ARTICLE IN PRESS

YGYNO-972623; No. of pages: 8; 4C: 5

Available online at www.sciencedirect.com

Gynecologic Oncology xx (2008) xxx xxx www.elsevier.com/locate/ygyno

Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer

David Durrant a , Joanna E. Richards a , Winston T. Walker a , Kristen A. Baker a , Daniele Simoni b , Ray M. Lee a,
a

Massey Cancer Center, and Department of Internal Medicine, Virginia Commonwealth University, Richmond, VA, USA b Dipartimento di Scienze, Farmaceutiche, Universita di Ferrara, Ferrara, Italy Received 17 December 2007

Abstract Objective. Stilbene derivative, cis-3, 4, 5-trimethoxy-3-aminostilbene (stilbene 5c), is highly potent to induce cell death in ovarian cancer cells. This study is to investigate its mechanism to induce cell death. Methods. UCI101 ovarian cancer cells were used for this study. Cell death was analyzed by Alamar blue staining. Cell cycle was analyzed by flow cytometry after PI staining. Mitochondrial potential and reactive oxygen species were determined by MitoTracker green and DCF-DA, respectively. Immunofluorescent staining was done with tubulin antibody following by confocal microscope examination. Cell lysates were collected after treatment with stilbene 5c for Western blotting analysis of various cell cycle regulators and signal transduction mediators. Results. Stilbene-treated cells die in both cell cycle-dependent and -independent pathways. Low concentration (30 nM) induces cell death without cell cycle arrest. This process involves disruption of mitochondrial potential and production of ROS by a Bcl-2-independent pathway. Higher concentration of stilbene 5c arrests cell cycle in G2/M phase, which is supported by dephosphorylation of Cdc2 and Cdc25C, and transiently elevation of spindle checkpoint BubR1. Although phosphorylation of Chk1 and Chk2 both increases after treatment, loss of Chk1 suppresses, whereas loss of Chk2 enhances, stilbene 5c-induced cell death. Phosphorylation of Akt and Stat3, but not MAPK, is suppressed after stilbene 5c treatment. Conclusion. These studies provide a mechanistic insight in using stilbenes in ovarian cancer. Stilbenes could be potentially useful agents for ovarian cancer therapy and induce cell death through mitochondrial damage and cell cycle arrest. 2008 Elsevier Inc. All rights reserved.
Keywords: Stilbenes; Ovarian cancer; Therapy; Signal transduction; Cell cycle; Mitochondria

Introduction Ovarian cancer is the leading cause of death from gynecological malignancies [1]. Vast majority of women with ovarian cancer are presented with advanced disease when diagnosed. Women with epithelial ovarian cancer undergo a surgical cytoreductive procedure followed by postoperative chemotherapy. While the primary and larger tumor nodules may be removed surgically, micronodular disease and floating tumor colonies in
Corresponding author. Massey Cancer Center, Virginia Commonwealth University, 1101 E. Marshall Street, Richmond, VA 23298, USA. Fax: +1 804 828 8079. E-mail address: rlee5@vcu.edu (R.M. Lee). 0090-8258/$ - see front matter 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2008.02.031

the peritoneal cavity cannot be treated adequately by surgery. This has led to a dismal 5-year survival rate for advanced stage disease. There is an urgent need to develop a new therapy for ovarian cancer. Standard chemotherapeutic agents for ovarian cancer are platinum-based DNA alkylating agents and taxane group microtubule inhibitors (MIs) [2]. There is a lot of interest in developing new MIs agents because of their successful track record. There are three types of MIs based on their binding sites in tubulin [3]. Vinca alkaloids such as vincristine, vinblastin and vinorelbin, are commonly used in hematological malignancies and solid tumors. Taxanes such as paclitaxel and docetaxel are commonly used in ovarian cancer. The third group is colchicine-site inhibitors (CSIs), but currently none of them has been approved by FDA for

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
2 D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx

cancer therapy. Investigation in this group of compounds was facilitated by solving the structure model of tubulin bound with colchicine [4], which provides a useful tool for computer-based drug design and optimization. The most advanced CSIs in clinical development are combretastatin-A4-phosphate (CA4P) [57] and ZD6126 [8,9]. Both of them have completed phase I clinical studies, and CA4P is in phase II study for solid tumors, particularly in anaplastic thyroid cancer [10]. They have the capability of disrupting tumor vessels and were developed as vascular disrupting agents (VDAs). Using dynamic contrast-enhanced (DCE) MRI or PET scan, both CA4 and ZD6126 were shown to suppress tumor perfusion within 24 h after drug administration. However, the caveats of these two drugs are cardiotoxicity and neurotoxicity [68]. Neurotoxicity has been an unavoidable side effect of all MIs, but cardiac toxicity is unique for CSIs. Tachycardia, bradycardia, hypertension, prolongation of QTc interval, decreased LV ejection fraction, and even elevated cardiac enzymes and characteristic EKG change like myocardiac infarction have been reported. Development of new CSIs with lower cardiac toxicity will be essential for CSIs to be widely utilized for cancer therapy. The mechanism of CSI-induced growth suppression is mediated by disruption of mitotic spindles to cause cell cycle arrest [11]. The initiation of mitosis is regulated by Cdc2/cyclin B complex [12]. Dephosphorylation of Cdc2 by phosphatase Cdc25 lead to activation of Cdc2 and initiation of the mitotic reactions [13]. Activation of Cdc25 is also mediated by a dephosphrylation event. Phosphorylated Cdc25 interacts with cytoplasmic 14-3-3 protein and is sequestered in the cytoplasm [14]. Dephosphorylation of Cdc25 leads to its release from 143-3 protein and allows its translocation into nucleus to dephosphorylate Cdc2. Proliferated centrosomes are separated by mitotic motor protein kinesin [15,16], and serve as organization centers for mitotic spindle formation. Cells treated with CSIs will not be able to form mitotic spindles to pull chromatids to either side of cells. Cells will be arrested in this stage by activation of spindle checkpoints, which involve protein Bub1, BubR1, and Mad2 to avoid the genetic disaster of inappropriate segregation of sister chromatids [1719]. With prolonged arrest, cells may exit the cell cycle arrest with a consequence of poor chromosome segregation [20]. In this circumstance cells will exhibit the morphology of multinuclear cells and undergo cell death through a delayed process called mitotic catastrophe [21]. Two other checkpoint proteins Chk1 and Chk2, initially thought to be only involved in DNA damage response [22], are also involved in spindle checkpoint regulation. During mitosis, both Chk1 and Chk2 are located in centrosomes [23]. However, it is still unclear what is their function in centrosomes. Inhibition of Chk2, but not Chk1, facilitate the induction of mitotic catastrophe [24], indicating its involvement in mitotic progression in addition to DNA damage response. CA4 belongs to the stilbene (1,2-diphenylethylene) family. We have studied a stilbene derivative cis-3, 4, 5-trimethoxy-3aminostilbene (stilbene 5c) [25], which competes with radioactive colchicine and fits into the colchicine-binding pocket of tubulin [26]. In this report, we investigate the mechanism of

stilbene 5c in triggering cell death in ovarian cancer cells. We found that stilbene 5c utilizes cell cycle-dependent mechanism as expected from its anti-tubulin function. However, there is also a cell cycle-independent pathway to cause cell death that is mediated by mitochondrial disturbance. In cells, mitochondria are not only the organelle for energy generation but also a key regulator of apoptosis [27,28]. Activation of apoptotic signals, such as Bax [29] and caspase [30], can make mitochondrial electron transport chain complex ineffective to enhance the production of reactive oxygen species (ROS). ROS subsequently induce peroxidation of lipids and proteins and a variety of structural damages and eventually cell death [31]. With the understanding of mechanism of stilbene 5c in induction of cell death, stilbene 5c could be exploited further to develop a new strategy to treat ovarian cancer.
Materials and methods Materials and cell lines
Stilbenes are synthesized based on the methods described (19). SKOV3 ovarian cancer cells are from ATCC and UCI-101 cells are supplied by Dr. Thai Cao, University of Utah as described [32]. Cells were grown in Iscove's Modified Dulbecco's medium (IMDM) supplemented with 10% fetal bovine serum and penicillium/streptomycin/glutamine in 5% CO2. Chk1 and Chk2 deficient cells derived from MDA-MB231 cells are kindly provided by Dr. Allen Eastman (Dartmouth University, Hanover, NH).

Cell cycle and cell death analysis

Cell cycle analysis was determined by staining trypsinized cells with 50 g/ ml propidium iodide (PI) in a buffer containing 3.8 mM sodium citrate, 0.1% Triton X-100 and 7 kU/ml RNAse followed by flow cytometry analysis. To determine tumor growth suppression and IC50 of different CSIs, cells were grown in 96-well plates and treated with 0, 0.01, 0.03, 0.1, 0.3, 1.0, and 3.0 M for 48 h before harvested for Alamar blue staining. In this staining, 1/10 volume of Alamar Blue solution was added to each well and optical density (OD) at 570 and 600 nm was determined by a microplate reader. The percentage of growth inhibition was calculated according to the manufacturer's formula as follows: [(117216 A570) (80586 A600)]/[(117216 Ao570) (80586 Ao600)] 100. In this formula, A570 is the absorbance of the treated samples at 570 nm; A600 is the absorbance of the treated samples at 600 nm; Ao570 is the absorbance of the untreated samples at 570 nm; and Ao600 is the absorbance of the untreated samples at 600 nm. The two constants, 117216 and 80586, are the extinction coefficients of Alamar Blue at 570 and 600 nm respectively. Each concentration was repeated in triplicates.

Mitochondrial potential and reactive oxygen species (ROS) analysis

Cells treated with stilbene 5c for various durations was incubated with MitoTracker green (Invitrogen, Carlsbad, CA) per manufacturer's suggestion. Cells were analyzed with flow cytometry within 30 min. For ROS quantification, cells were first treated with or without 50 M butylated hydroxyanisol (BHA) (Sigma Chemical, St. Louis, MO) for 1 h followed by addition of stilbene 5c. Cells were trypsinized and incubated with 10 M 2, 7-dichlorofluorescin diacetate (DCF-DA) (24) at various time points for flow cytometry study.

Immunofluorescent staining and Western blotting

Cells were fixed with 2% paraformaldehyde in phosphate buffered saline (PBS) and permeabilized with 0.2% Triton X-100/PBS for 2 min and rinsed with PBS. Non-specific reaction was blocked by 3% bovine serum albumin in TBS for 15 min, followed by incubation with tubulin monoclonal antibody for 1 h. Cells were washed with TBS 5 times and incubated with secondary antibody

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx conjugated with FITC for 30 min. Nuclei were stained with DAPI. Cells were examined by confocal microscope. For Western blotting analysis, cells were harvested with lysis buffer (50 mM Tris, pH 7.4, 100 mM NaCl, 1% Triton X-100, 1 mM EDTA and 1 mM phenylmethylsulfonyl fluoride and sonicated to break cellular DNA. The lysates were centrifuged at 12,000 g and the supernatants were analyzed with SDS-polyacylamide gel electrophoresis followed by electrotransfer to the Immobilon-P membrane (Millipore, Inc.) for Western blotting. All antibodies against various kinases are from Cell Signaling Technology, Inc. (Boston, MA). 3

merization [26] and disrupts mitotic spindles to arrest cells in early mitosis. Therefore stilbene 5c may have dual mechanisms in causing cell death, one related with cell cycle arrest and one not. Stilbene 5c does not damage DNA damage directly but its effect is enhanced by Chk2 deficiency The mechanism of cell cycle-independent effect of stilbene 5c was explored next. One possibility is that stilbene 5c directly damages DNA to induce cell death. To address this possibility, we treated UCI10 1 cells with stilbene 5c and harvested cells at 0, 4, 8, 16 h to study phosphorylation of H2AX, a DNA damage marker. There was no increase in H2AX phosphorylation at 4 8 h after stilbene 5c treatment (Fig. 2a). At 16 h, a significant increase of H2AX phosphorylation appeared, but this could be directly from cell cycle arrest in mitotic stage and does not represent DNA damage response. In contrast, cells treated with adriamycin induces changes in H2AX phosphorylation. A direct DNA alkaline unwinding assay to measure DNA strand break [33] also failed to demonstrate that stilbene 5c directly induces DNA damage, whereas a positive control with irradiation led to enhanced DNA unwinding due to DNA damages (data not shown). We then investigated the effect of DNA damage responders such as Chk1 and Chk2. We treated UCI101 cells with stilbene 5c, 6c (cis-3, 4, 5-trimethoxy-3-hydroxystilbene, a compound replacing the 3 amino group of stilbene 5c with hydroxyl group) or paclitaxel and harvested cell lysates for Western blotting. Phosphorylation of both Chk1 and Chk2 was enhanced at 16 and 36 h after treatment with any of

Results Stilbene 5c induces apoptosis and cell cycle arrest We first tested UCI-101 and SKOV3 ovarian cancer cells for their sensitivity to stilbene 5c, and compared with other CSIs, including colchicine, vincristine, CA4 and another potent stilbene derivative GG251. CA4 is the most active among the compounds we tested and has IC50 at 2 nM for UCI-101 cells and 8 nM for SKOV3 cells, whereas other CSIs induce growth suppression with IC50 between 30100 nM (Figs. 1a,b). This finding confirms that stilbene 5c is a potent cell growth inhibitor. Studying cell cycle distribution in UCI-101 cells showed that stilbene 5c was ineffective at 10 nM, but induced a significant subG1 population at 30 nM. Interestingly, there was no G2 arrest associated with the appearance of subG1 population (Fig. 1c), suggesting that induction of subG1 population is independent of cell cycle arrest. When the concentration of stilbene 5c was further increased to 100 nM, majority of cells were arrested in G2/M phase, which is consistent with our previous observation that stilbene 5c inhibits microtubule poly-

Fig. 1. Stilbene 5c induces growth suppression in ovarian cancer cells. (a) Dose response in UCI-101 cells. UCI-101 cells in 96-well plates were treated with various concentrations (0, 0.01, 0.03, 0.1, 0.3, 1.0, and 3.0 M) of stilbene 5c, vincristine, colchicine, CA4 and a stilbene derivative GG251 for 48 h. Tumor cell growth was evaluated with Alamar blue staining and the percentages of tumor growth suppression were calculated. Shown are the averages of three measurements. (b) Dose response of SKOV3 cells. (c) Flow cytometry of UCI-101 treated with various concentrations of stilbene 5c. Cells were stained with PI before FACScan analysis. Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
4 D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx

Fig. 2. Stilbene 5c induces cell death through cell cycle-dependent and -independent pathway. (a) Effect of stilbene 5c in H2AX phosphorylation. UCI-101 cells were treated with stilbene 5c and harvested at 0, 4, 8, and 16 h. Cell lysates were probed with phospho-H2AX antibody by Western blotting. (b) Effect of stilbenes 5c, 6c and paclitaxel on mitosis-related cell cycle regulators. UCI-101 cells were treated with 0.1 M of test compounds for 4, 8, 16 and 36 h and harvested for Western blot analysis using various phospho- and total protein antibodies. Actin was used as loading control. (c) Effect of Chk1 and Chk2 on effect of stilbene 5c. MDA-MB231 cells and their corresponding cells with deficiency of Chk1 or Chk2 were first tested with Western blotting for Chk1 and Chk2 expression. Cells were treated with stilbene 5c at 0, 10, 30 and 100 nM and harvested at 24 h for PI staining and flow cytometry analysis.

the three MIs (Fig. 2b). Because it is still a late event similar to H2AX phosphorylation, we believe that this event is also related with cell cycle arrest in mitosis rather than a DNA damage response. This is consistent with several reports that Chk1 has a G2/M check point function that is independent of DNA damage response [34,35] and Chk1 is required for spindle check point [36]. We then used MDA-MB231 and its Chk1 or Chk2 deficient derivatives to study the role of Chk1 and Chk2 in stilbene 5c sensitivity. Western blotting confirmed the lack of

expression in Chk1 or Chk2 (top, Fig. 2c). Dose response analysis showed that Chk2 deficient cells are more sensitive to stilbene 5c by a factor of 3 folds. The cell cycle-independent subG1 population appeared in Chk2/ cells at 10 nM in contrast to the control MDA-MB231 cells at 30 nM stilbene 5c, indicating that loss of Chk2 enhanced the sensitivity to stilbene 5c (Fig. 2c). In contrast, loss of Chk1 makes MDA-MB231 cells more resistant to cell death, and subG1 population disappeared even at 100 nM stilbene 5c, although the G2 arrest was still

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx 5

drial potential before and after stilbene treatment. Cells were treated with stilbene 5c or 6c and stained with MitoTracker green at 0, 1, 2, 4 and 16 h for flow cytometry analysis. Fluorescence uptake of MitoTracker decreased significantly within 1 h and stayed low until 16 h, indicating that disruption of mitochondrial potential appears quickly after stilbene treatment (Figs. 3a,b). Quantification of ROS by DCF-DA revealed that cells treated with stilbene 5c have a higher level of ROS. Addition of antioxidant BHA suppressed the production of ROS (Figs. 3c,d). This is consistent with the current theory that mitochondria are the key organelles to produce ROS when cells are treated with apoptotic stimuli, and that stilbene 5c may cause mitochondrial damage leading to loss of mitochondrial potential and production of ROS. We overexpressed Bcl-2 in UCI-101 cells to test how Bcl-2 affects the sensitivity to stilbene 5c. However, overexpression of Bcl-2 cannot abolish the effect of stilbene 5c (Fig. 3e). Mechanism of stilbene 5c-induced cell cycle arrest Next we explored the cell cycle-dependent mechanism of stilbene 5c. When cells are treated with MIs such as vincristine or paclitaxel, cells are arrested in G2/M phase. Condensed chromosomes in early mitotic stage will not be able to be aligned in the midplane due to lack of mitotic spindles. Cells

Fig. 3. Stilbenes disrupt mitochondrial damages. (a, b) Effect on mitochondrial potential. UCI-101 cells were incubated with 100 nM of stilbene 5c (a) or 6c (b). Cells were harvested at 0, 1, 2, 4, and 16 h after treatment and incubated with MitoTracker green to measure mitochondrial potential. Shown are the overlays of histograms in each time point. (c) Stilbene 5c-induced ROS production. UCI101 cells were treated with 30 nM stilbene 5c and harvested at 0, 3 and 6 h. Cells were stained with DCF-DA to quantify ROS production. (d) Inhibition of ROS production by BHA. UCI-101 cells were first incubated with 50 M BHA for 1 h followed by treatment with 30 nM stilbene 5c for 0, 3 and 6 h before adding DCF-DA to quantify ROS production. (e) Overexpression of Bcl-2 cannot suppress stilbene 5c-induced cell death. UCI-101 cells transfected with the control vector or Bcl-2 overexpressing vector (UCI-Bcl-2) were examined with Western blotting and incubated with 0.1 M stilbene 5c followed by staining with PI and flow cytometry analysis.

present (Fig. 2c). We conclude that stilbene 5c-induced cell cycleindependent cell death is not related with direct DNA damage, but Chk2 inhibits, whereas Chk1 activates, a pathway that enhances cell death. Stilbene 5c interferes with mitochondrial potential and increases ROS production What could be the cell cycle-independent mechanism that stilbene 5c induces cell death? Since stilbene 5c does not induce DNA damage directly, we investigated the effect of stilbene 5c in mitochondria. MitoTracker was first used to study mitochonFig. 4. Stilbene 5c-induced cell cycle arrest is mediated by disrupting microtubule. (a) Immunofluorescent staining with tubulin antibody. UCI-101 cells were treated with or without 100 nM stilbene 5c for 16 h and subjected to immunofluorescent staining with anti-tubulin (-tubulin in green, -tubulin in red) antibodies. (b) Changes in kinases involved in mitotic cell cycle regulation. UCI-101 cells were treated with 100 nM stilbene 5c, 6c or paclitaxel for 0, 4, 8, 16, and 36 h. Cells were harvested and Western blot was probed with p-Cdc25c, Cdc25c, p-Cdc2, Cdc2, and BubR1 antibodies. Actin was used as loading control.

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
6 D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx

will be arrested in early mitotic stage with chromosome scattered in cells. We stained the UCI-101 cells with tubulin, -tubulin and DAPI, and confirmed that majority of cells are blocked in early prophase after stilbene 5c treatment. In contrast, the control cells exhibited normal mitotic spindles during cell division (Fig. 4a). We next investigated kinases that regulate mitotic entry. Western blotting of cell lysates treated with stilbene 5c, 6c or paclitaxel shows that Cdc2 is completely dephosphorylated at 16 h after stilbene 5c treatment (Fig. 4b), even though the total amount of Cdc2 protein had a transient increase at this time point. Phosphorylation of Cdc2 then recovers to a small degree at 36 h. Similarly, phosphorylation of Cdc25c disappeared at 16 h and recovered to some degree at 36 h (Fig. 4b). These findings indicate that cells have completed progression from G2 to M phase and arrested in early mitotic stage after treatment with stilbene 5c. The level of mitotic spindle checkpoint BubR1 increases at 48 h and decreases to baseline at 36 h after stilbene 5c treatment (Fig. 4b). This finding is consistent with the function of BubR1 to execute spindle checkpoint function. Similar study with other microtubule inhibitors, stilbene 6c and paclitaxel, showed similar findings in the cell cycle kinase profile, suggesting that these are common events after cells are treated with MIs. Stilbene 5c affects signal transduction pathways Paclitaxel is one of the standard agents in ovarian cancer therapy. There are many reports that paclitaxel induces changes in signal transduction cascade [3739], but changes induced by CSIs are less well characterized. Modification of various kinase activities can also modulate the sensitivity to paclitaxel [37 39]. We investigated the cellular response to stilbenes 5c, 6c and paclitaxel in three major kinase cascades, Jak/Stat3, PI3K/Akt and Raf-1/MEK/ERKs. Phosphorylation of Stat3 has no change in the first 8 h, decreases at 16 h, and then recovers at 36 h (Fig. 5). Previously, Kotha et al. reported that resveratrol, a trans-hydroxylstilbene, inhibited Src and Stat3 signaling to induce apoptosis in cancer cells with high Stat3 activity [40].

With the pro-proliferative property of Stat3, stilbene 5c and other stilbene derivatives may provide a strategy to induce cell death through blocking Stat3 signaling. The second pathway we tested is PI3K/Akt. Phosphorylation of Akt at Ser473 decreased after 16 h and stayed low at least until 36 h, which is also a favorable trend in induction of cell death. Finally studies with MAPKs show that there is no major change in ERK1 and ERK2 phosphorylation, suggesting that ERK activity may not be related with stilbene 5c-induced cell death. No difference was observed among cells treated with stilbene 5c, 6c or paclitaxel. Discussion We report the mechanistic studies of stilbene 5c-induced cell death. In our previous study, we demonstrated that stilbene 5c is a potent microtubule inhibitor at the colchicine site. Stilbene 5ctreated cells exhibit disruption of mitotic spindle leading to cells with multiple micronuclei (Fig. 4), a finding suggesting that cell death could be due to mitotic catastrophe [21]. Investigation of kinases involved in cell cycle regulation confirmed that stilbene 5c-treated cells are arrested in mitotic stage based on dephosphorylation of Cdc2 and Cdc25c. Elevation of spindle checkpoint protein BubR1 is also consistent with this theory. What is surprising is the detection of a subG1 population without association of G2/M arrest in cells treated with 30 nM stilbene 5c. This finding suggests that there is a cell cycle-independent mechanism for stilbene 5c. To explore the mechanism of cell cycle-independent cell death, initially we hypothesized that cell cycle-independent cell death could be related to DNA damage. We tested this hypothesis with a DNA damage marker, H2AX phosphorylation. However, the result was considered negative even though phosphorylation of H2AX was dramatically increased 16 h after stilbene 5c treatment because late phosphorylation could well be an event in mitosis. Further testing with alkaline unwinding study failed to support any direct DNA damage. The study with Chk1 and Chk2, however, is interesting. Phosphorylation of Chk1 and Chk2 increases at 16 and 36 h, but not at early time points of 4 and 8 h. It is known that Chk1 and Chk2 are involved in mitosis [41], and Chk1 phosphorylates Cdc25B independent of DNA damage [27,28]. Phosphorylation of Cdc25B prevents Cdc2 activation and cell cycle entry into mitosis [21]. In our study, phosphorylation of Chk1 and Chk2 in cells treated stilbene 5c starts at 16 h and becomes more prominent at 36 h, a time point after that of Cdc2 and Cdc25c dephosphorylation (Fig. 4). Hence the Chk1 and Chk2 phosphorylation is unrelated to Cdc25 dephosphorylation and Cdc2 activation. In contrast, it could be a feedback mechanism that cells activate Chk1 and Chk2 to induce Cdc25 phosphorylation and attempt to exit from mitosis. The Chk1 and Chk2 deficient cells yield another interesting finding. Chk2-deficient cells are more sensitive, whereas Chk1deficient cells are more resistant to stilbene 5c-induced cell death (Fig. 2c). The mechanism of this phenomenon is probably related with the mechanism of cell death through mitotic catastrophe. It has been reported that Chk2 is a negative regulator of mitotic catastrophe [24]. Cell death of solid tumor cells

Fig. 5. Stilbene 5c induces changes in signal transduction cascade. UCI-101 cells were treated with 100 nM stilbene 5c, 6c or paclitaxel for 0, 4, 8, 16, and 36 h. Cells were harvested and Western blot was probed with p-Stat, Stat, p-AKT (Ser473), Akt, p-MAPK or MAPK antibodies. Actin was used as loading control.

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx 7

is generally not due to apoptosis, but rather through necrosis, autophagy, mitotic catastrophe and senescence. Among these mechanisms of cell death, mitotic catastrophe, characterized by formation of multiple micronuclei, is most commonly utilized in cells treated with MIs. Activation of Chk2 leads to downregulation of Cdc2/cyclin B complex and prevents cell cycle entry, and inhibition of Chk2 enhances activation of mitotic catastrophe [24]. Multiple micronuclei, a hallmark of mitotic catastrophe, are a frequent finding in cells after prolonged treatment with stilbene 5c. CA4 was also reported to induce cell death through mitotic catastrophe [42]. However, it remains unclear how Chk1-deficient cells are more resistant to stilbene 5c. Stilbene 5c induces disruption of mitochondrial potential and production of ROS within the first 4 h, suggesting a direct mitochondrial damaging effect. This is apparently not related with common apoptosis pathway since overexpression of Bcl-2 in cells failed to suppress the effect of stilbene 5c. What could be the mechanism of this mitochondrial effect? One possibility is that stilbene 5c directly suppresses mitochondrial oxidative phosphorylation, particularly Complexes I and III, which are the two sites of mitochondrial electron transfer chain to produce most ROS. Several compounds that block mitochondrial electron transport chain are known, for example, rotenone for Complex I and antimycin for Complex III. Rotenone has some degree of similarity to trans-stilbene in structure. Resveratrol, a trans-stilbene, has been shown to induce expression of mitochondrial oxidative phosphorylation proteins and improve mitochondrial function through activation of the protein deacetylase, SIRT1 [43]. We have tested the protein level of oxidative phosphorylation complex in cells treated with stilbene 5c but could not detect any difference (data not shown). A functional study of the oxidative phosphorylation may help to address this question. The current study focuses on mechanistic investigation of stilbene 5c in tissue culture cells. The in vivo efficacy of tumor growth suppression was reported in a separate manuscript [45]. In that study, stilbene 5c was given by daily intraperitoneal injection into nude mice with subcutaneous xenograft model. Tumor growth was suppressed by a marginal benefit of only 30%. This could be attributed to the fact that stilbene 5c is similar to other vascular disrupting agents to shut down tumor perfusion, leading to a hypoxic status in tumor. Hypoxic tumor could secrete angiogenic agents such as VEGF to induce mobilization of bone marrow endothelial progenitor cells, which home in tumor vascular bed to reestablish tumor circulation [44]. In that manuscript, we combined VEGF inhibitor bevacizumab and stilbene 5c to treat the same xenograft model, and achieve a significantly better therapeutic response (80% tumor growth suppression). That study provides a useful guidance for how to use stilbene 5c in future clinical application. In conclusion, stilbene 5c induces cell death through both cell cycle-independent and -dependent pathways. The former is through mitochondrial damage and ROS production, and the latter is related with mitotic arrest and cells eventually die from mitotic catastrophe. Understanding the mechanism of stilbene 5c will facilitate future development of this compound and

designing combination strategies to achieve a better therapeutic effect.

Conflict of interest statement The authors declare that they have no conflict of interest to disclose.

Acknowledgments We thank Dr. Allen Eastman, Dartmouth University, Hanover, NH for providing the Chk1 and Chk2 deficient cells for the study, Dr. David Gewirtz for advice in DNA alkaline unwinding study. The work was supported by Massey Cancer Center and Virginia Commonwealth University pilot program (R.M.L.). The flow cytometry and confocal microscope are supported by Cancer Center Core Grant (P30 CA16059). References
[1] Jemal A, et al. Cancer statistics. 2007CA Cancer J Clin 2007;57(1):4366. [2] Mano MS, et al. Current management of ovarian carcinosarcoma. Int J Gynecol Cancer 2007;17(2):31624. [3] Jiang N, et al. Advances in mitotic inhibitors for cancer treatment. Mini Rev Med Chem 2006;6(8):88595. [4] Ravelli RB, et al. Insight into tubulin regulation from a complex with colchicine and a stathmin-like domain. Nature 2004;428(6979):198202. [5] Pettit GR, et al. Isolation and structure of the strong cell growth and tubulin inhibitor combretastatin A-4. Experientia 1989;45(2):20911. [6] Stevenson JP, et al. Phase I trial of the antivascular agent combretastatin A4 phosphate on a 5-day schedule to patients with cancer: magnetic resonance imaging evidence for altered tumor blood flow. J Clin Oncol 2003;21 (23):442838. [7] Rustin GJ, et al. Phase I clinical trial of weekly combretastatin A4 phosphate: clinical and pharmacokinetic results. J Clin Oncol 2003;21(15):281522. [8] Beerepoot LV, et al. Phase I clinical evaluation of weekly administration of the novel vascular-targeting agent, ZD6126, in patients with solid tumors. J Clin Oncol 2006;24(10):14918. [9] Blakey DC, et al. Antitumor activity of the novel vascular targeting agent ZD6126 in a panel of tumor models. Clin Cancer Res 2002;8(6):197483. [10] Dowlati A, et al. A phase I pharmacokinetic and translational study of the novel vascular targeting agent combretastatin a-4 phosphate on a singledose intravenous schedule in patients with advanced cancer. Cancer Res 2002;62(12):340816. [11] Pellegrini F, Budman DR. Review: tubulin function, action of antitubulin drugs, and new drug development. Cancer Invest 2005;23(3):26473. [12] Stark GR, Taylor WR. Control of the G2/M transition. Mol Biotechnol 2006;32(3):22748. [13] Izumi T, Maller JL. Elimination of cdc2 phosphorylation sites in the cdc25 phosphatase blocks initiation of M-phase. Mol Biol Cell 1993;4(12):133750. [14] Lopez-Girona A, et al. Nuclear localization of Cdc25 is regulated by DNA damage and a 14-3-3 protein. Nature 1999;397(6715):1725. [15] Blangy A, et al. Phosphorylation by p34cdc2 regulates spindle association of human Eg5, a kinesin-related motor essential for bipolar spindle formation in vivo. Cell 1995;83(7):115969. [16] Valentine MT, Fordyce PM, Block SM. Eg5 steps it up! Cell Div 2006;1:31. [17] Li W, et al. BUBR1 phosphorylation is regulated during mitotic checkpoint activation. Cell Growth Differ 1999;10(11):76975. [18] Chan GK, et al. Human BUBR1 is a mitotic checkpoint kinase that monitors CENP-E functions at kinetochores and binds the cyclosome/ APC. J Cell Biol 1999;146(5):94154. [19] Gillett ES, Sorger PK. Tracing the pathway of spindle assembly checkpoint signaling. Dev Cell 2001;1(2):1624. [20] Dowling M, et al. Mitotic spindle checkpoint inactivation by trichostatin a defines a mechanism for increasing cancer cell killing by microtubuledisrupting agents. Cancer Biol Ther 2005;4(2):197206.

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031

ARTICLE IN PRESS
8 D. Durrant et al. / Gynecologic Oncology xx (2008) xxxxxx [34] Loffler H, et al. Chk1-dependent regulation of Cdc25B functions to coordinate mitotic events. Cell Cycle 2006;5(21):25437. [35] Kaneko YS, et al. Cell-cycle-dependent and ATM-independent expression of human Chk1 kinase. Oncogene 1999;18(25):367381. [36] Zachos G, et al. Chk1 is required for spindle checkpoint function. Dev Cell 2007;12(2):24760. [37] Clark AS, et al. Constitutive and inducible Akt activity promotes resistance to chemotherapy, trastuzumab, or tamoxifen in breast cancer cells. Mol Cancer Ther 2002;1(9):70717. [38] Mabuchi S, et al. Inhibition of phosphorylation of BAD and Raf-1 by Akt sensitizes human ovarian cancer cells to paclitaxel. J Biol Chem 2002;277 (36):33490500. [39] Blagosklonny MV, et al. Taxol-induced apoptosis and phosphorylation of Bcl-2 protein involves c-Raf-1 and represents a novel c-Raf-1 signal transduction pathway. Cancer Res 1996;56(8):18514. [40] Kotha A, et al. Resveratrol inhibits Src and Stat3 signaling and induces the apoptosis of malignant cells containing activated Stat3 protein. Mol Cancer Ther 2006;5(3):6219. [41] Tang J, Erikson RL, Liu X. Checkpoint kinase 1 (Chk1) is required for mitotic progression through negative regulation of polo-like kinase 1 (Plk1). Proc Natl Acad Sci U S A 2006;103(32):119649. [42] Nabha SM, et al. Combretastatin-A4 prodrug induces mitotic catastrophe in chronic lymphocytic leukemia cell line independent of caspase activation and poly(ADP-ribose) polymerase cleavage. Clin Cancer Res 2002;8(8): 273541. [43] Lagouge M, et al. Resveratrol improves mitochondrial function and protects against metabolic disease by activating SIRT1 and PGC-1alpha. Cell 2006;127(6):110922. [44] Shaked Y, et al. Therapy-induced acute recruitment of circulating endothelial progenitor cells to tumors. Science 2006;313(5794):17857. [45] Durrant D, et al. cis-3, 4, 5-trimethoxy-3-aminostilbene disrupts tumor vascular perfusion without damaging normal organ perfusion. Cancer Chemother Pharmacol in press. [21] Castedo M, et al. Cell death by mitotic catastrophe: a molecular definition. Oncogene 2004;23(16):282537. [22] Niida H, Nakanishi M. DNA damage checkpoints in mammals. Mutagenesis 2006;21(1):39. [23] Zhang S, Hemmerich P, Grosse F. Centrosomal localization of DNA damage checkpoint proteins. J Cell Biochem 2007;101(2):45165. [24] Castedo M, et al. The cell cycle checkpoint kinase Chk2 is a negative regulator of mitotic catastrophe. Oncogene 2004;23(25):435361. [25] Roberti M, et al. Synthesis and biological evaluation of resveratrol and analogues as apoptosis-inducing agents. J Med Chem 2003;46(16): 354654. [26] Cao, TM, et al., Stilbene derivatives that are colchicine site microtubule inhibitors have anti-leukemic activity and minimal systemic toxicity. Am J Hematol in press. [27] Kroemer G, Reed JC. Mitochondrial control of cell death. Nat Med 2000;6(5):5139. [28] Desagher S, Martinou JC. Mitochondria as the central control point of apoptosis. Trends Cell Biol 2000;10(9):36977. [29] Harris MH, et al. Role of oxidative phosphorylation in Bax toxicity. Mol Cell Biol 2000;20(10):35906. [30] Ricci JE, et al. Disruption of mitochondrial function during apoptosis is mediated by caspase cleavage of the p75 subunit of complex I of the electron transport chain. Cell 2004;117(6):77386. [31] Kowaltowski AJ, Vercesi AE. Reactive oxygen generation by mitochondria. In: Lemasters JJ, Neiminen A, editors. Mitochondria in pathogenesis. New York: Kluwer Academic/Plenum Publishers; 2001. p. 281300. [32] Karimi M, et al. Silencing human NKG2D, DAP10, and DAP12 reduces cytotoxicity of activated CD8+ T cells and NK cells. J Immunol 2005;175 (12):781928. [33] Demasters G, et al. Potentiation of radiation sensitivity in breast tumor cells by the vitamin D3 analogue, EB 1089, through promotion of autophagy and interference with proliferative recovery. Mol Cancer Ther 2006;5(11):278697.

Please cite this article as: Durrant D, et al, Mechanism of cell death induced by cis-3, 4, 5-trimethoxy-3-aminostilbene in ovarian cancer, Gynecol Oncol (2008), doi:10.1016/j.ygyno.2008.02.031