All reagents were of analytical grade and commercially purchased. Primary antibodies against MTA1 were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). E-cadherin was obtained from Epitomics Inc. (Burlingame, CA, USA). α-tublin, β-actin and alkaline phosphatase-conjugated anti-rabbit/mouse/goat IgGs were purchased from Sigma (Deisenhofen, Germany). FITC/Cy3-conjugated IgG was obtained from Proteintech Group Inc. (Chicago, IL, USA). 4,6-Diamino-2-pheylindoledi (DAPI), fibronectin (FN) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were purchased from Sigma. The SP histostain-plus kit was obtained from ZhongShan Biotech Co. (Beijing, China). All other chemicals were of analytical grade and obtained from standard suppliers.

The human paired poorly-metastatic prostate adenocarcinoma cell lines PC-3M-2B4 (2B4) and highly-metastatic prostate adenocarcinoma cell lines PC-3M-1E8 (1E8) were kindly provided by Dr Jie Zheng (Beijing University, Beijing, China) (33). All cells were cultured in RPMI-1640 medium (Invitrogen Life Technologies, Carlsbad, CA, USA) with 10% (v/v) fetal bovine serum (FBS; Sijiqing Co., Hangzhou, China), 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C and 5% CO₂ using a humidified incubator (Heraeus, Hanau, Germany).

For western blot analysis, cells were lysed in RIPA buffer containing 50 mM Tris/HCl (pH 7.2), 150 mM NaCl, 1% NP-40, 0.1% SDS and 0.5% (w/v) sodium deoxycholate. Equivalent amounts of cell extracts (150 μg) were separated on 10% SDS-polyacrylamide gel (SDS-PAGE) and transferred onto a polyvinylidene difluoride membrane (PVDF). Filters were blocked in 25 mM Tris (pH 8.0) containing 125 mM NaCl, 0.1% Tween 20 and 5% skimmed milk for 1 h, and then incubated with the diluted primary antibody (MTA1, 1:200; E-cadherin, 1:2000; β-actin, 1:1000) at 4°C overnight. After incubation, the secondary antibodies were added (1:1000 dilution). The immunoreactive bands were visualized with alkaline phosphatase and 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium (BCIP/NBT).

Cell lines were cultured in 6-well tissue culture plates and flasks at 37°C and 5% CO₂ using a humidified incubator (Heraeus). Cells were transfected with 200 pmol/ml of siRNA duplex using 5 μl Lipofectamine 2000 when cells were 20% confluent following the manufacturer’s instructions of Lipofectamine™ 2000 (Invitrogen Life Technologies) siRNA treatment. The controls included untransfected cells and transfected cells with a scrambled negative control siRNA (Ruibo Biotechnology Co., Guangzhou, China). For plasmid transfection, 4×10⁵ cells/6-well plate were prepared using 4 μg of plasmid and 10 μl Lipofectamine 2000 per well. The full length PCMV-SPORT6 E-cadherin plasmid was purchased from Yrbio (Hunan, China). It was reconstructed to pcDNA3.1 E-cadherin in our laboratory (Cancer Biology Medical Center, Tongji Hospital) and the empty vector pcDNA3.1 was preserved. Cells were harvested after 48 h and protein levels were measured using western blotting. E-cadherin siRNA sequences were as follows: forward, 5′-CAGACAAAGACCAGGACU-AdTdT-3′; reverse, 3′-dTdT GUCUGUUUCUGGUCCUGAU-5′.

Exponentially growing cells were detached from culture plates by trypsinization and seeded into 6-well plates (1×10⁵ cells/well in complete media) with 10% FBS in DMEM. The plates were maintained at 37°C in 5% CO₂ at saturated humidity in an incubator overnight. The following day, the confluent cell monolayers were converged and scrape-wounded using a micropipette tip. Floating cells were removed after three washes with serum-free medium, while scratched cells were observed under an inverted phase microscope to identify the scratch widths. Cells were cultured in serum-free culture media and images were captured under phase contrast microscopy every 12 h after wounding. For evaluation of ‘wound closure’, five randomly selected points along each wound were marked, and the horizontal distance of the migrating cells from the initial wound was measured. Each experiment was performed twice or in triplicate. Data collected were as the mean ± SD. The measurements were obtained by measuring the area of the wound using Image J software (available from the National Institute of Health website at http://imagej.nih.gov/ij/download/).

Transwell chambers with polycarbonate membrane filters with 24-well inserts, 6.5 mm diameter and 8 μm pore size (Corning Life Sciences, Corning, NY, USA) were coated with 1.5 mg/ml Matrigel™ (BD Biosciences, Franklin Lakes, NJ, USA). The lower chamber was filled with 600 μl RPMI-1640 containing 20% FBS and mouse embryonic fibroblast cell line (NIH 3T3) supernatant, acting as a chemotactic factor (CF), or just DMEM with 1% fetal calf serum (FCS). A total of 1×10⁴ cells/200 μl were seeded into the upper compartment and incubated for 48 h at 37°C and 5% CO₂. Following removal of the non-migratory cells from the upper surface of the filter, invasive cells that had migrated through to the lower surface of the filter were fixed in 2.5% (v/v) glutaraldehyde for 15 min and stained with crystal violet for 10 min. The number of invasive cells were calculated by counting at least three randomly chosen visual fields at 100-fold magnification (Leica, Solms, Germany). Three independent experiments were conducted for each set.

Adhesion was determined using an MTT assay. Exponentially growing cells were detached from culture plates by trypsinization. After washing, cells were resuspended in serum-free medium. Equal numbers of cells (4×10⁴ cells/well) were seeded into 96-well plates precoated with 1 μg/ml FN (Sigma), which proved to be the most evident in pre-experiments (data not shown). All FN-coated wells were compared with cells seeded in 1% (w/v) bovine serum albumin (BSA). After 1 h of incubation at 37°C, the plate was immersed in PBS containing 1 mmol/l MgCl₂ to remove non-adherent cells. The adhered cells were then measured using an MTT assay at 490 nm wavelength. Subsequently, 30 μl of MTT (5 mg/ml) was added 4 h prior to the end of the incubation period. Following incubation, the samples were centrifuged at 2200 × g for 5 min. Once the supernatant was removed, 150 μl of dimethyl sulphoxide (DMSO) was added to resolve the crystals and the optical density (OD) values of each well was measured at 490 nm with a microplate reader after 15 min. The OD values reflect the proportion of cells adhered to the FN-coated 96-well plate. The rate of adhesion was calculated using the following equation: (the OD value of test/the OD value of control) × 100. All experiments were conducted four times and repeated twice.

For immunofluorescence cytoskeleton staining, 1×10⁴ cells were seeded on glass coverslips (13 mm diameter). Once the coverslips were fixated with 4% paraformaldehyde (Merck, Haar, Germany) in PBS, cell membranes were permeabilized with 0.2% (v/v) Triton X-100 in PBS, and non-specific binding sites were blocked with 5% (w/v) BSA in PBS for 30 min at 37°C. Treatment with the first antibody (α-tublin, 1:50) was conducted at 4°C overnight. Subsequently, the cells were washed three times with cold PBS and incubated with Cy3-conjugated IgG (1:50 dilution) in PBS for 30 min at room temperature. Nuclei were stained for 5 min at room temperature with DAPI (2 μg/ml methanol). Cells were then rinsed with PBS and were observed under confocal microscopy (Olympus, Tokyo, Japan).

Tissues samples of normal prostate as well as localized and metastatic prostate cancer were obtained from the Department of Pathology at Tongji Hospital Affiliated to Huazhong University of Science and Technology (Hubei, China). Immunohistochemical staining for E-cadherin and MTA1 expression were conducted following standard procedures. The 5 μm paraffin sections cut on poly-L-lysine-coated microscopy slides were deparaffinized and rehydrated in graded alcohol. The sections were heated in 0.01 M citrate buffer (pH 6.0) in a 85–95°C microwave oven (3 times for 5 min each) and the non-specific binding sites were blocked with goat serum. Sections were incubated overnight at 4°C with rabbit monoclonal anti-human E-cadherin (1:500 dilution) and goat polyclonal anti-human MTA1 (1:100 dilution) and then washed with PBS. Biotinylated goat anti-rabbit immunoglobulin (Dako, Kyoto, Japan) was then added to the sections for 30 min at room temperature. Peroxidase-conjugated avidin (Dako) was applied once the sections had been washed with PBS. Peroxidase activity was detected by exposure of the sections to the solution of 0.05% 3,3′-diaminobenzidine (DAB) and 0.01% H₂O₂ in Tris-HCl buffer (DAB solution) for 3–6 min at room temperature. Finally, the slides were counterstained with hematoxylin. Negative control slides were processed similarly, but were incubated with PBS instead of primary antibody.

Data were analyzed using the ANOVA test. Statistical analysis was conducted using SPSS version 13.0 software. P<0.05 was considered to indicate a statistically significant difference. The correlation between the indicators was analyzed using Spearman correlation analysis.

As shown in Fig. 1A, E-cadherin was present in two prostate cancer cell lines, but at different expression levels. The protein expression levels of E-cadherin in 2B4 cells were over 2.5- to 3-fold higher compared with those in 1E8 cells. Since the 2B4 cells expressed the highest levels of endogenous E-cadherin, we selected this cell line to investigate the effects of heterologous E-cadherin expression on the cellular biological properties of prostate carcinoma cells in vitro. Western blot analysis demonstrated that E-cadherin siRNA is able to specifically downregulate E-cadherin expression and significantly increase MTA1 expression (Fig. 1B).

To determine whether E-cadherin affects the migration ability of 2B4 cells, a wound healing assay was conducted. The wound healing ability of cells reflects their movement and migration on the surface on which they are anchored to for growth. At 0, 12 and 24 h after wounding, transfected E-cadherin siRNA cells demonstrated an increased level of wound healing compared with the mock and control siRNA cells (Fig. 2). At 24 h after wounding, the wound of the E-cadherin siRNA transfected group had healed, while that of the cells from the control siRNA and mock groups had not closed (P<0.01 compared with the control cells).

The 2B4 cells were transfected with siRNA against E-cadherin for 48 h and a solid-phase adhesion assay was used to detect the alteration in cell adhesion ability. For the cells with a similar concentration of surface-coated FN, silencing of E-cadherin by siRNA significantly inhibited the adhesion process compared with the mock and control siRNA treated cells (Fig. 3A). 2B4 cells were transfected with siRNA against E-cadherin for 48 h and a Matrigel-coated transwell model was used to detect the alteration in cell invasion ability. The number of cells at the lower side of the membrane reflects the invasion ability of the cells being investigated. The number of cells that invaded to the lower side of the membrane were 320.22±12.50, 293.56±18.32 and 521.13±21.67 for the mock, control siRNA- treated and E-cadherin-siRNA-treated cells, respectively. Compared with the mock and control-siRNA treated group, the invasion ability of the cells treated with E-cadherin siRNA was significantly greater (Fig. 3B and = C). Each assay was repeated three times. The decreased adhesion ability as well as an increased invasion and migration ability, influences a change in the architecture and composition of the cytoskeleton. To examine whether the inhibition of E-cadherin expression affects the subcellular distribution and organization of the cytokeratin filaments, immunofluorescence analyses were conducted using E-cadherin siRNA-transfected, control-transfected siRNA and mock cells. In the E-cadherin siRNA-transfected cells we observed an increased elongated spindle-shape morphology with extended pseudopodia between adjacent cells and an increase in cellular polarity strength compared with the control siRNA and mock cells (Fig. 3D). These results suggest that E-cadherin plays a key role in controling the malignant phenotypes of 2B4 cells, including the adhesion ability, invasion ability and cytoskeleton organization.

The 2B4 cells were treated with E-cadherin siRNA and full length E-cadherin plasmid for various time periods between 24 and 48 h, and western blot analyis was used to detect E-cadherin and MTA1 protein expression. As demonstrated in Fig. 4A, 48 h after treatment with E-cadherin siRNA, E-cadherin expression was significantly downregulated, while MTA1 expression was significantly upregulated. Additionally, following empty vector pCMV-SPORT6 transfection, MTA1 expression was significantly downregulated and E-cadherin was significantly upregulated 48 h after treatment (Fig. 4B). All experiments were repeated three times.

E-cadherin expression was detected predominantly in the cytoplasm and membrane of the prostate cancer tissues. MTA1 protein positive signaling was detected in the cytoplasm or nucleus of cancer cells. A representative result of the immunohistochemistry for MTA1 and E-cadherin in prostate cancer is shown in Fig. 5. The samples probed with E-cadherin antibody demonstrated 2+ or 3+ membrane staining of the epithelial cells. Cytoplasmic staining was 0− or 1+. The positive staining in the prostate lesions had distinct circumferential E-cadherin immunoreactivity of high intensity, which revealed that E-cadherin is a cell membrane protein.

To further assess whether E-cadherin expression levels correlate with the expression of MTA1 in prostate cancer metastasis, we analyzed the results for E-cadherin and MTA1 immunoreactivity. Intact E-cadherin immunostaining was observed in 70% (14/20) of benign prostatic hyperplasia (BPH), 53.6% (15/28) of primary cancer and 16.7% (2/12) of metastatic lesion tissues. MTA1 expression levels in BPH, primary cancer and metastatic lesion tissues were 30 (6/20), 75 (21/28) and 25% (3/12), respectively. E-cadherin expression was decreased in the prostate cancer tissues compared with the BPH tissues (P<0.001), while MTA expression was increased in the prostate cancer tissues compared with the BPH tissues (P<0.05; Table I). Furthermore, we analyzed the correlation between the expression levels of E-cadherin and MTA1 in the tumor group. As shown in Table II, we observed a statistically significant correlation between E-cadherin and MTA1 immunoreactivity in the prostate cancer group (P<0.05; r_s=−0.434, respectively).