Dox (Solarbio, Beijing, China) was dissolved in sterile water in a 10 mM stock solution. WP1066 (Selleckchem, Shanghai, China) and 3-(4,5-dimethyl-2-thiazolyl)-2,5-di-phenyl-2-H-tetrazolium bromide (MTT; Solarbio) were lysed in DMSO (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) in 50 mM and 5 mg/ml stock solutions, respectively.

The human MDA-MB-468 cell line, mouse breast cancer 4T1 cell line and human MCF-7 cell line were obtained from American Type Culture Collection (Manassas, VA, USA). The cells were cultured in RPMI-1640 supplemented with 10% FBS (both from HyClone; GE Healthcare Life Sciences, Logan, UT, USA) and 1% penicillin/streptomycin (Beijing Solarbio Science & Technology Co., Ltd.). All cultures were maintained at 37°C in a humidified 5% CO₂ incubator. The MCF-7/Epr human epirubicin-resistant breast cancer cell line was obtained from Shanghai Bogoo Biotechnology (Shanghai, China), which was originally selected with 1,000 ng/ml epirubicin.

Tumorsphere culture was performed in low attachment dishes (Corning Incorporated, Corning, NY, USA). The cells were seeded at a density of 8,000 cells per well into a 6-well plate supplemented with 1X B27 and 2X N-2 (both from Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), 20 ng/ml EGF, 20 ng/ml IGF, 10 ng/ml FGF-basic (all from PeproTech, Inc., Rocky Hill, NJ, USA) and 5 μg/ml heparin (Beijing Solarbio Science & Technology Co., Ltd.). The cells were maintained at 37°C in a humidified incubator containing 5% CO₂. Following 7 days of culture, the wells were examined under an inverted microscope at ×400 magnification.

The Dox-resistant tumor cells and parental tumor cells were harvested by cytospin, respectively, and fixed in 4% paraformaldehyde for 10 min. The slides were washed three times in 1X PBS and, using goat sera (Wuhan Boster Biological Technology, Ltd., Wuhan, China), the cells were blocked for 1 h. The cells were incubated with CD133 (1:500 dilution, cat. no. 18470-I-AP; ProteinTech Group, Inc., Chicago, IL, USA) and ABCG2 (1:500 dilution, cat. no. ab130244; Abcam, Cambridge, MA, USA) primary antibodies at 4°C. Following a wash in PBS, the cells were incubated with immunoglobulin G/tetramethylrhodamine goat anti-rabbit antibody (1:200 dilution, cat. no. ZF0316; OriGene Technologies, Inc., Beijing, China) for 1 h at 4°C. The cells were then washed with PBS, stained with DAPI for 10 min, washed, and covered with mounting medium. Immunofluorescent staining was observed and images were captured.

Total protein was extracted from the Dox-resistant TNBC cells and parental cells, respectively. Briefly, the cells were lysed with RIPA lysis buffer combined with PMSF (both from Beijing Solarbio Science & Technology Co., Ltd.) and then centrifuged at 12.000 × g for 15 min at 4°C. The supernatants were collected for western blot analysis and protein concentration was determined using the BCA method (CWBio, Beijing, China). Total protein (25 μg) was separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto a polyvinylidene fluoride membrane (EMD Millipore, Billerica, MA, USA). The membranes were blocked with 5% non-fat dry milk in TBST for 1 h and then blotted overnight at 4°C with the following primary antibodies: Mouse monoclonal β-actin (1:3,500 dilution, cat. no. 66009-1-Ig; PeproTech, Inc.); mouse monoclonal Stat3 (1:1,000 dilution, cat. no. 9139s); mouse monoclonal phosphorylated (p)-Stat3 (1:1,000 dilution, cat. no. 4113s) (both from Cell Signaling Technology, Inc., Danvers, MA, USA); rabbit monoclonal Oct-4 (1:1,000 dilution, cat. no. ab200834); rabbit monoclonal c-Myc (1:5,000 dilution, cat. no. ab32072); rabbit monoclonal Sox2 (1:1,000 dilution, cat. no. ab92494); mouse monoclonal ABCG2 (1:1,000 dilution, cat. no. ab130244) (all from Abcam) and rabbit polyclonal CD133 (1:1,000 dilution, cat. no. 18470-1-AP; PeproTech, Inc.) antibodies. The membranes were then blotted for 1 h at room temperature with anti-mouse (1:5,000 dilution, cat. no. A0216) and anti-rabbit (1:3,500 dilution, cat. no. A0208) secondary antibodies (both from Beyotime Institute of Biotechnology, Shanghai, China). The signal was detected using enhanced chemiluminescence (Immobilon Western Chemiluminescent Horseradish Peroxidase Substrate, cat. no. WBKLS0100; EMD Millipore) and recorded on X-ray film.

The dissociated MDA-MB-468-Dox cells, 4T1-Dox cells and their parental cells were counted and transferred into a tube, respectively. The cells were washed and resuspended in PBS at a concentration of 10⁶/ml. The cell suspensions were incubated with PE-rat anti-mouse CD44 antibody, which also cross-reacts with human cells (1:100 dilution, cat. no. 553134; BD Biosciences, Franklin Lakes, NJ, USA) for 30 min at room temperature. Following washing with PBS, the samples were analyzed using a FACSCalibur. Isotype-matched PE-rat IgG2b (1:100 dilution, cat. no. 553989; BD Biosciences) served as negative control.

The MDA-MB-468-Dox cells, 4T1-Dox cells, MCF-7-Dox cells and their parental cells were seeded into 96-well plates at a density of ×10³ cells/well in RPMI-1640 medium supplemented with 10% FBS. The cells were treated with Dox at various concentrations (0, 0.05, 0.1, 0.5, 1, 5 and 10 μM) for 48 h at 37°C. MTT solution (5 mg/ml) was added to the plates (20 μl per well). The plates were incubated at 37°C for an additional 4 h. The supernatant was discarded, and 150 μl DMSO was added to each well. The plates were thoroughly mixed and read at 570 nm on a microplate reader.

Each experiment was performed at least three times. Data are presented as the mean ± standard deviation. A one-way analysis of variance was performed for comparisons between two subgroups, and the Fisher's least significant difference post hoc test was used for determination of the significance among the means of multiple groups using SPSS version 17.0 software (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

Dox has been documented to induce the resistance of ER-positive MCF-7 breast cancer cells by exposure to a high dose of Dox. To establish Dox-resistant TNBC cell lines, human MDA-MB-468 and murine 4T1 cells were treated with a long-term low dose of Dox. Based on the dose-dependent curve of Dox, the concentration of 0.05 μM Dox was used to stimulate MDA-MB-468 cells and 0.1 μM of Dox was used to stimulate the 4T1 cells. Following continuous treatment with Dox for 4 weeks, morphological differences were observed between the parental cells and the Dox-resistant TNBC cells (Fig. 1A and B). These Dox-treated cells were termed MDA-MB-468-Dox cells and 4T1-Dox cells, and were banked for further investigations.

The cytotoxicity of Dox in the Dox-resistant cell lines and their parental cells was examined using an MTT assay. The dose-dependent effects of Dox in the above two pairs of cell lines are shown in Fig. 1C and D. Compared with the parental cancer cells, the cytotoxicity of Dox towards the Dox-resistant TNBC cells was significantly decreased. The EC₅₀ of Dox in the parental MDA-MB-468 cells was 0.47 μM, whereas the EC₅₀ of the MDA-MB-468-Dox was increased to 0.92 μM (P<0.05; Fig. 1C). Similarly, the EC₅₀ of the 4T1-Dox cells was significantly increased from 0.98 to 5.37 μM (P<0.05; Fig. 1D). These data suggested that the sensitivity of TNBC cells against Dox was significantly decreased by long-term exposure with Dox at a low concentration.

To compare Dox resistance in TNBC cells to ER-positive cells, a human epirubicin-resistant breast cancer cell line MCF-7/Epr was induced into a Dox-resistant cell line, termed MCF-7-Dox, by prolonged exposure to 0.9 μM Dox for 4 weeks (Fig. 1E). The MTT assay demonstrated that the ER-positive MCF-7-Dox cells, compared with the original cell line, was less sensitive to Dox (Fig. 1F). The results confirmed the results of a previous report that a high dose of Dox is required for maintenance of the drug-resistant phenotype of MCF-7/ADR cells (5).

As cells that are intrinsically resistant to chemotherapy may have CSC characteristics (8,21,22,24,28,36), the present study examined whether MDA-MB-468-Dox and 4T1-Dox cells have properties of the CSC phenotype through examining specific CSC markers expressed on the surface of the Dox-resistant cells. The Dox-resistant breast cancer cells were analyzed for the expression of the CSC surface marker CD44 by flow cytometry. As shown in Fig. 2A–C, the expression of CD44 in the Dox-resistant TNBC cells was higher than that in the parental cells. As expected, exposure to Dox increased the expression of CD44 in MCF-7-Dox cells, which was consistent with the observation by Calcagno et al (5).

The expression levels of CD133 and ABCG2 in the MDA-MB-468-Dox, 4T1-Dox and MCF-7-Dox cells were observed by western blot analysis. The levels of CD133 and ABCG2 were markedly higher in the MDA-MB-468-Dox and 4T1-Dox cells, compared with those in their parental cells (Fig. 2D). The high expression levels of CD133 and ABCG2 in 4T1-Dox cells were also observed by immunofluorescence (Fig. 2E). In the MCF-7-Dox cells, continuous stimulation with Dox upregulated the expression of CD133, but did not affect the expression of ABCG2. These data suggested that the Dox-resistant TNBC cells exhibited typical CSC molecular properties with high expression levels of CD44/CD133/ABCG2, which differed from the CSC biomarker characteristics of MCF-7-Dox cells.

The present study also performed mammosphere formation assays to evaluate the sphere-forming ability of Dox-resistant 4T1 cells. As shown in Fig. 2F, the volume of the sphere formed by Dox-resistant 4T1 cells was larger than that of the parental 4T1 cells. The numbers of mammospheres of the Dox-resistant 4T1 cells were also higher compared with those of the control 4T1 cells. These data obtained from the expression of surface molecular markers and mammosphere formation capacity indicated that continuous stimulation with Dox at a low dose induced the enrichment of CSCs in TNBC cells.

The high expression levels of pluripotent transcription factors, including Oct-4, Sox2 and c-Myc, has been reported in CSCs, which may promote stem cell self-renewal and differentiation (3,18,24,26). To clarify the roles of pluripotent transcription factors in regulating the enrichment of CSCs induced by Dox, the present study examined the expression of these genes in BCSCs induced by Dox. As shown in Fig. 3A, compared with their parental cancer cells, the expression levels of Oct-4 and c-Myc were increased in the Dox-resistant TNBC cells, whereas the expression of Sox2 was not altered significantly. However, the expression of Sox2 in MCF-7-Dox cells was increased. These results indicated that the upregulated expression of Sox2 may be a major factor in the CSC activities of MCF-7-Dox cells. Notably, Dox induced the expression of Oct-4 and c-Myc, but did not affect the expression of Sox2 to promote the CSC enrichment in the TNBC cells.

In addition to specific pluripotency transcriptional factors, Stat3 is also reported to be a key mediator for mammosphere formation and the CSC process in breast cancer (3,15,18,49,50). To investigate whether Stat3 is involved in CSC enrichment induced by Dox, the present study examined the expression and activation of Stat3. As shown in Fig. 3B, the phosphorylation of Stat3 was elevated in Dox-resistant breast cancer cells, whereas there was no significant change in the total expression of Stat3. This finding suggested that the activation of Stat3 was also involved in the regulation of CSC formation induced by Dox in breast cancer cells.

From the above data, it was observed that the pluripotency transcriptional factors Oct-4 and c-Myc, and the activation of Stat3 were involved in the Dox-induced CSC enrichment in TNBC. The CSCs transcriptional factor c-Myc is a recognized transcriptional target of Stat3 in embryonic stem cells and hematopoietic stem cells (18,47). Oct-4 also acts as a downstream factor of Stat3 during mammosphere culture (3). Therefore, the Stat3 inhibitor WP1066 (51–56) was used to examine whether the inactivation of Stat3 affects the interaction between Stat3 and Oct-4/c-Myc in the process of Dox-induced CSC enrichment. Treatment with WP1066 (1.25 μM) not only decreased the phosphorylation of Stat3 (Fig. 4A), but also markedly downregulated the expression of Oct-4 and c-Myc (Fig. 4A). These data suggested that continuous Dox stimulation promoted the enrichment of CSCs through the activation of Stat3, which increased the expression of pluripotency transcriptional factors Oct-4 and c-Myc.

An MTT assay was performed to evaluate whether WP1066 restored the sensitivity of MDA-MB-468-Dox and 4T1-Dox cells to Dox. Following incubation with WP1066 (1.25 μM), the EC₅₀ of MDA-MB-468-Dox cells was decreased from 0.32 to 0.14 μM (P<0.05; Fig. 4B). Similarly, the EC₅₀ of 4T1-Dox was decreased from 5.37 to 2.24 μM (P<0.05; Fig. 4C). There was no significant difference in EC₅₀ values between the WP1066 treatment group and parental MDA-MB-468 or 4T1 cells (P>0.05). These data suggested that WP1066 treatment sensitized the MDA-MB-468-Dox and 4T1-Dox cells to Dox, as demonstrated by EC₅₀ values in WP1066-treated cells, which were similar or close to those in parental cells. In addition, following treatment with WP1066 at the concentration of 1.25 μM, the maximal response to Dox in the 4T1-Dox cells was lower than that in the parental 4T1 cells, indicating that this concentration of WP1066 did not completely restore the sensitivity of 4T1-Dox cells, which may be improved by increasing the dose and incubation time of WP1066.

The present study also examined whether WP1066 affects the expression of CSC markers on Dox-resistant cells. As shown in Fig. 5A, the increased expression of CD133 and ABCG2 on the MDA-MB-468-Dox cells was reduced by WP1066 treatment (Fig. 5A). The effect of WP1066 treatment on the CD44-positive cell population was also analyzed by flow cytometry. WP1066 treatment reduced the CD44-positive cell population in the MDA-MB-468-Dox cells (Fig. 5B) and 4T1-Dox cells (Fig. 5C). Analysis of the peak channel of the flow cytometry assay showed that, under treatment with WP1066, the mean peak value of 4T1-Dox was shifted left from 731 to 533. In addition, the mean peak value of MDA-MB-468-Dox was shifted left from 896 to 771. These results indicated that the inhibition of Stat3 downregulated the Dox-induced enrichment of the CSC population of TNBC cells, and led to reversion of TNBC cell resistance to Dox.