BRCA1 promoter hypermethylation in sporadic epithelial ovarian carcinoma: Association with low expression of BRCA1, improved survival and co-expression of DNA methyltransferases

The breast cancer susceptibility gene 1 (BRCA1) inactivation in sporadic epithelial ovarian carcinoma (EOC) is common and low BRCA1 expression is associated with promoter hypermethylation. The clinical validation of BRCA1 methylation as a prognostic marker in EOC remains unresolved. The aim of the present study was to determine the aberrant promoter methylation of BRCA1 in benign and malignant ovarian tumor tissues, to establish the association with the clinicopathological significance and the prognostic value. Additionally, the contribution of DNA methyltransferase (DNMT) expression to BRCA1 promoter hypermethylation was determined. The rate of BRCA1 methylation was observed to be 35.2% (50/142) in the EOCs; however, no methylation (0/32) was observed in the benign tumors. BRCA1 methylation was significantly associated with the downregulation of BRCA1 expression (P<0.001) and the frequency of BRCA1 methylation was greater in the carcinomas of patients whose tumor was bilateral than that of patients with a unilateral carcinoma (P=0.015). BRCA1 methylation was significantly associated with the preoperative serum carbohydrate antigen-125 level (P=0.013), improved overall survival (P=0.005) and disease-free survival (P=0.007). In addition, a significant correlation was observed between the co-expression of DNMTs and the methylation status of BRCA1. Thus, the present study provided support for BRCA1 promoter hypermethylation as a prognostic marker for survival in sporadic EOC, and co-expression of DNMTs was observed to contribute to BRCA1 promoter hypermethylation.


Introduction
Epithelial ovarian carcinoma (EOC) is a predominant, lethal gynecologic malignancy, with a five-year survival rate of <25% for patients that are diagnosed with stage III-IV disease (1,2). The treatment for EOC remains challenging, regardless of advances in surgical procedures and chemotherapy. The current ovarian cancer classification scheme distinguishes tumors based on the histopathological subtype, grade and surgical stage. Although established differences are observed in the clinical behavior, all of the ovarian neoplasia are subject to the same treatment paradigm. Therefore, the development of predictive markers that will direct the treatment selection, in addition to novel targeted therapies, are required for the treatment of EOC.
Aberrant DNA methylation is currently recognized as a common molecular abnormality in cancer. The abnormal promoter methylation may be a potential molecular marker for cancer diagnosis, prognosis and treatment (3)(4)(5). Recent studies have focused on the identification of methylated genes and the evaluation of their potential application as biomarkers for the prognosis and diagnosis of EOC (6,7). However, it is particularly unclear which genes should be used to identify a methylator phenotype that may become instrumental in the clinical diagnosis, prognosis and treatment of cancer.
The breast cancer susceptibility gene 1 (BRCA1) is a significant breast and ovarian cancer susceptibility gene; it has been mapped to chromosome 17q21 and encodes a nuclear protein, which is comprised of 1863 amino acids (8). BRCA1 contributes to the regulation of transcriptional activation, DNA repair, apoptosis, cell-cycle checkpoint control and chromosomal remodeling (9,10). BRCA1 dysfunction, frequently observed in high-grade serous ovarian carcinomas, commonly results from germline and somatic mutations, as well as promoter methylation. Therefore, the identification of tumors exhibiting BRCA1 defects has therapeutic and prognostic implications (11)(12)(13)(14)(15)(16). The hypermethylation of the BRCA1 promoter is important in silencing BRCA1 in sporadic EOCs, and previous studies have demonstrated that methylation of BRCA1 is associated BRCA1 promoter hypermethylation in sporadic epithelial ovarian carcinoma: Association with low expression of BRCA1, improved survival and co-expression of DNA methyltransferases with chemotherapy sensitivity and patient survival (17)(18)(19)(20). BRCA1 CpG island hypermethylation predits sensitivity to poly (ADP-ribose) polymerase-1 inhibitors (21), thus indicating a potential role for BRCA1 methylation as a biological marker in the clinical treatment of EOC.
DNA methyltransferases (DNMTs) are a family of enzymes responsible for the transfer of methyl groups to cytosine (22,23) and the DNMT family includes DNMT1, 2, 3a, 3b and 3L (24,25). DNMT2 (whose substrate and DNA methylation activity is unclear) (26), was shown to methylate transfer RNA (27,28) and DNMT3L (which is essential for establishing maternal genomic imprints, however, lacks key methyltransferase motifs) is potentially a regulator of methylation rather than an enzyme that methylates DNA (29). It has been suggested that that global genomic DNA methylation patterns are established and predominantly maintained by the combined action of three enzymatically active DNMTs, 1, 3a and 3b. DNMT1 is traditionally referred to as a maintenance enzyme as it copies methylation following replication, whereas DNMT3a and DNMT3b are de novo enzymes, which establish novel patterns of methylation during differentiation (30,31). Aberrant expression of DNMTs and disruption of DNA methylation patterns are closely associated with numerous types of cancer. Our previous study demonstrated that DNMT3a expression was higher in EOC (32), indicating an important role for the differential expression of DNMTs in aberrant DNA methylation patterns of EOC. However, there is currently no data concerning the association between DNMT expression and DNA methylation in EOC.
In the present study, the expression and methylation of BRCA1 was investigated in the same cohort of ovarian tumor patients as our previous study (32) and the association between BRCA1 expression and methylation was analyzed. The clinicopathological and prognostic differences, based on BRCA1 promoter methylation in EOC patients, were analyzed to determine the role of BRCA1 methylation in EOC and to understand the clinical significance. Furthermore, the association between BRCA1 methylation and DNMT protein expression was analyzed to investigate the role of DNMTs in BRCA1 methylation.

Patients and methods
Patients and tissue samples. The present study included 142 EOC cases and 32 ovarian benign tumor cases, which were selected from the Department of Surgical Oncology and General Surgery at the China Medical University-Affiliated First and Second Hospitals (Liaoning, China). All of the cases included in the present study were clinical cases that were routinely examined and diagnosed between 2002 and 2010. None of the patients had a family history of cancer and they were surgically staged according to the current International Federation of Gynecologists and Obstetricians (FIGO) classification system. The histological diagnosis was determined based on criteria set out by the World Health Organization (33). Approval for the present study was obtained from the Institute Research Medical Ethics Committee of the China Medical University (Liaoning, China). Consent was obtained from the families of all patients.
Immunohistochemical analysis. Formalin-fixed and paraffin-embedded tissue samples were cut into 4-µm sections and mounted onto poly-L-lysine-coated glass slides. For immunohistochemical staining, the sections were deparaffinized in xylene, rehydrated in a series of alcohol and washed using tap water. The sections were autoclaved in 10 mM sodium citrate buffer (pH 6.0; China National Pharmaceutical Group Corporation, Beijing, China) for 10 min for antigen retrieval. Endogenous peroxidase activity was blocked by incubating the sections in 3% H 2 O 2 (China National Pharmaceutical Group Corporation) at 37˚C for 20 min. The sections were blocked for non-specific binding by 10% normal goat serum at 37˚C for 30 min and were incubated at 4˚C overnight with mouse monoclonal anti-BRCA1 antibodies (1:100; MS110, Calbiochem, Darmstadt, Germany). The following day, the sections were washed three times with 0.01 mol/l phosphate-buffered saline (PBS; pH 7.4) for 15 min and incubated with a goat anti-mouse IgG secondary antibody (catalogue no. sc-2039; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) for 30 min at 37˚C. Subsequently, the sections were incubated with a streptavidin horseradish peroxidase solution (LSAB TM kit, Dako, Glostrup, Denmark) for 30 min, washed with PBS and stained with 3,3'-diaminobenzidine. The sections were then counterstained with Mayer's hematoxylin (Sigma-Aldrich, St. Louis, MO, USA), dehydrated and mounted. The negative controls were generated using PBS as a replacement for the anti-BRCA1 antibodies.
Evaluation of immunohistochemistry. The immunostained sections were independently reviewed and scored by two investigators who were blinded to the clinicopathological characteristics of the patients; a positive correlation was observed between them. The nuclear positivity of the BRCA1 proteins was evaluated using semi-quantitative scoring criteria according to the staining intensity (0, negative; 1, weak; 2, moderate; and 3, strong) and the proportion of positive cells (0, negative; 1, positive in ≤10%; 2, positive in >10% and ≤50%; 3, positive in >50% and ≤80%; 4, positive in >80% of tumor cells). The two scores were multiplied together for each case and the expression was graded as: 0, negative score; 1-4, weak expression score; 5-8, moderate expression score; and 9-12, strong expression score.

Methylation-specific polymerase chain reaction (MSP).
The promoter methylation status of the BRCA1 gene was analyzed using MSP. The bisulfite modification, primer sequences and PCR conditions were performed as previously described (29). Lymphocyte DNA, which was treated with SssI bacterial methylase (New England BioLabs, Hithcin, UK) served as the positive control for the methylated alleles and the DNA from normal lymphocytes served as the control for the unmethylated alleles; the negative controls, without DNA, were included in each experiment. All PCR reactions were conducted in duplicate and a methylated band, which was detected in either or both duplicates, was recorded as positive for promoter methylation.
Statistical analysis. Pairwise correlations between the categorical variables were investigated using the χ 2 test or Fisher's exact test where appropriate. Overall survival (OS) and disease-free survival (DFS) were estimated using the Kaplan-Meier method, and the log-rank test was used to compare the patient survival time between or among the groups. The Cox proportional hazards model (Cox regression) was used with backward elimination to identify the significant, independent, prognostic factors. OS was defined as the time interval from initial surgery to mortality or, for surviving patients, the time interval between the initial surgery and the final follow-up. DFS was defined as the time interval between the initial surgery and the date of disease progression or recurrence, or the final follow-up. The statistical tests were two-sided and P<0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using the SPSS statistical software program (SPSS, Chicago, IL, USA).

Results
Patient characteristics. In the present study, the tissue sections were obtained from 174 ovarian tumor samples for the evaluation of BRCA1 protein expression and promoter methylation. The clinicopathological data from the patients are shown in Table I. Briefly, the mean age of the patients at the time of surgery was 53 years (range, 20-74 years). Twenty-seven (23.3%) patients exhibited lymph node-metastasized disease at the time of surgery and 110 (80.9%) patients exhibited serous carcinoma as the predominant histological diagnosis, followed by mucinous carcinoma (8.5%), clear cell carcinoma (5.6%) and undifferentiated carcinoma (5.6%). The preoperative serum levels of carbohydrate antigen (CA)-125 and CA19-9 were elevated prior to surgery (higher than in the levels of healthy individuals) in the majority of patients but none of the patients received any neo-adjuvant chemotherapy. Follow-up data were available for 85 patients. The mean and median OS times were 56. 1   the 142 carcinomas, however, no methylation was detected in the benign tumors. The frequency of BRCA1 methylation was significantly higher in the EOC samples than that observed in the benign tumor samples (P<0.001, Pearson's χ 2 test). The BRCA1 expression in the 142 carcinoma samples was as follows, 67 (47.2%) were negative and 62 (43.7%) were weakly positive, nine (6.3%) were moderately positive and four (2.8%) were strongly positive for BRCA1. Among the 32 cases of benign tumors; nine (28.1%) were negative and 15 (46.9%) were weakly positive, four (12.5%) were moderately positive and four (12.5%) were strongly positive for nuclear BRCA1. The expression of BRCA1 in the carcinoma samples was significantly reduced compared with that observed in the benign tumors (P=0.031, Pearson's χ 2 test).
In the EOC samples, a reduction of the BRCA1 protein was significantly associated with BRCA1 methylation (P<0.001, Pearson's χ 2 test; Table II). Representative results of the immunohistochemical staining and BRCA1 promoter methylation in EOC tissues are shown in Fig. 1.

BRCA1 promoter methylation is associated with clinicopathological parameters in EOC.
The association between BRCA1 promoter methylation and clinicopathological parameters is shown in Table III. The data indicated that BRCA1 methylation was significantly associated with tumor localization. In addition, the frequency of BRCA1 methylation was observed to be higher in the patients whose tumor was bilateral than that of patients whose tumor unilateral (P=0.015,Pearson's χ 2 test). Moreover, the methylation of the BRCA1 gene was marginally associated with the FIGO clinical stage (P=0.085, Pearson's χ 2 test).
CA-125, CA19-9 and carcinoembryonic antigen (CEA) are the predominant biomarkers for confirming the diagnosis and management of EOC. Although they commonly aid with Table II. Correlation between BRCA1 expression and BRCA1 methylation in 142 sporadic EOC cases.  the diagnosis of ovarian malignancies, there are significant limitations relating to their sensitivity and specificity. The association between BRCA1 gene promoter methylation and preoperative serum levels of CA-125, CA19-9 and CEA were analyzed in the present study and BRCA1 methylation was observed to be significantly associated with the CA-125 level. The frequency of BRCA1 methylation was greater in patients exhibiting higher preoperative CA-125 levels than that observed in patients exhibiting lower CA-125 levels (P=0.013, Pearson's χ 2 test).  Table III. Correlation between BRCA1 promoter methylation and clinicopathological features of sporadic EOC patients.    95% CI, 1.218-3.508) are significantly associated with a reduced OS and DFS (Table IV). Subsequently, multivariate Cox regression models using the clinical stage, tumor size, location of neoplasia and BRCA1 methylation revealed that the clinical stage alone provided the independent prognostic factor (OS: P=0.012;      (Table V).

Discussion
The loss of expression of tumor suppressor genes is known to occur via biallelic inactivation (35). The current consensus is that the somatic mutation of BRCA1 is rare in sporadic EOC, thus, the loss of BRCA1 expression is considered to be due to a combination of allelic loss and methylation or biallelic methylation (36,37). Promoter methylation is significant in silencing BRCA1. The varying experimental methods and, in particular, the region of the BRCA1 promoter that has been investigated across studies have identified a range of frequencies of BRCA1 promoter methylation, from 5 to 40% of clinical specimens (38). In the present study, the methylation rate of BRCA1 in EOC samples is consistent with that observed in previous studies and BRCA1 promoter methylation was significantly associated with a decreased expression of BRCA1. These results corroborated the hypothesis that BRCA1 hypermethylation was a predominant cause of BRCA1 loss of expression. Previous studies regarding the association between BRCA1 gene promoter methylation and patient survival remain controversial. Montavon et al (6) and Yang et al (39) indicated that BRCA1 methylation was not associated with patient survival, however, another study demonstrated a survival disadvantage in patients whose neoplasms were methylated at BRCA1 (17). In the present study, a significant association was observed between BRCA1 methylation and an improved survival rate in patients with an advanced FIGO stage (III-IV). The inconsistency of these results may be due to the varying regions of the BRCA1 promoter that were investigated in addition to the different populations that were involved in the studies. The data from the present study resulted in the hypothesis that the cells from the patients with methylation of the BRCA1 promoter may exhibit an impaired ability to repair DNA damage via downregulation of BRCA1 expression. The patients may experience an increased sensitivity to platinum chemotherapy, resulting in an improved survival outcome. These results were consistent with previous studies, which demonstrated that BRCA1 gene promoter methylation may be an effective indicator of the patient response to chemotherapy (20,21). As there were limitations in the present study due to the small number of cases, further investigation is required to clarify the role of BRCA1 gene methylation in chemoresponsiveness and patient survival in EOC cases.
The present study indicated that BRCA1 methylation was significantly associated with the tumor location. Furthermore, the frequency of BRCA1 methylation was observed to be greater in patients with bilateral ovarian cancer than in unilateral cancer patients, which may indicate a contrast between the unilateral and bilateral ovarian cancers regarding their biological characteristics, genetics and mechanisms of carcinogenesis. Moreover, these data indicated that BRCA1 methylation was significantly associated with the preoperative serum CA-125 levels. CA-125 was established as a prognostic marker in cancer, specifically in ovarian carcinoma (40) and the present results indicated that BRCA1 methylation and CA-125 levels, in combination, may be used as biomarkers for the diagnosis and prognosis of EOC.
The causes of DNA methylation at specific CpG islands in cancer are ambiguous and potentially multifactorial. DNMTs were a predominant cause of DNA methylation and the differential expression of DNMTs has been identified in numerous types of cancer (41)(42)(43), including EOC, which we previously reported (32). It was hypothesized in the present study that the over-expression of DNMT proteins may contribute to BRCA1 methylation; therefore, the association between BRCA1 methylation and DNMT protein expression was analyzed. The results indicated that there was no significant correlation identified between the methylation status of BRCA1 and the individual protein expression of DNMT1, 3a or 3b. However, the co-expression of DNMTs was identified to be significantly associated with BRCA1 methylation. The results of the present study supported the hypothesis that genomic methylation patterns may be established depending on the interaction of these three enzymes. Further studies are required to clarify how these DNMTs interact with each other.
In conclusion, BRCA1 gene promoter methylation was identified to be significant in reducing protein expression in a Chinese population and the co-expression of DNMTs may have contributed to BRCA1 gene promoter hypermethylation. Moreover, BRCA1 gene promoter hypermethylation was observed to be significantly associated with a favorable survival rate, thus supporting the hypothesis that BRCA1 methylation may be an important prognostic marker for EOC.