*2.3. Cell Culture*

The MCF-7 human epithelial breast adenocarcinoma cell line and the T47D human epithelial breast ductal carcinoma cell line were obtained from the America Type Culture Collection (ATCC, Manassas, VA, USA) and maintained in supplemented RPMI medium with 10% FCS. The OVCAR-3 human epithelial ovarian adenocarcinoma cell line (ATCC, Manassas, VA, USA) was maintained in RPMI medium supplemented with 20% FCS and 5 μg/mL insulin. The COV434 (ECACC 07071909) human ovarian granulosa cancer cell line was maintained in supplemented DMEM/F12 medium. Media in each 75 cm<sup>2</sup> flask of cells were replaced every 2–3 days and each cell line was subcultured twice a week. Cells that had undergone fewer than 25 passages were used for all experiments when they were 80% confluent, and in the exponential growth phase.

#### *2.4. Determination of MCF-7 E*ff*ective Concentration (EC) Values*

MCF-7 cells (20,000 cells per well) were exposed to increasing concentrations of chemotherapeutics and tocopherols for 24 h and cell viability was examined in a crystal violet assay. The e ffective concentration that killed 50% and 25% of MCF-7 cells was calculated by a non-linear regression analysis performed using GraphPad Prism (Version 5.00, San Diego, CA, USA). The experiment was repeated on three separate occasions.

#### *2.5. E*ff*ect of Dox, 4-Hydroperoxycyclophosphamide (4-Cyc),* α *or* γ *Tocopherol on ROS Generation*

MCF-7, T47D, OVCAR-3 or COV434 cells (20,000 cells per well) were added to dark, clear bottom 96-well microplates for 24 h to adhere before adding each test agen<sup>t</sup> to triplicate wells. Cells were exposed to 100 μL 10 μM DCFDA for 45 min at 37 ◦C in a humidified 5% CO2 incubator in the dark. The DCFDA solution was removed, and cells were exposed to 100 μL of chemotherapeutics or tocopherols (Table 1) for 24 h. Concentrations of chemotherapeutics and γToc were the e ffective concentrations that killed 25% of MCF-7 cells (EC25). Since a cytotoxic concentration of αToc was not determined, the highest concentration tested was selected for further examination.

Controls were cells in medium only (background negative control), and cells exposed to low (12.5 μM) or high (50 μM) concentrations of TBHP (positive controls) [38], or 0.8% DMSO as a vehicle control for the tocopherols. Each experiment was repeated on three separate occasions (*n* = 3).


**Table 1.** 24 h MCF-7-derived EC25 chemotherapeutics and tocopherols values.

Dox—Doxorubicin, 4-Cyc—4-hydroperoxycyclophosphamide, αToc—α-Tocopherol, γToc—γ-Tocopherol.

#### *2.6. ROS Measurement by DCFDA Assay*

The ROS production was detected by recording fluorescence immediately after addition of test agents (time 0), every hour for a 3 h incubation period, and after 24 h continuous incubation. Fluorescence was measured according to protocol described by Figueroa et al., [38]. Fluorescence readings were made using a plate spectrofluorometer (GloMax ® Explorer, Promega, Sydney, Australia). Relative fluorescence units (RFU) for each culture well were calculated by subtracting background readings (cells in media only), from all fluorescence values obtained from DCFDA loaded cells in media + test reagents. Each concentration of DCFDA and TBHP was examined in triplicate wells. Plates were sealed to maintain sterility during fluorescence readings and kept at 37 ◦C in a humidified 5% CO2 incubator in the dark between readings.

### *2.7. Crystal Violet (CV) Assay*

Aftermeasurement of ROS, cell viability was determined using crystal violet (4-[(4-dimethylaminophenyl)- phenyl-methyl]- *<sup>N</sup>*,*<sup>N</sup>*-dimethyl-aniline) to stain DNA [39–42]. In short, 20,000 cells per well were cultured for 24 h to allow adherence, then loaded with DCFDA and exposed to test reagents. ROS were measured 0, 1, 2, 3 and 24 h after adding chemotherapeutics and tocopherols. Media containing test agents and non-adherent dead cells were removed, the cells were rinsed with PBS, and the PBS was replaced with 50 μL of crystal violet stain (0.5%) for 10 min. Cells were rinsed with demineralised water to remove any excess stain, then left to air-dry overnight. A total of 60 μL destain solution of 33% acetic acid was added for 10 min before absorbance was read at 570 nm with correction at 630 nm [41]. Linear correlations between optical density and cell number have been reported [41,42], therefore the numbers of viable cells remaining after exposure to test agents were determined by a comparison with a CV standard curve using densities of 0–80,000 cells per well ( *R*<sup>2</sup> = 0.99) generated for the same replicate experiment. Since CV stains DNA, the optical density values included contributions from any stained and adherent DNA, such as that included in condensed nuclei in the early stages of apoptosis or other forms of cell death, and adherent apoptotic bodies characteristic of the later stages of apoptosis.

#### *2.8. DAPI Staining and Scoring of Cell Nuclei*

6-diamidino-2-phenylindole (DAPI), a blue fluorescent dye, binds A–T-rich regions in dsDNA and has been used to visualise condensed, deformed or fragmented nuclei formed during both necrosis and apoptosis [8,43]. Early apoptosis is characterised by cell shrinkage and increased membrane permeability, which facilitates uptake of nuclear dyes such as DAPI. This causes condensed chromatin to appear as 'bright' dye-dense areas, whereas during late apoptosis the nucleus fragments and forms smaller apoptotic bodies.

MCF-7, T47D, OVCAR-3 or COV434 cells (30,000 cells per well) were added to Nunc Lab-Tek II –CC2 chamber units (Promega, Sydney, Australia). After an initial 24 h adherence period to the glass microscope slide, cells were exposed to 300 μL of chemotherapeutics with or without tocopherols (Table 1) and incubated for another 24 h. The test reagents were removed, and the cells rinsed with PBS before fixation with 4% paraformaldehyde in PBS for 25 min at 4 ◦C. The cells were rinsed with PBS, then incubated with 1 μg/mL DAPI prepared in sterile PBS for 30 min in the dark at room temperature. After rinsing with PBS, cells were mounted in bu ffered glycerol and examined using an Olympus fluorescence microscope with filter Chroma 31,000 at excitation 340–380 nm, Dichroic 400 and emission 435–485 nm [44]. Four digital images of each well were taken at 20× magnification and the experiment was repeated on three separate occasions (*n* = 3) for each of the four cell types.

Scoring DAPI-stained nuclei in digital images is a subjective pastime. The scoring criteria were determined by reviewing published reports [45–47] and by observation of MCF7 images from the present study. Very small, bright (DAPI-intense) objects, that appeared as though the nucleus had fragmented into smaller apoptotic bodies (Figure 1A,D), or small very bright objects, usually with irregular shapes suggestive of condensed nuclei (Figure 1A,D,H,I), were collectively scored as

condensed nuclei. Larger, relatively dull objects with regular spherical outlines (Figure 1E–G), or sometimes crescent-shaped outlines, particularly for the COV434 cells (Figure 1C), were scored as normal nuclei. Objects with irregular outlines that were smaller and brighter than normal nuclei were scored as uncertain. In some cases, they were only slightly larger than condensed nuclei, but if they had less DAPI (were duller), they were considered to have less condensed DNA and were classified as uncertain (Figure 1B,F). Groups of dull objects with a similar morphology to groups of apoptotic bodies were included in the condensed category (Figure 1A). Objects scored as uncertain were not clearly apoptotic or condensed nuclei, but neither were they unequivocally normal nuclei. Only complete nuclei or objects were included in the count and objects on the edges of the images were excluded.

Each image was allocated a code and deidentified, shu ffled into a random order and scored blind. The scores were recorded onto each image before re-identifying and entering the numbers of condensed, uncertain or normal nuclei into a spreadsheet. The numbers of condensed nuclei (includes apoptotic bodies that were given a score of 1 for each group) were expressed as a percentage of the normal nuclei for each image (i.e., the uncertain nuclei were excluded from this calculation). There were three independent experiments, except for some cases (OVCAR and MCF7 medium control, MCF7 Dox + 4-Cyc + αToc, MCF7 Dox + 4-Cyc + γToc and MCF7 Dox) in which the experiment was repeated on four separate occasions. The e ffects of the chemotherapeutics and tocopherols on the nuclear morphology of MCF7, T47D and OVCAR cells were very similar, and hence representative examples have been shown in Figure 1. The mean ± stdev (*n* = 3 or 4) was calculated for the percentages of condensed nuclei. Data were subjected to 1Way ANOVA with Tukey post-hoc test.
