Harmful effect of ERβ on BCRP-mediated drug resistance and cell proliferation in ERα/PR-negative breast cancer



The role of estrogen receptor β (ERβ) in breast cancer is still under investigation. Various studies have provided evidence that ERβ behaves as a tumor suppressor in breast cancer, whereas some studies of estrogen receptor α (ERα) negative breast cancer reported a positive correlation between high ERβ expression and poor prognostic phenotypes, such as induced proliferation, invasion and metastasis. In the present immunohistochemistry study of 99 ERα/progesterone receptor (PR)-negative breast cancer samples, nuclear expression of ERβ was positively associated with membranous expression of breast cancer resistance protein (BCRP), Ki67 (proliferation marker) and tumor size. Moreover, both endogenous and exogenous ERβ upregulated BCRP expression which induced BCRP-mediated drug resistance and enhanced proliferation of ERα−/PR− breast cancer cells in the presence of 17β-estradiol, whereas these effects were reversed by additional use of tamoxifen (TAM). In addition, the regulation of BCRP via specific binding between ERβ and estrogen response element (ERE) was demonstrated in an electrophoretic mobility shift assay. Overall, our findings manifest that ERβ might act as a tumor promoter of cell proliferation and BCRP-mediated drug resistance in ERα−/PR− breast cancer. TAM routinely used for patients with ERα+/PR+, ERα+/PR− and ERα−/PR+ breast cancer might also be effective in ERα−/PR− but ERβ+ breast cancer. Therefore, the detection of ERβ in clinic is valuable and should not be neglected in breast cancer, especially for the ERα−/PR− phenotype.


ATP-binding cassette


breast cancer resistant protein






electrophoretic mobility shift assay


estrogen receptor


estrogen response element


estrogen receptor α


estrogen receptor β


50% inhibitory concentration


multidrug resistance




proliferating cell nuclear antigen




progesterone receptor


Breast cancer continues to be a leading cause of women's death worldwide. It is well known that estrogen exposure is closely associated with breast cancer in both epidemiological and experimental studies [1]. One potential explanation of this association is that estrogen-induced activity of estrogen receptors (ERs) stimulates tumor growth and development. ERs include estrogen receptor α (ERα) and estrogen receptor β (ERβ). The harmful role of ERα in estrogen-mediated breast cancer has been well accepted [2], whereas the effect of ERβ on breast cancer remains controversial [3-5]. Most studies have provided evidence that ERβ acts as a negative modulator of ERα and indicates a good prognosis with prolonged disease-free survival [6-8]. However, these studies have predominantly focused on ERα+ breast cancer and the role of ERβ co-expression with ERα. Several investigators have reported that there is a positive correlation between ERβ expression and poor prognostic phenotypes, such as accelerated proliferation and basal phenotype in ERα-negative breast cancer [9-11]. Moreover, the highest content of proliferating cells was seen in ERα−/ERβ+ cancer cells [12].

There have been few reports on the relationship between ERs and multidrug resistance (MDR) in breast cancer, particularly for ERβ. It is known that MDR as well as proliferation and invasion/metastasis represent malignant transformation of tumors. MDR is a major obstacle of successful chemotherapy and decreases the survival rate in patients with breast cancer [13, 14]. Breast cancer resistance protein (BCRP), as a member of the ATP-binding cassette (ABC) superfamily, has been verified to result in MDR to various chemotherapeutic agents, such as adriamycin, epirubicin, mitoxantrone (MX), topotecan, methotrexate, doxorubicin and flavopiridol [15, 16]. Its overexpression has been observed in many cell lines and tissue samples of breast cancer [17, 18]. Our previous study found that regulation of BCRP relied on ligand-ERα complex binding to the estrogen response element (ERE) of BCRP promoter via the classical genomic pathway [19, 20]. Therefore it is of interest to investigate whether ERβ also plays a role in regulation of BCRP expression.

Herein we investigated the effects of ERβ on BCRP and cellular proliferation in ERα−/progesterone receptor (PR)− breast cancer cell lines as well as tissue samples of patients, and investigated the underlying mechanism involved.


Expression of ERβ, BCRP, Ki67 and Her-2 protein in ERα−/PR− breast invasive ductal carcinoma tissue

Cell membrane staining was considered BCRP positive and nuclear staining of the cell was judged as ERβ positive in our experiment. As shown in Table 1 and Fig. 1, 49 of 99 patients were BCRP positive (49.49%), 57 of 99 were ERβ positive (57.58%), 45 of 99 were Her-2 positive (45.45%) and 60 of 99 were Ki67 positive (60.61%) (data on Her-2 and Ki67 were obtained from the Department of Pathology of Qilu Hospital of Shandong University). Expression of ERβ was positively associated with that of BCRP (= 0.025), Ki67 (= 0.012) and tumor size (= 0.024) in ERα−/PR− breast invasive ductal carcinoma, while the expression of ERβ was not associated with that of Her-2 (= 0.838) and lymph node metastasis (= 0.400) (Table 1).

Table 1. Results of immunohistochemistry
  n ERβP-value
  1. LN, lymph node.

Ki67 ≤13916230.012
> 1604119
Size ≤34520250.024
> 3543717
Figure 1.

Immunochemistry staining by anti-BCRP and anti-ERβ antibody to identify the expression of BCRP and ERβ in ERα/PR-negative breast invasive ductal carcinoma tissue. (A), (B) Nuclear staining was positive for ERβ (200 × and 400 ×, respectively). (C), (D) Nuclear staining was negative for ERβ (200 × and 400 ×, respectively). (E), (F) Membranous staining pattern of BCRP (200 × and 400 ×, respectively). (G), (H) Expression of BCRP was negative (200 × and 400 ×, respectively).

Upregulation of BCRP level by 17β-estradiol (E2) in a dose-dependent manner, which was antagonized by tamoxifen (TAM)

In the current study, we chose two kinds of ERα−/PR− human breast cancer cell lines, MDA-MB-453 (Her-2+) and MDA-MB-468 (Her-2−). MDA-MB-453 cells transfected with both BCRP and ERβ were treated with different concentrations of E2 (0, 10−13, 10−11 and 10−9 m) or a combination of E2 (10−9 m) and TAM (10−6 m) for 24 h to obtain total RNA and 72 h for proteins. RT-PCR and western blot analysis revealed that BCRP level was dose-dependently upregulated by physiological concentrations of E2 in 453/ERβ/P-BCRP cells (P refers to BCRP promoter including the ERE region), and the increase was significantly counteracted by the anti-estrogen TAM (Fig. 3A,B). However, because of the promoter genotypes, E2 did not lead to any change of BCRP level in 453/ERβ/C-BCRP cells (C refers to CMV promoter) (Fig. 3E,F). Moreover, the BCRP level was not affected by E2 in ERβ-negative MDA-MB-453, 453/P-BCRP and 453/C-BCRP cells (data not shown). These findings suggest that the effects of E2 depend on both the ERβ and BCRP promoter.

To confirm the association of endogenous ERβ and BCRP, we chose ERβ+/BCRP+ MDA-MB-468 cells to construct 468/SiN cells (SiN refers to pSUPER vector including non-targeting sequences, SiN as a negative control of SiERβ) and 468/SiERβ cells (knockdown of endogenous ERβ in MDA-MB-468 cells) (Fig. 2B). Treated as presented above, the BCRP level of 468/SiN cells was elevated by E2 treatment, which was antagonized by TAM (Fig. 3C,D). This result is consistent with that of the exogenous model of 453/ERβ/P-BCRP cells. In addition, we inhibited ERβ expression by transfecting SiERβ in MDA-MB-468 cells (468/SiERβ) and observed no dramatic change of basal BCRP expression, which implies that ERβ might have no effect on regulating BCRP in the absence of E2 (Fig. 2B). As expected, the functions of E2 and TAM were also lost after knockdown of ERβ in MDA-MB-468 cells (Fig. 3G,H).

Figure 2.

Detection of PR, ERα, ERβ and BCRP levels in MDA-MB-453 and MDA-MB-468 cells. (A) Expression of PR, ERα, ERβ and BCRP in MDA-MB-453 and transfected 453 cells. (B) Expression of PR, ERα, ERβ and BCRP in MDA-MB-468 and transfected 468 cells.

Figure 3.

BCRP level was regulated by E2 in a dose-dependent manner, which was antagonized by TAM. Cells were cultured in a range of concentrations of E2 or combined treatment with E2 and TAM. (A)–(D) E2 elevated mRNA and protein levels of BCRP, and TAM antagonized this effect in MDA-MB-453/ERβ/P-BCRP and MDA-MB-468/SiN cells. (E)–(H) No significant change was observed in MDA-MB-453/ERβ/C-BCRP and MDA-MB-468/SiERβ cells under the same experimental conditions. (I)–(L) Densitometry analysis was then performed. *< 0.05 and **< 0.01 versus the corresponding untreated control cells. #< 0.05 and ##< 0.01 versus the corresponding cells with 10−9 m E2 treatment. Data shown are from three independent experiments.

ERβ specifically binds to the ERE in the promoter of the BCRP gene

Further research is needed to ascertain whether the ERE in the BCRP promoter can actually be recognized by exogenous ERβ. An electrophoretic mobility shift assay (EMSA) was performed with an rabbit anti-ERβ and a 3′-biotin labeled oligonucleotide probe harboring the putative ERE of the BCRP promoter (Fig. 4A,B). Probes used for EMSAs are shown in Table 2. The specific binding complex was observed in 453/ERβ cells (Fig. 4A) but not in 453 cells (Fig. 4B). Addition of 10- and 100-fold molar excess of unlabeled ERE probe competed partly and completely respectively with the labeled probe–protein complexes (Fig. 4A, lanes 3 and 4), while no competition for binding was observed with 50-fold excess of unlabeled mutant ERE (Fig. 4A, lane 5). These results indicate that transfected ERβ bound specifically with the ERE probe. A supershift assay was performed to detect specific binding of ERβ protein with the ERE probe. The presence of ERβ within the complex was verified by a specific ERβ antibody, which resulted in the appearance and a reduction in the intensity of the DNA–protein complex (Fig. 4A, lane 6). Additionally, ERβ expression was detected in 453/ERβ cells but not in 453 cells (Fig. 4C). Together, these preliminary data suggest that binding only occurs with specific recognition of the consensus ERE on the human BCRP gene by ERβ protein, in support of our hypothesis that ERβ protein is directly involved in formation of the complex.

Table 2. Probes used for EMSAs. Mutated sites are shown as underlined letters
Figure 4.

EMSA analysis of specific binding between transfected ERβ and ERE of the BCRP promoter. (A) A gel shift assay was conducted using nuclear extracts from MDA-MB-453/ERβ cells. Lane 1, probe control alone; lane 2, probe with nuclear extract; lanes 3 and 4, competition of biotin-labeled ERE probe with 10- and 100-fold excess of unlabeled ERE probes, respectively; lane 5, competition of biotin-labeled ERE probe with 50-fold excess of unlabeled mutated ERE (EREm) probe; lane 6, probe with nuclear extract in the presence of anti-ERβ antibody. The upper arrow indicates a super shift band caused by this antibody, whereas the middle arrow shows the ERβ–ERE binding complex. (B) Nuclear extracts from ERβ-negative MDA-MB-453 cells. Lane 1, probe with nuclear extract; lane 2, competition of biotin-labeled ERE probe with 10-fold excess of unlabeled ERE probe; lane 3, competition of biotin-labeled ERE probe with 50-fold excess of unlabeled mutated ERE (EREm) probe; lane 4, probe with nuclear extract in the presence of anti-ERβ antibody. (C) Western blot was performed to detect the expression of ERβ in MDA-MB-453 and MDA-MB-453/ERβ breast cancer cells.

Effects of E2 on BCRP-mediated MX efflux activity

To further determine whether the function of BCRP is affected by treatment with E2 alone or E2 combined with TAM, the intracellular accumulation and retention of MX were evaluated by flow cytometry analysis following these treatments. MX is a fluorescent compound and a known high-affinity substrate of BCRP, which can be used to detect transport activity of BCRP. In 453/ERβ/P-BCRP and 468/SiN cells with E2 treatment, a fluorescence peak shift to the left indicates a significant decrease in fluorescence intensity (Fig. 5A,B, dotted line) whereas this effect is reversed by additional use of TAM (long dashed line). This finding implies that E2 enhances the MX efflux, which is antagonized by TAM. As for 453/ERβ/C-BCRP and 468/SiERβ cells, no obvious shift was observed in comparison with untreated cells (Fig. 5C,D). These findings are consistent with the results of RT-PCR and western blot analysis.

Figure 5.

Effect of E2 on the BCRP function of selected cells was determined by MX efflux assay. Cellular content of MX was measured by fluorescence-activated cell sorting. The lines stand for the fluorescence intensity of MX in selected cells for 2-h efflux. (A), (B) E2 treatment led to a fluorescence peak shift to the left compared with the corresponding untreated control cells, and TAM reversed the effect of E2 which was indicated by the peak shift to the right in 453/ERβ/P-BCRP and 468/SiN cells. (C), (D) No obvious shift was detected compared with untreated controls in 453/ERβ/C-BCRP and 468/SiERβ cells. Bold line, no treatment; long dashed line, combined treatment with E2 and TAM; dotted line, E2 treatment alone. (E) Intracellular fluorescence intensity analysis was performed, and representative data from three independent experiments are shown. **< 0.01 versus the corresponding untreated control cells. #< 0.05 versus the corresponding cells with 10−9 m E2 treatment.

Effects of E2 on BCRP-mediated drug resistance

The BCRP-mediated drug resistance was further confirmed by cytotoxicity assays. Evaluation of sensitivity to anticancer drugs was accomplished by MTT assay. The dose–response curves representing the relationship between drug concentration and cell viability indicated that E2 increased MX resistance of 453/ERβ/P-BCRP and 468/SiN cells, and the restoration of sensitivity to MX occurred when these cells were concomitantly treated with TAM (Fig. 6A,B). As shown in Table 3, E2 treatment resulted in elevated 50% inhibitory concentration (IC50) values of MX in 453/ERβ/P-BCRP, 468/SiN and MDA-MB-468 cells. This enhancement of MX resistance by E2 was markedly reduced by the addition of TAM. As expected, no significant change was observed in MDA-MB-453, 453/ERβ/C-BCRP and 468/SiERβ cells (Table 3). These findings are in agreement with the change of BCRP level and further support our hypothesis that upregulation of BCRP expression by E2 induces resistance of anticancer agents in ERβ-only breast cancer cells.

Table 3. IC50 values of MX in main cell groups
 UntreatmentE2E2 + TAM
  1. IC50 is the drug concentration (μm) resulting in 50% inhibition of cell growth. The relative drug resistance (RR) value of untreated cells was set to 1. Results represent mean values ± SEM of three experiments done in triplicate. *P < 0.05 and **P < 0.01 versus corresponding untreated control cells; #P < 0.05 and ##P < 0.01 versus cells with E2 treatment alone.

MDA-MB-4530.6190 ± 0.072510.6005 ± 0.06680.970.6094 ± 0.09770.98
453/ERp/P-BCRP1.648 ± 0.0697110.92 ± 1.1886.63*1.816 ± 0.07581.10#
453/ERp/C-BCRP4.141 ± 0.485213.282 ± 1.1660.794.142 ± 0.63981.00
MDA-MB-4683.308 ± 0.2473123.34 ± 1.2807.06**4.350 ± 0.35811.31#
468/SiN3.214 ± 0.2482122.22 ± 1.2656.93**2.524 ± 0.39740.79##
468/SiERp3.893 ± 0.471914.193 ± 0.55511.082.829 ± 0.39770.73
Figure 6.

The dose–response curves of the relationship between MX concentration and cell survival rate. Cytotoxicity was assessed by using the MTT assay. The cells were treated with five concentrations of MX. Cell survival rate was calculated relative to untreated controls. (A), (B) E2 treatment resulted in a decrease of MX cytotoxicity in 453/ERβ/P-BCRP as well as 468/SiN cells, and the decrease was reversed by TAM. (C), (D) No significant differences were observed under the same experimental treatments in 453/ERβ/C-BCRP and 468/SiERβ cells. *< 0.05 versus the corresponding untreated control cells. #< 0.05 versus the corresponding cells with 10−9 m E2 treatment.

E2-induced proliferation in ERβ+ cells, which was reversed by TAM

The influence of E2 and TAM on cell proliferation was evaluated by western blot and the 5-ethynyl-2′-deoxyuridine (EdU) proliferation assay. Proliferating cell nuclear antigen (PCNA), a protein synthesized in the early G1 and S phases of the cell cycle, is a common marker for cell proliferation. Western blot showed that PCNA expression was elevated when treated with E2 in 453/ERβ and 468 cells, and the elevation was significantly impaired by TAM (Fig. 7A,B). Moreover, the result of the EdU proliferation assay revealed that the ratio of proliferative cells increased significantly with E2 treatment, and the additional use of TAM antagonized this effect of E2 (Fig. 7C,D). Additionally, no remarkable change of proliferation was observed in 453 and 468/SiERβ cells (data not shown). These findings further confirmed the immunohistochemisty results concerning the relationship between ERβ and tumor size (Table 1).

Figure 7.

E2-induced cell proliferation and reversal of this effect by TAM. (A), (B) PCNA was detected by western blot in MDA-MB-453/ERβ and MDA-MB-468 cells, respectively. E2 treatment enhanced the expression of PCNA, and the addition of TAM antagonized this effect of E2. (C), (D) The cellular proliferation was detected by EdU proliferation assay. E2 treatment elevated the ratio of proliferative cells in MDA-MB-453/ERβ and MDA-MB-468 cells, and TAM reversed E2-induced cell proliferation. Blue dots, all cells; red dots, proliferative cells. *< 0.05 versus the corresponding untreated cells. #< 0.05 and ##< 0.01 versus the corresponding cells with 10−9 m E2 treatment.


The major obstacle in the treatment of breast cancer is the rapid development of MDR in patients. MDR is usually related to the elevated expression of ABC transporters, which can pump substrate drugs out of cells, reducing intracellular drug concentration. BCRP, as a primary type of ABC transporter, is involved in MDR to several kinds of anticancer drugs, in which some are not the specific substrates of BCRP [21, 22]. Although the BCRP protein is synthesized in the cytoplasm, it must be transported to the cell membrane for its pump function [23]. Therefore, the location of BCRP may impact its function. In this experiment, only the brown staining on cell membrane was judged as BCRP positive.

Prior studies have found that hormone nuclear receptors were involved in the regulation of BCRP. It was reported that estrogen downregulated BCRP expression by novel post-transcriptional mechanisms through binding to ERα [24]. Wang et al. [25] also found that E2 downregulated BCRP expression through an ER, but possibly ERβ in human placental BeWo cells. After the genomic structure and characterization of the BCRP promoter were demonstrated, a novel estrogen/progesterone response element (ERE/PRE) was also revealed in the promoter region of BCRP [26-28]. It has also been reported that E2–ERα complex upregulated, while progesterone–PR complex downregulated, BCRP expression through binding to ERE/PRE in breast cancer cells [19, 20, 29]. Nevertheless, the correlation of ERβ and BCRP has not previously been reported in breast cancer. Only two studies determined the role of E2 in downregulation of BCRP through binding to ERβ via a non-genomic pathway at the blood–brain barrier [30, 31].

In the present study, we chose ERα−/PR− breast cancer samples and cell lines to investigate the effect of ERβ on regulating BCRP. The immunohistochemistry results show that membranous BCRP expression accounts for nearly 50% in ERα−/PR− breast invasive ductal carcinoma samples, and positively correlates with the nuclear expression of ERβ. We also observed that ERβ was positively associated with Ki67 and tumor size. It has previously been demonstrated that ERα induces proliferation of breast cancer cells in the presence of E2 [32]. Published data from several studies have demonstrated that ERβ has an anti-proliferative function when re-introduced into ERα+ breast cancer cells [33, 34]. However, it has also been reported that in ERα− breast cancers ERβ correlated with the proliferation marker Ki67, and the highest content of proliferative cells was seen in ERα−/ERβ+ cancers, which are consistent with our findings [9-12]. Taken together, these findings implicate ERβ as a marker for chemotherapy resistance and cell proliferation of breast cancer.

To further elaborate the molecular mechanisms by which ERβ regulates BCRP expression and cell proliferation, we constructed different cell models. ERβ and BCRP plasmids were transfected into MDA-MB-453 (ERα−/ERβ−/PR−) cells. By treating the transfected cells with various ranges of E2, a dose-dependent upregulation of BCRP level was observed in 453/ERβ/P-BCRP cells but not in 453/ERβ/C-BCRP, 453/P-BCRP and MDA-MB-453 cells. These findings indicate that the regulation of BCRP depends on both ERβ and ERE. Furthermore, we verify that endogenous ERβ has the same effect on BCRP expression in MDA-MB-468 cells (ERα−/ERβ+/PR−/BCRP+). However, no remarkable change of the basal BCRP level was observed after knockdown of ERβ in MDA-MB-468 cells, which indicates that, while the presence of the receptor alone might not affect BCRP expression, the E2–ERβ complex can. Additionally, the role of E2 in upregulating BCRP was suppressed by combined treatment of TAM. Apart from this, we also found that both exogenous and endogenous ERβ induced proliferation of ERα−/PR− breast cancer cells (453/ERβ and 468) in the presence of E2, which provides confirmation of the immunohistochemistry result about the relationship between ERβ and Ki67 as well as tumor size. Moreover, additional use of TAM reversed E2-induced cellular proliferation. Collectively, these findings demonstrated that anti-estrogen therapy might also be valuable for ERα−/PR− but ERβ+ breast cancer, at least for pre-menopausal women with a high level of estrogen. Nowadays, endocrine therapy is not routinely used for ERα−/PR− breast cancer patients. However, our results suggest that detection of ERβ might be meaningful for endocrine therapy of these patients.

In one instance ERβ was observed in over 40% of the ERα− samples analyzed [35], and ERβ variants such as ERβcx have been noted in up to 60% of ERα− samples [11]. Moreover, Skliris et al. [36] have recently reported that approximately 18% of breast cancers express only ERβ, which is much lower than our detection (over 50%). This difference of ERβ expression may be caused by ethnic variability within studies. Due to the large number of ERβ-only cases in breast cancer, identification of ERβ effects and how to target ERβ for treatment of ERα−/PR− breast cancer is very important. Several studies of ERα− breast cancer demonstrated a positive correlation between high ERβ expression and poor prognostic phenotypes [9-11]. Our study provides additional evidence that ERβ is positively associated with membranous BCRP expression, indicating that ERβ might be a potential marker of BCRP-mediated MDR. In addition, some reports showed that ERβ expression is good for TAM responsiveness in ERα-negative breast cancer [37, 38], which is consistent with the current observation that TAM antagonized the harmful effect of E2 on drug resistance and proliferation in cell models.

The present study also provides evidence for the first time that expression of ERβ protein is positively correlated with BCRP in ERα−/PR− breast cancer. The cell line experiments reveal that transcriptional activation is involved in the regulation of BCRP expression by E2. This process was modulated by E2–ERβ complex binding to the ERE in the BCRP promoter via the classical pathway to stimulate transcription of the human BCRP gene, resulting in enhanced BCRP-mediated drug resistance in ERα−/PR− breast cancer cells.

It is well known that estrogen stimulates the development of breast cancer through ERα, and ERβ co-expressed with ERα is suspected to behave as a tumor suppressor [39-41]. Our study provides evidence that ERβ alone may act as a tumor promoter in MDR and cell proliferation, and may at least partially explain poor treatment outcomes of ERα−/PR− breast cancer. There may be potential clinical benefit in the reversal of BCRP-mediated MDR and in making an attempt to suppress tumor cell proliferation of ERα−/PR− breast cancer through the targeting of ERβ.

Materials and methods

Patients, tissue samples and immunohistochemistry

This study was conducted on breast tissue samples of 99 patients at the Department of Pathology of Qilu Hospital of Shandong University from 2007 to 2009. These samples were histopathologically diagnosed as ERα−/PR− invasive ductal carcinoma, in which 45 samples were Her-2+ and 54 were Her-2−. Paraffin-embedded 4-μm thick breast cancer sections were dewaxed and subjected to antigen retrieval by microwave irradiation for 20 min in 0.01 m citrate buffer (pH 6.0). Endogenous peroxidase was blocked with 0.3% H2O2 in methanol for 20 min. Then slides were incubated in primary antibodies to BCRP (clone BXP21, dilution 1 : 40, Abcam, Cambridge, MA, USA), ERβ (dilution 1 : 100, Millipore, Billerica, MA, USA), ERα (ZSGB-Bio, Beijing, China), PR (ZSGB-Bio), Her-2 (ZSGB-Bio) and Ki67 (ZSGB-Bio) overnight at 4 °C. After being washed with NaCl/Pi, the primary antibody was detected with the appropriate secondary antibody for 25 min at 37 °C. Following washes, slides were incubated in a two-step plus Poly-HRP Anti-Mouse/Rabbit IgG Detection System (PV-9000, ZSGB-Bio) according to the manufacturer's recommendations, visualized using DAB (PV-9000, ZSGB-Bio), and then rinsed in distilled water and counter-stained with hematoxylin. A dual semi-quantitative scale combining staining intensity and percentage of positive cells was used to evaluate the protein staining. Briefly, staining of BCRP was scored semi-quantitatively for intensity (0, no expression; 1, weak; 2, moderate; 3, strong) and for the percentage of positive cells (0, < 10%; 1, 10–40%; 2, 40–70%; 3, (70%) [29]. The staining of ERβ was scored semi-quantitatively for intensity (0, no expression; 1, weak; 2, moderate; 3, strong) and for the percentage of positive cells (1, 0–1/100; 2, 1/100–1/10; 3, 1/10–1/3; 4, 1/3–2/3; 5, 2/3–1), according to a previously published scoring standard of ERα [42]. The score of immunohistochemistry (IHC) ≥ 2 was defined as positive and < 2 as negative. The staining of Ki67 was scored for the percentage of positive cells (0, 0–5%; 1, 6–25%; 2, 26–50%; 3, > 50%). The optimal cutoff value was identified as (1 for low Ki67 expression level and > 1 for high expression. Scoring of Her-2 was done on a 0–3 scale [43]. Positive (3+) was defined as strong complete membranous staining in more than 30% of the tumor cell population; borderline (2+) was defined as moderate membranous staining in more than 10% of tumor cells; 1+ was defined as either weak or barely perceptible membranous staining in more than 10% of the tumor cells. Furthermore, fluorescence in situ hybridization (FISH) analysis was performed for Her-2 neu gene amplification in all 2+ cases by IHC. Scores of 0, 1+ and 2+ by IHC but negative by FISH were considered as negative for Her-2/neu expression, 3+ and positive by FISH of 2+ cases as positive for Her-2/neu expression. Normal breast epithelium and capillary were used as positive controls. All the slides were examined and scored independently by two pathologists without any knowledge of the patient's clinical data. The procedures were in accordance with the ethical standards of the responsible committee on human experimentation.

Cell lines and cell culture

The human breast cancer cell lines including MDA-MB-453 (ERα−/PR−/Her-2+) and MDA-MB-468 (ERα−/PR−/Her-2−) were purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured in phenol red-free L-15 supplemented with 10% heat-inactivated fetal bovine serum (Gibco, Carlsbad, CA, USA) at 37 °C in a humidified 5% CO2 atmosphere. All cells were free from mycoplasma infection.

Plasmid construction

The pcDNA3.1(−)-C-BCRP and pcDNA3.1(−)-P-BCRP plasmids were constructed as previously reported [20]. The only difference between the two plasmids is the promoter region. The expression of BCRP in the former plasmid is driven by a CMV promoter, and the endogenous BCRP promoter including ERE drives BCRP expression in the latter one.

The pEGFP-C1-CMV-ERβ vector was kindly provided by RongLin Zhai (Tongji Medical College, Wuhan).

SiERβ sequences were synthesized as follows: forward, 5′-GATCCCCAAGCCCAAATGTGTTGTGGCCTTCAAGAGAGGCCACAACACATTTGGGCTTTTTTTGGAAA-3′, and reverse, 5′-AGCTTTTCCAAAAAAAGCCCAAATGTGTTGTGGCCTCTCTTGAAGGCCACAACACATTTGGGCTvTGGG-3′. SiERβ sequences were cloned into the pSUPER.neo+GFP expression vector. Non-targeting (Notarget) small interfering RNAs were designed as negative controls using the following sequences: forward, 5′-GATCCCCTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAATTTTTGGAAA-3′, and reverse, 5′-AGCTTTTCCAAAAATTCTCCGAACGTGTCACGTTCTCTTGAAACGTGACACGTTCGGAGAAGGG-3′. pSUPER-siN including non-targeting sequences was constructed as the control vector.

Transfection of breast cancer cells

Use TurboFect™(Fermentas, Burlington, Canada) to stably transfect breast cancer cells. Subsequently, the stable transformants were selected by growth in medium with 300 μg·mL−1 geneticin (G418; Invitrogen, Carlsbad, CA, USA). The main cell groups were as follows: MDA-MB-453/ERβ, MDA-MB-453/ERβ/C-BCRP, MDA-MB-453/ERβ/P-BCRP, MDA-MB-468/SiN and MDA-MB-468/SiERβ. PR, ERα, ERβ and BCRP levels were also detected in these cells by western blot (Fig. 2A,B).

RT-PCR analysis

The RNA Iso-Plus kit (TaKaRa, Kyoto, Japan) was used to extract total cellular RNA, and the first strand cDNA was synthesized by Omniscript Reverse Transcriptase (Toyobo, ShangHai, China). Products were amplified using the following primers: BCRP, sense, 5′-TGGCTGTCATGGCTTCAGTA-3′, antisense, 5′-GCCACGTGATTCTTCCACAA-3′, 235-bp product; β-actin internal control, sense, 5′-CTCCATCCTGGCCTCGCTGT-3′, antisense, 5′-GCTGTCACCTTCACCGTTCC-3′, 268-bp product. RT-PCR results were analyzed by a gel figure analysis system (Gel Doc 2000, Bio-Rad, Hercules, CA, USA).

Preparation of cell lysates

Total proteins were extracted from cells using RIPA lysis buffer (Beyotime, Beijing, China). Nuclear and cytoplasmic total proteins were prepared using a nuclear and cytoplasmic protein extraction kit (Beyotime).

Western blot analysis

Cell lysates from each group were separated by electrophoresis on 10% SDS/PAGE. These proteins were blotted onto poly(vinylidene difluoride) membranes (Millipore) and blocked with 5% fat-free milk powder in NaCl/Tris at room temperature for 1 h. After incubation overnight at 4 °C with the relevant primary antibodies, the blots were washed and probed again with species-specific secondary antibodies coupled with horseradish peroxidase. The protein–antibody complexes were visualized by chemiluminescence with the ECL Advance Western Blotting Detection kit (Millipore).

The primary antibodies detecting proteins were mouse anti-BCRP (BXP21) (diluted 1 : 200, Abcam), rabbit anti-ERβ (diluted 1 : 200, Millipore), rabbit anti-ERα (MC-20) (diluted 1 : 400, ZSGB-Bio), rabbit anti-PR (diluted 1 : 500, ProMab, Richmond, Canada), PCNA (I88) pAb (diluted 1 : 1000, Bioworld Technology, Shanghai, China) and mouse monoclonal antibody to β-actin (TA-09) (diluted 1 : 2000, ZSGB-Bio) as a total protein internal control.

Electrophoretic mobility shift assay (EMSA)

Nuclear proteins were extracted from both ERβ-transfected and non-transfected MDA-MB-453 breast cancer cells. EMSA probes synthesized by Sangon Biotech Co. Ltd (Shanghai, China) were as follows: ERE probes with and without a 3′-biotin label, and the ERE mutated probe without a label. Sequences of the normal and mutated ERE probes are shown in Table 2. Total nuclear proteins (2 μg) were incubated in an EMSA/gel-shift binding buffer (Beyotime) and 0.2 nmol of labeled DNA probe at room temperature for about 25 min. In competition assays, 10- and 50-fold excess of unlabeled ERE probe or 50-fold excess of unlabeled mutated ERE probe was added into the mixture mentioned above and incubated in the same conditions. In the supershift assay, 2 μL of undiluted anti-ERβ antibody was added into the same mixture, but the incubation time was 15 min longer than in the other assays. The complexes were electrophoretically separated by a 5% non-denaturing polyacrylamide gel and then transferred onto a positively charged nylon membrane (Hybond-N+) in 0.5 × Tris/borate/EDTA buffer gel. Transferred complexes were crosslinked to the membrane for 15 min, detected with horseradish-peroxidase-conjugated streptavidin and visualized using an enhanced chemiluminescence kit (Millipore).

Intracellular MX efflux assay

All cells were cultured in the presence or absence of 10−9 m E2 (Invitrogen) or concomitant treatment with 10−6 m TAM. After treatment for 72 h, the suspension cells were incubated either in complete medium alone for the blank histogram or in complete medium containing 25 μm MX, a known high-affinity substrate of BCRP, in order to detect BCRP transport activity. The efflux histogram was generated from cells which were incubated with 25 μm MX for 30 min and then allowed to efflux for 2 h in complete medium alone. The cells were washed twice with ice-cold NaCl/Pi and then resuspended in 500 μL of serum-free and phenol red-free medium to be analyzed by flow cytometry (BD AccuriC6, San Jose, CA, USA) with a 635-nm red diode laser and 670-nm band-pass filter.

Cytotoxicity assay

The MTT (Sigma, San Francisco, CA, USA) assay was used to assess the chemosensitivity of selected transfection cells to MX. Cells were seeded into a 96-well plate at a density of 5 × 103 cells per well for 24 h and then incubated in the presence or absence of 10−9 m E2 or concomitant treatment with 10−6 m TAM for 72 h. Subsequently, different concentrations (0, 0.4, 2, 10, 50 μm) of MX were added for another 48 h before the MTT assay was performed following the manufacturer's instructions. Optical densities were measured using the spectrometric absorbance at 570 nm against a background of 630 nm on a Bio-Rad microplate reader. IC50 values and relative drug resistance values were calculated to determine the chemosensitivity of the cells.

EdU proliferation assay

The cells were seeded at the proper density in 96-well plates with different treatments for 72 h. The Cell-Light™ EdU DNA Cell Proliferation Kit (Ribobio, Guangdong, China) was used to stain and detect proliferation of the cells according to the manufacturer's instructions. The cells were analyzed with a fluorescence microscope (Leica, Weitsbaden, Germany) using the Cy5 channel for detecting proliferative cells and Hoechst33342 for all cells.

Statistical analysis

Statistical analysis was performed using spss 17.0 and graphpad prism 5 (GraphPad Software, San Diego, CA, USA). All data are presented as the mean ± SEM, and Student's t test was used to determine statistical significance. Comparisons between two groups were performed using the paired t test. < 0.05 was considered statistically significant.


We thank Dr Ronglin Zhai (Tongji Medical College, Wuhan, China) for kindly providing us with the pEGFP-C1-ERβ plasmid and Dr Jason M. Cuellar (NYU Hospital for Joint Diseases, New York, USA) for his critical reading of this paper. This work was supported by Graduates' Independent Innovation Foundation of Shandong University to Weiwei Li (yzc12159).