The human tongue-derived OSCC cell line HSC-4 and Ca9-22 was purchased from the RIKEN BioResource Center (Ibaraki, Japan). HSC-4 and Ca9-22 cells were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% (v/v) foetal bovine serum (FBS) in a humidified incubator at 37°C and 5% CO₂. DMEM and FBS were purchased from Gibco (Life Technologies, Tokyo, Japan). HGF and curcumin were purchased from Sigma-Aldrich. All antibodies used in this study were commercially available and included antibodies against c-Met and phosphorylated-cMet (phospho-c-Met, Tyr1234/1235) (Cell Signaling Technology), α-tubulin (Sigma-Aldrich), E-cadherin and vimentin (Merck Millipore), and ERK and phospho-Tyr204-ERK (Tyr204) (both from Santa Cruz Biotechnology).

Cell migration was determined using a scratch wound healing assay, as described previously (21,22), except for the following modifications. Briefly, semi-confluent cells in 12-well plates were treated for 4 h with 10 µg/ml mitomycin C to block cell proliferation, and an artificial wound was made by scraping the bottom of the dish. The cells were subsequently wounded with a sterile 200-µl pipette tip to generate a cell-free gap, ~0.3 mm in width. Cells were then washed with phosphate-buffered saline (PBS) and photographed to record the wound width at 0 h (~0.3 mm). Finally, one group of cells was cultured for 48 h in DMEM supplemented with 10% (v/v) FBS as a control, while the other groups were treated with 20 ng/ml HGF (23). For curcumin pre-treatment, cell monolayers were scratched and incubated for 2 h with DMEM supplemented with 0.5% (v/v) FBS and 15 µM curcumin before HGF treatment. The concentration of curcumin was selected based on the study by Tong et al, who reported that the effective concentration of curcumin is 10–20 µM (24). At the end of the incubation, photographs were taken, and the wound width was measured to evaluate cell migration.

In vitro cell invasion assays were performed following the manufacturer's recommendations. HSC-4 cells (3×10⁵ cells/ml) incubated in the presence or absence of HGF were seeded in a 6-well plate fitted with a BioCoat Matrigel Invasion Chamber (Corning; Becton Dickinson). Cells were then cultured for 48 h in DMEM supplemented with 10% (v/v) FBS in the presence or absence of 20 ng/ml HGF and 15 µM curcumin. After 48 h, non-invading cells were removed from the surface of the membrane by scrubbing, and the invading cells were fixed with 100% methanol and then stained using the Diff-Quick staining kit after fixation with methanol. Finally, the number of invading cells in six random fields was counted using a microscope equipped with a ×200 objective and the Image Pro Express software (Meyer Instruments) to evaluate the invasion index.

The gelatinolytic activity of HSC-4 cells was examined by gelatin zymography. Proteins from serum-free conditioned medium were diluted in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and separated using a 12.5% (w/v) polyacrylamide gel containing 0.1% (w/v) gelatin. For each sample, the amount of material loaded onto the gel was corrected for the number of cells in culture and corresponded to the proteins secreted by ten cells. Electrophoresis was carried out for 2 h at 60 mA and 4°C, and the gel was incubated overnight in 50 mM Tris-HCl buffer (pH 7.5) containing 10 mM CaCl₂ and 1X Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific). Finally, the gel was stained using 0.25% Coomassie Blue in a methanol:acetic acid:water (50:10:40) solution and destained using the same solution without dye. Negative staining (white bands against a dark background) indicated proteolysis and, therefore, gelatinolytic activity.

Following treatment, HSC-4 and Ca9-22 cells were collected, washed with PBS, and lysed using RIPA buffer (10 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% (v/v) Nonidet P-40, 0.5% (w/v) deoxycholic acid, 0.1% (w/v) SDS, 5 mM ethylenediaminetetraacetic acid, 1X Halt Protease Inhibitor Cocktail, and 1X Halt Protein Phosphatase Inhibitor (Thermo Fisher Scientific, Inc.). The protein concentration of the cell lysates was determined using a BCA Protein Assay kit (Thermo Fisher Scientific, Inc.), and equal amounts of protein were subjected to SDS-PAGE. The separated proteins were transferred onto a PVDF membrane (GE Healthcare), which was then blocked for 1 h at room temperature with 5% (w/v) bovine serum albumin in Tris-buffered saline (TBS)/Tween-20 (TBS-T) to prevent non-specific binding. The membrane was incubated overnight at 4°C with antibodies diluted in TBS-T, washed, and incubated with HRP-conjugated secondary antibodies diluted in TBS-T. Finally, antibody-antigen complexes were detected using the ECL Plus Western Blotting Detection Reagent (GE Healthcare).

All data are presented as the mean ± standard deviation from three independent experiments unless stated otherwise. The statistical significance of a difference between two groups was evaluated using unpaired Student's t-test. P<0.05 was considered to indicate a statistically significant difference.

To assess the effect of curcumin on oral cancer cells, we first examined HGF-induced cell motility of HSC-4 cells using an invasion assay (Fig. 1). HGF-induced cells exhibited a 2-fold increase in the number of invasive cells compared to control cells (P<0.05). In stark contrast, curcumin pre-treatment significantly inhibited HGF-induced invasion of HSC-4 cells (P<0.05), with a number of invasive cells that was comparable to that of the control cells.

To further assess the effect of curcumin on HGF-induced cell motility, we also examined cell migration using a wound healing assay (Fig. 2). In agreement with our cell migration data, the migration of HGF-induced HSC-4 cells was increased >2-fold compared to control cells (P<0.001). However, as in the case of cell invasion, curcumin pre-treatment dramatically reduced the migration of HGF-induced cells (P<0.05), which was similar to that of control cells. Collectively, these findings strongly suggested that curcumin pre-treatment could inhibit the HGF-induced motility of HSC-4 cells, resulting in reduced cell invasion and migration.

The concomitant down-regulation of the epithelial marker E-cadherin and up-regulation of the mesenchymal marker vimentin is recognised as a hallmark of cells undergoing EMT (25,26) and results in the loss of cell polarity, an important step in EMT induction (27,28). Therefore, we next performed western blot analyses to assess the expression level of these markers and determine whether the observed inhibitory effects of curcumin on HGF-induced cell motility involved alterations in the EMT process (Fig. 3). As expected, E-cadherin and vimentin expression levels were decreased and increased, respectively, in response to HGF stimulation, which confirmed that HGF-induced HSC-4 and Ca9-22 cells were undergoing EMT. Remarkably, curcumin pre-treatment abrogated HGF-induced changes in E-cadherin and vimentin expression, which strongly suggested that curcumin could inhibit HGF-induced EMT in oral cancer cells.

As mentioned previously, HGF signalling plays an important role in EMT induction. A crucial step is the homodimerisation and autophosphorylation of the c-Met receptor tyrosine kinase upon HGF binding (29,30). Therefore, we performed western blot analyses to examine the effect of curcumin pre-treatment on the level of phospho-c-Met. HGF-induced HSC-4 and Ca9-22 cells exhibited an increased level of phospho-c-Met compared to control cells. Remarkably, curcumin pre-treatment abolished the increase in phospho-c-Met level resulting from HGF stimulation (Fig. 4), which strongly suggested that the inhibitory effects of curcumin on HGF-induced EMT involve the down-regulation of HGF signalling.

The role of HGF signalling in the induction of EMT is mediated via the downstream activation of the AKT and ERK effector pathways (31). Accordingly, we assessed the activation status of the ERK pathway to further dissect the mechanisms of EMT inhibition by curcumin. Western blot analyses showed that ERK phosphorylation, a marker of ERK activation, was increased in HGF-induced HSC-4 cells compared to control cells (Fig. 5). In stark contrast, curcumin pre-treatment completely blocked HGF-induced ERK phosphorylation. Collectively, these findings indicated that curcumin could inhibit HGF-induced EMT by repressing c-Met and ERK activation.

Matrix metalloproteinases (MMPs) are major regulators of the extracellular matrix and are known to play important roles in tumour invasion and metastasis (32). Furthermore, activation of the MMP2 and MMP9 gelatinases has been associated with an induction of the ERK pathway (33). Therefore, we used gelatin zymography to evaluate gelatinolytic activity in the conditioned medium of HSC-4 cells (Fig. 6). While control cells exhibited two major types of gelatinolytic activity, which were consistent with pro-MMP9 and pro-MMP2, HGF-induced cells exhibited a marked increase in pro-MMP9 production. Notably, curcumin pre-treatment inhibited the production of pro-MMP9 by HGF-induced cells, whereas the production of pro-MMP2 remained unaffected. This observation suggested that inhibition of HGF signalling by curcumin might decrease cell motility by repressing the production of gelatinolytic activity.