Sixty-seven biopsy specimens of primary OSCC were obtained from patients undergoing surgical resection at the Department of Oral and Maxillofacial Surgery, Kanazawa University Hospital between 1989 and 2009. The patients (38 male and 29 female subjects) ranged in age from 32 to 91 years (mean age: 60 years). Informed consent for experimental use of the samples was obtained from the patients according to the hospital’s ethical guidelines.

Immunohistochemical staining was performed by the labeled streptavidin-biotin (LSAB) method after deparaffinization and rehydration as described by Nozaki et al(25). The sections were reacted with the following primary antibodies: anti-claudin-7 monoclonal antibody (Invitrogen Corp., Camarillo, CA, USA) diluted 200-fold with PBS at 4°C overnight. Sections were then reacted with a secondary antibody, biotin-labeled goat anti-rabbit immunoglobulin polyclonal antibody (Dako Japan, Kyoto, Japan) at RT for 60 min. A section of normal oral epithelium previously identified as strong staining was used as a positive control with each batch. As a negative control, PBS treated sections instead of claudin-7 antibody was used.

All cell lines were maintained at 37°C in a humidified incubator containing 5% CO₂. The OSCC cell lines HSC-4, OSC-20, OSC-19, OTC-04, HOC313 and TSU were maintained in minimal essential medium (MEM; Sigma-Aldrich, Ayrshire, UK) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. The cell lines were derived from OSCC with the following grades of invasiveness, according to the Yamamoto-Kohama criteria (20): HSC-4 and OSC-20 cells from grade 3 as described for the low invasive type; OSC-19 and OTC-04 from grade 4C, as described for the mild invasive type; HOC313 and TSU from 4D as described for the high invasive types.

RT-PCR analysis was performed using a modified method by Conboy et al(26). RNA was extracted from cultured cells using an RNeasy kit (Qiagen, Hilden, Germany). A 1-μg sample in 10 μl of RNase free water was incubated for 5 min at 60°C and then quickly chilled on ice for 5 min. The RNA samples were reversed-transcribed into first-strand cDNA at 40°C for 40 min in RT solution from the RNeasy kit. The cDNA samples were amplified following addition of the PCR mixture solution and the following primers for claudin-7, 5′-aat gta cga ctc ggt gct cg-3′ (forward) and 5′-att ccc agg aca gga aca gg-3′ (reverse); for E-cadherin, 5′-agc cat ggg ccc ttg gag-3′ (forward) and 5′-cca gag gct ctg tca cct tc-3′ (reverse); for Snail, 5′-acc act atg ccg cgc tct ttc ctc g-3′ (forward) and 5′-gac agg aga agg gct tct cgc cag t-3′ (reverse) and for β-actin, 5′-gaa aat ctg gca cca cac ctt-3′ (forward) and 5′-ttg aag gta gtt tcg tgg at-3′ (reverse). PCRs were carried out under the following conditions: 3 min at 94°C, followed by cycles (30 for claudin-7, 30 for E-cadherin, 30 for Snail and 20 for β-actin) of 1 min at 94°C, 1 min at 58°C, and 1 min at 72°C. All reactions were completed with a final incubation at 72°C for 10 min. The lengths for the amplified fragments for claudin-7, E-cadherin, Snail and β-actin genes were 288, 544, 637 and 592 bp, respectively. PCR products were detected by 3.0% agarose gel electrophoresis and staining with ethidium bromide.

Cultured cells on 80% confluent plates were used for protein samples. The protein samples (30 μg) that were extracted from the whole cellular structure using M-PER (Mammalian protein extraction reagent) (Pierce, Rockford, IL, USA) were heated at 95°C for 5 min before electrophoresis and then subjected to 10% SDS-PAGE. After electrophoresis, the samples were transferred onto PVDF membranes (ATTO Co., Tokyo, Japan) and incubated for 1 h with 200-fold diluted polyclonal anti-rabbit antibody against claudin-7 (Invitrogen, Carlsbad, CA, USA), a 2,000-fold diluted polyclonal anti-mouse antibody against E-cadherin (BD Biosciences, San Jose, CA, USA), a polyclonal anti-rabbit antibody against Snail (Abgent, San-Diego, CA, USA) and 5000-fold diluted polyclonal anti-mouse antibody β-actin (Sigma, St. Louis, MO, USA) respectively. The membrane was washed three times with PBS and then incubated for 1 h with 2000-fold diluted horseradish peroxidase-conjugated anti-rabbit IgG (Amersham, Buckinghamshire, UK) to detect claudin-7 and Snail, 2000-fold diluted horseradish peroxidase-conjugated anti-mouse IgG (Amersham) to detect E-cadherin and β-actin, respectively. The blots were revealed by enhanced chemiluminescent detection carried out according to the manufacturer’s recommendations.

Statistical analysis was performed with the SPSS for window version 16.0 (SPSS Inc., Chicago, IL, USA). The expression of claudin-7 in tumor cells was evaluated as present or absent. Only cases in which at least 25% of the tumor cells were immunoreactive were scored as positive. The Mann-Whitney’s U test and χ² test were used to analyze the association of claudin-7 expression with clinicopathological factors. Survival rates of claudin-7 -positive and -negative patients were calculated by the Kaplan-Meyer method, and examined for statistical significance using the log-rank test. Differences were considered significant at p-values of <0.05. Uni- and multi-variate analyses for the 5-year overall survival of individual parameters were performed.

The relationship between the clinicopathological parameters and expression of claudin-7 is summarized in Table II. Claudin-7 immunostaining was observed especially in membranes including the cytoplasm and nucleus of tumor cells (Fig. 1). Immunohistochemical staining showed that 35 specimens (52.2%) were positive for claudin-7. There was a significant negative correlation between claudin-7 and T-category (p<0.05), Lymph node metastasis (p<0.01) and mode of invasion (p<0.001); the number of claudin-7 positive cases were 11 (91.7%) for grade 1, 12 (75.0%) for grade 2, 9 (52.9%) for grade 3, 3 (18.8%) for grade 4C, and 0 (0%) for grade 4D. Moreover, the number of claudin-7-positive cases was 5 (26.3%) with lymph node metastasis and 30 (62.5%) without lymph node metastasis. Therefore, claudin-7 expression showed a negative correlation with lymph node metastasis (p<0.01). The overall 5 year-survival rate was 77.1% in patients showing claudin-7 expression and 59.4% in patients without claudin-7 expression (P<0.05) (Fig. 2). Table III summarizes the univariate and multivariate analyses of the correlation between the clinicopathological and immunohistochemical variables, with respect to overall survival. Multivariate analysis revealed that only the 3-4D mode of invasion was significant and independent variables with relative risks of 7.44, although univariate analysis revealed that T-category, N-category, Cell differentiation, mode of invasion, expression of claudin-7 were also significant variables.

We further examined the levels of claudin-7 mRNAs, E-cadherin and Snail in six cell lines by RT-PCR. Expressions of claudin-7 and E-cadherin were significantly lower in the HOC313 cells and TSU cells (grade 4D) while expression of Snail was higher in grade 4D than in the other cell lines (Fig. 3).

The expression of claudin-7 and E-cadherin protein was significantly lower in the HOC313 and TSU cell lines (grade 4D) while the expression of Snail was approximately consistent with the other cell lines (Fig. 4).