Background
Ovarian cancer continues to be a major worldwide gynecological malignancy. Approximately 25,000 new cases are diagnosed each year in the USA, and 15,000 patients die of this malignancy [
1]. Currently, no sufficiently accurate screening tests to diagnose this malignancy are available. Consequently, it is detected only in its late stages leading to minimal survival rates after diagnosis. At stage III, ovarian cancer metastasizes and spreads to the surrounding organs such as the peritoneum and stomach. By stage IV, ovarian cancer spreads to distant metastatic organs such as the lungs and liver. Cisplatin is a well established platinum drug used to treat various cancers, including ovarian cancer [
2,
3]. Patients treated with cisplatin often relapse or do not respond to the treatment. In addition, at higher doses cisplatin exerts side effects such as nephrotoxicity and ototoxicity in patients [
4]. Several reports suggest that signal transducer and activator of transcription 3 (STAT3) overexpression is positively associated with cisplatin resistance [
5].
The STATs are a novel class of transcription factors that are positively associated with the growth and survival of cells [
6]. STAT3 is a receptor tyrosine kinase that is activated either by upstream receptor kinases such as Janus activated kinases (JAKs) or cytokines such as interleukin (IL)-6 [
7]. When IL-6 binds to its receptors, it activates STAT3 by phosphorylating it at Tyr-705. Activation of STAT3 at Tyr-705 leads to formation of a homodimer that translocates to the nucleus, where it binds to the promoter regions of several genes that transactivate STAT3-responsive genes such as Mcl-1, survivin and cyclin D1 [
8‐
10]. It is also phosphorylated at Ser-727, which is not required for DNA binding activity but is important for its maximal transcriptional activity. STAT3 activates vascular endothelial growth factor (VEGF), thereby promoting neovascularization in tumors [
11]. It also regulates hypoxia-inducible factor 1α (HIF-1α) and vascular epithelial growth factor (VEGF) during hypoxia, leading to hypoxia-induced angiogenesis [
12,
13].
Previously published reports suggest that STAT3 is overexpressed in various tumors, including ovarian tumors [
10]. A recent clinical study scored 322 patients for overexpression of phosphorylated (p)-STAT3 and observed that 303 patients were positive for hyperactivation of STAT3, accounting for 94% of the study group [
14]. Furthermore, various reports indicate the role of STAT3 in resistance of ovarian cancer to chemotherapy [
5]. Since STAT3 is involved in various aspects of cancer growth ranging from tumor initiation, angiogenesis, and metastasis, it represents an attractive target for intervention.
3,3'-Diindolylmethane (DIM), an active metabolite of indole-3-carbinol, is present in cruciferous vegetables [
15]. Accumulating epidemiological evidence indicates an inverse relationship between intake of cruciferous vegetables and the risk of ovarian cancer [
16]. Several studies, including those from our laboratory, have suggested that DIM possesses chemopreventive and therapeutic properties [
17‐
19]. Moreover, DIM was shown to be non-toxic to normal cells [
20]. A recently concluded DIM clinical trial demonstrated that 50% of cervical cancer patients showed improvement [
21]. It is also currently in clinical trials for prostate cancer [
22]. The effects of DIM were recently discussed in detail by Banerjee
et al. [
23]. In our previous study, we showed that DIM exhibits antiproliferative properties in ovarian cancer cells by causing G2/M cell cycle arrest [
17]. However, the mechanism by which DIM inhibits proliferation of ovarian cancer cells was not clear.
In the present study, we provide evidence that DIM induces apoptosis in ovarian cancer cells by blocking the activation of STAT3 and its downstream effector molecules, while IL-6 treatment or overexpression of STAT3 significantly protects ovarian cancer cells from DIM-induced apoptosis. Our results also show that DIM suppresses angiogenesis and metastasis. In addition, DIM potentiates the effect of cisplatin in inducing apoptosis and inhibiting angiogenesis and metastasis. As a proof of concept, in vivo efficacy of DIM alone and in combination with cisplatin also was evaluated.
Methods
Chemicals
BR-DIM was a kind gift from Dr Michael Zeligs (Bio Response, Boulder, CO, USA). Cisplatin was obtained from Novaplus (Bedford, OH, USA). Antibodies against cleaved (Cl)-caspase 3, Cl-poly(ADP-ribose) polymerase (PARP), p-STAT3 (Tyr-705), STAT3, Mcl-1, and survivin were obtained from Cell Signaling Technology (Danvers, MA, USA). VEGF antibody was obtained from R&D systems (Minneapolis, MN, USA), Lamin B was from Santa Cruz Biotechnologies (Santa Cruz, CA, USA), and p-STAT3 (Ser-727) and HIF-1α were obtained from Abcam Inc. (Cambridge, MA, USA). Actin antibody, IL-6, MCDB105 and Medium 199 were procured from Sigma Aldrich (St Louis, MO, USA). RPMI and McCoy 5A were purchased from Mediatech (Manassas, VA, USA). NE-PER nuclear fractionation kit was from Thermo Scientific. EZ-TFA transcription factor assay kit was obtained from Upstate (Millipore, Billerica, MA, USA). Dual luciferase kit was bought from Promega (Madison, WI, USA). The VEGF Elisa kit was from Invitrogen (Carlsbad, CA, USA) and FuGENE 6 was obtained from Roche (South San Francisco, CA, USA). IL-6 secretion ELISA kit was from ebiosciences (San Diego, CA, USA).
Cell cultures
SKOV-3, OVCAR-3 and TOV-21G cells lines were procured from American Type Culture Collection (ATCC; Manassas, VA, USA). SKOV-3 cells were maintained in McCoy's 5A medium supplemented with 10% fetal bovine serum (FBS). OVCAR-3 cells were maintained in RPMI medium supplemented with 20% FBS, 10 mM sodium pyruvate, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 10 mg/l bovine insulin and 4.5 g/l glucose. Human normal ovarian surface epithelium cells (NOSE) were a kind gift from Dr Jinsong Liu at MD Anderson, Houston, TX, USA. NOSE cells were previously transfected with the SV40 early region expressing large T and small t antigens as described elsewhere [
24]. NOSE and TOV21G cells were maintained in 1:1 mixture of MCDB105 and Medium 199 supplemented with 15% FBS. The A2780 cell line (a kind gift from Dr Thomas Hamilton, Fox Chase Cancer Center, Philadelphia, PA, USA) was maintained in RPMI media supplemented with 10% FBS and 2.7 units/ml insulin. OVCAR-429 and OVCAR-433 (a kind gift from Dr Laurie Hudson, University of New Mexico) were maintained in Dulbecco's modified Eagle medium (DMEM) with 10% FBS. A 1% antibiotic mixture was used in all the above media. All the cell lines were maintained at 37°C in a humidified incubator circulated with 5% CO
2/95% air. The cell survival assay was performed as described by our group previously [
17].
Annexin V apoptosis assay
SKOV-3 cells were plated at a density of 0.3 × 10
6 cells per well in a six-well plate and allowed to attach overnight. Cells were then treated with or without DIM. After 24 h cells were exposed to IL-6 for 15 minutes, washed, suspended in binding buffer, and incubated for 15 minutes with annexin V-FITC (BD Biosciences, San Jose, CA, USA). Fluorescence was measured using a C6 Accuri flow cytometer (Ann Arbor, MI, USA) with a minimum of 10,000 events per sample as previously described by our group [
25].
Western blot analysis
SKOV-3, OVCAR-3, TOV-21G and A2780 cells were exposed to varying concentrations of DIM alone or in combination with cisplatin. Cells were collected, lysed, and about 20 to 80 μg protein was subjected to SDS gel electrophoresis followed by immunoblotting as previously described by our group [
26].
Nuclear fractionation
SKOV-3, OVCAR-3, TOV-21G, or A2780 cells were plated at a density of 1 × 106 in 100 mm culture dishes and exposed to different concentrations of DIM for 24 h. Nuclear fraction was extracted using NE-PER kit from Thermo Scientific according to the manufacturer's instructions.
STAT3 DNA binding activity
DNA binding activity of STAT3 was measured by Universal EZ-TFA transcription factor assay colorimetric kit. SKOV-3 or OVCAR-3 cells were treated with or without DIM for 24 h. Nuclear extracts were used to determine the specific STAT3 DNA binding activity as previously described by our group [
27].
STAT3 luciferase reporter assay
Transcriptional activity of STAT3 was determined in SKOV-3 and OVCAR-3 cells by transfecting the cells with 2 μg pLuc-TK/STAT3, which encoded firefly luciferase under the control of STAT3 promoter, and with 0.2 μg of a pRL-TK, which constitutively expressed Renilla luciferase, the latter as a transfection efficiency control. At 24 h after transfection, cells were treated with or without DIM for 24 h. Whole cell lysates were collected using passive lysis buffer provided by dual luciferase reporter assay kit. Renilla and firefly luciferase activities were measured by a luminometer. Firefly luciferase activities were corrected for Renilla values and then normalized relative to dimethylsulfoxide (DMSO) control as previously described [
27].
IL-6 treatment
SKOV-3, OVCAR-3, or TOV-21G cells were treated with 75 μM DIM for 24 h followed by incubation with 10 ng/ml IL-6 for 15 minutes. Cells were then processed for apoptosis assay or western blotting as described above.
MG132 treatment
Since HIF-1α is proteasomally degraded, SKOV-3 and OVCAR-3 cells were pretreated with 10 μM MG132 for 1 h and then treated with 75 μM DIM for 6 h. In another experiment, after DIM treatment, cells were treated with or without 10 ng/ml IL-6 for 15 minutes. Samples were processed for western blotting as described above.
STAT3α overexpression
A total of 0.3 × 106 SKOV-3 cells were plated in McCoy's 5A medium containing 10% FBS without antibiotics and allowed to attach overnight. Complexes were prepared by incubating 2 μg STAT3α plasmid with 6 μl FuGENE 6 transfection reagent in 100 μl McCoy media without serum or antibiotic for 1 h. These complexes were then added to the cells. At 6 h after transfection, the media was replaced by regular media. After 24 h of transfection, cells were further treated with or without DIM for 24 h.
STAT3 small hairpin (sh)RNA
A total of 0.3 × 106 SKOV-3 cells were plated in McCoy's 5A medium containing 10% FBS without antibiotics and allowed to attach overnight. Complexes were prepared by incubating 2 μg STAT3 shRNA with 6 μl FuGENE 6 transfection reagent in 100 μl McCoy media without serum or antibiotic for 1 h. These complexes were then added to the cells. At 6 h after transfection, media was replaced by regular media. After 24 h of transfection, cells were processed for apoptosis assay as described above.
Estimation of IL-6 secretion by ELISA
About 10,000 SKOV-3, OVCAR-3 or OVCAR-429 cells were plated per well in a 96-well plate. Cells were starved overnight followed by treatment with DIM for 24 h. After 24 h, medium was collected and processed for measuring secreted IL-6 levels using ELISA kit according to manufacturer's instructions.
Aortic ring assay
Aortic ring spouting assay was performed as previously described [
28]. In brief, 1 mm long rings were excised from rat thoracic aorta. The rings were submerged in 350 μL Matrigel (BD Biosciences) containing 50 ng/ml IL-6. After a 24-h incubation, DIM, cisplatin, or both were added to the rings and incubated for an additional 3 to 5 days. The aortic rings that formed microvascular-like sprouts were photographed under light microscope (Olympus Inc., PA, USA) and the results were quantified by ImageJ V.1.43 software provided by NIH.
Estimation of VEGF secretion by ELISA
Secreted VEGF levels in DIM treated SKOV-3 and OVCAR-3 cell culture medium were measured using ELISA kit according to manufacturer's instructions.
Wound healing assay
Wound healing assay was performed as described previously [
29]. Confluent monolayers of SKOV-3; OVCAR-3; and TOV-21G cells in six well plates were scratched with a 1 ml pipette tip and incubated in respective medium containing 50 μM DIM. Cells were photographed under a light microscope (Olympus) at 0, 24 and 48 h and the results were quantified by ImageJ software (NIH).
Transwell cell invasion assay
Cell invasion was performed according to the manufacturer's instructions in transwell Boyden's chambers with 8.0 μm pore size filters (BD Biosciences). Briefly, cells were serum starved overnight and harvested by trypsinization. A suspension of 20,000 SKOV-3 cells in 600 μL McCoy medium containing 1% serum were seeded on the upper well of the Boyden's chamber and the lower chamber was filled with 1.5 ml of media containing 1% serum. After incubation for 2 h, DIM or cisplatin or both were added to the upper chamber whereas 10% FBS and 20 ng/ml VEGF was added to the lower chamber as chemoattractant. After incubation for 24 h, cells from the upper chamber were removed by wiping with a cotton swab, and the filter was fixed with 10% trichloroacetic acid (TCA) and stained with 0.4% (w/v) sulforhodamine B (SRB) solution. The filters containing stained cells were removed from the transwell chambers and individually transferred to individual wells in a 96-well plate. The SRB dye retained on the filter was extracted with 10 mM Tris buffer and the absorbance was measured at 570 nm using a microplate reader (BioTek Instruments, VT, USA). Assays were performed in duplicates and data was expressed as percent migration with control.
In vivoxenograft experiment
Female athymic nude mice, 4 to 6 weeks old, were purchased from Charles River Laboratories (Wilmington, MA, USA). The use of mice and their treatment was approved by Institutional Animal Care and Use Committee (IACUC), Texas Tech University Health Sciences Center, and all the experiments were carried out in strict compliance with regulations. Mice were fed with antioxidant-free AIN-76A special diet for a week before starting the experiment. About 5 × 10
6 SKOV-3 cells were injected subcutaneously into both right and left flanks. Eight mice were assigned randomly to each group. Since each mouse was implanted two xenografts, each group had 16 tumors. Once each mouse achieved a tumor of about 90 mm
3, the control group received PBS whereas mice in the treatment group received 3 mg DIM suspended in PBS by oral gavage every day. At day 34, 5 mg/kg cisplatin was injected intraperitoneally to the treatment group mice. Beginning on the 7th day after cell implantation, tumor volume was measured three times a week using vernier calipers until day 48, as previously described by our group [
27]. On day 48 the mice were killed and tumors were removed for western blot analysis.
SRB cell survival assay
Around 5,000 cells in 0.1 ml medium were plated per well in 96 well plates and allowed to attach overnight. Desired concentrations of DIM were added to the cells and incubated at 37°C for 24 h. The cells were then processed and stained with 0.4% SRB solution and the absorbance was read at 570 nm using a Biotek plate reader as described by our group previously [
17].
Statistical analysis
All the statistical analyses were performed using Prism 5.0 (GraphPad Software Inc., San Diego, CA, USA). The data represents mean values with SD. The Student's t test was used to compare the control and treated groups. In experiments involving more than three groups, non-parametric analysis of variance followed by Bonferroni post hoc multiple comparison test was used. All statistical tests were two sided. Differences were considered statistically significant when the P value was less than 0.05.
Discussion
Our results demonstrate that DIM suppresses the growth of ovarian cancer cells and potentiates the effect of cisplatin
in vitro and
in vivo by targeting STAT3 signaling without being toxic to normal ovarian cells. To the best of our knowledge, this is the first report demonstrating STAT3 as a target of DIM. Accumulated evidence indicates the involvement of STAT3 in the transformation of normal cells into malignant ones [
30,
33]. STAT3 is overexpressed in ovarian tumors and is associated with ovarian tumorigenesis [
14]. A recent study demonstrated that STAT3 is activated in 94% of ovarian cancer patients [
14]. Furthermore, STAT3 is also implicated in resistance to chemotherapy in ovarian cancer [
5]. Our studies showed that DIM-induced apoptosis in various ovarian cancer cells was mediated by substantially suppressing the phosphorylation of STAT3 at Tyr-705 and Ser-727. Tyrosine phosphorylation is mainly associated with oncogenic status of STAT3. The effects of DIM were not specific to only SKOV-3 cells as similar observations were made in various other ovarian cancer cell lines such as OVCAR-3, TOV-21G, A2780, OVCAR-429 and OVCAR-433. These studies agree with previous reports that demonstrate that inhibiting constitutive activation of STAT3 by STAT3 inhibitor AG490 inhibits tumor growth [
34].
Unphosphorylated STAT3 resides in the cytoplasm and is activated by phosphorylation. Once phosphorylated at Tyr-705, STAT3 dimerizes and translocates into the nucleus [
35,
36]. Around 86% of ovarian tumor tissues have activated STAT3 in the nucleus and not in the cytoplasm [
14]. Our results clearly show that the nuclear translocation and activation of STAT3 was substantially reduced by DIM treatment. Inhibition of nuclear translocation in turn inhibits DNA binding activity and transcriptional activity of STAT3 [
37,
38]. Specific DNA binding of STAT3 leads to transcriptional activation of several downstream molecules such as survivin and Mcl-1. These results demonstrate that DIM treatment blocks the transcriptional and DNA binding activity of STAT3 and downregulates the expression of survivin and Mcl-1. Our observations confirm previous studies indicating that STAT3 inhibition downregulates survivin or Mcl-1 in various cancers [
27,
39,
40]. It is noteworthy that a gene expression profiling of a previous study showed that survivin is a target of DIM [
41]. Survivin overexpression is implicated in resistance to cisplatin-induced apoptosis [
42,
43]. Our results showed that survivin expression was substantially upregulated by cisplatin treatment in ovarian cancer cells. However, DIM treatment completely abolished the overexpression of survivin by cisplatin, suggesting a potential role of DIM in reducing cisplatin-mediated resistance.
Various cytokines and growth factors can activate STAT3. IL-6 is well known to activate STAT3 by phosphorylation at Tyr-705. DIM not only eliminated constitutive activation of STAT3, but also blocked IL-6 mediated activation of STAT3. In addition, to convincingly establish STAT3 as a target of DIM, cells were transfected with STAT3 expression plasmid to activate STAT3 in ovarian cancer cells. Induction of apoptosis by DIM was almost completely blocked in cells overexpressing IL-6 or STAT3, demonstrating that induction of apoptosis in our model was mediated through STAT3 downregulation. Our data indicating that DIM induces apoptosis by inhibiting STAT3 was also supported by our STAT3 knockout studies, which demonstrated the induction of apoptosis when STAT3 was knocked out by shRNA. These results agree with a previous study showing the role of STAT3 in suppressing apoptosis in pancreatic cancer cells [
27]. Furthermore, DIM treatment dramatically reduced the secretion of IL-6 not only in ovarian cancer cells but also in tumors. Interestingly, a recent study reported higher IL-6 levels in ovarian cancer patients [
31]. This study indicated the IL-6-STAT3-HIF axis as an important target for therapy in ovarian cancer [
31]. Our results show that DIM treatment alone or in combination with cisplatin inhibited neovascularization induced by IL-6, suggesting the antiangiogenic potential of DIM. HIF-1α and VEGF play an important role in angiogenesis. HIF-1α is short lived and degrades proteasomally; however, MG132 blocks the degradation of HIF-1α. Our results show that HIF-1α expression retained by MG132 was suppressed as early as 6 h after DIM treatment. IL-6 also activated HIF-1α, which clearly indicates that HIF-1α perhaps is regulated through STAT3 in ovarian cancer cells. It is obvious that inhibition of HIF-1α by DIM was mediated through STAT3 in our model [
13,
44]. VEGF is another important regulator of angiogenesis and its expression in cancer cells has been shown to correlate with activation of STAT3 [
33]. Previous reports have suggested that overexpression of STAT3 increases VEGF expression, leading to angiogenesis [
33]. Hence disrupting the activation of STAT3 would inhibit VEGF, blocking angiogenesis [
45]. Our results demonstrated that DIM inhibits both VEGF expression and VEGF secretion. This explains the mechanism of the antiangiogenic effects of DIM in ovarian cancer cells in our study.
Cisplatin is a frontline drug used in the treatment of advanced ovarian cancer [
46]. Nonetheless, only 15% of patients treated with cisplatin achieve long-term survival; the rest experience persistent or recurrent disease [
47]. Moreover, cisplatin is associated with cytotoxicity and resistance to chemotherapy [
48,
49]. STAT3 activation or overexpression is associated with cisplatin resistance [
5]. Our results demonstrate that DIM potentiates the effect of cisplatin at one-quarter of its IC
50 concentration by inhibiting the activation of STAT3. Combination treatment blocked phosphorylation of STAT3 at both Tyr-705 and Ser-727 when compared to cisplatin treatment alone. Likewise, VEGF secretion and angiogenic sprouting were noticeably reduced in combination treatment as compared to cisplatin alone. These results agree with previous studies that demonstrated inhibition of STAT3 by an analog of aspirin potentiated the effects of cisplatin [
50]. The CI provides qualitative information about the nature of drug interaction. CI values less than 1, equal to 1 or greater than 1 indicate synergistic, additive or antagonistic effects, respectively. Our CI values for the combination treatment at both the concentrations of DIM with cisplatin were less than 1, showing the synergistic effects of DIM with cisplatin.
Oral administration of 3 mg DIM per day substantially suppressed the growth of established ovarian tumors, indicating the tumor regression potential of DIM. However, complete regression of tumors was not observed with DIM. When given with cisplatin treatment, DIM further retarded the growth of tumors. The tumors from DIM or DIM plus cisplatin-treated mice clearly demonstrated inhibition of STAT3 signaling and increased apoptosis. Similar to our
in vitro observations, cisplatin treatment caused a drastic increase in the expression of survivin in the tumors, which was completely suppressed by DIM in combination treatment. Interestingly, mice that received DIM did not show any significant change in body weight when compared with the weight of control mice. However, the mice that received cisplatin and DIM showed a decrease in weight, though this was not as drastic as in the mice that received only cisplatin, suggesting that DIM may reduce the systemic toxicity associated with cisplatin. DIM is a major indole compound present in cruciferous vegetables, and is consumed on a daily basis [
51]. A recent single dose phase I clinical trial suggested that 200 mg DIM is well tolerated not only by healthy volunteers but also in patients with cervical cancer [
21,
52]. In the same study 200 mg single dose produced a mean C
max of 104 ng/ml and mean area under curve (AUC) of 553 h ng/ml. The half-life of this dose was 2.6 ± 0.7 h. Interestingly a single dose of DIM alone caused a significant clinical improvement in patients with stage II and stage III cervical intraepithelial neoplasia (CIN) [
21]. Administration of 2 mg/kg/day DIM via an oral route showed improvement in pap smear results, human papillomavirus (HPV) status, and improved CIN by 1 to 2 grades [
21]. Several pharmacokinetic studies on DIM have stated that up to 300 mg in a single dose of DIM can be tolerated by humans [
52], indicating that our dose of DIM falls within the accepted and tolerated dose. Cisplatin in our studies was given intraperitoneally at a dose of 5 mg/kg twice a week, which is 14 mg/m
2 when converted to the equivalent human dose [
53]. The weekly dose of cisplatin ranges from 40 to 140 mg/m
2 in humans. We administered cisplatin twice a week, with an additive effect of 28 mg/m
2, which is much lower than the regular dosage typically given to patients with cancer. If a low dose of cisplatin can be given to patients without the loss of any therapeutic effect but with reduced side effects, it would represent a significant breakthrough in clinical practice. Nevertheless, further clinical studies are needed to show that DIM can reduce the side effects of cisplatin.
In conclusion, our results firmly establish that DIM induces apoptosis in ovarian cancer cells by inhibiting STAT3 signaling. Our results also provide evidence that DIM inhibits the invasion of ovarian cancer cells and angiogenesis by inhibiting HIF-1α and VEGF, which are regulated by STAT3. Importantly, DIM potentiated the effect of cisplatin both in culture and in vivo by inhibiting STAT3. Taken together, the findings from our study provide support for the use of DIM alone or in combination with cisplatin in preclinical and clinical settings in the management of ovarian cancer patients.
Acknowledgements
This work was supported in part by R01 grants CA106953 and CA129038 (to SKS) awarded by National Cancer Institute, NIH. We thank Dr Jacqueline Bromberg (Sloan Kettering Memorial Cancer Center, NY, USA) for providing STAT3α and STAT-Luc plasmids, Dr Thomas Hamilton (Fox Chase Cancer Center, PA, USA) for providing A2780 cells, Dr Jinsong Liu (MD Anderson, TX, USA) for providing NOSE cells, Dr Laurie Hudson (University of New Mexico, NM, USA) for providing OVCAR429 and OVCAR433 cells, and Dr Michael Zeligs (Bio Response, CO, USA) for providing BR-DIM for our studies. The technical help of Srinivas Reddy Boreddy and Kartick C Pramanik with the in vivo experiment is greatly appreciated.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Both PKK and SKS were responsible for designing the study, analyzing the data and writing the manuscript. Experiments were performed by PKK. Both authors read and approved the final manuscript.