The TUBO and TUBO-P2J cell lines were cultured in vitro in high glucose DMEM (HyClone™; Thermo Fisher Scientific, MA, USA) supplemented with heat-inactivated 10% foetal bovine serum (FBS; Thermo Fisher Scientific) and 1% penicillin-streptomycin (HyClone; Thermo Fisher Scientific); the cells were maintained under 5% CO₂. Snail shRNA constructs targeting the Mus musculus sequences of Snai1 (NM_011427) were purchased from Sigma (Sh1-sequence: GTACCGGATGTGTC TCCCAGAACTATTTCTCGAGAAATAGTTCTGGGAGAC ACATTTTTTTG, Sh3-sequence: CCGG GATCTTCAACTG CAAATATTGCTCGAGCAATATTTGCAGTTGAAGATCT TTTTG). 5-Aza-dC (DNA methyltransferase inhibitor), SAHA (HDAC inhibitor), epirubicin and doxorubicin was purchased from Sigma.

The lentiviral transduction particles are produced from a library of sequence-verified lentiviral plasmid vectors for mouse snail gene (Sigma). Transduction of lentiviral particles were performed according to the manufacturer's instructions. Transduced cells were selected with 500 µg/ml of puromycin for 3–14 days. Purimycin-resistant colonies and separated single cells were picked. Each single cell was expanded to evaluate the knockdown status of the Snail gene via PCR and western blotting.

Cells (1×10⁶/dish) were seeded on 10 cm culture plate and incubated for 5 h for the cells to attach to the dishes. 5-Aza-dc (2.5 µM) and/or SAHA (0.5 or 2 µM) were treated and incubated at 37°C in a humid atmosphere of 5% CO₂ for 3 or 7 days, then cells were subcultured with trypsinization to prevent overgrowth.

Total RNA extracted from cultured cells was used as a template for reverse transcriptase reactions. Aliquots of cDNA were amplified using the mouse and human primer pairs (Table I). After an initial denaturation at 94°C for 5 min, the following was performed: 30 cycles of denaturation at 94°C for 30 sec, annealing at 55–60°C for 30 sec, and extension at 72°C for 30 sec. PCR was performed with a 2720 thermo cycler (Applied Biosystems, CA, USA). The reaction products were analysed in 1.5% agarose gels.

Cell lysates (40 µg/lane) were electrophoresed on polyacrylamide-SDS gels and then transferred to polyvinylidene fluoride membranes. Immunoblotting was performed by primary antibodies. E-cadherin, Snail-1 and GAPDH antibodies were obtained from Cell Signaling Technology (MA, USA) and estrogen receptor α (ESR1) antibody was purchased from Santa Cruz (TX, USA). HRP-conjugated secondary antibodies were purchased from Jackson Laboratories (Jackson ImmunoResearch, PA, USA).

To detect CD24 and CD44 expression, 2×10⁵ cells were stained with 0.5 µg/ml of PE-Cy5-conjugated CD24 and FITC-conjugated CD44 antibody (BioLegend) for 30 min at room temperature in the dark. For aldehyde dehydrogenase (ALDH) activity, cells were stained using the Aldefluor kit (Stem Cell Technologies Inc., Vancouver, Canada) according to the manufacturer's protocol. Stained cells were acquired with the BD FACSCanto II cytometry system and analysed with FlowJo software (version 10).

The migration and invasion assays were performed using 8.0-µm pore size 24-well insert systems (BD Falcon) with 2 mg/ml of Matrigel coating (invasion) or not (migration). Then, 5×10⁴ cells (migration) or 5×10⁵ cells (invasion) were added to the upper chamber and incubated for 4–6 h (migration) or 72 h (invasion). After incubation, the upper surface of the membrane was wiped with a cotton-tipped applicator to remove residual cells. Cells in the bottom compartment were fixed and stained with H&E. Cells in ten randomly selected fields at ×40 magnification were counted.

Six-well plates were covered with a layer of 0.5% agar in medium supplemented with 10% foetal bovine serum. Cells were prepared in 0.3% agar and seeded in triplicate at 3 different dilutions ranging from 1×10³ to 5×10⁵. The plates were incubated at 37°C in a humid atmosphere of 5% CO₂ for 2 weeks. Each experiment was repeated at least 3 times. Colonies were photographed between 10 and 12 days at an original magnification of ×40 under phase contrast.

The chemo-susceptibility was measured with an In Vitro Toxicology assay kit (TOX6, Sigma). Briefly, 0.5–1×10⁴ cells/well were seeded and attached in 96-well culture plates. Indicated doses of chemo-drugs were administered for 72 h. Cells were fixed in 10% trichloroacetic acid for 1 h at 4°C, stained with SRB for 15 min, and washed 3 times with 1% acetic acid. The incorporated dye was solubilized with 10 mM Tris base, pH 8.8. Absorbance was spectrophotometrically measured at 565 nm using an EL800 microplate reader (Bio-Tek Instruments, Winooski, VT, USA).

Cells were harvested with trypsin and suspended in 15 ml conical tubes at 1×10⁴ cells/ml. After irradiation with the indicated doses with X-ray generator (RapidArc^®; Varian Medical Systems Inc., Palo Alto, CA, USA), cells were diluted 10-fold and seeded in 6-well plates. After 72 h, the cells were stained with 0.5% crystal violet for 10 min and fixed with 4% paraformaldehyde. Colonies were counted under a microscope. For in vivo radiation treatment, TUBO-P2J/NC (1.5×10⁵) and TUBO-P2J/shSnail (2×10⁵) cells were implanted subcutaneously and tumour tissues were locally irradiated at 15-Gy with an X-ray generator under anaesthesia (ketamine 90 mg/kg and xylazine 10 mg/kg).

All of the procedures involving animals were approved by the INJE University College of Medicine Institutional Animal Care and Use Committee. All of the mice (BALB/C, 6–8 weeks of age) used in this study were purchased from Orient Bio (Taejun, Korea). To evaluate the tumour formation ability and growth, 1×10⁵ or 1×10⁴ cells were subcutaneously implanted on the backside of the mice. Tumour development was followed every day, and tumour sizes were measured two times per week. Tumour volumes were measured along three orthogonal axes (x, y, and z) and calculated as tumour volume = (xyz)/2. To analyse experimental metastasis, 2×10⁴ cells were intravenously injected through the tail vein and lung tissues were collected at day 19.

Differences between groups were analysed using an unpaired t-test. Error bars represent ± SD. All statistical analyses were conducted using Graph-Pad Prism Version 4.0 (GraphPad Software). Unless specified, statistically significant differences of p<0.05, 0.01, and 0.001 are noted.

Previously we reported on the the highly metastatic mouse breast cancer cell line TUBO-P2J, derived from the non-metastatic TUBO cell line (10). TUBO cells were morphologically epithelial in nature. In contrast, TUBO-P2J cells were spindle-shaped mesenchymal-type cells. TUBO-P2J cells showed a gain of Snail and loss of E-cadherin. In flow cytometry analysis and soft agar colony assays, the TUBO-P2J cell line was characterized as stem cell-like with a prominent CD44^high population (Fig. 1A) and 5 times higher colony number (Fig. 1B) than TUBO cells. Chemo and radiation susceptibility tests revealed that TUBO-P2J cells were more resistant to chemotherapy and radiation therapy (Fig. 1C and D). Snail expression levels were associated with poor prognosis phenotypes, including metastatic potential, soft agar colony formation capacity and resistance to chemotherapy and radiotherapy.

Before Snail silencing, we compared the migration potentials of TUBO-P2J cell line with other breast cancer cell lines, such as MCF7, MDA-MD-231, HCC70, 4T1, and original TUBO cell lines (Fig. 2A). The migration activity of TUBO-P2J was higher than that of human breast cancer MDA-MB-231, which is a well-known cell line with a highly metastatic potential. TUBO-P2J cell line also showed the highest level of snail expression among the tested breast cancer cell lines (Fig. 2B). In addition, TUBO-P2J cell line did not express estrogen receptor α1 (ESR1) similarly to MDA-MB-231, triple-negative breast cancer cell line (Fig. 2C). Even though TUBO-P2J is a mouse tumour cell line, it is derived spontaneously in vivo and could be exactly compared with parent TUBO cells which showed less malignant phenotypes, including metastatic potential, stemness, chemoresistance and radioresistance (10). Based on these data, we chose the TUBO-P2J cell line for Snail silencing experiments with efficient lentiviral transfection system.

To examine the role of Snail in maintaining the mesenchymal phenotype and the suppression of E-cadherin, we produced a stable Snail knockdown TUBO-P2J cell line (TUBO-P2J/shSnail) by lentiviral transfection. Two independent stable shSnail-expressing clones (Sh1 and Sh3, one clone from each construct) and a stable negative control shRNA clone (NC) were analysed for Snail mRNA expression. Sh1 and Sh3 showed 50–90% downregulation of Snail mRNA, whereas the control clone did not exhibit any significant reduction in Snail mRNA (Fig. 3A). Knockdown of Snail in TUBO-P2J cells was confirmed with western blot analysis (Fig. 3B). Snail protein levels were correlated with mRNA levels. The morphology of Snail knockdown cells (Sh1 and Sh3) and control cells (NC) was not different, as all of the cells showed a spindle-shaped mesenchymal phenotype (Fig. 3C). Next, we evaluated the expression of E-cadherin with RT-PCR and western blot analysis (Fig. 3D and E). In TUBO-P2J cells, Snail knockdown did not induce transcription of E-cadherin. Based on Snail expression levels, we chose the Sh1 clone as TUBO-P2J/shSnail for further experiments.

Lim et al (12) reported that Snail represses the transcription of E-cadherin through binding to E-boxes of this gene and inducing DNA methylation of its promoter by recruiting HDAC1 and DNMT. To evaluate the reason that E-cadherin was not expressed in TUBO-P2J/shSnail cells, cells were treated with 5-aza-dC and/or SAHA. By the third day of 5-azaC treatment, E-cadherin mRNA was detected in all TUBO-P2J cells regardless of Snail knockdown (Fig. 4A). However, E-cadherin protein was detected only in shSnail cells by treatment of 5-azaC and SAHA for 7 days (Fig. 4B). These data suggest that Snail might suppress the expression of E-cadherin through a translation step co-operating with HDAC in TUBO-P2J cells.

Because processes of EMT have been linked with metastasis, we next evaluated the metastatic potential of PLKO and Sh1 cells. By silencing Snail, migration was significantly decreased from 195.3±7.5 (mean ± SE) cells/fields to 86.7±4.7 cells/fields (56% decrease) (Fig. 5A) and invasion was decreased even more from 127.1±47.6 cells/fields to 10.7±2.7 cells/fields (93% decrease) (Fig. 5B). An in vivo metastasis assay using tail vein injection also showed that lung colony numbers of shSnail cells were reduced to one-third of that of control cells (Fig. 5C). These data suggest that inactivation of Snail could reduce the metastatic potentials of breast cancer cells without driving MET.

As Snail is expressed in CD44^high compared to CD44^low cells, we examined whether Snail could affect the expression of CD44, a breast cancer stem cell marker. Snail silenced TUBO-P2J cells showed a slight decrease in the CD44^high population (37.5%) compared to their control counterparts (52.9%) (Fig. 6A). Aldehyde dehydrogenase (ALDH) activity was also reduced significantly by Snail silencing (Fig. 6B). Next, we used a soft agar colonization assay to evaluate the role of Snail in anchorage-independent growth. Data showed that the sizes of the colonies were not different with Snail silencing, but the numbers were significantly reduced by 40% compared to that of control cells (Fig. 6C). In vivo tumourigenicity was tested with subcutaneous implantation of TUBO-P2J, TUBO-P2J/NC, and TUBO-P2J/shSnail (1×10⁴ or 1×10⁵ cells/mouse). In all of the mice implanted with parent TUBO-P2J and TUBO-P2J/NC cells, tumours developed at day 14, but tumours developed in two mice out of seven at 30 days when 1×10⁴ of TUBO-P2J/shSnail cells were implanted (Table II). Moreover, a decrease in tumour size was observed in tumours established with Snail silenced cells compared with those from parent and control cells (Fig. 6D). These in vitro and in vivo data demonstrate that Snail has an important role in the maintenance of breast cancer stem cell properties.

As Snail is expressed in chemo- and radioresistant cells, TUBO-P2J, we examined whether Snail could induce chemo- and radioresistance. Our data revealed that silencing Snail increased the susceptibility to epirubicin and doxorubicin (Fig. 6E and F); however, susceptibility to radiation therapy was not affected by Snail silencing (Fig. 6G and H). These results suggested that Snail silencing might be a good strategy to overcome chemoresistance.

To test whether the phenotype changes in Snail silent cells were mediated by other EMT related genes, we evaluated mRNA levels of EMT related genes such as SLUG (Snail2), vimentin, fibrinectin, Twist, DDR2, and FoxC2 (Fig. 7A). Data revealed that silencing of Snail reduced the expressions of fibronectin and DDR2 but did not influence on the expressions of other EMT related genes including vimentin, SLUG, and Twist. In addition, silencing of Snail did not change the activation status of β-catenin (Fig. 7B). These data suggested that the malignant phenotypes of TUBO-P2J cell lines which are increased stemness, metastasis, and chemoresistance are mediated by the increased expression of Snail and the role of Snail in this model was not linked with the expression of vimentin and SLUG and Wnt/β-catenin signaling pathway.