The SKOV3 ovarian cancer cell line was obtained from the Shanghai Cell Bank of Chinese Academy of Sciences and maintained in McCoy's medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂. Later cells were dissociated using 0.25% trypsin-ethylenediaminetetraacetic acid (EDTA) for 1–2 min at 37°C and maintained under stem cell conditions using serum-free Dulbecco's modified Eagle's medium (DMEM)/F12 supplemented with 5 mg/ml insulin (Sigma-Aldrich), 10 ng/ml human recombinant epidermal growth factor (Invitrogen, Carlsbad, CA, USA), 10 ng/ml basic fibroblast growth factor (Invitrogen), 12 ng/ml leukemia inhibitory factor (Gibco, Paisley, UK) and 0.3% bovine serum albumin (Sigma-Aldrich). The selected cancer cells formed non-adherent spheres grown in this condition. The medium was changed every 2 days by centrifuging at 800 rpm for 5 min to remove the dead cell debris. Regular cell culture plates were used for the experiment. The other tumor cells were maintained under standard conditions (DMEM/F12 supplemented with 10% FBS without growth factors) and formed attached differentiated cells.

Appropriate number of the UA or UA in combination with cisplain in SKOV3 cells and SKOV3 sphere cells along with equal number of their respective controls were cultured in 96-well plates at 37°C in 5% CO₂ incubator for 48 h. The experimental concentration of UA and cisplatin was 3.125, 6.25, 12.5, 25, 50 and 100 µg/ml, respectively. At the endpoints, cells were incubated with thiazolyl blue tetrazolium bromide (Sigma-Aldrich) at a concentration of 0.5 mg/ml for further 4 h. Resulting formazan crystals were dissolved with 100 µl of dimethyl sulfoxide, and proliferation was monitored by the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay and optical density (OD) reading at 490 nm. Then the inhibition rate (IR) and half maximal inhibitory concentration (IC₅₀) of the two kinds of cells were calculated.

SKOV3 sphere cells were treated with UA and UA combined with cisplain at IC₅₀ concentrations for 48 h. The experiment was divided into four groups: SKOV3 cells, SKOV3 sphere cells, SKOV3 sphere cells with UA and SKOV3 sphere cells with UA plus cisplatin. Then, cells in each group were collected and fixed in 70% cold ethanol at −20°C overnight. After washing twice with PBS, cells were resuspended in PBS. RNaseA (0.02 mg/ml) and propidium iodide (PI) (0.02 mg/ml) were added to the fixed cells for 1 h at 4°C. The DNA content of cells was then analyzed using a flow cytometer. The percentage of cells in the different cell cycle phases was calculated using BD FACSDiva™ software (BD Biosciences, San Jose, CA, USA).

SKOV3 sphere cells were treated with UA at IC₅₀ concentrations and UA plus cisplatin for 48 h. Then, cells in four groups were digested with 0.25% trypsin without EDTA and washed twice with PBS and then re-suspended in the binding buffer, with the cell density adjusted to 2×10⁵/ml. The cell suspension (195 µl) was obtained, and Annexin V-fluorescein isothiocyanate (FITC) (5 µl) and PI (10 µl) were added, respectively. The mixture was kept at room temperature for 30 min and then measured for apoptosis using flow cytometry.

SKOV3 sphere cells were treated with UA and combined with cisplain at IC₅₀ concentrations for 48 h. The experiment was divided into four groups as above. Invasion assays were performed in triplicate using 8.0 µm Transwell invasion chambers (Corning Costar, Rochester, NY, USA) coated with Matrigel (100 µg per filter) (Becton-Dickinson, Franklin Lakes, NJ, USA) as described in the manufacturer's instructions. Cells (1×10⁵/well; 200 µl per chamber) in serum-free media were seeded onto top chambers. Complete medium (600 µl) with 10% FBS was added to the lower chambers. Following 48-h incubation, cells that had invaded through the surface of the membrane were fixed in methanol and stained with crystal violet. Cells that did not invade into the lower chamber were scraped from the top of the Transwell plate with a cotton swab. Invading cells from three random microscopic fields per filter were selected for cell counting. Procedure for the migration assay was similar to that for the invasion assay, differing in that for the migration assay the Transwell chambers were not coated with Matrigel and 5×10⁴ cells/well were seeded onto top chambers and incubated for 16 h.

Before and after the treatment of IC₅₀ concentration of UA and cisplain in vitro, the expression of marker gene mRNA of SKOV3 cells and sphere cells was measured. Total RNA was extracted from SKOV3 sphere cells and SKOV3 cells using the RNeasy Mini kit (Qiagen, Valencia, CA, USA). In total, 500 ng of total RNA from each sample was utilized for reverse transcription using the iScript cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA, USA). Real-time PCR was performed on cDNA using iQ SYBR-Green with Mastercycler ep realplex real-time PCR system (Eppendorf, Hamburg, Germany). All reactions were performed in a 25-ml volume. PCR was performed by an initial denaturation at 95°C for 5 min, followed by 40 cycles for 30 sec at 95°C, 30 sec at 60°C and 30 sec at 72°C. PCR using water instead of the template was used as a negative control. Specificity was verified by melting curve analysis and agarose gel electrophoresis. The threshold cycle (Ct) values of each sample were used in the post-PCR data analysis. 18S RNA was used as an internal control for mRNA-level normalization. The primer sequences for each gene analyzed are summarized in Table I.

The in vivo evaluation of UA was performed using a xenograft model of ovarian cancer SKOV3 sphere cells. Athymic BALB/c-nu female nude mice (5–6 weeks old, obtained from Beijing HFK Bioscience Co., Ltd., Beijing, China) were housed in a specific pathogen-free room within the animal facilities at the Laboratory Animal Center. Animals were allowed to acclimatize to their new environment for 1 week prior to use. The dissociated sphere SKOV3 cells (5×10⁶) were resuspended in PBS, and injected s.c. into the left side of flank of nude mice. Engrafted mice were inspected for the appearance of tumor by visual observation and palpation until the tumor formed. From the 20th day of injection, mice were randomly assigned to three treatment groups (n=4 for each group) and injected intraperitoneally (i.p.) with normal saline, UA (60 mg/kg body weight, daily) and UA combined (60 mg/kg body weight, daily) with cisplatin (2.5 mg/kg body weight, daily) treatment for 14 consecutive days. Body weight and tumor mass were measured every 2 days. Tumor volume was determined using a caliper and calculated according to the formula (width² × length × π)/6. Mice were sacrificed by cervical dislocation under anesthesia after 2 weeks of treatment. Animal welfare and experimental procedures were performed strictly in accordance with high standards for animal welfare and other related ethical regulations approved by the Shanghai University of Traditional Chinese Medicine.

Immunohistochemical studies were performed on the xenograft tumors after they were removed from nude mice. The tumors were fixed in 40 mg/ml paraformaldehyde, paraffin-embedded and cut into 4µm serial sections. Next, endogenous peroxidases were quenched and the sections were washed carefully with phosphate-buffered saline (PBS) three times. The sections were blocked with 2% goat serum and rabbit serum, respectively, in PBS at 37°C for 45 min, then incubated with mouse anti-proliferating cell nuclear antigen (PCNA) antibody (1:200 dilution; Abcam), rabbit anti-Ki-67 antibody (1:200 dilution; Millipore), mouse anti-vimentin antibody (1:300 dilution; Abcam), mouse anti-fibronection antibody (1:300 dilution; Abcam) overnight at 4°C. Later, the sections were incubated with horseradish peroxidase-conjugated secondary antibodies separately and avidin-biotin complex followed by diaminobenzidine (Vectastain ABC; Vector Laboratories Burlingame, CA, USA). The sections were immersed in 2% ammonia water and hematoxylin was used for counterstaining.

Positive PCNA and Ki-67 staining was mainly in the nuclei, while positive expression of vimentin and fibronection was primarily a cytoplasmic pattern in tumor cells. For evaluation of positive expression, staining intensity was scored as 0 (negative), 1 (weak), 2 (medium) or 3 (strong). Strong positive (scored as 3), strong staining intensity (90% of positive cells); moderate positive (scored as 2), moderate staining intensity (50–89% of positive cells); weak positive (scored as 1), weak staining intensity (10–49% of positive cells); absent (scored as 0), no staining intensity and no positive or only a few positive cells (11).

Data are presented as the mean ± standard deviation. Student's t-test was performed to evaluate the difference between mean values. P<0.05 was considered to indicate a statistically significant result. All experiments were performed in triplicate.

It has been reported that ovarian cancer stem-like cells could be enriched and exhibit characteristics expected of CSCs (12–15). In the present study, attempts were made to isolate a self-renewing stem cell population from the SKOV3 cell line. The SKOV3 cell line was cultured with McCoy's medium supplemented with 10% FBS (Fig. 1A). Under serum-free condition, the SKOV3 cells were able to form non-adherent spheres. The formation of sphere cells was observed on day 3 after plating (Fig. 1B and C). These cluster cells were small, non-adherent and non-symmetric. Primary spheres could be enzymatically dissociated to single cells, which in turn give rise to secondary spheres. This procedure could be repeated, and the tumorigenic spheres grow faster than the cells under differentiating conditions (Fig. 1D). The stem/progenitor cell phenotype of the sphere cells was further confirmed by the expression of putative stem cell markers. Quantitative real-time PCR showed that the expression of Nanog, Oct-4, Sox-2, CD133, CD117 and ABCG2 in sphere cells was higher than that in differentiated cells (Fig. 1E; P<0.01).

UAs are triterpenoid compounds that exist widely in food, medicinal herbs and other plants (Fig. 2A). To demonstrate its effects on ovarian cancer cells, MTT assay was performed, which showed that the proliferation rate was significantly decreased in the UA-treated cells when compared with the non-treated cells (P<0.05 or P<0.01). UA inhibited the proliferation of the cells in a concentration-dependent manner (Fig. 2C). The IR of UA on the SKOV3 cells is much higher than that of the sphere cells, while IC₅₀ of SKOV3 cells is 12.04 mg/l, whereas that of sphere cells is 74.54 mg/l. To determine whether UA could enhance the cisplatin cytotoxicity to ovarian cancer cells, SKOV3 sphere cells were exposed to different concentrations of UA, cisplatin and a combination of UA and cisplatin. The IC₅₀ of cisplatin-treated SKOV3 and SKOV3 sphere cells was 70.38 and 217.73 mg/l, respectively. When SKOV3 sphere cells were co-treated with 75 mg/l of UA, the cell viability was significantly decreased compared with treatment with UA alone or cisplatin alone (Fig. 2B). Thus, UA may regulate cisplatin chemosensitivity in ovarian cancer sphere cells.

To further demonstrate whether UA induces apoptosis of SKOV3 sphere cells, SKOV3 sphere cells were treated with control, UA, and combinations of UA (75 mg/l) with cisplatin (70 mg/l), respectively. Fig. 3A shows that UA obviously destructed the morphology of sphere cells. Real-time PCR showed that after treatment with UA and cisplatin, stem cell genetic marker mRNA expression quantity of SKOV3 sphere cells was reduced; stem cell marker mRNA expression of UA combined with cisplain group reduced more obviously (P<0.05 or P<0.01 or P<0.001) (Fig. 3B). To further quantify the apoptotic effects of treatment with UA and cisplatin, SKOV3 cells and sphere cells were stained with Annexin V-FITC and PI, and subsequently analyzed using flow cytometry for cell apoptosis. Consistent with growth inhibitory effects, UA combined with cisplatin caused a significant increase in the distribution of cells at the S phase in a dose-dependent manner. Besides evident S arrest, distinct G0–G1 peaks were observed in SKOV3 sphere cells after treatment (Fig. 3C and D). Proportions of Annexin V-stained cells were higher in cisplatin- and UA-treated cells than in the control SKOV3 sphere cells. Obvious increase in the number of apoptotic cells was detected for cells treated with cisplatin and UA compared to UA alone (Fig. 3E and F).

In our previous study, we reported that enrichment of ovarian CSCs is accompanied by EMT (9). To investigate the influence of UA and cisplatin on other mitogen-dependent processes, two assays were employed in the next step to compare the motility and invasion of the SKOV3 sphere cells with those of the negative control group. The EMT markers were also examined to demonstrate biological changes after treatment with UA and cisplatin. The Transwell invasion assay (Fig. 4B) showed that sphere cells had significant elevation in their invasive ability, and Transwell migration assay (Fig. 4A) was further used to assess their motility. Results from the migration assay (Fig. 4A) indicated that the migration of SKOV3 sphere cells treated with UA combined with cisplatin was significantly lower than their control counterparts 16 h after plating. The invasion assay (Fig. 4B) similarly showed reduced number of invaded cells in the UA combined with cisplatin group examined at 48 h. In summary, treatment with UA and cisplatin was associated with attenuation of the motility and invasion of SKOV3 sphere cells, in vitro. As shown in (Fig. 4C), mesenchymal markers of SKOV3 sphere cells such as Snail, Slug, Twist, vimentin, N-cadherin and fibronectin were expressed significantly higher than SKOV3 cells. Twist, vimentin, N-cadherin and fibronectin of UA alone group and UA combined with cisplatin group were significantly reduced compared to the sphere cells group, while epithelial markers CK19 and E-cadherin did not change significantly.

SKOV3 sphere cells were used to generate xenograft tumors in athymic nude BALB/c-nu mice to determine whether UA could strengthen the effects of chemotherapy in vivo. SKOV3 sphere cells (5×10⁶) formed tumors with a 13 day tumor latency (Fig. 5A). Representative hematoxylin and eosin staining of xenograft ovarian cancer of each group is shown in Fig. 5B. As expected, treatment with UA alone suppressed tumor growth compared with normal saline control. Tumors from mice treated with cisplatin in combination with UA were smaller at day 28 than those treated with UA alone. Furthermore, when measured both in tumor size and in tumor weight, the combined treatment of cisplatin and UA displayed a cancer prohibition effect when compared to the control group (Fig. 5C and D).

Immunohistochemical assays were further performed in tumors removed from the nude mouse xenograft model. In tumors treated with UA alone, Ki-67 and PCNA staining, respectively, showed moderate intensity with the scores of 1.6 and 1.4. As expected, mouse group treated with cisplatin in combination with UA had a much lower level of Ki-67 and PCNA staining (score 0.8 and 1.0). In the group treated with UA alone, almost all cancer cells displayed weak vimentin and fibronection staining with scores of 1.8 and 1.2, while treatment with cisplatin plus UA displayed scores of 0.8 and 0.4 (Fig. 6). These results demonstrated that UA downregulates the expression of Ki-67, PCNA, vimentin and fibronection of ovarian cancer in vivo.