Background
Ovarian cancer is the most common cause of death from gynecologic malignancies in the United States [
1]. Most women with epithelial ovarian cancers are diagnosed with advanced, metastatic disease characterized by widespread peritoneal carcinomatosis and abdominal ascites [
2]. The development of resistance to platinum chemotherapy (carboplatin and cisplatin) is common in advanced disease [
3]. Therefore, identifying new drugs to improve platinum response is critical to prolonging the life of women with refractory disease.
Thymoquinone (TQ) is a product of the medicinal plant
Nigella sativa which has promising anti-tumor efficacy in preclinical models of human cancer [
4‐
7]. Multiple molecular mechanisms of action have been described for the demonstrated ability of TQ to reduce tumor growth and survival in these preclinical studies. These include activation of tumor suppressor genes such as PTEN and p21, reducing pro-inflammatory and angiogenic signals via inhibition of NF-κB signaling, an important molecular link between inflammation and cancer [
8‐
13], and induction of DNA damage through generation of reactive oxygen species (ROS) [
4‐
6]. Early clinical trials have shown promising lack of toxic effects in patients with symptoms of cardiovascular disease such as hypertension and hypercholesterolemia [
6]. Only one Phase 1 trial has been reported for thymoquinone administration in 21 cancer patients, with no toxic or therapeutic effects detected over treatment times ranging from 1 to 20 weeks [
14]. Definitive trials for establishing safe and effective doses of TQ in cancer patients are currently lacking, but are well supported by preclinical data [
4‐
7].
Several mechanisms of resistance to platinum compounds in cancer cells have been identified [
15]. First, cisplatin treatment is known to induce NF-κB [
16], and NF-κB inhibitors potentiate the anti-tumor activity of various cytotoxic agents [
17]. Second, cisplatin induces double-strand DNA breaks by intercalating into DNA [
15], and its effects are reduced in ovarian cancer cells with intact DNA repair capacity [
18]. We have shown previously that drugs which promote DNA damage or inhibit DNA repair (e.g. histone deacetylase inhibitors) can sensitize ovarian cancer cells to cisplatin and DNA-damaging drugs [
19,
20]. Since TQ has multiple cellular effects that could potentiate cisplatin response, we hypothesized that TQ would sensitize ovarian cancer cells cultured
in vitro and in our syngeneic model to the cytotoxic effects of cisplatin.
Most preclinical models are limited by the fact that drug effects are tested on cancer cells in the absence of the supporting tumor microenvironment, essential for cancer progression in vivo. For this reason, we generated a mouse syngeneic model using ID8 mouse ovarian cancer cells grown intra-peritoneally in C57BL/6 mice [
21]. The cells have a stably integrated NF-κB reporter plasmid, allowing for quantification of tumor NF-κB activity in response to drug treatment during intraperitoneal abdominal carcinomatosis accompanied by ascites formation.
In this study, we show that combined TQ and cisplatin treatment induced synergistic anti-tumor effects in cultured ID8-NGL cells, and reduced tumor burden, proliferative and apoptotic markers in ID8-NGL-derived tumors. These combinatorial effects were associated with enhanced expression of the DNA damage marker, pH2AX(ser139), compared to either drug alone. Although TQ-mediated inhibition of NF-κB was observed in vitro, our syngeneic model showed an unexpected increase in tumor NF-κB activity and ascites volume with TQ treatment alone. These results emphasize the potential of targeting DNA damage as a therapeutic approach in ovarian cancer, but also that strongly caution TQ may have an overall deleterious effect through promotion of ascites formation. Since TQ is a likely candidate for future clinical trials in cancer patients [
5‐
7], this finding has considerable potential relevance to the clinic.
Materials and methods
Cell culture
Mouse ovarian cancer cells stably expressing a NF-κB reporter plasmid, ID8-NGL [
21], were cultured in 10 % FBS-supplemented DMEM High-Glucose medium with 400 μg/ml G418, and passaged by standard techniques. The human ovarian cancer cell lines, OVCAR3 and NCI/ADR-RES were cultured as previously described [
22]. Cultured ID8-NGL cells were treated with increasing concentrations of cisplatin (Sigma Chemical Co., Cat# 479306) and/or the NF-κB inhibitor, thymoquinone (TQ; Sigma Chemical Co., Cat# 274666).
Cell viability assays
Sulforhodamine B (SRB) assays were used to determine cell proliferation and cytotoxicity in response to TQ and/or cisplatin as previously described [
22]. Absorbance was measured at 510 nm using a Spectramax M5 spectrophotometer (Molecular Devices, Sunnyvale, CA) in the High-Throughput Screening Core of the Vanderbilt Institute of Chemical Biology. The interaction between fixed ratios of TQ and cisplatin was measured with the Combination Index (CI) method [
23]. A CI level of <0.9, CI = 0.9–1.1 and CI > 1.1 indicates synergy, additivity and antagonism respectively, between drug combinations.
Animal model and drug treatment
Wild-type C57BL/6 mice were injected intra-peritoneally (IP) with 1 × 10
7 ID8-NGL cells suspended in 200 μl sterile PBS [
21]. Thirty days after ID8-NGL injection, mice were randomized into the following treatment groups: vehicle (PBS), TQ (20 mg/kg thrice weekly), cisplatin (2 mg/kg weekly) and the TQ/cisplatin combination via the IP route for 30 days (
n = 5). No signs of drug toxicity were observed in the single or combination treatment mice. Tumor progression was monitored by body weight and abdominal girth measurements. At time of sacrifice, abdominal ascites fluid was extracted with hypodermic syringe, and volume measured. Tumor implants in the peritoneal wall and the mesentery were measured, then harvested and snap frozen or formalin-fixed for further analysis. The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committee at Vanderbilt University.
Luciferase assays
Luciferase activity was measured in harvested tumors following tissue homogenization in 1 ml reporter lysis buffer, and in whole cell protein extracts from cells grown in vitro, using the Promega Luciferase Assay system (Cat#4030). Activity was analyzed using a GloMax Luminometer (Promega, Madison, WI). Results were expressed as relative light units (RLU) normalized for protein content, as measured by the Bradford assay (Bio-Rad, Cat# 500-0002).
RNA extraction and quantitative RT-PCR (QPCR)
RNA from snap-frozen tumors was isolated using the RNeasy Mini kit (Qiagen, Valencia, CA) and QPCR performed as described [
21]. Steady-state mRNA levels of the established NF-κB targets, TNF-α and IL-1β, were expressed relative to corresponding GAPDH levels using the comparative 2
ΔΔCt method [
24]. Primers sequences used were as previously described [
21,
25,
26].
Immunofluorescence
Processing, embedding and sectioning of formalin-fixed ID8-NGL tumor tissue, and hematoxylin and eosin staining for histology, were performed in The Allergy/Pulmonary & Critical Care Med Division Immunohistochemistry Core at Vanderbilt [
27]. Immunofluorescence analysis of formalin-fixed paraffin-embedded tumor tissue or methanol-fixed cultured ID8-NGL cells was performed using standard techniques [
21,
19]. The following primary antibodies: mouse monoclonal anti-pH2AX(ser139) (EMD Millipore, Cat# 05-636, 1:250 dilution); rabbit polyclonal anti-Ki67/Mib-1 (Abcam, Cat# ab16667; 1:200 dilution); and rabbit polyclonal anti-cleaved caspase-3 (Cell Signaling Technology, Cat# 9661; 1:100 dilution), were used. Secondary antibody used was goat anti-rabbit Alexa Fluor 488 (Life Technologies, Cat# 11070) (all 1:200 dilution). Images were acquired and analyzed as previously described [
21,
19]. For quantifying the percentage of, where applicable, tumor cells or macrophages positive for these proteins, at least 5 independent fields were assessed with at least 200 cells counted per sample.
Western blotting
In ID8-NGL cells treated with TQ (25 μM) and/or cisplatin (1 μM), or in drug-treated tumors, whole cell protein isolation, subcellular fractionation, western blotting and signal detection were performed as described [
28,
29]. Primary antibodies used were rabbit polyclonal anti-PARP (Cell Signaling Technology; Cat# 9542; 1:1000 dilution), mouse monoclonal anti-PCNA (Santa Cruz, Cat# PC10, 1:1000 dilution), rabbit polyclonal anti-Bax (EMD Millipore, Cat# 06-499, 1:1000 dilution), and mouse monoclonal anti-pH2AX(ser139) (EMD Millipore, Cat# 05-636, 1:500 dilution). Mouse monoclonal anti-β-actin (Sigma Chemical Co., Cat# A5441 1:10000 dilution) and rabbit polyclonal anti-H2AX (EMD Millipore, Cat# 06-627, 1:500 dilution) were used as loading control, where appropriate.
Statistical analysis
Unless otherwise indicated, values shown for in vitro experiments were the mean + SE of 3 independent experiments, with comparison of groups performed by 2-tailed Student’s t test. Comparison of groups in in vivo experiments was performed by 2-tailed Mann–Whitney test. A p value <0.05 is considered statistically significant.
Discussion
Platinum resistance is common in ovarian cancer, and identifying novel drug combinations to enhance efficacy of platinum drugs such as carboplatin and cisplatin is a promising strategy. In this study, we tested the ability of TQ to sensitize ovarian cancer cells to cisplatin treatment in established preclinical models. First, we demonstrated that the combination of TQ and cisplatin had synergistic inhibitory effects on cell viability and survival in cultured human and mouse ovarian cancer cells. Second, compared to either drug alone, there were enhanced effects of the combination on established indices of tumor burden (peritoneal implants and mesenteric tumor mass), proliferation and apoptosis in a syngeneic mouse ovarian cancer model [
21].
TQ is known to have a myriad of cellular effects in tumor cells, including promotion of DNA damage through generation of reactive oxygen species and inhibition of NF-κB activity [
4‐
6]. Our results indicate that TQ was able to enhance DNA damage induced by cisplatin in both cultured cells and in tumors, consistent with previous studies from this laboratory showing that drugs which promote DNA damage sensitize ovarian cancer cells to the cisplatin-mediated cytotoxicity [
19,
20]. In contrast, TQ-mediated inhibition of NF-κB activity could not explain the enhancement of the cisplatin response in cultured cells and tumors. First, cisplatin only weakly induced NF-κB activity in vitro, an established mechanism of cisplatin-resistance [
16]. Second, in contrast to cultured cells in vitro, TQ induced tumor NF-κB reporter activity and mRNA levels of the key NF-κB targets, IL-1β and TNF-α, in drug-treated mice.
We acknowledge that the tumor microenvironment could impact drug response in our syngeneic model. For example, it is likely that the unexpected increase in ascites formation and activity of the NGL NF-κB reporter in tumors following the 30-day TQ treatment was mediated through drug effects on the tumor microenvironment. We have previously shown that TQ treatment for 10 days effectively reduces NGL reporter activity within tumors
in vivo [
21]. Furthermore, thrice-weekly TQ treatment results in persistent inhibition of NF-κB activity in cultured ID8-NGL cells in vitro for up to 21 days (unpublished observations). Identifying drug-induced alterations in specific inflammatory cell populations such as macrophages and dendritic cells, and possibly related mechanisms by which prolonged TQ exposure can induce NF-κB activity and ascites formation, will be the focus of future studies.
A possible microenvironmental role for TQ-induction of NF-κB activity and ascites following longer periods of treatment is supported by a recent study showing that the proteasome inhibitor, bortezomib, induces pro-inflammatory effects on the tumor microenvironment, particularly macrophages, subsequently promoting tumor progression in a mouse lung cancer model [
34]. Short-term exposure to bortezomib produces opposing effects resulting in inhibition of tumor cell growth [
34], similar to the anti-tumorigenic properties of 10 day TQ treatment [
21]. It is not known whether the deleterious effects of long-term TQ and bortezomib, both relatively non-specific inhibitors of NF-κB with additional cellular effects [
4‐
6,
34], are due to their actions on NF-κB activity in these preclinical models. However, because of the interest in using systemic NF-κB inhibitors as mono-therapy or in combination with other chemotherapeutic drugs in clinical trials in ovarian cancer patients [
35], possible side effects that limit drug efficacy, promote toxicity or have frankly deleterious effects are highly relevant clinically.
In conclusion, our results emphasize the strong potential of targeting DNA damage as a therapeutic approach to improve cisplatin response in ovarian cancer. However, the choice of drug to combine with cisplatin is critical, since we have shown that TQ, a very promising therapeutic agent in preclinical cancer models, had an unexpected deleterious effect through promotion of ascites formation in a syngeneic model. Since TQ is a potential candidate for future clinical trials in cancer patients [
5‐
7], this finding provides caution for its future use in patients with ovarian cancer.
Acknowledgments
We acknowledge The Vanderbilt Imaging Core, and The Allergy/Pulmonary & Critical Care Med Division Immunohistochemistry Core.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AJW oversaw the experiments, performed tumor cell injections into C57BL/6 mice, performed dissections and collection of ascites/peritoneal lavages, performed in vitro drug treatment experiments and western blots, analyzed the data, and drafted the manuscript. JS was responsible for mouse husbandry, maintained cell lines, processed tumor tissue for paraffin-embedding, processed ascites/peritoneal lavage fluid for RNA and protein extraction and preparation of cytospin slides, performed immunofluorescence staining in tumor and ascites samples, and prepared RNA for cDNA synthesis and QPCR analysis. WB was responsible for mouse husbandry, assisted in mice dissections and ascites processing, performed QPCR analysis, performed cytospin counts, and assisted in preparation of the manuscript. FY conceived the study, provided the C57BL/6 mice, consulted on experimental design and data analysis and shared final editorial oversight of the manuscript. DK conceived the study, consulted on experimental design and data analysis, and shared final editorial oversight of the manuscript. All of the authors have read and approved the final version.