0.05 was considered to indicate a statistically significant difference.

Our group synthesized a series of quinazoline derivatives targeting tyrosine kinase signaling to inhibit tumor invasion and migration. One of these compounds, HMJ-30 (Fig. 1A), was proposed to actively target IGF-1R. This hypothesis was tested using western blotting, immune-precipitation, kinase activity assays and a computational modeling program. HMJ-30 reduced protein levels of p-IGF-1R (Tyr 1131) and p-tyrosin (4G10) which are determined by immune-precipitation and western blotting (Fig. 1B). Several small molecule inhibitors of IGF-1R tyrosine kinases have been developed and they are currently undergoing clinical investigations (50). However, this approach may be unsatisfactory as co-inhibition of IR is expected to cause hyperglycemia, and this potential effect results from a high protein sequence similarity between the kinase domains of IGF-1R and IR (51). In the present study, p-IR and p-tyrosine (4G10) protein expression levels were evaluated using immune-precipitation and western blotting, and the results revealed that there was no statistical significance between HMJ-30-treated groups and the control (Fig. 1C). The effects of HMJ-30 on p-IGF-1R (Tyr1131) and tyrosine kinase in U-2 OS cells were further investigated. p-IGF-1R (Tyr1131) protein expression was significantly downregulated (Fig. 1D). Tyrosine kinase activity was significantly downregulated, as measured using a kinase assay (Fig. 1E). PPP, which has been reported to act as a non-competitive, potent and specific inhibitor of IGF-1R in vitro and in vivo (52), was used as a positive control, and was also revealed to reduce kinase activity.

In order to predict the major target site of HMJ-30, a docking simulation of HMJ-30 and IGF-1R was performed using the program Discovery Studio Modeling 2.5 (Accelrys). The three-dimensional crystalline structure of the IGF-1R kinase domain in complex with a benzimidazole was downloaded from the RCSB Protein Data Bank website. The computational modeling of the HMJ-30 and IGF1R interaction indicated that HMJ-30 is able to bind to the ATP-binding site of IGF-1R. The interaction between HMJ-30 and IGF-1R is involved in the F-H interaction of 6-fluoro with Ile¹¹³⁰, the F-H interaction of 2-(3-fluoro) phenyl with Met¹⁰⁵² and Ala¹⁰⁰¹, the hydrophobic interactions of the quinazoline ring with Met¹¹²⁶ and Thr^1127, and the hydrophobic interactions of 2-(3-fluoro)phenyl with Met¹⁰⁴⁹ and Glu¹⁰⁵⁰ (Fig. 1F). These interactions made HMJ-30 bind readily to IGF-1R, with low potential energy. Furthermore, the structure of HMJ-30 superimposes well onto the benzimidazole inhibitor (A)-3-[(5-(1H-imidazol-1-yl)-7-methyl-1H-benzimidazol-2-yl)-4-[(pyridin-2-yl-methyl) amino]pyridin-2(1H)-one (Fig. 1F, bottom). Therefore, HMJ-30 may be an IGF-1R inhibitor.

To determine the effect of HMJ-30 on U-2 OS cell invasion, wound and Transwell assays were used. HMJ-30 inhibited cell migration in a concentration-dependent manner (Fig. 2A and B). HMJ-30 also inhibited cell invasion in a concentration-dependent manner (Fig. 2C and D). Tumor cell viability was inhibited, however, only following treatment with 40 μM HMJ-30 for 24 h (Fig. 2E). These data suggested that a concentration of HMJ-30 <40 μM primarily inhibited tumor migration and invasion, but toxicity in U-2 OS cells may require a higher concentration and an increased incubation time.

Previous studies have reported that tumor invasion requires the degradation of basement membranes and proteolysis of the extracellular matrix (ECM) (4). Multiple proteolytic enzymes, including MMPs and serine proteinases, are involved in tumor-host interactions to transmigrate limiting basement membranes and the ECM (53). MMP-2 and MMP-9 serve an important function in cancer cell invasion, and degrade type IV collagen and gelatin which are primarily found in the basal lamina (54). To further evaluate the invasive activity, protein levels of MMP-2 and MMP-9 were determined by gelatin zymography, sandwich ELISA assay and western blotting. HMJ-30 significantly inhibited MMP-2 and MMP-9 activity (Fig. 3A and B) and protein levels (Fig. 3C and D).

The tissue inhibitors of metalloproteinase (TIMP) family, which includes TIMP-1 to TIMP-4, is a group of endogenous MMP inhibitors that serve a central post-translational regulatory role in the majority of known MMPs (55). TIMP-1, the inducible form, has been identified as an inhibitor of the majority of metalloproteinases. TIMP-2, a primarily constitutive protein, has been observed interacting with MMP-1, -2, -9 and -14 (55,56). The present study evaluated the protein levels of TIMP-1 and TIMP-2, which were revealed to be significantly increased in HMJ-30-treated U-2 OS cells (Fig. 3D). These results suggested that HMJ-30 may also be involved in the regulation of MMP-2 and MMP-9 activity.

During the progression and invasion of solid tissue epithelial cancers, tumor cells will lose their epithelial characteristics and gain mesenchymal properties, thus undergoing EMT (23). In contrast, previous studies have reported that the reactivation of epithelial characteristics and the loss of mesenchymal cell markers causes cells to undergo a reverse EMT process, which is also known as MET. The reverse process has been revealed to be associated with the inhibition of invasion and migration in cancer cells (57–59). To evaluate whether or not HMJ-30 induces MET in U-2 OS cells, morphological changes and mesenchymal/epithelial cell markers were investigated. HMJ-30-treated U-2 OS cells acquired a flattened, less elongated and cobblestone-like shape, which is typical of epithelial cell appearance (Fig. 4A). Furthermore, the cells were more aggregated and formed tighter clusters compared with the control group. Expression of vimentin and fibronectin (mesenchymal markers) and of E-cadherin and β-catenin (epithelial markers) were evaluated by immunofluorescent staining and western blotting. HMJ-30-treated U-2 OS cells gained E-cadherin and β-catenin (epithelial markers) and lost vimentin and fibronectin protein (mesenchymal markers) (Fig. 4B and C). These results indicated that MET acts as an important anti-invasive function in HMJ-30-treated U-2 OS cells.

Activated IGF-IR initiates signaling through two primary cascades, the PI3K/AKT and Ras/MAPK pathways (7). These cascades serve critical functions in diverse cellular processes (60). MAPKs are part of a phosphorylation regulatory system composed of three sequentially activated kinases, ERK1/2, p38 kinase and c-Jun amino-terminal kinases (JNKs) (60). To elucidate the potential molecular signaling pathways in HMJ-30-treated U-2 OS cells, the levels of associated proteins in the PI3K/AKT and Ras/MAPK signaling pathways were assessed using western blotting. Key phosphorylated proteins in these signaling pathways were reconfirmed via sandwich ELISA assay. HMJ-30 caused a decrease in the protein levels of Ras, p-ERK1/2, p-p38, and p-JNK in U-2 OS cells (Fig. 5A). Inhibitors of MEK 1 and MEK 2 (U0126) and p38 (SB203580) also reduced the protein levels of p-p44/22 MAPK (ERK) and p-p38, respectively (Fig. 5B and C). HMJ-30 caused a decrease in the protein levels of PI3K (p85) and p-AKT in U-2 OS cells (Fig. 5D). Inhibitors of PI3K (wortmannin) also reduced the protein expression of p-AKT (Fig. 5E).

The PI3K and AKT signaling pathway allows the translocation of NF-κB to the nucleus to regulate gene transcription. The accumulation of IκB, resulting from the decrease of the phosphorylation of IκB by IKKs, prevents the translocation of NF-κB and the activation of transcription by retaining NF-κB in the cytoplasm (61). NF-κB serves an important function in the modulation of the expression of oncogenes, including MMP-2/9, cyclin D1, cell survival proteins and vascular endothelial growth factor; the activities of which are associated with the growth as well as the invasive and metastatic propertoes of cancer cells (61,62). HMJ-30 caused a decrease in protein levels of IKK and NF-κB (p65) and an increase in the protein level of IκB, as confirmed by western blotting (Fig. 5F). HMJ-30 caused a decrease in the protein levels of IKK and NF-κB (p65) and an increase in the protein levels of IκB (Fig. 5F). Furthermore, HMJ-30 caused a decrease in NF-κB activity, and the inhibitor of NF-κB, PDTC, also reduced the activity of NF-κB, as determined by EMSA (Fig. 5G). These results suggested that HMJ-30 inhibited PI3K/AKT, Ras/MAPK, and IκB/NF-κB signaling pathways in U-2 OS cells.