The human ovarian cancer cell lines, SK-OV-3 and A2780 were cultured in RPMI-1640 (Gibco; Thermo Fisher Scientific, Inc.), and PA-1 were cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.), containing 10% (FBS) (Gibco; Thermo Fisher Scientific, Inc.), were purchased from American Type Culture Collection. All the cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. Emodin was purchased from Sigma-Aldrich (Merck KGaA), dissolved in dimethyl sulfoxide (DMSO) and stored at −20°C for further use. The antibodies against N-cadherin (cat. no. 13116; 1:1,000), E-cadherin (cat. no. 14472, dilution, 1:1,000), vimentin (cat. no. 5741; 1:1,000), CD133 (cat. no. 64326; 1:1,000), OCT4 (cat. no. 2840; 1:1,000) and β-actin (cat. no. 3700; 1:1,000) were purchased from Cell Signaling Technology, Inc.

Briefly, suspensions of the SK-OV-3, A2780 and PA-1 cell lines (2,000-5,000 cells/100 µl) were added to 96-well microplates. Approximately 12 h later, the cells were treated with different concentrations (0, 2.5, 5, 10, 20, 40 and 80 µM) of emodin for 24, 48 or 72 h. Then, the cell culture medium in each well was replaced with 20 µl MTT solution (5 mg/ml) and the samples were incubated at 37°C for a further 2-4 h. Subsequently, 150 µl DMSO was added to each well to dissolve the formazan crystal produced by the living cells. Finally, the optical density of each well was measured using a Spectra MAX M5 microplate spectrophotometer (Molecular Devices, LLC) at 570 nm. At least three independent experiments were performed.

The clonogenic ability of the ovarian cancer cell lines was assessed using the colony formation assay following treatment with emodin. Briefly, the ovarian cancer cell lines (400–600 cells/well) were seeded in 6-well plates and cultured for a further 12 h. The cells were treated to various doses of emodin (0, 5, 10, 20 and 40 µM) and cultured for a further 12 days. The cell culture medium was replaced with fresh medium containing corresponding concentrations of emodin every 3 days. Subsequently, the culture medium was discarded, and the cell colonies were washed three times with PBS, fixed with methanol for 15 min at 25°C and stained with 0.5% crystal violet for 20 min at 25°C. Finally, the number of cell colonies (>30 cells) was counted manually and images were captured using an inverted microscope (Axiovert 200; Zeiss AG).

The ovarian cancer cell lines were seeded in 6-well plates and cultured for 12 h. When the monolayer of the cells reached 80% confluence, the wound was created using a sterile 200 µl micropipette tip. Then, the cells were washed with fresh medium and cultured in medium (with 0.5% FBS) and different concentrations of emodin (0, 10, 20 and 40 µM) for 48 h. Next, the cell culture medium was discarded, the cells were washed three times with PBS, then images were captured using a light microscope (Axiovert 200; Zeiss AG). The migration rates of the treated cells were quantified using the following equation: Migration rate=(1-W₀/W_emodin)/(1-W₀/W_ctrl), where W₀ represents the width of the wound at 0 h; W_ctrl represents the width of the wound at 48 h in the control group and W_emodin represents the width of the wound at 48 h in emodin treatment group. The migration rate of the control cells was regarded as 100%.

For the analysis of the migratory effect of emodin, 1×10⁵ SK-OV-3, A2780 and PA-1 cells, suspended in 100 µl serum-free medium, were added to the upper chamber, while 600 µl complete medium, containing 10% FBS, was added to the bottom chamber. Different concentrations of emodin (0, 10, 20 and 40 µM) were added to medium in the upper chambers. After migration for ~48 h at 25°C, non-migratory cells remaining in the upper chamber were discarded using a cotton swab and the migrated cells located on the bottom of membrane were washed, fixed with methanol for 15 min at 25°C, then stained with 0.5% crystal violet for 20 min at 25°C. Images of the migrated cells were captured using a light microscope and five random fields of view were counted to compare the migration rate.

A Matrigel assay was performed to determine the invasion ability of the cells treated with emodin. Briefly, the upper surface of a 24-well Transwell plate (MilliporeSigma) was pre-coated with ~70 µl Matrigel (BD Biosciences) diluted (1:4) in serum-free cell culture medium. After Matrigel polymerization for 1 h at 37°C, 100 µl serum-free medium, containing 1×10⁵ SK-OV-3, A2780 and PA-1 cells was added to upper chamber, while 600 µl the complete medium was added to the lower chamber, which served as a chemoattractant. Different concentrations of emodin (0, 10, 20 and 40 µM) were added to both the upper and lower chambers. After invasion for 48 h, the non-invasive cells remaining in the upper chamber were discarded using a cotton swab and the invasive cells on the bottom chamber were washed, fixed with methanol for 15 min at 25°C, then stained with 0.5% crystal violet or 20 min at 25°C. The images of the invasive cells were captured using a light microscope and five random fields of view were selected to calculate the invasive rate.

To investigate the effect of emodin on relevant signaling pathways, the changes in the protein expression level of key proteins in the SK-OV-3, A2780 and PA-1 cell lines were evaluated using western blot analysis. After treatment with the indicated concentrations of emodin (0, 10, 20 and 40 µM) for 48 h, the SK-OV-3, A2780 and PA-1 cell lines were harvested and lysed with RIPA buffer (Beyotime Institute of Biotechnology) to obtain the total protein. Concentration of total protein was determined by BCA method. An equal amount of protein (~30 ug), from the differently treated samples were separated using SDS-PAGE (10%), then transferred onto PVDF membranes (Amersham Bioscience; Cytiva). The membranes were blocked with skimmed milk (5% in TBST buffer (Tween-20, 0.5%) for 1 h at 37°C and incubated with the primary antibodies overnight at 4°C, then incubated with the corresponding secondary antibodies (horseradish peroxidase-conjugated goat Anti-Rabbit IgG H&L; cat. No.: ab205718, dilution: 1:10,000, Abcam) at 37°C for 2 h. Immunoreactive protein bands were visualized and detected using an enhanced chemiluminescence kit (MilliporeSigma). The β-actin antibody was used as a loading control. Protein expression levels were determined by ImageJ (v 1.8.0, National Institutes of Health).

All the animal experiments were approved by the Institutional Animal Care and Treatment Committee of Gannan Medical University (Jiangxi, China) and were conducted according to the approved guidelines. All mice (Balb/c nude; female; 6-8 weeks old; weight, 18-20 g) used in the experiments were purchased from Beijing HFK Bioscience Co. Ltd. The mice were kept in a specific-pathogen-free condition facility, in an air-conditioned room at 25±2°C, with a relative humidity of 40-70%, and a 12-h light/dark cycle, with free access to food and water. A total of 18 female mice (6–8 weeks old; weight, 18-20 g) were subcutaneously injected with 100 µl suspension (normal saline) of SK-OV-3 cells (0.5×10⁷ cells per mouse). Several days later, the injected site (right flank) a successful mouse tumor model was established if there were noticeable signs of growth. When the tumor volume reached ~100 mm³, the mice were randomly divided into three groups (vehicle, and 20 and 40 mg/kg emodin; n=6). The mice were intraperitoneally administrated with 20 and 40 mg/kg emodin or vehicle once daily. The tumor volume and body weight of the mice was measured and recorded every 3 days during the treatment process, and the tumor volume was calculated as follows: Volume (mm³)=LxW²/2, where L (mm) represents the length of the tumor and W (mm) represents the width of the tumor. All the mice were sacrificed with CO₂ (70% volume displacement rate) at the endpoint of the experiment, which was defined by the tumor volume (~1,500 mm³) in the control group. Tumor inhibitory rate (%)=(V(emodin)-V(vehicle))/V(vehicle), where V represents tumor volume. The tumors from each group were collected, weighted and fixed with 4% paraformaldehyde for 48 h at 25°C for further evaluation. In addition, the major organs (heart, liver, spleen, the lungs and the kidneys) from the mice in each group were collected, fixed in 4% paraformaldehyde for 24 h at 25°C, then sectioned (3–5 µm) for hematoxylin and eosin staining. The sections were stained with hematoxylin (1%) and eosin (1%) at 25°C for 2-5 min. Finally, sections were photographed by light microscope.

The collected tumor tissue samples were fixed in 4% paraformaldehyde for 24 h at 25°C, embedded in paraffin and sliced into thin sections (3–5 µm). After dewaxing and rehydration, the sections were incubated with 3% hydrogen peroxide at 25°C for 30 min to block the endogenous peroxidase activity, then treated with 5% bovine serum albumin (Thermo Fisher Scientific, Inc.) at 37°C for 30 min to block non-specific binding. Subsequently, the sections were incubated with primary antibodies against Ki67 (cat. no. ab15580: 1:500), CD133 (cat. no. ab278053; 1:1,000), and cleaved caspase3 (cat. no. ab32042; 1:1,000; all Abcam) overnight at 4°C. The sections were washed with PBS 3 or 4 times, incubated with the biotinylated secondary antibody (Abcam, ab64256, dilution: 1:1,000) at 37°C for 1 h, then treated with streptavidin horseradish peroxidase at 37°C for 20-30 min. Images were captured under a light microscope (Axiovert 200; Zeiss AG). Positive stained cells were calculated by Image J (V1.8.0, National Institutes of Health) from the average of four random fields of view.

A total of 18 mice were intravenously injected with a 100 µl suspension SK-OV-3 cells in saline (1×10⁶ cells per mouse) via the tail vein to produce an experimental lung metastasis tumor mouse model. The mice were randomly divided into three groups (vehicle, and 20 and 40 mg/kg emodin groups; n=6) 2 days following injection and were intraperitoneally injected with 20 or 40 mg/kg emodin, or vehicle once daily. The mice from the experimental and control groups were euthanized on day 18, and their lung tissues were isolated, weighted and collected for observation of visible metastatic nodules. The dots on the lung surface were counted manually and confirmed as ovarian cancer metastases.

At the study endpoint, the lung tissues from each group were collected and prepared as single-cell suspensions using mechanic dispersion with surgical scissors and enzymatic method of I-collagenase (1 mg/ml). Then, 1×10⁶ freshly prepared cells were suspended in 100 µl PBS and stained with different combinations of FITC-CD24 (Biolegend, cat. no. 311103) and APC-CD44 (Biolegend, cat. no. 103011). The cells were analyzed using flow cytometry (FACS Canto II, BD Biosciences) and the data were analyzed using FlowJo software (FlowJo LLC, V7.6.1). The proportions of CD44⁺/CD24⁻ cells were considered to be ovarian CSCs.

The data are presented as the mean ± standard deviation of at least of three independent experiments. SPSS v16.0 (SPSS, Inc.) software was used to perform the statistical analyses. One-way ANOVA followed by Dunnett's or Tukey's post hoc test was used for multi-group comparisons. P<0.05 was considered to indicate a statistically significant difference.

The SK-OV-3, A2780 and PA-1 cell lines were used to evaluate the antitumor effect of emodin on ovarian cancer. Firstly, the time- and dose-dependent effect of emodin on these cells was performed by treating the cells with different concentrations of emodin (0, 2.5, 5, 10, 20, 40 and 80 µM) for various time points (24, 48 and 72 h). As shown in Fig. 1A, the SK-OV-3, A2780, and PA-1 cell lines treated with 10-20 µM emodin for 24 h exhibited weakened viability (P<0.05). However, after treatment for 48 and 72 h, cell viability was significantly (P<0.01) suppressed. In addition, treatment with 40 and 80 µM emodin significantly reduced the viability of the SK-OV-3, A2780 and PA-1 cell lines at 24, 48 and 72 h. This suggests that emodin exhibited a time- and dose-dependent (>2.5 µM emodin) effect in ovarian cancer cell lines. To further investigate whether emodin could suppress the proliferation ability of the ovarian cancer cell lines, a colony formation assay was performed after the cells were treated with emodin. As shown in Fig. 1B, the number of the colonies in the SK-OV-3, A2780 and PA-1 cell lines treated with emodin was significantly reduced in a dose-dependent manner. In addition, emodin significantly inhibited the number of colonies in the SK-OV-3, A2780 and PA-1 cell lines in comparison to the control group. Taken together, these results suggest that emodin significantly inhibited cell viability and colony formation in the ovarian cancer cell lines.

A wound healing assay was performed using the SK-OV-3, A2780 and PA-1 cell lines to investigate the in vitro anti-migration ability of emodin in ovarian cancer cells. As displayed in Fig. 2A and B, emodin significantly inhibited the wound healing rate of the SK-OV-3 and A2780 cell lines in a dose-dependent manner, suggesting an anti-migratory effect of emodin in ovarian cancer cells. In addition, Transwell and Matrigel assays were performed to further investigate the anti-migratory and anti-invasive abilities of emodin in ovarian cancer cells. The results demonstrated that the migration abilities of the SK-OV-3, A2780 and PA-1 cell lines were significantly suppressed in the presence of emodin compared with that in the control group (Fig. 2C and D). As indicated in Fig. 2E and F, treatment with different concentrations of emodin distinctly inhibited the invasive abilities of the SK-OV-3, A2780 and PA-1 cell lines. Furthermore, as there is an association between cancer cell migration and invasion, and EMT (35), it was investigated whether emodin inhibited the migration and invasion abilities of the ovarian cancer cells via EMT. As demonstrated in Fig. 2G-I, the protein expression levels of N-cadherin and vimentin were decreased in the SK-OV-3, A2780 and PA-1 cell lines following treatment with emodin, whereas the protein expression levels of E-cadherin were increased, indicating EMT was affected in the ovarian cancer cell lines. Taken together, these results suggest that emodin suppressed the migration and invasion abilities of the ovarian cancer cells by affecting EMT.

To investigate whether the antitumor activity of emodin in vivo was consistent with its anti-proliferation effect in vitro, SK-OV-3 tumor-bearing mice were treated with different doses of emodin (20 and 40 mg/kg) for 18 days following injection of the cells. As shown in Fig. 3A, tumor growth was significantly reduced following treatment with 20 and 40 mg/kg emodin compared with that in the control group. Mice were sacrificed at the study endpoint, and tumors from each group were isolated. The tumor size and average weight in the 20 and 40 mg/kg treatment groups were notably smaller compared with that in the control group (Fig. 3B and C). Furthermore, no adverse effects, such as toxic death, skin ulceration and body weight loss were observed, during the treatment of the SK-OV-3 tumor-bearing mice with emodin. The body weight of the mice was not significantly different between the three treatment groups (Fig. 3D). In addition, after pathological evaluation of the major organs (heart, liver, spleen, the lungs and the kidneys), no notable pathological changes were found in the emodin treatment groups, suggesting that emodin is safe to use in vivo (Fig. 3E).

The antitumor effect of emodin was also analyzed using IHC from the tumor sections of each group. As displayed in Fig. 4A and B, increased expression of cleaved-caspase 3 and decreased number of Ki-67 positive cells was observed in the emodin treatment groups. Furthermore, compared with that in the control group, CD133, a key marker of ovarian CSCs (36), was significantly downregulated, suggesting a reduction in the development of ovarian CSCs following treatment with emodin.

Subsequently, western blot analysis was used to verify the anti-CSC effect of emodin. As demonstrated in Fig. 4C and D, the protein expression level of E-cadherin was increased following treatment with emodin, whereas the expression level of N-cadherin was decreased, suggesting that emodin inhibited EMT. In addition, the stemness of the tumors in the emodin treatment groups was significantly suppressed, as evidenced by the decrease in protein expression level of OCT4, which has been validated as a master regulator in the maintenance of the cancer stem-like phenotype (37).

A pulmonary metastasis mouse model was established using SK-OV-3 cells to evaluate the anti-metastasis properties of emodin. Following treatment with emodin for 18 days, the mice were sacrificed, the weight of the lung tissues was measured, and the number of metastatic lung nodules were counted. As indicated in Fig. 5A and B, compared with that in the control group, the number of metastatic nodules on the lung was markedly reduced in the emodin treatment groups. In addition, the weight of the lungs in the emodin treatment groups was significantly inhibited compared with that in the control group.

It is reported that poor survival and distant metastasis in patients with ovarian cancer are caused by the renewal of CSCs, and increased CSCs have been associated with tumor recurrence or relapse, distant metastasis and chemoresistance (38,39). As aforementioned, the protein expression of the CSC markers, CD133 and OCT4 was significantly decreased in the emodin treatment groups. Various studies have indicated that the CD44⁺/CD24⁻ population may represent stem cell-like properties of ovarian cancer cells (40–42). To further evaluate whether emodin inhibited lung metastasis of ovarian cancer cells by destruction of the ovarian CSCs, flow cytometry was used to measure the proportion of ovarian CSCs (CD44⁺/CD24⁻) in metastatic lung tissues following emodin treatment for 18 days. As shown in Fig. 5D and E, the percentage of CD44⁺/CD24⁻ cells was 32.65±2.95% in the control group, while treatment with 20 and 40 mg/kg emodin significantly reduced the proportion of CD44⁺/CD24⁻ cells to 27.82±3.21 and 17.89±2.76%, respectively (P<0.001), indicating that emodin reduces pulmonary metastasis of ovarian cancer cells by killing cancer stem cells.