Presented at the 99th Annual Meeting of the United States and Canadian Academy of Pathology, Washington, DC, March 20-26, 2010.
We thank Karen Muller in the Department of Scientific Publications at The University of Texas MD Anderson Cancer Center for editing this article.
Enhancer of zeste homolog 2 (EZH2), a member of polycomb group proteins, is involved in the regulation of cell cycle progression and has been implicated in various human malignancies, including breast cancer, and also has been associated with aggressive tumor behavior. However, the clinical significance of EZH2 expression in inflammatory breast cancer (IBC), a rare but aggressive type of breast carcinoma, has not been explored. In this retrospective study, the authors examined EZH2 expression in IBC tumors and evaluated the relation between EZH2 expression and patient survival.
Tissue microarrays of 88 surgically resected IBC tumors were stained immunohistochemically for EZH2, and the authors evaluated the association of EZH2 expression status with clinicopathologic factors and clinical outcome.
The median follow-up for the entire cohort was 45.7 months, and the 5-year overall survival (OS) rate was 45%. EZH2 was expressed frequently in IBC tumors (75.7%) and was associated significantly with unfavorable prognostic factors, such as higher tumor grade, negative estrogen receptor status, and triple-negative status (ie, negative for the estrogen, progesterone, and human epidermal growth factor 2 receptors). Univariate survival analysis indicated that patients who had EZH2-positive IBC had a significantly lower 5-year OS rate than patients who had EZH2-negative IBC (P = .01). In multivariate analysis, only positive EZH2 status remained an independent predictor of worse OS.
Inflammatory breast cancer (IBC) is rare, accounting for approximately 2.5% of all breast cancers in the United States,1 and it is the most aggressive type of breast cancer. Characterized by diffuse dermal lymphatic plugging by tumor emboli, patients with IBC typically present with rapid onset and progression of breast enlargement with overlying skin changes, such as diffuse erythema, edema or peau d'orange, induration, tenderness, and warmth, mimicking the inflammatory process. At the time of diagnosis, approximately two-thirds of patients with IBC have locoregional lymph node metastasis, and up to one-third have distant metastasis.1 Despite improvement in the overall survival (OS) rate for IBC patients with current multimodality treatment,1, 2 the clinical outcome of these patients remains poor. In 1 study, the 5-year OS rate for patients with IBC was 40.5% compared with 63.2% for patients with locally advance non-IBC, although the patients had received similar multimodal treatment.3 Understanding biologic insights that underlie the aggressive behavior of IBC and identifying novel therapeutic strategies are highly desirable for improving the clinical outcome of patients with IBC.
The polycomb group (PcG) proteins are epigenetic chromatin modifiers that regulate transcription through nucleosome modification, chromatin remodeling, and interaction with other transcription factors, and these proteins are involved in the maintenance of embryonic and adult stem cells, cell cycle regulation, and cell division.4-11 Deregulation of PcG proteins can lead to cancer, and targeting PcG proteins has been suggested as a potential therapeutic strategy.12 Enhancer of zeste homolog 2 (EZH2) is a key member of the PcG proteins. Aberrant expression of EZH2 has been linked significantly to cell cycle progression and increased tumor cell proliferation (as estimated by Ki-67 expression and mitotic counts) in various types of cancer.13-17 EZH2 reportedly is implicated in cancer initiation and in the progression of a variety of human malignancies, including lymphoma,18, 19 gastric cancer,20 hepatocellular carcinoma,21 prostate cancer,17, 22-24 melanoma,17, 25 bladder cancer,26 endometrial cancer,17 and breast cancer.13, 14, 16, 17, 27-30 In studies with breast carcinoma, not otherwise specified (NOS), Kleer et al reported that high EZH2 levels were associated significantly with increased tumor size, high disease stage, younger age at diagnosis, and negative hormone receptor status. Moreover, those authors observed that EZH2 messenger RNA (mRNA) expression levels were significantly higher in invasive carcinomas that metastasized within 5 years of primary diagnosis compared with invasive carcinomas that did not metastasize.14 Collett et al demonstrated a significant correlation between EZH2 positivity (assessed by immunochemistry) and locally advanced breast cancer or metastatic spread at diagnosis.13 Raaphorst et al compared EZH2 expression in benign breast tissue, invasive breast carcinomas, and their precursors and observed a positive association between EZH2 expression and histologic grade of breast carcinoma, indicating that deregulation of EZH2 is related to loss of differentiation.16 Therefore, EZH2 has been proposed as a marker of aggressive breast cancer associated with poor clinical outcome. However, to our knowledge, the expression of EZH2 and its clinical importance in IBC has not been evaluated previously. In this retrospective study, we examined EZH2 expression in IBC tumors and evaluated the relation between EZH2 expression and patient survival. We demonstrated a high frequency (75.7%) of EZH2 expression in IBC tumors and observed that positive EZH2 status was associated with worse clinical outcome.
MATERIALS AND METHODS
This study included patients with primary IBC who were treated at The University of Texas MD Anderson Cancer Center from September 1994 to August 2004 and who had tumor tissue available for analysis and clinical follow-up information. In all patients, the initial histologic diagnosis of invasive breast carcinoma had been established on pretreated tumor tissues that were obtained by core-needle biopsy. Estrogen receptor (ER) status, progesterone receptor (PR) status, and human epidermal growth factor receptor 2 (HER2) status had been established on formalin-fixed, paraffin-embedded invasive tumor tissues. All patients then received multimodal treatment, including neoadjuvant chemotherapy, surgery, and radiation therapy.
The neoadjuvant chemotherapy included 4 to 6 cycles of an anthracycline-based, intravenous chemotherapy regimen of doxorubicin (50 mg/m2), cyclophosphamide (500 mg/m2), and 5-fluorouracil (500 mg/m2) given every 21 days. Most patients (89%) also received paclitaxel (225 mg/m2) for 3 weeks. At the completion of neoadjuvant chemotherapy, all patients underwent mastectomy; most patients also underwent axillary lymph node dissection. All patients received postmastectomy radiation to the chest wall and draining lymphatics, 31 patients (35.2%) received adjuvant endocrine therapy, and all patients with HER2-positive tumors received adjuvant trastuzumab. We included 88 patients who had a residual tumor at surgery and whose residual tumors had enough volume for constructing a tissue microarray (TMA). Patients who achieved a pathologic complete response were not eligible. Pathologic characteristics of the 88 tumors and prognostic and predictive biomarker status were reviewed and recorded retrospectively. This study was approved by The University of Texas MD Anderson Cancer Center Institutional Review Board.
Tumor Samples and TMAs
Formalin-fixed, paraffin-embedded primary IBC tumors obtained from mastectomy were selected for constructing TMAs. Samples with minimal residual tumor burden or dispersed tumor cells were excluded to ensure a good quality of the TMAs. We used a manual tissue puncher/array (Beecher Instruments, Sun Prairie, Wis) to construct TMAs. Three cores, each 1.0 mm in greatest dimension, were obtained from each tumor.
Immunohistochemistry Staining and Evaluation
Immunohistochemical staining for EZH2 was performed on TMA slides. Briefly, 4-μm-thick paraffin sections were dewaxed and treated with 3% H2O2 in methanol for 30 minutes to block endogenous peroxidase activity. Antigen retrieval was conducted by steaming the slides in 10 mM citrate buffer, pH 6.0, for 45 minutes. The sections were incubated for 30 minutes with normal nonimmune serum to eliminate nonspecific staining and then incubated with monoclonal antibody against EZH2 (clone 6A10; dilution, 1:200; Leica Novocastra, Bannockburn, Ill) for 1 hour at room temperature. The antigen was detected with biotin-labeled secondary antibody by the avidin-biotin-peroxidase technique using 3,3′-diaminobenzidine as the chromogen. The tissues were then counterstained with Mayer hematoxylin solution and were visualized under a lightfield microscope.
EZH2 staining was interpreted and recorded independently by 2 pathologists (Y.G. and L.H.) in a blinded manner. Positive EZH2 status was defined as nuclear staining in at least 10% of invasive cancer cells. Any discrepancies between the findings of the 2 pathologists were resolved by discussion at a multihead microscope.
ER and PR status had been determined by immunohistochemical staining with monoclonal antibodies against ER (6F11; Novocastra Laboratories, Burlingame, Calif) and against PR (1A6; Novocastra Laboratories) as part of the initial routine pathology evaluation. Positive hormone receptor status was defined as nuclear staining in at least 10% of invasive cancer cells. HER2 status had been tested by fluorescence in situ hybridization using the PathVysion kit (Vysis, Downers Grove, Ill) or by immunohistochemical staining with a monoclonal antibody (AB8; NeoMarkers, Fremont, Calif).
The chi-square test or the Fisher exact test was used to evaluate associations between EZH2 status and clinicopathologic variables. We used the Kaplan-Meier method to estimate OS and disease-specific survival (DSS). OS was calculated from the date of initial pathologic diagnosis of the primary tumor to the date of death from any cause or the date of last follow-up. For the DSS analysis, only disease-related death was considered an event. The patients who survived or who died from other causes were censored at the date of last follow-up or death, respectively. A Cox proportional hazards regression model was then used to test the statistical significance of several potential prognostic factors for OS and DSS. The factors analyzed included EZH2 expression; age; race; lymph node status; histologic type; tumor grade; lymphovascular invasion; ER, PR, and HER-2 status; and triple-negative (ER-negative, PR-negative, and HER2 negative) status. This modeling was done in a univariate fashion. From this model, we estimated the hazard ratio for each potential prognostic factor with a 95% confidence interval (CI). Then, all potential prognostic factors with a P value < .10 from the univariate analysis were included in a saturated model, and backward elimination was used to remove factors from the model based on the likelihood ratio test in the multiple regression analysis. The analysis was performed using SAS (version 9.2; SAS Institute Inc., Cary, NC). All reported P values were 2-sided, and P < .05 was considered statistically significant.
Of the 88 patients who were included in this study, 74 (84.1%) were whites, 10 (11.4%) were Hispanics, and 4 (4.5%) were blacks and Asians. Age at the time of initial diagnosis ranged from 23 years to 75 years (median age, 49.5 years). Sixty-seven patients had stage IIIB disease, 13 had stage IIIC disease, and 8 had stage IVdisease. The median follow-up was 45.7 months (range, 0.2-164.1 months). Lymph node involvement was observed in 67 of 80 patients (83.8%). Histologically, 78 of 88 tumors (88.6%) were ductal type, 60 of 86 (68.9%) were high grade tumors, and 65 of 74 (87.8%) demonstrated lymphovascular invasion.
Positive ER expression was observed in 37 of 86 of tumors (43%), positive PR status was observed in 30 of 86 tumors (34.9%), and HER2 overexpression and/or amplification was observed in 33 of 85 tumors (38.8%). Triple-negative status was observed in 21 of 88 tumors (23.9%). Results of EZH2 staining were available for 74 tumors, of which 56 (75.7%) were EZH2-positive with mostly moderate-to-strong staining intensity; in the remaining tumors, the tissue cores either were missing during staining or contained too few tumor cells to be interpreted properly. Figure 1 provides images of positive and negative staining results in representative tumors.
EZH2 Expression and Clinicopathologic Variables
To determine whether EZH2 expression was associated significantly with other clinicopathologic variables that have known potential prognostic value, EZH2 was evaluated in relation to different factors (Table 1). Positive EZH2 expression was associated significantly with higher tumor grade (P = .03) and negative ER status (P < .0001). Specifically, EZH2 was expressed more frequently in high-grade tumors (84%; 43 of 51 tumors) than in low-grade and intermediate-grade tumors (57%; 12 of 21 tumors) and also was expressed more frequently in ER-negative tumors (93%; 40 of 43 tumors) than in ER-positive tumors (53%; 16 of 30 tumors). A significant association also was observed between positive EZH2 expression and triple-negative status (P = .004), and all triple-negative tumors were EZH2-positive.
Table 1. Relation Between Enhancer of Zeste Homolog 2 Status and Clinicopathologic Factors
EZH2 Expression Status: No. of Patients (%)
Abbreviations: ER, estrogen receptor; EZH2, enhancer of zeste homolog 2; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; PR, progesterone receptor.
These P values were calculated using the Fisher exact test (other P values were calculated using the chi-square test).
There was a trend toward an association between positive EZH2 expression and ductal histologic type (P = .06) and between positive EZH2 expression and positive lymph node status at surgery (P = .10) (Table 1). EZH2 expression was not associated significantly with lymphovascular invasion, PR status, or HER2 status.
EZH2 Expression and Patient Outcome
Of the 85 patients who had follow-up information available, 52 patients had died (46 died of IBC, 1 died of another cause, and 5 died of unknown causes), and 33 patients remained alive. The median OS duration was 4.04 years (95% CI, 2.85-8.79 years). The 5-year and 10-year OS rates for the entire cohort of patients were 45% and 37%, respectively. The median DSS duration was 4.56 years (95% CI, 3.17-10.17 years). The 5-year and 10-year DSS rates for the entire cohort of patients were 49% and 40%, respectively.
In univariate analysis, positive EZH2 expression was associated significantly with a lower OS rate (P = .01) and a lower DSS rate (P = .03) (Fig. 2). The 5-year OS rate for patients who had EZH2-positive tumors was 32% compared with 64% for patients who had EZH2-negative tumors, and the corresponding 5-year DSS rates were 36% and 65%, respectively (Table 2).
Table 2. Relation Between Overall and Disease-Specific Survival, Enhancer of Zeste Homolog 2 Status, and Clinicopathologic Factors
Univariate analyses were performed to determine whether any other clinicopathologic factors were associated with the clinical outcome of patients with IBC. We observed that lower OS and DSS rates were associated significantly with negative ER status (P = .007 and P = .02, respectively) and with triple-negative status (P = .01 and P = .02, respectively) (Table 2). There was no significant association between OS and DSS rates and histologic type, tumor grade, lymphovascular invasion, lymph node status, PR status, or HER2 status. In multivariate analysis, we observed that only EZH2 status remained an independent predictor of OS (hazard ratio, 2.74; 95% CI, 1.22-6.15; P = .01), and DSS (hazard ratio, 2.44; 95% CI, 1.08-5.54; P = .03)
Despite advances in the management of IBC over the past 3 decades, the survival of patients with IBC remains much lower than that of patients with non-IBC. The identification of novel molecular markers that constitute the basis for target therapy is highly desirable.
Previous experiments demonstrated that the EZH2 expression level increases as breast cancer develops and progresses14, 16, 27, 31 and that EZH2 is a marker of the aggressiveness of breast carcinoma-NOS. In the current study of IBC, an “intrinsically” aggressive form of breast cancer associated with advanced stage and frequent metastasis, we observed that positive EZH2 status remained predictive of a relatively worse clinical outcome in patients with IBC. To our knowledge, this is the first study to explore the clinical significance of EZH2 expression in IBC. It is noteworthy that the tumors included in this study represented refractory or residual IBC tumors after neoadjuvant systemic therapy.
We observed that EZH2 was expressed frequently in this cohort of tumors (75.7%) and was associated significantly with some unfavorable prognostic factors, including higher tumor grade, negative ER status, and triple-negative status. We also observed a trend suggesting an association between EZH2 expression and positive lymph node status at surgery. To our knowledge, the association between EZH2 expression and triple-negative status has not been evaluated previously, although a relation between EZH2 expression and the basal-like phenotype in breast carcinoma-NOS was described by Collett et al.13 We did not observe a significant association between EZH2 expression and HER2 status; this finding is consistent with a report by Kleer et al14 but is in contrast to the results reported by Collett et al.13
In studies of breast carcinoma-NOS, controversy exists regarding the value of EZH2 in predicting survival. Kleer et al reported that higher EHZ2 protein levels were associated with a shorter disease-free interval after initial surgical treatment, a lower OS rate, and a higher probability of disease-specific death and that EZH2 status was an independent predictor of survival.14 Collett et al observed a significant association between EZH2 expression and reduced survival in univariate (but not multivariate) analysis.13 Raaphorst et al failed to identify a prognostic value of EZH2 in a Kaplan-Meier survival analysis of 172 patients.16 In a cohort of 688 patients with lymph node-negative breast cancer who did not receive adjuvant systemic therapy, Reijm et al reported that EZH2 was not associated significantly with metastasis-free survival; however, those authors observed that higher EZH2 expression was associated significantly with shorter progression-free survival in 278 patients who received first-line tamoxifen monotherapy for recurrent disease.15
In the current study with IBC, we observed that positive EZH2 status was associated significantly with a worse OS and DSS in both univariate and multivariate analyses, indicating that EZH2 positivity can be used to identify a subset of patients with IBC who have relatively worse clinical outcomes. Other factors that were predictive of a poor clinical outcome in univariate analysis were negative ER status and triple-negative status. It is interesting that tumor grade, lymphovascular invasion, and lymph node status did not predict worse survival. This probably is attributable to the fact that IBC, unlike non-IBC, is characterized “intrinsically” by a high frequency of high tumor grade, lymphovascular invasion, and lymph node involvement.
In agreement with findings by others,13, 14 we observed that EZH2 status was correlated inversely with ER status. The ER-positive rate in the current study (43%) was lower than the rate in breast carcinoma-NOS. This low ER-positive rate may have resulted from the high frequency of EZH2 expression in IBC tumors. Functional interactions between EZH2 and ER were investigated in previous studies. Reijm et al reported that EZH2 mRNA expression levels were associated inversely with ER protein levels in 1318 breast cancer specimens; those authors also observed that the down-regulation of EZH2 mRNA expression in the ER-positive human breast cancer cell line MCF7 using small interfering RNA was associated significantly with the up-regulation of ER.15 On the other hand, EZH2 and ER may function independently. For example, a combination of EZH2 silencing and antiestrogen ICI 164384 treatment caused 67% inhibition of MCF7 cell growth compared with controls; this reflected an adding-up effect of EZH2 silencing alone (38% inhibition) and ICI 164384 treatment alone (25% inhibition), thus indicating that the effect of EZH2 silencing on MCF7 growth is independent of the effect of ER inhibition by ICI 164384.15 Clinically, an association of EZH2 with outcome was observed both in patients with ER-positive cancer and in patients with ER-negative breast cancer.14 Recently, a study demonstrated a critical role of EZH2 in the proliferation and growth of ER-negative breast cancer both in an animal model and in vitro: EZH2 knockdown decreased the volume of tumors and improved survival in mice and delayed the transition from G2 phase to mitosis.28
The mechanisms underlying EZH2 biology in breast cancer remain to be elucidated. EZH2 exerts its biologic role mainly as a transcriptional repressor by forming a polycomb-repressive complex with other PcG proteins and regulating a variety of genes involved in cell fate, differentiation, and proliferation.31-33 EZH2 has histone lysine methyltransferase activity and trimehylates lysine residue 27 in histone 3 (H3K27), a mechanism of epigenetic silencing. In addition, EZH2 can transactivate genes that are commonly targeted by estrogen and Wnt signaling pathways; EZH2 interacts with components of the estrogen and Wnt pathways (ERα and β-catenin) and connects the estrogen and Wnt signaling circuitries.34 EZH2 also appears to regulate a broad cohort of genes implicated in immunity and inflammation network,35 DNA repair mechanisms,30, 36 and p53-regulated cell cycle control;37 and EZH2 is linked functionally with breast cancer 1 (BRCA1) expression.28
In view of the biologic role of EZH2 and the high frequency of EZH2 expression in IBC tumors, blockade of EZH2 may offer a promising therapeutic approach for these patients. One anti-EZH2 strategy may be to manipulate microRNAs (miRNAs) that are involved in EZH2 regulation. For example, miRNA-101 reportedly decreased the invasiveness of prostate cancer cells,38 and diminished miRNA-101 expression led to increased EZH2 expression and dysregulation of epigenetic pathways.39 Furthermore, miRNA-26a, which targets EZH2, reportedly inhibited cancer cell proliferation and induced tumor-specific apoptosis in a mouse model of hepatocellular carcinoma.40, 41
In conclusion, the current results indicated that EZH2 is expressed frequently in IBC tumors and is associated not only with unfavorable prognostic variables but also with significantly decreased survival. These findings indicate that the clinical outcome of patients with IBC can be stratified further on the basis of EZH2 status. Our results extend previous studies in patients with breast carcinoma-NOS. Given the high frequency of expression of EZH2 in IBC tumors, targeting EZH2 may provide a novel way to improve the clinical outcome of patients with IBC. A further study that includes patients with pretreated IBC tumors is warranted to further refine these findings.
This study was supported by the faculty research fund of The University of Texas MD Anderson Cancer Center (to Y.G.).