Human ovarian cell lines, SKOV3 and A2780, were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The cells were maintained as a monolayer in high glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 100 IU/ml penicillin G and 100 μg/ml streptomycin (Life Technologies, Shanghai, China) at 37°C in a humidified 5% CO₂ incubator.

For the sphere-forming culture, the cells were collected and washed to remove the serum, prior to being suspended in serum-free DMEM/F12 supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin, 20 ng/ml human recombinant epidermal growth factor, 10 ng/ml human recombinant basic fibroblast growth factor, 2% B27 supplement without vitamin A and 1% N2 supplement (Invitrogen, Carlsbad, CA, USA). The cells were subsequently cultured in ultra-low attachment 6-well plates (Corning Inc., Corning, NY, USA) at a density of ≤2,000 cells/well.

The spheres were collected by gentle centrifugation (500 × g for 5 min), dissociated with trypsin-EDTA and mechanically disrupted with a pipette. The resulting single cells were centrifuged (500 × g for 5 min) to remove the enzyme, and re-suspended in serum-free medium where they were allowed to re-form spheres. The spheres were passaged every 8 days until they reached a diameter of 100 μm. Dissociated single sphere-forming cells (SFCs) were diluted to a density of 500 cells/ml. The diluted cell suspension was plated onto ultra-low attachment 96-well plates at 2 μl/well (Corning Inc.). Serum-free medium (150 μl) was then added. Wells with only one cell were marked and observed every day.

Four-week-old BALB/c-nu male mice (Shanghai Laboratory Animal Center, Chinese Academy of Sciences, Shanghai, China) were housed and maintained in accordance with the Institutional Guidelines of Hunan Normal University (Changsha, Hunan, China). The study was approved by the ethics committee of Hunan Normal University. The SKOV3 parental cells and the third passages of the SFCs were used in the tumorigenicity experiments. Trypan blue staining was used to assess cell viability. Various numbers of viable single cells in serum-free DMEM/Matrigel (1:1; BD Biosciences, Shanghai, China) were subcutaneously injected into the mice using a 100-μl microsyringe [Sangon Biotech (Shanghai) Co., Ltd., Shanghai, China]. The mice were humanely sacrificed 8 weeks after the injection, and the tumors were harvested for further examination.

The SFCs from the SKOV3 cell line and the parental cells were seeded in 96-well plates (precoated with Matrigel) at a density of 5,000 cells per well. The cells were exposed to increasing concentrations of DFOG. After 48 h, MTT reagent (Sigma-Aldrich, St. Louis, MO, USA) was added to each well according to the manufacturer’s instructions. Absorbance was measured at 570 nm.

The FOXM1 cDNA plasmid was purchased from OriGene Technologies Inc. (Rockville, MD, USA). The SFCs derived from the SKOV3 cell line were transfected with cDNA using Lipofectamine 2000 (Life Technologies), as previously described (17).

Western blot analysis was carried out, as previously described (18). Monoclonal mouse anti-FOXM1, monoclonal mouse anti-CD133 and monoclonal mouse anti-CD44 (Cell Signaling Technology, Inc., Danvers, MA, USA), as well as monoclonal mouse anti-aldehyde dehydrogenase 1 (ALDH1) and monoclonal mouse anti-β-actin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) antibodies were used as primary antibodies. The cells were lysed by being incubated in lysis buffer for 20 min at 4°C. The protein concentration was determined using the Bio-Rad assay system (Bio-Rad, Hercules, CA, USA). Total proteins were fractionated using SDS-PAGE and transferred onto a polyvinylidene fluoride membrane (Millipore, Billerica, MA, USA). Signals were detected using an Enhanced Chemiluminescence Advance western blot analysis system (Amersham Pharmacia Biotech Inc., Piscataway, NJ, USA).

Statistical analysis and database management was undertaken using SPSS version 15.0 software (SPSS, Inc., Chicago, IL, USA). Data are represented as the mean ± standard deviation. Multiple group comparisons were made using one-way analysis of variance, and pairwise comparisons were performed using the least squares difference method. A two-tailed t-test was used when appropriate. Values of P<0.05 were considered to indicate a statistically significant difference.

Ovarian cells were plated in stem cell-conditioned culture medium in 6-well plates at a density of 2,000 cells/well, which allowed the formation of discrete colonies. Under these conditions, the cells grew as non-adherent, three-dimensional sphere clusters. Fig. 1A shows the anchorage-independent spheres that formed in the SKOV3 and A2780 cells. The spheres were passaged after 8 days, when they had reached ~50 μm in diameter. The SKOV3 and A2780 spheres were serially passaged for >12 generations, indicating their self-renewal capability in vitro. To corroborate the finding that a sphere could be generated from a single cell, single SKOV3 cells were plated on a 96-well plate and the wells with one cell were visualized everyday. Fig. 1B shows the process by which a single SKOV3 cell formed a sphere.

Next, the ovarian CSC markers, CD133, CD44 and ALDH1, were evaluated using western blot analysis; the results showed enrichment of the CD133⁺, CD44⁺ and ALDH-high populations in the SFCs derived from the SKOV3 cells compared with the parental cells (Fig. 1C).

To confirm that the SFCs from the SKOV3 cells exhibited enhanced tumor-initiating capability, BALB/c-nu mice were transplanted with varying numbers of SKOV3 SFCs. SKOV3 parental cells were used as controls. The results indicated that as few as 1,000 SFCs were sufficient for tumor development, whereas, at least 2×10⁵ parental cells were necessary to consistently generate a tumor in the same model over a longer period of time (Table I). The tumor nodules formed by the SFCs of the SKOV3 cell line displayed similar histology to that observed with the parental cells (Fig. 1D). These findings indicated that the tumorigenic efficacy of the SFCs was higher than that of the parental cells, and that non-adherent tumor spheres from the ovarian cancer SKOV3 cell line cultured in stem cell-conditioned medium possess OCSLC properties.

It has been reported that CSCs have the characteristics of extensive proliferation, and GEN has been shown to inhibit the proliferative activity of pancreatic cancer stem cells (11). In the present study, the MTT results showed that DFOG (0.1, 1.0 and 10.0 μmol/l) and GEN (10.0 μmol/l) preferentially inhibited the proliferation of the SFCs derived from the SKOV3 cells (P=0.030; Fig. 2A), indicating that DFOG is able to preferentially suppress the proliferative ability of OCSLCs.

DFOG (0.1, 1.0 and 10.0 μmol/l) reduced the number of spheroids formed in the SFCs from the SKOV3 cells in a concentration-dependent manner (P=0.022; Fig. 2B). These results indicate that DFOG can suppress the self-renewal of OCSLCs.

We previously demonstrated the molecular role of FOXM1 in mediating the biological effects of GEN and DFOG in human ovarian cancer cell lines, including the SKOV3 cell line (17). A previous study reported that the overexpression of FOXM1 leads to EMT and the formation of a cancer stem cell phenotype in pancreatic cancer cells (12). Based on these findings, the present study next sought to compare the status of FOXM1 protein expression in the parental cells and SFCs. The results showed that FOXM1 expression was higher in the SFCs compared with the parental cells (Fig. 3A). In addition, FOXM1 expression in the SFCs was downregulated by DFOG (Fig. 3B).

In pancreatic cancer stem cells, the overexpression of FOXM1 has been shown to result in increased sphere-forming capacity and the increased expression of the CSC surface marker, CD44 (12). In the present study, DFOG inhibited the protein expression of CD133, CD44 and ALDH1 in the SFCs (Fig. 3B).

Western blot analysis showed that the upregulation of FOXM1 by pcDNA3.1-FOXM1 transfection resulted in the overexpression of FOXM1, CD133, CD44 and ALDH1 proteins in the SFCs from the SKOV3 cells (Fig. 4A). The overexpression of FOXM1 reversed the DFOG-induced downregulation of FOXM1, CD133, CD44 and ALDH1 protein expression levels (Fig. 4A), and reduced the inhibition of self-renewal to a certain extent (Fig. 4B). These results provide mechanistic evidence, indicating that DFOG-inhibited self-renewal is in part due to the inactivation of FOXM1 in OCSLCs derived from SKOV3 cells.