Reagents for cell culture were purchased from Invitrogen (Carlsbad, CA, USA), Lonza (Basel, Switzerland), or Cellgro (Manassas, VA, USA). HS578T, MDA-MB-231, and MDA-MB-436 were obtained from the Tissue Culture Shared Resource of Georgetown University Medical Center and maintained in the Dulbecco’s modified Eagle’s medium (DMEM) (Lonza) containing 10% heat inactivated fetal bovine serum (Omega Scientific, Inc., Tarzana, CA, USA) and 100 U/ml penicillin/streptomycin (Lonza). SUM149PT was maintained according to the manufacturer’s recommendation (Asterand, Detroit, MI, USA). The viability of cultured cells was monitored by the trypan blue dye exclusion method using the Luna Automated Cell Counter (Logos Biosystems, Gyunggi-Do, Korea). Receptor tyrosine kinase inhibitors were purchased from the following sources: AEW541 from Cayman Chemical (Ann Arbor, MI, USA); AG1024 from Enzo Life Sciences (Farmingdale, NY, USA); BMS-754807 and OSI-906 from MedKoo Biosciences (Chapel Hill, NC, USA); ABT-869, AV-951, BAY 73-4506, BMS-536924, BMS-599626, brivaninb, cediranib, CYC116, E-7080, ENMD-2076, GSK1838705A, GSK1904529A, JNJ-38877605, LDN193189, MGCD265, motesanib, MP-470, NVP-TAE684, OSI-930, PF-2341066 (crizotinib), PHA-665752, SB431542, SB525334, SU11274, Tie2 kinase inhibitor, XL184, and XL880 from Selleck Chemicals (Houston, TX, USA); axitinib, dovitinib, gefitinib, GW-2580, lapatinib, lestaurtinib, masitinib, pazopanib, sorafenib, sunitinib, tandutinib, vandetanib, and vatalanib from LC Labs (Woburn, MA, USA). Genistein and MG132 was purchase from Sigma (St. Louis, MO, USA). Stock solutions of compounds were made in dimethyl sulfoxide (DMSO) and stored at −20°C in small aliquots.

MDA-MB-231 cells (2,500 cells/ well) in 96-well plates were treated with increasing amount of gefitinib and increasing amount of RTKIs in duplicates in a 6×5 matrix (Fig. 1A). In an initial screening, the highest concentration of RTKIs was 10 μM. The highest concentrations of RTKIs were reduced when significant reduction of cell viability was observed in single agent treatments. The synergism was determined by calculating classification index (CI) with equation of A × B / AB, where A and B are the cell viability with individual agent and AB is the cell viability with the combination (32). We further indexed as follows: strong synergism as index 3 when the CI>1.3 at >5 combination points; medium synergism as index 2 when the CI>1.3 at 3 or 4 combination points; weak synergism as index 1 when the CI>1.3 at 1 or 2 combination points. Cell viability was determined at ~72 h after treatment of compounds by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay as described previously except for using 4 mg/ml of MTT solution (25,33).

Cells were subcultured into 6-well plates with appropriate densities: 500–1,000 cells/well for HS578T and 3,000 cells/well for MDA-MB-231. The day after subculture, the cells were treated with indicated concentrations of compounds for 24 h, and then the cells were supplemented with fresh normal growth media without compounds. The cells were further cultured for 10–14 days after treatment with replacement of fresh normal growth media twice per week. The survived colonies were stained as described previously (34). After intensive washing, the images of colonies were captured by scanner. The relative number of colonies was determined as follows: crystal violet stain of colonies was solubilized by solubilization buffer [1:1 mixture (v/v) of 0.1 M sodium phosphate (NaH₂PO₄, pH 4.5) and ethanol] and the observance of solubilized crystal violet was measured by ELx808 microplate reader (BioTek, Winooski, VT, USA).

Western blot analyses were performed as described previously (25). Antibodies used in this study were as follows: MET (sc-161), ERK1 (sc-94), and PARP (sc-7150) from Santa Cruz (Santa Cruz, CA, USA); p-EGFR (Y1068) (#2237), EGFR (#4405), p-MET (Y1234/Y1235) (#3123), phospho-AKT (Ser473) (#9271), AKT (#9272), p-ERK1/2 (T202/Y204) (#4370), p-p70 S6K (T389) (#9205), p70 S6K (#9202), p-S6 (S235/S236) (#4856), S6 (#2217) and XIAP (#2045) from Cell Signaling (Danvers, MA, USA); α-tubulin, β-actin, and horseradish peroxidase-conjugated secondary antibodies from Sigma.

Transfection of siRNA was performed with Lipofectamine 2000 (from Invitrogen) as described previously (35). In brief, HS578T (0.4–0.6×10⁵ cells/well) or MDA-MB-231 (1.0×10⁵ cells/well) cells in 6-well plates were transfected with 100 pmoles of siRNA mixed with 2.5 μl of Lipofectamine 2000 in serum-free DMEM. After 4-h incubation, cells were supplemented with equal volume of DMEM containing 20% FBS and 200 U/ml penicillin/streptomycin to maintain normal growth condition and further incubated for 3 days. After 3-day incubation, cells were further supplemented with equal volume of DMEM containing 20% FBS and 200 U/ml penicillin/streptomycin and incubated for ≤2 more days. The proliferation of cells was determined by counting viable cells which were stained by acridine orange (AO)/propidium iodide (PI) with the Luna-FLDual Fluorescence Cell Counter (Logos Biosystems). The siRNAs were purchased from Bioneer (Seoul, Korea) with following sequences: control-siRNA, 5′-GAC GAG CGG CAC GUG CAC AUU-3′; and RPS6-siRNA, 5′-GAA GCA GCG UAC CAA GAA A(dTdT)-3′.

Cells were treated as indicated and the cells, both attached and floating, were harvested to analyze the cell cycle at the Flow Cytometry and Cell Sorting Shared Resource of Georgetown University Medical Centers described previously (33).

The two-tailed Student’s t-test was applied for statistical analysis. ^*P<0.05; ^**P<0.01; ^***P<0.001.

Since our previous study identified synergistic effects of EGFR and PI3K/AKT inhibition in a subset of TNBC cells (25), we reasoned that combination of kinase inhibitors with EGFR inhibition might induce synthetic lethality in TNBC cells. We noted that TNBC cells of MSL subtype showed innate resistance to EGFR inhibition. Because overcoming resistance is an unmet need to treat human cancer (36,37), we performed an MTT screening for synthetic lethality with a small library of various RTKIs in MDA-MB-231 cells in the presence of an EGFR inhibitor gefitinib (Fig. 1A). We identified various MET inhibitors (Table I) as potential agents that induced synthetic lethal effects with gefitinib in MDA-MB-231 cells (Fig. 1B). Gefitinib and MET inhibitor combinations synergistically reduced the viable cells in MDA-MB-231 cells in a range of various molar ratios of these drugs (Fig. 1C).

One distinct feature of TNBC is overexpression of EGFR (6,11,12). In addition, a recent study reported that high level expression of MET is an adverse prognostic factor in TNBC patients (38). We determined the level of these proteins in a set of TNBC cell lines by western blot analysis. As reported previously (25), the level of EGFR was high in all TNBC cell lines tested compared to the luminal breast cancer cell line MCF7 (Fig. 2A). The level of MET was also higher in TNBC cells than in MCF7 (Fig. 2A).

Since the levels of both EGFR and MET are elevated in human TNBC cell lines, we further determined the cytotoxic effect of EGFR and MET inhibitors as a single agent toward four different human TNBC cell lines. The cells were treated with the EGFR inhibitor gefitinib or the MET inhibitor SU11274 (39) for ~72 h and the viable cells were determined by MTT cell viability assay. Consistent with a previous report (25), three cell lines (HS578T, MDA-MB-231 and MDA-MB-436) of MSL subtype were relatively resistant to gefitinib compared to a BL2 subtype cell line SUM149PT (Fig. 2B). Gefitinib reduced the viable SUM149PT cells in a dose-dependent manner. On the contrary, the effect of gefitinib was limited on three cell lines of MSL subtype. Additionally, three MSL subtype cell lines were more resistant to SU11274 than SUM149PT cell line (Fig. 2C). Notably, near complete loss of viable cells was observed in SUM149PT cells treated with 10 μM of SU11274, while the effect of SU11274 was less potent toward HS578T, MDA-MB-231 and MDA-MB-436 cells.

Since MET inhibitors were identified as potent synthetic lethal agents in combination of gefitinib, we further determined whether the combination of EGFR/MET inhibitors has any beneficial effect in treatment of human TNBC cell lines of MSL subtype. Three types of TNBC cells were treated with increasing concentrations of both EGFR (gefitinib) and MET (SU11274) inhibitors for ~72 h and the viable cells were determined by MTT assay. Combination of gefitinib and SU11274 with fixed molar ratio of 1:1 markedly reduced the viable cells in HS578T, MDA-MB-231, and MDA-MB-436 cells (Fig. 3A). In addition to gefitinib, lapatinib and BMS-599626 (EGFR/HER2 dual inhibitors) also showed marked synergism with SU11274 (Fig. 3B and C). These results suggest that combination of EGFR/MET inhibitors synergistically reduces the cell viability of MSL subtype cell lines of human TNBC.

The effect of gefitinib/SU11274 combination was further evaluated by clonogenic cell survival assay. TNBC cells were subcultured in 6-well plates in an appropriate density and treated with drug combinations for 24 h. After wash out of drugs, cells were further cultivated in normal growth media. As shown in Fig. 4A, gefitinib alone could not suppress the number of survived colonies in either HS578T or MDA-MB-231 cells. On the contrary, SU11274, as a single agent, significantly reduced the number of surviving colonies. Consistent with MTT assay, gefitinib/SU11274 combination reduced the colony formation in both cell lines.

The effect of gefitinib/SU11274 combination on the cell cycle distribution was further analyzed. Cells were treated with drugs for 24 h and the cells, both attached and floating, were collected to determine the cell cycle. As shown in Fig. 4B, gefitinib alone could not significantly affect the cell cycle distribution of HS578T and MDA-MB-231 cells. However, treatment of SU11274 markedly induced the accumulation of G2 accompanying by reduction of both G1 and S phase in both cell lines. The cell cycle distribution induced by SU11274 was sustained in the cells which were treated with gefitinib/ SU11274 combination.

To detect apoptotic cell death, we further analyzed the Poly (ADP-ribose) polymerase (PARP) cleavage and the level of X-linked inhibitor of apoptosis protein (XIAP) by western blot analysis. HS578T and MDA-MB-231 cells were treated with compounds for 24 h and the lysates were subjected to western blot analysis. As shown in Fig. 4C, no apparent induction of PARP cleavage was observed. The level of XIAP protein was also marginally reduced by gefitinib/SU11274 combination.

To determine signaling pathways mediating the gefitinib/SU11274 effect, we performed a series of western blot analyses. HS578T and MDA-MB-231 cells were treated with increasing concentrations of drugs for 24 h, either single agents or combination, then the lysates from these cells were subjected to western blot analysis. Interestingly, single agent treatment, either gefitinib or SU11274 for 24 h, reduced the level of phospho-ribosomal protein S6 (RPS6) (S235/S236) in HS578T and MDA-MB-231 cells in a dose-dependent manner (Fig. 5A). In addition, gefitinib/SU11274 combination synergistically reduced the level of phospho-RPS6 in these cells. Surprisingly, the level of RPS6 protein itself was reduced by these drugs as single agents and further reduced by combination treatment.

Unexpectedly, 24-h treatment of gefitinib did not reduce the level of phospho-EGFR (Y1068) in these cells, while gefitinib/ SU11274 combination reduced the level of phospho-EGFR (Y1068) only in MDA-MB-231 cells (Fig. 5B). As expected, SU11274 reduced the level of phospho-MET (Y1234/Y1235) in these cells (Fig. 5B). In addition, the level of phospho-MET was also reduced by gefitinib in both cell types. However, neither gefitinib nor SU11274 could reduce the levels of phospho-AKT (S473) and phospho-ERK1/2 (T202/Y204). The gefitinib/SU11274 combination could not reduce either phospho-AKT or phospho-ERK1/2 (Fig. 5B). These results suggest that 24-h treatment of gefitinib or SU11274 could not inhibit the AKT and ERK pathways in these cell lines.

To determine the level of RPS6 over time, cells were treated with gefitinib/SU11274 combination for several time intervals and the level of RPS6 proteins was detected by western blot analysis (Fig. 6A). Interestingly, the decrease of both phospho-RPS6 (S235/236) and RPS6 itself was evident as early as 1 h after treatment (Fig. 6A). In addtion, the decrease of RPS6 protein level was sustained for up to 16 h. The level of phospho-AKT (S473) was decreased at 1 h after combination treatment. However, the decrease of phospho-AKT (S473) was reversed over time in both cell lines (Fig. 6A).

The effect of proteasome inhibition on the level of RPS6 was also determined by western blot analysis (Fig. 6B). Cells were treated with 10 μM of either gefitinib or SU11274 and combination of both drugs for 4 h in the presence of the proteasome inhibitor MG132. Consistently, gefitinib/SU11274 combination markedly reduced the level of RPS6 in both cell lines. However, the treatment of MG132 did not affect gefitinib/SU11274-mediated reduction of RPS6 (Fig. 6B). Contrary to 24-h treatment, 4-h treatment of 10 μM gefitinib reduced the level of phospho-EGFR (Y1068) in these cells. In addition, gefitinib/SU11274 combination further reduced the level of phospho-EGFR. These results suggest that gefitinib/ SU11274 combination induces irreversible reduction of RPS6 in a proteasome-independent manner.

Since gefitinib/SU11274 combination synergistically reduced the level of RPS6 in MSL subtype TNBC cells, we questioned whether RPS6 is important to the proliferation of these cells. To address this, we knocked down the RPS6 protein by specific siRNA. HS578T cells and MDA-MB-231 cells were transfected with either control- or RPS6-siRNA and cultivated for up to 5 days. The difference of cell proliferation was traced by viable cell counting at indicated days. As shown in Fig. 7A, knockdown of RPS6 profoundly reduced the proliferation of both HS578T and MDA-MB-231 cells as early as 3 days after siRNA transfection. Western blot analysis confirmed the knockdown of RPS6 under these conditions (Fig. 7B). Taken together, our data suggest that gefitinib/SU11274 combination reduced the proliferation of a subset of TNBC cells through downregulation of RPS6 proteins.