Sunitinib (>99% purity) was purchased from Melonepharma with a Cas no. 341031-54-7. FITC-HM-3 with 98.3% purity was obtained from GL Biochem Ltd. (Shanghai, China) (catalog no. 340664). HM-3 with 99.5% purity was synthesized by GL Biochem Ltd. (catalog no. 140214-2). All laboratory chemicals were of molecular biology grade. Tumor cell lines were obtained from Shanghai Cell Biology Institutes (Shanghai, China).

HCT-116 colorectal cancer cells were cultured in McCoy's 5A medium containing 50 IU/ml penicillin, 100 μg/ml streptomycin, 2,200 μg/ml NaHCO₃ and 10% (v/v) fetal bovine serum (FBS) (Gibco). SMMC-7721 and Hep G-2 hepatocellular carcinoma cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 50 IU/ml penicillin, 100 μg/ml streptomycin, 2,000 μg/ml NaHCO₃ and 10% (v/v) FBS. All cells were incubated at 37°C in a humidified atmosphere of 5% CO₂ (17) while MDA-MB-231 were grown in the absence of CO₂. Cells were detached through incubation with trypsin/EDTA and sub-cultured every 2–5 days (18).

Equal quantities of extracted proteins were loaded on a 10% SDS-polyacrylamide gel and were electrophoretically separated. The separated proteins in the gel were electrically transferred to a polyvinylidene fluoride (PVDF) membrane (Roche), according to the procedure described by Laemmli (19). Membranes were blocked with 5% defatted milk in Tris-buffered saline (TBS) with 0.5% Tween-20 for 30 min. Then, the membrane was incubated with a primary antibody overnight at 4°C. The first antibodies included rabbit monoclonal anti-αv integrin (catalog no. 4711S); rabbit monoclonal anti-β3 integrin (catalog no. 4702); mouse monoclonal anti-β1 integrin (catalog no. 4706S) (Cell Signaling Technology); rabbit polyclonal IgG for α5 integrin (catalog no. 130609) (Santa Cruz Biotechnology). β-actin was used as the internal control. After incubation with the primary antibody, membranes were washed twice for 1 min and once for 10 min with TBS containing 0.5% Tween-20, and were then incubated with goat anti-rabbit IgG with horseradish peroxidase-conjugated (catalog no. 3223449; Biotech, Manufacturing Co. Ltd.) or goat anti-mouse IgG with horseradish peroxidase-conjugated (catalog no. 422331022; Bioss, Beijing, China) at 25°C for 45 min. After incubation with the secondary antibody, membranes were washed three times with TBS with 0.5% Tween-20 for 10 min. Finally, development and detection of bands were carried out by chemiluminescence (20,21). Briefly, the membranes were incubated with enhanced chemiluminescence detection reagents with SuperSignal West Pico Chemiluminescent Substrate (Pierce) for 5 min, and quantification was carried out using ImageJ software (22).

The adhesion of FITC-HM-3 to three tumor cell lines was evaluated by the flow cytometry technique, a method described by Janouskova et al (22) with minor modifications. Briefly, the cultured tumor cells were spread on a 6-well plate and were incubated at 37°C overnight. After digestion and wash with phosphate-buffered saline (PBS), 1 ml PBS (containing 1% BSA) was used to resuspend the cells with a density of 1×10⁶ cells/ml. The cells were incubated with 10 μl of 1 mg/ml FITC-HM-3 at 4°C for 30 min in the dark. Then, the cells were collected by centrifugation at 800 rpm for 5 min. After another wash with PBS, the cells were re-suspended in 800 μl PBS for flow cytometric detection. The mean fluorescence intensity that characterized drug affinity to integrins αvβ3 was measured using the FlowJo software version 7.6.1.

During this assay, cells were placed in the upper layer of a cell permeable membrane in presence of 0.5% BSA and HM-3. The well below the cell permeable membrane was filled with cell medium that contained 5% FBS. Cell migration was performed for 18 h. Then, migrated cells were stained and counted. The main advantage of this assay was the need of low levels of angiogenic inducers. Images were captured and the migration inhibition rate (MIR) was obtained, following the equation: MIR% = (1 − N_test/N_control) × 100%, in which MIR is the migration inhibition rate, N_test the number of cells in test samples and Ncontrol the number of cells in the control sample.

In consideration of the above mentioned results in pre-experiment, in vivo image was carried out to identify drug distribution and tumor targeting effect in mice. Two female BALB/c nude mice (6 weeks and 20 g) were subcutaneously injected with HCT-116 (1×10⁶ cell/ml) (22). Another two mice were tumor-free as controls. After tumor grew to 300 mm³, detection was performed in which mice were intravenously injected with FITC-HM-3 (6 mg/kg) and anesthetized with isoflurane inhaler. Images were captured by Caliper IVIS Spectrum system (Caliper Life Sciences, Waltham, MA, USA) with excitation wavelength of 490 nm and emission wavelength of 520 nm. Images were captured at indicated time points.

Animal experiments were carried out in female BALB/c nude mice (5 weeks and 18–21 g). All mice were subcutaneously injected in the right flank with HCT-116 cells (a density of 1×10⁶ cell/ml and a total volume of 0.2 ml). After tumor volume reached 70–100 mm³, mice were put into different groups, which included the negative control group (n=12) that were intravenously injected with 0.2 ml normal saline every day for 21 days; positive controls (n=6) that were fed by oral gavage with 0.2 ml sunitinib (60 mg/kg) every day for 17 days; mice in the three HM-3 groups (n=6) were intravenously injected with HM-3 (3, 12 or 48 mg/kg) every day for 21 days (Table I). During the treatment period tumor volume and body weight of mice were measured every two days for determination of drug efficacy and signs of toxicity. Animals were sacrificed after 21 days from the first injection and the tumor mass was extracted, dissected, blotted on filter paper and immediately weighed. Immunohistochemistry (IHC) was performed where (CD31, HIF-1α and VEGF) antibodies were used for detection of tumor angiogenesis. Tumors were immediately fixed in 4% formaldehyde and embedded in paraffin, then, sectioned for immunohistochemical staining of CD31, HIF-1α and VEGF with rabbit anti-CD31, HIF-1α and VEGF polyclonal antibodies (Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.). Briefly, staining for the mentioned angiogenesis factors was performed on sections that were incubated with their specific primary antibodies, biotinylated goat anti-rabbit secondary antibodies, horseradish peroxidase-labeled streptavidin, and the results were visualized with diaminobenzidine (DAB) chromogen. Images were captured under a microscope with a magnification of ×200 (23).

The experiments involving animals conformed to the ethical standards of China Pharmaceutical University and the care of animals was in accordance with the guidelines of care and use of laboratory animals of China Pharmaceutical University. In vivo imaging was performed under isoflurane anesthesia, and all efforts were carried out to minimize suffering.

The data were analyzed using the statistics software SPSS statistics 17.0 (Softonic, San Francisco, CA, USA) and are expressed as mean ± SD. Statistical significance was assessed using the Student's t-test. P<0.05 was considered statistically significant; P<0.01 was considered statistically very significant.

The expression levels of integrin subunit αv, β3, α5 and β1 on three human tumor cell lines were detected by western blot assays. As shown in Fig. 1, integrin αv and β1 were expressed on HCT-116, Hep G-2 and SMMC-7721 cells and integrin α5 was expressed on Hep G-2 and SMMC-7721 cells. Integrin β3 was expressed on HCT-116 cells (Fig. 1). Only HCT-116 expressed a substantial amount of integrin αvβ3. Hep G-2 and SMMC-7721 expressed a substantial amount of αv, but trace mount of β3, probably they had integrin αvβ5. All three cell lines obviously expressed integrin subunit β1 whereas only Hep G-2 and SMMC-7721 expressed a substantial amount of α5. HCT-116 did express α5 only slightly, probably it expressed other integrin subunits as β1 subunit can combine with various α subunits. These results were in line with a previous study (24). Similar study was performed to detect the expression levels of integrin subunits on HUVECs and other human tumor cell lines (A549, MCF-7, HeLa, BEL-7402, MGC-803, HT-29, MDA-MB-231 and U87). Various cells (e.g. A549, MCF-7 and HeLa) expressed high levels of integrin αvβ3 and α5β1 whereas HT-29 expressed low levels of integrin αvβ3 and α5β1 (data not shown). This difference in integrin expression levels may influence the antitumor effect of anti-angiogenic reagents that use integrins as targets.

In the flow cytometric assays, the fluorescence of cells in the control group that were treated with free FITC molecules was used to define the 'gate' and 99.7% cells with autofluorescence were included in this gate (Fig. 2A). After FITC-HM-3 incubation, an average of 85.9% of HCT-116 cells bound with FITC-HM-3, the fluorescence shifted rightward and out of the gate (three tests in Fig. 2B–D). HCT-116 cells displayed high adhesion with FITC-HM-3 and this correlated to its high expression of integrins αvβ3 (Fig. 1) as integrin αvβ3 is a target for HM-3. Similar studies were performed for Hep G-2 and SMMC-7721. SMMC-7721 appeared to possess very low adherence with FITC-HM-3 (10.3%) which was 21 times compared to the control (0.5%), whereas Hep G-2 displayed a 1.2% binding with FITC-HM-3 that was approximately four times the control cells (0.3%) (Fig. 2E). As these two cell lines expressed high levels of integrin α5β1, this result confirmed that HM-3 bound with a high affinity to integrin αvβ3, but not to α5β1 (25). The present study demonstrated that the three tumor cells varied in their expression levels of integrin subunits and in their capacity to adhere with FITC-HM-3.

Cell proliferation and migration are both important procedures during angiogenesis. A previous study found that HM-3 did not have inhibitory effect on the proliferation of HUVECs, HCT-116 and Hep G-2 cells (data not shown). The inhibitory effect of HM-3 in the migration of HCT-116 and Hep G-2 cells was evaluated in the present study. The Transwell assay is commonly used to evaluate the migratory response of endothelial cells or cancer cells to angiogenic inducers or inhibitors (26). The regulatory effect of HM-3 at low or high doses on the migration of two different tumor cell lines (HCT-116 and Hep G-2) was investigated. Cells were chosen based on their high and low affinity with integrins in the flow cytometric assays (Fig. 2E). The positive control sunitinib was chosen based on its broad spectrum effect in tumor growth, angiogenesis and metastasis and its inhibition of multiple receptor tyrosine kinases (RTKs) (27). Typical images of migrated HCT-116 cells under different conditions are shown in Fig. 3A–G. HM-3 at a concentration of 2 μg/ml displayed high inhibition of HCT-116 migration with a cell migration number of 159 cells, while the cell migration number in the negative control group was 585 cells (Fig. 3H). The inhibition rates of 2 μg/ml HM-3 in the migration of HCT-116 cells was 72.8% (Fig. 3I). The inhibition rate (MIR) of sunitinib at 0.015 and 2 ng/ml were 35.2 and 47.4%, respectively (Fig. 3I). Furthermore, HM-3 showed lower inhibitory effects in Hep G-2 migration compared with HCT-116 migration at the corresponding concentrations, e.g. the inhibition rate by HM-3 at 2 μg/ml and sunitinib at 2 ng/ml were 21.6 and 7.9%, respectively (Fig. 4). However, both HM-3 and sunitinib showed a similar dose-efficacy relationship in inhibition of HCT-116 and Hep G-2 migration (Figs. 3I and 4B). These cell migration results were in line with the above presented western blot results (Fig. 1) and flow cytometry results (Fig. 2E). The inhibition of cell migration by HM-3 was also performed for HUVECs (16) and HM-3 showed a similar dose-efficacy relationship. At 8 μg/ml, HM-3 inhibited HUVEC migration with an inhibition rate of 67.0% whereas at 32 μg/ml HM-3 promoted HUVEC migration with a promotion rate of 10.0%.

As HCT-116 cells showed high integrin expression, high FITC-HM-3 binding and highly inhibitory effect in cell migration by HM-3, this cell line was selected for further in vivo study.

Being a peptide, HM-3 is prone to be degraded by in vivo proteases (14,28). The in vivo half-life of HM-3 in rat is only 27 min (28). To assess the kinetics of HM-3 in vivo and also the distribution of HM-3 in HCT-116 tumor-bearing and tumor-free mice, FITC-HM-3 was intravenously injected and the fluorescence signal was detected at various time points. Within 10 min of the injections, FITC-HM-3 appeared to be widely distributed in different tissues, including the GIT, breast, limbs, lung and also in the tumor mass (Fig. 5A and B). The signal for FITC-HM-3 in tumor-free mice decreased quickly and at the time point of 20 min, the signal substantially decreased (Fig. 5B). However, at 20 min, the signals in HCT-116 tumor-bearing mice were still high (Fig. 5A). At 28 min, there was still obvious signal for FITC-HM-3 and the signal existed in the position of tumor mass (Fig. 5A). These results confirmed that HM-3 had a short in vivo half-life and also it targeted tumor tissue in vivo.

In vivo dose-efficacy relationship of HM-3 was evaluated on the HCT-116 xenograft model in nude mice. The grouping and drug treatment strategies are shown in Table I. During the drug treatment period, tumor volumes (Fig. 6) and mouse weight in different groups (data not shown) were measured every two days. At the initiation of therapy the tumor volume ranged between 75 and 100 mm³. The positive control sunitinib (60 mg/kg) was daily administered by intra oral gavage for 17 days. Sunitinib at this dose inhibited tumor growth (Fig. 6A) with an inhibition rate of 72.0% on day 21 (Fig. 6B and C). HM-3 at 3 mg/kg also showed inhibition with a tumor volume inhibition rate of 74.0% on day 21 (Fig. 6A) and a tumor mass inhibition rate of 71.5% (Fig. 6B and C). However, HM-3 did not show a dose-dependent inhibition of tumor growth. At a dose of 12 mg/kg, HM-3 showed a tumor volume inhibition of 61.9% and a tumor mass inhibition of 59.2% (Fig. 6). At 48 mg/kg, HM-3 showed a tumor volume inhibition of 17.9% and tumor mass inhibition of 36.0% (Fig. 6). Data in the present study are in accordance with the previous study. In a previous study, on an SMMC-7721 xenograft model in nude mice, HM-3 at a dose of 1.5 or 3 mg/kg showed a high tumor growth inhibition with inhibition rates of 57.4 and 53.0%. However, with increase of HM-3 dosage, the inhibition rate decreased gradually and at 24 mg/kg, the inhibition rate was only 9.5% (16).

HM-3 was not a cytotoxic reagent and during the animal experiment, even the high dose of HM-3 had no sign of toxicity as 100% of the mice survived in the tested groups until sacrificed for tumor extraction. The present study agreed with a previous (20) declaring that HM-3 had no apparent toxic effect on the animals.

IHC analysis of CD31 (Fig. 7), VEGF (Fig. 8) and HIF-1α (Fig. 9) was performed to evaluate the expression levels of these angiogenesis-related factors in HCT-116 tumors. Typical images for each experimental condition are presented as panels A–D in Figs. 7–9. In Fig. 7, microvascular density (MVD) of the tumor tissue from the control group was 62.2 in average whereas those for sunitinib (60 mg/kg) group and HM-3 (3 mg/kg) group were 12.8 and 11.0, demonstrating that these two anti-angiogenic reagents, though with different working mechanisms, can substantially inhibit tumor angiogenesis during tumor growth (Fig. 7E). The MVD of the tumor tissue from HM-3 (48 mg/kg) group was 25.8, showing a decreased anti-angiogenic effect of HM-3 at this dose (Fig. 7E). This is also in line with the tumor growth inhibition experiments (Fig. 6). In Fig. 8, percent of cells expressing VEGF were counted after staining of VEGF-expressing cells immunohistochemically. On average 20.3% of total cells were positively stained on the sections from control group tumors. On the sections from sunitinib (60 mg/kg) and HM-3 (3 mg/kg) groups, the percent of positively stained cells were 29.8 and 28.2. Similarly, 25.1% of cells were positively stained on the sections from HM-3 (48 mg/kg) group (Fig. 8E). This indicates that the anti-angiogenic effect of HM-3 was not due to the deceased levels of growth factors such as VEGF, but to a direct inhibitory effect of HM-3 on HUVEC migration, which is an important procedure during angiogenesis. Similarly, sunitinib inhibited angiogenesis since it can 'cut-off' the signal transduction pathway of VEGFR2, which is the main VEGF receptor during angiogenesis, and so higher VEGF levels in the sunitinib (60 mg/kg) group did not increase angiogenesis. In Fig. 9, percent of cells positively stained with HIF-1α were counted after probing of sections from different groups with anti-HIF-1α antibody and staining. On the sections from control group tumors, an average of 67.9% cells were positively stained whereas for the sections from sunitinib (60 mg/kg) group, HM-3 (3 mg/kg) and HM-3 (48 mg/kg) groups, the percent of positively stained cells were 78.5, 82.3 and 79.1 (Fig. 9E). This result indicated that with the decrease of blood vessel density compared with the control group, the tumors of the drug treatment groups all had a higher degree of hypoxia. The higher hypoxia state made the cells in tumor mass express higher levels of HIF-1α and thereafter higher levels of VEGF.