Seventy-eight tissue samples from Chinese patients with pathologically confirmed OL between January 1996 and December 2010 were obtained from the Department of Pathology, School of Stomatology, Shanghai Jiao Tong University. Nineteen of the 78 OL patients later developed OSCC. Tumor tissues from the 19 patients were also available for the study. The diagnosis of all the samples was verified by examining a hematoxylin and eosin section from each tissue block used in the study. Clinicopathological parameters and the follow-up information were obtained through chart review. Informed consent was received from each enrolled patient, and the research was carried out with the approval from the Ethics Committee of Shanghai Jiao Tong University (Shanghai, China).

The WSU-HN4 cell line previously described (27) was cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco-BRL, Grand Island, NY, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco-BRL) at 37°C in a humidified 5% CO₂ atmosphere. For the EDTA (Sigma-Aldrich; Merck KGaA, Taufkirchen, Bayern, Germany) activation assay, WSU-HN4 cells were washed twice in phosphate-buffered saline (PBS) and incubated in PBS, 2.5 mM EDTA or 2.5 mM CaCl₂ (Sigma-Aldrich; Merck KGaA) for 15 min at 37°C. The cells were then recovered in DMEM for 4 h and subjected to western blot analysis or immunofluorescence staining.

Extracted proteins from cells were assessed using the BCA protein assay reagent kit (Pierce; Thermo Fisher Scientific Inc., Waltham, MA, USA). Protein samples were resolved by 10% SDS-PAGE and immunoblotted with rabbit anti-Notch1 XP™ (1:1,000 dilution; clone D1E11; cat. no. 3608; Cell Signaling Technology, Inc., Danvers, MA, USA) and appropriate HRP-conjugated antibodies anti-rabbit and anti-mouse IgG HRP-linked secondary antibodies (1:2,000 dilution; cat. no. 7074; cat. no. 7076; Cell Signaling Technology, Inc). Mouse anti-β-actin (β-actin; clone AC-74; Sigma-Aldrich; Merck KGaA) was used to normalize protein loading. For immunofluorescence staining, the cells were washed twice in PBS and fixed with 10% formalin in PBS for 10 min. The cells were then treated with 0.25% Triton X-100 (Sigma-Aldrich; Merck KGaA) for 15 min and were blocked with 2.5% normal goat serum in PBS for 40 min. The first antibody (1:500 dilution) was the same as that used previously for the western blotting. The second antibody was goat fluorescein-conjugated anti-rabbit IgG (1:500; DI-1488; Vector Laboratories, Inc., Burlingame, CA, USA). Slides were mounted with Vectashield containing DAPI (Vector Laboratories, Inc.).

Briefly, 4-µm formalin-fixed, paraffin-embedded tissue sections were de-paraffinized and rehydrated. Heat-mediated antigen retrieval using 0.01 M sodium citrate buffer (pH 6.0) was performed and endogenous peroxidase was quenched with 3% hydrogen peroxide for 20 min at room temperature. After incubation with 5% normal goat serum to reduce nonspecific binding, the sections were incubated with rabbit anti-Notch1 (D1E11) XP™ (1:500 dilution; clone D1E11; cat. no. 3608; Cell Signaling Technology, Inc.) according to the manufacturer's protocol.

Notch1 immunoreactivity in both animal and human samples was semi-quantitatively evaluated according to staining intensity and distribution using the immunoreactive score. The grading for the percentage of positive cells was as follows: 0, negative; 1, <10%; 2, 11–50%; 3, 51–80%; and 4, >80% positive cells. For tinctorial strength, scores were distributed as: no staining, 0 points; light yellow, 1 point; brownish yellow, 2 points; dark brownish yellow, 3 points. Finally, the result was classified into four grades by multiplying the two scores aforementioned: 0 points, negative (−); 1–4 points, weakly positive (+); 5–8 points, positive (++); 9–12 points, strongly positive (+++). Scores for >5 points were regarded as positive (28,29).

A total of 48-week-old male SD rats were purchased from the Nanjing Medical University Animal Research Center. The animal experiments were approved by the Animal Ethics Committee and all the procedures were performed following the institutional animal welfare guidelines. After a week of acclimation, the rats were separated into three experimental groups (Groups 1, 2 and 3, n=10 for each group) and one group for control (Group 4, n=10) randomly in separate cages. The rats in the experimental groups were treated with 40 µg/ml 4-NQO (Sigma-Aldrich; Merck KGaA) in the drinking water, while the rats in the control group were treated with an equivalent volume of propylene glycol in water (30,31). The treatment lasted for 8 weeks (Group 1), 16 weeks (Group 2) or 24 weeks (Groups 3 and 4). Then, the rats were euthanized, and the tongues were resected and immediately fixed in 5% paraformaldehyde solution for 12–24 h and paraffin-embedded for further analysis.

Statistical analysis was performed using SPSS 17.0 (SPSS, Inc., Chicago, IL, USA). The Chi-square or Fisher's exact tests were used to determine the relationship between the expression of Notch1 and clinicopathological parameters. OSCC-free survival (OFS) was determined as the outcome variable. The association between Notch1 expression and OFS was estimated using the log-rank test. All the variables were subjected to a univariate and multivariate analysis using Cox proportional hazards regression model. The hazard ratios (HRs) with their corresponding 95% confidence intervals (CIs) and P-values were reported. All the tests were two sided and a P-value <0.05 was considered as statistically significant.

Of the 78 patients with OL, 33 (42%) were male and 45 (58%) were female with age ranging from 27 to 85 years old (mean, 56 years old). Forty-nine (63%) of the OL lesions were located on the tongue and 29 (37%) at other anatomic locations in the oral cavity. During the median 74.18 months follow-up period, 19 of the 78 (24%) patients developed OSCC. The clinicopathological characteristics are listed in Table I. No statistically significant difference was observed in age, sex, OL location, smoking and alcohol consumption statuses between those who developed OSCC and those who did not, except for a trend towards more OL with severe dysplasia in the group that developed OSCC (P=0.063, Chi-square test, Table I).

To ensure the quality of the Notch1 detection method, we first assessed the anti-Notch1 antibody by immunofluorescence in HN4 cells (Fig. 1A-a-c). Then, we artificially induced the dissociation of the heterodimer by treating the cells with EDTA to activate Notch1 cleavage. After treatment with 2.5 mM EDTA, the immunofluorescence staining for Notch1 was enriched in the nucleus of HN4 cells (Fig. 1A-d-f). Notch1 cleaved at S2 and S3 sites could be detected by a Notch1-specific antibody (clone D1E11) or Notch1 cleaved at S3 could be directly detected by Notch1 Val1744-specific antibody (clone D3B8). Western blot assays enabled the detection of Notch1 as the expected transmembranous form (~120 kDa) in the HN4 cells (Fig. 1B). A smaller size band expected to be cleaved-NICD was detected after EDTA treatment, but not for the cells treated with CaCl₂, which neutralized the function of EDTA (Fig. 1B). These data ascertained the usability of this Notch1 detection method.

By performing IHC staining, we assessed the expression of Notch1 in 78 OL specimens and 19 corresponding OSCC specimens from patients who progressed from OL to OSCC. The weak/moderate/strong levels of Notch1 staining were demonstrated in Fig. 2A-C. In OL, Notch1 was generally expressed in a diffused manner in the suprabasal layer of the epithelium and mostly confined to the nucleus and cytoplasm with variable intensities (Fig. 3A-C). Strong overall expression of Notch1 was observed in 10 (13%) cases (Table I). Notably, in 24 (31%) of the OL samples, membranous staining of Notch1 was observed (Fig. 3D and E, Table I). We noted that lesions with severe dysplasia (Fig. 3D) tended to have stronger membrane staining and cytoplasmic staining of Notch1 than in the lesions with low-to-medium grade dysplasia (Fig. 3E). This membrane staining revealed variable intensity in the spinous layer, granular layer and throughout the whole epithelium compared with the more strictly suprabasal layer expression of Notch1 in the other OL samples.

In the 19 malignant-transformed OL (or OSCC) samples derived from OL cases, 11 (14%) exhibited positive membranous Notch1 expression (Table I). To be specific, in the 19 cases, the lesions in Fig. 3D and E turned into the malignancies in Fig. 3F and G, respectively during the follow-up studies. Moreover, Notch1 expression was more diffused on the membrane and in the cytoplasm than in the nucleus in these OSCC samples. Strong membranous Notch1 expression was often observed at the invasive front or at the border of cancer nests. Moreover, a significant decreased nuclear Notch1 expression was detected in malignant-transformed (MT) OL samples, compared to the untransformed (UT) OL samples (P=0.001, t-test; Fig. 4A). Meanwhile, a decrease of unclear Notch1 expression (P<0.001, grouped t-test) and an increase of membranous Notch1 expression (P=0.002, grouped t-test) were observed in the 19 OSCC samples compared to their matched OL samples (Fig. 4B and C). Even within the same OSCC sample, we observed a decrease in nuclear Notch1 and an increase in membranous Notch1 expression in cancer nests compared to the adjacent non-cancerous epithelium (Fig. 4D).

The relationship between the membranous Notch1 expression in OL and clinicopathological parameters was also analyzed. A statistically significant association was observed between membranous Notch1 expression and a more severe dysplastic status (P<0.001, Chi-square test, Table II). For further analysis, we determined the role of membranous Notch1 expression in the malignant transformation of OL. As expected, the grade of dysplasia and expression levels of membranous Notch1 were associated with OSCC-free survival (P=0.016 and 0.001, respectively, log-rank test, Fig. 4E and F). After 5 years, only 7 (13%) of the 54 patients without membranous Notch1 expression developed OSCC, whereas 9 (38%) of the 24 patients with membranous Notch1 expression developed OSCC. When combined with grade of dysplasia, the patients with membranous Notch1 expression and severe dysplasia exhibited even shorter OSCC-free survival than the other patients (P=0.000, log-rank test, Fig. 4G). As determined by univariate analysis, both the expression level of membranous Notch1 and the grade of dysplasia were associated with OSCC development (P=0.002 and P=0.022, respectively, Table III). By multivariate analysis, however, only membrane Notch1 expression was an independent factor that was significantly associated with OSCC development (P=0.019, Table III).

An OL/OSCC model was successfully established in Sprague-Dawley rats by 4-nitroquinoline-1-oxide (4-NQO; Fig. 5A) induction as aforementioned to study the expression of Notch1 in different stages of carcinogenesis. Administration of 4-NQO was performed as outlined in the general scheme in Fig. 5B. We found that the experimental groups (Groups 1–3) treated with 4-NQO developed different pathological changes, including mild dysplasia (Fig. 5C-b), moderate-to-severe dysplasia (Fig. 5C-c) and OSCC (including in situ carcinoma and invasive carcinoma; Fig. 5C-d) with the increase in the duration of 4-NQO administration, and no visible lesions were detected in the control group (Group 4, Fig. 5C-a). The pathological lesions of samples were confirmed by two trained pathologists using H&E staining in a double-blind fashion (Table IV). The rates of positive expression of Notch1 in normal mucosa, dysplasia and carcinoma were 20% (2/10), 64.7% (11/17) and 84.6% (11/13), respectively (P<0.01, Chi-square test, Table V). Representative Notch1 IHC images are shown in Fig. 5D. We found that Notch1 was mainly localized in the membrane and cytoplasm in this murine model. Notch1 staining was negative in the normal tongue mucosa or was mainly distributed in the stratum basale (Fig. 5D-a), whereas it extended from the stratum basale to the stratum corneum during the progression of cancer (Fig. 5D-a-d). The IHC scores of membranous Notch1 expression were also increased during carcinogenesis (Fig. 5E).