Total 39 human breast cancer tissue samples were obtained from Second Affiliated Hospital of Chongqing Medical University. All patients were underwent surgery at Second Affiliated Hospital of Chongqing Medical University during 2008-2011 and all tissue samples were identified as breast cancer by clinicopathological diagnose and divided into three groups by TNM standard: 3 for stage I, 27 for stage II and 9 for stage III. Patients were informed before all experiments with ethical approval.

All tissue samples were fixed with formalin and embedded in paraffin. Immunohistochemical staining with streptavidin-peroxidase assay was conducted as manufacturer instructions (Minneapolis, MN). Briefly, 4 μm thick tissues sections were deparaffinized, heated at 95˚C sodium citrate for antigen retrieval, and the endogenous peroxidase activity was quenched by 3% H₂O₂. Nonspecific binding sites were blocked by goat serum. Slides were exposed to Rabbit anti-GPER polyclonal antibody (Abcam) and Rabbit anti-CXCR1 polyclonal antibody (Abcam) with 1:250 dilution or PBS as negative control for 2 h at 25˚C. Sections were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (Hyclone) for 20 minutes at 37˚C to detect tissue-associated rabbit antibodies. Nuclei were counterstained with hematoxylin. Reduction mammoplasty tissue was used as positive control.

Evaluation of immunohistochemical staining was processed by semiquantitative method under phase-contrast microscopy system. Briefly, five 400X yields were randomly chosen from every specimen, 100 cells were counted for each yield and the positive staining was identified by yellow or brown staining of nuclei, cytoplasm or membrane. Staining scores were consisted by two parts. The staining proportion scores: 0, negative staining in all cells; 1, ≤25% cells stained; 2, 26% ~ 50% cells stained; 3, 51% ~ 75% cells stained; 4, >75% cells stained. The staining intensity scores: 0, negative staining; 1, light yellow; 2, yellow; 3, brown. Adding two parts of scores as the staining scores: 0, (−); 2 - 3, (±); 4 - 5, (+); 6 - 7, (++). (+) and (++) were considered as positive staining.

Human breast cancer cell lines MDA-MB-231 and MCF-7 were maintained in RPMI1640 (GIBCO) supplemented with 10% fetal bovine serum (GIBCO) and 1% antibiotics (Penicillin-Streptomycin, GIBCO). Cells were cultured at 37˚C under a 5% CO₂ atmosphere and all experiments were conducted with log phase cells.

The expression level of total GPER, CXCR1, ER, Akt (tAkt), phosphorylation Akt (pAkt), total ERK (tERK), phosphorylation ERK (pERK) and NF-kB (p65) were evaluated by western-blotting analysis using Quantity one system. MDA- MB-231 cells were treated with 10⁻⁸ mol/L E2 and proteins were extracted after different time course. Total cell extracts were separated on SDS-PAGE and immunoblotted with rabbit anti-GPER, CXCR1, ER, tAkt, -pAkt, -tERK, -pERK, -p65 polyclonal antibodies (Abcam), followed by anti-rabbit secondary antibodies conjugated with horseradish peroxidase (Hyclone) respectively. For p65, the MDA-MB-231 cells were treated with 10⁻⁸ mol/L E2 alone or combination with 2 μM lycopene and incubated for 4 h before the cells harvested for lysis. The digitized fluorescent bands were integrated by ECL and normalized against the amount of β-actin, which was used as a loading reference.

The expression level of IL-8 in MDA-MB-231 cells was evaluated by ELISA analysis (Minneapolis, MN). MDA-MB-231 cells were treated with 10⁻⁸ mol/L E2 for 1 - 24 h and the supernates were collected for ELISA analysis. The measurements were according to manufacturer instructions. Briefly, 50 µl of Standard, blank control and sample were added for each well, incubated at 37˚C for 0.5 h and wash. 50 µl IL-8 Conjugates were added for all wells and incubated at 37˚C for 0.5 h followed by adding 100 µl Substrate Solution after wash with 10 min incubation at 37˚C. 50 µl stop solution was added to each well and optical density of each well was determined with microplate reader at 450 nm using blank well as reference.

Cell invasion was assayed with 24-well cell invasion Transwell assay plates comprising a polycarbonate membrane with 8-μm pores (BD). The inner and outer chamber membranes were coated with 5 μg/ml of human fibronectin (R&D) at 37 ˚C for 2 h, and then rinsed with PBS. MDA-MB-231 cells were treated with E2 for 8 h before seeding on the inner chamber of the Transwell plate at a concentration of 1 × 10⁶ cells/ml in 200 μl serum-free RPMI1640. The outer chamber was filled with 800 μl full culture medium which contained 10% FBS and incubated for 48 h at 37 ˚C. The number of invaded cells was counted in fields randomly chosen from 5 equally divided zones of the membrane in triplicates under phase-contrast and fluorescence channel with the phase-contrast microscope system.

The SPSS version 17.0 for Windows (SPSS Inc., Chicago , USA ) was used for the statistical analysis. All measurements are expressed in terms of mean ± standard deviation ( x ¯ ± s). Comparisons between 2 samples of unequal variance were performed by the χ² test, the Spearman test and individual sample t test. P-values of 0.05 or less were considered to be statistically significant.

The predominant staining patterns of GPER and CXCR1 were cytoplasmic in carcinoma tissues regardless of the histopathological type (Figure 1). No nuclear staining was observed. GPER was positive in 68.9% of all specimens in which CXCR1 was 71.7%, revealing the extensive expression of GPER and CXCR1 in primary breast carcinomas (Table 1). The association between GPER and CXCR1 was significant (p = 0.04), whereas 96% of the CXCR1+ specimens were positive for GPER, implying the expression of GPER may be consistent with CXCR1. Furthermore, the expression level of GPER and CXCR1 were significantly associated with the developing stage of the breast cancer tissues, the stage I showed relatively low staining of GPER and CXCR1 which significantly increased in stage II. The stage III tissues exhibited much higher GPER and CXCR1 expression than stage I and II which means GPER and CXCR1 expression level may have positive relationship with the malignity of breast cancer (Figure 1).

In order to distinguish the different expression level of GPER, CXCR1 and estrogen nuclear receptors in metastatic and non-metastatic breast cancer cells, MDA-MB-231 and MCF-7 were chosen for western-blotting analysis (Figure 2). Results showed that the expression level of estrogen receptors ERα and ERβ were almost undetectable in MDA-MB-231 cells whereas MCF-7 cells had significant high expression (Figure 2(A)). Compared to the expression level of β-actin, the relative expression amounts in MCF-7 were 4 folds for ERα and 18 folds for ERβ to MDA-MB-231 respectively (Figure 2(B)). The MDA-MB-231 and MCF-7 cells had very similar expression level of GPER and the difference was not significant. But, interestingly, the CXCR1 expression level was much stronger in MDA-MB-231 than MCF-7 cells and the relative expression level of CXCR1 was nearly 3 folds in MDA-MB-231 to MCF-7 cells (Figure 2(B)). As a kind of metastatic breast cancer cell line, MDA-MB-231 cells naturally had stronger migration and invasion activities than MCF-7 cells despite the lower ERα, ERβ and similar GPER expression, the difference in CXCR1 expression level may play a important role to determine the malignance in MDA-MB-231 and MCF-7 cells and this different malignance was independent from the expression level of estrogen receptors ERα and ERβ.

Signaling through GPER by estrogen stimulation occurs via transactivation of the epidermal growth factor receptor (EGFR) [23] [24] . In order to probe such

regulation and dissect the possible signaling pathway, we examined Akt and ERK activation in MDA-MB-231 cells in response to treatment with estrogen. Akt and ERK activation stimulated by 10⁻⁸ mol/L E2 after exposure at different lengths of time were investigated in MDA-MB-231 cells. The ratio of phosphorylation proteins/total proteins was used as levels of activation of Akt and ERK. Result showed that a rapid activation of Akt was observed by this concentration of E2 within 15 mins, estrogen induced a significant increase of Akt phosphorylation and the peak level of p-Akt could be observed at 15 min in MDA-MB-231 cells (Figure 3); The activation of ERK by estrogen was even faster than Akt, the peak level of p-ERK appeared at 5 min.

NF-kB is one of the redox-sensitive transcription factors. Upon stimulation, activated NF-kB will translocate to the nucleus and bind to promoters of its target genes and regulates the expression of genes involved in many cellular events including proliferation and migration through the activation of the Akt and ERK cascades [25] [26] [27] [28] . To elucidate whether the activation of GPER by estrogen affected activation and nuclear translocation of NF-kB through Akt and ERK cascades, NF-kB (p65)-associated proteins in cytoplasma and nucleus were measured in MDA-MB-231 cells upon estrogen stimulation by western-blotting. We analyzed the effects of E2 alone and in combination with lycopene, on total NF-kB expression level in MDA-MB-231 cells. As shown in Figure 4, a 4 h-treatment of the MDA-MB-231 cells with 10⁻⁸ mol/L E2 resulted in an increase

of the NF-kB DNA-binding activity; pre-treated with lycopene with 2 μM concentration for 4 h, the expression of NF-kB was significantly inhibited and this inhibition was turned over by combination with E2, the expression level of NF-kB was significantly recovered and even higher than control.

Furthermore, the ratio between the content of NF-kB in nucleus and cytoplasma was used to present the nuclear translocation of NF-kB in MDA-MB-231 cells upon estrogen stimulation. Results showed that before exposure to 10⁻⁸ mol/L E2, NF-kB in nucleus was nearly 65% of cytoplasma; after activated by 10⁻⁸ mol/L E2, NF-kB gradually translocated to nucleus according to the exposure time and the peak level of NF-kB in nucleus appeared at 4 h with almost 1.5 folds amount of NF-kB compared to the cytoplasma (Figure 5).

Overexpression of IL-8 has been detected in many human tumors, including breast cancer, and is associated with poor prognosis [29] . We investigated the effect of estrogen on IL-8 production via NF-kB transcription in MDA-MB-231 cells by ELISA. Results showed that the pre-incubation of MDA-MB-231 cells with 10⁻⁸ mol/L E2 for different time length had significant effect on IL-8 protein production in culture medium and IL-8 production reached the peak level at 8 h treatment with estrogen (Figure 6), demonstrating the key role of estrogen in IL-8 production by NF-kB modulation.

To determine the role of estrogen in breast cancer cells, the metastatic potential of MDA-MB-231 cells was investigated by using in vitro Matrigel invasion assay. Results showed that the invasive potential of MDA-MB-231 cells was significantly promoted by estrogen stimulation and the stimulated cells exhibited 3 folds increase in invasion above control cells (Figure 7). Thus, it seems like estrogen had significant effect on in vitro invasion activities of MDA-MB-231 cells.