Human breast cancer cell lines (MCF-7, MDA-MB-231 and T47D) and human promyelocytic leukemia cell line HL-60 cultured in RPMI-1640 supplemented with 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere of 5% carbon dioxide in air.

ATO, nitroblue tetrazolium (NBT), 12-O-tetradecanoylphorbol-13-acetate (TPA) and N-acetyl-L-cysteine (NAC) were purchased from Sigma-Aldrich (St. Louis, MO, USA). CN-A, ISIR-005, 12OH-ISIR-005 and Fusicoccin J (FC-J) were prepared as previously described (19,27). The structures of CN-A, ISIR-005, 12OH-ISIR-005 and FC-J are shown in Fig. 1. YM155 was obtained from Selleckchem (Houston, TX, USA). Methyl cellulose was purchased from Wako Pure Chemical Industries (Osaka, Japan). Human apoptosis array kit and anti-p27 antibody were obtained from R&D Systems (Minneapolis, MN, USA). Anti-caspase-7, anti-survivin, anti-p21 and anti-XIAP antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). Anti-α-tubulin antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

NBT reduction was assayed colorimetrically as previously described (28). Briefly, HL-60 cells were incubated in 1 ml of serum-free medium containing 1 mg/ml NBT and 100 ng/ml TPA at 37°C for 30 min. The reaction was stopped by adding HCl. Formazan deposites were solubilized in DMSO, and the absorption of the formazan solution at 560 nm was measured in a spectrophotometer.

Cells were seeded at 1–3×10⁴ cells/ml in a 24-well multidish. After culture with or without test compounds for the indicated times, viable cells were examined by a modified MTT (3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide) assay as previously reported (25).

MCF-7 cells (2×10³ cells/well) and MDA-MB-231 cells (4×10³ cells/well) were plated in RPMI-1640 supplemented with 10% fetal bovine serum and 1.0% methylcellulose in a 24-well ultra-low attachment multidish (Corning Inc., Corning, NY, USA). Colonies containing 10 or more cells were counted 12 days after seeding.

Cells were packed after washing with cold PBS and then lysed at a concentration of 1×10⁷ cells/ml in lysis buffer CelLytic™ M (Sigma-Aldrich) supplemented with a proteinase inhibitor cocktail and phosphatase inhibitor cocktail 1/2 (Sigma-Aldrich). Equal amounts of protein were separated on 5–20% SDS-polyacrylamide gels (Wako Pure Chemical Industries). Proteins were electrophoresed on gels and transferred to an Immobilon-P membrane (Millipore, Bedford, MA, USA) using the primary antibodies. An anti-rabbit or anti-mouse IgG HRP-linked antibody (Cell Signaling Technology) was used as a secondary antibody (1:2,000 dilution). Bands were identified by treatment with Immune-Star™ HRP chemiluminescence (Bio-Rad Laboratories, Hercules, CA, USA) for 5 min at room temperature and detected using a Fuji Lumino Image Analyser LAS-4000 system (Fuji Film Co., Ltd., Tokyo, Japan) (28). All western blots shown are representative of at least 3 independent experiments.

Cells were plated in 100-mm plastic dishes at a density of 4×10⁴ cells/ml and incubated with CN-A (3.5 μg/ml) and ATO (4 μM) at 37°C for 96 h. The cells were washed with PBS twice and solubilized at 1×10⁷ cells/ml in lysis buffer (Human Apoptosis Array kit). The lysates were resuspended and rocked gently at 2–8°C for 30 min. After centrifuging, the supernatant was transferred into a clean tube. Then, 400 μg of total protein was used for the Human Apoptosis Array kit according to the manufacturer’s protocol as previously described (29).

Cell invasion activity was measured by real-time monitoring of cell invasion using xCELLigence Real-Time Cell Analyzer (Roche-Diagnostics Japan, Tokyo, Japan). For continuous monitoring of cell invasion 5×10⁴ cells were seeded in a 5% (v/v) Matrigel-coated CIM-Plate 16 with 10% serum serving as the chemoattractant in the lower chamber according to the manufacturer’s protocol.

We and others have reported that CN-A and ATO (at low doses) alone could induce the differentiation of human myeloid leukemia cells (4,21). Firstly, we found that CN-A and 0.25 μM ATO synergistically induced the NBT reduction (one of typical differentiation markers of human leukemia cells) (Fig. 2). Then we examined whether these combined treatment also could be effective in the suppression of the proliferation of solid tumor cells including breast cancer cells. CN-A and ATO synergistically inhibited the proliferation of human breast cancer cell line MCF-7 cells (Fig. 3A). Although ATO alone even at a higher concentration (6 μM) inhibited the growth of MCF-7 cells to ~50% of control and CN-A (1.25 μg/ml) alone slightly inhibited the growth of MCF-7 cells, in the presence of CN-A (1.25 μg/ml) ATO at 1.5–4 μM, which is in or close to the range of clinically achievable concentrations (30,31), could inhibit the growth of MCF-7 cells to <50% of control (Fig. 3A). Unexpectedly, MDA-MB-231 cells were more sensitive to ATO: at 1 μM ATO alone could inhibit the proliferation of MDA-MB-231 cells to <50% of control (Fig. 3B). CN-A also effectively enhanced the ATO-induced growth inhibition of MDA-MB-231 cells (Fig. 3B).

Since anchorage-independent growth is well correlated with tumorigenic potential, we next examined whether this combined treatment with CN-A and ATO could effectively inhibit the anchorage-independent growth of these breast cancer cells. Although ATO (1 μM) or CN-A (1 μg/ml) alone inhibited colony formation of MCF-7 cells to ~67 or 64% of controls, respectively, combined treatment with ATO and CN-A inhibited colony formation to 10% of controls (Fig. 4A). Although ATO (0.2 μM) alone slightly inhibited colony formation of MDA-MB-231 cells (<10% inhibition) and CN-A (2.5 μg/ml) alone inhibited colony formation of MDA-MB-231 cells to ~69% of controls, combined treatment with ATO and CN-A inhibited colony formation to 23% of controls (Fig. 4B).

We next examined whether the active CN-A analogue and ATO also cooperatively inhibited the growth of MCF-7 cells. Although ISIR-005, a synthetic CN-A-derivative, at 6 μg/ml slightly inhibited the growth of MCF-7 cells (~20% inhibition) after the 5-day treatment, combined treatment with ISIR-005 plus ATO synergistically inhibited the growth of MCF-7 cells (Fig. 5A). On the other hand, FC-J, a CN-A-related natural product, at 6 μg/ml scarcely inhibited the growth of MCF-7 cells after the 5-day treatment and also could not enhance ATO-induced growth inhibition (Fig. 5B). Furthermore, similar results were obtained from ATO treatment plus 12OH-ISIR-005, an inactive analogue of ISIR-005 (Fig. 5C).

We next examined whether the combined treatment with CN-A and ATO inhibited growth of MCF-7 cells through the induction of apoptosis. Although cleaved caspase-3 is used for one of the markers of apoptosis (32), MCF-7 cells lack expression of caspase-3 as a result of a 47-bp deletion in exon 3 of the CASP3 gene (33). Since there were reports that in MCF-7 cells apoptosis was induced through the activation of caspase-7 instead of caspase-3 (34), we examined whether the combined treatment with CN-A and ATO could induce cleaved caspase-7. MCF-7 cells were cultured with 4 μM ATO, 3.5 μg/ml CN-A or both ATO and CN-A for 96 h. Although ATO alone scarcely induced cleaved caspase-7 and CN-A alone only weakly induced cleaved caspase-7, this combined treatment markedly induced cleaved caspase-7 in MCF-7 cells (Fig. 6). These results suggest that the combined treatment with CN-A and ATO induced apoptosis through the activation of caspase-7.

In order to investigate further the mechanism by which both CN-A and ATO induce apoptosis, various apoptosis-related proteins were examined using apoptosis array analysis. MCF-7 cells were cultured with 4 μM ATO, 3.5 μg/ml CN-A or both ATO and CN-A for 96 h. Whole cell lysates were used for each apoptosis array spotted with 35 antibodies specific to apoptosis-related proteins. As mentioned above, the expression of caspase-3 was not detected in this array (Fig. 7A spot number 1). The expression of death receptors [DR5 (TRAIL receptor 2) and Fas] and inhibitors of cell cycle (p21/CIP1 and p27/Kip1) were significantly induced (Fig. 7A and B). On the other hand, among the inhibition of apoptosis (IAP) family, expressions of cIAP-1 and survivin were clearly decreased but the expression of XIAP was not significantly changed (Fig. 7A and B). We further examined expressions of survivin, XIAP, p21/CIP1 and p27/Kip1 proteins by western blot analysis (Fig. 8). We found that the expression of survivin in MCF-7 cells was dramatically decreased in the presence of both CN-A and ATO, although the expression of survivin was only slightly decreased by CN-A or ATO alone (Fig. 8A and B). On the other hand, the expression of XIAP was not significantly modulated by CN-A and/or ATO (Fig. 8B).

CN-A or ATO alone clearly induced p21/CIP1 and p27/Kip1 proteins. The combined treatment with CN-A and ATO did not further increase the p21/CIP1 and p27/Kip1 proteins (Fig. 8A).

As mentioned above, our results suggest that the induction of cleaved caspase-7 and the inhibition of survivin are important events in the corporative inhibition of growth of MCF-7 cells by both CN-A and ATO. Since survivin is a direct inhibitor of caspasse-3 and -7 (35), we examined the effect of the survivin inhibitor YM155 (36), on the CN-A-induced or ATO-induced growth inhibition of MCF-7 cells. As shown in Fig. 9A, YM155 and ATO synergistically inhibited the growth of MCF-7 cells. Whereas ATO at 2 μM or YM155 at 2.5 nM alone scarcely inhibited the growth of MCF-7 cells, the combined treatment with ATO and YM155 inhibited growth to <30% of control (Fig. 9A). On the other hand, the combined treatment with CN-A and YM155 showed only additive growth inhibition of MCF-7 cells (Fig. 9B). These results suggest that the synergistic growth inhibition by both CN-A and ATO was, at least in part, due to the synergistic inhibition of survivin induced by both CN-A and ATO.

Next, we determined whether oxidative stress was involved in the combined treatment with CN-A and ATO -induced growth inhibition. An antioxidant compound N-acetylcysteine (NAC) that can broadly scavenge reactive oxygen species (ROS) was used to reduce ROS-induced cellular stress. When cell growth rate was assessed, pretreatment with NAC significantly reduced the combination treatment-induced cell growth inhibition (Fig. 10). Furthermore, we found that hydrogen peroxide at low doses could enhance the CN-A-induced or ATO-induced growth inhibition of MCF-7 cells (data not shown). These data suggest that oxidative response plays an essential role in the combination treatment-induced apoptosis.

Finally, we examined whether the combined treatment with CN-A and ATO also could be effective in the suppression of the invasive capacity of MDA-MB-231 cells. For the detection of MDA-MB-231 cell invasion, we used the impedance-based xCELLigence Real-Time Cell Analysis (RTCA) technology. Matrigel (5%, v/v) as extracellular matrix component was added on the top of the microporous membranes of upper chambers. Under this condition, MDA-MB-231 cells could invade time-dependently to lower chambers, whereas MCF-7 and T47D cells could not invade (Fig. 11A). We found that CN-A alone could dose-dependently inhibit the invasion activity of MDA-MB-231 cells (Fig. 11B). As previously reported using standard Matrigel-coated Transwell assay, we confirmed that ATO alone also inhibited the invasion activity in this assay (Fig. 11C). Furthermore, we found that the combined treatment with CN-A and ATO completely suppressed the invasion activity (Fig. 11C).