Objectives: To investigate clinical implications of epithelial mesenchymal transition (EMT) expression in oral cavity squamous cell carcinoma (OSCC). Materials and Methods: E-cadherin and vimentin expression was studied in 50 newly diagnosed cases of OSCC who underwent surgical excision. EMT expression at non cold spot infiltrative margin and cold spot was studied and correlated with prognostic factors and disease-free survival (DFS). Results: EMT expression at the cold spot and non-cold spot infiltrative margin showed significant results with nodal status (P < 0.001, P < 0.009 respectively). On multivariate analysis, only EMT at the cold spot correlated significantly with prognostic factors (P < 0.030). The factors affecting DFS on Kaplan Meier index were EMT expression and differentiation (P < 0.002, P < 0.016 respectively) which proved significant in cox regression analysis. Conclusion: The study reveals that EMT expression at the cold spot is a significant biomarker for predicting lymph-node metastasis and tumor recurrence in OSCC.
Keywords: Cold spot assessment, epithelial mesenchymal transition, oral cavity squamous cell carcinoma
|How to cite this article:
Rajashekar K, Khonglah Y, Raphael V, Shunyu B. Cold spot assessment in Epithelial-Mesenchymal Transition: a histological predictor of tumor recurrence in oral cavity squamous cell carcinoma. Indian J Pathol Microbiol 2022;65:274-9
|How to cite this URL:
Rajashekar K, Khonglah Y, Raphael V, Shunyu B. Cold spot assessment in Epithelial-Mesenchymal Transition: a histological predictor of tumor recurrence in oral cavity squamous cell carcinoma. Indian J Pathol Microbiol [serial online] 2022 [cited 2022 May 4];65:274-9. Available from: https://www.ijpmonline.org/text.asp?2022/65/2/274/343210
Head and neck squamous cell carcinomas rank sixth among the most common cancers worldwide. Oral cavity squamous cell carcinoma (OSCC) is one of the most prevalent tumors of the head and neck region and constitutes more than 90% of all oral cancers.
Oral cancer is one of the deadliest human cancers due to its aggressiveness and high mortality rate. Metastatic behavior is critical for survival. Patients of OSCC with distant metastasis have a 5-year survival rate that is three times less than that of patients with adjacent lymph node metastasis. Hence, accurate prediction of metastasis in OSCC would have an immediate clinical impact in the assessment of treatment.
Emerging evidence reveals that epithelial to mesenchymal transition (EMT) is linked to acquisition of the invasive and metastatic potential.,,,, EMT implies loss of non-motile parent epithelial cells and acquisition of motile mesenchymal cells. Loss of E-cadherin expression, which leads to the reduced cell to cell adhesion, and elevated levels of mesenchymal markers, such as vimentin, are distinctive events in EMT and are commonly observed in metastatic carcinomas.
This study is done to analyze the EMT behavior in OSCC and to compare it with the outcome for prognostication of the patients. It is also the first study attempting to assess the EMT phenotype in relation to tumor heterogeneity.
|Materials and Methods|
The medical records of patients treated for primary oral cavity squamous cell carcinoma were reviewed from January 2015 to June 2017. Eligibility criteria included the following: (1) oral cavity cancer, (2) squamous cell carcinoma, (3) patients who underwent surgical excision of tumor as initial treatment. Patients were excluded if they had second primary at the time of the presentation or were subjected to prior chemotherapy/radiotherapy before surgery. In total, 50 patients (29 men, 21 women; mean age 51.3 years, range 27–72 years) were enrolled. Postoperative radiotherapy was performed in 21 cases. Follow-up was available for 32 cases. The following prognostic factors data was obtained from the histopathological report: Histological grade/differentiation, pTNM staging according to American Joint Committee on Cancer (AJCC) classification, stage grouping, depth of invasion, resected margin, perineural invasion, and lymphovascular invasion.
Representative tumor area selection and Immunohistochemical staining
The hematoxylin and eosin sections of the tumor were studied and the infiltrative margin was marked under microscopy. The paraffin embedded biopsy blocks were retrieved from our pathology archives. The paraffin embedded sections were subjected to immunohistochemical staining (IHC) using E-cadherin (ready to use; BioGenex, Fremont, CA. REF: AM390-5M, LOT: AM3900517) and Vimentin (ready to use; BioGenex, Fremont, CA. REF: AM074-5M, LOT: AM0740616).
Immunohistochemical staining staining assessment
Variable expression of epithelial and mesenchymal markers within tumors was noted in oral cavity squamous cell carcinoma. This variability in expression can be attributed by intra tumoral heterogeneity of OSCC. Considering tumor heterogeneity and to reduce discrepancy in interpretation in results it was decided to assess areas where there is the maximum loss of E-cadherin expression at the infiltrative margin and the corresponding Vimentin expression was noted. We coined the term cold spot for this representative region. [Cold spot is defined as follows in this study: Cold spot represents the area in the infiltrative margin where there is maximum loss of E-cadherin expression. The rest of the margin, which showed some E-cadherin expression, was termed as non-cold spot infiltrative margin. The expression of E-cadherin and vimentin in 100 cells was manually observed under 40x objective in this region.
To reduce interobserver variation E-cadherin and vimentin expressions were determined by three independent observers without the knowledge of the patient’s data. Considering tumor heterogeneity, the pattern of E-cadherin and vimentin expression of each slide under light microscopy was studied as follows: (1) Study of E-cadherin and vimentin expression in overall infiltrative margin, (2) study of E-cadherin and vimentin expression in cold spot, (3) scoring of immunostain by immunoreactivity score (IRS), and (4) assessment of EMT phenotype.
Scoring of immunostain
E-cadherin stains the cell membrane, whereas vimentin stains the cytoplasm. Immunostaining is graded by an IRS. IRS was calculated by multiplying staining intensity (SI) and percentage of positive cells (PP), [IRS = SI x PP]. SI is assessed to be negative (=0), weak (=1), moderate (=2), or strong (=3). The reactivity is determined by the percentage of PP. Negative specimens had a score of 0. Score of 1 is given when 1–10% are positive cells, score of 2 when 11–30% are positive, score of 3 when 31–50% are positive, score of 4 when 51–80% are positive, and score of 5 when >80% are positive. Staining of E-cadherin and vimentin was individually defined as low (IRS = 0–4), medium (IRS = 5–9), or high (IRS = 10–15).
Assessment of epithelial mesenchymal transition phenotype
According to epithelial or mesenchymal character, the tumor was categorized as follows: complete EMT phenotype (vimentin + ve and E-cadherin -ve), incomplete EMT phenotype (vimentin -ve/+ve and E-cadherin -ve/+ve), epithelial EMT phenotype (vimentin -ve and E-cadherin +ve).
The study was approved by the Institute Research and Ethics Committee (T-346/17/086).
To determine statistically significant results between EMT phenotype (both at the noncold spot infiltrative margin and cold spot) with prognostic factors using Chi-square, Fischer’s exact test and multivariate analysis. DFS rates were calculated by Kaplan-Meier method, and multivariate analysis performed by Cox regression analysis using variables shown to be significant in univariate analysis (Kaplan-Meier method). All the analyses were performed using Statistical Package for the Social Science (SPSS) version 24 software (SPSS, Chicago, IL), and significance threshold set at P value less than 0.05.
Immunoreactivity score score and epithelial mesenchymal transition expression
E-cadherin and vimentin expression has been assessed by three independent observers in all 50 cases using an IRS. On averaging the results, the E-cadherin and vimentin expression pattern is described in [Table 1]. It was observed that there is an increase in the percentage of cases with loss of expression in E-cadherin and gain of expression in vimentin at the cold spot in comparison with the entire infiltrative margin. This was further confirmed by increased expression of complete EMT phenotype at the cold spot than in the non-cold spot infiltrative margin in [Table 2]. EMT expression was calculated from the IRS score of E-cadherin and vimentin, explained above. Epithelial and incomplete phenotypes were considered negative for EMT expression and complete EMT phenotype was considered positive for EMT expression [Figure 1], [Figure 2], [Figure 3].
|Figure 1: Epithelial phenotype (a) Positive E-cadherin expression and (b) negative vimentin expression. (IHC 10x) E-cadherin shows homogenous staining of cell membrane and vimentin shows cytoplasmic staining|
|Figure 2: Incomplete phenotype (a) negative E-cadherin expression and (b) negative vimentin expression (IHC 10x)|
|Figure 3: Complete phenotype (a) negative E-cadherin expression and (b) positive vimentin expression in epithelial cells (IHC 10x)|
The relationship of EMT phenotype at the non-cold spot infiltrative margin and cold spot with prognostic factors were assessed. EMT phenotype at the cold spot showed significance with nodal status and tumor differentiation (P = 0.003, P = 0.005, respectively) [Table 3]. EMT phenotype at the infiltrative margin showed significance only with nodal status (P = 0.019) [Table 4].
|Table 4: EMT expression in non-cold spot infiltrative margin with prognostic factors|
In multivariate analysis only EMT at the cold spot remained significant with prognostic factors (P = 0.030), whereas EMT at the non-cold spot infiltrative margin was not significant (P = 0.184) [Table 5].
Follow-up was available for 32 cases and the rest of 18 cases were lost to follow-up. The mean follow-up period for 32 cases was 29.5 months (range, 13-42 months). A total of 14 out of 32 cases showed recurrence (13 cases local recurrence, 1 case distant metastasis). Overall, 3-year DFS was 56.3%. On Kaplan–Meier index, complete EMT expression at the non-cold spot infiltrative margin (P < 0.007), cold spot (P < 0.002), and differentiation (P < 0.016) were significantly associated with DFS [Table 6]. The variables shown to be significant in univariate analysis (Kaplan-Meier method) are subjected to multivariate analysis to assess DFS. Multivariate analysis by Cox regression showed significant results both with EMT at cold spot and tumor differentiation (P < 0.012) and EMT at non-cold spot infiltrative margin and tumor differentiation (P < 0.014). However, the significance was relatively higher at the cold spot in comparison to the non-cold spot infiltrative margin.
Note: One case of primary spindle cell squamous cell carcinoma was observed. It showed a complete EMT phenotype throughout the tumor. (E-cadherin was lost and vimentin was expressed).
OSCC displays a wide range of metastatic behavior that cannot be predicted by tumor size, standard histology, or even individual gene or protein expression/activity. Study of tumor markers expression for the prediction of metastatic behavior plays a significant role in treatment modality. EMT activation plays a critical role in acquisition of malignant phenotype, which is important for migration and metastasis of cancer cells. The concept of EMT associated with acquisition of malignant phenotype is based on observations of various biomarkers like cell surface proteins, mesenchymal markers, transcription factors, and micro-RNA. The loss of epithelial cell adhesion molecule, such as E-cadherin and acquisition of mesenchymal markers, such as vimentin and S-100 is associated with increased metastatic potential.,,,,, Identification of such tumor markers has increased our understanding of the pathogenesis of head and neck cancer and provided insight into tumor biology, prognosis, and response to therapy.
Study of E-cadherin and vimentin in OSCC was significantly associated with lymph node metastasis, tissue location, differentiation, and other prognostic factors.,,,,,, Head and neck SCC patients with EMT phenotype have a significantly higher incidence of distant metastasis compared to patient lacking this phenotype. EMT at the invasive front is analyzed in many studies,, but the study of EMT biomarkers given tumor heterogeneity in OSCC was not found.
In the present study, EMT phenotype expressions were studied at the entire infiltrative margin correlated with prognostic factors. Each tumor infiltrative margin was studied to analyze the pattern of E-cadherin and vimentin expression. As previously mentioned three patterns of IHC were observed, (1) epithelial phenotype where E-cadherin was retained in the tumor cells and vimentin was negative favoring epithelial nature of tumor cells, (2) incomplete EMT phenotype where both E-cadherin and vimentin expression is present (favoring type II EMT – seen in tissue regeneration and fibrosis) or both E-cadherin and vimentin are negative, (3) complete EMT phenotype, E-cadherin was lost in tumor cells and vimentin was acquired (type III EMT favors cancer progression and metastasis). These patterns of IHC expression were variable in different tumors and it was noticed that cases generally considered as negative for EMT expression had focal areas which expressed a complete EMT phenotype. To understand tumor heterogeneity expression in OSCC, these focal expressions were termed as cold spots and its significance with prognostic factors analyzed. An arbitrary cut off of at least 100 cells with EMT expression was required to call the focal expression as cold spot.
In the present study, EMT phenotype expressions were studied at the non-cold spot infiltrative margin and cold spot and correlated with prognostic factors. Complete EMT phenotype was observed in 14 cases in non-cold spot infiltrative margin and in 24 cases at cold spot. EMT phenotype at the non-cold spot infiltrative margin and cold spot showed significance with nodal status, suggesting that EMT can be used as a biomarker in assessing nodal metastasis in early lesions. In addition, EMT at the cold spot also showed significance with differentiation. This significance was attributed to moderately differentiated lesions which exhibited complete phenotype only in 7 cases at the non-cold spot infiltrative margin in contrast to 15 cases at cold spot. EMT at the cold spot is a better marker in assessing the malignant potential of moderately differentiated lesions.
In multivariate analysis only EMT at cold spot showed significance with all prognostic factors, proving to be a better biomarker than the entire infiltrative margin in risk assessment of OSCC.
Factors affecting DFS were EMT phenotype (at non-cold spot infiltrative margin and cold spot) and differentiation. Although both EMT at the non-cold spot infiltrative margin and cold spot showed significance with DFS, in 14 cases with recurrence complete EMT phenotype was observed in 13 cases at cold spot and 9 cases at the non-cold spot infiltrative margin. On studying the differentiation pattern in the 14 cases with recurrence, 1 well differentiated, 3 poorly differentiated and 10 moderately differentiated lesions were observed. One well-differentiated lesion and 3 poorly differentiated lesions expressed complete EMT phenotype both at cold spot and non-cold spot infiltrative margins. In moderately differentiated lesions, 5 out of 10 cases expressed complete EMT at non cold spot infiltrative margin and 9 out of 10 cases expressed complete EMT in cold spot. Although the number of recurrent cases in well and poorly differentiated lesions were low in this study, which could lead to sampling bias, it was interesting to note that in moderately differentiated lesions complete phenotype expression at cold spot (9 out of 10 cases) was higher than in non-cold spot infiltrative margin (5 out of 10 cases). These data again suggest the importance of assessment of cold spot in prognosis of moderately differentiated lesions.
One case of well-differentiated carcinoma with recurrence, TNM staging of T3N0M0, with no perineural and lymphovascular invasion, but which had complete EMT phenotype emphasizes the importance of EMT in assessing early lesions. In 14 recurrent cases nodal status was positive only in 7 cases, EMT at cold spot was positive in 13 cases. These results stress the importance of assessment of EMT in analyzing early lesions where the standard prognostic markers fail to guide patient management. However, to assess the significance of EMT and to derive a conclusion in relation to recurrence in well-differentiated carcinomas, a larger and diverse study population needs to be analyzed.
Limitations of the study
The study size includes only fifty patients and follow up was available for only 32 cases. Cold spot was analyzed with an arbitrary value of 100 cells in current study, it’s essential to understand what percentage of cells with focal expression of EMT phenotype can affect the prognosis of the patient in low and moderately differentiated lines. Due to the loss of follow up of 18 cases, EMT assessment with post-surgery radio-therapy was not analyzed in the study. Hence, more studies are required to further validate these results.
In conclusion, EMT assessment in initial biopsy specimens can help in identifying high risk patients. Tumor heterogeneity, which is an important parameter, needs to be considered while assessing EMT.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
Sommers CL, Walker-Jones D, Heckford SE, Worland P, Valverius E, Clark R, et al. Vimentin rather than keratin expression in some hormone-independent breast cancer cell lines and in oncogene-trans- confers resistance to apoptosis in three murine formed mammary epithelial cells. Cancer Res 1989;49:4258–63.
Source of Support: None, Conflict of Interest: None