Background

The processes of invasion and metastasis that cause mortality in patients are extraordinarily distinctive features of breast cancer progression [1]. Although lymph-node metastasis, large tumor size, and poorly-differentiated histopathological grade are commonly considered to be established prognostic markers related to metastasis [2], distant metastasis still occurs in 20-30% of the patients with negative lymph-node involvement [3]. Thus far, Human Epidermal Growth Factor Receptor 2 (HER-2/neu) [4], c-myc [5] and HOXB9 [6] have emerged as predictors of the risk of metastasis in breast cancer. The aberrant expression of these factors may induce the expression of growth and angiogenic factors in tumors, leading to increased local concentrations of these factors within the tumor microenvironment and thus favoring tumor progression [6]. Recently, a new genomic test (gene-expression profiling) has been suggested to predict the clinical outcome more accurately than the traditional clinical and pathological standards [7, 8]. However, it is an open question as to whether this method will enter into the clinical routine for staging and grading [9]. Although these new markers and methods have been implicated, the molecular mechanism of breast cancer metastasis remains far from being fully understood due to the heterogeneity of this cancer and represents a new prerequisite for develo** better treatment strategies.

The polycomb (PcG) proteins constitute a global system with important roles in multi-cellular development, stem cell biology and cancer [10]. B-lymphoma Moloney murine leukemia virus insertion region-1 (Bmi-1), a member of the PcG family of transcription repressors, has emerged as a Myc-cooperating oncogene in murine lymphomas [11, 12]. Bmi-1 can not only lead human mammary epithelial cells (HMECs) to bypass senescence and immortalize, but it also can play a key role in human breast cancer [13, 14]. Moreover, a significant correlation has been observed between Bmi-1 expression and axillary lymph node metastasis in invasive ductal breast cancer [15]. These findings suggest that Bmi-1 could be involved in the carcinogenesis and metastasis of breast cancer. Although increasing evidence has shown that Bmi-1 expression is associated with unfavorable prognosis [16, 17], other studies have not confirmed these findings [18, 19]. Bmi-1 protein is detected in only 25% of African breast cancer patients and is associated with a low histological grade [18]. Additionally, higher Bmi-1 mRNA expression has been observed in early-stage patients without lymph node metastasis [20]. In contrast, up-regulation of Bmi-1 was shown to be associated with the invasion of nasopharyngeal carcinomas and to predict poor survival [84]. The blots were probed with mouse anti-Bmi-1, anti-E-cadherin, anti-β-catenin, anti-fibronectin and anti-vimentin antibodies (BD, Transduction Laboratories, Lexington, UK) as well as with rabbit anti-p-GSK, anti-t-GSK (Cell Signaling Technology, Inc. USA), anti-snail (Abcam, Cambridge Science Park, Cambridge, UK), anti-p-Akt (Santa Cruz Biotechnology, CA, USA) and goat anti-t-Akt (Santa Cruz Biotechnology, CA, USA) antibodies. The membranes were stripped and re-probed with mouse anti-α-tubulin (Sigma Aldrich, Inc. St Louis, Missouri, USA) to confirm equal loading of the samples.

Wound Healing Assay

Cells were seeded in six-well plates and cultured under permissive conditions until 90% confluence. After starving the cells for 24 h in medium without EGF or FBS, the confluent cell monolayer was lightly and quickly scratched with a pipette tip to produce a straight line. The debris was removed and the edge of the scratch was smoothed with PBS washing. The wound healing assays were done in growth factor-free medium, further excluding any effect due to a potential proliferation difference. The open gap was then inspected and photographed microscopically at indicated times, and is shown in the Figures at a 200X magnification. The migration activity was calculated as the number of cells entering into the rectangle. Experiments were repeated a minimum of three times.

Proliferation Assay

1 × 105 cells were plated on a P60 plate. Every 24 h, cells were trypsinized and counted under a light microscope at least three times until the sixth day. Experiments were repeated a minimum of three times.

Boyden Chamber Assay

This assay measures the ability of cells to invade a Matrigel matrix overlying a membrane containing 8-μm pores. Cells were seeded in medium deprived of EGF or FBS in the top chamber (BD), whereas medium containing EGF or FBS was added to the bottom chamber. After an appropriate cultivation time, the chambers were fixed with 1% paraformaldehyde and stained with hematoxylin. The number of cells in ten random fields of view was enumerated at 200X or 400X magnification for each filter. Three independent experiments were performed and the data are presented as the mean ± SD.

Three Dimensional Matrigel Culture

Matrigel (1.2 mg/ml, BD) was coated on the bottom of a 24-well plate. After Matrigel polymerization, cells were seeded into the well with growth medium containing 2% Matrigel. The cells were cultivated at 37°C incubation and alterations to the morphologic phenotype were monitored at 200X magnification every other day. Experiments were repeated a minimum of three times.

Anchorage-Independent Growth in Soft Agar

The soft agar assay was used to determine the propensity for anchorage-independent growth. Cells were plated in a 60-mm dish using 2 ml of growth medium, including 0.33% agar on the top of a bottom layer containing 0.66% agar. The cells were fed every two days with 1 ml medium. Colonies were photographed and counted in ten random fields of view at 200X magnification using light microscopy. Each experiment was done in triplicate.

Confocal Immunofluorescence Microscopy

Cells were seeded onto glass slides for 24 h, washed with PBS, fixed in 4% paraformaldehyde and permeabilized with 0.5% Triton X-100 for five minutes. After blocking with BSA, cells were stained with anti-snail primary antibody followed by FITC-conjugated anti-rabbit IgG. To visualize the nucleus, 4' 6-Diamidino-2-phenylindole (DAPI) staining was also performed, as previously described [85]. Immunofluorescence was detected by fluorescence microscopy (Olympus).

Mouse Injections, Necropsy, Histopathology

The ability to form tumors and metastasize was analyzed by injecting cells with repressed Bmi-1 into nude mice. Mice were bred and maintained under SPF conditions in the Department of Animal Center, Cancer Center, Sun Yat-Sen University, as approved by the China Care Committee Institute. Ten healthy female nude mice, which were four- to six- weeks old, were randomly assigned to each group. Each mouse was injected in the fat pad with 2 × 106 cells in PBS solution. Tumor growth was measured by caliper, and tumor volume was calculated according to the formula: length × width2 × 0.52, as described previously [86]. All mice were sacrificed on the sixth week after injection. The primary tumor and lung tissues of each mouse were removed, weighed and embedded in 10% paraffin. Each tissue was chopped into small pieces. Total protein was extracted to detect Bmi-1 expression from the primary xenografts. Each section from the primary xenografts and lung tissues was subjected to H&E staining, according to standard protocols, for histological examination and metastasis evaluation. The nodes of lung metastasis were quantified by counting metastatic lesions in ten sections (10 μm per section in a series). Data were collected by counting the total numbers of metastatic lesions from ten sections. Sections of primary tumors and lung lesions were used to detect the expression of the markers (Bmi-1, β-catenin, fibronectin) by IHC, as described previously.

Statistical analysis

The Chi-Square test was employed to evaluate the differences in Bmi-1 expression between the two categories of tissues. For assessment of the correlation between clinical features and Bmi-1 expression in breast cancer, P values were calculated by using the Chi-Square test or the Fisher's exact test. Relative risks (RRs) of death associated with Bmi-1 expression and other predictive variables were estimated by using the univariate and multivariate Cox proportional hazards model. The overall survival curve was plotted using Kaplan-Meier survival analysis and compared by the log-rank test. Result variations for the chamber invasion assays, wound healing assay, soft agar assay, tumor volume, tumor weight and lung metastasis lesions in mice, described as mean ± SD, were assessed using the two-tailed Student's t test. A value of P < 0.05 was considered significant (two tailed) by using SPSS 16.0.