Human breast cancer cell lines (MCF7 and BT-474) and a human ovarian carcinoma cell line (ES-2) were obtained from the American Type Culture Collection (ATCC; Manassas, USA). MCF7 cells were cultured in Dulbecco's Modified Eagle's Medium supplemented with 10% fetal bovine serum (FBS), 0.1 mM non-essential amino acids and 1.0 mM sodium pyruvate. BT-474 cells were cultured in ATCC Hybri-Care Medium (ATCC 46-X) supplemented with 10% FBS and 30 ng/mL epidermal growth factor. ES-2 cells were maintained in McCoy's 5a Medium supplemented with 10% FBS. Cells were maintained at 37°C in a humidity-controlled incubator with 5% CO₂. Paclitaxel and doxorubicin were obtained from Sigma-Aldrich (St. Louis, USA).

Clonogenic assay is a versatile tool for in vitro screening of drugs for antineoplastic activity [8]. The clonogenic ability is an estimate of the reproductive viability of cancer cells (capacity of a single-cell suspension to produce progeny; or a single cell to form a colony of 50 or more cells) under different circumstances. We performed the colony-forming assay as previously described [9]. MCF7 cells were briefly treated with paclitaxel (30 nM) for 1 hour, doxorubicin (200 nM) for 3 hours or vehicle control. Cells were re-suspended and a total of 700 cells were seeded into 6-well plates containing standard culture media. After 10 days, colonies were stained with 3% crystal violet and examined by microscopy. Colonies containing at least 50 individual cells were counted.

This assay was carried out using modified Boyden chambers as previously described [10]. Cells were cultured in 6-well plates and were treated with paclitaxel for 1 hour, doxorubicin for 3 hours, or a combination of both in different sequences. Cells were then re-suspended and plated into the transwell insert. Transwell filters were polycarbonate membranes with 8-μm pores (Corning Life Sciences, Corning, USA). For invasion assay, the upper surfaces of the transwell membranes were coated with Matrigel matrix. Cells (MCF7, 8 × 10⁴ cells/insert; BT-474, 2 × 10⁵ cells/insert; ES-2, 1 × 10⁴ cells/insert) in the transwell insert were allowed to migrate/invade toward the underside of the membrane for an additional 48 hours (MCF7 and BT-474 cells) or 18 hours (ES-2 cells). Non-migrating/non-invading cells were removed by wiping the upper side of the membrane, and cells on the lower surface of the filter were stained with Diff-Quick (Sysmex, Kobe, Japan). Number of migrated/invaded cells were counted from five random fields and averaged from at least three independent experiments.

MCF7 cells were treated with paclitaxel (30 nM) for 1 hour or doxorubicin (200 nM) for 3 hours. Thereafter, the media was replaced and cells were maintained in standard culture media further for 24 to 48 hours. Whole cell lysates were prepared using the M-PER Protein Extraction Reagent (Thermo Fisher Scientific, Waltham, USA), and the protein concentrations were determined using the Bradford assay kit (Bio-Rad, Hercules, USA). Aliquots of 30 μg of total protein were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis and then transferred to a nitrocellulose membrane. The membranes were blocked in 5% (w/v) non-fat milk in TBST buffer, followed by primary antibody incubation at 4°C overnight. The primary antibodies used were anti-phospho-IκB kinase (IKK) α/β (Ser176/180) (#2697; Cell Signaling Technology, Danvers, USA), anti-E-cadherin (#5296; Cell Signaling Technology), anti-Twist1 (sc-15393; Santa Cruz Biotechnology, Dallas, USA), and phospho-myosin light chain (MLC) 2 (Ser19) (#3671; Cell Signaling Technology). Anti-β-actin antibody (Sigma-Aldrich) was used as the loading control. Immunoreactive proteins were detected by using horseradish peroxidase-conjugated secondary antibodies and were visualized by using the enhanced chemiluminescence system (Merck Millipore, Burlington, USA) [11].

MCF7 cells were treated with paclitaxel (30 nM) or doxorubicin (200 nM) for 30 minutes. Cell lysates were prepared and RhoA activity assay was performed using the RhoA G-LISA Activation Assay Kit (Cytoskeleton, Inc., Denver, USA) according to the manufacturer's instructions. RhoA activity was quantified by determining the absorbance reading at 490 nm.

Two RhoA inhibitors, Y16 and rhosin/G04, were obtained from MedChemExpress (Monmouth Junction, USA). These inhibitors have been reported to specifically inhibit RhoA activity and RhoA-mediated signaling functions [12]. Cells were pretreated with or without Y16 (50 μM) or rhosin/G04 (50 μM) for 24 hours, and then incubated with doxorubicin (MCF7, 200 nM; BT-474, 800 nM) or vehicle control for an additional 3 hours. Then, cells were re-suspended and migration and invasion assays were performed as previously described.

Experiments were performed in triplicates, and data were reported as the mean ± standard error of the mean. Comparisons were made by using the two-tailed Student's t-test. Data were analyzed using GraphPad Prism version 6.01 for Windows (GraphPad Software, La Jolla, USA). Statistical significance was determined at a probability value of < 0.05.

Anthracyclines and taxanes have been established as the cornerstones of chemotherapy in the management of breast cancer. To investigate the sublethal effects of these two agents, first we determined the half maximal inhibitory concentrations (IC₅₀) of paclitaxel and doxorubicin by dose-response cell viability curve. The IC₅₀ values of paclitaxel and doxorubicin in MCF7 cells were 30 nM and 200 nM, respectively. These concentrations were used in subsequent experiments.

In our institution, paclitaxel was administered weekly, in a standard manner of one-hour infusion. Doxorubicin was administered by three-hour slow infusions instead of bolus administration to reduce cardiac exposure to and penetration by anthracyclines. Slow infusion is a well-established strategy which may prevent or diminish the risk of early or delayed cardiotoxicity [13]. Therefore, in this study, the exposure time was based on our clinical practice.

MCF7 cells, following a brief treatment with paclitaxel or doxorubicin, exhibited reduced clonogenic ability (Figure 1A). Colony numbers were 78.2% ± 3.6% and 70.7% ± 4.1% compared to the control in the paclitaxel and doxorubicin group, respectively. This suggests that a brief treatment with these cytotoxic drugs is sufficient to compromise the cell growth of breast cancer cells.

Cell migration in the extracellular matrix is a multistep process involving dynamic changes in the cytoskeleton, cell-substrate adhesion, and the extracellular matrix components. Strikingly, the migratory ability of breast cancer cells was enhanced following a brief treatment with doxorubicin but not paclitaxel (Figure 1B). There was a 1.56-fold increase in the migratory ability of MCF7 cells following a three-hour treatment with doxorubicin. Conversely, the migratory ability decreased to 64.2% ± 4.0% compared to control after one-hour treatment with paclitaxel.

Next, we determined the invasive capacity by adding a layer of Matrigel matrix on top of the transwell membrane to mimic the process of extracellular matrix invasion and extravasation of breast cancer cells. Consistent with the results of the migration assay, the invaded cells were 65.2% ± 7.0% and 155.8% ± 15.9% compared to control following treatments with paclitaxel and doxorubicin, respectively (Figure 1C).

The effects of combination treatment in different sequences were also examined. The invasion ability of MCF7 cells was 72.9% ± 12.5% compared to control after one-hour paclitaxel treatment followed by three-hour doxorubicin treatment (Figure 1D). In the other group with a reverse sequence of treatment, such as, doxorubicin followed by paclitaxel, the invasion ability was 101.7% ± 10.4% of the control. This observation suggests that the invasion-inhibiting effect of paclitaxel may surpass the invasion-promoting effect of doxorubicin when the two agents are sequentially used in combination. Furthermore, although the difference was not statistically significant (p = 0.11), the invasion-promoting effect of doxorubicin tends to be more prominent during sequential treatment of cells with doxorubicin followed by paclitaxel than paclitaxel followed by doxorubicin.

To corroborate our findings in another breast cancer cell line, BT-474 cells were treated with doxorubicin for three-hours and migration and invasion assays were subsequently performed. As shown in Figure 2A and B, brief treatment with doxorubicin (200 nM) did not significantly change the migratory or invasive capacity of BT-474 cells. Nonetheless, a brief treatment with 800 nM doxorubicin significantly augmented the migration and invasion of BT-474 cells. Taken together, a brief treatment with doxorubicin suppressed the clonogenic ability but enhanced the migratory and invasive capacity of BT-474 breast cancer cells.

We further examined whether these phenomena were limited to breast cancer cells or could be observed in other cancer cells. ES-2 ovarian cancer cells exhibited IC₅₀s for paclitaxel and doxorubicin similar to those of MCF7 cells (data not shown). Interestingly, a brief treatment with paclitaxel did not affect the migratory or invasive capacity, but doxorubicin treatment promoted cell migration or invasion in ES-2 ovarian cancer cells (Figure 2C and D). This suggests that the migration- and invasion-promoting effects induced by doxorubicin were not unique to breast cancer cells.

To identify the mechanisms which might account for the increase in cell migration and invasion in doxorubicin-treated breast cancer cells, the expression of pertinent cell motility factors was determined. Doxorubicin treatment in breast cancer cells has been shown to activate an NF-кB transcription signature [14]. Furthermore, doxorubicin might induce epithelial-mesenchymal transition by downregulation of E-cadherin and upregulation of Twist1 [15]. However, when cells were briefly treated with paclitaxel or doxorubicin, there was no significant change observed in IKK phosphorylation or Twist1 expression (Figure 3A). The expression of E-cadherin increased at 24 hours and returned to the baseline following either paclitaxel or doxorubicin treatment.

MLC is directly and indirectly phosphorylated by Rho-associated coiled-coil-containing protein kinase (ROCK) at Ser19, thereby increasing the actomyosin contractile force and cell protrusions, resulting in cell migration [16]. A recent study showed that doxorubicin treatment caused the formation of long filopodia-like protrusions in MCF7 spheroid cells [17]. Interestingly, we found that the expression of phosphorylated MLC2 was significantly downregulated following one-hour treatment with paclitaxel, while MLC2 phosphorylation increased progressively from 24 to 48 hous following a three-hour treatment with doxorubicin (Figure 3A).

ROCK is a RhoA effector downstream of RhoA activation [18]. We next investigated whether the upstream RhoA activity was enhanced after treatment with doxorubicin. RhoA activity in MCF7 cells was determined following a brief treatment with paclitaxel or doxorubicin. Notably, paclitaxel decreased RhoA activity whereas doxorubicin significantly increased RhoA activity (Figure 3B), consistent with the changes in MLC2 phosphorylation.

Rescue experiments were performed to investigate the significance of RhoA-GTPase activity in relation to our findings. Pretreatment with RhoA inhibitors, Y16 or rhosin/G04, significantly abolished the increase in cell migration (Figure 4A and B) and invasion (Figure 4C and D) induced by doxorubicin treatment in MCF7 and BT-474 breast cancer cells. This finding suggests that the migration- and invasion-promoting effects induced by a brief treatment with doxorubicin are mediated, at least in part, by the upregulation of the RhoA/MLC pathway.