Prognostic value of the immunohistochemical detection of epithelial-mesenchymal transition biomarkers in oral epithelial dysplasia: A systematic review

Background Oral potentially malignant disorders (OPMDs) comprise a range of clinical-pathological alterations that are frequently characterized as architectural and cytological derangements upon histological analysis. Epithelial-mesenchymal transition (EMT) has been proposed as a critical mechanism for the acquisition of the malignant phenotype in neoplastic epithelial processes. This study aims to systematically review the current findings on the immunohistochemical expression of epithelial-mesenchymal transition markers in oral potentially malignant disorders and to evaluate their possible application as biomarkers associated with the progression of oral epithelial dysplasias. Material and Methods A systematic search was performed in the following databases: PubMed, EMBASE, Chinese BioMedical Literature Database, and Cochrane Library. Articles that evaluated the relationship between the expression of EMT markers and the degree of oral epithelial dysplasia were selected for the systematic review. The quality of each eligible study was evaluated by independent reviewers that used operationalized prognostic biomarker reporting guidelines (REMARK). Results Seventeen articles met all inclusion criteria and were selected. The EMT markers analyzed exhibited an important association with the prognosis of the cases evaluated. The results showed a progressive increase in the expression of nuclear transcription factors and markers of mesenchymal differentiation, as well as negative regulation of epithelial and cell adhesion markers, according to the stage of oral epithelial dysplasia. Conclusions The dysregulation of expression of important EMT components in oral dysplastic epithelium is a potential prognostic marker in OPMDs. Key words:Oral potentially malignant disorder, oral epithelial dysplasia, epithelial-mesenchymal transition, biomarker, prognosis.


Introduction
Oral potentially malignant disorders (OPMDs) comprise a range of clinical-pathological alterations that are frequently characterized as architectural and cytological derangements upon histological analysis (1)(2)(3). OP-MDs exhibit an increased risk of malignant transformation. A recent meta-analysis estimated an overall risk of malignant transformation of 10.5% for oral epithelial dysplasia (4). The potential of malignant transformation is believed to be related to the degree of epithelial dysplasia observed (2)(3)(4). Several studies have been conducted to identify possible markers that can trigger the development of OPMDs as well as predict their progression (2,5,6). The epithelial-mesenchymal transition (EMT) is a mechanism frequently dysregulated in cancer (5)(6)(7)(8)(9)(10)12). However, studies evaluating the role of EMT in the development/progression of dysplastic processes in oral epithelium are scarce. The EMT, which is an essential event during embryogenesis, is a biological process in which epithelial cells lose their characteristics and shift to a mesenchymal cell-like phenotype. This process has been proposed as a critical mechanism for the acquisition of the malignant phenotype in neoplastic epithelial processes (5,8,9,12). EMT is mediated by nuclear transcription factors and can orchestrate intracellular alterations such as the negative regulation of epithelial markers, as well as the positive regulation of some mesenchymal markers (7,8,10). A variety of biomarkers are known to be associated with EMT. These biomarkers can be divided into five groups according to their characteristics/functions: 1. cell surface markers; 2. cytoskeletal markers; 3. extracellular proteins; 4. transcription factors; 5. epigenetic markers and microRNAs. 8 Recent studies indicate that important EMT markers may be involved in the malignant transformation of cases diagnosed as OPMD (5,11,12). In this study, a systematic review was conducted to clarify the prognostic value of EMT biomarkers in OPMDs. The review includes publications that evaluated the immunodetection of EMT markers and their possible use as prognostic factors/predictors in oral epithelial dysplasia.

Material and Methods
We performed a systematic review to conduct this investigation. The dependent variables were biomarkers of EMT available by immunohistochemistry and the Conclusions: The dysregulation of expression of important EMT components in oral dysplastic epithelium is a potential prognostic marker in OPMDs.
were defined as those dysplasias that developed cancer at the same site as the initial biopsy when followed over time. Biomarkers of EMT were defined according to the study by Zeisberg & Neilson (8).
In the screening of titles or abstracts, citations were retained if they were original studies, except for reviews or meeting reports, that explored associations between the immunoexpression of the EMT markers, clinicopathological characteristics, and outcome/progression of the oral epithelial dysplasia cases analyzed. The articles were selected independently by two reviewers (EFM and JCP). Any disagreement was resolved by consensus.
In the final full-text screening, studies were included if they met the following criteria: (I) designed as a prospective or retrospective cohort study; (II) oral epithelial dysplasia was diagnosed and histopathological grading was performed by pathological examination; (III) data about the association between the EMT biomarker evaluated and histopathological grading were reported. When studies were based on overlapping data, the more comprehensive set was selected.
-Data extraction and analysis Quality assessment was performed in duplicate for each eligible study by three independent reviewers using the Newcastle operationalized prognostic biomarker reporting guidelines (REMARK) (15). Any disagreement was resolved by consensus. Studies receiving a score of less than 6 were not included in the systematic review. Three reviewers (EFM, JASL, RPM) independently extracted the data from the selected articles.
-Risk of bias in individual studies Methodologically, the authors appraised all of the included studies according to a checklist based in Meta -Analysis of Statistics Assessment and Review Instrument (MAStARI) (17). Two reviewers (EFM and RPM) answered 9 questions for descriptive studies as Y for "yes", N for "no", U for "unclear" and NA for "not applicable". After that, the risk of bias was categorized as high when the study reached up to 49% of a "yes" score, moderate when the study reached 50% to 69% of a "yes" score, and low when the study reached more than 70% of a "yes" score. Disagreements were solved by discussion between the four authors (EFM, RPM, JCP, JASL).

-Study selection and characteristics
The database searches and manual search developed in this systematic review retrieved 378 studies. After reading the titles and abstracts, 54 studies were considered potentially eligible and their full text was read by three reviewers. Seventeen articles met all inclusion criteria and were selected for this systematic review (5,(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33). Regarding methodological characteristics, all articles included in this systematic review involved 1,178 patients, with a mean of 69.2 participants per study. Our analysis revealed that 743 (63%) of the cases evaluated were lesions diagnosed in different stages of epithelial dysplasia. Among these cases, 48.7% were graded as mild epithelial dysplasia and 51.3% as moderate or severe epithelial dysplasia. The main characteristics and findings of the studies are shown in Table 1.
Group I (n = 21) -mild ED Group II (n = 15) -moderate ED Group III (n = 13) -severe ED NE NE β-catenin γ-catenin Tissue (F) β-and γ-catenin were differently expressed according to dysplasia grade. As the degree of dysplasia became more severe, it was observed a reduction of catenins expression, loss of the exclusive membranar expression and cytoplasmic delocalization. Tissue (F) Cytokeratin-13 loss was found to be an independent diagnostic marker for high-grade dysplasia.
Cytokeratin-13 loss may be useful for distinguishing EDs from reactive atypical epithelium.    (30), and N-cadherin (31). The methods used for immunohistochemical analysis in the different studies selected were identified in this systematic review ( Table 2). Immunoreactivity for E-cadherin, β-catenin, and APC was assessed using a semiquantitative method based on the proportion of positive cells and staining intensity: Immunoreactivity score = proportion positive score x intensity score. The proportion of positive staining was scored as follows: 0 (0-20% of epithelial cells stained positive), 1 (21-40% of positive cells), 2 (41-60% of positive cells), 3 (61-80% of positive cells), or 4 (>80% of positive cells). The intensity was scored as: 0 (negative staining), 1 (weak intensity), 2 (moderate intensity), or 3 (strong intensity). Immunoreactivity for vimentin was assessed using the Label Index (Li) score, which presents the proportion of positively-stained cells as a percentage. Inoue et al.
Podoplanin The intensity of immunoexpression was scored as follows: 0 (no expression in any part of the epithelium); 1 (intensity weaker than that in lymphatic vessels as a positive control, and expression observed in the basal layer; 2 (intensity weaker than that in lymphatic vessels, and expression was observed in more than two layers; 3 (intensity equal to that in lymphatic vessels, and expression observed in basal layer); and 4 (intensity equal to that in lymphatic vessels, and expression observed in more than 2 layers. The score was based on examination of the whole section. Silva et al.

Twist p-Akt
Five histological fields were randomly chosen, and 1000 cells were counted in each slide. The analysis was performed by counting positive cells in cytoplasmic and/or nuclear compartments.
Rani et al.
Laminin-5 The extent of positive immunostaining was calculated by subtracting the negative control strongly positive cytoplasmic or nuclear pixel count from the matching area with a strongly positive cytoplasmic or nuclear pixel count on the study slides. This resultant normalized positive immunohistochemistry pixel count was divided by the area (square millimeters) to provide a comparable indicator of immunopositivity among individual tissue samples. Anura et al. (2014) E-cadherin The epithelium was demarcated into proliferative and differentiative layers to evaluate the differential expression pattern of the molecules using intensity scores. To evaluate E-cadherin expression, random points corresponding to each layer were selected from randomly selected microphotographs of the studied groups. Expression intensity scores were given to each point in the range of '0-10', where: '10-8' (intensely deep color), '7-6' (deep color), '5-3' (faint) and '2-0' (very faint color). The intensity of Ecadherin expression in cell membrane and cytoplasm was scored separately. de Freitas Silva et al.

E-cadherin Twist
The analysis of Twist immunoexpression was performed taking in count the percentage of positive staining cells in the whole examined area. For Twist evaluation, the percentage of positive cells was classified into four categories: 1 (<25%); 2 (25%-50%); 3 (51%-75%); 4 (>75%). For E-cadherin, immunohistochemical staining was evaluated as: 1 (less than 25% positive cells); 2 (heterogeneous staining in 25%-50% of cells); 3 (homogenous staining in 50%-75% of cells); 4 (strong staining in more than 75% of cells). β-catenin The cellular localization of β-catenin was classified as membranar, cytoplasmic or nuclear, depending on the immunolocalization pattern. Intensity of membranar staining was assessed as mild, moderate or intense. Five randomly selected fields with presence of dysplastic and/or neoplastic epithelium were assessed. In these photomicrographs, 1000 epithelial cells were counted per case and the number of positive cells was determined with the Image J software. Silva et al.
β-catenin A qualitative β-catenin analysis was performed based on the percentage of positive cells, considering as positive only specimens with 25% or more cells presenting β-catenin expression. The cellular localization (membranar, cytoplasmic and nuclear) and the epithelial layer (basal, parabasal, and spinous cell layers) were also considered. β-catenin immunoexpression was analyzed by two independent blind and calibrated examiners under light microscopy at 200-fold magnification. CK13 was scored as follows: 0, strong and diffuse expression; 1, weak and/or patchy expression; or 2, no expression. CK17 was scored as follows: 0, no expression; 1, weak and/or patchy expression; or 2, strong and diffuse expression. Discrepant slides were re-evaluated using a dual vision microscope to achieve a consensus. Chandolia et al.  -Quality assessment and Risk of bias We used the criteria established by the REMARK guidelines to evaluate the studies included in the systematic review ( Table 3). The selected studies provided details about the objective/hypothesis of the study, characteristics of the patients included in the sample, analysis method used, and relationship of the EMT markers with the degree of epithelial dysplasia in the cases analyzed. In addition, the studies discussed implications for future studies and clinical value. Based on the MAStARI assessment, 2 articles (18,32) were classified as carrying a high risk of bias, mainly because the answers for questions 3 and 4 (related to co -founding factors, description of the groups and follow-up, respectively) were "No". Fifteen studies were classified as with low risk for bias (5,(19)(20)(21)(22)25,(26)(27)(28)(29)(30)(31)33) (Table 4).

Checklist Criteria Introduction
The hypotheses and objectives of the study were clearly explained.

Cohort description
Retrospective or prospective cohort with a well-defined study population. Medical treatment of the cases was explained.

Patient data
The basic data such as age, gender, clinical stage and histopathologic grade was provided. Evaluation method Well-described method including the microscopic field/s and the cutoff point. Inter-observer variability was evaluated.

Prognostic analysis
The survival end point was defined and/or the relationship between the tumour budding and lymph node metastasis was studied.

Statistical analysis
Estimated effect (e.g., hazard ratio, relative risk with their confidence interval), which reveal the relationship between tumour budding and the survival end point/s The independence of prognostic value was reported by multivariate analysis Classical prognostic factors The prognostic value of the classical prognostic factors (e.g., stage and grade) were reported.
The relationship between tumour budding and classical prognostic factors was reported.

Interpretation of the prognostic value and discussion
Comparison of the current findings with other studies. Strengths and limitations of the current data.
Recommendation for further research.

Author (year) Questions* % Yes # Risk of Bias
. Is the study based on a random or pseudorandom sample? Q2. Are the criteria for inclusion in the sample clearly defined? Q3. Are confounding factors identified and strategies to deal with them stated? Q4. Are outcomes assessed using objective criteria? Q5. If comparisons are being made, was there sufficient description of the groups? Q6. Is follow up carried out over a sufficient time period? Q7. Are the outcomes of people who withdrew described and included in the analysis? Q8. Are outcomes measured in a reliable way? Q9. Is appropriate statistical analysis used?
*Y=Yes, N=No, NA=Not applicable (which was not considered on the percentage calculation) # Risk of bias was categorized as high (H) when the study reaches up to 49% score "yes", moderate (M) when the study reached 50% to 69% score "yes", and low (L) when the study reached more than 70% score "yes". Table 3: Evaluation criteria used to assess the quality of studies evaluated (adapted from REMARK guidelines). Table 4: Analysis of the risk of bias of the articles included in the review was performed with the MAStARI (Meta-Analysis of Statistics Assessment and Review Instrument) critical appraisal tool.
In our systematic review, β-catenin was a biomarker strongly associated with the progression of oral epithelial dysplasia. Ishida et al. (18) detected strong immunoexpression of β-catenin in the cell nucleus of cases diagnosed as oral epithelial dysplasia, with a significant difference between dysplasia samples and normal oral epithelium (P < 0.01) and between oral leukoplakia with dysplasia and the group without dysplastic alterations ( and in the groups with mild and moderate dysplasia, exhibiting a predominant cytoplasmic localization. However, in cases of severe epithelial dysplasia, immunoexpression was also detected in the superficial layers of the epithelium, indicating dysregulation of protein expression in advanced stages of epithelial dysplasia. The authors also showed a statistically significant correlation in the immunoexpression of Twist and E-cadherin, with the observation of an inverse relationship between the immunoexpression of these proteins (r = 0.512; P < 0.001). Rani et al. (24) observed low expression of laminin-5 in cases of epithelial dysplasia, with predominantly weak and cytoplasmic immunostaining. In that study, the expression of laminin-5 did not differ significantly according to the degree of dysplasia. A significant increase in laminin-5 immunopositivity was only found in cases of OSCC compared to the epithelial dysplasia groups, suggesting a role of this protein in the invasion of the already established neoplastic process.

Discussion
The present study analyzed the role of EMT in cases diagnosed with oral epithelial dysplasia. The malignant transformation of OPMDs is an important process in oral carcinogenesis, which is still poorly understood. Thus, the identification of markers that are associated with the development and progression of oral epithelial dysplasia is of paramount importance for establishing a possible relationship with the prognosis of these lesions.
The histopathological classification of oral epithelial dysplasia can assist in the monitoring and definition of the most appropriate treatment; however, other factors may be determinant in the process of malignant transformation (34,35). Several biomarkers are being analyzed by immunohistochemistry in order to evaluate their participation in oral epithelial dysplasia (3,6). In clinicopathological practice, immunohistochemistry is relatively easy to apply and more accessible than other techniques. The EMT has been indicated as an important process in carcinogenesis that is directly associated with the aggressiveness of OSCC. During this process, epithelial cells lose their capacity of cell-cell adhesion. In addition, reorganization of the cytoskeleton and significant changes in signaling occur that define the shape and structure of the neoplastic cell. EMT is believed to be the result of reprogramming of gene expression mediated by transcription factors such as Twist, Snail, and Slug (10,(36)(37)(38). This transition increases cell motility and enables the development of an invasive phenotype (37). Despite advances in the understanding of the function of EMT in already established OSCC, it is necessary to identify the role of EMT markers in the development and progression of oral epithelial dysplasia.
The loss of cell adhesion is likely to play a key role in EMT. This phenomenon can be observed in our systematic review by the loss of membrane expression of cell adhesion markers such as claudin-1 and, particularly, E-cadherin in cases diagnosed as epithelial dysplasia (5,10,21,26,27,32,33,36,38). Furthermore, E-cadherin was immunoexpressed in the cytoplasm of epithelial dysplasia cases. In this respect, studies have shown that the cytoplasmic expression of this protein is frequently associated with tumors in advanced stages (39,40). It is possible that the loss of membrane expression and the onset of E-cadherin translocation to the cytoplasm, and consequently the loss of basic functions associated with cell adhesion, occur during the early stages of oral carcinogenesis and are already present in cases of oral epithelial dysplasia, progressing with increasing severity of the latter. Twist is an important nuclear transcription factor for EMT. De Freitas Silva et al. (5) indicated a repressive effect of Twist on the expression of E-cadherin in epithelial dysplasia and oral cancer. Furthermore, previous studies have elucidated the role of Twist in the already established neoplastic process, which negatively regulates the expression of E-cadherin and the overexpression of mesenchymal markers (5,10,37). Other transcription factors with a potential role in the development and progression of oral epithelial dysplasia need to be better investigated. Zheng et al. (41) reported the overexpression of the transcription factors Snail and Slug to be associated with a poor prognosis of the neoplastic process. The Twist, Snail and Slug proteins seem to a play a similar role in the repression of E-cadherin and, apparently, the expression of these proteins is also related to the regulation of β-catenin (42). The binding of E-cadherin to β-catenin in the cytoplasm/membrane represses tumor progression, maintaining cell-cell adhesion and inhibiting EMT, cell motility and tumor metastasis (42). The negative regulation or loss of E-cadherin and β-catenin expression, as well as the immunoexpression of β-catenin in the nucleus, is frequently observed in several type of cancer, including head and neck cancer. In our systematic review, the nuclear expression of β-catenin was a frequent finding in the samples of cases diagnosed as advanced oral epithelial dysplasia and might be an important marker associated with the progression of this condition (18,21,28). Prgomet, Andersson, & Lindberg44 evaluated the expression of WNT5A, β-catenin and E-cadherin by immunohistochemistry in 21 tissue samples. Each sample contained areas of mucosa with normal appearance, oral epithelial dysplasia, and OSCC. In that study, membrane expression of β-catenin was lower in OSCCs than in dysplasia or normal-appearing mucosa regions, while cytoplasmic expression of β-catenin increased with the severity of dysplasia and was detected in half of the OSCCs. Similar findings of reduced membrane expression of β-catenin and progression of oral carcinogenesis were reported in the studies included in our systematic review. The cytoplasmic accumulation and subsequent nuclear translocation of β-catenin might be the result of activation of the canonical Wnt signaling pathway or impairment of this pathway due to mutations in some of its components. The cytoplasmic/nuclear overexpression of β-catenin is known to be associated with malignant transformation in different types of cancer (28,45).

Conclusions
In conclusion, the dysregulation of expression of EMTassociated proteins in dysplastic oral epithelium is a potential prognostic marker. The results showed a progressive increase in the expression of nuclear transcription factors and markers of mesenchymal differentiation, as well as negative regulation of epithelial and cell adhesion markers, according to the severity of oral epithelial dysplasia. It should be noted that the studies included in this systematic review examined OPMDs that were not continuously followed up from the time of diagnosis to malignant transformation. Longitudinal studies are necessary to identify the possible association of EMT with the risk of malignant transformation of OPMDs.