記住我
Abstract
Background: Enhancer of zeste homolog 2 (EZH2) is one of the major epigenetic modifiers involved in the transcriptional repression of target genes through trimethylation of H3K27 (lysine 27 residue of histone H3). Deregulated expression of both EZH2 and H3K27me3 has been implicated in the biological behavior and prognostic outcome of various malignancies. Aim: To assess the role of EZH2 and H3K27me3 in the carcinogenesis of urothelial carcinoma of urinary bladder. Materials and Methods: One hundred fifty consecutive urothelial carcinoma cases of urinary bladder (54.7% high-grade) were included in this study. Immunohistochemical analysis for EZH2 and H3K27me3 was performed on whole tissue sections. A multiplication score obtained by multiplying staining intensity and proportion of positively stained neoplastic cells was used for assessment. Results: EZH2 showed a significant correlation with the tumor grade and lamina propria invasion (p < 0.001). The cases with high EZH2 expression showed a significantly high proliferative index (Mean- 32.7%; p < 0.001). In contrast, negative and low expression of H3K27me3 was significantly more common in high-grade cases (p = 0.006). The expression of H3K27me3 was significantly associated with lamina propria (p = 0.01) and deep muscle invasion (p = 0.007). EZH2 showed a significantly higher expression in the high-grade invasive areas as compared to the high-grade non-invasive areas of the same tumor (p = 0.03). Conclusions: This study establishes an important role of the key epigenetic regulators EZH2 and H3K27me3 in the pathobiology of urothelial carcinomas. Strong expression of EZH2 and weak expression of H3K27me3 are associated with higher grade, proliferative index and invasive behavior.
Keywords: Epigenetics, EZH2, H3K27me3, urothelial carcinoma
How to cite this article: Introduction 
Urothelial carcinoma of the bladder is a significant cause of morbidity and mortality in the elderly population. Histologically, the urothelial carcinoma is divided into low-grade and high-grade morphology with or without lamina propria/detrusor muscle invasion.[1] While low-grade, non-invasive urothelial carcinomas are notorious for recurrence, the high-grade invasive ones show biologically aggressive behavior. Consequently, the genetics and molecular profiling of urothelial carcinoma are studied in great detail to explore the molecular pathways of the carcinogenesis, thereby intending to decipher the molecular-targeted therapy of this malignancy. The epigenetic pathway of urothelial carcinogenesis has recently been popular among the urothelial pathologists and researchers in an attempt to understand the carcinogenic pathways better. These epigenetic alterations have immense biological, morphological, prognostic, and predictive values.
Enhancer of zeste homolog 2 (EZH2) is a component of polycomb repressive complex (PRC2) and has an intrinsic histone methyltransferase activity. This, therefore, adds methyl groups to the 27th lysine residue of histone 3 (H3) (H3K27me3).[2],[3] The addition of the methyl group alters the configuration of the heterochromatin and brings about transcriptional repression of the protein. Notably, the tumor suppressor genes are suppressed by EZH2 activation, thereby promoting proto-oncogenic activity. The role of EZH2/H3K27me3 and its correlation with biological aggressiveness has been studied in different malignancies, including breast, prostate, and brain. H3K27me3 mutation has also gained acceptance in the recent classification of brain tumors acknowledging its significant role in cancer biology. The role of EZH2/H3K27me3 in urothelial malignancies has been highlighted in only a handful of studies.[2],[3],[4],[5] While the majority of these studies have highlighted the role of EZH2 in urothelial malignancy, only a few have explored its correlation with H3K27me3.[6] The role of EZH2 in the determination of biological tumor behavior is also variable in different studies.
We, therefore, tend to explore the role of these epigenetic markers in our setting and correlate them with the tumor grade, invasiveness, proliferative activity, and other parameters known to be related to tumor prognosis.
[TAG:2]Materials and Methods[/TAG:2]
A total of 150 consecutive cases of urothelial carcinoma of urinary bladder who underwent transurethral resection of bladder tumor (TURBT) or cystectomy were included in this study (January 2017–March 2020). Non- urothelial carcinoma, metastatic carcinoma, and benign biopsies of the urinary bladder were excluded from the study. The hematoxylin and eosin (H&E) stained slides of all the cases were reviewed by three pathologists trained in genitourinary pathology (SuM, SP, and AKA) for the concordance of the diagnosis and assessment of other histological parameters. The demographic data was obtained from the patient's records. This study was performed as a part of a project approved by the Institutional Ethical Committee.
Immunohistochemical analysis
The slides of all 150 cases were reviewed by three pathologists to ascertain the tumor grade, invasion, and heterogeneity. A single block is chosen in case of a morphologically homogeneous tumor (tumor with a uniformly similar grade, invasiveness, and histomorphology). In contrast, one/multiple (more than one) blocks were chosen in the tumor with heterogeneous areas [different areas with different grades, histomorphology (papillary along with non-papillary areas), and invasiveness (both non-invasive and invasive areas)]. Serial sections (3-4 μm thick) were cut from the blocks for immunohistochemical analysis and mounted on poly-L-lysine-coated glass slides. A section of lymph node (external positive control) was included in all batches of staining to assess the quality of staining.
Immunohistochemical staining for EZH2 (1:500 dilution, Cell Signalling Technology, Danvers, MA), H3K27me3 (1:1000 dilution, Cell Signalling Technology, Danvers, MA), and Ki67 (1:150 dilution, Dako, Carpinteria, CA) were performed on all slides. The antigen retrieval was performed in 0.01 mol/L citrate buffer (pH 6.0) in a 600-watt microwave oven for 30 min.
Quantification of immunohistochemistry
EZH2 and H3K27me3 scoring were performed by three experienced pathologists (SuM, SP, and AKA) blinded to the clinicopathological information. A labeling index was calculated in each case as a percentage of positively stained nuclei. For this, the slides were scanned under X40 magnification to look for the distribution of the staining and the hot-spot (the areas with the highest density of positive nuclei). At least 10 high power fields were counted under X400 magnification, starting with the hot-spot. The vascular endothelial cells and inflammatory cells were carefully excluded while counting. A semiquantitative scoring was ascribed depending on the percentage of nuclear positivity and the intensity of immunostaining. The labeling index was further sub-grouped as 0 for <10% of positive cells, 1 for 10%–33% of positive cells, 2 for 33%–66% of positive cells, and 3 for 67%–100% of positive cells. The staining intensity was also scored as 0 for no staining, 1 for weak staining, 2 for moderate staining, and 3 for strong staining. An overall score was ascribed, combining all the areas in the heterogeneous tumor where multiple representative areas were sampled/multiple blocks were chosen. The final immunostaining score was calculated by multiplying the percentage score with the intensity score. Hence, the minimum score was 0, and the maximum score was 9. Based on the multiplication scores, the expression pattern was divided into 3 sub-groups: Nil (score = 0), Low-expression (score = 1–4), and High-expression (score = 6–9). This was decided considering 5 to be the median value of the multiplication score. The Ki67 labeling index was calculated by standard protocol.
Comparative multiplication score calculation
Twenty-one cases of high-grade, invasive urothelial carcinoma among the 150 cases showed separately lying high-grade, non-invasive and high-grade, invasive components. The invasive and non-invasive areas in these cases were distinct from each other without transitioning into one another. We separately performed the multiplication scores of EZH2 and H3K27me3 in the frankly invasive and frankly non-invasive areas of the individual case to compare the variation of immunoexpression between the invasive and the non-invasive regions.
Statistical analysis
Statistical tests were performed using SPSS version 23.0 software (SPSS Inc., Chicago, IL, USA) and Microsoft Excel. The relationships among the qualitative variables were assessed by Chi-square and Fisher's exact test. To test the association of immunostaining score with other qualitative variables, the Mann–Whitney U test and Kruskal–Wallis test were used. Spearman's correlation coefficients were used for the assessment of the correlation between two continuous variables.
Recurrence-free survival (RFS) was calculated from the date of surgery to the date of first recurrence (clinical/radiological). Patients who did not show any evidence of recurrence at the last follow-up were considered as a censored event in the analysis. To assess the association of EZH2 and H3K27me3 expression with RFS, survival curves were calculated according to the Kaplan–Meier method, and the differences between curves were assessed using the log-rank test.
In all analyses, two-sided P values <0.05 were considered significant.
Results Clinicopathological features
The mean age of the patients with urothelial carcinoma was 61.1 years (range: 23–90 years). Only 13.3% subjects (n = 20) were young adult (age <50 years). There was a male preponderance (M:F- 6.9:1). The majority of the cases showed high-grade histomorphology (n = 82) (54.7%). Lamina propria invasion was present in 50% cases (n = 75), whereas 21.6% of cases (n = 21) showed the involvement of the detrusor muscle (n = 97). Only three cases showed the presence of lymphovascular emboli and perineural invasion (2% each).
Expression of EZH2
The mean EZH2 multiplication score in the low-grade tumors (1.98; SD = 2.51) (median 2.0) was very significantly different from the mean EZH2 multiplication score in the high-grade tumors (5.24; SD = 3.3) (median 6.0) (p value <0.001). Similarly, the mean EZH2 multiplication score in non-invasive tumors (2.03; SD = 2.47) (median 2.0) was very significantly different from the mean EZH2 multiplication score in the invasive tumors (5.41; SD = 3.31) (median 6.0) (p value <0.001). However, there was no significant difference between the mean EZH2 multiplication scores of detrusor-invasive (3.76; SD = 3.25) (median 2.0) and detrusor-non-invasive tumors (3.2; SD = 3.28) (median 4.0) (p value = 0.6).
EZH2 expression was not seen in normal mucosa or cystitis cystica (benign mucosa). Approximately, 58.8% of cases of the low-grade tumor (n = 40/68) and 86.6% of the high-grade tumor (n = 71/82) showed expression of EZH2. High expression was noted in 59.8% of cases of the high-grade tumor (n = 49/82). The EZH2 expression in the low-grade tumor was limited to the basal layer of the papillae. Besides, the intensity of EZH2 expression in the basal layer was higher than the intensity of the tumor cells in the superficial layer of the low-grade tumor. In contrast, only 10.3% of cases of the low-grade tumor (n = 7/68) showed a high expression of EZH2 [Table 1]. Interestingly, the invasive fronts within a tumor showed higher intensity as well as the proportion of immunopositivity as compared to other non-invasive tumor areas. Similarly, the high-grade areas of the same tumor showed higher expression of EZH2 in contrast to the low-grade areas signifying heterogeneity in its expression. About 61.3% of lamina propria-invasive tumors (n = 46/75) showed a high expression of EZH2. In contrast, only 13.3% of lamina propria-non-invasive tumors (n = 10/75) showed high EZH2 expression. The mean proliferative indices (MIB-1) in high, low, and nil EZH2 expressions were 32.7%, 8.9%, and 2.6%, respectively, depicting a very significant association (p value < 0.001) [Table 1].
Table 1: Correlation of EZH2 and H3K27me expression with different histopathological parametersIn summary, the high expression of EZH2 along with high-grade and invasiveness of the tumor was very significant in terms of statistics (p value < 0.001). Nevertheless, no significant association of EZH2 was identified with deep muscle invasion.
Expression of H3K27me3
The mean H3K27me3 multiplication scores were significantly different between the low-grade (mean 8.19; SD = 1.89) (median 9.0) and high-grade tumors (mean 6.67; SD = 2.84) (median 9.0) (p value = 0.006), non-invasive (mean 8.17; SD = 1.98) (median 9.0) and invasive tumors (mean 6.68; SD = 2.82) (median 9.0;) (p value = 0.01), and detrusor-non-invasive (mean 7.66; SD = 2.23) (median 9.0) and detrusor-invasive tumors (mean 5.71; SD = 3.42) (median 6.0) (p value = 0.007).
Mucosa from the non-tumorous bladder showed diffuse and strong positivity for H3K27me3. Among the neoplastic cases, 94.1% cases of the low-grade tumor (n = 64/68) showed strong expression, whereas low (n = 3/68) and nil (n = 1/68) expressions were present in only 4.4% and 1.5% cases, respectively. In contrast, 76.8% high-grade tumors showed high expression of H3K27me3 (n = 63/82), whereas low (n = 14/82) and nil expressions (n = 5/82) were seen in 17.1% and 6.1% cases, respectively. There was a significant difference of H3K27me3 expression between the muscle-invasive and non-muscle-invasive tumors (p-value = 0.007) as 88.1% (n = 67/76) of non-muscle invasive tumors showed high expression of H3K27me3 (66.7%; n = 14/21 cases of muscle-invasive tumors showed high H3K27me3 expression) [Table 1]. The invasive fronts of the tumors showed a distinctly weaker/loss of staining as compared to the other non-invasive areas. The pattern was almost opposite to that of EZH2 staining and this was overtly visible in the tumor with both low- and high-grade intratumoral areas. There was no significant correlation of H3K27me3 expression with the proliferative activity of the tumor. [Figure 1] and [Figure 2] depict the histomorphology, EZH2, and H3K27me3 expression and MIB-1 labeling of a low-grade and a high-grade tumor, respectively. [Figure 3] highlights the intratumoral heterogeneity in terms of similar parameters.
Figure 1: Histomorphology of a non-invasive papillary urothelial carcinoma, low-grade (a; Hematoxylin and eosin; 200×) with low EZH2 (b; 200×), and high H3K27me3 expression (c; 200×) alongside low Ki67 labeling index (d; 200×). Note the positivity of EZH2 and H3K27me3 in the stromal lymphocytes and endothelial cells that acted as internal control
Figure 2: Histomorphology of an invasive high-grade urothelial carcinoma (a; Hematoxylin and eosin; 200×) with high EZH2 (b; 200×), and low H3K27me3 expression (c; 200×) alongside high Ki67 labeling index (d; 200×)
Figure 3: Intratumoral heterogeneity in a high-grade papillary urothelial carcinoma with low-grade and high-grade areas. The histomorphology (a; Hematoxylin and eosin; 200×), EZH2 (b; 200×), H3K27me3 (c; 200×), and Ki67 labeling index (d; 200×). Note the exact opposite pattern of EZH2 and H3K27me3 expression in the low-grade and high-grade areasIn contrast to EZH2, loss of expression or low expression of H3K27me3 appeared to be more critical in the context of tumor biology. The mean multiplication scores of the high-grade or invasive tumors show lower values than the low-grade or non-invasive tumors.
Comparative analysis between the high-grade non-invasive and high-grade invasive areas of the same tumor
The median EZH2 score (6.0) (mean = 6.3; SD = 2.83) in the high-grade invasive area was significantly different as compared to the median EZH2 score (3.5) (mean = 4.05; SD = 3.25) in the high-grade non-invasive area (p-value = 0.03). In contrast, there was no significant difference of immunoexpression of H3K27me3 between the high-grade non-invasive and high-grade invasive areas of the same tumor (p-value = 0.56).
Follow-up data and survival analysis
Follow-up data for 47 cases (range 4.47–169.71 weeks) was available, out of which 23 patients had a recurrence. The estimated median RFS was 78.28 weeks. As there were a limited number of cases in each sub-group of immunoexpression (negative, low, and high) and the low and negative group showed similar trends, they were clubbed together for survival analysis. The cases with low/negative expression of H3K27me3 had a significantly shorter RFS as compared to the cases with high expression (estimated median RFS – 30.0 weeks vs. 88.29 weeks; P value = 0.04). However, there was no significant difference between the cases with strong expression of EZH2 and cases with low/negative expression in terms of RFS (estimated median RFS – 72.43 weeks vs. 78.29 weeks; P value = 0.35). The Kaplan–Meier curves have been depicted in [Figure 4].
Figure 4: Recurrence free survival (RFS) represented by Kaplan–Meier plots for H3K27me3 and EZH2 Discussion Urothelial carcinoma is known for recurrence and aggressive biological behavior. The low-grade and high-grade urothelial carcinomas are significantly different in terms of clinical, biological, histomorphological, and molecular characters. The epigenetic alterations of the urothelial carcinoma have recently been explored and gained considerable importance among the pathologists and molecular biologists as targeted therapies are available against different epigenetic modifiers.[2],[7],[8]
The polycomb repressive complex (PRC2) is significantly deregulated in malignancies, although its role in neoplastic transformation remains elusive.[9],[10] EZH2 is a well-studied protein of PRC2 that exerts its effects via methylation of H3K27 residue. The role of EZH2 in urothelial carcinoma has been studied in only a few previous studies.[2],[3],[4],[5] In addition to EZH2, we also tend to evaluate the role of H3K27me3 in urothelial carcinoma.
Our study showed a positive correlation of EZH2 with the tumor grade signifying its role in tumor biology. This result corroborates with the findings of Hinz et al. and Warrick et al.[2],[3] Interestingly, a significant proportion of low-grade urothelial carcinoma did not show any expression of EZH2. In contrast, the expression of EZH2 was noted in the tumor containing both low-grade and high-grade areas. This finding has also been reported by Warrick et al.[3] and probably signifies the role of EZH2 in the lineage of the tumor rather than stepwise carcinogenesis. Importantly, EZH2 expression also corroborated to the labeling index of the tumor. This draws our attention to the parallelism of EZH2 with MIB-1 and possibly explains the role of this epigenetic modifier in the cell-cycle regulation of the tumorigenesis.
Other interesting finding in our study is the overexpression of EZH2 in the invasive fronts of the tumor. This is statistically significant and highlights that there is overexpression of EZH2 at the invasive high-grade area compared to the non-invasive high-grade area despite these areas being part of the same tumor. This feature reiterates and reinforces the observation of the previous authors.[2],[3],[4],[5] This feature is important as this indicates the tumor heterogeneity at the epigenetic echelon and signifies the role of EZH2 in the shift in biological behavior from non-invasive to invasive.
H3K27me3 is the product of EZH2 activity, although its expression pattern appears to be different from EZH2. We observed a loss/weak expression of H3K27me3 to be significant than its overexpression, unlike EZH2. In fact, H3K27me3 expression was found to negatively correlate with the tumor grade and invasion with a significant association of high tumor grade and invasive property with the loss/low expression of H3K27me3. This is evident by the opposite patterns of immunoexpression of EZH2 and H3K27me3 in high-grade urothelial carcinomas. An almost similar observation was voiced by Ellinger et al.[6] in their seminal article comparing H3K27me3 and H3K9 expression in invasive urothelial carcinomas. We believe that this observation has immense significance in understanding the biology of EZH2 and H3K27me3.
Most of the studies with follow-up data have described the association of strong/high EZH2 expression with poor clinical outcomes in the form of shorter RFS, cancer-specific survival, and overall survival in different sub-groups of urinary bladder carcinomas.[2],[11],[12] In contrast, two other studies did not find any significant association of EZH2 protein expression with clinical outcome.[3],[13] Although the present study showed a relatively shorter RFS in patients with strong EZH2 expression, the difference was not statistically significant. In contrast to EZH2, only limited data is available regarding the role of H3K27me3 expression in determining the clinical outcome. The low/negative expression of H3K27me3 was associated with shorter RFS in the present study. This is in contrast to the previous reports by Liu et al.,[14] who found a longer cancer-specific survival in cases with low expression of H3K27me3. Moreover, the study by Ellinger et al.[6] failed to find any association of H3K27me3 expression with clinical outcome. However, most of the previous studies, including ours, have demonstrated the association of low/negative expression of H3K27me3 with established poor prognostic parameters like high histological grade, lamina propria, and deep muscle invasion.[6] Hence, the association of low/negative expression of H3K27me3 with shorter RFS is essentially in keeping with the aggressive tumor biology.
The index study establishes a positive correlation of EZH2 with the grade and labeling index of the tumor, whereas a negative correlation is observed with H3K27me3. This phenomenon, in contrast to popular belief, helps us to postulate a radically opposite role of these two proteins in tumor biology. Secondly, both these proteins correlated with the invasiveness of the tumor with overexpression of EZH2 in the invasive front while the loss of the H3K27me3 in similar areas. Obviously, this speaks volumes about the role of these proteins in the alteration of tumor biology. Thirdly, as EZH2 showed remarkable parallelism with proliferation index and grade of the tumor, which can be used as a surrogate marker in difficult cases of urothelial carcinoma where tumor grading becomes difficult. In this regard, Warrick et al.[3] also has recommended its usage in carcinoma-in situ cases. Lastly, the tumor markers showed heterogeneous expression in different tumor areas, and hence, we postulate its absence of impact in stepwise carcinogenesis.[15]
Conclusions The index study signifies the role of epigenetic markers (EZH2 and H3K27me3) in urothelial carcinogenesis. These markers correlate with tumor grade invasiveness and RFS. It also unravels the prospects of future epigenetic research in the field of urothelial carcinoma for pharmacological targeting of these epigenetic molecules in preventing the morbidity and mortality. Specific inhibitor especially against EZH2 are available and being use in clinical trials for various other malignancies.
Acknowledgments
RM has collected the data and prepared the first draft of the manuscript. SuM, AKA, and SP had reviewed all the slides, interpreted and scored the immunohistochemistry, and analyzed the data with preparation of the final manuscript. SwM and PN were involved in the clinical management, follow-up, and critical review of the manuscript.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/ijpm.ijpm_1267_21
留言 (0)