SALL4 stemness agent expression in oral squamous cell cancer and its clinical significance

Abstract Cancer stem cells are currently considered an important cause of tumoral heterogeneity, treatment failure and recurrence. SALL4 is a marker of stemness in many types of human cancers. In the present study, the expression of SALL4 was analysed in oral squamous cell carcinomas (OSCCs) using immunohistochemistry. This expression was linked to clinical and pathological characteristics of OSCC patients. Using tissue microarrays of OSCCs, a significant majority of these tumours were shown to positively express SALL4 protein in their cytoplasm and nuclei. This expression positively correlated with tumour grade and tumour location within the oral cavity. Furthermore, SALL4 expression was correlated with aberrant expression of beta-Catenin (AEB) in OSCCs. Morphologic evidence presents significant overlap between cells expressing the two proteins. Given the above, and in light of past evidence of SALL4 modulating the Wingless/Wnt/beta-Catenin pathway in other tumours, this study presents a possible role for the SALL4 protein in OSCC development and progression, and asserts the need for further studies into that role.


Introduction
Head and neck squamous cell cancer (HNSCC), of which oral squamous cell cancer is the most common, is one of the top ten most common malignancies in the world [1,2]. World-wide, this tumour has 5-year survival rates of less than 50%, with persistent treatment failures and frequent cases of relapse and recurrence [1,2]. This dismal prognosis is caused by locoregional recurrence at the primary site, late presentation, metastases in neck lymph nodes and low response to current treatment modalities [3]. One of the causes of failures of conventional chemo-, radiation and biologic therapies is inter-tumoral heterogeneity resulting in treatment failure and tumour progression [4]. Presence of cancer stem cells (CSCs) is an important cause for this intra-tumoral heterogeneity [4]. While the existence of CSCs in the past was disputed, it has now been confirmed to be present in a number of solid tumours including breast, brain, lung, colon cancers and melanoma [4,5]. CSCs can renew themselves, allowing cancerous tumours an unlimited reserve pool of undifferentiated cells that are able to resist therapy, and mount a second wave of recurrent tumours and metastases [6][7][8][9].
Identifying these CSCs and charting their mechanisms of actions is of prime importance in the management of malignant tumours [14]. Since stem cells are not present in most adult tissues, they could be ideal targets for cancer-specific diagnosis and treatment [15].
SALL4 is a C2H2-zinc finger transcription factor, similar to other stem cell factors such as SOX2, OCT4 and NANOG [16]. SALL4 is known to be a marker of cancer stem cells [17]. It plays a vital role in protecting the properties of embryonic stem cells, and regulates cell growth, proliferation and apoptosis in embryonic stem cells and cancer stem cells [17,18]. SALL4 expression is silenced in differentiated tissues [15]. There is evidence to suggest that a resurgence of SALL4 expression later in adult life could be associated with malignant transformation in different tissues [19]. This protein is also recognized as a potential biomarker for assessing cancer prognosis. A high expression of SALL4 is significantly associated with increased overall cancer mortality and recurrence [19,20]. Knocking down SALL4 resulted in the inhibition of tumour formation in xenograft tumour models [21]. The levels of expressions of this protein in the serum and tissues correlated with lymph node metastases, differentiation degree and tumour stage [22]. Furthermore, SALL4 was shown to induce the acquisition of properties of the EMT phenomenon and to promote invasion and metastases in endometrial cancer cells [23]. SALL4-knockdown restored chemotherapy sensitivity, reversed EMT and diminished cell metastases and suppressed the downregulation of E-cadherin and the upregulation of N-cadherin [23].
In the head and neck region, SALL4 has been shown to modulate the effect of radiation therapy on nasopharyngeal tumour cells, where inhibition of the SALL4 gene reduced proliferation, sensitized cells to radiation both in-vitro and in-vivo, increased radiation-induced DNA damage, increased apoptosis of tumour cells and resulted in tumour cell cycle arrest [24]. While expression of SALL4 in oral squamous cell cancers has been reported previously [25][26][27], quantifying that expression, correlating it with clinicopathologic characteristics and analysing its functional properties remain largely uncharted [18].
Functional deregulation of the Wnt signalling pathway can promote oral cancer development and progression [28]. Beta-Catenin, a known downstream target, and a key component of the canonical Wingless/Wnt signalling pathway [29], may have a role in the stemness of cancer cells, when it interacts with reprogramming genes KIf4, Oct4 and Sox2 [30]. This interaction will further enhance the expression of SALL4 during the induction of a somatic cell, and reprogramming it into a pluripotent stem cell [30]. However, little is known about the expression of SALL4 and beta-Catenin, and their relationship in oral squamous cell cancer (OSCC).
Considering the above, this study analysed the expression of the SALL4 and beta-Catenin proteins and assessed any possible correlations and interactions in OSCC.

Ethics statement
The study was approved by the Institutional Review Board (IRB) of the College of Medicine at Alfaisal University.

OSCC tissues
Following approval of the Institutional Review Board (IRB) of the College of Medicine at Alfaisal University, two commercial human oral microarrays (Catalogue no. OR601c, Biomaxc, US, Rockville, Maryland), with 50 cases of oral squamous cell carcinomas (OSCCs) and ten cases of non-neoplastic tissue, each, were used for this study. All cases were confirmed OSCCs. Major parameters of these tumours include age, sex, anatomic location within the oral cavity, pathologic diagnosis and grade.

Immunohistochemistry (IHC)
A rabbit polyclonal antibody against SALL4 (ab 29112; Abcam, UK), and a rabbit polyclonal antibody against beta-Catenin (ab 16051, Abcam, UK) were diluted in antibody diluent (Agilent, Santa Clara, California) per manufacturer's instructions; and applied to 5-mm sections from formalin-fixed, paraffin-embedded tissue specimens, using the avidin-biotin peroxidase method (Vectastatin Elite ABC kit; Vector Laboratories, Burlingame, CA), following the manufacturer's instructions. The immunohistochemical (IHC) stain was performed manually at room temperature. Negative controls were used with omission of primary antibody. Separate positive controls of mixed tumour (for SALL4) and appendix (for beta-Catenin) were used for test optimization and run validation.

Staining evaluation
Immunohistochemical staining of both markers, for all tumours, were evaluated by two pathologists (AO) and (ER), independently, and a consensus was reached for discordant cases. All tissues, normal controls, tissues adjacent to cancer or malignant tissue, were compared with the staining of the positive control. The degree of SALL4 positivity was scored by applying a semi-quantitative immunoreactivity scoring (IRS) system as described by Baccelli and co-workers [31]. The staining intensity was categorized into four grades: 0, no immunostaining; 1, weak staining (light yellow); 2, moderate staining (buffy); and 3, strong staining (brown). The percentage of positive cells was categorized into five grades: 0 (0%); 1, (1-10%); 2 (11-50%); 3 (51-80%); and 4 (>80%). The staining intensity and percentage of positive cells were multiplied to obtain an IRS for SALL4 expression, in the range of 0-12 for each individual case [31]. In line with studies done in other tissues [16,32], a case was scored as positive for SALL4 with an IRS of ≥3 defined as positive expression.
Beta-Catenin immunostaining was assessed using a previously described scoring method [33,34]. Staining was evaluated semi-quantitatively by comparing the intensity and cellular localization of the positive signals with those in the adjacent normal epithelium as an internal positive control. Staining was then recorded as normal or abnormal: Normal described uniform membranous staining, strong and similar to that in the internal controls. Partial loss (mixed areas of positive and negative cells with normal membranous staining) or complete loss (uniformly negative) of normal membranous staining, or altered cellular distribution in the cytoplasm and/or nucleus were defined as abnormal [33,34].

Statistical analysis
The χ 2 test was used to compare SALL4 expression and oral squamous cell cancer (N = 50). The χ 2 test was also used to examine the relationship between SALL4 positive cases (N = 37) and various clinical, demographical and pathological criteria (gender, age, tumour grade and tumour location). Furthermore, the χ 2 test was used to examine the relationship between beta-Catenin positive cases (N = 43) and tumour grade and location within the oral cavity; and to analyse the correlation between the two markers in cancer cells.
All statistical analyses were performed with IBM SPSS Statistics software package, version 25.0. p < 0.05 was considered to indicate a statistically significant difference.

Results
Two copies of the human microarray (Catalouge no. OR601c, Biomax, US, Rockville, MD) with cases of oral squamous cell carcinomas were used for this study. The age of the cancer patients in this array was in the range of 35 to 80, with a median of 53 years. The OSCC grades were distributed as follows: 27 patients are grade I; 19 patients are Grade II and 4 patients are grade III. The group comprised of 17 females and 33 males. The tumours included SCCs from the tongue (n = 31), the bucca cavioris (n = 8), gum (n = 4), lip (n = 4), face (n = 1), jaw (n = 1) and oral cavity (n = 1). Non-malignant cases in the microarray (n = 10), were all from the oral cavity, and consisted of 7 from cancer-adjacent normal tongue tissue and 3 from non-malignant mucous membrane of the maxilla. The expression of SALL4 in normal epithelial mucosal oral cells was negative in all normal samples (IRS#0-2). In OSCC cases, on the other hand, 37/50 cases were positive for SALL4 (74% with an IRS of ≥3, χ 2 = 11.52; p = 0.001, Figure 1A-C). The rest of the OSCC cases (13 in total) had weak or negative SALL4 expression (IRS# 0-2). All of the OSCC cases included in this study (37/37, 100% of cases) expressed SALL4 in the cytoplasm and nuclei of the cells (Arrow heads, Figure 1A-C, Table  1). The expression of beta-Catenin in normal oral mucosa was membranous and weak in intensity (Figure 2). The majority of OSCC cases had partial loss of membranous staining (thin arrows, Figure 3A,B), with abnormal expression of beta-catenin in the cytoplasm and nuclei of those cells (Arrow heads, Figure 3C,D).

Association between SALL4 immunohistochemical expression and clinic-pathological parameters of OSCCs
The association between SALL4 expression and clinical-pathological parameters was further analysed in the OSCC cases (Table 2). Positive expression for SALL4 was significantly correlated with OSCC grade (χ 2 = 11.243; p = 0.004). SALL4 expression also correlated with the OSCC location within the mouth. (χ 2 = 65.351; p < 001). There was no significant association between positive staining for SALL4 and other clinicopathological parameters, including age (p > 0.05) and sex (p > 0.05) ( Table 2).

Association between beta-Catenin immunohistochemical expression and OSCCs grade and location within the oral cavity
The association between beta-Catenin expression and clinical-pathological parameters was further analysed in the OSCC cases (Table 3). Positive staining for beta-Catenin was significantly correlated with tumour    grade (χ 2 = 13.29; p = 0.001). Beta-Catenin expression also correlated with the OSCC location within the mouth (χ 2 = 68; p < 0001).

Association between SALL4 and beta-Catenin
OSCC cases with positive expression for SALL4 were significantly associated with aberrant expression of beta-Catenin (84%, χ 2 = 16.892; p < 0.001, Table 4). The aberrant expression of beta-Catenin in those cases was in the form of either loss of membranous expression, or a loss of membranous expression with a switch to cytoplasmic or nuclear expression (Arrow heads Figure 3A-D).
In addition to the above, a strong morphological correlation could be clearly seen between the expression of SALL4 and beta-Catenin, when compared between tumours from the same positions on the array. Sections 4 A, 5 A, 6 A and 7 A were all stained with beta-Catenin, whereas their counterparts 4B, 5B, 6B and 7B were stained with SALL4. More cells expressed SALL4 (thick arrows, Figures 4B and 5B) than beta-Catenin, however, all the cells that expressed beta-Catenin also expressed SALL4 (thin arrows, Figures 4-7). This overlap is clearly evident, despite the higher intensity of SALL4 expression than beta-Catenin. The expression of both markers was predominantly in the cytoplasm and nuclei of the tumour cells (Arrow heads, Figures 1 and 3-7).   Figure 5. two representative sections taken from the same tumour, a grade 2 oScc from the Bucca cavioris, on the microarray. (a) beta-catenin staining; (B) Sall4 staining. more cells have stained with Sall4 than with beta-catenin (thick arrow, (B)). however, cells staining with beta-catenin were also stained with Sall4 (thin arrows, (a) and (B)). Both markers has predominant nuclear and cytoplasmic staining of tumour cells (arrow heads, (a) and (B)). (magnification X10).

Discussion
In this study, OSCC cells exhibited positive expression of Sal-like protein 4 (SALL4). SALL4 expression was significantly correlated with the grade of the OSCC cases and with location of these tumours within the oral cavity. While SALL4 expression was reported in oral cancer tissues in past work [25][26][27], none of the above studies quantified the expression of this protein in oral cancer tissues, analysed it across the clinical and pathological characteristics of patients, or linked it with the Wingless/Wnt pathway, a major pathway in oral tumorigenesis. SALL4 protein is a master regulator which contributes to cell stemness in biological development and tumour growth [17]. A cancer stem cell is currently an accepted explanation for the poor survival rates seen in cancer in general, and in oral squamous cell carcinomas in particular. The 5-year survival rates of oral cancer range between 15 and 50% [1,2]. Following initial treatment and cancer cell death, cancer stem cells provide a reserve pool, which may give rise to newly formed cancer cells, thus resurrecting the tumour (Satpute et al. 2013). The role of this sturdy group of cells goes beyond recurrence, and they are possibly involved in resisting conventional therapies, and metastases [35][36][37][38][39]. Identifying these cancer stem cells and analysing their properties are proving to be promising tools in the management of cancer [4].
SALL4 plays, initially, a central role in embryo-fetal development. SALL4 expression is then silenced in differentiated tissues. Evidence suggests that resurgence of SALL4 expression later in adult life is seen in several cancers; worsening the prognosis [19,20]. Levels of expressions in serum and tissues of the SALL4 protein highlighted a correlation with lymph node metastases, differentiation and tumour stage [22]. SALL4 has been shown to be over-expressed in many other tumours, including breast lung cancer (74.5% of cancer tissue at the mRNA level) [40], gastric cancer (two-fold or higher increase in 79.2% cancer tissue at the mRNA level) [41] and oesophageal squamous cell cancer (three-folds increased mRNA expression in 62% of cancer tissue when compared to normal) [42].
This study also found correlation between the expression of SALL4 and beta-Catenin. Beta-Catenin is a critical element of the Epithelial-Mesenchymal Transition (EMT) Phenomenon, and a key player in the canonical Wnt molecular pathway; an important process in oral carcinogenesis. Aberrant beta-Catenin expression (ABE) in the cytoplasm and nuclei, as seen in the OSCC cells of this study, represents the activated state of beta-Catenin [43,44].
The dual expression of the two proteins by OSCC cells, shows the close relationship between stemness, of which SALL4 is a representative, and the EMT process represented by beta-Catenin. Interestingly, knockdown of SALL4 resulted in reversal of the EMT phenomenon and reduced cellular invasion and metastases elsewhere in the body [23]. Other aspects of the interaction of the two proteins exist. In a normal state, SALL4 is directly regulated by the canonical Wnt signalling pathway, because within a minimal SALL4 promoter region of 31 bp exists a consensus TCF/LEF-binding site [45]. Targeting SALL4 in oesophageal squamous cell cancer cells resulted in decreased tumorigenicity involving EMT, through modulation of the Wnt/beta-Catenin pathway  Represents two sections taken from the same tumour, a grade 2 oScc from the tongue, on the microarray. (a) beta-catenin staining; (B) Sall4 staining. Significant overlap in the expression of Sall4 and beta-catenin in the same tumour cells could be seen (thin arrows, (a) and (B)). predominanty nuclear and cytoplasmic expression in the tumour cells for both markers (arrow heads, (a) and (B)) could be seen. (magnification X10). [46]. Furthermore, knockdown of SALL4 decreased the protein expression of Wnt3 and beta-Catenin in osteosarcoma cells [47]. And finally Chen and co-workers have shown that SALL4 may promote cell growth and cervical cancer development by upregulating the activity of the Wnt/beta-Catenin signalling pathway by directly binding to the CTNNB1 and trans-activating it [48]. The above, taken collectively, with evidence presented in this current study, may provide a platform to further explore the nature of the relationship between the two proteins in OSCC.

Conclusions
In summary, this study provided evidence of SALL4 positive expression in OSCC and correlated this expression with the tumour grade and location within the oral cavity. Furthermore, SALL4 expression correlated with that of beta-Catenin. This positive expression and the correlations presented, may shed some light on a possible role of this protein in oral carcinogenesis, and the possibility that it may be used in the diagnosis, prognosis and treatment of oral cancer.