Abstract
Background/Aim: Oral squamous cell carcinoma (OSCC) is a highly invasive malignancy with poor prognosis. Recent reports suggest that Sonic Hedgehog (SHH) plays a key role in tumor progression and worsens the response to therapy, possibly through an association with a cancer stem cell (CSC) phenotype. The objective of our study was to investigate the relationship between SHH expression and CSC markers in OSCC. Materials and Methods: A total of 67 OSCC specimens were immunostained for SHH and CSC markers using specific antibodies and expression was correlated with clinicopathological parameters. Results: SHH expression was significantly correlated with CD133 (p=0.026, r=0.272) and SRY-box transcription factor 2 (SOX2; p<0.001, r=0.793). SHH and SOX2 expression were associated with worse survival in OSCC (p=0.003 and p=0.003, respectively). In multivariate analysis SHH and CD44 were independent prognostic biomarkers in patients with OSCC (p=0.001 and p=0.008, respectively). Conclusion: Our study revealed that SHH overexpression is closely associated with CSC markers, contributing to tumor progression and worse outcomes of patients with OSCC.
Oral squamous cell carcinoma (OSCC) is one of the most common types of cancer of the head and neck region, with an incidence of over 350,000 cases per year worldwide (1, 2). OSCCs are characterized by aggressive clinical behavior and poor response to therapy (2). It was found that OSCC consists of two biologically and morphologically different fractions, differentiated and undifferentiated tumor cells, which might reflect different biological tumor behaviors (3, 4). There are reports which revealed that the undifferentiated subpopulation of oral cancer cells often shows properties similar to those of normal adult stem cells with the innate capacity to self-renew and the potential to differentiate into progenitor cells (3, 4). These cells also possess unique features such as the ability to initiate the growth and progression of tumor and are described as carcinoma stem cells (CSCs) (3-5). Increasing experimental studies support the notion of CSCs in the development of OSCC, finding that a small proportion of stem cells have the capability to sustain tumor formation, growth and differentiation into various heterogeneous cancer cells (3-6). Recent studies indicated that CSCs have increased motility that allow them to invade surrounding tissue prior to metastasis to other organs (3, 4, 7).
Cluster of differentiation (CD44) is an important multifunctional transmembrane cell surface glycoprotein which is involved in cell aggregation, differentiation, migration and proliferation (8). CD44 was the first CSC marker described in many types of cancer, including OSCC, and a high frequency of CD44-positive cells in OSCCs is strongly correlated with recurrence and tumor aggressiveness (9, 10). A higher frequency of CD44-positive cells was observed in head and neck carcinomas (HNSCCs) that recurred than in tumors without recurrence (4). Previous studies suggest a direct correlation between CD44 expression, cancer stem cells and the aggressiveness of HNSCC (3).
CD133 (also known as prominin-1) is a five-transmembrane glycoprotein that was proposed as a marker for CSCs but its cellular function is not clear (5, 11). Its prognostic value for patients was also found in cancer such as colorectal, gastric and lung (12-14). In an OSCC study, it was shown that CD133 expression identified tumor cells with stemness features (5, 11). According to latest data, positive correlation between high CD133 expression and octamer-binding transcription factor 4 (OCT4), NANOG in oral cancer may identify a patient’s resistance to chemotherapy (11).
Another molecule, SRY-box transcription factor 2 (SOX2), is a stem cell transcription factor (15, 16). SOX2 is a member of the SRY-related high-mobility group box family and it was initially found to maintain the embryonic stem cell pluripotency (4, 15, 16). There are data which revealed the role of SOX2 in oral carcinogenesis (5, 15, 17). Some studies showed that SOX2 promotes cell proliferation, migration and tumor metastasis in several solid tumor types, including OSCC (4, 5). Several reports demonstrated that nuclear SOX2 accumulation is closely associated with tumor recurrence and poor prognosis in patients with HNSCC (4, 5, 18). The clinical utility of SOX2 in determining the outcome of patients with OSCC is still controversial but some authors showed that SOX2 might be considered as biomarker for oral cancer risk assessment (15).
Signaling pathways, including Hedgehog (HH), WNT and NOTCH, that control normal stem cell self-renewal and differentiation are frequently overactivated in oral CSCs (5, 17). Therefore, identification of the crucial pathways for CSC maintenance in tumor tissue is required (4, 5, 17). The Sonic Hedgehog (SHH) pathway, one of components of the HH pathway, is involved in proliferation, differentiation and migration of normal stem cells (19). However, the SHH pathway is also involved in tumorigenesis and metastasis of human carcinomas (19, 20). Selective inhibition of SHH or one of its components is used with good clinical effect in basal cell carcinoma and medulloblastoma (21). Previously, we showed that SHH overexpression is an independent negative prognostic biomarker in OSCC (22). Interestingly, recent studies of other tumor types revealed that the HH pathway plays an important role in invasion and tumor progression via its interaction with CSCs (23). The HH pathway is a known regulator of CSCs in esophageal, gastric, breast, lung and prostate cancer (24-29). Clinical impact and correlations between CSC markers and the HH pathway in OSCC remain still unclear. To address this issue, we investigated the relationship between SHH expression and known CSC markers (CD44, CD133 and SOX2) in OSCC in order to define their role in clinical outcome of patients with OSCC.
Materials and Methods
Patients. Medical records and excisional surgical specimens of 67 patients with primary OSCC diagnosed between 2011 and 2015 at the Department of Pathomorphology and Oncological Cytology of Wroclaw Medical University were used in our study. All patients underwent surgery as primary treatment of OSCC only from the following sites: tongue, floor of the mouth, oral vestibule and hard palate. Patients with other localization of tumor, non-OSCC histopathology or radiotherapy/chemotherapy before surgery were excluded from this study. In our study, we considered the following clinicopathological parameters: Age, gender, tumor grade, tumor location, pathological tumor (pT) stage, lymph node metastasis status, clinical stage and survival. Comprehensive clinicopathological data was retrieved from medical documentation. Specimens were histopathologically graded and tumors were categorized as well (grade 1), moderately (grade 2), or poorly differentiated (grade 3) according to the World Health Organization classification (30). Pathological tumor stage (pT), lymph node metastasis status and clinical staging were established according to the seventh edition of the American Joint Committee on Cancer staging criteria (31). Survival status of patients was sourced from the Lower Silesian Cancer Registry. All patients were followed-up until March 2021. The study was performed in accordance with the declaration of Helsinki, and it was accepted by the Ethics Committee of Wroclaw Medical University, Poland (protocol number: KB-230/2016). Informed consent from patients was waived by the Ethics Committee because it was a retrospective study without impact on the original treatment of patients.
Immunohistochemistry. Immunohistochemical studies were performed on paraffin-embedded OSCC tissues from the selected blocks using the Universal Dako REAL EnVision Detection System, Peroxidase/DAB+, Rabbit/Mouse (Dako, Copenhagen, Denmark) and the following antibodies: anti-SHH (clone EP1190Y, dilution: 1:200; Abcam, Cambridge, UK), anti-CD44 (clone 156-3C11, 1:200; ThermoFisher Scientific, Waltham, MA, USA), anti-CD133 (orb99113, 1:400; Biorbyt, Cowley, UK) and anti-SOX-2 (orb11398, 1:250; Biorbyt). All paraffin-embedded OSCC blocks were cut into 4 μm slices, deparaffinized and antigen-retrieval was carried out in citrate buffer (pH=6.0) by microwaving at 800 W for 2×5 min for each antibody. Afterwards, the slides were slowly cooled for 30 min. Endogenous peroxidase was blocked with Dako REAL Peroxidase Blocking Solution (Dako). Tissue slides were incubated with specific antibodies overnight at 4°C. Afterwards, tissue sections were washed with 0.1 M Tris-buffer (pH=7.4) and incubated for 30 min at room temperature with Dako REAL EnVision/HRP, Rabbit/Mouse secondary antibody (Dako). After that, the slides were incubated with 3,3’-diaminobenzidine (Dako) as a chromogen for 8 min at room temperature. Then sections were counterstained with hematoxylin and mounted. Internal positive controls were performed according to the manufacturer’s instructions. Negative controls were prepared with Tris-buffer in place of primary antibodies.
The intensity of SHH, CD44, CD133, SOX2 immunostaining and the percentage of tumor tissue with positive reaction in the tissue slides were evaluated semiquantitatively with usage of a double-headed light microscope Olympus BX-51 (Olympus, Tokyo, Japan) by two authors blinded to clinical data. Immunoreactivity of SHH, CD44 and CD133 expression was assessed by determining membranous immunostaining based on the intensity of immunostaining and the percentage of positive tumor tissue area. SOX2 expression was evaluated by counting 1,000 tumor cells in 10 randomly selected high-power fields in relation to the total number of tumor cells. Positive reaction was judged when immunostaining of SHH, CD44, CD133, SOX2 was found in 10% or more tumor cells. The intensity of reaction was scored as: negative (no color), weak (light brown color), moderate (dark brown color), strong (very dark brown color). For the purpose of statistical analysis, SHH, CD44, CD133 and SOX2 expression was divided into two groups: negative - less than 10% positive tumor tissue area, and positive - 10% or more positive tumor tissue area.
Statistical analysis. The statistical associations between expression of SHH, CD44, CD133 and SOX2 and clinicopathological data were evaluated by chi-squared, Mann–Whitney and Kruskal–Wallis tests. Correlations between analyzed proteins (percentage of immunostained tumor tissue area) were studied using Spearman’s test. Overall survival (OS) was calculated as the number of months from the day of surgery to the last follow-up examination or death. Survival curves were plotted using the Kaplan–Meier method and compared with the log-rank test. In order to identify prognostic factors for OS, Cox regression analyses were performed. The statistical significance limit was p<0.05. All statistical analyses and curves were carried out in the STATISTICA v13.0 software (Statsoft, Krakow, Poland).
Results
Patient characteristics. The main clinicopathological features of patients included in this study are detailed in Table I. The median age of patients at diagnosis was 63 years. The median follow-up time for overall survival was 108.7 (range=100.9-118.1) months for alive patients and 12.4 (range=1-118.4) months for deceased patients. Death was documented in 51/67 (76.1%) patients.
Clinicopathological characteristics of the study group.
SHH, CD44, CD133 and SOX2 expression in oral squamous cell carcinoma. The presence of SHH was noted in 48/67 (71.6%) cases and ranged from 10% to 90% (median=40%) of tumor tissue (Figure 1A). Stromal cells did not reveal SHH immunopositivity. The majority of OSCCs showed strong intensity of reaction (56.3% of all positive cases). Increased SHH expression was associated with the presence of lymph node metastasis but this correlation did not reach statistical significance (p=0.083). There were no correlations between SHH expression and other clinicopathological parameters (Table II).
Immunohistochemical staining of Sonic Hedgehog (SHH), cluster of differentiation 44 (CD44), cluster of differentiation 133 (CD133) and SRY-box transcription factor 2 (SOX2) in oral squamous cell carcinoma (OSCC) tissues (EnVision technique). A: Intensive, membranous immunoreactivity of SHH in OSCC tissue. Scale bar=100 μm. B: Membranous CD44 expression in tumor tissue. Scale bar=50 μm. C: Membranous CD133 immunostaining in OSCC tissue. Scale bar=50 μm. D: Nuclear accumulation of SOX2 in OSCC cells. Scale bar=50 μm.
Association between clinicopathological parameters and protein expression in patients with oral squamous cell carcinoma.
Membranous CD44 immunostaining was found in 32/67 (47.8%) cases. The presence of CD44 was noted to range from 10% to 70% (median=30%) of tumor tissue area (Figure 1B). CD44 expression was absent from stroma cells, however, CD44 expression was observed in tumor cells on the tumor/stromal border but was weaker than in the central part of the tumor. Most positive cases presented moderate intensity of immunostaining (53.1%). The presence of CD44 was significantly associated with more advanced clinical stage (p=0.042). No correlations between CD44 expression and other clinicopathological parameters were found (Table II).
CD133 expression was found in 33/67 (49.3%) OSCCs. CD133 expression ranged from 10% to 60% (median=30%) of positive tumor cells (Figure 1C). Moderate intensity of reaction dominated in positive cases (66.7%). No associations between CD133 expression and clinicopathological parameters of OSCC were observed (Table II).
Nuclear SOX2 staining was observed in 45/67 (67.2%) OSCCs. SOX2 immunoreactivity was observed in 10% to 70% (median=30%) of tumor tissue (Figure 1D). The majority of OSCCs showed strong intensity of immunoreaction (55.6% of all positive cases). The presence of SOX2 was associated with gender, SOX2 expression being significantly more frequent in male patients (p=0.036). Statistically significant correlation was revealed between SOX2 expression and lymph node metastasis (p=0.024). There were no statistically significant associations between SOX2 expression and other clinicopathological parameters (Table II).
Correlations between SHH and CSC markers in OSCC. We conducted correlation analyses between SHH and CSC markers in analyzed OSCCs. Strong, positive correlation was found between SHH and SOX2 (p<0.001, r=0.793). CD133 expression positively correlated with SHH expression (p=0.026, r=0.272) as well as with SOX2 expression (p=0.049, r=0.242) (Table III). The Spearman correlation test did not reveal a statistically significant association between SHH and CD44, although a trend for positive correlation between these proteins was observed (p=0.082, r=0.214). No correlations were found between CD44 and CD133 (Table III).
Spearman’s rank correlation between expression levels of Sonic Hedgehog (SHH), cluster of differentiation 44 (CD44), cluster of differentiation 133 (CD133) and SRY-box transcription factor 2 (SOX2) in oral squamous cell carcinomas.
Prognostic significance of SHH and CSC markers in patients with OSCC. Survival analysis was performed on all cases included in this study using the Kaplan–Meier method and compared with the log-rank test. OS in months was calculated from the day of surgery to the last follow-up examination or death. Kaplan–Meier survival curves were plotted according to SHH, CD44, CD133, SOX2 expression. Kaplan–Meier analysis indicated that SHH and SOX2 expression were significantly correlated with shorter OS (p=0.003 and p=0.003, respectively). Kaplan-Meier analysis revealed also that CD44 is associated with worse survival, but this correlation did not reach statistical significance (p=0.063) (Figure 2). In multivariate analysis, tumor grade, pT stage, SHH and CD44 expression were shown to be significant independent prognostic factors for survival of patients with OSCC (p<0.001, p<0.001, p=0.001 and p=0.008, respectively) (Table IV).
Kaplan–Meier analysis of overall survival curves for: Sonic Hedgehog (SHH) (A), cluster of differentiation 44 (CD44) (B), cluster of differentiation (CD133) (C), SRY-box transcription factor 2 (SOX2) (D) in patients with oral squamous cell carcinoma (p-values of log-rank test).
Univariate and multivariate Cox regression analysis of overall survival in patients with oral squamous carcinoma.
Discussion
CSCs are considered new targets in cancer therapy because of their chemo/radioresistance and potential for recurrence and metastasis (4, 5). Recent reports showed that HH pathway is involved in maintenance of CSCs from various tumors (25, 28, 29) but still little is known about their relationships in OSCC. In this study, we investigated associations between SHH expression and CSC markers in OSCC in order to establish their impact on the clinical outcome of patients with OSCC.
Prognostic significance of SHH immunoreactivity was observed in various malignancies, suggesting that SHH could be a valuable prognostic biomarker (32, 33). In the current study, multivariate Cox regression analysis confirmed that SHH overexpression is an independent biomarker for poor prognosis in OSCC. Increased SHH expression was associated with shorter OS of patients with OSCC. SHH overexpression was observed in 71.6% of cases, which is comparable to data from other authors (34, 35). Moreover, similarly to previous data (34, 36), high SHH expression was associated with lymphatic metastasis, confirming the important role of SHH in tumor progression and cancer invasion. Some investigators found that SHH expression is also associated with clinical stage (34, 35, 37). Based on previous data (28, 38), results from this study suggest that SHH might be involved in CSC migration and metastasis leading to poor prognosis of patients with OSCC.
In this context, the new aspect of our study was to investigate relationships between SHH and known CSC markers in OSCC. For this purpose, we evaluated expression of three known CSC markers: CD44, CD133 and SOX2.
Growing evidence indicates that CD44 expression is related to the progression and metastasis of many tumor types including OSCC (9, 10). Similarly to Saghravanian et al. (39), we observed that increased CD44 expression was correlated with more advanced clinical stage of OSCC. Previous reports suggested that CD44 expression in OSCC tissue might characterize a subpopulation of OSCC cells with stemness traits which possess high tumorigenic potential and drug resistance (10, 40). It needs to be highlighted that experimental study by Yan et al. showed that CD44 expression is imperative for preservation of the CSC phenotype (41). The correlation between CD44 expression and worse prognosis of OSCC revealed in the current study is similar to previous data (42). Our results suggest that CD44 expression may be defined as independent prognostic biomarker in OSCC and that CD44 expression is needed for preservation of CSC subpopulation in OSCC.
Similarly to other authors (11), we assumed that oral carcinoma stem cells were identified by CD133 biomarker expression in OSCC tissue. The lack of correlation between CD133 expression and clinicopathological parameters was observed in our study. Nevertheless, there are other reports suggesting that it is still unclear whether CD133 expression is correlated with clinicopathological parameters of OSCC and is associated with aggressive biological behavior of OSCC (11). In addition, there is study indicating that OSCC cells which share certain characteristics of CSCs harbor CD133-positive cells potentially responsible for OSCC aggressiveness (11). According to previous reports, we might conclude that identification of CD133-positive OSCC in current study might reflect increased self-renewal and higher levels of stemness genes and tumorigenicity (43).
In the current study, SOX2 expression observed in OSCC was comparable to previous reports (6, 16). In accordance with other authors, our findings revealed SOX2 expression to be significantly associated with lymphatic metastasis, indicating that SOX2 is involved in OSCC progression (16, 44). Furthermore, high SOX2 expression was closely related to shorter survival, consistent with other studies (15, 45). However, contrary to our results, Zullig et al. (46) and Fu et al. (47) observed better prognosis of patients with OSCC with increased SOX2 expression.
To the best of our knowledge, this is the first study that investigated associations between SHH and CSC markers in patients with OSCC. Our study revealed correlations between SHH and selected markers of CSCs CD44, CD133, SOX2 in OSCC, which suggests that SHH may play an important role in the maintenance of CSCs in oral carcinoma. Yang et al. (48) found associations between another protein of the HH signaling pathway, glioma-associated oncogene 1 (GLI1), and CD44 and SOX9 in esophageal squamous cell carcinoma, which suggests that cross-talk between these proteins might reflect the pathway of CSC activation. According to published data about GLI1 expression in breast cancer by Ni et al. (49), the positive correlation observed in the current study between SHH and SOX2 expression suggests that components of the HH pathway might promote the expansion of CSCs in OSCC (50). It was found that the HH pathway is able to activate the transcription of key genes that contribute to the stem cell phenotype, such as NANOG, OCT4, SOX2, and B cell-specific Moloney murine leukemia virus integration site 1 (BMI1) (50). Positive correlation observed in our study between CD133 and SHH expression as well as between CD133 and SOX2 expression in OSCC suggests that SHH is strongly activated in tumor tissue with CSC phenotype (32). The biological and clinical behavior of OSCC might not only be related to correlation of SHH with stemness proteins but also with other proteins associated with oral CSCs. This might be supported by other authors that found positive correlation of OCT4 and NANOG with an increased expression of CD133 was associated with worse prognosis for patients with oral cancer (11). Therefore, it seems reasonable to assume that associations between proteins observed in our study may indicate that progression of OSCC requires cooperation between different molecules presented by CSCs. The HH pathway in oral CSCs may be a promising target in cancer therapy to reduce tumor progression, recurrence, and metastasis (50-53).
In conclusion, our study revealed that SHH is closely associated with CSC markers in OSCC, contributing to tumor progression and adverse outcomes of patients. Moreover, significant correlation between SHH and CSC markers suggests that SHH expression might be a potential marker of CSCs in OSCC. Further studies are required to determine specific interactions between SHH and CSCs in oral carcinoma. Based on these findings, patients with CSC phenotype OSCC may benefit from SHH-targeted immunotherapy.
Acknowledgements
The Authors would like to offer special thanks to Professor Michal Jelen for sharing specimens used during this study and his pathological assistance. This research was supported by Wroclaw Medical University (grant no. STM.B132.17.0007).
Footnotes
Authors’ Contributions
PC and JB designed the study. PC collected tissue samples and clinicopathological data. PC and AL-N performed laboratory work. PC and JB conducted the assessment of immunohistochemical samples. PC performed statistical analysis and interpreted data. PC and JB participated in results analysis and writing the article. All Authors have read and agreed to the published version of the article.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received August 18, 2021.
- Revision received October 10, 2021.
- Accepted October 11, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.