Abstract
Background/Aim: Advanced ovarian clear-cell carcinoma (CCC) fails to respond to standard chemotherapy, and has a poor prognosis. Since hypoxia-inducible factor-1 (HIF-1) stimulates various genes involved in cancer, we aimed to examine the efficacy of silibinin, an active component of milk thistle belonging to Asteraceae, in suppressing HIF-1 activity, and elucidate the underlying mechanism in human CCC cell lines. Materials and Methods: Human ovarian CCC cell lines HAC-2, OVISE, and RMG-1 were treated with 500 μM silibinin for 4 h under normoxic and hypoxic conditions. Using DNA microarray, we analysed genes whose expression modulated more than 2-fold in response to hypoxia, whereas HIF-1α expression was measured using ELISA. Results: Silibinin treatment decreased HIF-1α protein in all cell lines, and eIF4E2 and RPS6 mRNA in HAC-2 and RMG-1 cells. Conclusion: Silibinin suppressed HIF-1α protein under hypoxic conditions in CCC cell lines and could be a potential anti-cancer drug.
Clear-cell carcinoma (CCC) is the second most common histological type of cancer in Japan, accounting for 30% of all ovarian carcinomas (1, 2). In comparison with serous carcinoma (3), which is the most frequent carcinoma type, advanced cases of CCC have an unfavourable prognosis due to poor response to standard anti-cancer chemotherapy comprising of paclitaxel and carboplatin (4).
Currently, novel strategies, using poly ADP-ribose polymerase (PARP) inhibitor and immune-checkpoint inhibitors, are being gradually introduced (5, 6). However, these agents are mainly targeted towards serous carcinoma, instead of CCC. Hypoxia is a classical condition accompanying cancer growth. In hypoxic environments, hypoxia-inducible factor-1 (HIF-1) plays a pivotal role as a transcription factor as well as a putative anti-cancer therapeutic agent (7-12). However, down-regulation of HIF-1 is challenging, since it is also essential for maintaining systemic homeostasis. Previously, everolimus, a mammalian target of rapamycin (mTOR) inhibitor, was assessed as a therapeutic agent for ovarian cancer in a phase II trial. Currently, everolimus is applicable as a therapeutic alternative in renal cancer (13, 14), neuroendocrine neoplasms (15-17), and breast cancer (18-21); however, its applicability in ovarian cancer remains unknown (3, 22).
Silibinin is a component of silymarin, a standardised extract of the milk thistle seeds (23-25). It was originally used in European folk medicine to treat liver dysfunction; it also exhibited anti-cancer activity in the ovarian endometrioid and serous carcinoma cell lines via extracellular signal-regulated kinase and Akt (26). Studies have suggested that silibinin has multiple targets in the cell (27, 28). In the SEC cell line, the combined use of silibinin and paclitaxel showed significant suppression of cell proliferation and significant up-regulation of tumour suppressor genes P53 and P21, and is more effective in treating ovarian cancer. This was verified in a study on drug resistance caused by long-term use of paclitaxel (29). A study showed that cisplatin and taxol could restore the sensitivity of the combinatorial administration of cisplatin and taxol in drug-resistant human ovarian cancer cells and reduce drug-induced hepatotoxicity in drug-resistant ovarian endometrioid carcinoma cell lines (30). Thus, silibinin may be employed as a part of combinatorial therapies against ovarian cancer.
Hence, our study aimed to explore the effect of silibinin on suppressing CCC growth, and evaluate its efficacy as a potent anti-cancer drug.
Materials and Methods
Cell lines and reagents. The human ovarian CCC cell lines HAC-2, OVISE, and RMG-1 were kindly provided by Professor Kazunari Kiguchi, Department of Obstetrics and Gynaecology, St. Marianna University School of Medicine (Kanagawa, Japan). The cells were cultured in RPMI-1640 medium (R8758, Sigma-Aldrich, Tokyo, Japan) supplemented with 10% foetal bovine serum (FBS) and 1% penicillin-streptomycin. HeLa, a human cervical adenosquamous cell carcinoma cell line, was provided by the Support Centre for Medical Research and Education, Isehara Research Promotion Division, Tokai University. HeLa cells were cultured in minimum essential medium (M4655, Sigma-Aldrich) supplemented with 10% FBS and 1% penicillin-streptomycin. All cells were cultured at 37°C at 5% CO2 in a humidified atmosphere.
Silibinin (1006211), purchased from Cayman Chemical (Ann Arbor, MI, USA), was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 1.0×106 μM to make the stock solution. Deferoxamine mesylate (DFO) (5764) was purchased from TOCRIS (Bristol, England), and dissolved in DMSO to a concentration of 100 mM to prepare the stock solution.
Lactate dehydrogenase (LDH) activity-based cytotoxicity assay. To evaluate the cytotoxicity of silibinin, LDH released from the lysed cells was quantitatively measured using the LDH Cytotoxicity Assay Kit (601170/Cayman Chemical) according to the manufacturer’s instructions. All the cell lines were treated with 500 μM silibinin, followed by 4 h incubation at 37°C in a 5% CO2 incubator. The value of LDH obtained by subtracting the blank (medium only) was used for the calculation: (%)=experimental LDH release/maximum LDH release ×100 (maximum LDH release: 45 min before the end of culture, 0.1% Triton X-100 (X100PC-5X5ML, Sigma-Aldrich).
HIF-1α cell-based ELISA. All the cell lines were treated with 500 μM silibinin, and cultured in a humidified atmosphere at 5% CO2 (normal atmospheric condition) for 0 h or 4 h, as well as hypoxic conditions (2% oxygen concentration). As a positive control, the cells were cultured for 4 h in a culture medium containing 200 μM DFO to induce HIF-1α expression. After culture, the HIF-1α expression level was compared by luminescence intensity (RFU), using the HIF-1α Cell-Based ELISA Kit (CBA-281, CELL BIOLABS, San Diego, CA, USA), according to the assay protocol.
RNA extraction and quantification. Total cellular RNA was isolated from the cells using the RNeasy Mini RNA isolation kit (74104, Qiagen, Hilden, Germany). Total RNA was eluted from RNeasy Mini columns using 30 μL RNase-free water. RNA quantity was estimated using a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific NanoDrop Technologies, Wilmington, DE, USA), and RNA integrity was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).
DNA microarray analysis. Total RNA from the 500 μM silibinin-treated and untreated (only DMSO-treated) HAC-2 and RMG-1 cells was extracted as mentioned above. Genes that showed more than two-fold fluctuation after silibinin treatment, as assessed by performing SurePrint G3 Human GE Microarray 8x60K Ver. 3.0 (G4858A # 72363, Agilent Technologies), compared to the untreated group, were considered to be modulated. In the present study, we emphasised on the HIF-related factors.
Statistical analysis. Statistical analyses of the data were performed using IBM SPSS Statistics version 25. Data are presented as the mean±SD from three independent experiments, except for DNA microarray, in which two independent experiments were performed. Data were analysed by the Student’s t-test or one-way analysis of variance (ANOVA), and results with p<0.05 were considered statistically significant.
Results
LDH release ratio. The LDH release ratios of HAC-2 and RMG-1 cells were 10% or less of those of the control (untreated cells). However, the amount of LDH released was significantly increased (p=0.0029) in the HeLa cells of silibinin-treated group (Figure 1).
Evaluation of silibinin cytotoxicity. Each cell line was treated with silibinin (500 μM) and incubated for 4 h under normal atmospheric conditions and hypoxic condition, at oxygen concentration of 2%; the cell-culture supernatant was used to measure the LDH concentration. In HAC-2, OVISE, and RMG-1 cells, LDH release was approximately 10% of that released by the control; however, in HeLa cells, LDH release was significantly higher in the silibinin-treated group. *1; p=0.0004, *2; p=0.0029.
HIF-1α protein expression. HIF-1α protein expression was induced under hypoxic conditions (Figure 2); however, its level did not increase in the untreated HAC-2 cells. In HeLa cells, hypoxia induced HIF-1α protein expression at the same level as by DFO stimulation (data not shown), but the induction of HIF-1α expression in HAC-2 and OVISE cells was mild under 4-h hypoxic conditions. In RMG-1 cells, the HIF-1α expression level was not affected by the 4-h hypoxic condition. However, HIF-1α protein level was significantly reduced in HAC-2 and OVISE cells treated with silibinin under 4-h hypoxic conditions.
HIF-1α protein expression in silibinin-treated cells. In RMG-1 cells, the expression level of HIF-1α protein was significantly down-regulated by silibinin treatment regardless of oxygen concentration. In HAC-2, OVISE, and HeLa cells, HIF-1α protein levels were significantly decreased by silibinin treatment under hypoxic conditions. *1; p=0.0018, *2; p=0.01, *3; p=0.0012, *4; p=0.00004, *5; p=0.0218.
DNA microarray analysis. Among the three cell lines, HAC-2 showed the highest basal level of HIF-1α protein expression; in RMG-1 cells, HIF-1α protein was significantly reduced in response to silibinin treatment, regardless of oxygenation conditions (Figure 2). Therefore, these cell lines were used for DNA microarray analysis to identify genes exhibiting ≥2-fold change in expression between the untreated/control group and the silibinin-treated group (Figure 3). The number of genes with more than 2-fold increase was 5111 in HAC-2, 8064 in RMG-1, and 2673 in both the cell lines (Figure 4A). Conversely, the number of genes with more than 2-fold decrease was 2600 in HAC-2, 1318 in RMG-1, and 883 in both the cell lines (Figure 4B).
Scatter plot of microarray data. A: HAC-2 cells; B: RMG-1 cells. Compared to that in the untreated group, the green line showed more than a 2-fold change in expression in response to silibinin treatment.
Number of genes with more than 2-fold change (FC) in expression as per microarray data. A: The number of genes with more than 2-fold increase in expression owing to silibinin treatment was 5111 in HAC-2, 8064 in RMG-1, and 2673 in both the cell lines. B: The number of genes that showed more than 2-fold decrease in expression owing to silibinin treatment was 2600 in HAC-2, 1318 in RMG-1, and 833 in both cell lines.
Interestingly, we observed that the expression of eukaryotic translation initiation factor 4E (eIF4E) and ribosomal protein S6e (RPS6), which promote HIF-1α mRNA translation, was significantly reduced in both cell types. In the silibinin-treated group, the expression of HIF-1α and other HIF-1-regulated genes was not altered by more than 2-fold (Figures 5 and 6). Furthermore, the expression of human PIK3R1 encoding phosphatidylinositol 3-kinase regulatory subunit alpha was reduced by approximately 2.8-fold. The HIF-1α-related genes that showed more than 2-fold increase in expression in response to silibinin treatment, were EPO, ERBB2, HK3, PDHA1, and RPS6KB1 (Figure 6).
Colouring of pathways by KEGG Mapper – Colour Pathway (HIF-1 signalling pathway). A: HAC-2 cells, B: RMG-1 cells. For the HIF-1 signalling pathway, the expression ratio of each gene in the microarray is displayed in different colours.
HIF-1 signalling pathway-related genes that demonstrated more than 2-fold change in HAC-2 and RMG-1 cells. HAC-2 and RMG-1 cells shared 11 HIF-1 signalling pathway-related genes with more than a 2-fold change in expression upon silibinin treatment.
Discussion
Currently, approximately 70% of patients with ovarian cancer suffer from relapses after undergoing TC (Paclitaxel and Carboplatin) therapy, the proportion being even higher for patients at stages III and IV (31-33). In a previous study on epithelial ovarian cancer, we reported that HIF-1α expression and HIF-1 activation was significantly up-regulated in CCC (34). Therefore, this experimental study was based on the hypothesis that ovarian CCC may be suppressed by silibinin via its anti-cancer effect on HIF-1.
The amount of LDH released in response to silibinin treatment was 10% or less of that released by the control in all target cell lines, indicating that silibinin exerted negligible adverse effect on HIF-1α expression (Figure 1). Since the gene expression varies remarkably among the cell types, the function of silibinin may be influenced by the properties related to each cell type. Silibinin reportedly inhibits the Warburg effect in cancer cells through multiple mechanisms, including inhibiting HIF-1 activation (35, 36). It has also been reported to inhibit methionine and restore sensitivity to cisplatin and taxol in endometrioid cell lines, besides reducing drug-induced hepatotoxicity (37). Notably, its combination with paclitaxel to target incurable ovarian endometrioid carcinoma improves the outcome of primary ovarian cancer (38); a phase II clinical trial on prostate cancer was also recently conducted (39, 40).
Treatment with silibinin significantly attenuated the expression of eIF4E and RPS6 (Figures 3, 4, 5 and 6). EIF4E is a cap-binding protein that recognises the cap structure (5-cap’) at the 5’-end of mRNA. EIF4E binds to the cap structure along with the helicase eIF4A, and the large scaffolding protein eIF4G, thereby forming the ternary complex eIF4F to initiate cap-dependent translation (41). Since cap-dependent translation is involved in cell proliferation (26, 37, 38), its dysfunction would cause cancer cell development. In fact, several cancer cells overexpress eIF4E (26, 40), which in turn induces cancer cell development (34, 39). Phosphorylation of eIF4E-binding proteins (4E-BPs) by the PI3K/Akt/mTOR pathway, an HIF-1 activation pathway independent of oxygen concentration, hampers the eIF4E binding to 4E-BPs, and therefore, eIF4E binds to 4eIF4G, which is a scaffolding protein that bridges between eIF4E and 4E-BPs. Formation of eIF4F complex activates the cap-dependent translation (41). Down-regulation of these two genes by silibinin treatment can suppress the cap-dependent translation and ribosome biosynthesis, thereby exerting anti-cancer effects (Figure 5). A slight decrease in the PIK3R1 expression level was also observed (Figure 6). This gene codes for one of the regulatory subunits of class IA PI3K, and mutations in this gene play roles in various types of human solid tumours, including breast cancer as shown earlier (42).
Although no change was observed in the HIF-1α expression level (Figure 5), the amount of HIF-1α protein was significantly decreased (Figure 2). Moreover, HIF-1α synthesis was considerably suppressed despite HIF-1α overexpression. Silibinin is anticipated to be highly efficacious in suppressing HIF-1. Akt, as a factor influencing silibinin in vitro, is involved in cancers of breast, cervix, colon, lungs, ovary (serous carcinoma), and skin (23). The role on HIF-1α has been reported in cancers of cervix, colon, lungs, and skin (32); levels of phosphorylated 4E-BP1 and phosphorylated p70S6K proteins are reduced in cervical cancer and liver cancer cells (23). Thus, although the effect of silibinin on the mTOR pathway has been reported in various cancer types, the suppression of eIF4E and RPS6 mediated by silibinin in ovarian CCC lines has not been reported.
In this study, HIF-1 suppression required higher doses of silibinin in ovarian CCC than in serous carcinoma (data not shown). Even at a high dose, silibinin may be used in combination with current chemotherapy and molecular-targeted therapies. Previous in vitro studies on ovarian cancer reported that co-administration of paclitaxel with silibinin promoted apoptosis and reduced invasion; however, it targeted serous carcinoma cell lines (32, 33). Since silibinin can significantly mitigate the adverse effects compared to the current anti-cancer drugs, its long-term application may significantly contribute to the improved quality of life in patients with CCC. In the near future, we intend to conduct an in vivo experimental study on ovarian CCC to determine the appropriate silibinin concentration in combination with the existing chemotherapy regimen to achieve maximal beneficial effect. The effect of silibinin on HIF-1 degradation also remains to be investigated.
HIF-1 is an essential molecule that plays a crucial role in the maintenance of homeostasis in the body. Hence, for the clinical application of silibinin in anti-cancer therapy of ovarian CCC, an appropriate drug delivery system that does not affect cellular homeostasis needs to be established. Our study may contribute remarkably to develop novel treatment strategies for treating CCC, which has a high incidence in the Japanese population, low sensitivity to anti-cancer agents, and poor prognosis. Although the effects of silibinin on HIF-1 activity were not investigated in this study, and our results remain to be validated in future studies, we believe that silibinin is a potential candidate for use in combination therapy and postoperative chemotherapy to cure ovarian CCC.
Acknowledgements
Some analyses were performed by intégrale Co., Ltd. (Tokushima, Japan), and Macrophi Inc. (Kagawa, Japan), and some were performed and supported by Support Centre for Medical Research and Education, Tokai University. This study was supported in part by 2017 Tokai University School of Medicine Research Aid. We would like to thank Editage (www.editage.com) for English language editing.
Footnotes
Authors’ Contributions
MM (Mariko Miyazawa) and MM (Masaki Miyazawa) performed molecular biological experiments, analyzed and evaluated the data. MM (Mariko Miyazawa) drafted the manuscript, MY designed and supervised the experiments, assisted with the writing of the manuscript. NO and TK interpreted the data and contributed advice about clinicopathological point of view. MY, TH, MM (Mikio Mikami) and HI interpreted the data an contributed advice in terms of treatment in obstetrics and gynecology. All Authors read and approved the final manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received October 19, 2020.
- Revision received November 9, 2020.
- Accepted November 13, 2020.
- Copyright © 2020 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
