Induction of Apoptosis in Human Breast Cancer Cells via Caspase Pathway by Vernodalin Isolated from Centratherum anthelminticum (L.) Seeds

Background Centratherum anthelminticum (L.) seeds (CA) is a well known medicinal herb in Indian sub-continent. We recently reported anti-oxidant property of chloroform fraction of Centratherum anthelminticum (L.) seeds (CACF) by inhibiting tumor necrosis factor-α (TNF-α)-induced growth of human breast cancer cells. However, the active compounds in CACF have not been investigated previously. Methodology/Principal Findings In this study, we showed that CACF inhibited growth of MCF-7 human breast cancer cells. CACF induced apoptosis in MCF-7 cells as marked by cell size shrinkage, deformed cytoskeletal structure and DNA fragmentation. To identify the cytotoxic compound, CACF was subjected to bioassay-guided fractionation which yielded 6 fractions. CACF fraction A and B (CACF-A, -B) demonstrated highest activity among all the fractions. Further HPLC isolation, NMR and LC-MS analysis of CACF-A led to identification of vernodalin as the cytotoxic agent in CACF-A, and -B. 12,13-dihydroxyoleic acid, another major compound in CACF-C fraction was isolated for the first time from Centratherum anthelminticum (L.) seeds but showed no cytotoxic effect against MCF-7 cells. Vernodalin inhibited cell growth of human breast cancer cells MCF-7 and MDA-MB-231 by induction of cell cycle arrest and apoptosis. Increased of reactive oxygen species (ROS) production, coupled with downregulation of anti-apoptotic molecules (Bcl-2, Bcl-xL) led to reduction of mitochondrial membrane potential (MMP) and release of cytochrome c in both human breast cancer cells treated with vernodalin. Release of cytochrome c from mitochondria to cytosol triggered activation of caspase cascade, PARP cleavage, DNA damage and eventually cell death. Conclusions/Significance To the best of our knowledge, this is the first comprehensive study on cytotoxic and apoptotic mechanism of vernodalin isolated from the Centratherum anthelminticum (L.) seeds in human breast cancer cells. Overall, our data suggest a potential therapeutic value of vernodalin to be further developed as new anti-cancer drug.


Introduction
Breast cancer is one of the most common malignancies in women. Global breast cancer incidence has increased at an annual rate of 3.1% over the last three decades to more than 1.6 million cases in year 2010 [1]. In Malaysia, breast cancer is the most common cancer among females. There were 3,242 female breast cancer cases diagnosed in 2007, accounted for 18.1% of all cancer cases reported and 32.1% of all female cases (National Cancer Registry Report 2007). Different subtypes of breast cancers arise from different gene mutations occurring in luminal or basal progenitor cell population, causing difficulty in breast cancer diagnosis and treatment [2]. Being both genetically and histopathologically heterogeneous, the mechanisms underlying breast cancer development remains uncertain [3]. Owing to this, conventional chemotherapy, surgery or radiation shows very limited effects. On the other hand, specific natural or synthetic chemical compounds have been widely applied for cancer chemoprevention to inhibit or revert carcinogenesis and to suppress the malignancy of cancer [4].
Medicinal plants have been used for centuries to treat a variety of diseases and maintain health before the advent of modern medicine [5,6]. The accumulation and developing knowledge of the medicinal properties of plants by personal experimentation, local custom, anecdote, and folk tradition leads to the formation of numerous traditional medical systems and therapies, including traditional Chinese medicine (TCM), Ayurvedic medicine, indigenous medicine, naturopathy and aromatherapy [7,8,9]. In modern medicine, plants have been a source for new anti-cancer drugs. For example, vinblastine was traditionally obtained from Catharanthus roseus, taxol was isolated from the bark of the Pacific yew tree Taxus brevifolia, camptothecin was isolated from the bark and stem of Camptotheca acuminata [10,11,12]. The advancement of technology such as gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS) have speed up the process of drug screening and discovery [13]. LC-MS is a highly sensitivity and selectivity method used in drug development at many different stages including profiling of secondary metabolites in plants, impurities detection, metabolic stability or degradant analysis [14,15].
Centratherum anthelminticum (L.) Kuntze, commonly known as kalajiri, somraj, black cumin or bitter cumin, is a robust leafy plant belongs to Asteraceae family of the flowering plants ( Figure 1). Scientific synonyms for this plant include Vernonia anthelmintica and Conyza anthelmintica. This plant can be found in India, Himalaya mountain, Khasi mountain, Sri Lanka, Afghanistan, and is widely used as a traditional herb against fever, cough and diarrhea in the region. Recent experimental analyses have proven that extracts from seeds of C. anthelminticum possess various pharmacological properties. The methanolic extract from the C. anthelminticum seeds demonstrates antiviral properties [16] whereas acetone and ethyl acetate extracts demonstrate antifilarial activity against Setaria cervi [17]. Besides, petroleum ether and alcohol extracts show analgesic, antipyretic and anti-inflammatory effect in rat model [18,19].
Different extracts from C. anthelminticum seeds also show antimicrobial and antifungal properties when screened on various pathogens in vitro [20]. A recent report also suggests C. anthelminticum seeds phenols inhibit liposomal peroxidation and protect oxidative damage to genomic DNA of Bacillus, therefore can function as an anti-oxidant agent [21].
In 2004, Lambertini et al. reported the in vitro anti-proliferative effect of extracts from C. anthelminticum on human breast cancer cells [22]. We recently reported that the chloroform, but not hexane or methanol fractions from C. anthelminticum (L.) seeds (CACF) exhibited anti-oxidant property by inhibiting tumor necrosis factor-a (TNF-a)-induced human cancer cell growth by interrupting the activation of nuclear factor-kappa B (NF-kB) [23]. However, the active compounds in CACF were not examined in the previous reports. In this study, we showed that CACF inhibited MCF-7 breast cancer cell growth. Administration of CACF caused morphological changes, disrupted cytoskeletal structures and DNA fragmentation. Bioassay-guided fractionation led us to the identification of vernodalin as the cytotoxic agent in CACF. To the best of our knowledge, this is the first report on the cytotoxic and the apoptotic mechanism of vernodalin isolated from C. anthelminticum seeds in human breast cancer cells.

Plant Material
The seeds of C. anthelminticum were purchased from the medicinal plant cultivation zone of Amritum Bio-Botanica Herbs Research Laboratory Pvt. Ltd, Betul Madhya Pradesh India. The seeds were identified by the quality control department of the company itself. Voucher specimen (CA-9) was deposited at the Pharmacology Department of University Malaya, Malaysia.

Cell Culture
The human breast cancer cell line, MCF-7 was purchased from Cell Lines Service (300273; Eppelheim, Germany) and MDA-MB-231 cell line was obtained from American Type Culture Collection (HTB-26; ATCC, Manassas, VA). Human mammary epithelial cells were purchased from ScienCell (7610; Carlsbad, CA) and maintained in mammary epithelial cell medium (ScienCell). MCF-7 and MDA-MB-231 cells were grown in Dulbecco's Modified Eagle Medium (DMEM, Life Technologies, Inc, Rockville, MD) supplemented with 10% heat-inactivated fetal bovine serum (Sigma-Aldrich, St. Louis, MO), 2 mM glutamine, 1% penicillin and streptomycin. Cells were cultured in tissue culture flasks (Corning, USA) and were kept in CO 2 incubator at 37uC in a humidified atmosphere with 5% CO 2 . For experimental purposes, cells in exponential growth phase (approximately 70-80% confluency) were used.

Extraction and Isolation
The powdered seeds of C. anthelminticum (100 g) were extracted successively with hexane (36250 ml) (Merck, Darmstadt, Germany), chloroform (CHCl 3 ) (36250 ml) (Merck, Darmstadt, Germany) and methanol (MeOH) (36250 ml) (Merck, Darmstadt, Germany), in a Soxhlet apparatus for 24 hours. The resultant extracts were filtered using Whatman No. 1 filter paper (Whatman, England) and dried under vacuum to yield 20.1, 7.7, 11.6 g, respectively of the extracts. Then the dried fractions were kept at 220uC until further use. In our previous paper, the chloroform extract (CACF) showed highest activity on MTT assay, therefore CACF was chosen for this study [23].
The cytotoxic active fractions CACF-A, CACF-B and CACF-C were then further purified using either preparative high performance liquid chromatography (HPLC) (Gilson GX-281/322 system) using a Waters Novapak C 18 column (256100 mm, 6 mm) or by recrystallisation. The major active compound of the fractions CACF-A and CACF-B were obtained by preparative  HPLC. The fractions were eluted at a flow rate of 12 ml per min over 75 min. The gradient started at 10% solvent (A) (acetonitrile with 0.1% formic acid) and 90% solvent B (water with 0.1% formic acid) for 5 min. The gradient then changed from 10% to 60% (A) over 5 minutes, followed by 60-100% (A) for 50 minutes and finally an isocratic elution of 100% (A) from 60 to 75 min. The fractions yielded 10 mg of compound (1) (colorless oil) eluted at about 30 min. Fraction CACF-C was purified by recrystalliza-

Identification of Compounds
The 1 H and 13 C Nuclear Magnetic Resonance (NMR) spectroscopy were carried out on a JOEL NMR 400 and 100 MHz with TMS as internal standard. HR-ESI-MS was measured on a LCMS-IT-TOF mass spectrometer (Shimadzu IT-TOF). MS was recorded on Shimadzu GC-MS model QP2010 Plus spectrophotometer.

Spectral Data
Vernodalin (1) was obtained as colorless oil. The 1 H and 13 C NMR spectral data obtained for CDCl 3 solution of the compound were in agreement to those previously identified as vernodalin [24]. The positive ion HR-ESI-MS spectrum showed a molecular ion peak at m/z 361 [M+H] + consistent to molecular formula C 19   MTT Cell Viability Assay 1610 4 cells per well were seeded into 96-well plate overnight. Cells were treated with various concentrations of compound or extract (dissolved in dimethyl sulfoxide, DMSO) for 24 hours. As negative control, cells were treated with vehicle (DMSO) only. Next, cells were incubated with 50 ml of 4,5-dimethylthiazol-2-yl-2,5-diphenyltetrazolium bromide (MTT) (2 mg/ml) at 37uC for 2 hours. After dissolving the formazan crystals in DMSO, plates were read in Chameleon TM multitechnology microplate reader (Hidex, Turku, Finland) at 570 nm against 620 nm. This experiment was performed in triplicates and repeated for 3 times. Mean values 6 SD for each concentration was determined. Calculation of cell viability was described previously with slight modification [25]. Cell viability (in percentages, %) was showed as ratio of absorbance (A 570 nm ) in treated cells relative to absorbance in control cells (DMSO) (A 570 nm ). The IC 50 was defined as the concentration of sample needed to reduce 50% of absorbance relative to the vehicle (DMSO)-treated control.

Real Time Cell Growth Assay
Cell proliferation was measured using xCELLigence Real-Time Cellular Analysis (RTCA) system (Roche, Germany), which allows us to monitor the cell viability and cell growth continuously at multiple time point. Briefly, background measurements were taken after adding 50 ml of the culture medium to the wells. Next, cells were seeded at density 1610 4 on a specialized 16-well plate with electrodes for 18 hours to allow cells grow to the log phase. Cells were treated with 100 ml of CACF or vernodalin in various concentrations (mg/ml) dissolved in cell culture media and continuously monitored for up to 72 hours. Cell sensor impedance was expressed as an arbitrary unit called the Cell Index. Cell index were recorded every 5-10 minutes by RTCA analyzer. To eliminate variation between wells, cell index values were normalized to the value at the beginning of treatment time-point.

Real Time Cell Invasion Assay
The kinetics of cell invasion was assayed using the xCELLigence Real-Time Cell Analyzer (RTCA DP; Roche). CIM-plates (Roche) were pre-coated with 30 ml of matrigel (BD Biosciences) diluted 1:10 in DMEM for 1 h at 37uC. The upper chambers contained pre-warmed serum-free DMEM, whereas the lower chambers contained either DMEM with 10% FCS or DMEM medium only (negative control). Indicated concentrations of vernodalin were added into the medium of upper and lower chambers. 1610 4 MDA-MB-231 cells were seeded into each well of the upper chambers. Cells were allowed to settle for 30 min at room temperature before being placed back to the RTCA DP in a humidified incubator at 37uC with 5% CO 2 . Readings were

Apoptosis Assay
For in vitro fluorescent staining, 1610 4 cells per well were seeded in 96 well-plate overnight. Cells were then treated with CACF at various concentrations for 12 hours. Live cells were stained with FITC-annexin V (BD Biosciences, San Jose, CA) for 15 minutes before fixed with 4% paraformaldehyde. Cells were washed 3 times with PBS and the fluorescent images were acquired using Cellomics ArrayScan high content screening (HCS) reader (Thermo Scientific, Pittsburgh, PA). Compartmental analysis bioapplication module was used to quantify the fluorescence intensity of FITC-annexin V.
For apoptosis assay by flow cytometry, cells were seeded at 1610 5 per ml on 25 cm 2 flask overnight before treated with vernodalin at various concentrations for 24 hours. Determination of apoptotic cells by fluorescent staining was done as described previously [25]. Briefly, cells were incubated with FITC-annexin V and propidium iodide (PI) (BD Biosciences) in binding buffer for 15 minutes in dark. Stained cells were immediately subjected to flow cytometry analyses using FACS Canto II flow cytometer (BD Biosciences).
Cytoskeletal Rearrangement Analysis 1610 4 MCF-7 cells per well seeded overnight in 96-well plate were exposed to DMSO (negative control) or CACF at various concentrations for 12 hours. Cells were fixed, washed with wash buffer before probed using phalloidin conjugated with DyLight TM 554 and Hoechst 33258 according to the manufacturer's instruction. Cells were visualized and images were acquired using Cellomics ArrayScan HCS reader (Thermo Scientific). Morphology bioapplication module was used to quantify the fluorescence intensity of phalloidin. Reactive Oxygen Species (ROS) Analysis 1610 4 cells per well were seeded onto 96-well plate. Cells were treated with vernodalin or DMSO (negative control) at indicated concentrations for 12 hours. Dihydroethidium (DHE) dye contained in Cellomics ROS kit was added into live culture for 30 minutes. Cells were fixed and washed with wash buffer as described by the manufacturer's instruction. Stained cells were visualized and acquired using Cellomics ArrayScan HCS reader (Thermo Scientific). Target activation bioapplication module was used to quantify the fluorescence intensities of DHE dye in the nucleus. Nuclear Morphology, Membrane Permeability, Mitochondrial Membrane Potential Dym (MMP) and Cytochrome C Release Analysis Cellomics Multiparameter Cytotoxicity 3 Kit (Thermo Scientific) was used. Cells were plated at 1610 4 cells per well on 96-well plate overnight. DMSO (solvent) or vernodalin was added at various concentrations and further incubated for 24 hours. MMP dye (Excitation 552/Emission 576) and the cell permeability dye (Excitation 491/Emission 509) were added to live cells and incubated for 1 hour. Cells were fixed with 4% formaldehyde for 15 minutes. Fixed cells were permeabilized with 0.1% Triton X-100 in phosphate buffer saline (PBS). Samples were blocked with 3% bovine serum albumin and incubated with cytochrome c primary mouse antibody for 1 hour. Samples were washed three times with wash buffer I (16PBS) before addition of goat antimouse secondary antibodies conjugated with DyLight TM 649. Cells were rinsed three times with wash buffer II (16PBS with 1% Tween-20). Nucleus was stained with Hoechst 33258. Stained cells were visualized and images were captured using Cellomics ArrayScan HCS reader (Thermo Scientific). Cell health profiling bioapplication module was used to quantify the fluorescence intensities of each dye.

Bioluminescent Assays for Caspase-3/7,-8 and -9 Activities
A time-dependent study of caspase-3/7, -8 and -9 activities was performed in triplicates using assay kits Caspase-GloH 3/7, 8 and 9 (Promega, Madison, WI) on a white 96-well microplate. A total of 1610 4 cells was seeded per well and incubated with 100 ml of vernodalin (final concentration 6.25 mg/ml) for 1, 3, 6, 12, 18, 24 and 30 hours. Caspase activities were investigated according to the manufacture protocol. Briefly, 100 ml caspase-Glo reagent was added and incubated at room temperature for 30 minutes. Presences of active caspases from apoptotic cells cleaved the aminoluciferin-labeled synthetic tetrapeptide thus release substrate for the luciferase enzyme. The caspase activities were measured using a Tecan InfiniteH200 Pro (Tecan, Mä nnedorf, Switzerland) microplate reader.

Statistical Analysis
Experimental values were expressed as the means 6 standard deviation (SD) of the number of experiments indicated in the legends. Analysis of variance (ANOVA) was performed using GraphPad Prism 5 software. Statistical significance was defined when P,0.05.

CACF Inhibits Survival of Human Breast Cancer MCF-7 Cells
We first determined the cytotoxic effect of CACF on cell survival using a well-characterized human breast cancer cell line, MCF-7. MTT assay was used to determine cell viability. The survival of MCF-7 decreased significantly in a concentration dependent manner with IC 50 value at 6.861.2 mg/ml (Figure 2A). No significant cell inhibitory effect was observed in DMSO (solvent)-treated samples. As a positive control, we treated MCF-7 cells with doxorubicin, a cancer chemotherapy drug, which showed IC 50 value at 2.060.8 mg/ml. To verify MTT results, we repeated the experiments using Alamar blue staining for cell viability. We found comparable results between MTT and Alamar blue staining assays ( Figure S1A).
MTT assays are end point assays which only detect cell viability at certain time-point. Next, we observed the subtle changes or the pattern of cell growth after CACF-treatment for 3 consecutive days using real-time cell proliferation assay (RTCA). In control wells (vehicle, DMSO only), we observed an exponential increased of cell growth as reflected by increased in normalized cell index (nCI) values. Whereas MCF-7 treated with doxorubicin at concentration 6.25 mg/ml resulted in cell growth inhibition ( Figure 2B). A dose-dependent attenuation of cell proliferation was observed in CACF-treated MCF-7 ( Figure 2B). As shown in Figure 2B, we observed a sudden decrease in nCI values about 1-2 hours after treated with 25 or 12.5 mg/ml of CACF, indicating acute toxicity at high dosages. Together, our results showed that CACF inhibited cell growth of MCF-7 breast cancer cells in doseand time-dependent manners.

Morphological Assessment of CACF-treated MCF-7 Cells
Next, we examined if CACF treatment resulted in cell death through apoptotic pathway. We treated MCF-7 cells with control (DMSO solvent) or CACF for 12 hours before staining live cells with apoptosis marker annexin V conjugated to FITC. Exposure of 6.25 and 12.5 mg/ml of CACF led to higher annexin V staining compared to control, suggesting apoptotic activities ( Figure 3A).
Under light miscroscope, we observed that MCF-7 cells exposed to CACF resulted in reduction of cell size and cell-cell contact areas ( Figure 3B). To further investigate this, we examined cytoskeletal F-actin structure by staining the cells with phalloidin conjugated to DyLight TM 554, which detect polymerized actin (Factin). Control cells demonstrated well-organized actin filament bundles or stress fibers in the cytoplasm ( Figure 3B). On the contrary, CACF treatment (6.25 and 12.5 mg/ml) on MCF-7 cells caused a drastic reduction in phalloidin stain ( Figure 3B) and loss of stress fibers in the cytoplasm. Furthermore, F-actin was no longer distributed evenly at the cell periphery, but appeared as punctuate stain at the plasma membrane. This result suggests that CACF treatment led to the disruption of cytoskeletal structure in MCF-7 cells.
Because apoptotic activity is usually associated with DNA cleavage, we examined the effect of CACF on nuclear morphology of MCF-7 cells using Hoechst 33258. After CACF treatment for 24 hours, a population of condensed and fragmented nuclei was observed ( Figure 3C). The number of cells with fragmented nuclei increased with higher dosages of CACF administrated while no detectable DNA damage was detected in control cells. Together, these data indicated that CACF treatment triggered apoptotic pathway as evidenced by higher annexin V staining, cell shrinkage, disrupted cytoskeleton and DNA damage in MCF-7 cells.

Cytotoxic Activities of CACF Fractions
To identify the active compound in CACF extracts which possesses cytotoxicity activity against human breast cancer cells, we performed HPLC analysis. The cytotoxic extract CACF was fractionated into six fractions using preparative HPLC ( Figure 4A and 4B). Among these fractions, CACF-A and CACF-B showed highest activity on MCF-7 cells with IC 50 values of 5.860.6 mg/ml and 5.560.3 mg/ml, respectively. However, CACF-C showed moderate activity on MCF-7 cells with IC 50 value of 38.261.6 mg/ml while the other fractions (CACF-D to F) exhibited low cytotoxic activity (IC 50 .100 mg/ml). Compounds in fractions CACF-A, -B and -C were further isolated as described in Materials and Methods.

Vernodalin Inhibits Cell Growth of MCF-7 and MDA-MB-231 Cells
To examine the in vitro anti-cancer efficacy of vernodalin, we included a highly invasive and metastatic variant of human breast cancer cell-line, MDA-MB-231, apart from MCF-7 (non-metastatic). Both MCF-7 and MDA-MB-231 cells were exposed to various concentrations of vernodalin for 24 hours. Cell viability was determined by MTT assays. The IC 50 values for vernodalin treated MCF-7 and MDA-MB-231 were 2.560.3 mg/ml and 3.460.6 mg/ml, respectively ( Figure 7A). On the other hand, the IC 50 of normal mammary epithelial cells was 12.760.5 mg/ml, relatively more resistant to cell killing by vernodalin. We further verified the results using Alamar blue proliferation assay ( Figure  S1B). Meanwhile, 12,13-dihydroxyoleic acid, another compound isolated showed no cytotoxicity effect on both breast cancer cell lines, IC 50 .100 mg/ml (data not shown).
Next, we monitored real-time vernodalin-mediated cell growth inhibition for 3 consecutive days by RTCA. As shown in Figure 7B and 7C, a dose-dependent cell-growth inhibition was observed in vernodalin-treated MCF-7 and MDA-MB-231 cells. Reduced nCI was observed at concentration 3.125 mg/ml, whereas no significant increment in nCI at 6.25 mg/ml or higher after vernodalin treatment compared to control ( Figure 7B and 7C). These results showed that vernodalin inhibited cell growth of breast cancer cells, MCF-7 and MDA-MB-231 in a dose-and time-dependent manner.

Vernodalin Inhibits Invasive Potential of Metastatic Breast Cancer Cells
Next, we examined whether vernodalin has an influence on the invasive potential of metastatic breast cancer cells MDA-MB-231 by using modified Boyden chamber (CIM plates, Roche) with matrigel as a substrate. The presence of chemoattractant FCS strongly induced non-treated MDA-MB-231 cell migration to the lower chamber. In contrast, no signal or cell invasion was detected in lower chamber without addition of FCS. On the other hand, dose-dependent inhibition of cell invasion was observed upon vernodalin treatment (Figure 8). Of note, we previously showed that CACF treatment resulted in reduced phalloidin staining ( Figure 3B), implying that vernodalin-mediated cell invasion inhibition was probably mediated through disruption of actin polymerization.

Vernodalin causes Apoptosis and Cell Cycle Arrest
To examine whether cells undergo apoptosis, untreated or vernodalin-treated MCF-7 and MDA-MB-231 breast cancer cells were stained with annexin V and PI. Flow cytometry analysis of stained cells can distinguish cells into four groups, namely viable (annexin V-PI-), early apoptosis (annexin V+ PI-), late apoptosis (annexin V+ PI+) and necrotic (annexin V-PI+) cells. As shown in Figure 9A and 9B, vernodalin exposure at different concentrations (3.125, 6.25 and 12.5 mg/ml) resulted in higher population of early apoptotic population (30. Next we examined the cell cycle distribution by staining vernodalin treated breast cancer cells with propidium iodide and analyzed the percentages of G0/G1, S, G2/M cell population using flow cytometry. MCF-7 cells treated with 6.125 or 12.5 mg/ ml vernodalin showed higher G0/G1 population (72.5611.7% and 71.664.9%, respectively) compared with 64.862.6% in the control (Figure 10). MDA-MB-231 cells treated with 6.125 or

Vernodalin Induces ROS Generation
ROS is produced especially when cells undergo chemical or environmental stress and could be one of the causative factors leading to cell cycle arrest or apoptosis. Next, we examined the ROS level in control or vernodalin-treated breast cancer cells by staining with DHE dye. ROS convert non-fluorescent DHE to fluorescent ethidium, which then intercalates into DNA. Hoechst 33258, a DNA binding dye is used to identify the nuclei of individual cells then the DHE fluorescence is quantified using the Cellomic HCS machine to evaluate the oxidative stress level. As shown in Figure 11A and 11B, ROS production was at the basal level in control DMSO-treated MCF-7 or MDA-MB-231 cells. In contrast, treatment with vernodalin (12 hours) resulted in dosedependent increased of ROS production as shown by increased DHE staining in the nucleus ( Figure 11A and 11B).

Effect of Vernodalin on Nuclear Morphology, Membrane Permeabilization, MMP (Dym) and Cytochrome c Release
Since high ROS production could lead to plasma membrane, DNA, mitochondrial damage, we further examined the nuclear morphology, membrane permeability, mitochondrial membrane potential (MMP, Dym) and cytochrome c release and localization. As shown in Figure 12A and 12B, 24 hours of exposure to vernodalin revealed a concentration-dependent increment of membrane permeability, attenuation of MMP and increased cytochrome c in the cytosol compared to control. In some vernodalin-treated cells, cytochrome c was localized in the nucleus and we could observe nuclear condensation and fragmentation in these cells ( Figure 12A).Whereas in control samples, nucleus remained rounded and uniform in size. Moreover, plasma membrane was intact as shown by the weak staining of permeability dye (green, Figure 12A). Cytochrome c (cyan) was distributed homogenously in the cytosol, which colocalized with MMP dye (red), indicating that cytochrome c was not released from the mitochondria in control cells ( Figure 12A).

Effect of Vernodalin Treatment on Caspase-3/7, -8, -9
Apoptosis is a complex activity that mobilizes a number of molecules and is classified into caspase-dependent or caspaseindependent mechanisms. Caspase-dependent pathway can be further divided into extrinsic or intrinsic pathway, as determined by involvement of caspase-8 or caspase-9, respectively. Both intrinsic and extrinsic pathway involved activation of caspase-3/7 which is important for inducing downstream DNA cleavage molecules. To examine the molecular mechanism underlying apoptosis process, we stained cells with aminoluciferin-labeled substrate of caspase and determined the caspase-3/7, -8, -9 activities by measuring the luminescence intensities every three hours. As shown in Figure 13A, we observed a gradual increased of caspase-9 and caspase-3/7 activity, which peaked at 18 hours in both MCF-7 and MDA-MB-231 cells treated with 6.25 mg/ml of vernodalin ( Figure 13A and 13B). The activity of caspase-3/7 increased significantly from 6 to 12 hours, but remained high even after 30 hours of treatment, indicating a more latent effect of vernodalin in MDA-MB-231 cells ( Figure 13B). In contrast, there were no significant changes in the activity of caspase-8 for the time span of 30 hours in vernodalin treated-MCF-7 or MDA-MB-231 cells. Our data suggested that vernodalin induced activation of intrinsic caspase pathway in both breast cancer cell lines. To examine whether caspases and the downstream PARP molecule were involved in vernodalin-induced apoptosis, we performed Western blot analysis using cell lysates of untreated/ vernodalin-treated MCF-7 or MDA-MB-231 cells. Results indicated that vernodalin dose-dependently caused cleavage of caspase-7 and -9 in MCF-7, whereas caspase-3, -7 and -9 were activated in MDA-MB-231 cells ( Figure 13C). On the other hand, PARP cleavage was also detected in both cells, suggesting involvement of caspase cascade and PARP inactivation in vernodalin-mediated apoptosis ( Figure 13C).

Vernodalin Downregulates Anti-apoptotic Molecules
Cell survival is maintained by pro-survival (anti-apoptotic) molecules such as Bcl-2 and Bcl-xL. To examine if the vernodalin initiated apoptosis by affecting the cellular level of these molecules, we performed Western blot analysis using control or vernodalintreated breast cancer cells. Cells were also treated with a standard drug doxorubicin as a positive control of apoptosis induction. Our data showed that vernodalin dose-dependently reduced the expression level of Bcl-2 and Bcl-xL ( Figure 14).

Discussion
In this study, chloroform extract of the seed of C. anthelmintica (CACF) was tested for its cytotoxicity effect on MCF-7 human breast cancer cells. We showed that CACF effectively inhibited cell growth of MCF-7. CACF-treated cells exhibited morphological hallmarks of apoptosis including cell shrinkage, lost of cytoskeletal F-actin structure, higher stain with apoptotic marker annexin V and DNA damage. In fact, defects in apoptotic pathway are thought to contribute to a number of human malignancies [27]. Thus, anti-cancer agents that induce apoptosis is one of the efficient strategies in cancer chemotherapy [28].
Bioassay-guided isolation is a procedure whereby extract is chromatographically fractionated until a pure and active compound is isolated. Each fraction produce during the process is evaluated in bioassay system (e.g. in vitro cell culture, in vivo rat or zebrafish model etc) and the active fractions are chosen for further purification [29,30,31]. In this study, bioassay guided isolation of CACF extract led us to the identification of a potent compound vernodalin. Vernodalin is a sesquiterpene lactone isolated from various plant species including the seeds of C. anthelmintica [32]. Vernodalin exhibits anti-malarial and anti-bacterial activities and is a constituent of Vernonia amygdalina (Compositae), a plant ingested by wild chimpanzees sometimes suffering from parasite-related diseases in the Mahale Mountains National Park, Tanzania [33,34,35]. So far, only two studies have been done on anti-cancer effect of vernodalin, which demonstrated cytotoxic activity on melanoma and ovarian cancer cell lines [36] and human carcinoma of the nasopharynx (KB) [37]. These two studies mainly examined cell viability of vernodalin treated cancer cells by MTT based assays and less information was available on anticancer mechanism by vernodalin. To the best of our knowledge, this is the first report on cytotoxicity and mechanism of vernodalin on human breast cancer cells.
12,13-dihydroxyoleic acid (2), a compound isolated from CACF-C (a moderate active fraction) did not show inhibitory activity on breast cancer cells. The moderate activity of CACF-C could be due to the presence of trace quantity of vernodalin ( Figure 5B). The structure of the compound was confirmed by comparing its spectral data (MS, 1 H NMR, 13 C NMR) with those reported for a synthetic 12,13-dihydroxyoleic acid [38]. The MS spectrum ( Figure S2A) showed ions at m/z 182 and 131 corresponding to the allylic cleavage and indicated the double bond at C 9 and C 10 ( Figure S2B). Fragmentation ions at m/z 213 and 157 corresponding to the alpha cleavage on the either side of the OH groups which confirmed their positions at C 12 and C 13 [38,39]. A characteristic peak at m/z 85 (100%) as a result of ions formed through C 4 and C 5 cleavage and losing two protons form [C 4 H 7 O 2 N ]. This is the first time for the compound (2) to be reported as secondary metabolite in plant. 12,13-dihydroxyoleic acid (2) is normally synthesized by acetolysis of vernolic acid, an epoxy fatty acid obtained from C. anthelmintica oil [40]. Hydroxy fatty acids are important in industry for the production of oleochemicals [38,39].
We showed that vernodalin induced cell cycle arrest in breast cancer cells. Cell cycle progression is a hallmark for cell proliferation. Deregulation of cell cycle has been linked with cancer initiation and progression [41]. Thus, cell cycle has emerged as one of the attractive therapeutic target in the treatment of cancer. Nevertheless, siRNA or small molecule inhibitors that target cell cycle have been developed, for example flavopiridol is the first cell cycle inhibitor to be tested in clinical trials [42,43]. To date, most of the chemotherapeutic agents caused cell cycle arrest either at G0/G1 or the G2/M stage, whereas cell cycle arrest at the S-phase is rare. For instance, synthetically methoxylated analogue of resveratrol induces G1 cell cycle arrest of human breast carcinoma MCF-7 cells [44] whereas curcumin induced G2/M cell cycle arrest in cisplatin resistant ovarian cancer cells [45]. The cell cycle is controlled by a group of cyclin family proteins called cyclin-dependent kinase (CDKs) enzymes [46]. The regulation of CDKs activities is achieved by their association with cyclin partners and kinases, phosphatases and specific inhibitors [46]. Future works are needed to examine the detail mechanism of cell cycle arrest in vernodalin-treated breast cancer cells.
ROS are either free radicals or reactive anions containing oxygen atoms, such as oxygen ions and peroxides. ROS could be a by-product of aerobic respiration, tissue-specific enzyme or microsomal cytochrome P450 metabolism of xenobiotic compounds [47]. High level of ROS can destroy the integrity of plasma membrane, affects dynamic of actin cytoskeleton and causes DNA damage, cumulatively known as oxidative stress [48,49,50]. Interestingly, we observed that ROS production in vernodalin-treated human breast cancer cells were 2-4 fold higher compared to control. Although harmful to cells, the anti-cancer effect of several conventional treatments such as ionizing radiation, etoposide or arsenates rely on their ability to stimulate ROS production, which modulate cellular redox balance leading to oxidative stress, destabilization of mitochondria and subsequently induction of apoptosis [51]. For example, etoposide caused severe ROS accumulation preferentially in the human glioblastomaastrocytoma cells and elevated ROS rendered these cells highly sensitive to cell death [52].
Studies have shown that mitochondria played a key role in the apoptotic process [53,54,55]. Changes in the MMP (Dym) increase the release of apoptogenic factors such as cytochrome c from the outer mitochondria membrane space into the cytosol. Released cytochrome c then form apoptosome with pro-caspase-9, apoptotic protease activating factor-1 (Apaf-1) and ATP, which in turn activates downstream apoptotic signal such as caspase-3/7 [56,57]. However, malignant tumor cells predominantly produce ATP through glycolysis rather than oxidation of pyruvate in mitochondria like most normal cells, a phenomena known as Warburg effect [53]. Hence, tumor mitochondria are less susceptible to mitochondria membrane permeabilization rendering them more resistant to mitochondrial pathway of apoptosis [53]. In this report, we showed that vernodalin induced attenuation of MMP (Dym) possibly through ROS production, which promote mitochondria membrane permeabilization and subsequent induction of apoptosis.
Caspases is a family of cysteine proteases that is divided into executioner caspases such as caspase-3 or -7, and initiator caspases, such as caspase-8 and -9 [57]. Initiator caspase-8 is known to be activated through extrinsic pathway, whereas caspase-9 is activated in the event of mitochondrial cytochrome c leakage [57]. Both initiator caspases can activate caspase-3 or -7, which commit cells to apoptosis [58]. Incubation with vernodalin causes a time-dependent activation of caspase-9, while caspase-8 activities remained at basal level. The increase in caspase-9 activity was concomitant with the increase in caspase-3/7 activity. These results suggest that vernodalin induced apoptosis via mitochondrial-dependent intrinsic pathway. Of note, caspase-3/7 cleaves several target proteins, one of which is DNA repair enzyme, PARP. Interestingly, DNA fragmentation was detected in vernodalin-treated MCF-7 and MDA-MB-231 cells. Since MCF-7 is deficient in caspase-3 expression, it is possible that DNA fragmentation could be mediated by activation of caspase-7 and PARP cleavage, as shown previously by other studies [59,60].
Members of the Bcl-2 family are major regulators of cell death or cell survival. Bcl-2 and Bcl-xL act as apoptosis inhibitors in the cells. Our data showed that vernodalin treatment reduced expression of pro-survival/anti-apoptotic proteins Bcl-2 and Bcl-xL, implying the relevance of Bcl-2 family proteins for breast cancer cell survival. Another study by Shimizu et al. highlighted the importance of Bc1-2 and Bcl-xL in protecting mitochondria against loss of function during apoptosis and some forms of necrotic cell death [61]. Presence of Bcl-2 in mitochondria blocks cell death by inhibiting apoptosis-associated release of cytochrome c from the mitochondria [62], or by regulating ion flux [63]. Whereas Bcl-xL interacts with Apaf1 to prevent apoptosis by inhibiting Apaf1 dependent activation of caspase 9 [64]. Therefore, downregulation of Bcl-2 and Bcl-xL upon vernodalin treatment could lead to loss of MMP which facilitated cytochrome c release and activation of caspase cascade.
Estrogen stimulates proliferation of various breast cancer cells via estrogen receptors (ER). Studies show that compounds such as phytoestrogens, alkylphenols, organochlorine pesticides and phthalates could bind to estrogen receptors and mediate estrogen responses [65,66], whereas polycyclic aromatic hydrocarbons (PAHs) or dioxin binds to aryl hydrocarbon receptor (AhR) which forms complex with ER [67]. To investigate whether the antiproliferative effect by vernodalin was ER-dependent, we performed TR-FRET assays to examine the binding ability of vernodalin to ER-a and ER-b. However, we did not find any significant reduction of TR-FRET signal by vernodalin even at the highest concentration, 200 mg/ml (data not shown). In fact, MTT or Alamar blue cell viability assays showed comparable IC 50 values between the two breast cancer cell-lines, MCF-7 (ER positive) and MDA-MB-231 (ER negative) 24 h after treatment with vernodalin (Figure 7 and S1B). Based on these findings, we propose that the presence of ER has no significant effect on vernodalin-induced cell growth inhibitory activity.
In conclusion, this report showed that CACF has profound activity against MCF-7 human breast cancer cell line. Through bioassay guided isolation, we identified vernodalin as the active compound responsible for the anti-cancer property in CACF. Our collective data suggest that vernodalin inhibits cell growth of MCF-7 and MDA-MB-231 breast cancer cells through induction of cell cycle arrest and apoptosis. Vernodalin induces apoptosis by generating ROS and downregulating pro-survival molecules Bcl-2 and Bcl-xL. These processes subsequently lead to attenuation of MMP and cytochrome c release. Release of cytochrome c activates caspase cascade and PARP cleavage to execute apoptotic program through fragmentation of chromatin DNA. The findings in this report indicate potential therapeutic value of vernodalin and further research in animal tumor models is necessary to confirm its anti-cancer activity in vivo.