ERβ1 inhibits metastasis of androgen receptor-positive triple-negative breast cancer by suppressing ZEB1

Eighty two TNBC tissue samples were collected from patients who underwent tumor resection in The First Affiliated Hospital of Wenzhou Medical University from April 2005 to March 2014. The use of clinical tissues for this study was approved by the Jinling Hospital’s Ethics Committees and conducted in accordance with the Helsinki Declaration. All patients gave their informed consent prior to inclusion in the study.

Human TNBC cell lines, MDA-MB-231 cells and Hs578T cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in RPMI-1640 or Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum (FBS) at 37 °C in 5% CO₂. For the ligand experiments, cells were maintained in phenol red-free media containing 5% dextran-coated charcoal (DCC)-treated FBS. BALB/c athymic nude mice (female, 6 weeks old) were purchased from the Department of Comparative Medicine, Jinling Hospital (Nanjing, China) and maintained in a pathogen-free facility. 5α-Dihydrotestosterone (DHT) was purchased from Sigma-Aldrich (St. Louis, USA).

Total RNA was extracted from cultured cells using TRIzol (Invitrogen, USA). For qRT-PCR analysis, cDNA was synthesized using the PrimeScript^TM RT Master Mix (Perfect Real Time) Kit (RR036A, Takara, China), and PCR was performed using the Power SYBR Green PCR Master Mix (Life Technology, USA). The primer sequences were as follows: ERβ1 forward 5’- CGATGCTTTGGTTTGGGTGAT-3’ and reverse 5’-GCCCTCTTTGCTTTTACTGTC-3’; AR forward 5’-CCTGGCTTCCGCAACTTACAC-3’ and reverse 5’-GGACTTGTGCATGCGGTACTCA-3’; E-cadherin forward 5’-TGAAGGTGACAGAGCCTCTGGAT-3’ and reverse 5’-TGGGTGAATTCGGGCTTGTT-3’; N-cadherin forward 5’-CACTGCTCAGGACCCAGAT-3’ and reverse 5’-TAAGCCGAGTGATGGTCC-3’; ZEB1 forward 5’-GCCAATAAGCAAACGATTCTG-3’ and reverse 5’-TTTGGCTGGATCACTTTCAAG-3’; Snail forward 5’-CACTATGCCGCGCTCTTTC-3’ and reverse 5’-GGTCGTAGGGCTGCTGGAA-3’; Twist forward 5’-AGTCCGCAGTCTTACGAGGA-3’ and reverse 5’-GCCAGCTTGAGGGTCTGAAT-3’; and GAPDH forward 5’-AAATCAAGTGGGGCGATGCTG-3’ and reverse 5’-GCAGAGATGATGACCCTTTTG-3’. The 2^−ΔΔCt method was used to determine the relative mRNA expression.

The human full-length cDNA of ERβ1 and ZEB1 were cloned into the pFLAG-CMV expression vector (Sigma) and verified by sequence analysis before transfection, respectively. The MDA-MB-231 and Hs578T cells were transfected with an empty vector or pFLAG-CMV-ERβ1 by Lipofectamine 2000 (Invitrogen, USA) following selection by 1 mg/ml G418 (Gibco). AR shRNA (shAR) and scrambled shRNA control (shNC) were cloned into the pGpU6/GFP/Neo vector purchased from GenePharma (Shanghai, China) and transfected into cells according to the manufacturer’s protocol. The shRNA sequences were as follows: shAR, 5’-CACCAATGTCAACTCCAGGAT-3’, and shNC, 5’-AGTGCACGTGCATGTCCTA-3’.

Cells were seeded in 6-well plates and incubated to generate confluent cultures. Wounds were scratched in the cell monolayer using a 200 μl sterile pipette tip, and cells were rinsed with PBS. The migration of the cells at the edge of the scratch was photographed at 0 and 24 h. The invasion ability of cells was determined using 24-well transwell chambers (Costar, USA) coated with matrigel (BD Biosciences, San Jose, CA). After transfection, approximately 1 × 10⁵ cells/200 μl were resuspended in medium without serum and plated in the top chamber of each transwell, and 800 μl of medium supplemented with 10% FBS was injected into the lower chamber. After 24 h incubation, the inserts were fixed with 100% methanol, subsequently stained with crystal violet and photographed under a microscope.

Cells lysates were resolved by SDS-PAGE electrophoresis (30 μg/sample) and electro-transferred onto polyvinylidene fluoride (PVDF) membranes. After incubation in blocking buffer, the membranes were probed overnight at 4 °C with the following primary antibodies: GAPDH (CST), 1:5000; ERβ1(sc-6822) (Santa Cruz), 1:200; AR (D6F11) (CST), 1:2000; E-cadherin(24E10) (CST),1:1000; N-cadherin(D4R1H) (CST), 1:1000; ZEB1(D80D3) (CST), 1:1000; Snail (C15D3) (CST), 1:1000; and Twist (ab50581) (Abcam), 1:500. The subsequent steps were performed as previously described [20].

Cells were fixed with 3% paraformaldehyde for 10 min, permeabilized with 0.1% SDS solution in PBS for 10 min and then blocked for 20 min. Fixed cells were stained with primary antibodies and then with a secondary antibody coupled to Dylight 649 for 30 min. Cell nuclei were stained with Dapi-Fluoromount-G for 15 min. The expression was defined as follows: −, no immunofluorescence; ±, weak immunofluorescence; +, moderate immunofluorescence; ++, strong immunofluorescence; and +++, very strong immunofluorescence. The samples with scores ++ or +++ were defined as high expression, and the remaining samples were defined as low expression.

Following deparaffinization, sections were rehydrated and subjected to antigen retrieval using citrate buffer (BioGenex, USA). The slides were incubated with primary antibodies, including ERβ1(ab27720) (Abcam) and other antibodies as described in the previous section, at 4 °C overnight. The following steps were performed as previously described [20]. The cut-offs for positivity at 1% and 20% were defined as AR-positive and ERβ1-positive, respectively [9, 21]. The percentages of positive cells and staining intensities were scored as previously described [20]. IHC scoring was performed without prior knowledge of the clinical response.

The cells were treated with 1% formaldehyde for 8 min to crosslink histones to DNA. After washing with cold PBS, the cell pellets were resuspended in lysis buffer and sonicated for 8 s 7 times. The lysate was divided into three fractions, which were incubated with IgG antibody as a negative control, RNA polymerase II antibody as a positive control, or AR antibody (ab74272) (Abcam), at 4 °C overnight. To collect the immunoprecipitated complexes, protein A-Sepharose beads (Pharmacia Biotech) were added and incubated for 1 h at 4 °C. After washing, the beads were treated with RNase (50 μg/ml) for 30 min at 37 °C and then proteinase K overnight. The crosslinks were reversed by heating the sample at 65 °C for 1.5 h. DNA was extracted by the phenol/chloroform method, ethanol-precipitated, and resuspended, and it was then used for PCR, the primer sequences were shown in Table 1.

All animal experiments were conducted in accordance with the Guide for Care and Use of Laboratory Animal, and all experimental protocols were approved by the Animal Ethics Committee. The control and ERβ1-expressing MDA-MB-231 and Hs578T cells (1× 10⁷cells/ml, 100ul/mouse) were injected into 6-week-old BALB/c female nude mice via the tail vein to establish a tumor lung metastasis model. The lungs of mice were removed on the eighth week after injection.

Previous studies had shown an association of ERβ1 expression with tumor metastasis and the suppression of factors known to be involved in bone metastasis in prostate cancer [22]. We hypothesized that ERβ1 may regulate metastasis in TNBC. To explore the functions of ERβ1 on tumor metastasis in AR-positive TNBC, we transfected two AR-positive TNBC cell lines, MDA-MB-231 and Hs578T, with the empty vector or the ERβ1 expression vector. Eight weeks later, stable ERβ1-expressing cell lines were obtained and confirmed by Western blot analysis and qRT-PCR (Fig. 1a and b). Wound-healing and transwell assays were then performed to evaluate the migration and invasion of these cells. The results showed that ERβ1 repressed the migration and invasion of two cell lines, MDA-MB-231-ERβ1 and Hs578T-ERβ1, when compared to the control (P < 0.05) (Fig. 1c and d). Our results suggested that ERβ1 inhibits the migration and invasion abilities of AR-positive TNBC cells.

We next investigated the underlying mechanisms by which ERβ1 suppresses the metastasis of AR-positive TNBC. As shown in Fig. 2a, compared with the control cells, ERβ1 upregulated E-cadherin epithelial marker expression (1.75-fold and 1.60-fold increase) and downregulated N-cadherin mesenchymal marker expression (56% and 32% decrease) in MDA-MB-231-ERβ1 and Hs578T-ERβ1 cells, respectively. The results of qRT-PCR revealed that ERβ1 was also positively associated with E-cadherin mRNA and inversely associated with N-cadherin mRNA (Fig. 2b). Moreover, the results of immunofluorescence showed higher expression of E-cadherin and lower expression of N-cadherin on the cell surface of ERβ1-expressing cells compared to control cells (Fig. 2c), suggesting that ERβ1 promoted cell-cell adhesion by regulating the expression of E-cadherin and N-cadherin. Several transcription factors, such as Snail, Twist and ZEB1, had been reported to promote EMT in multiple tumors, including breast cancer [23, 24]. For example, BCL6 induced EMT by promoting the ZEB1-mediated transcription repression of E-cadherin in breast cancer cells. Therefore, we examined if the inhibition of these factors was involved in the regulation of ERβ1-induced E-cadherin and N-cadherin expression. Only ZEB1 expression was reduced in both ERβ1-expressing cell lines compared to control cells (Fig. 2d). In addition, qRT-PCR revealed that ERβ1 inhibited ZEB1 expression at the transcriptional level (Fig. 2e). ZEB1 overexpression abrogated the increase of E-cadherin expression and the decrease of N-cadherin expression caused by ERβ1 overexpression (Fig. 2f). Taken together, these data showed that ERβ1 regulates E-cadherin and N-cadherin expression through inhibiting ZEB1.

To further investigate the anti-metastatic effect of ERβ1 in vivo, we established a lung metastasis mouse model using control TNBC cells and ERβ1-expressing TNBC cells. Lung metastasis was determined by comprehensive analysis of macroscopical observation and HE staining of lung tissue sections. The results showed that there was higher occurrence of lung metastasis (4/8 in MDA-MB-231 and 3/8 in Hs578T) in the control cells and a lower occurrence of lung metastasis (1/8 in MDA-MB-231-ERβ1 and 1/8 in Hs578T-ERβ1) in ERβ1-expressing cells (Fig. 3a and b). Fig. 3c showed the representative HE staining of lung tissues from mice treated with control or ERβ1-expressing MDA-MB-231 and Hs578T cells. Our results showed that ERβ1 inhibits metastasis of AR-positive TNBC cells in vivo.

AR is expressed in both MDA-MB-231 and Hs578T TNBC cell lines. It has been reported that AR expression correlates with metastases in TNBC. We further examined if AR is involved in the regulation of ERβ1-mediated metastasis inhibition in TNBC cells. Wound-healing and transwell assays showed that AR activation by DHT inhibited cell migration and invasion ability in both MDA-MB-231-ERβ1 and Hs578T-ERβ1 cells compared to control cells (Fig. 4a and b). Moreover, ZEB1 protein was downregulated after AR activation, which was accompanied by upregulated E-cadherin and downregulated N-cadherin, in both ERβ1-expressing cell lines (Fig. 4c). As mentioned above, ZEB1 expression responded to the modulation of ERβ1. We measured the expression of ERβ1 after AR activation and found that activated AR increased ERβ1 protein and mRNA expression levels (Fig. 4c and d). These data suggested that AR activation enhances ERβ1 expression, which improves the suppressive effect of ERβ1-mediated metastasis in ERβ1-expressing TNBC cells.

We next investigated the possible mechanisms by which AR regulated ERβ1 expression. First, we detected the expression of ERβ1 after AR knockdown by shAR in the presence of DHT. AR knockdown led to a decrease in the expression of ERβ1 protein and mRNA compared to the negative control (Fig. 5a and b), which indicated that DHT caused the above changes through activation of the AR signaling pathway. It is well known that AR, as a transcription factor, regulates the expression of many genes in many different cellular processes [25, 26]. Because we showed that ERβ1 expression was regulated by activated AR at both protein and mRNA levels, we hypothesized that AR functioned as a transcription factor that directly bound to the promoter region of ERβ1. To better determine if AR binds to the promoter of ERβ1, a ChIP assay was performed using ERβ1-expressing MDA-MB-231 cells exposed or not exposed to DHT. The results showed that AR, as a transcription factor, directly bound to the promoter of ERβ1 and that the recruitment of AR to the promoter of ERβ1 was increased with DHT treatment (Fig. 5c and d). Furthermore, no difference in AR protein and mRNA expression was observed between the control and ERβ1-expressing TNBC cells (Fig. 5e and f). Our results suggested that activated AR promotes the expression of ERβ1 by functioning as a transcription factor that directly binds to the promoter of ERβ1.

We next examined the correlation of ERβ1 with AR, ZEB1, E-cadherin and clinical characteristics in 82 TNBC samples. The protein levels of ERβ1, AR, ZEB1, and E-cadherin were determined by IHC (Fig. 6a and b). Our results showed that ERβ1 and AR were overexpressed in 31.7% (26/82) and 23.2% (19/82) of the TNBC clinical samples, respectively (Table 2). There was a negative correlation between ERβ1 and ZEB1 (r = −0.330, P = 0.003) and a positive correlation between ERβ1 and E-cadherin (r = 0.391, P < 0.001) (Table 2). Moreover, ERβ1 expression was negatively correlated with lymph node metastasis (r = −0.368, P = 0.001), and positively correlated with the expression of AR (r = 0.309, P = 0.005). There was no association of ERβ1 expression with age and tumor size. Besides, the 82 TNBC clinical samples were divided into ERβ1-negative or ERβ1-positive groups based on the level of ERβ1. The ERβ1-positive group had higher percentage of E-cadherin-positive, ZEB1-negative and AR- positive samples (Fig. 6c).