Expression of EZH2 in uveal melanomas patients and associations with prognosis

Purpose To analyze the prognostic value and potential target for therapeutic intervention of enhancer of zeste homologue 2 (EZH2) in uveal melanomas (UM) patients. Method We analyzed EZH2 expression in 89 primary UM patients by immuno- histochemistry to observe the clinicopathological and prognostic value of EZH2. Results The high levels of mitoses count and Ki67 labeling index had significant correlation with overexpression EZH2 (R = 0.408, P Conclusion Our critical finding is that overexpression EZH2 in UM can be served as predictive marker and is associated with adverse clinical outcomes. Further observation of EZH2 as a potential therapeutic target in UM is necessary.


INTRODUCTION
Uveal melanoma (UM) is the most common primary intraocular malignant tumor in adults. UM eventually spread to the liver in up to 50% of patients and nearly half of the patients have subclinical metastasis at the time of diagnosis [1,2]. There are several treatment measures on UM in clinic nowadays, the main alternatives include enucleation, proton beam radiotherapy and plaque radiotherapy. Although the progress and availability of alternative therapeutic models, the survival rates of UM patients are nearly unchanged in 30 years [3]. Once metastasis occurs, median survival is only 6 month with or without treatment [4].
The increased risk of developing metastatic disease in UM was reported to be associated with clinic and morphological factors such as larger tumor size, presence of epithelioid cells, closed vascular loops, mitotic activity, nodular growth and extracellular matrix patterns [5][6][7][8]. Molecular studies have also shown that cluster differentiation could be made, classifying tumors according to their low and high risk of metastasis. In our previous study, the poor outcome of UM was related with overexpression of high mobility group A1 protein (HMGA1) [9]. Genetic studies reported the loss of chromosome 3 was the risk factor of poor outcome of UM [10,11]. However, none of the above indicators can be considered as effectively therapeutic targets in UM except HMGA1. Therefore, it is worth identifying reliable biomarkers of uveal melanoma for its early diagnosis and effective therapy.
Enhancer of zeste homologue 2 (EZH2) is a member of the polycomb group of genes, regulating the cell cycle through nucleosome modification, chromatin remodeling, and interacting with other transcription factors [12], the expression of EZH2 Research Paper delays upon tissue maturation and differentiation [13]. EZH2 overexpression has been reported to be related with increased tumor cell proliferation and worse outcome in several carcinomas including breast cancer [14], endometrial carcinomas [15] and hormonerefractory prostate cancer [16], which indicated that the expression of EZH2 protein might serve as the potential biomarker in carcinomas. To the best of our knowledge, no report was published on the role of EZH2 protein in UM prognosis.
In the present study, a series of UM cases were analyzed the prognostic value and potential target for therapeutic intervention of EZH2 by using immunohistochemistry. The correlation of EZH2 protein with cell proliferation marker Ki67/MIB-1 and relevant clinical parameters were explored herein.

Correlation between EZH2 expression and other characteristics in UM patients
Detection of EZH2 immunoreactivity in UM was shown in Figure 1A with clear nuclear staining. High level of expression, negative case without nuclear staining, positive control and negative control were shown in Figure 1B-1E, respectively. EZH2 nuclear expression was detected in 51 UM samples (57%), 31(35%) expressed high levels among them ( Table 1).

DISCUSSION
In this study, we used immunohistochemistry to analyze the EZH2 expression in UM patients and evaluate its prognostic value and potential therapeutic target. Our critical finding is that overexpression of EZH2, associated with presence of epithelioid cell, high level of mitoses counts and Ki67 LI, is related with poor clinical outcomes in UM patients.
Common tumor treatment alternatives including enucleation and plaque brachytherapy can be helpful for preserving the effected eye and vision in some patients. However, UM patient survival has been nearly no change in three decades despite the progress of therapy methods [3]. Hence, an effective candidate is necessary in clinic. EZH2 is a cell cycle regulator played an important role in G2-M transition and  EZH2 overexpression was found to be linked to shorter survival time of UM patients in this study, consistent with previous studies in other malignancies [20,21]. In addition, high EZH2 expression and metastatic mortality were associated with other UM parameters such as presence of epithelioid cell pattern, high mitoses count and Ki67 LI. This finding confirmed the role of EZH2 as a reliable marker for prognosis in UM.
Tumor cell pattern is also a helpful prognostic factor for survival time in UM. UM cellular morphology was

Figure 2: Correlation between EZH2 expression and other parameters in UM. The Ki67 LI (A) and mitoses counts (B) were
significantly higher in high EZH2 expression group than in low EZH2 expression group. Significant correlations have been found between the expression of EZH2 and Ki67 LI (C), and mitoses counts (D). EZH2 expression levels were significantly higher in cases showing epithelioid cell pattern than in cases lacking this cell pattern (E).

Figure 3: Kaplan-Meier curves for survival in UM patients. Survival of UM patients were evaluated according to EZH2
expression and Ki67 LI. Significant differences in disease-free survival rates were observed (A). Significant differences in disease-specific survival rates were also observed (B). P values were calculated with a log-rank test. divided into spindle cell and mixed cell types according to Modified Callender system [22]. McLean IW et al. have reported that the spindle cell type had the best prognosis, mixed cell type a worse prognosis [22,23]. In this study, we classified the tumor cell type as mentioned above and found the similar correlation in univariate logistic regression analysis. However, when adjusted for other compounding factors, epithelioid cell pattern was no longer significantly associated with survival. It may can be explained that tumor cell type is not play a decisive role in prognostic.
In our patient cohort, clinic features were not significantly associated with the survival rate except mitoses counts. Correspondingly, the conclusions of different studied were various [24][25][26][27]. The parameters used as predictive markers may not be the most reliable method evidently. Currently, several studies have reported that one of EZH2 functions is involved in target gene activation and high EZH2 expression is related with many signaling pathways, such as the pRB-E2F, estrogen receptor and c-Myc signal transduction pathways [17,28,29]. These results collectively, suggest that the regulation of EZH2 expression is rather complicated, it may serve as a promising marker for UM treatment and can help doctors to judge the tumor prognosis.
In conclusion, we assessed the prognostic value of EZH2 expression and its correlation with clinical and histopathologic parameters by immunohistochemistry. We found that overexpression EZH2 in UM is associated with adverse clinical outcome, hence, it might able to play a part in therapeutic target for reducing the tumor metastasis and improving survival time of UM patients.

Patients and tumor samples
Eighty-nine primary UM samples were collected after surgical enucleation from 1998 to 2006. Full ophthalmologic and systemic clinical examinations, such as liver function tests, chest X-ray and abdominal ultrasonography, were performed regularly before and after operation. Computed tomography or magnetic resonance imaging was used to confirm the possibility of metastasis in other parts of the body. This study conformed to the Declaration of Helsinki and was approved by the ethics committee of the appropriate institutes. All the patients had given their informed consent for experimental research previously. All the patients were diagnosed at ocular oncology clinic of Qilu hospital or Beijing Tongren hospital and the enucleation was conducted for UM during the indicated period. The duration of the follow-up period began with the date of UM diagnosis to the date of death or last follow-up (January 1 st 2016), whichever came first. Eighty-nine patients with complete follow-up data have been examined in survival analyses.

Histopathologic examination
As our previous study [9], the diagnosis of melanoma was confirmed using sample stained for hematoxylin and eosin (H&E) and/or MelanA. The tumors were histologically examined for cell type, localization, size, mitotic count, necrosis, and scleral invasion. Spindle and mixed cell types were assessed using the modified Callender system [22]. Extravascular matrix patterns were assessed using the periodic acid-Schiff reagent without hematoxylin counterstaining, and the sections were viewed under a green filter [30]. The mitotic count was measured by counting the number of mitoses in 40 highpower fields (HPF) in the H&E sections [31]. UM size, node, and metastasis (TNM) was classified according to the AJCC 7th edition system criteria [32].

Immunohistochemistry
Histological sections of formalin-fixed paraffinembedded samples were analyzed for the presence of Ki67 and EZH2 by the labeled streptavidin-biotin method. After deparaffinization and antigen retrieval using an autoclave oven technique, sections were incubated at 4°C overnight and incubated with EZH2 Rabbit Polyclonal Antibody (clone ZMD.309) (1:100; #187395, Thermo Fisher Scientific, Waltham, MA USA 02451) and Ki67 antigen mouse monoclonal antibody (1:75, DakoCytomatin, Glostrup, Denmark) at 4°C. Antigen-antibody complexes were detected by the cobalt-3, 3'-diaminobenzidine reaction. Squamous cell carcinomas known to be positive for EZH2 expression were used as positive controls [33]. Sections incubated in phosphate-buffered saline without the primary antibody served as negative controls.
Images of several HPF (×400) were captured from regions with different staining intensities, including high, moderate, low, and negative staining for each case. The photographs were printed on plain paper, and a grid was drawn over them. A total of 1000 cells were counted and expressed as a percentage of tumor cells with positive nuclei. The percentage of EZH2 positive tumor cells was scored on a scale from 0 to 4 (0, no staining; 1+, ≤10%; 2+,≤30%; 3+, ≤50%; 4+, >50%). The expression levels of EZH2 were divided into two groups according to score: low (score: 0, 1+); high (score: 2+, 3+, and 4+) [9,34]. The Ki67 labeling index (LI) was determined by counting the number of positive cells in a total of 800-1000 tumor cells observed in regions of highest staining (hot spot) at several HPF(×400). The results were expressed as a percentage of tumor cells with positive nuclei.

Statistical analysis
Statistical evaluation was performed using SPSS software (IBM SPSS Statistics 21; SPSS Inc., Chicago, IL). All data were described as mean ± standard deviation (SD) if applicable. The Pearson x 2 test or Fisher's exact test was used to compare qualitative variables. Associations and differences among the different parameters were analyzed using the Mann-Whitney U test and x 2 test. The relation between the expression levels of EZH2 and Ki67 LI were performed by Pearson's correlation coefficient. The Kaplan-Meier method and log-rank test were used for survival analyses. Cox proportional hazards regression model was used to calculate mortality hazard ratios and 95% confidence intervals (CIs).
To control for confounding variables, we used multivariate Cox proportional hazards regression models. To assess independent association between EZH2 expression and key severity markers (epithelioid cells, mitosis count and Ki67 LI), multivariate logistic regression analysis was done and odds ratio (OR) was adjusted for age and gender. Probability values (P)<0.05 were considered to be statistically significant.

Author contributions
Y.C., Y.Q. and W.W. designed the plan of this study; Y.Q. and W.W. collected the cases and tumor samples. Y.C. and Y.L. took the immunohistochemical photographs of tumor samples. Y.C., Y.L. and X.H. analyzed the data; Y.C. and Y.Q. used statistic software tool to made the tables and figures in this manuscript; Y.C. wrote the manuscript.