HSC2 and HSC4 cells were purchased from Cell Bank, RIKEN BioResource Center (Ibaraki, Japan). Both cell lines were exposed to step-wise increasing concentrations of 5-FU (Wako, Osaka, Japan). The cells that survived in culture medium including 5-FU for a year, were cloned by limiting dilution. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA, USA), 100 μg/ml streptomycin, 100 U/ml penicillin (Invitrogen) in a humidified atmosphere containing 5% CO₂.

Cells (5×10³ cells per well) were seeded on 96-well plates (Becton-Dickinson Labware, Franklin lakes, NJ, USA) in DMEM supplemented with 10% FBS. Twenty-four hours later, the cells were treated with 5-FU (0, 2, 4 and 8 μg/ml). After 48 h, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was added to each well (25 μl/well) and incubated for 4 h. The blue dye taken up by cells was dissolved in dimethyl sulfoxide (100 μl/well), and the absorbance was measured with a spectrophotometer (Bio-Rad Laboratories, Hercules, CA, USA) at 490 nm. All assays were run in triplicate.

Cells (5×10³ cells per well) were seeded on micro-cover glass (Matsunami Glass Ind., Osaka, Japan) in DMEM containing 10% FBS. After incubation for 24 h, the culture medium was replaced with DMEM with 10% FBS and 2 μg/ml of 5-FU. After further incubation for 48 h, the cells on the micro-cover glass were washed twice with phosphate-buffered saline (PBS), air dried, and fixed in 4% paraformaldehyde at room temperature for 30 min. TUNEL assay was performed using a DeadEnd™ Colorimetric TUNEL System according to the manufacturer’s instructions (Promega Corporation, Madison, WI, USA). Briefly, the cells on the micro-cover glass were incubated in 20 μg/ml proteinase K for 15 min. Endogenous peroxidase of cells on the micro-cover glass was blocked by incubating in a 3% hydrogen peroxide solution for 5 min after cells were rinsed in distilled water. After being washed with PBS (0.05 M phosphate buffer containing 0.145 M sodium chloride, pH 7.4), the cells were incubated with equilibration buffer and then TdT enzyme in a humidified chamber at 37°C for 60 min. They were subsequently put into pre-warmed working strength stop wash buffer for 10 min. After being rinsed in PBS, the cells were incubated with antidigoxigenin-peroxidase conjugate for 30 min. Peroxidase activity in each cell was demonstrated by the application of diaminobenzidine. Hematoxylin was used as a counterstain. At least 1,000 cells were counted under a microscope in several random fields of each micro-cover glass. The number of apoptotic cell was calculated the number of TUNEL positive cells divided by the total number of counted cells and the result was expressed as a percentage.

In the same manner as above, TUNEL assay was performed in 4-μm-thick paraffin sections of tumor using a DeadEnd Colorimetric TUNEL System according to the manufacturer’s instructions (Promega).

Cells were lysed with RIPA Buffer (Thermo scientific, Rockford, IL, USA). Whole cell lysates were subjected to electrophoresis on 10% SDS-polyacrylamide gels, and then transferred to a PVDF membrane. The membranes were incubated with anti-E-cadherin rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-N-cadherin rabbit monoclonal antibody (Epitomics, Burlingame, CA, USA), anti-Twist mouse monoclonal antibody (Epitomics). The antibody was detected using a chromogenic immunodetection system, WesternBreeze (Invitrogen) according to the manufacturer’s instructions. Also, anti-a-tubulin monoclonal antibody (Santa Cruz Biotechnology) was used for normalization of western blot analysis.

Female athymic nude mice with CAnN. Cg-Foxnlnu/CrlCrlj genetic background (CLEA Japan, Inc., Tokyo, Japan) were purchased at 4 weeks of age and kept under sterile conditions in a pathogen-free environment. The mice were provided with sterile water and food ad libitum and all manipulations were carried out aseptically inside a laminar flow hood. The mice were maintained and were handled in accordance with the Guidelines for Animal Experimentation of Yamaguchi University.

The effect of 5-FU treatment was assessed by inoculation of cells into 5-week-old female athymic nude mice. Cells (1×10⁶) were suspended in 0.1 ml of serum-free medium and injected into the subcutaneous tissue of mice (average weight 15.0 g) using a 27-gauge needle. Tumors at the inoculation site were monitored and measured. When the tumors reached 100–150 mm³ in volume, they were divided into 4 groups, and treated with 5-FU for 3 weeks. Briefly, 5-FU (15 mg/kg) was injected into the peritoneal cavity for 3 weeks (7 times/week). Also, mice in control group were received saline (100 μl) by peritumoral injection. The tumors were measured every two days and the tumor volumes were calculated. At 21 days, mice were sacrificed by cervical dislocation and the tumors were dissected out, fixed in neutral-buffered formalin and embedded in paraffin for further study.

All statistical significance was set at P<0.05. Statistical analyses were performed using StatView software (version 5.0J, SAS Institute Inc., Cary, NC, USA).

Both HSC2 and HSC4 cells were treated with increasing concentrations of 5-FU up to 10 μg/ml to obtain 5-FU-resistant cells. In vitro cell growth inhibition assay was performed occasionally to determine the resistance nature of the cells and HSC2/FU and HSC4/FU were established almost one year later. There were no significant differences in cellular proliferation between HSC2 and HSC2/FU, and between HSC4 and HSC4/FU (Fig. 1).

To compare the growth inhibitory effect of 5-FU between parental and resistant cell lines, we treated the cells with various concentrations of 5-FU. The inhibitory concentration (IC₅₀) data indicated that the 5-FU-resistance levels of HSC2/FU (IC₅₀, 14.0 μg/ml) were 14-fold greater than that of the HSC2 (IC₅₀, 1.0 μg/ml) (Fig. 2A), and the 5-FU-resistance levels of HSC4/FU cells (IC₅₀, 7.0 μg/ml) were 5-fold greater than that of the HSC4 (IC₅₀, 1.4 μg/ml) (Fig. 2B).

TUNEL staining showed a remarkable increase in the number of HSC2 cells exposed to 5-FU that were stained brown, which indicated the occurrence of apoptosis; whereas 5-FU-induced apoptosis was less evident in the HSC2/FU cells. Similarly, 5-FU-induced apoptosis occurred at an increased rate in HSC4 compared to HSC4/FU cells (Fig. 3A). In both cases, the number of apoptotic cells were significantly less in the 5-FU-resistant OSCC cells compared to the parental cell lines (Fig. 3B).

The 5-FU-resistant cells, HSC2/FU and HSC4/FU, were morphologically distinct from their parental cell lines (Fig. 4). The resistant cells had lost the characteristics of cell-cell adhesion, developed spindle-shaped morphology and showed pseudopodia; which are typical phenotypes of EMT.

Unlike epithelial cells that express high levels of E-cadherin, mesenchymal cells express N-cadherin. Our western blot analysis showed decreased E-cadherin and increased N-cadherin and Twist in the HSC2/FU and HSC4/FU cells in vitro compared to the parental cell lines (Fig. 5).

Tumor xenograft studies demonstrated that HSC2/FU and HSC4/FU tumors are resistant to 5-FU when compared to HSC2 and HSC4 tumors (Fig. 6). Briefly, tumor volume of HSC2 and HSC4 was significantly decreased by 5-FU administration though HSC2/FU and HSC4/FU tumors were not affected by 5-FU injection. Moreover, TUNEL assay showed a significant decreased number of apoptotic cells in 5-FU-resistant tumors after treatment with 5-FU (Fig. 7).

We carried out immunohistochemistry experiments to evaluate the expression pattern of EMT-related factors (E-cadherin, N-cadherin and Twist) between 5-FU-sensitive tumors and 5-FU-resistant tumors. The expression of E-cadherin was markedly decreased in 5-FU-resistant tumors, and the expression of N-cadherin and Twist was increased in 5-FU-resistant tumors (Fig. 8).