Human glioblastoma U251MG and U87MG cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) containing 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc.), 100 U/ml penicillin and 100 µg/ml streptomycin (HyClone; GE Healthcare Life Sciences, Logan, UT, USA). The cells were incubated at 37°C in a humidified atmosphere of 95% air and 5% CO₂.

Luteolin and phalloidin were purchased from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany). Insulin-like growth factor-1 (IGF-1) was purchased from PeproTech, Inc. (Rocky Hill, NJ, USA) Anti-phosphorylated (p-)insulin-like growth factor-1 receptor (IGF-1R) (cat. no. sc-81499; 1:500) was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Anti-matrix metalloproteinase (MMP)-2 (cat. no. ab-7033; 1:1,000), anti-MMP-9 (cat. no. ab-76003; 1:1,000), anti-tissue inhibitor of metalloproteinase (TIMP)-1 (cat. no. ab-109125; 1:1,000) and anti-TIMP-2 (cat. no. ab-157386; 1:1,000) were purchased from Abcam (Cambridge, UK). Anti-E-cadherin, anti-N-cadherin, anti-vimentin, anti-β-catenin, anti-vimentin (EMT kit; cat. no. cst-9782; 1:1,000), anti-p-protein kinase B (AKT) (cst-4060; 1:1,000), anti-AKT (cat. no. cst-9272; 1:1,000), anti-p-mammalian target of rapamycin (mTOR) (cat. no. cst-2971; 1:1,000), anti-mTOR (cat. no. cst-2983; 1:1,000), anti-β-actin (cat. no. cst-4970; 1:1,000), and horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulin G (IgG; heavy and light chain; cat. no. cst-7074; 1:5,000) secondary antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). The goat anti-mouse IgG-HRP were purchased from BioworldTechnology (cat. no. BS12478; 1:10,000; St. Louis Park, MN, USA).

Cell proliferation was assessed by Cell Counting Kit-8 (CCK-8) assay (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). Briefly, U251MG and U87MG cells were seeded into 96-well plates at a density of 2×10³ cells per well, and cultured for at least 24 h to adhere. Then, the cells were treated with different concentrations of luteolin (0, 5, 10, 20, 40, 80 and 100 µM). Following treatment for 24 h at 37°C, 10 µl CCK-8 reagent was added to the cells followed by incubation for 2 h at 37°C. Then the absorbance was measured at a wavelength of 450 nm using a Bio-Rad ELISA microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The proliferation rate was calculated as follows: [1-optical density (OD) 450 values of treated groups/OD450 values of control group] × 100%.

This was performed as previously described by Etienne-Manneville (13). U251MG and U87MG cells were seeded on collagen-covered 6-well plates at a density of 0.3×10⁶ (U251MG) and 0.4×10⁶ (U87MG) per plate in DMEM containing 10% FBS. Following incubation for 24 h, the medium was replaced by DMEM containing 0.5% FBS, and the cells were treated for 24 h with 0, 5, 10 or 20 µM luteolin at 37°C. In each plate, three areas were scratched, creating three gaps of similar widths with a 200 µl standard pipette tip. At the indicated time points (0, 6, 12 and 24 h), phase-contrast images of the plates were obtained using a ZEISS inverted microscope (Zeiss GmbH, Jena, Germany; magnification, ×4). The widths of gaps treated with the different concentrations of luteolin and at different time points were measured by Image-Pro Plus software (version 6.0; Media Cybernetics, Inc., Rockville, MD, USA). The widths of the three scratches on each plate were averaged to obtain the mean gap width at a given time. Statistical analysis disclosed the mean gap width (in arbitrary units) of luteolin-treated cells relative to the control (DMSO) at different time points (mean ± standard error of the mean; n=3).

U251MG and U87MG cells were plated at a density of 0.3×10⁶ (U251MG) and 0.4×10⁶ (U87MG) cells in 6-well plates or 35 mm dishes, respectively, and were allowed to grow overnight in DMEM containing 10% FBS. The medium was then replaced by medium without FBS, and the cells were treated for 24 h with different concentrations of luteolin (0, 5, 10 or 20 µM) with or without 100 ng/ml IGF-1. The cells were then lysed with solubilization buffer [50 mmol/l Tris-HCl (pH 7.6), 20 mmol/l MgCl₂, 200 mmol/l NaCl, 0.5% NP40, 1 mmol/l DTT and protease inhibitors] on ice for 10 min, and the lysate (20–100 µg) was subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes (Merck KGaA). Following blocking with 5% nonfat dried milk for 1 h at room temperature, the membrane was incubated with the appropriate primary antibodies overnight at 4°C. Then, the immunoreactive bands were visualized using an Enhanced Chemiluminescence kit (EMD Millipore, Billerica, MA, USA) using horseradish peroxidase-conjugated IgG secondary antibodies at room temperature for 2 h.

U251MG cells were seeded on glass coverslips in DMEM containing 10% FBS, placed in 6-well plates at a density of 1×10⁵ cells per well, for 24 h. The medium was then replaced by DMEM containing 0.5% FBS, and 20 µM luteolin was added. Following further incubation for 24 h at 37°C, the cells were fixed with 4% paraformaldehyde for 30 min at room temperature, and were stained using rhodamine-conjugated phalloidin (1:1,000; Sigma-Aldrich; Merck KGaA) for 30 min at room temperature. Slides were washed with PBS, mounted, imaged with 20 fields and counted using a ZEISS fluorescence microscope (Zeiss GmbH) (magnification, ×40). Cells with stress fibers (mean ± standard error of the mean; n=3) were expressed as a percentage of 100 cells counted from each slide. Cell area (mean ± standard error of the mean; n=3) was measured using ZEN2 software (Zeiss GmbH).

Statistical analysis was performed using SPSS 19.0 software (IBM SPSS, Armonk, NY, USA). Data are presented as the mean ± standard deviation and evaluated using one-way analysis of variance followed with Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

Luteolin demonstrated a clear anti-proliferative effect on U251MG and U87MG cells (Fig. 1). U251MG and U87MG glioblastoma cells were treated with various concentrations of luteolin (0, 5, 10, 20, 40, 80 and 100 µM) for 24 h, and the effect was examined by CCK-8 assay. At concentrations >40 µM, luteolin significantly inhibited the proliferation of U251MG cells, and significantly inhibited the proliferation of U87MG cells at 80 µM. However, at concentrations <20 µM, there was no apparent anti-proliferative effect. Therefore, to exclude the cytotoxic effects of excess luteolin, the following experiments selected a concentration range of luteolin <40 µM to determine the associated effects on glioblastoma cells.

Next, the effect of luteolin on the motility of glioblastoma was assessed. This was performed using a scratch-induced migration assay in which the width of the gap formed by a scratch was monitored at different times following the infliction of the wound (13). The U251MG and U87MG cells were pretreated with luteolin (5, 10 or 20 µM) or with DMSO (as a control) for 24 h, and maintained in DMEM containing 0.5% FBS to block cell proliferation, which would otherwise account for gap closure. The gap width was then monitored at the indicated time points of 0, 6, 12 and 24 h. Luteolin treatment significantly decreased the migration ability of glioblastoma cells in a time- and concentration-dependent manner (Fig. 2). There was no difference in width length at the lowest concentration of 5 µM, whilst at concentrations of 10 and 20 µM the gap closure was attenuated. It appeared to move at a slower pace compared with the control cells, closing a smaller portion of the wound gap. Statistical analysis of the results indicated that luteolin caused a significant decrease in U251MG and U87MG cell migration.

MMPs are gelatinizes which are capable of degrading the extracelluar matrix (ECM) and facilitating the migration and invasion of cancer cells. In contrast, TIMPs are the endogenous inhibitors of MMPs, and prevent the breakdown of the ECM. Therefore, the protein expression of MMPs (MMP-2 and MMP-9) and TIMPs (TIMP-1 and TIMP-2) was analyzed (Fig. 3A). The results revealed that luteolin significantly decreased the expression of MMP-2 and MMP-9 (Fig. 3B and C, respectively), and significantly increased the expression of TIMP-1 and TIMP-2 (Fig. 3D and E, respectively). These results demonstrated that luteolin inhibited the migration of glioblastoma cells partially via downregulation of MMPs and upregulation of TIMPs.

The process of cancer cell invasion is enabled by EMT, which is the initiator of the metastatic cascade (14). Changes in protein expression levels of E-cadherin, N-cadherin, β-catenin and Vimentin, all EMT-associated proteins, were assessed. Following luteolin treatment for 24 h, the protein levels of N-cadherin, β-catenin and Vimentin were visibly decreased; while the protein level of E-cadherin increased (Fig. 4A). A previous study indicated that N-cadherin is upregulated in cancer cells while E-cadherin is downregulated; known as the ‘cadherin switch’ (15). N-cadherin interacts with the fibroblast growth factor receptor, leading to overexpression of MMP-9 and cellular invasion (16). The effects on luteolin on U251MG cell morphology and cytoskeleton organization were then examined using rhodamine-labeled phalloidin, which interacts with polymeric F-actin. The cells were treated with luteolin (20 µM) or with DMSO as a negative control for 24 h under the same conditions as those used in the scratch-induced migration assay, and then imaged. Control U251MG glioblastoma cells were characterized by small cell bodies and long extensions, while the luteolin-treated cells exhibited flat morphology and were visibly larger (Fig. 4B). In addition, the glioblastoma cells had visibly fewer pseudopodia following luteolin treatment compared with the control (Fig. 4B; as indicated by the green arrow). The majority of the polymeric actin in the control cells appeared to be concentrated in membrane ruffles, while in luteolin-treated cells, polymeric actin was organized into stress fibers (Fig. 4B; as indicated by the white arrow). The cell area of U251MG cells was significantly higher following treatment with 20 µM luteolin (Fig. 4C) and the percentage of U251MG cells exhibiting pseudopodia was significantly decreased following treatment with 20 µM luteolin (Fig. 4D).

The PI3K/AKT/mTOR signaling pathway is involved in regulating the migration of cancer cells. Therefore, the effect of luteolin on this pathway was examined. Luteolin treatment visibly decreased the protein levels of p-AKT and p-mTOR in a concentration-dependent manner (Fig. 5A). IGF-1R is a transmembrane heterotetramer with a cytoplasmic tyrosine kinase domain that activates the PI3K/AKT and RAS-rapidly accelerated fibrosarcoma-mitogen-activated protein kinase (MAPK) signaling pathways (17). IGF-1R is overexpressed in multiple types of cancer (18). Previous studies have demonstrated that picropodophyllin, an inhibitor of IGF-1R phosphorylation, inhibits the growth of human glioblastoma cell lines and causes tumor regression not only in subcutaneous xenografts, but also in intracerebral xenografts, along with reduced phosphorylation of IGF-1R and AKT (19). Therefore, the protein levels of p-IGF-1R, which was upstream of PI3K, were assessed. These also decreased following treatment with luteolin for 24 h (Fig. 5A). To further explore the involvement of p-IGF-1R and this pathway in the anti-migratory effect of luteolin, IGF-1 was used to upregulate p-IGF-1R. The U251MG and U87MG cells were serum-starved overnight and then treated for 24 h with 20 µM luteolin prior to 1 h stimulation with 100 ng/ml IGF-1. The results revealed that IGF-1 recovered the level of p-AKT, p-mTOR, MMP-2, MMP-9, N-cadherin and Vimentin, which decreased following treatment of luteolin (Fig. 5B). In addition, the protein levels of TIMP-1, TIMP-1 and E-cadherin also decreased following treatment with IGF-1 (Fig. 5B). These results suggested that luteolin-induced inhibition of migration in glioblastoma cells partially occurred through the p-IGF-1R/PI3K/AKT/mTOR signaling pathway.