Potent cytotoxic effects of Calomeria amaranthoides on ovarian cancers

Background Ovarian cancer remains the leading cause of death from gynaecological malignancy. More than 60% of the patients are presenting the disease in stage III or IV. In spite of combination of chemotherapy and surgery the prognosis stays poor for therapy regimen. Methods The leaves of a plant endemic to Australia, Calomeria amaranthoides, were extracted and then fractionated by column chromatography. In vitro cytotoxicity tests were performed with fractions of the plant extract and later with an isolated compound on ovarian cancer cell lines, as well as normal fibroblasts at concentrations of 1-100 μg/mL (crude extract) and 1-10 μg/mL (compound). Cytotoxicity was measured after 24, 48 and 72 hours by using a non-fluorescent substrate, Alamar blue. In vivo cytotoxicity was tested on ascites, developed in the abdomen of nude mice after inoculation with human OVCAR3 cells intraperitoneally. The rate of change in abdomen size for the mice was determined by linear regression and statistically evaluated for significance by the unpaired t test. Results Two compounds were isolated by chromatographic fractionation and identified by 1H-NMR, 13C-NMR and mass spectrometry analyses, EPD, an α-methylene sesquiterpene lactone of the eremophilanolide subtype, and EPA, an α-methylene carboxylic acid. Cytotoxicity of EPD for normal fibroblasts at all time points IC50 was greater than 10 μg/mL, whereas, for OVCAR3 cells at 48 hours IC50 was 5.3 μg/mL (95% confidence interval 4.3 to 6.5 μg/mL). Both, the crude plant extract as well as EPD killed the cancer cells at a final concentration of 10 μg/mL and 5 μg/mL respectively, while in normal cells only 20% cell killing effect was observed. EPA had no cytotoxic effects. Changes in abdomen size for control versus Cisplatin treated mice were significantly different, P = 0.023, as were control versus EPD treated mice, P = 0.025, whereas, EPD versus Cisplatin treated mice were not significantly different, P = 0.13. Conclusions For the first time both crude plant extract from Calomeria amaranthoides and EPD have been shown to have potent anti-cancer effects against ovarian cancer.


Background
Calomeria amaranthoides, described both by Ventenat and Smith in 1804 [1,2] as Humea elegans belonging to the genus Haeckeria in the tribe of Inuleae was grown in France and England from seeds originating from the Blue Mountains, New South Wales (NSW) in Australia. The plant is of a monotypic genus, endemic to NSW and Victoria, Australia [3].
In 2004 the genus Haeckeria was reassessed by Orchard as C. amaranthoides and since then C. amaranthoides belongs to the genus Calomeria of the family Asteraceae (Compositae) [4]. As a biennial plant it can grow to more than three metres high, with flowers as waving plume bushes and wrinkly leaves with an aromatic scent. It is also called incense plant.
The plant family of Asteraceae are known for their natural products. One type includes sesquiterpene lactones (SL) which to date is of great interest for their potential as anti-cancer agents as reviewed by Heinrich et al. and Zhang et al. [5,6].
Ovarian cancer is the fifth leading cause of death in women and remains the leading cause of death from gynaecological malignancy in many countries, in spite of chemotherapy with Platinum derivates and/or Taxol after surgery. Of the malignant epithelial tumors (>90% of all ovarian cancers), the serous papillary variants form the largest subgroup [7,8]. Due to its dismal prognosis there is an urgent need for new treatment strategy for ovarian cancer.
For the first time we have studied C. amaranthoides for its possible anti-tumor activity. An SL (EPD) and a structurally related sesquiterpene (EPA) have been found, extracted and purified. Among them EPD has shown in vitro and in vivo (mice) high toxicity in ovarian cancers.

Methods
A voucher specimen of Calomeria amaranthoides, collected near Old Bell's Line of Road, Mount Tomah NSW 2758, Australia, is held in the John Ray Herbarium, University of Sydney, Collection number: Silvester 110118-01.
Leaves of C. amaranthoides, gathered in the Blue Mountains (Mount Tomah, NSW, Australia) were airdried while protected from sunlight.

High-performance liquid chromatography (HPLC)
HPLC analyses were carried out using the Akta purifier (Amersham Pharmacia Biotech, Sweden) with a HPLCcolumn (150 mm × 4.6 mm i.d. plus pre-column; Grace, The Netherlands), filled with HS Silica (particle size 3 μm), UV detection at 214 nm, 254 nm and 280 nm. Ten μL of the fractionated extract was injected, after dilution to 100 μL with eluent A: hexane (99.5 mL)-dioxane (0.5 mL). The first 10 minutes the column was eluted at a flow rate of 0.5 mL/min with eluent A, followed by 30 minutes with eluent B: hexane (85 mL)diethyl ether (10 mL)-ethanol (5 mL). 1 H-NMR and 13 C-NMR analyses 1 H-NMR and 13 C-NMR spectroscopy was performed on those plant fractions with clear cytotoxicity effects. 1 H-NMR, 13 C-NMR and Correlation Spectroscopy (COSY) were performed using a Varian Gemini 300 MHz instrument (Palo Alto, CA, USA). The spectra were measured in parts per million (ppm) and were referenced to tetramethylsilane (TMS = 0 ppm).
Electrospray ionisation in positive and negative mode (ESI) mass spectrometry analyses were performed using a TSQ 7000 Liquid Chromatography Mass Spectrometer (LC-MS/MS; Thermo, San Jose, CA, USA), equipped with Xcalibur data acquisition and processing software. Short-Column Vacuum Chromatography (SCVC) was performed using a column packed with TLC-grade silica gel H60 (Merck, Darmstadt, Germany)) and applying a step-wise gradient of solvents with increasing polarity. Substances were detected by TLC performed on silica gel coated TLC plates (H60 F254, Merck, Germany) and by 1 H-NMR spectroscopy. Structures of purified compounds were determined by mass spectrometry and 1 H-NMR and 13 C-NMR spectroscopy.

Graphs and Statistics
Graphing and statistical evaluations were carried out with GraphPad Prism 5 for Windows.

Cell lines and cell cultures
Cells used in the assays were five ovarian cell lines (JV, JG, JC, JoN, NF), which were earlier established [9,10], two cell lines OVCAR 3 and SKOV 3 from the American Type Culture Collection (ATCC) as well as epithelial cells from the ovary (serous papillary cystadenomas) [11] and human dermal fibroblasts primary cultures [12].

In vitro cytotoxicity tests with different fractions of C. amaranthoides
In vitro cytotoxicity tests were performed using a nonfluorescent substrate, Alamar blue (BioSource Invitrogen, UK), as described by Pagé et al. [13]. Ovary cells (1 × 10 4 or 5 × 10 4 ) were seeded in 24-wells plates (Costar, USA) and grown in RPMI-1640, supplemented with 6 mM L-glutamine, 10% fetal calf serum (FCS) (Gibco, Invitrogen, UK) and penicillin (100 units/mL) and streptomycin (100 μg/mL), while normal fibroblasts were grown in Dulbecco's modified Eagle medium (DMEM), also supplemented with L-glutamine and FCS. The cultures were maintained in a humidified atmosphere of 5% CO 2 at 37°C. Cell cultures, in triplicates, in exponential growth were treated with the different dried fractions of the plant extract, redissolved in dimethyl sulfoxide (DMSO) and added at final concentrations of 1, 10 and 100 μg/mL. The control cultures had 0.02% (1 μg/mL) 0.2% (10 μg/ mL) and 2% (100 μg/mL) DMSO added to the medium. In 2 mL medium/well 10% Alamar blue was added and 100 μl of the supernatants of the 24-well plates after 24, 48 and 72 hrs incubations were pipetted into 96-well plates (Costar, USA). Cell viability was measured with a 96-well plate reader (Molecular Devices Ltd, UK). In a later stage, after identifying fractions with high cytotoxic effects, the final concentrations of extracts tested ranged from 1-10 μg/mL, with final concentrations of 0.02 up to 0.2% DMSO.

In vivo pilot experiment
An in vivo pilot experiment was performed with 20 BALB/c nude mice (Charles River Laboratories, France). In order to mimic advanced ovarian cancer the mice were injected intraperitoneally (i.p.) with 10 7 OVCAR 3 cells (ATCC) into the abdominal cavity to form ascites. Three groups of mice were examined: 6 control mice (no treatment), 6 mice treated with Cisplatin and 6 mice treated with EPD after ascites had formed. Cells of ascites of two mice were frozen and stored for future experiments. To study reduction of the swollen abdomen 5 mg/kg Platosin (Cisplatin, Pharma Chemie, The Netherlands) and the isolated compound EPD at a final concentration of 20 mg/kg were administered i.p.
Bioassays with ovarian cancer cells indicated fraction 4 (309 mg, 0.09% of the dried plant; out of the twelve fractions, see above) as the fraction with most of the cytotoxicity and its main chemical constituent was identified as EPD. A second main non-cytotoxic constituent, present mostly in Fractions 7 to 9 was identified as EPA (137 mg, 91% purity by NMR and MS analyses).
A small sample of freshly dried leaves (1.63 g) was extracted with dichloromethane (100 mL), filtered and the dichloromethane removed under reduced pressure leaving a dark green residue (62.6 mg, yield 3.9%). Quantitative 1 H-NMR analysis of a CDCl 3 solution showed EPD 44%, EPA 31% and a complex mixture of unidentified constituents 25%.
A small sample of dried leaves (10.31 g), that had been stored in the dark under ambient conditions for 3.5 years was extracted with CHCl 3 (100 mL, 48 hours) filtered and the CHCl 3 removed under reduced pressure leaving a dark green-brown residue (0.62 g, yield 6.0%). Quantitative 1 H-NMR analysis of a CDCl 3 solution showed that EPD and EPA were almost completely absent and a very complex mixture of unidentified constituents made up the bulk of the material.  EPA, is an α-methylene carboxylic acid [15]. The remaining impurities in the purified sample of EPD and EPA (Figures 1A and 1B) were identified as waxes and lipids. No other sesquiterpenoid substances of similar structure to EPD and EPA were detected.

In vitro cytotoxicity tests
Cell viability of normal skin fibroblasts and of cells of the ovarian cell line JC treated with the crude plant extract for 24, 48 and 72 hours at final concentrations of 1, 10 and 100 μg/mL was as follows: The screening test for the fibroblasts with doses of 1, 10 and 100 μg/mL measured for 1 μg/mL: after 24 hours showed cell viability of 104%; after 48 hours 97%; and after 72 hours 98%; for 10 μg/ml: after 24 hours cell viability showed 100%; after 48 hours 96%; and after 72 hours 80%; and for 100 μg/mL: after 24 hours cell viability showed 98%; after 48 hours 83%; and after 72 hours 65%. At all time points (24, 48 and 72 hours) IC 50 was greater than 100 μg/mL.
A similar type of biological assay was performed with the purified compound EPD at final concentrations of 1, 5 and 10 μg/mL for 24, 48 and 72 hours (Table 1). Percent of cell reduction for normal fibroblasts at 72 hours at the highest dose (10 μg/mL) was approximately 30%, while IC 50 was greater than 10 μg/mL. Screening tests for OVCAR 3 and SKOV 3 cells showed that more than 50% and 80% of cells were killed at doses of 5 and 10 μg/mL, respectively.

In vivo pilot experiment
Control mice only injected with the OVCAR 3 cells, were killed when the ascites became a burden. EPD (at final concentration of 20 mg/kg b.w.) was administered i.p. twice/week for six weeks and Cisplatin (at final concentration of 5 mg/kg b.w.) was administered i.p. during 4 weeks, once/week. In general a similar cytotoxic effect was observed between EPD and Cisplatin on the OVCAR 3 cells. However, mice treated with EPD could be kept for a much longer period of time than those mice treated with Cisplatin, for the latter the mice had lost weight significantly and had to be sacrificed after the fourth week. Moreover, following EPD treatment for  6 weeks, three mice were kept alive for another month to see if the reduced abdomen would stay of normal size. Two mice kept their normal size abdomen, whereas, after 6 weeks the abdomen of the third mouse started to increase in size ( Table 2). The rate of change in abdomen size for the mice was determined by linear regression (Figure 2) and statistically evaluated for significance by the unpaired t test. Control versus Cisplatin treated mice were significantly different, P = 0.023, as were control versus EPD treated mice, P = 0.025, whereas, EPD versus Cisplatin treated mice were not significantly different, P = 0.13.

Discussion
The chemical constituents composition of aerial parts of C. amaranthoides have been examined once before by Zdero et al. [16]. None of the constituents reported by them were identified in the C. amaranthoides described in this study. The three constituents reported [16] are isomeric with the two major constituents reported in this study, EDP and EPA. The different constituents reported previously may be due to incomplete isolation and analyses or possibly the result of variation in constituent profiles of plant phenotypes. Another possible explanation is degradation on storage. Our studies have shown that freshly dried plant material is necessary as dried plant material stored for over three years was found to yield less than one-tenth of the normal yield of EDP and EPA.
For the first time the anti-cancer activity of C. amaranthoides has been examined. Two major sesquiterpenes with the eremophilanolide structure sub-type were identified by 1 H-NMR and 13 C-NMR and by mass spectrometry and by comparison with published 1 H-NMR partial spectra as eremophila-1(10)-11(13)-dien-12,8βolide (EPD or Xanthanodien) and eremophila-1(10),11 (13)-dien-12-oic acid (EPA) [14,15]. Belonging to the family of Asteraceae, this family has contributed a large number of natural products including SL's. The alphamethylene gamma-lactone ring is responsible for their bioactivity. Various SL's have demonstrated their anticancer capability in in vitro cell culture and by prevention of metastasis in in vivo animal models [6]. Thus, it is not surprising that C. amaranthoides extract can kill cancer cells, given the fact that one of the two isolated sesquiterpenes, EPD, shows high toxicity.
In the present study, EPA, the other sesquiterpene isolated and identified, did not show cytotoxic effects on the ovarian cancer at concentrations up to 10 μg/mL of purified compound.
Besides the cytotoxic effects of the crude extract of C. amaranthoides with clear effects at 10 μg/mL (cell reduction >80%), the isolated biologically active compound EPD has been shown to have high cytotoxicity (>50%) for ovarian cancer cells at lower concentrations of 5 μg/mL (72 hours) and increased (> 60%) with a dose of 10 μg/mL (at 48 hours; Table 1). Interestingly, both the crude plant extract and EPD did show only a slight cytotoxic effect (20%-30%) on normal fibroblasts in vitro at a concentration of 10 μg/mL (at 72 hours). The in vivo pilot experiment with BALB/c nude mice (Table 2, Figure 2) did show that both EPD and Cisplatin reduced the size of the abdomen. The difference, however, was that mice treated with Cisplatin were in poor condition and became wasted compared with the EPD treated mice.
Ovarian cancer has a poor prognosis. With more than 60% of the patients presenting the disease in stage III or IV, combination chemotherapy with Platinum and Taxol after cytoreductive surgery gives the most tolerated standard regimen [19,20]. In spite of the introduction of new drugs into the management of ovarian cancer there is still need for more novel treatments.

Conclusion
The compound EPD has shown unique cytotoxicity effects on both in vitro (ovarian cancer cell lines) as well as in vivo (mice). Interestingly, it had low cytotoxic effects on normal cells.
More studies in vivo are required to verify the mechanisms and mode of action of EPD, and to further validate the potential of EPD as an anti-cancer drug in ovarian cancer and other types of cancer.