The MDA-MB-231 human breast cancer cell line was purchased from the Korean Cell Line Bank (Seoul, Korea) and routinely maintained in DMEM (Welgene, Daegu, Korea) supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA) and 1% antibiotic-antimycotic solution (Welgene).

The protocol used for tumorsphere culture was as previously described (5,6,8,14). Briefly, cells were suspended in serum-free DMEM/F12 (Welgene) supplemented with 1:50 B27 (Gibco BRL, Grand Island, NY, USA), 10 μg/ml insulin (Welgene), 20 ng/ml recombinant human epidermal growth factor (EGF; R&D Systems, Minneapolis, MN, USA), 10 ng/ml recombinant human fibroblast growth factor (FGF; R&D Systems), and 1% antibiotic-antimycotic solution (Welgene) and cultured in non-adherent plates.

To examine cell proliferation rates, MDA-MB-231 cells were plated at different concentrations (3,000–20,000 cells/well) into 96-well plates under non-adherent (see above), or monolayer culture conditions. After 4 days, premixed cell proliferation reagent WST-8 (Dojindo Laboratories, Kumamoto, Japan) was added to each well and the absorbance of the water-soluble formazan produced by viable cells was measured at 450 nm according to the manufacturer’s instructions.

To compare the chemo-sensitivity of cells in the TS and 2D culture systems, several chemotherapeutics including doxorubicin (Sigma, St. Louis, MO, USA), paclitaxel (Sigma), lapatinib (a dual ErbB1/ErbB2 inhibitor; LC Laboratories, Woburn, MA, USA), U0126 (a MEK inhibitor; LC Laboratories), or LY294002 (a PI3K/AKT inhibitor; LC Laboratories) were added into cells grown in either adherent or non-adherent 96-well plates and cell viabilities were measured 3 days later. To examine the combinational effects of doxorubicin and lapatinib, cells were treated with doxorubicin in a range of 0.2–1 µM in the presence of 5 µM lapatinib.

For cell cycle analysis, cells grown in 2D or TS culture for 4 days were trypsinized after washing with PBS, centrifuged at 1,000 rpm for 3 min and fixed in cold 70% ethanol. After centrifugation, the cells were washed with PBS containing 2% FBS and stained in the dark with 20 µg/ml propidium iodide (Sigma) and 200 µg/ml RNase A (Sigma) for 30 min at room temperature. The cells were analyzed by FACS Calibur II flow cytometry (Becton Dickinson Biosciences, San Jose, CA, USA). To measure intracellular doxorubicin accumulations in cells grown in 2D or tumorsphere cultures, cells were seeded into adherent or non-adherent 6-well plates for 3 days. Cells were treated with 0.5 µM doxorubicin for 30 min, trypsinized, and washed twice with PBS containing 2% FBS. After resuspending in PBS containing 2% FBS, the cells were analyzed by FACSCalibur II flow cytometry (Becton Dickinson Biosciences).

Cells cultured under 2D and TS conditions were treated with 0.3 µM doxorubicin, 5 µM LY294002 or 10 µM SB203580 (a p38 inhibitor; LC Laboratories) for 3 days, and harvested for RNA isolation. Total RNA was extracted using the easy-BLUE™ Total RNA Extraction kit (iNtRON Biotechnology Inc., Sungnam, Korea) and cDNA was synthesized with reverse transcriptase (Takara, Shiga, Japan). RT-PCR for cyclin D1, MDR-1, and GAPDH were conducted as previously described (15). Densitometric analysis was performed using Scion Image software (Scion Co., Frederick, MD, USA). The utilized primer sequences for the RT-PCR reactions were as follows: Cyclin D1 (forward) 5′-AGCTCCTGTGCTGCGAAGT GGAAAC-3′ and Cyclin D1 (reverse) 5′-AGTGTTCAAT GAAATCGTGCGGGG-3′; MDR-1 (forward) 5′-GCC TGGCAGCTGGAAGACAAATACACAAAATT-3′ and MDR-1 (reverse) 5′-CAGACAGCAGCTGACAGTCCAA GAACAGGACT-3′; GAPDH (forward) 5′-ATCCCATCAC CATCTTCCAG-3′ and GAPDH (reverse) 5′-TTCTAGACG GCAGGTCAGGT-3′. The real-time PCR reactions for multi-drug resistance-associated protein-1 (MRP-1) and GAPDH were performed using QuantiMix SYBR green kit (Philekorea, Daejeon, Korea) in Eco Real-time PCR (Illumina, San Diego, CA, USA). mRNA expression level of MRP-1 was calculated after normalizing with GAPDH. The utilized primer sequences for the real-time PCR reactions were as follows: MRP-1 (forward) 5′-GCGAGTGTCTCCCTCAAACG-3′ and MPR-1 (reverse) 5′-TCCTCACGGTGATGCTGTTC-3′; GAPDH (forward) 5′-CTGCTCCTCCTGTTCGACAGT-3 and GAPDH (reverse) 5′-CCGTTGACTCCGACCTTCAC-3′.

Cells grown in 2D or TS conditions were lysed with RIPA buffer (50 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 7.5 and 2 mM EDTA). Phosphatase and protease inhibitor cocktails (GenDepot, Barker, TX, USA) were added immediately before use. Lysates were cleared of debris at 13,000 rpm for 10 min, and protein concentrations were determined using bicinchoninic acid reagent (Sigma). Equal amounts of protein were separated by SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes, which were then blocked with 5% non-fat skim milk in 1X TBS-0.1% Tween-20 (TTBS) for 2 h and incubated with a primary antibody (EGFR, p-EGFR, AKT, p-AKT, ERK 1/2, p-ERK 1/2, p38, p-p38 or GAPDH; Cell Signaling, Beverly, MA, USA) overnight. HRP-conjugated secondary antirabbit antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted 1:5,000 was incubated with blots for 1 h at room temperature. Blots were developed using Luminescent Image Analyzer LAS-4000 (Fujifilm, Tokyo, Japan).

Statistical significance was determined using the Student’s t-test. All experiments were conducted in triplicates, and results are presented as mean ± SD. P-values of <0.05 were considered significant.

First, we cultured MDA-MB-231 breast cancer cells in non-adherent culture condition for 4 days to test their ability to form tumorspheres. As previously reported (14), tumorspheres generated from MDA-MB-231 cells exhibited much looser structures (Fig. 1A) than those derived from MCF-7 cells (data not shown). To evaluate the cell growth rates of tumorspheres and monolayer cultures, cells were plated at different concentrations (3,000–20,000 cells/well) into 96-well plates, in either non-adherent plates, or regular tissue culture plates. After 4 days of 2D or TS culture, cell viabilities were assayed by measuring WST-8 absorbance. The overall WST-8 readings of suspension cultures were at least three times less than those of 2D cultures, seeded with the same cell numbers (Fig. 1B). The cell cycle analysis revealed that this slow cell growth rate in tumorspheres correlated with the accumulation of cells at the G0/G1 phase, showing that 79.81±9.4% cell population in tumorspheres was in the G0/G1 phase, whereas 68.76±5.5% of the 2D cultured cell population was in the G0/G1 phase (Fig. 1C). Moreover, the mRNA expression of cyclin D1, which is the first regulatory protein to drive the G1/S phase transition, was also decreased in tumorspheres (Fig. 1D). Taken together, these results confirmed that the TS culture condition enhanced the quiescent MDA-MB 231 cell population as compared with the 2D culture condition.

To investigate whether culturing cells as spheres affects chemo-sensitivity, MDA-MB-231 culture under TS or 2D conditions were exposed to different concentrations of doxorubicin (0.2–1 µM) or paclitaxel (5–60 nM) for 3 days and then assessed for cell viability. It was found that tumorspheres were resistant to both doxorubicin and paclitaxel (Fig. 2A and B). In fact, the IC₅₀ of doxorubicin for tumorspheres was at least three fold higher than its IC₅₀ for 2D cultured cells (Fig. 2A). Similarly, tumorspheres exhibited significant resistance to paclitaxel with an IC₅₀ value ten times that of 2D cultured cells (Fig. 2B).

We further analyzed intracellular doxorubicin accumulation using flow cytometry (16). Cells grown from either monolayer or tumorsphere cultures were treated with 0.5 µM doxorubicin for 30 min and then doxorubicin fluorescence was analyzed by flow cytometry. The treatment of 2D cultured cells with doxorubicin for 30 min caused a right-shift of fluorescence intensity of doxorubicin as compared with untreated MDA-MB-231 cells, confirming the intracellular accumulation of doxorubicin within cells. On the other hand, the fluorescence intensity of tumorspheres was lower than that of monolayer cultured cells (Fig. 2C), implying that the decreased accumulation of intracellular doxorubicin may contribute to chemoresistance of tumorspheres.

It has been well established that the intracellular accumulation of doxorubicin is associated with the expression of the ATP-binding cassette (ABC) family of drug transporters (17), and therefore, we examined the mRNA expression levels of multidrug resistance (MDR-1) and MRP-1. It was observed that the expression of MDR-1 was similarly induced by doxorubicin in TS and 2D cultured cells (Fig. 2D). However, mRNA expression of MRP-1 was induced in tumorspheres by doxorubicin but not in monolayer cultures (Fig. 2E), suggesting that MRP-1, and not MDR-1, was involved in the chemoresistance demonstrated by tumorspheres.

To elucidate the molecular mechanisms responsible for the chemoresistance of tumorspheres, we first analyzed which signaling pathways are upregulated in the tumorspheres. Western blot analysis showed that the EGFR signaling pathway was more activated in TS than in 2D cultured cells (Fig. 3A). Furthermore, the phosphorylation of EGFR in tumorspheres was associated with concomitant increases in the phosphorylations of ERK1/2 and AKT.

We next explored the role of the EGFR signaling pathway in the formation of tumorspheres. Cells were treated with lapatinib, U0126, or LY294002 to block the EGFR, MAPK, and PI3K/AKT signaling pathways, respectively, and then cell viabilities were assessed. Of note, responses to lapatinib were similar for TS and 2D cultured cells (Fig. 3B), whereas TS cultured cells were less sensitive to U0126 and to LY29400 (Fig. 3C and D). These results suggested that the EGFR signaling pathway plays an important role in mediating the survival of cells in the quiescent state rather than pathways downstream of EGFR, such as, the MAPK and PI3K/AKT pathways.

We next examined whether the blockade of EGFR signaling by lapatinib enhanced cytotoxic effect of doxorubicin on quiescent MDA-MB-231 cells. Cells were treated with different concentrations of doxorubicin in the presence of 5 µM lapatinib for 3 days and the cell viability was measured. Noteworthy, the chemosensitivity of tumorspheres to doxorubicin was more enhanced by lapatinib than that of 2D cultured cells (Fig. 4A and B). As shown in Fig. 4A, treatment with lapatinib mildly increased cytotoxicity of doxorubicin in the monolayer culture but the chemosensitivity of tumorspheres was dramatically increased in the presence of lapatinib (Fig. 4B). In fact, the 60% cell viability after treatment with 1 µM of doxorubicin was significantly decreased to <5% when TS cultured cells were treated with 1 µM doxorubicin in the presence of 5 µM lapatinib (Fig. 4B).

Since we found that increased MRP-1 expression was responsible for the chemoresistance of tumorspheres (Fig. 2E), we tested whether treatment with lapatinib affected the expression of MRP-1 in doxorubicin-treated tumorspheres. As shown in Fig. 4C, doxorubicin-induced MRP-1 expression was significantly suppressed in the presence of lapatinib, suggesting that lapatinib sensitizes tumorspheres to doxorubicin by inhibiting the expression of MRP-1.

To obtain more insight into the mechanism underlying the inhibitory effects of lapatinib against doxorubicin-induced MRP-1 expression in tumorspheres, we first analyzed the effects of lapatinib on EGFR and its downstream signaling pathways. Treatment with lapatinib was found to inhibit the phosphorylation of EGFR and that of AKT and p38 in doxorubicin-treated tumorspheres (Fig. 5A). Furthermore, blockades of the PI3K/AKT or p38 MAPK signaling pathways with 5 µM LY294002 or 10 µM SB203580 remarkably decreased the expression of MRP-1 in doxorubicin-treated tumorspheres (Fig. 5B and C), indicating the involvement of PI3K/AKT and p38 MAPK signaling pathways in the expression of MRP-1 in tumorspheres. These observations suggest that lapatinib inhibits doxorubicin-induced MRP-1 expression by inhibiting EGFR signaling and its downstream PI3K/AKT and p38 MAPK signaling pathways.