Estrogen-Responsive Finger Protein as a New Potential Biomarker for Breast Cancer Free

Breast cancer is the most common type of cancer and continues to be the most frequent cause of cancer-related deaths in women in the Western world. Whether or not human primary breast cancers are estrogen dependent is a critical factor that determines patient prognosis and availability of antiestrogenic endocrine therapy (1). Two thirds of breast carcinomas are positive for estrogen receptor α (ERα) and a great majority of these tumors initially respond to antiestrogens such as tamoxifen and aromatase inhibitors. However, it is also true that these ERα-positive breast carcinomas frequently acquire resistance to endocrine therapy, although ERα remains to be expressed (1, 2). The molecular mechanisms through which breast carcinomas become hormone-refractory are still largely unclear.

Identification and functional studies of ERα target molecules may provide a clue for understanding the mechanism that alters tumor phenotypes. We have previously isolated estrogen-responsive finger protein (Efp), which is a member of RING finger-B box-Coiled Coil family (3). Efp also is one of the downstream targets of ERα (3–6). Efp-deficient mice displayed underdeveloped uteri and reduced estrogen responsiveness (7), and therefore, Efp is considered to be essential for estrogen-dependent proliferation. It has also been shown that Efp promotes the growth of breast tumor by functioning as a ubiquitin ligase (E3) that targets the negative cell cycle checkpoint 14-3-3σ (8).

Expression of Efp was previously reported in breast carcinoma tissues at mRNA (5) and protein levels (9). However, information on the expression of Efp in human breast carcinoma tissues is still very limited, and the biological significance of Efp remains unclear at this juncture. Therefore, in this study, we examined expression of Efp in 30 cases of breast carcinoma tissues using immunohistochemistry and laser capture microdissection/real-time PCR. We subsequently examined immunolocalization of Efp in 151 cases of human breast carcinoma tissues and correlated these findings with various clinicopathologic factors including clinical outcome of the patients.

Patients and tissues. Thirty specimens of invasive ductal carcinoma were obtained from patients who underwent mastectomy in 2001 in the Department of Surgery at Tohoku University Hospital, Sendai, Japan. Specimens for RNA isolation were snap-frozen and stored at −80°C, and those for immunohistochemistry were fixed with 10% formalin and embedded in paraffin wax. Informed consent was obtained from all patients before their surgery and examination of specimens used in this study.

One hundred fifty-one specimens of invasive ductal carcinoma of the breast were obtained from female patients who underwent mastectomy from 1982 to 1989 in the Department of Surgery, Tohoku University Hospital, Sendai, Japan. All specimens were fixed with 10% formalin and embedded in paraffin wax, and snap-frozen tissues were not available for examination in these cases. These patients did not receive any preoperative radiotherapy and chemotherapy as well as any postoperative hormone therapy. Information on patient age, menopausal status, stage, tumor size at operation, lymph node status, histologic grade, and relapse and survival times was retrieved from the review of patient charts. The mean follow-up period was 105 months (3-157 months).

Research protocols for this study were approved by the Ethics Committee at Tohoku University School of Medicine.

Immunohistochemistry. Anti-human Efp antibody was generated as previously described (3). Polyclonal antibody for 14-3-3σ (N-14) and monoclonal antibodies for ERα (1D5), progesterone receptor (MAB429), Ki-67 (MIB-1), and p53 (DO7) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), Immunotech (Marseille, France), Chemicon (Temecula, CA), DAKO (Tokyo, Japan), and Novocastra Laboratories (Newcastle, United Kingdom), respectively. A Histofine Kit (Nichirei, Tokyo, Japan), which employs the streptavidin-biotin amplification method, was used for immunohistochemistry, and the antigen-antibody complex was visualized with 3.3′-diaminobenzidine solution [1 mmol/L 3,3′-diaminobenzidine, 50 mmol/L Tris-HCl buffer (pH 7.6), and 0.006% H₂O₂]. For a negative control for Efp immunohistochemistry, an immunohistochemical preabsorption test was done.

Efp immunoreactivity was classified into three groups: ++, >50% positive carcinoma cells; +, 1% to 50% positive cells; and −, no immunoreactivity, according to a previous report (10). Immunoreactivity of ERα, progesterone receptor, and Ki-67 was scored in more than 1,000 carcinoma cells for each case, and the percentage of immunoreactivity [i.e., labeling index (LI)] was determined. Cases that were found to have ERα LI of more than 10% were considered ERα-positive breast carcinomas (11).

Laser capture microdissection/real-time PCR. Laser capture microdissection was conducted using the Laser Scissors CRI-337 (Cell Robotics, Inc., Albuquerque, NM). A detailed procedure has been described previously (12, 13). Briefly, ∼1,000 carcinoma or intratumoral stromal cells were separately collected under the microscope from breast carcinoma frozen tissue sections embedded in Tissue-Tek O.T.C. The Light Cycler System (Roche Diagnostics GmbH, Mannheim, Germany) was used to semiquantify the level of Efp mRNA expression in this study. The primers used for real-time PCR were the following: Efp sense, 5′-CGTGGAGTGGTTCAACAC-3′, and Efp antisense, 5′-GAGCAGATGGAGAGTGTGG-3′; glyceraldehyde-3-phosphate dehydrogenase sense, 5′-TGAACGGGAAGCTCACTGG-3′, and glyceraldehyde-3-phosphate dehydrogenase antisense, 5′-TCCACCACCCTGTTGCTGTA-3′. To verify amplification of the correct sequences, PCR products were purified and subjected to direct sequencing. Efp mRNA levels were normalized to those of glyceraldehyde-3-phosphate dehydrogenase, and subsequently, the fold change of Efp mRNA level in each sample was evaluated using the mRNA level in MCF7 cells as a positive control. Negative control experiments lacked cDNA substrate to check for the possibility of exogenous contaminant DNA, and no amplified products were detected under these conditions.

Statistical analyses. Statistical analyses were done using one-way ANOVA and Bonferroni test or a cross-table using χ² test. Overall and disease-free survival curves were generated according to the Kaplan-Meier method, and statistical significance was calculated using log-rank test. Univariate and multivariate analyses were evaluated by a proportional hazard model (Cox) using PROC PHREG in our SAS software. Differences with P < 0.05 were considered significant.

Expression of Efp in 30 breast cancer tissues: Immunohistochemistry. Efp immunoreactivity was detected in the cytoplasm of breast carcinoma cells, but not in intratumoral stromal cells (Fig. 1A and B). The number of cases expressing immunoreactive Efp in each group of 30 breast carcinoma tissues is summarized as follows: ++, n = 9 (30.0%); +, n = 13 (43.3%); and −, 17 (56.7%). Efp immunoreactivity was also positive in epithelial cells of morphologically normal mammary glands (Fig. 1C). Immunohistochemical preabsorption test for Efp showed no specific immunoreactivity in a negative control (Fig. 1D).

Laser capture microdissection/real-time PCR. To examine the localization of Efp mRNA in breast carcinoma tissues, we did laser capture microdissection/real-time PCR analyses. Expression of Efp mRNA was detected in carcinoma cells, but not in intratumoral stromal cells (Fig. 2A). As shown in Fig. 2B, a significant association was detected between Efp immunoreactivity and Efp mRNA level (P = 0.0012, ++ versus − and ++ versus +, respectively) in 30 cases of breast cancer tissues examined.

Correlation between Efp immunoreactivity and clinicopathologic variables in 151 breast carcinomas. Results of associations between Efp immunoreactivity and clinicopathologic variables in 151 breast carcinomas were summarized in Table 1. The number of cases expressing immunoreactive Efp in each group is summarized as follows: ++, n = 46 (30.5%); +, n = 64 (42.4%); and −, n = 41 (27.2%). Efp immunoreactivity was significantly associated with lymph node status (P = 0.0027), ERα status (P = 0.0013), or ERα LI (P = 0.0023, ++ versus −; P = 0.0045, + versus −). On the other hand, negative correlation was detected between Efp immunoreactivity and histologic grade (P = 0.0064) or 14-3-3σ immunoreactivity (P < 0.0001). There was, however, no significant relationship between Efp immunoreactivity and other clinicopathologic variables, including patient age, menopausal status, stage, tumor size, progesterone receptor LI, Ki-67 LI, and 53 status, in this study. Similar tendencies described above were confirmed in increased rankings of positivity for Efp immunoreactivity into five groups (0%, 1-25%, 26-50%, 51-75%, and 76-100% positive cells; data not shown).

Correlation between Efp immunoreactivity and clinical outcome of the 151 breast cancer patients. Efp immunoreactivity was significantly associated with an increased risk of recurrence (P < 0.0001; Fig. 3A). A similar tendency was also detected when Efp immunoreactivity was further categorized into five groups (0%, 1-25%, 26-50%, 51-5%, and 76-100% positive cells; Fig. 3B). Following univariate analysis by Cox (Table 2), lymph node status (P < 0.0001), Efp immunoreactivity (P < 0.0001), tumor size (P = 0.0019), and 14-3-3σ immunoreactivity (P = 0.0314) were shown as significant prognostic variables for disease-free survival in 151 breast carcinoma patients examined. A multivariate analysis revealed that lymph node status (P < 0.0001), Efp immunoreactivity (P = 0.0011), and tumor size (P = 0.0349) were independent prognostic factors with relative risks over 1.0, whereas 14-3-3σ immunoreactivity was not significant (P = 0.0681; Table 2).

Overall survival curve was shown in Fig. 3C. A significant correlation was detected between Efp immunoreactivity and adverse clinical outcome of the patients (P < 0.0001), and a similar tendency was also detected when Efp immunoreactivity was categorized into five groups (Fig. 3D). Using a univariate analysis (Table 3), Efp immunoreactivity (P < 0.0001), lymph node status (P < 0.0001), tumor size (P = 0.0084), and p53 status (P = 0.0230) turned out to be significant prognostic factors for overall survival in this study. However, multivariate analysis revealed that only Efp immunoreactivity (P = 0.0030) and lymph node status (P = 0.0065) were independent prognostic factors with a relative risk over 1.0; other factors were not significant in this study (Table 3).

The significant association between Efp immunoreactivity and clinical outcome of the patients was detected regardless of ERα status in this study (Fig. 4A-D).

In this study, Efp immunoreactivity was significantly associated with Efp mRNA level, and was detected in carcinoma cells in 110 of 151 human breast carcinomas (72.8%). Previously, Ikeda et al. (5) reported Efp mRNA expression in 9 of 15 (60.0%) breast carcinoma tissues using RNase protection assay, and Thomson et al. (9) reported Efp immunoreactivity in breast carcinoma cells in 64 of 91 (70.3%) cases. The frequency and cellular localization of Efp in our present study were in good agreement with these reports (5, 9), and widespread distribution of Efp may suggest an important role of Efp in breast carcinomas. Efp immunoreactivity was also detected in normal glandular epithelia in our study. It is also consistent with previous studies (5, 9) and Thomson et al. (9) suggested the involvement of Efp in mammary gland differentiation.

Efp is known as a downstream product of ERα (3–7). Efp gene has an estrogen-responsive element at the 3′-untranslated region (3). The estrogen-responsive element of Efp responded to ERα in transfected estrogen receptors in 293T cells (5), and Efp mRNA was rapidly induced by estrogen treatment within 0.5 hour in MCF7 cells (5). In this study, Efp immunoreactivity was significantly associated with ERα status and ERα LI in 151 breast carcinoma tissues. Therefore, it is suggestive that Efp is mainly produced in carcinoma cells through ERα as a result of estrogenic action in breast carcinoma. On the other hand, we also found Efp immunoreactivity in 22 of 42 ERα-negative breast carcinomas. It may be partly explained that Efp expression was induced by a low or undetectable level of ERα in these cases. However, Ikeda et al. (14) analyzed human 5′-flanking region of human Efp gene, and reported the possible regulation of Efp promoter by multiple elements and/or interacting factors. Therefore, other factors rather than ERα may be also involved in the expression of Efp in some breast carcinomas.

In our study, Efp immunoreactivity was significantly associated with an increased risk of recurrence or worse prognosis (P < 0.0001, respectively). Both univariate and multivariate analyses have shown that Efp immunoreactivity was a potent prognostic factor for both recurrence and overall survival in breast carcinomas, and that the effect is similar to that of lymph node status, a well-established diagnostic modality (15). Efp knockout mice showed a smaller increase in uterine weight and a lower cell cycle progression from G₀/G₁ to S phase compared with the wild-type (7), suggesting a pivotal role of Efp in ERα-induced cell growth in the uterus. In addition, Urano et al. (8) showed that overexpression of Efp caused tumor cell growth in MCF7 breast cancer cells. Therefore, taken together with these previous reports and our present results, it is suggested that Efp plays an important role in the proliferation of breast carcinoma cells. It is well known that biologically active estrogen, estradiol, is locally produced in breast carcinoma tissues from circulating inactive steroids, and acts on these cells via ERα (16). Therefore, residual cancer cells following surgical treatment in Efp-positive breast carcinomas may grow rapidly in the presence of local estrogens, thereby resulting in an increased recurrence and poor prognosis in these patients.

14-3-3σ induces G₂ arrest and inhibits the progression of cell cycle (17) by sequestering the mitotic initiation complex Cdc2-cyclin B1 in the cytoplasm, blocking nuclear entry (18). Expression of 14-3-3σ was examined by several groups in breast carcinoma tissues. However, results of these studies seem to be inconsistent. Ferguson et al. (19) reported that 14-3-3σ mRNA was detected only in 3 of 48 (6.3%) breast carcinoma tissues by Northern blot analysis. Ferguson et al. (19) and Umbricht et al. (20) showed hypermethylation of CpG islands in the 14-3-3σ gene in more than 90% of breast carcinomas, and postulated that loss of 14-3-3σ expression was an early event in neoplastic transformation of the breast. Simooka et al. (21) detected 14-3-3σ immunoreactivity in 23% of invasive ductal carcinomas and reported that loss of 14-3-3σ expression was relatively low compared with the methylation status of 14-3-3σ gene in breast carcinoma previously reported. On the other hand, Urano et al. (8) showed that 14-3-3σ is a primary target for proteolysis by Efp, and 14-3-3σ protein was regulated by Efp-mediated posttranslational modification. However, Moreira et al. (22) recently reported that expression level of 14-3-3σ was similar in nonmalignant breast epithelial tissue and matched malignant tissue with only sporadic loss of expression observed in 3 of the 68 (4.4%) tumors examined. The lack of expression of 14-3-3σ in the three breast carcinomas was not associated with increased expression of Efp, and they suggested that loss of expression of 14-3-3σ protein was a sporadic event in the breast carcinoma (22). In our present study, immunoreactivity of 14-3-3σ was detected in 58 of 151 (38.4%) breast carcinomas, and was inversely associated with Efp immunoreactivity. These results seem to support the down-regulation of 14-3-3σ by methylation of the gene and/or proteolysis by Efp in breast carcinoma tissues; however, these are not necessarily consistent with the findings by Moreira et al. (22). Further examinations, including validation of the immunohistochemical results by another laboratories, are required to clarify the expression of 14-3-3σ in breast carcinoma tissues.

In summary, Efp immunoreactivity was detected in carcinoma cells in 72.8% of breast cancer tissues, and it was associated with the mRNA level. Efp immunoreactivity was significantly associated with lymph node status or ERα status, and was inversely correlated with histologic grade or 14-3-3σ immunoreactivity. Moreover, Efp immunoreactivity was significantly associated with poor clinical outcome of the patients. These present results suggest that Efp is mainly involved in the estrogen-dependent growth of breast carcinomas, and Efp immunoreactivity is a potent prognostic factor in breast carcinoma patients.

We appreciate the skillful technical assistance of Dr. Yasuhiro Miki (Department of Pathology, Tohoku University School of Medicine, Sendai, Japan) and Toshiki Hishinuma (Saitama Medical School, Hidaka, Japan). We thank Dr. Bruce Blumberg (Department of Developmental and Cell Biology, University of California, Irvine, CA) for critical reading and comments on the manuscript.