Leukemia & Lymphoma, September 2014; 55(9): 2056–2063 © 2014 Informa UK, Ltd. ISSN: 1042-8194 print / 1029-2403 online DOI: 10.3109/10428194.2013.858816
ORIGINAL ARTICLE: CLINICAL
Polycomb protein EZH2 expression in diffuse large B-cell lymphoma is associated with better prognosis in patients treated with rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone Hyun Jung Lee, Dong Hoon Shin, Kyung Bin Kim, Nari Shin, Won Young Park, Jung Hee Lee, Kyung Un Choi, Jee Yeon Kim, Chang Hun Lee & Mee Young Sol 1Department of Pathology, School of Medicine, Pusan National University, Yangsan, Korea and 2Medical Research Institute,
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Pusan National University, Yangsan, Korea
Abstract Polycomb group (PcG) proteins are evolutionarily conserved regulators of gene expression that contribute to normal lymphocyte development, and are involved in malignant transformation of these cells. Recently, BMI1 and EZH2 have been shown to be involved in lymphomagenesis and oncogenesis. We tried to elucidate the role of EZH2 as a prognostic factor for diffuse large B-cell lymphoma (DLBCL). High-level expression of EZH2 (EZH2 ⱖ 70%) was associated with superior overall survival (OS) of 85.8% compared to low expression (EZH2 ⬍ 70%), with OS of 44.5% (p ⴝ 0.005). Subgroup analysis showed that the activated B-cell (ABC) subtype with high EZH2 expression had the highest overall survival (p ⴝ 0.011). In analysis of EZH2 expression within low International Prognostic Index (IPI) score, high EZH2 expression had a significant statistical correlation with longer OS (p ⴝ 0.034). With high IPI score, high EZH2 expression tended to be associated with longer OS (p ⴝ 0.130). Our results showed that EZH2 expression had a high prognostic relevance to survival outcomes. We demonstrated that DLBCL was associated with increased expression of the EZH2 PcG protein and Ki67. The distribution of EZH2 expression was wider than that of Ki67. In summary, increased EZH2 expression of tumor cells was associated with improvements in OS. Keywords: Polycomb protein, EZH2, diffuse large B-cell lymphoma
Introduction Enhancer of Zeste homolog 2 (EZH2) is a member of the polycomb group (PcG) protein family, which are evolutionarily conserved regulators of gene expression [1]. EZH2 represses gene activities through its methyltransferase activity, which specifically trimethylates the lysine 27 residue of histone H3 (H3K27me3) [2]. PcG proteins ensure correct embryonic development by suppression of homeobox genes, body plan formation (axial patterning
through the repression of Hox genes), hematopoiesis and checkpoints affecting cell cycle entry [3]. As a member of the PcG protein family, EZH2 is also known to control B-cell development via H3K27me3 and immunoglobulin heavy chain (IgH) rearrangement [4]. Recently, PcG proteins have been suggested to be involved in cancer development and progression, and the role of EZH2 has been studied in several cancers including prostate cancer, breast cancer, lung cancer and lymphoma [1,5]. EZH2 may function as an oncogene and contribute to tumorigenesis by silencing tumor suppressor proteins, most of which work on cellcycle mechanisms [1–3]. Diffuse large B-cell lymphoma (DLBCL), not otherwise specified, is the most common lymphoid malignancy, constituting approximately 30% of all adult lymphomas and with rapidly rising incidence [6]. Because DLBCL has heterogeneous biologic features and clinical behaviors, current attempts to evaluate prognosis in DLBCL rely mainly on clinical parameters, as used most notably in the International Prognostic Index (IPI) score [7]. Although the subdivision of DLBCL into germinal center B-cell (GCB) and activated B-cell (ABC) types using cDNA gene expression microarray shows a better prognosis for the GCB subtype, subgrouping using immunohistochemical markers has provided variable results, making this classification difficult to apply in daily clinical practice [8,9]. In an effort to develop reliable but easily accessible prognostic markers, the Ki67 protein is an attractive candidate, because DLBCL is a highgrade lymphoma which exhibits active tumor cell proliferation, and Ki67 is a well-established cell proliferation marker. However, its prognostic role in DLBCL remains unsatisfactory. Many studies have presented variable results for Ki67 as prognostic marker [7,8,10]. Therefore, there is a need for additional markers to identify high-risk patients at diagnosis and to improve the definition of specific gene expression pathways for novel therapeutic interventions.
Correspondence: Dong Hoon Shin, MD, Department of Pathology, School of Medicine, Pusan National University, Beomeo-ri, Mulgeum-eup, Yangsan 626-770, Korea. Tel: 051-510-8050. Fax: 051-510-8040. E-mail: [email protected] Received 24 August 2013; revised 2 October 2013; accepted 21 October 2013
2056
EZH2 expression in diffuse large B-cell lymphoma Considering that EZH2 is expressed in centroblasts in normal germinal centers of lymph nodes and is involved in cell cycle regulation, we hypothesized that EZH2 could reflect the proliferation capacity of DLBCL and provide prognostic information better than that of Ki67. In this study, immunohistochemistry and double immunofluoresence staining were used to examine the distribution and frequency of the protein expression of EZH2, in comparison with Ki67, among patients with DLBCL. We also analyzed the prognostic value of EZH2 expression at the time of diagnosis as well as its relationship to other standard clinicopathologic parameters.
Materials and methods
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Patient selection A retrospective analysis was performed in 84 patients diagnosed with DLBCL at Pusan National University Hospital and Pusan National University Yangsan Hospital (Busan, Korea) from January 2003 to December 2012. These patients fulfilled the following inclusion criteria: (i) histologically proven diagnosis of DLBCL according to the World Health Organization (WHO) “classification of tumours of haematopoietic and lymphoid tissues” [11]; (ii) positive CD20 expression; (iii) availability of paraffin-embedded tumor specimens before treatment; (iv) no previous treatment; (v) no previous neoplasm or no second primary malignancy; (vi) no severe coexisting disease; (vii) and available clinical information and follow-up data. The patients were between 17 and 88 years of age. Antibodies to CD10, BCL6, MUM1, CD20 and CD3 were applied for immunophenotyping analysis. GCB and ABC subtypes were classified using an algorithm proposed by Hans et al. [6–10]. Clinical data were made available by referring clinicians for all patients listed in Table I. All patients analyzed in this study were treated with rituximab in combination with CHOP (R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone) as first-line chemotherapy. R was administered at the standard dose of 375 mg/m2 on day 1 with the CHOP regimen. The standard CHOP regimen consisted of cyclophosphamide 750 mg/m2 (600 mg/m2, age ⱖ 60), doxorubicin 50 mg/m2 (30 mg/m2, age ⱖ 60) and vincristine 1.4 mg/m2 (1 mg/m2, age ⱖ 60), as well as oral prednisone 100 mg/day on days 1–5. The patients were followed up from the date of diagnosis until death or to the last visit to the outpatient department (median follow-up 15 months, range 1–114 months). At the time of last follow-up, 53 patients (63.0%) were alive and 30 patients (35.7%) had died, and one patient (1.1%) was unknown (lost to follow-up).
Immunohistochemistry Immunohistochemistry was carried out on formalin-fixed, paraffin-embedded sections (4 μm thick). Staining was performed using the BondMax autostainer and reagents (Vision BioSystems, Mount Waverley, Australia). Deparaffinization was performed automatically in the autostainer with BondWash solution at 72°C for 30 min. Slides were then incubated with Epitope Retrieval Solution 1 (Leica Microsystems,
2057
Table I. Major clinical features of patient cohort (n ⫽ 84). Clinical parameter Gender Male Female Age (years) Median (range) ⱕ 60 ⬎ 60 Ann Arbor stage I II III IV LDH (UI/L) ⬎ 460 ⬍ 460 NA IPI Low risk (0–2) High risk (3–5) NA B symptoms Yes No NA Metastatic site Nodal Extranodal
No. of patients 46 38 59 (17–88) 41 43 13 33 17 21 46 37 1 45 38 1 14 68 2 28 56
LDH, lactate dehydrogenase; NA, not available; IPI, International Prognostic Index.
Wetzlar, Germany) for 20 min at 100°C, peroxide block for 5 min, primary monoclonal antibody for 15 min, postprimary reagent for 8 min and polymer for 8 min. The following primary antibodies were used: Ki67 (1:200, clone MIB1; Dako, Denmark), EZH2 (1:200, clone ZMD.309; Invitrogen, Carlsbad, CA). Ki67 and EZH2 expressions were observed in the nucleus of tumor cells. The percentage of cells with Ki67 and EZH2 expression was calculated from the number of cells with Ki67 and EZH2 positive nuclei and the total number of malignant cells in the highest labeling field under high magnification (⫻ 400). A minimum of 1000 epithelial cells was counted for each control or tumor sample. Two pathologists (H.J.L. and D.H.S.) independently evaluated all slides, while being blinded to the clinicopathologic information.
Confocal microscopy Double immunofluorescence staining with antisera against EZH2 and Ki67 was performed on 4 μm sections from formalin-fixed, paraffin-embedded tissue. After deparaffinization, antigen retrieval was performed at 95°C in pH 6.0 citrate buffer for 40 min. Sections were blocked with 1% bovine serum albumin for 20 min. Double immunolabeling was carried out by incubating the sections overnight with a mixture of two primary antibodies against EZH2 and Ki67. EZH2 was detected by incubation with anti-rabbit secondary antibody conjugated with ALEXA 594 (Invitrogen). Ki67 was detected by anti-mouse antibody conjugated with ALEXA 488 (Invitrogen). All washing steps were performed using phosphate buffered saline (PBS). Sections were analyzed with an Olympus Confocal Laser Scan microscope FV-1000 (Olympus, Tokyo, Japan). The acquired images were analyzed using FV ASW 3.1 software (Olympus).
2058
H. J. Lee et al.
Figure 1. Immunohistochemical staining of EZH2 in diffuse large B-cell lymphoma. (A) High EZH2 expression (⫻ 400), (B) low EZH2 expression (⫻ 400).
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Statistical analysis χ2
Pearson’s test was used to analyze the relationship of clinical characteristics with EZH2 expression. Survival curves and univariate analysis were performed using the Kaplan–Meier method. Follow-up information was also obtained for survival analysis. Overall survival (OS) was defined as the time interval between the date of diagnosis and the date of death from any cause or the date of last follow-up. Cases lost to follow-up or deaths from any other cause were defined as censored data for analysis of survival rates. The prognostic value of the possible factors influencing OS was assessed by multivariate analysis using the Cox proportional hazards model. p-Values less than 0.05 were considered statistically significant. Statistical analysis was performed using the SPSS version 19.0 software package (SPSS Inc., Chicago, IL).
Results Cut-off scores for EZH2 and Ki67 The percentage of tumor cells with nuclear EZH2 and Ki67 staining was evaluated. The median percentage of cells with nuclear expression of both EZH2 and Ki67 was 70%. From the lower to upper quartile range, various cut-off levels of EZH2 and Ki67 expressions were evaluated in increments of 10% (30% to 100%). The most significant differences in OS, using the log-rank test, occurred at a 70% cut-off for EZH2 expression and 80% cut-off for Ki67 expression. The receiver operating characteristic (ROC) curve was used to determine the cut-off point for EZH2 and Ki67 expression. ROC curve analysis established 65% of EZH2 expression as the cut-off value for OS with an area under the curve of 0.679 (95% confidence interval [CI], 0.559–0.789, p ⫽ 0.007), and 75% of Ki67 expression as the cut-off value for OS with an area under the curve of 0.581 (95% CI, 0.444–0.717, p ⫽ 0.225). Therefore, we chose the value of 70% as the optimal cut-off for EZH2 and an 80% cut-off value for Ki67. EZH2 expression was divided into either high (ⱖ 70%) or low (⬍ 70%). Out of the 84 total, EZH2 expression was high in 59 (70.3%) patients and low in 25 (29.7%) patients (Figure 1). The incidence of IPI score ⱖ 3 tended to be higher, with statistical significance (p ⫽ 0.002), in patients with ABC subtype than in patients with GCB subtype. Analysis
of the relationship between EZH2 expression level and CD10, BCL2, BCL6, MUM1 and Ki67 showed no significant correlation between EZH2 and other biomarkers (Table II).
EZH2 and Ki67 expression and OS As shown in Figure 2(A), log-rank test analysis revealed that patients with a high EZH2 expression (EZH2 ⱖ 70%) had a superior OS of 85.8% compared to those with low expression (EZH2 ⬍ 70%), with OS of 44.5% (p ⫽ 0.005). In contrast, low Ki67 expression (Ki67 ⬍ 80%) was associated with superior OS at 84.7% compared to those with high expression (Ki67 ⱖ 80%), with OS of 45.5% [p ⫽ 0.012, Figure 2(B)]. The impact of EZH2 expression on OS according to low and high IPI scores (IPI ⬍ 3 and IPI ⱖ 3, respectively) was analyzed. In patients with low IPI scores, high EZH2 expression had a statistically significant correlation with longer OS (p ⫽ 0.034) [Figure 3(A)]. In patients with high IPI scores, high EZH2 expression tended to correlate with longer OS (p ⫽ 0.130) [Figure 3(B)]. Possible prognostic factors for OS in all 84 patients were analyzed. The results of univariate analysis are summarized in Table III. IPI score ⱖ 3, elevated lactate dehydrogenase (LDH), low EZH2 expression level Table II. Comparison of clinical features according to immunophenotype and EZH2 expression. GCB subtype (n ⫽ 33)
ABC subtype (n ⫽ 51)
Low High Low High EZH2 EZH2 EZH2 EZH2 Characteristic (n ⫽ 8) (n ⫽ 25) p-Value (n ⫽ 16) (n ⫽ 35) p-Value CD10 ⫹ ⫺ BCL6 ⫹ ⫺ BCL2 ⫹ ⫺ MUM1 ⫹ ⫺ Ki67 High Low
8 0
23 2
0.425
0 17
0 34
0.000
5 3
17 8
0.782
11 6
20 14
0.692
8 0
24 1
0.580
16 1
34 0
0.159
8 0
22 3
0.319
16 1
33 1
0.618
2 6
15 10
0.458
6 11
15 19
0.555
GCB, germinal center B-cell; ABC, activated B-cell.
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
EZH2 expression in diffuse large B-cell lymphoma
2059
Figure 2. Kaplan–Meier curves showing overall survival (OS) of patients with diffuse large B-cell lymphoma. (A) OS in patients with low and high EZH2 expression, (B) OS in patients with low and high Ki67 expression.
and high Ki67 expression level had significant adverse prognostic impacts on OS. Table IV shows multivariate analysis by the Cox regression model, which revealed that low EZH2 expression had a marginally significant independent prognostic value for OS (p ⫽ 0.056).
EZH2 and Ki67 expression status by immunophenotype subgroup High EZH2 expression in GCB and ABC subtypes showed no significant difference (75.8% vs. 66.7%, p ⫽ 0.373). In patients with the GCB subtype, EZH2 expression had no effect on OS [Figure 4(A)]. However, in patients with the ABC subtype, low EZH2 expression was related to shorter OS [p ⫽ 0.016, Figure 4(B)]. Patients in the ABC group with high EZH2 expression had the highest OS, compared with the three other groups (p ⫽ 0.011). With respect to Ki67, high Ki67 expression in GCB and ABC subtypes showed no significant difference (36.4% vs. 63.6%, p ⫽ 0.659). Patients with high Ki67 expression and with
GCB subtype experienced shorter OS [p ⫽ 0.002, Figure 4(C)]. In patients with the ABC subtype, Ki67 expression had no effect on OS [Figure 4(D)].
Relationship between EZH2 and mitotic activity Mitotic number was counted, and the average number of mitoses was 17.89. Mitotic activity was arbitrarily divided into either high (ⱖ 10/10 high powered fields [HPFs]) or low (⬍ 10/10HPFs). EZH2 expression was significantly correlated with high mitotic activity (p ⫽ 0.05). No significant correlation was found between Ki67 expression and mitotic activity (p ⫽ 0.896). The impact of mitotic activity on OS was analyzed, and there was no statistically significant correlation with OS (p ⫽ 0.421).
Immunofluorescence expression of EZH2 and Ki67 in normal and neoplastic cells of DLBCL We analyzed PcG expression in normal germinal center B-cells and DLBCL cells. We confirmed that centroblasts in
Figure 3. Kaplan–Meier curves showing overall survival (OS) according to EZH2 expression in all patients grouped according to IPI scores. (A) OS in patients with IPI score 0–2 with low and high EZH2 expression, (B) OS in patients with IPI score ⱖ 3 with low and high EZH2 expression.
2060
H. J. Lee et al.
Table III. Analysis of prognostic factors for survival in univariate analysis. Variable Age ⬎ 60 Gender, male Elevated LDH level IPI score ⱖ 3 ABC subtype Low EZH2 expression High Ki67 expression
Table IV. Analysis of prognostic factors for survival in multivariate analysis.
OS, p-value 0.341 0.926 ⬍ 0.0001 ⬍ 0.0001 0.203 0.005 0.012
95% Confidence interval Parameter IPI score (3–5) Low EZH2 expression High Ki67 expression
Relative risk
Lower
Upper
p-Value
7.011 2.407 1.294
2.306 0.977 0.518
21.317 5.298 3.233
0.001 0.056 0.582
IPI, International Prognostic Index.
LDH, lactate dehydrogenase; IPI, International Prognostic Index; ABC, activated B-cell; OS, overall survival.
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Discussion normal germinal center B-cells expressed EZH2 and Ki67 (Figure 5). Neoplastic centroblasts showed high nuclear staining for EZH2. Staining in these DLBCL cells seemed comparable to the pattern for Ki67. Neoplastic cells simultaneously expressed Ki67 and EZH2, resulting in a yellow nuclear staining after combining the two signals. In this study, EZH2 expression was more widely distributed than Ki67 expression (Figures 6 and 7).
As DLBCL is the most common lymphoma and is biologically and pathologically heterogeneous, it is difficult to predict its prognosis, and there is no definite prognostic marker yet. With cDNA gene expression microarray, DLBCL can be subdivided into GCB and ABC types, by which distinction patients with GCB type DLBCL demonstrate improved OS. However, cDNA microarray is expensive and impractical in clinical practice, and the simpler and
Figure 4. Kaplan–Meier curves showing overall survival (OS) according to EZH2 expression and Ki67 expression. (A) OS in GCB-type patients with low and high EZH2 expression, (B) OS in ABC-type patients with low and high EZH2 expression, (C) OS in GCB-type patients with low and high Ki67 expression, (D) OS in ABC-type patients with low and high Ki67 expression.
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
EZH2 expression in diffuse large B-cell lymphoma
2061
Figure 5. Immunofluorescencence detection in normal germinal center B-cells. (A) Blue staining for 4¢,6-diamidino-2-phenylindole (DAPI), (B) green staining for Ki67, (C) red staining for EZH2, (D) double immunofluorescence for EZH2 and Ki67 (red and green, respectively), yellow cells indicate centroblasts. All images were taken from one representative experiment in normal germinal center B-cells (⫻ 600).
Figure 6. Immunofluorescence detection in DLBCL. (A) Blue staining for 4¢, 6-diamidino-2-phenylindole (DAPI), (B) green staining for Ki67, (C) red staining for EZH2, (D) double immunofluorescence for EZH2 and Ki67 (red and green, respectively) confirms tumor cells. Yellow cells indicate centroblasts. All images were taken from one representative experiment in DLBCL (⫻ 600).
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
2062
H. J. Lee et al.
Figure 7. Immunofluorescence detection in DLBCL. (A) Blue staining for 4¢,6-diamidino-2-phenylindole (DAPI), (B) green staining for Ki67, (C) red staining for EZH2, (D) double immunofluorescence for EZH2 and Ki67 (red and green, respectively) confirms tumor cells. Yellow cells indicate centroblasts. All images are from one representative experiment in DLBCL (⫻ 600, ⫻ 5.5).
cheaper method of immunohistochemistry was tried as an alternative to cDNA microarray analysis. Although there is some debate, immunohistochemical subtyping is generally regarded to have good correlation with cDNA microarray data [11,12]. With the introduction of R, however, the GCB/ ABC subtyping no longer correlated with prognosis, reflecting the advanced survival of patients with ABC type DLBCL [11,12]. While the most recent WHO criteria on DLBCL consider high proliferation index as a likely indicator of poor prognosis [6], Ki67 expression has shown variable results, and new prognostic markers seem necessary to stratify patients with DLBCL and to guide treatment in the post-R era. EZH2 is a cell cycle regulator necessary for G1/S and G2/M transition, and is an E2F-regulated gene [13,14]. It is important in B-cell development and is highly expressed in lymphoid progenitor cells and activated germinal center B-cells [13,14]. The association between EZH2 expression and proliferation of neoplastic cells has been demonstrated in other malignancies such as breast cancer, prostate cancer and melanoma [1,5]. In these malignancies, EZH2 expression in tumor cells is associated with aggressive clinical behavior and a worse prognosis. In this study, we examined the expression of EZH2 and its prognostic significance in 84 patients diagnosed with DLBCL. This study is meaningful in that no study has previously investigated the correlation between EZH2 expression and clinicopathologic factors or survival in patients with DLBCL. The results of this study show that Ki67 expression in ⱖ 80% of tumor cells was correlated with worse OS, when
all 84 patients or patients with GCB type were considered. In the ABC type, Ki67 expression did not show prognostic significance. Ki67 expression was arguably regarded as a poor prognostic factor before the advent of R [7–10], and its expression was reported to be higher in the ABC type than in the GCB type, accordant with a worse prognosis of ABC type [7–10]. However, in the post-R era, the survival outcomes have greatly improved for patients with ABC-type DLBCL, and the value of Ki67 as a prognostic marker seems to be limited [7–10]. As in previous studies, our results did not show a correlation between Ki67 and poor prognosis of the ABC type. In contrast, EZH2 expression levels exhibited high prognostic relevance to survival outcomes, in that patients with high EZH2 expression tended to experience longer survival. When patients were divided into GCB and ABC groups, this relationship between EZH2 expression and survival outcomes remained in the ABC group. Furthermore, EZH2 expression was marginally significant in multivariate analysis, while Ki67 expression was not significant. Recent studies indicate that EZH2 is expressed in proliferating lymphoma cells, which are also called neoplastic centroblasts [3]. We also observed that EZH2 is expressed in centroblasts in normal germinal centers, and EZH2 and Ki67 expressions are highly concordant in these centroblasts. However, in DLBCL, EZH2 was stained in more neoplastic cells compared to Ki67, with the mean value of EZH2 expression being higher than that of Ki67 expression in both ABC and GCB types (Table V). This observation was confirmed by double immunofluorescence staining,
EZH2 expression in diffuse large B-cell lymphoma Table V. Mean values of EZH2 and Ki67 expression. Parameter Average percent EZH2 expression ABC type GCB type Average percent Ki67 expression ABC type GCB type
Mean value (%) 72.50 70.78 75.15 67.74 69.41 65.15
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
ABC, activated B-cell; GCB, germinal center B-cell.
which revealed a wider distribution of EZH2 positive neoplastic cells and narrower distribution of Ki67 positive neoplastic cells. A similar distribution pattern of EZH2 and Ki67 expression was observed in a study on oral squamous cell carcinoma, in which EZH2 was more widely distributed than Ki67 in those neoplastic cells [15]. This broader expression of EZH2 protein may imply that EZH2 more accurately reflects the proliferative potential of neoplastic cells than does Ki67. Our results showing that high mitotic number in DLBCL was better correlated with high EZH2 expression rather than Ki67 expression also supports that EZH2 may reflect tumor cell proliferation more accurately than Ki67. The better OS outcome associated with high EZH2 expression probably suggests that proliferating lymphoma cells expressing EZH2 are more susceptible to chemotherapy. It is well established that overexpression of EZH2 increases the cell cycle G1/S and G2/M transition and rate of progression through the cell cycle [13,14], and knockdown of the EZH2 gene leads to significant cell cycle arrest in G2/S in a DLBCL cell line [5]. Because vincristine is an M phase inhibitor, this CHOP agent may be highly toxic toward lymphoma cells entering the M phase. In a study of colon cancer, a similar observation was made between EZH2 expression and improved survival outcomes for patients who received chemotherapy [16]. Why EZH2 is more broadly expressed than Ki67 in lymphoma cells is unclear. A recent study reported that amplification of the EZH2 gene in oral cancer was concordant with high expression of EZH2 protein [15]. Two other studies have detected a specific mutation of the EZH2 gene in non-Hodgkin lymphoma [17,18]. Whether these genetic changes are correlated with EZH2 protein overexpression and are involved in the pathogenesis of DLBCL is yet to be determined. In conclusion, this study showed that EZH2 expression in DLBCL was an independent prognostic factor by multivariate analysis. The IPI score was also an independent prognostic factor, whereas the Ki67 index was not a significant factor. These results indicate that EZH2 is a good prognostic marker in addition to the IPI score, which has been regarded as the most reliable prognostic factor. Potential conflict of interest: Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.
2063
This work was supported by a clinical research grant from Pusan National University Yangsan Hospital.
References [1] Bracken AP, Helin K . Polycomb group proteins: navigators of lineage pathways led astray in cancer. Nat Rev Cancer 2009;9: 773–784. [2] Martin C, Zhang Y. The diverse functions of histone lysine methylation. Nat Rev Mol Cell Biol 2005;6:838–849. [3] van Kemenade FJ, Raaphorst FM, Blokziji T, et al. Coexpression of BMI-1 and EZH2 polycomb-group proteins is associated with cycling cells and degree of malignancy in B-cell non-Hodgkin lymphoma. Blood 2001;97:3896–3901. [4] Su IH, Basavaraj A , Krutchinsky AN, et al. Ezh2 controls B cell development through histone H3 methylation and Igh rearrangement. Nat Immunol 2003;4:124–131. [5] Velichutina I, Shaknovich R, Geng H, et al. EZH2-mediated epigenetic silencing in germinal center B cells contributes to proliferation and lymphomagenesis. Blood 2010;116:5247–5255. [6] Stein H, Warnke RA , Chan WC, et al. Diffuse large B-cell lymphoma, not otherwise specified. In: Swerdlow SH, Campo E, Harris NL, et al., editors. World Health Organization classification of tumours of haematopoietic and lymphoid tissues. Lyon: IARC: 2008. pp 233–237. [7] Zhi-Ming Li, Jia-Jia Huang, Yi Xia, et al. High Ki-67 expression in diffuse large B cell lymphoma patients with non-germinal center subtype indicates limited survival benefit from R-chop therapy. Eur J Haematol 2012;88:510–517. [8] Salles G, de Jong D, Xie W, et al. Prognostic significance of immunohistochemical biomarkers in diffuse large B-cell lymphoma: a study from the Lunenburg Lymphoma Biomarker Consortium. Blood 2011;117:7070–7078. [9] Visco C, Li Y, Xu-Monette ZY, et al. Comprehensive gene expression profiling and immunohistochemical studies support application of immunophenotypic algorithm for molecular subtype classification in diffuse large B-cell lymphoma: a report from the International DLBCL Rituximab-CHOP Consortium Program Study. Leukemia 2012;26: 2103–2113. [10] Hasselblom S, Ridell B, Sigurdardottir M, et al. Low rather than high Ki-67 protein expression is an adverse prognostic factor in diffuse large B-cell lymphoma. Leuk Lymphoma 2008; 49:1501–1509. [11] De Jong D, Rosenwald A , Chhanabhai M, et al. Immunohistochemical prognostic markers in diffuse large B-cell lymphoma: validation of tissue microarray as a prerequisite for broad clinical applications–a study from the Lunenburg Lymphoma Biomarker Consortium. J Clin Oncol 2007;25:805–812. [12] Liu YH, Xu FP, Zhuang HG, et al. Clinicopathologic significance of immunophenotypic profiles related to germinal center and activation B-cell differentiation in diffuse large B-cell lymphoma from Chinese patients. Hum Pathol 2008;39:875–884. [13] Bracken AP, Pasini D, Capra M, et al. EZH2 is downstream of the pRB-E2F pathway, essential for proliferation and amplified in cancer. EMBO J 2003;22:5323–5335. [14] Gonzalez ME, Li X, Yoy K , et al. Down-regulation of Enhancer of Zeste-2 decreases growth of estrogen receptor negative invasive breast carcinoma and requires BRCA1. Oncogene 2009;28: 843–853. [15] Kidani K, Osaki M, Tamura T, et al. High expression of EZH2 is associated with tumor proliferation and prognosis in human oral squamous cell carcinomas. Oral Oncol 2009;45:39–46. [16] Fluge Ø, Gravdal K , Carlsen E, et al. Expression of EZH2 and Ki-67 in colorectal cancer and associations with treatment response and prognosis. Br J Cancer 2009;101:1282–1289. [17] Morin RD, Mendez-Lago M, Mungall AJ, et al. Frequent mutation of histone modifying genes in non-Hodgkin lymphoma. Nature 2011;476:298–303. [18] Yap DB, Chu J, Berg T, et al. Somatic mutations at EZH2 Y641 act dominantly through a mechanism of selectively altered PRC2 catalytic activity, to increase H3K27 trimethylation. Blood 2011; 117:2451–2459.
ORIGINAL ARTICLE: CLINICAL
Polycomb protein EZH2 expression in diffuse large B-cell lymphoma is associated with better prognosis in patients treated with rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone Hyun Jung Lee, Dong Hoon Shin, Kyung Bin Kim, Nari Shin, Won Young Park, Jung Hee Lee, Kyung Un Choi, Jee Yeon Kim, Chang Hun Lee & Mee Young Sol 1Department of Pathology, School of Medicine, Pusan National University, Yangsan, Korea and 2Medical Research Institute,
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Pusan National University, Yangsan, Korea
Abstract Polycomb group (PcG) proteins are evolutionarily conserved regulators of gene expression that contribute to normal lymphocyte development, and are involved in malignant transformation of these cells. Recently, BMI1 and EZH2 have been shown to be involved in lymphomagenesis and oncogenesis. We tried to elucidate the role of EZH2 as a prognostic factor for diffuse large B-cell lymphoma (DLBCL). High-level expression of EZH2 (EZH2 ⱖ 70%) was associated with superior overall survival (OS) of 85.8% compared to low expression (EZH2 ⬍ 70%), with OS of 44.5% (p ⴝ 0.005). Subgroup analysis showed that the activated B-cell (ABC) subtype with high EZH2 expression had the highest overall survival (p ⴝ 0.011). In analysis of EZH2 expression within low International Prognostic Index (IPI) score, high EZH2 expression had a significant statistical correlation with longer OS (p ⴝ 0.034). With high IPI score, high EZH2 expression tended to be associated with longer OS (p ⴝ 0.130). Our results showed that EZH2 expression had a high prognostic relevance to survival outcomes. We demonstrated that DLBCL was associated with increased expression of the EZH2 PcG protein and Ki67. The distribution of EZH2 expression was wider than that of Ki67. In summary, increased EZH2 expression of tumor cells was associated with improvements in OS. Keywords: Polycomb protein, EZH2, diffuse large B-cell lymphoma
Introduction Enhancer of Zeste homolog 2 (EZH2) is a member of the polycomb group (PcG) protein family, which are evolutionarily conserved regulators of gene expression [1]. EZH2 represses gene activities through its methyltransferase activity, which specifically trimethylates the lysine 27 residue of histone H3 (H3K27me3) [2]. PcG proteins ensure correct embryonic development by suppression of homeobox genes, body plan formation (axial patterning
through the repression of Hox genes), hematopoiesis and checkpoints affecting cell cycle entry [3]. As a member of the PcG protein family, EZH2 is also known to control B-cell development via H3K27me3 and immunoglobulin heavy chain (IgH) rearrangement [4]. Recently, PcG proteins have been suggested to be involved in cancer development and progression, and the role of EZH2 has been studied in several cancers including prostate cancer, breast cancer, lung cancer and lymphoma [1,5]. EZH2 may function as an oncogene and contribute to tumorigenesis by silencing tumor suppressor proteins, most of which work on cellcycle mechanisms [1–3]. Diffuse large B-cell lymphoma (DLBCL), not otherwise specified, is the most common lymphoid malignancy, constituting approximately 30% of all adult lymphomas and with rapidly rising incidence [6]. Because DLBCL has heterogeneous biologic features and clinical behaviors, current attempts to evaluate prognosis in DLBCL rely mainly on clinical parameters, as used most notably in the International Prognostic Index (IPI) score [7]. Although the subdivision of DLBCL into germinal center B-cell (GCB) and activated B-cell (ABC) types using cDNA gene expression microarray shows a better prognosis for the GCB subtype, subgrouping using immunohistochemical markers has provided variable results, making this classification difficult to apply in daily clinical practice [8,9]. In an effort to develop reliable but easily accessible prognostic markers, the Ki67 protein is an attractive candidate, because DLBCL is a highgrade lymphoma which exhibits active tumor cell proliferation, and Ki67 is a well-established cell proliferation marker. However, its prognostic role in DLBCL remains unsatisfactory. Many studies have presented variable results for Ki67 as prognostic marker [7,8,10]. Therefore, there is a need for additional markers to identify high-risk patients at diagnosis and to improve the definition of specific gene expression pathways for novel therapeutic interventions.
Correspondence: Dong Hoon Shin, MD, Department of Pathology, School of Medicine, Pusan National University, Beomeo-ri, Mulgeum-eup, Yangsan 626-770, Korea. Tel: 051-510-8050. Fax: 051-510-8040. E-mail: [email protected] Received 24 August 2013; revised 2 October 2013; accepted 21 October 2013
2056
EZH2 expression in diffuse large B-cell lymphoma Considering that EZH2 is expressed in centroblasts in normal germinal centers of lymph nodes and is involved in cell cycle regulation, we hypothesized that EZH2 could reflect the proliferation capacity of DLBCL and provide prognostic information better than that of Ki67. In this study, immunohistochemistry and double immunofluoresence staining were used to examine the distribution and frequency of the protein expression of EZH2, in comparison with Ki67, among patients with DLBCL. We also analyzed the prognostic value of EZH2 expression at the time of diagnosis as well as its relationship to other standard clinicopathologic parameters.
Materials and methods
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Patient selection A retrospective analysis was performed in 84 patients diagnosed with DLBCL at Pusan National University Hospital and Pusan National University Yangsan Hospital (Busan, Korea) from January 2003 to December 2012. These patients fulfilled the following inclusion criteria: (i) histologically proven diagnosis of DLBCL according to the World Health Organization (WHO) “classification of tumours of haematopoietic and lymphoid tissues” [11]; (ii) positive CD20 expression; (iii) availability of paraffin-embedded tumor specimens before treatment; (iv) no previous treatment; (v) no previous neoplasm or no second primary malignancy; (vi) no severe coexisting disease; (vii) and available clinical information and follow-up data. The patients were between 17 and 88 years of age. Antibodies to CD10, BCL6, MUM1, CD20 and CD3 were applied for immunophenotyping analysis. GCB and ABC subtypes were classified using an algorithm proposed by Hans et al. [6–10]. Clinical data were made available by referring clinicians for all patients listed in Table I. All patients analyzed in this study were treated with rituximab in combination with CHOP (R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone) as first-line chemotherapy. R was administered at the standard dose of 375 mg/m2 on day 1 with the CHOP regimen. The standard CHOP regimen consisted of cyclophosphamide 750 mg/m2 (600 mg/m2, age ⱖ 60), doxorubicin 50 mg/m2 (30 mg/m2, age ⱖ 60) and vincristine 1.4 mg/m2 (1 mg/m2, age ⱖ 60), as well as oral prednisone 100 mg/day on days 1–5. The patients were followed up from the date of diagnosis until death or to the last visit to the outpatient department (median follow-up 15 months, range 1–114 months). At the time of last follow-up, 53 patients (63.0%) were alive and 30 patients (35.7%) had died, and one patient (1.1%) was unknown (lost to follow-up).
Immunohistochemistry Immunohistochemistry was carried out on formalin-fixed, paraffin-embedded sections (4 μm thick). Staining was performed using the BondMax autostainer and reagents (Vision BioSystems, Mount Waverley, Australia). Deparaffinization was performed automatically in the autostainer with BondWash solution at 72°C for 30 min. Slides were then incubated with Epitope Retrieval Solution 1 (Leica Microsystems,
2057
Table I. Major clinical features of patient cohort (n ⫽ 84). Clinical parameter Gender Male Female Age (years) Median (range) ⱕ 60 ⬎ 60 Ann Arbor stage I II III IV LDH (UI/L) ⬎ 460 ⬍ 460 NA IPI Low risk (0–2) High risk (3–5) NA B symptoms Yes No NA Metastatic site Nodal Extranodal
No. of patients 46 38 59 (17–88) 41 43 13 33 17 21 46 37 1 45 38 1 14 68 2 28 56
LDH, lactate dehydrogenase; NA, not available; IPI, International Prognostic Index.
Wetzlar, Germany) for 20 min at 100°C, peroxide block for 5 min, primary monoclonal antibody for 15 min, postprimary reagent for 8 min and polymer for 8 min. The following primary antibodies were used: Ki67 (1:200, clone MIB1; Dako, Denmark), EZH2 (1:200, clone ZMD.309; Invitrogen, Carlsbad, CA). Ki67 and EZH2 expressions were observed in the nucleus of tumor cells. The percentage of cells with Ki67 and EZH2 expression was calculated from the number of cells with Ki67 and EZH2 positive nuclei and the total number of malignant cells in the highest labeling field under high magnification (⫻ 400). A minimum of 1000 epithelial cells was counted for each control or tumor sample. Two pathologists (H.J.L. and D.H.S.) independently evaluated all slides, while being blinded to the clinicopathologic information.
Confocal microscopy Double immunofluorescence staining with antisera against EZH2 and Ki67 was performed on 4 μm sections from formalin-fixed, paraffin-embedded tissue. After deparaffinization, antigen retrieval was performed at 95°C in pH 6.0 citrate buffer for 40 min. Sections were blocked with 1% bovine serum albumin for 20 min. Double immunolabeling was carried out by incubating the sections overnight with a mixture of two primary antibodies against EZH2 and Ki67. EZH2 was detected by incubation with anti-rabbit secondary antibody conjugated with ALEXA 594 (Invitrogen). Ki67 was detected by anti-mouse antibody conjugated with ALEXA 488 (Invitrogen). All washing steps were performed using phosphate buffered saline (PBS). Sections were analyzed with an Olympus Confocal Laser Scan microscope FV-1000 (Olympus, Tokyo, Japan). The acquired images were analyzed using FV ASW 3.1 software (Olympus).
2058
H. J. Lee et al.
Figure 1. Immunohistochemical staining of EZH2 in diffuse large B-cell lymphoma. (A) High EZH2 expression (⫻ 400), (B) low EZH2 expression (⫻ 400).
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Statistical analysis χ2
Pearson’s test was used to analyze the relationship of clinical characteristics with EZH2 expression. Survival curves and univariate analysis were performed using the Kaplan–Meier method. Follow-up information was also obtained for survival analysis. Overall survival (OS) was defined as the time interval between the date of diagnosis and the date of death from any cause or the date of last follow-up. Cases lost to follow-up or deaths from any other cause were defined as censored data for analysis of survival rates. The prognostic value of the possible factors influencing OS was assessed by multivariate analysis using the Cox proportional hazards model. p-Values less than 0.05 were considered statistically significant. Statistical analysis was performed using the SPSS version 19.0 software package (SPSS Inc., Chicago, IL).
Results Cut-off scores for EZH2 and Ki67 The percentage of tumor cells with nuclear EZH2 and Ki67 staining was evaluated. The median percentage of cells with nuclear expression of both EZH2 and Ki67 was 70%. From the lower to upper quartile range, various cut-off levels of EZH2 and Ki67 expressions were evaluated in increments of 10% (30% to 100%). The most significant differences in OS, using the log-rank test, occurred at a 70% cut-off for EZH2 expression and 80% cut-off for Ki67 expression. The receiver operating characteristic (ROC) curve was used to determine the cut-off point for EZH2 and Ki67 expression. ROC curve analysis established 65% of EZH2 expression as the cut-off value for OS with an area under the curve of 0.679 (95% confidence interval [CI], 0.559–0.789, p ⫽ 0.007), and 75% of Ki67 expression as the cut-off value for OS with an area under the curve of 0.581 (95% CI, 0.444–0.717, p ⫽ 0.225). Therefore, we chose the value of 70% as the optimal cut-off for EZH2 and an 80% cut-off value for Ki67. EZH2 expression was divided into either high (ⱖ 70%) or low (⬍ 70%). Out of the 84 total, EZH2 expression was high in 59 (70.3%) patients and low in 25 (29.7%) patients (Figure 1). The incidence of IPI score ⱖ 3 tended to be higher, with statistical significance (p ⫽ 0.002), in patients with ABC subtype than in patients with GCB subtype. Analysis
of the relationship between EZH2 expression level and CD10, BCL2, BCL6, MUM1 and Ki67 showed no significant correlation between EZH2 and other biomarkers (Table II).
EZH2 and Ki67 expression and OS As shown in Figure 2(A), log-rank test analysis revealed that patients with a high EZH2 expression (EZH2 ⱖ 70%) had a superior OS of 85.8% compared to those with low expression (EZH2 ⬍ 70%), with OS of 44.5% (p ⫽ 0.005). In contrast, low Ki67 expression (Ki67 ⬍ 80%) was associated with superior OS at 84.7% compared to those with high expression (Ki67 ⱖ 80%), with OS of 45.5% [p ⫽ 0.012, Figure 2(B)]. The impact of EZH2 expression on OS according to low and high IPI scores (IPI ⬍ 3 and IPI ⱖ 3, respectively) was analyzed. In patients with low IPI scores, high EZH2 expression had a statistically significant correlation with longer OS (p ⫽ 0.034) [Figure 3(A)]. In patients with high IPI scores, high EZH2 expression tended to correlate with longer OS (p ⫽ 0.130) [Figure 3(B)]. Possible prognostic factors for OS in all 84 patients were analyzed. The results of univariate analysis are summarized in Table III. IPI score ⱖ 3, elevated lactate dehydrogenase (LDH), low EZH2 expression level Table II. Comparison of clinical features according to immunophenotype and EZH2 expression. GCB subtype (n ⫽ 33)
ABC subtype (n ⫽ 51)
Low High Low High EZH2 EZH2 EZH2 EZH2 Characteristic (n ⫽ 8) (n ⫽ 25) p-Value (n ⫽ 16) (n ⫽ 35) p-Value CD10 ⫹ ⫺ BCL6 ⫹ ⫺ BCL2 ⫹ ⫺ MUM1 ⫹ ⫺ Ki67 High Low
8 0
23 2
0.425
0 17
0 34
0.000
5 3
17 8
0.782
11 6
20 14
0.692
8 0
24 1
0.580
16 1
34 0
0.159
8 0
22 3
0.319
16 1
33 1
0.618
2 6
15 10
0.458
6 11
15 19
0.555
GCB, germinal center B-cell; ABC, activated B-cell.
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
EZH2 expression in diffuse large B-cell lymphoma
2059
Figure 2. Kaplan–Meier curves showing overall survival (OS) of patients with diffuse large B-cell lymphoma. (A) OS in patients with low and high EZH2 expression, (B) OS in patients with low and high Ki67 expression.
and high Ki67 expression level had significant adverse prognostic impacts on OS. Table IV shows multivariate analysis by the Cox regression model, which revealed that low EZH2 expression had a marginally significant independent prognostic value for OS (p ⫽ 0.056).
EZH2 and Ki67 expression status by immunophenotype subgroup High EZH2 expression in GCB and ABC subtypes showed no significant difference (75.8% vs. 66.7%, p ⫽ 0.373). In patients with the GCB subtype, EZH2 expression had no effect on OS [Figure 4(A)]. However, in patients with the ABC subtype, low EZH2 expression was related to shorter OS [p ⫽ 0.016, Figure 4(B)]. Patients in the ABC group with high EZH2 expression had the highest OS, compared with the three other groups (p ⫽ 0.011). With respect to Ki67, high Ki67 expression in GCB and ABC subtypes showed no significant difference (36.4% vs. 63.6%, p ⫽ 0.659). Patients with high Ki67 expression and with
GCB subtype experienced shorter OS [p ⫽ 0.002, Figure 4(C)]. In patients with the ABC subtype, Ki67 expression had no effect on OS [Figure 4(D)].
Relationship between EZH2 and mitotic activity Mitotic number was counted, and the average number of mitoses was 17.89. Mitotic activity was arbitrarily divided into either high (ⱖ 10/10 high powered fields [HPFs]) or low (⬍ 10/10HPFs). EZH2 expression was significantly correlated with high mitotic activity (p ⫽ 0.05). No significant correlation was found between Ki67 expression and mitotic activity (p ⫽ 0.896). The impact of mitotic activity on OS was analyzed, and there was no statistically significant correlation with OS (p ⫽ 0.421).
Immunofluorescence expression of EZH2 and Ki67 in normal and neoplastic cells of DLBCL We analyzed PcG expression in normal germinal center B-cells and DLBCL cells. We confirmed that centroblasts in
Figure 3. Kaplan–Meier curves showing overall survival (OS) according to EZH2 expression in all patients grouped according to IPI scores. (A) OS in patients with IPI score 0–2 with low and high EZH2 expression, (B) OS in patients with IPI score ⱖ 3 with low and high EZH2 expression.
2060
H. J. Lee et al.
Table III. Analysis of prognostic factors for survival in univariate analysis. Variable Age ⬎ 60 Gender, male Elevated LDH level IPI score ⱖ 3 ABC subtype Low EZH2 expression High Ki67 expression
Table IV. Analysis of prognostic factors for survival in multivariate analysis.
OS, p-value 0.341 0.926 ⬍ 0.0001 ⬍ 0.0001 0.203 0.005 0.012
95% Confidence interval Parameter IPI score (3–5) Low EZH2 expression High Ki67 expression
Relative risk
Lower
Upper
p-Value
7.011 2.407 1.294
2.306 0.977 0.518
21.317 5.298 3.233
0.001 0.056 0.582
IPI, International Prognostic Index.
LDH, lactate dehydrogenase; IPI, International Prognostic Index; ABC, activated B-cell; OS, overall survival.
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
Discussion normal germinal center B-cells expressed EZH2 and Ki67 (Figure 5). Neoplastic centroblasts showed high nuclear staining for EZH2. Staining in these DLBCL cells seemed comparable to the pattern for Ki67. Neoplastic cells simultaneously expressed Ki67 and EZH2, resulting in a yellow nuclear staining after combining the two signals. In this study, EZH2 expression was more widely distributed than Ki67 expression (Figures 6 and 7).
As DLBCL is the most common lymphoma and is biologically and pathologically heterogeneous, it is difficult to predict its prognosis, and there is no definite prognostic marker yet. With cDNA gene expression microarray, DLBCL can be subdivided into GCB and ABC types, by which distinction patients with GCB type DLBCL demonstrate improved OS. However, cDNA microarray is expensive and impractical in clinical practice, and the simpler and
Figure 4. Kaplan–Meier curves showing overall survival (OS) according to EZH2 expression and Ki67 expression. (A) OS in GCB-type patients with low and high EZH2 expression, (B) OS in ABC-type patients with low and high EZH2 expression, (C) OS in GCB-type patients with low and high Ki67 expression, (D) OS in ABC-type patients with low and high Ki67 expression.
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
EZH2 expression in diffuse large B-cell lymphoma
2061
Figure 5. Immunofluorescencence detection in normal germinal center B-cells. (A) Blue staining for 4¢,6-diamidino-2-phenylindole (DAPI), (B) green staining for Ki67, (C) red staining for EZH2, (D) double immunofluorescence for EZH2 and Ki67 (red and green, respectively), yellow cells indicate centroblasts. All images were taken from one representative experiment in normal germinal center B-cells (⫻ 600).
Figure 6. Immunofluorescence detection in DLBCL. (A) Blue staining for 4¢, 6-diamidino-2-phenylindole (DAPI), (B) green staining for Ki67, (C) red staining for EZH2, (D) double immunofluorescence for EZH2 and Ki67 (red and green, respectively) confirms tumor cells. Yellow cells indicate centroblasts. All images were taken from one representative experiment in DLBCL (⫻ 600).
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
2062
H. J. Lee et al.
Figure 7. Immunofluorescence detection in DLBCL. (A) Blue staining for 4¢,6-diamidino-2-phenylindole (DAPI), (B) green staining for Ki67, (C) red staining for EZH2, (D) double immunofluorescence for EZH2 and Ki67 (red and green, respectively) confirms tumor cells. Yellow cells indicate centroblasts. All images are from one representative experiment in DLBCL (⫻ 600, ⫻ 5.5).
cheaper method of immunohistochemistry was tried as an alternative to cDNA microarray analysis. Although there is some debate, immunohistochemical subtyping is generally regarded to have good correlation with cDNA microarray data [11,12]. With the introduction of R, however, the GCB/ ABC subtyping no longer correlated with prognosis, reflecting the advanced survival of patients with ABC type DLBCL [11,12]. While the most recent WHO criteria on DLBCL consider high proliferation index as a likely indicator of poor prognosis [6], Ki67 expression has shown variable results, and new prognostic markers seem necessary to stratify patients with DLBCL and to guide treatment in the post-R era. EZH2 is a cell cycle regulator necessary for G1/S and G2/M transition, and is an E2F-regulated gene [13,14]. It is important in B-cell development and is highly expressed in lymphoid progenitor cells and activated germinal center B-cells [13,14]. The association between EZH2 expression and proliferation of neoplastic cells has been demonstrated in other malignancies such as breast cancer, prostate cancer and melanoma [1,5]. In these malignancies, EZH2 expression in tumor cells is associated with aggressive clinical behavior and a worse prognosis. In this study, we examined the expression of EZH2 and its prognostic significance in 84 patients diagnosed with DLBCL. This study is meaningful in that no study has previously investigated the correlation between EZH2 expression and clinicopathologic factors or survival in patients with DLBCL. The results of this study show that Ki67 expression in ⱖ 80% of tumor cells was correlated with worse OS, when
all 84 patients or patients with GCB type were considered. In the ABC type, Ki67 expression did not show prognostic significance. Ki67 expression was arguably regarded as a poor prognostic factor before the advent of R [7–10], and its expression was reported to be higher in the ABC type than in the GCB type, accordant with a worse prognosis of ABC type [7–10]. However, in the post-R era, the survival outcomes have greatly improved for patients with ABC-type DLBCL, and the value of Ki67 as a prognostic marker seems to be limited [7–10]. As in previous studies, our results did not show a correlation between Ki67 and poor prognosis of the ABC type. In contrast, EZH2 expression levels exhibited high prognostic relevance to survival outcomes, in that patients with high EZH2 expression tended to experience longer survival. When patients were divided into GCB and ABC groups, this relationship between EZH2 expression and survival outcomes remained in the ABC group. Furthermore, EZH2 expression was marginally significant in multivariate analysis, while Ki67 expression was not significant. Recent studies indicate that EZH2 is expressed in proliferating lymphoma cells, which are also called neoplastic centroblasts [3]. We also observed that EZH2 is expressed in centroblasts in normal germinal centers, and EZH2 and Ki67 expressions are highly concordant in these centroblasts. However, in DLBCL, EZH2 was stained in more neoplastic cells compared to Ki67, with the mean value of EZH2 expression being higher than that of Ki67 expression in both ABC and GCB types (Table V). This observation was confirmed by double immunofluorescence staining,
EZH2 expression in diffuse large B-cell lymphoma Table V. Mean values of EZH2 and Ki67 expression. Parameter Average percent EZH2 expression ABC type GCB type Average percent Ki67 expression ABC type GCB type
Mean value (%) 72.50 70.78 75.15 67.74 69.41 65.15
Leuk Lymphoma Downloaded from informahealthcare.com by Dicle Univ. on 11/14/14 For personal use only.
ABC, activated B-cell; GCB, germinal center B-cell.
which revealed a wider distribution of EZH2 positive neoplastic cells and narrower distribution of Ki67 positive neoplastic cells. A similar distribution pattern of EZH2 and Ki67 expression was observed in a study on oral squamous cell carcinoma, in which EZH2 was more widely distributed than Ki67 in those neoplastic cells [15]. This broader expression of EZH2 protein may imply that EZH2 more accurately reflects the proliferative potential of neoplastic cells than does Ki67. Our results showing that high mitotic number in DLBCL was better correlated with high EZH2 expression rather than Ki67 expression also supports that EZH2 may reflect tumor cell proliferation more accurately than Ki67. The better OS outcome associated with high EZH2 expression probably suggests that proliferating lymphoma cells expressing EZH2 are more susceptible to chemotherapy. It is well established that overexpression of EZH2 increases the cell cycle G1/S and G2/M transition and rate of progression through the cell cycle [13,14], and knockdown of the EZH2 gene leads to significant cell cycle arrest in G2/S in a DLBCL cell line [5]. Because vincristine is an M phase inhibitor, this CHOP agent may be highly toxic toward lymphoma cells entering the M phase. In a study of colon cancer, a similar observation was made between EZH2 expression and improved survival outcomes for patients who received chemotherapy [16]. Why EZH2 is more broadly expressed than Ki67 in lymphoma cells is unclear. A recent study reported that amplification of the EZH2 gene in oral cancer was concordant with high expression of EZH2 protein [15]. Two other studies have detected a specific mutation of the EZH2 gene in non-Hodgkin lymphoma [17,18]. Whether these genetic changes are correlated with EZH2 protein overexpression and are involved in the pathogenesis of DLBCL is yet to be determined. In conclusion, this study showed that EZH2 expression in DLBCL was an independent prognostic factor by multivariate analysis. The IPI score was also an independent prognostic factor, whereas the Ki67 index was not a significant factor. These results indicate that EZH2 is a good prognostic marker in addition to the IPI score, which has been regarded as the most reliable prognostic factor. Potential conflict of interest: Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.
2063
This work was supported by a clinical research grant from Pusan National University Yangsan Hospital.
References [1] Bracken AP, Helin K . Polycomb group proteins: navigators of lineage pathways led astray in cancer. Nat Rev Cancer 2009;9: 773–784. [2] Martin C, Zhang Y. The diverse functions of histone lysine methylation. Nat Rev Mol Cell Biol 2005;6:838–849. [3] van Kemenade FJ, Raaphorst FM, Blokziji T, et al. Coexpression of BMI-1 and EZH2 polycomb-group proteins is associated with cycling cells and degree of malignancy in B-cell non-Hodgkin lymphoma. Blood 2001;97:3896–3901. [4] Su IH, Basavaraj A , Krutchinsky AN, et al. Ezh2 controls B cell development through histone H3 methylation and Igh rearrangement. Nat Immunol 2003;4:124–131. [5] Velichutina I, Shaknovich R, Geng H, et al. EZH2-mediated epigenetic silencing in germinal center B cells contributes to proliferation and lymphomagenesis. Blood 2010;116:5247–5255. [6] Stein H, Warnke RA , Chan WC, et al. Diffuse large B-cell lymphoma, not otherwise specified. In: Swerdlow SH, Campo E, Harris NL, et al., editors. World Health Organization classification of tumours of haematopoietic and lymphoid tissues. Lyon: IARC: 2008. pp 233–237. [7] Zhi-Ming Li, Jia-Jia Huang, Yi Xia, et al. High Ki-67 expression in diffuse large B cell lymphoma patients with non-germinal center subtype indicates limited survival benefit from R-chop therapy. Eur J Haematol 2012;88:510–517. [8] Salles G, de Jong D, Xie W, et al. Prognostic significance of immunohistochemical biomarkers in diffuse large B-cell lymphoma: a study from the Lunenburg Lymphoma Biomarker Consortium. Blood 2011;117:7070–7078. [9] Visco C, Li Y, Xu-Monette ZY, et al. Comprehensive gene expression profiling and immunohistochemical studies support application of immunophenotypic algorithm for molecular subtype classification in diffuse large B-cell lymphoma: a report from the International DLBCL Rituximab-CHOP Consortium Program Study. Leukemia 2012;26: 2103–2113. [10] Hasselblom S, Ridell B, Sigurdardottir M, et al. Low rather than high Ki-67 protein expression is an adverse prognostic factor in diffuse large B-cell lymphoma. Leuk Lymphoma 2008; 49:1501–1509. [11] De Jong D, Rosenwald A , Chhanabhai M, et al. Immunohistochemical prognostic markers in diffuse large B-cell lymphoma: validation of tissue microarray as a prerequisite for broad clinical applications–a study from the Lunenburg Lymphoma Biomarker Consortium. J Clin Oncol 2007;25:805–812. [12] Liu YH, Xu FP, Zhuang HG, et al. Clinicopathologic significance of immunophenotypic profiles related to germinal center and activation B-cell differentiation in diffuse large B-cell lymphoma from Chinese patients. Hum Pathol 2008;39:875–884. [13] Bracken AP, Pasini D, Capra M, et al. EZH2 is downstream of the pRB-E2F pathway, essential for proliferation and amplified in cancer. EMBO J 2003;22:5323–5335. [14] Gonzalez ME, Li X, Yoy K , et al. Down-regulation of Enhancer of Zeste-2 decreases growth of estrogen receptor negative invasive breast carcinoma and requires BRCA1. Oncogene 2009;28: 843–853. [15] Kidani K, Osaki M, Tamura T, et al. High expression of EZH2 is associated with tumor proliferation and prognosis in human oral squamous cell carcinomas. Oral Oncol 2009;45:39–46. [16] Fluge Ø, Gravdal K , Carlsen E, et al. Expression of EZH2 and Ki-67 in colorectal cancer and associations with treatment response and prognosis. Br J Cancer 2009;101:1282–1289. [17] Morin RD, Mendez-Lago M, Mungall AJ, et al. Frequent mutation of histone modifying genes in non-Hodgkin lymphoma. Nature 2011;476:298–303. [18] Yap DB, Chu J, Berg T, et al. Somatic mutations at EZH2 Y641 act dominantly through a mechanism of selectively altered PRC2 catalytic activity, to increase H3K27 trimethylation. Blood 2011; 117:2451–2459.
