Background
Oral squamous cell carcinoma (OSCC) is a common malignant tumor of the head and neck region; however, patients with OSCC have a 5-year survival rate of less than 50 %. The traditional TNM staging and histopathological grading systems do not enable physicians to accurately predict a tumor’s aggressiveness or select treatment modalities on an individual basis [
1]. Therefore, research has been underway to investigate more effective and definitive tumor prognostic markers for patients with OSCC [
2,
3]. As the cancer stem cell (CSC) hypothesis has become one of the predominant theories explaining the tumor-initiating capacity and heterogeneity of tumor cells, various stem cell markers have been studied to determine their correlation with clinicopathological tumor features and long-term prognosis [
4‐
8].
The early transcription factors NANOG, OCT4, and SOX2, which play pivotal roles in the maintenance of pluripotency and self-renewal ability in both embryonic and adult stem cells, have also been reported as key regulation factors in the CSCs of head and neck squamous cell carcinoma (HNSCC) [
7,
8]. High expression of NANOG and OCT4 has been positively correlated with histologically high-grade carcinomas and clinically poor prognosis [
7‐
10]. CD44 was the first CSC marker described in a solid malignancy [
5], and a high frequency of CD44 positive cells in HNSCC strongly correlates with recurrence and tumor aggressiveness [
11]. Tumor protein 53 (p53) is considered a traditional tumor biomarker in squamous cell carcinoma (SCC); however, its usefulness as a prognostic indicator remains unclear [
1]. Wild-type p53 protein is barely detectable in normal tissues because of its short half-life of approximately 20 min. [
12] However, the tumor suppressor gene p53 is mutated in approximately 40–60 % of OSCC cases, and this mutant p53 plays an important role in tumor development and progression [
12,
13]. Interestingly, mutant p53 protein is usually detectable via immunohistochemistry (IHC) [
14,
15].
In order to predict long-term prognosis and define individual treatment modalities for patients with OSCC, more reliable tumor biomarkers (including immunohistochemical prognostic markers) are needed during the pretreatment workup period. The present study aimed to investigate the expression of various tumor markers via IHC, including stem cell and tumor-related biomarkers, to identify more reliable prognostic markers in OSCC biopsy specimens collected prior to cancer treatment. In the present study, we observed a positive correlation between the clinicopathological features of OSCCs and the immunohistochemical expression patterns of NANOG, human mutant p53, and CD44. In addition, the immunostaining intensities of these marker proteins (NANOG and mutant p53) were positively correlated with the overall survival rates of patients with OSCC.
Discussion
Since the emergence of the CSC hypothesis as a key theory to explain cancer progression, the expression patterns of several pluripotent stem cell markers and transcription factors have been studied in various cancer tissues or cell lines to determine their correlation with long-term prognosis. In the present study, the early transcription factor NANOG, the CSC marker CD44, and human mutant p53 were immunohistochemically investigated in the pretreatment biopsy specimens of 57 patients with OSCC to identify any relationship of their expression patterns with clinicopathological tumor features and patient prognosis. The results of this study demonstrated that the immunohistochemical expression patterns of NANOG and mutant p53 were directly associated with overall survival rates as well as clinicopathological features, including tumor stage, neck node metastasis, and histological grade. Patients with OSCC whose tumors had higher expression of NANOG and mutant p53 had less favorable clinicopathological features and lower survival rates. Enhanced expression of CD44 was also associated with poor prognosis and lower survival rates, but no statistical significance was observed. Recently, a common modality for predicting cancer prognosis has been the measurement of intra-tumor genetic heterogeneity [
19,
20]. Heterogeneity denotes that tumors are composed of multiple clonal subpopulations of different characteristic cells, and thus, histological examination frequently reveals differences in cell morphology and properties, among cells in the same cancer specimen [
20]. Enhanced intra-tumor heterogeneity was directly correlated with increased mortality; in particular, high heterogeneity and p53 mutation positivity were associated with higher mortality rates in a large cohort study of head and neck cancer [
19]. In the present study, serial sections of tumor specimens frequently revealed the co-expression of NANOG, mutant p53, and CD44, which was associated with poorer prognosis. The co-detection of these different marker proteins in the same cancer cell or same tumor region may represent high tumor heterogeneity.
NANOG is an early transcription factor and pluripotent marker that maintains the self-renewal of embryonic and mesenchymal stem cells [
8]. Recent studies revealed that NANOG is highly detected in poorly differentiated carcinomas and late-stage tumors [
8‐
10]. Moreover, a positive relationship between enhanced NANOG expression and lymph node metastasis of carcinoma was reported [
21]. Similarly, the present study illustrated that strong NANOG expression was positively correlated with clinically late-stage and histopathologically high-grade tumors. However, our findings on the relationship between neck node metastasis and NANOG expression slightly differed from those observed in previous studies. Weak or negative NANOG expression was associated with negative neck node metastasis, whereas higher NANOG expression was not associated with positive neck node metastasis in the present study. In IHC, NANOG is usually detected in the nuclei of pluripotent cells. However, in the present study, this transcription-related protein was sometimes observed in the cell cytoplasm, which is in agreement with previous reports in which NANOG was occasionally detected in the cytoplasm of highly primitive undifferentiated stem cells or tumor cells from patients with poor prognoses; these findings may be related to the pluripotency of stem cells and invasive or recurrent cancer cells [
22,
23]. In this study, cytoplasmic NANOG expression was usually detected in high-grade OSCCs (Figs.
1 and
2). In addition, the overall survival rate was significantly lower in patients with strong NANOG expression than in those with weak or negative NANOG expression. These results suggest that OSCCs with high NANOG expression, including cytoplasmic expression, may possess aggressive characteristics, which in turn indicate poor prognosis.
The wild-type p53 protein is a DNA-binding transcription factor that acts as a tumor suppressor by allowing the cell time to repair and recover from DNA damage or by inducing cell apoptosis in cases of serious damage [
12,
24]. A review of a large cohort study of OSCC revealed conflicting evidence for p53 as a tumor prognostic factor; some studies uncovered that positive p53 staining was associated with poor prognosis, whereas others reported no such association [
1]. These contradictory results may be related to the frequency of mutation of the p53 gene in cancer tissues, particularly carcinomas. Mutant p53 causes the inactivation and dysfunction of wild-type p53 and plays a pivotal role in the development and progression of carcinomas [
13,
25]. Co-detection of p53 and mouse double minute 2 (MDM2) by IHC might be helpful for discriminating the functional type of p53; high levels of p53 without increased MDM2 expression may indicate the inactivating type of p53, usually mutant p53 [
17]. Interestingly, mutant p53 is easily detected via immunohistochemical analysis [
14]. Although wild-type p53 can also be detected in cancerous tissues using IHC, wild-type p53 (normally a tumor suppressor gene) may be not related with poor prognosis of cancer [
26,
27]. In the present study, a human anti-mutant p53 antibody (Abcam™, ab32049), which does not recognize human wild-type p53 protein but does react with a synthetic peptide corresponding to human mutant p53 aa 1-100 (N terminal), was used to detect the mutated type of p53 in the pretreatment biopsy specimens of patients with OSCC. A positive p53 mutation signal was detected in 34 of 57 OSCC specimens (59.6 %). This result supports the finding from a previous study, which reported a p53 mutation rate of 40–60 % in carcinoma tissues [
1,
17]. In addition, we observed a positive correlation between high NANOG expression and mutant p53 positivity: OSCCs with moderate or strong expression of NANOG had a higher frequency of p53 mutation (Fig.
3a). In addition, OSCCs that expressed mutant p53 were more frequently high-grade tumors (Fig.
3b), whereas tumors that were negative for mutant p53 were associated with fewer neck node metastases (Fig.
4b). Moreover, patients with p53 mutations had a significantly poorer survival rate than those without p53 mutations (Fig.
5f). Interestingly, the co-detection of an enhanced NANOG signal and mutant p53 [NANOG(++)/p53(+)] was correlated with significantly lower patient survival versus tumors with weak NANOG expression and p53 negativity [NANOG(+/−)/p53(−)] (Fig.
5h).
CD44 is a cell surface glycoprotein that acts as a receptor for hyaluronic acid and as an adhesion molecule [
7,
28]. This cell surface protein plays a role in tumor cell invasion, metastasis, and angiogenesis by interacting with certain matrix metalloproteinases [
29]. CD44 was first described as a CSC marker in breast cancer and HNSCC [
4,
5], and it has since been used as a CSC marker and prognostic factor for SCC [
11,
30]. CD44 has exhibited positive correlations with tumor recurrence, high-grade SCCs, and poor prognosis [
30]. However, its limited usefulness as a CSC marker or prognostic factor in OSCC was also reported [
31,
32]. In the present study, OSCCs that expressed high levels of NANOG and mutant p53 also displayed significantly elevated expression of CD44 (Fig.
3a). In addition, high CD44 expression was associated with clinically late-stage and histologically high-grade tumors, although no statistical significance was found in this analysis. Only the group with weak CD44 expression displayed a statistically higher frequency of negative neck node metastasis (Fig.
4b). However, in this study, the CD44 expression pattern was not statistically correlated with overall survival rates, but high expression of CD44 tended to be associated with poor prognosis (Fig.
5g). These results indicate that CD44 expression could have limited use as a clinicopathological and prognostic marker for OSCC, as suggested in previous reports [
28,
32].
In this study, overall survival rates were analyzed in terms of the clinicopathological features of OSCCs as well as the immunohistochemical expression patterns of the three marker proteins. In the Kaplan-Meier analysis with the log-rank test, patients with late-stage tumors, positive neck node metastases, and high histopathological grade tumors had significantly lower survival rates. Similar results were obtained by univariate and multivariate Cox proportional hazards regression analysis for overall survival of the 57 patients with OSCC. Neck node metastasis, histological grade, tumor stage, and the expression intensity of NANOG and mutant p53 were directly associated with the overall survival rate of the patients; however, the relationship between survival and the expression pattern of mutant p53 was not statistically significant in multivariate Cox analysis (Tables
2–
3). In the three groups as divided treatment modality, the RT group displayed a much poorer prognosis than the Surg or Surg + RT group (Fig.
5d). This result is likely attributable to the characteristics of the patients in the RT group, as palliative radiotherapy is selected as a treatment option in cases of tumor inoperability or patient refusal to undergo surgery.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
HJL and YHK collected and analyzed the patient data and partially contribute in writing the manuscript. JSL evaluated the histological and immunohistochemical results of all tumor specimens. JHB and UKK assisted treatment of the patients and performed patient care. SJJ and GJR mainly conducted the immunohistochemistry for the present study and assisted statistical analysis of the obtained data. BWP performed the patient management, and was a major contributor in writing the manuscript. All authors read and approved the final manuscript.