The study population included 256 OCSCC patients who underwent primary surgical resection without previous radiotherapy and/or chemotherapy between October 1996 and August 2005. Clinicopathological information for each subject, including gender, age, tumor (T) stage, nodal (N) status, tumor node metastasis (TNM) stage and overall survival, was obtained retrospectively from clinical records and pathological reports. TNM status was classified according to the 1997 American Joint Committee on Cancer (AJCC) system. The study was approved by the Medical Ethics and Human Clinical Trial Committee at Chang Gung Memorial Hospital, Taipei, Taiwan. The patient group comprised 17 women and 239 men, with an average age of 50.9 years (range, 26–87 years). Thirty-nine patients were diagnosed with T1 tumors, 55 with T2, 64 with T3 and 98 with T4. A total of 153 patients had an N status of N0, 38 had N1, 48 had N2b, 13 had N2c and 4 had N3. Thirty-four patients had stage I tumors, 38 stage II, 61 stage III and 123 stage IV.

For tissue protein extraction, frozen samples (adjacent non-tumor and tumor tissues) were homogenized in RIPA lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40, 0.5% Na-deoxycholate and 0.1% SDS), and the protein concentrations were quantified using a Bio-Rad Protein assay (Bio-Rad, Hercules, CA, USA). Immunoblotting was performed according to standard procedures. Anti-TRF1 and -TRF2 polyclonal antibodies and the anti-GAPDH monoclonal antibody (Santa Cruz Biotechnology Santa Cruz, CA, USA) were used. The bound primary antibody was detected by incubation with HRP-conjugated secondary antibody (Bio/Can Scientific, Mississauga, ON, Canada). Blots were developed using the Western Lighting reagent, and protein bands were visualized using X-ray film.

OCSCC and adjacent non-cancerous tissue samples were identified by a pathologist based on diagnosis and microscopic morphology. Tumor tissues were fixed in 10% buffered formalin, embedded in paraffin and decalcified in 10% EDTA. Formalin-fixed paraffin-embedded tissue was sectioned to a thickness of 4 μm, and the sections were deparaffinized in xylene and rehydrated in a graded series of ethanol (100, 90, 80 and 70%). The sections were washed in phosphate-buffered saline (PBS) and treated with 3% H₂O₂ for 30 min to block the endogenous peroxidase activity. Antigens were retrieved by microwaving the sections in 10 mM citrate buffer, pH 6.0. The sections were then incubated with anti-TRF1 and -TRF2 antibodies (diluted 1:100) for 1 h, washed in PBS and incubated for 30 min with a horseradish peroxidase-Fab polymer conjugate (PicTure™-Plus kit; Zymed, South San Francisco, CA, USA). After the sections were washed in PBS, the immunoreactive bands were visualized by incubation with 3,3′-diaminobenzidine for 5 min. As a negative control, the primary antibody was omitted. Two pathologists blinded to the subjects’ clinical information independently evaluated the reactivity level of the immunostained tissues in 15–20 high-power fields. Criteria were developed for quantitating the immunoreactivity of the TRF1 and TRF2 staining in the adjacent non-tumor and tumor sections using a score range of 0 to +3. A percentage value of 0 indicated 0–25% of the area stained; +1, 25–50%; +2, 50–75% and +3, >75% stained. Similarly, the staining intensity was graded as +0, +1, +2 or +3. High expression of TRF1 and TRF2 was defined as ≥+2 in both scoring methods. Low expression of TRF1 and TRF2 was defined as ≤+1 in both scoring methods.

Fisher’s exact test was used to evaluate the correlation between TRF1 and TRF2 expression and various clinicopathological variables, including gender, age, N status, T stage and TNM stage. A p-value <0.05 was considered to indicate statistical significance. TRF1 and TRF2 expression and the clinicopathological variables were used in a Kaplan-Meier analysis of survival, and statistical significance (p<0.05) was assessed by the log-rank test. To determine the effects of specific prognostic factors on survival, a multivariate analysis was performed using Cox’s regression model.

To investigate the potential roles of TRF1 and TRF2 in the pathogenesis of OCSCC, their expression was assessed in representative and paired tumor and adjacent non-cancerous tissue samples by Western blot analysis using anti-human TRF1 and TRF2 polyclonal antibodies. The protein expression levels of TRF1 and TRF2 were lower in the tumor samples compared to those in the paired non-cancerous tissues (Fig. 1).

TRF1 and TRF2 expression in the tumor and adjacent non-cancerous tissues from the 256 OCSCC patients was also examined immunohistochemically. Representative results of TRF1 and TRF2 immunostaining are presented in Fig. 2A. Staining was stronger in the adjacent non-cancerous tissues (Fig. 2A-a and -f) than the tumor tissues (Fig. 2A-b and -c, TRF1; Fig. 2A-g and -h, TRF2). Moreover, TRF1 and TRF2 expression levels in the tumor samples were negatively correlated with T stage (Fig. 2A-b and -c, TRF1; Fig. 2A-g and -h, TRF2) and N stage (Fig. 2A-d and -e, TRF1; Fig. 2A-i and -j, TRF2). Notably, TRF1 and TRF2 were focally expressed in the nuclei of both tumor and non-cancerous cells (Fig. 2A-b and -c, TRF1; Fig. 2A-g and -h, TRF2).

To investigate whether the expression of TRF1 and TRF2 is associated with various prognostic factors, including age, gender and TNM pathologic classification, we classified the patients into two groups based on the immunohistochemical analysis: low (−/+) and high (++/+++) TRF1 or TRF2 expression. Low TRF1 expression was correlated with advanced T stage (p<0.001) and advanced TNM stage (p<0.001). Low TRF2 expression was correlated with advanced T stage (p<0.001) and advanced TNM stage (p<0.001), as well as positive lymph node metastasis (p=0.022). Neither TRF1 nor TRF2 expression was correlated with age or gender (Table I). These findings suggest that TRF1 and TRF2 expression levels may be linked to tumor progression in OCSCC.

In view of the finding that TRF1 and TRF2 expression levels were associated with T stage, we investigated whether TRF1 and TRF2 expression was correlated with patient prognosis. As shown in Fig. 2B, Kaplan-Meier overall survival analysis revealed that the prognosis of patients with low (−/+) tumor expression of TRF1 and TRF2 was significantly poorer than that of patients displaying higher (++/+++) expression (p<0.001). Univariate analysis revealed that advanced T stage (p<0.001), positive N stage (p<0.001), advanced TNM stage (p<0.001), low TRF1 expression and low TRF2 expression each predicted a significantly worse prognosis for OCSCC patients (Table II). The prognosis was not associated with age or gender. Cox regression analysis revealed that T stage (95% CI, 1.585–5.037; RR=2.826; p<0.001), N status (95% CI, 1.966–4.681; RR=3.034; p<0.001) and TRF1 expression (95% CI, 0.391–0.924; RR=0.601; p=0.02) were independent prognostic factors for survival. These results clearly indicate that the clinical prognosis for OCSCC patients is affected by the tumor expression of TRF1 and TRF2, and suggest that TRF1 and TRF2 may be good prognostic indicators in OCSCC.