The breast cancer cell lines MCF-7 (HTB-22™), MDA-MB-468 (HTB-132™) and MDA-MB-231 (HTB-26™), and a normal epithelial breast cell line (MCF-12A; CRL-10782™) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA), and maintained at the Department of Biomedical Sciences, Faculty of Medicine, Prince of Songkla University (Hat Yai, Thailand). The mouse fibroblast L-929 (CCL-1™; ATCC) cell line was provided by Professor Teerapol Srichana, Department of Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, Prince of Songkla University. The culture conditions of the cell lines followed methods described previously (14). Briefly, MCF-7 cells were grown in RPMI-1640 (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), and the MDA-MB-468, MDA-MB-231 and L-929 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Gibco; Thermo Fisher Scientific, Inc). The cultures were supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.), with 100 U/ml of penicillin and streptomycin. The MCF-12A cell line was cultured in DMEM and Ham's F12 medium (GE Healthcare, Chicago, IL, USA), supplemented with 100 ng/ml cholera toxin (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), 500 ng/ml hydrocortisone (Sigma-Aldrich; Merck KGaA), 0.01 mg/ml bovine insulin (Sigma-Aldrich; Merck KGaA), 20 ng/ml human epidermal growth factor (Invitrogen; Thermo Fisher Scientific, Inc.) and 5% horse serum (Invitrogen; Thermo Fisher Scientific, Inc.). All cultures were incubated at 37°C in a humidified atmosphere containing 5% CO₂.

Dried G. maderaspatana (L.) Poir plants were purchased from the Chaokromper Drug Store (Bangkok, Thailand). Their identity was confirmed by Dr Nijsiri Ruangrungsri, Department of Pharmacognosy and Pharmaceutical Botany, Faculty of Pharmaceutical Sciences, Chulalongkorn University (Bangkok, Thailand). A voucher specimen (no. 5182) was deposited at the Museum of Natural Medicine, Chulalongkorn University. The dried whole plant materials were ground to coarse powder using a mortar and pestle. The powder was then stored at room temperature (RT) prior to extraction. Crude compounds were extracted with dichloromethane from a powder weight of 1,500 g. The extract was evaporated under vacuum at 55°C, fractionated by silica gel column chromatography (silica gel no. 9385) and then eluted with different gradients of hexane-ethyl acetate 10:0 to 0:10 solvent systems, resulting in 132.2 mg of a compound resembling white needles (Rf 0.85, silica gel CH₂Cl₂-Acetone 9:1). The structure of the compound was elucidated by ¹H, ¹³C nuclear magnetic resonance spectroscopy and mass spectrometry. Its physical and spectral data were compared with previous reports (9,12). The compound was identified as the sesquiterpene lactone, frullanolide (Fig. 1), with the chemical formula C₁₅H₂₀O₂ (parent peak at m/z 232, colorless solid). The compound was stored at −20°C prior to testing with cancer cells.

All cell lines were seeded in 96-well plates at a density of 2×10⁴ cells/well in 100 µl culture medium/well. The cells were treated with different concentrations of the dimethyl sulfoxide (DMSO)-dissolved frullanolide compound (1.25, 2.5, 5.0, 10.0 and 20.0 µg/ml). After 72 h of incubation at 37°C, 100 µl MTT reagent (0.5 mg/ml) was added to each well and the cultures were incubated for an additional 30 min. The MTT reagent was removed and replaced by DMSO to ensure that solubilization was complete. Absorbance at 570 and 650 nm (reference wavelengths) was measured on a microplate reader. Half-maximal inhibitory concentration (IC₅₀) values were calculated from fitted response curves of the concentration and viability (%). Determination of cytotoxic activity followed the criterion; <5 µg/ml represented highly active; 5–10 µg/ml represented strongly active; and >10 represented weak cytotoxicity (15,16). Normal breast cells (MCF-12A) and L-929 fibroblasts were treated using the aforementioned procedure. Selectivity index (SI) values, indicating selectivity for tested cell lines, were calculated from the ratio of IC₅₀ values of the compounds obtained for normal vs. cancer cells. An SI score >3 represented good selectivity (17). This experiment was performed in triplicate.

The MCF-7 and MDA-MB-468 cells were seeded at densities of 4×10⁵ cells/well in 12-well plates. The MDA-MB-231 cells were seeded at 3×10⁵ cells/well. The cells were treated for 24 h with three concentrations of frullanolide (0.5×, 1× and 2× IC₅₀). Following removal of the compound, the treated cells were fixed with 70% ethanol at 4°C for 4 h and washed three times with cold PBS. Harvesting and fluorochrome binding of the cells were carried out as described previously (2). For cell cycle analysis, the fixed cells were resuspended in 400 µl propidium iodide (PI; 50 µg/ml) solution/1×10⁶ cells and incubated at RT for 5–10 min in the dark. A total of 5,000 cells were analyzed for each condition with a fluorescence-activated cell sorting (FACS) Calibur flow cytometer (BD Biosciences, San Jose, CA, USA), and histograms of cell population ratios in each phase of the cell cycle were acquired. For apoptosis detection, the apoptotic cells in the treated conditions were stained with Annexin V-fluorescein isothiocyanate (FITC)/PI following the manufacturer's protocol (FITC Annexin V Apoptosis Detection kit I; BD Pharmingen™, BD Biosciences). Dot plot graphs of the apoptotic cell ratios were created. All data were analyzed using WinMDI v.2.9 software (J. Trotter, The Scripps Institute, La Jolla, CA, USA).

Untreated cells and cells treated with frullanolide at 0.5× IC₅₀, were harvested at 0, 12, 24 and 48 h. The cells were lysed in radioimmunoprecipitation assay buffer (Thermo Fisher Scientific, Waltham, MA, USA). Protein samples were quantitated by Bradford Assay (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Total proteins (50 µg) of each sample were run separately on a 12% SDS-PAGE gel and transferred onto a nitrocellulose membrane in glycine-methanol buffer at 4°C (100 V, 2 h). The membranes were blocked in 5% low-fat milk in TBS with 0.1% Tween-20 (TBS-T) at RT for 1 h. Subsequently, the blots were incubated at 4°C overnight with primary antibodies; B cell lymphoma 2-associated X protein (Bax; cat no. 5023; 1:1,000 dilution), p21 (cat no. 2947; 1:1,000 dilution), p53 (cat no. 9282; 1:1,000 dilution) and β-actin (cat no. 4967; 1:1,000 dilution), in 1% low-fat milk in TBS-T. All antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). The blots were then washed 3 times (5 min each) and incubated with enhanced chemiluminescence (ECL) anti-rabbit immunoglobulin G horseradish peroxidase (cat no. PKNA934; GE Healthcare Life Sciences, Little Chalfont, UK) diluted to 1:5,000 in 1% low-fat milk in TBS-T at RT for 1 h. The protein bands were detected using an ECL chemiluminescent detection kit (Thermo Fisher Scientific, Inc.). Due to the limited amount of frullanolide compound available, each treatment was performed in triplicate. The triplicate samples were pooled prior to protein lysate preparation and western blotting.

The MTT results are presented as the mean ± standard deviation. To evaluate the difference between the cell lines, one-way analysis of variance (ANOVA) with Brown-Forsythe correction was performed, then IC₅₀ values of the cancerous cell lines were compared with MCF-12A cells using one-sided Dunnett's post hoc tests. Statistical analysis was performed with SPSS 20.0 software (IBM Corp., Armonk, NY, USA). P<0.05 was considered to indicate a statistically significant difference.

The IC₅₀ values of frullanolide for the breast cancer cell lines MCF-7, MDA-MB-468 and MDA-MB-231 were 10.74±0.86, 8.04±2.69 and 12.36±0.31 µg/ml, respectively. The IC₅₀ value for normal breast cells (MCF-12A) was 28.65±6.57 and 19.07±7.16 µg/ml for the fibroblast L-929 cell line. One-way ANOVA indicated significantly different values among the groups. It was tested whether the compound had lower IC₅₀ values in the cancerous cell lines than when applied to MCF-12A cells. In cancerous cell lines, frullanolide had significantly lower IC₅₀ values compared with MCF-12A, whereas the L-929 fibroblast cell line did not exhibit a significant difference. The results indicated a strong anti-proliferative effect of frullanolide on the breast cancer cell lines, but less of an effect on normal breast and fibroblast cell lines (Table I). The SI, which indicates the safety level of a compound toward normal breast cells, indicated that frullanolide has high cytotoxic activity against MDA-MB-468 (SI=3.56) but was less harmful to normal breast cells.

Cell cycle arrest was monitored using flow cytometry (Figs. 2 and 3). Following staining of the nuclei with PI, the FACS analysis revealed that frullanolide induced an increased proportion of SubG₁ cell debris for the MCF-7, MDA-MB-468 and MDA-MB-231 cell lines in a dose-dependent manner. The percentage of SubG₁ cells in the MCF-7 cell line increased from 79.77 at 0.5×IC₅₀ to 90.10% at 2×IC₅₀, compared with 51.05% in untreated cells (Figs. 2A and 3A). There was also an increase in SubG₁ cells in MDA-MB-468, from 50.97 at 0.5×IC₅₀ to 58.60% at 2×IC₅₀ dosage (Figs. 2B and 3B). Additionally, the SubG₁ proportion of the MDA-MB-231 cell line increased from 59.31 at 0.5×IC₅₀ to 76.62% at 2×IC₅₀ dose levels, compared with 37.47% in untreated cells (Figs. 2C and 3C). The cell proportions in the G₁, S and G₂/M phases in the histograms of the MCF-7 treated cells tended to decrease in a dose-dependent manner, from 11.87 to 6.91 (G₁), 3.22 to 1.14 (S) and 5.14 to 1.85% (G₂/M) (Figs. 2A and 3A). The distribution of MDA-MB-468 cells among the phases also changed in a dose-dependent manner (Fig. 2B). For MDA-MB-468, the number of treated cells in the G₁ and S phases gradually decreased until the highest dosage (2×IC₅₀). Conversely, cells appeared to accumulate in the G₂/M phase in treated cells (20.43, 24.13 and 20.37% at 0.5×, 1× and 2× IC₅₀, respectively), compared with untreated cells (14.81%; Fig. 3B). Notably, the G₁, S and G₂/M phases of the MDA-MB-231 cell cycles slightly decreased in a dose-dependent manner from 0.5× to 2×IC₅₀ dose levels (G₁, 25.53 to 13.19; S, 5.06 to 2.52; and G₂/M, 10.10 to 7.68%) as shown in Figs. 2C and 3C. These results indicated that frullanolide was associated with cell cycle arrest at G₂/M in MDA-MB-468 but not in the MCF-7 and MDA-MB-231 cell lines.

To assess the apoptotic action of frullanolide, Annexin V-FITC/PI was used to stain treated cells. A density plot was created based on the data obtained from the FACS analysis. The plot was divided into four quadrants (Fig. 4A-C), with the lower left quadrant presenting viable cells (Annexin V-negative, PI-negative), the lower right quadrant presenting cells that underwent early apoptosis (Annexin V-positive, PI-negative), the upper right quadrant presenting cells which underwent late apoptosis (Annexin V-positive, PI-positive), and the upper left presenting necrotic cells (Annexin V-negative, PI-positive). After 24 h of treatment of MCF-7 cells, the lowest viable counts along with the highest apoptotic counts (early and late apoptosis) were identified at the 0.5×IC₅₀ dose level (14.97 and 84.38%, respectively; Fig. 4A and D). These results indicated that frullanolide may be effective in MCF-7 cells at different dose levels, particularly at a low dose level (0.5×IC₅₀). Conversely, frullanolide treatment of MDA-MB-468 cells resulted in the lowest number of viable cells (11.96%) and the highest number of apoptotic cells (86.38%) at the highest dosage (2×IC₅₀; Fig. 4B and E). Although no significant differences were observed for the MDA-MB-231 cells regarding the number of viable and apoptotic cells among all treatment conditions, treatment exhibited effective cytotoxic activity compared with the control (Fig. 4C and F). Therefore, the data indicated that frullanolide may induce apoptosis in breast cancer cells at all tested concentrations.

To elucidate the possible mechanisms of apoptotic induction by frullanolide, the expression levels of three proteins (Bax, p21 and p53) were investigated by western blot analysis at 0.5×IC₅₀, for 12, 24 and 48 h in the three breast cancer cell lines. The overall effect of the frullanolide treatment in these breast cancer cell lines on detected protein expression is presented in Fig. 5. For all cell types treated with 0.5×IC₅₀ frullanolide, Bax protein expression decreased between 12 and 24 h. Its expression was prone to accumulation in treated MCF-7 and MDA-MB-231 cells at 48 h (Fig. 5A, C, J and L). However, expression levels of Bax protein increased at 12 h in treated MDA-MB-468 cells and its expression gradually reduced at 24–48 h (Fig. 5K). Elevated p21 expression was noted at 12 h, and gradually decreased between 24 and 48 h in all three breast cancer cell lines. The p21 expression was upregulated when compared with the controls (Fig. 5G-I). Notably, p53 protein was expressed differentially depending on the cell type. p53 expression was undetectable in MCF-7 cells (Fig. 5A and D), while distinct adverse alterations were identified in the treated MDA-MB-468 and MDA-MB-231 cells (Fig. 5E and F). The p53 protein levels in frullanolide-treated MDA-MB-468 cells increased at 12–24 h, and immediately declined by ~0.80-fold until the end of exposure (24–48 h) when compared with the control group (Fig. 5E). Therefore, it appeared conclusive that the frullanolide mechanism was involved in apoptotic pathways.