2.1 Cell lines and chemical agents

H322, H358 (courtesy of Dr. Jeou-Yuan Chen), H1650, and H1975 (American Type Culture Collection, Manassas, VA) cells were maintained in RPMI1640 medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (FBS) (Sigma–Aldrich, St. Louis, MO) and penicillin/streptomycin (Invitrogen, Carlsbad, CA) at 37 °C. Gefitinib (Ryss Lab, Inc., Union City, CA), CL-387,785 (EMD Chemicals, Gibbstown, NJ), and blebbistatin (Tocris, Ellisville, MO) were commercially obtained. All chemicals were dissolved in DMSO for a final concentration of 10 μM.

2.2 Cellular fractionation

The cells were rinsed twice with PBS and resuspended in hypotonic lysis buffer (10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.4% Triton X-100). The cells were lysed in lysis buffer with continuous rotation at 4 °C for 30 min. The lysate was centrifuged at top speed at 4 °C for 10 min. The supernatant was collected as the cytosolic fraction and the nuclear pellet was resuspended in hypertonic buffer (20 mM HEPES (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 10% glycerol) with pipetting followed by vigorous shaking at 4 °C for 2 h. The sample was centrifuged at top speed for 10 min and the supernatant was collected as the nuclear extract.

2.3 Western blotting

The cells were lysed in lysis buffer (150 mM NaCl, 1% Triton X-100, 50 mM Tris, pH 8.0). The cell extracts were resolved on 10% SDS-PAGE gels and transferred to PVDF membranes (PerkinElmer, Waltham, MA). The membrane was blocked with 5% skim milk in TBS with 0.1% Tween-20 for 30 min at room temperature prior to overnight incubation with primary antibodies recognizing tubulin (Sigma–Aldrich, St. Louis, MO), MYH9 (Abnova, Taiwan), EGFR, Lamin B, COX-2 (all from Santa Cruz, Santa Cruz, CA), β-actin (Millipore, Billerica, CA), p-EGFR (Tyr1068), p-EGFR (Tyr845), p-Stat3 (Tyr705), Stat3, p-Akt (Ser473), Akt, p-Erk (Thr202/Tyr204), and Erk (all from Cell Signaling, Danvers, MA). The membranes were washed thrice in TBST and incubated with horseradish peroxidase-conjugated secondary antibodies in TBST for 1 h. After consecutive washes, the membranes were visualized with a chemiluminescence kit (PerkinElmer, Waltham, MA).

2.4 Immunoprecipitation

The lysates collected as previously described were incubated with an anti-EGFR or anti-MYH9 (Santa Cruz, Santa Cruz, CA) antibody overnight at 4 °C. The next day, the reactions were incubated with protein A/G plus-agarose (Santa Cruz, Santa Cruz, CA) at 4 °C. After 3 h, the beads were washed twice with washing buffer (150, 300, or 500 mM NaCl, 10% glycerol, 0.1% NP-40, and 0.1% Triton X-100). The pull-down complexes were eluted by boiling the beads in SDS-sample dye and were subjected to Western blotting or Coomassie blue staining.

2.5 Cell proliferation analysis

The cells were incubated with the indicated compounds for 3 days. The cell proliferation rate was determined by the MTS assay. RPMI 1640/MTS (Promega, Madison, WI)/PMS (Sigma–Aldrich, St. Louis, MO) (8: 2: 0.1) medium was added to each well, and the cells were incubated for 1.5 h. The absorbance was measured at 490 nm by PowerWavex (BioTEK instruments, Winooski, VT). The rate of cell proliferation was calculated as the absorbance ratio of treated to vehicle-treated cells.

2.6 Flow cytometry

The cells were treated with the indicated compounds for 3 days, harvested by trypsin-EDTA, and stained with Annexin V-FITC (BD Pharmingen, San Diego, CA) and propidium iodide (PI; Sigma–Aldrich, St. Louis, MO) in binding buffer containing 2.5 mM Ca⁺². After 10 min, the samples were subjected to flow cytometry on a FACSCalibur machine (BD Bioscience, Franklin Lakes, NJ) for quantification of the apoptotic population using the CellQuest Pro software (BD Bioscience, Franklin Lakes, NJ). The cells in the early apoptotic stage were Annexin V-FITC positive and PI negative, whereas the cells in the late apoptotic stage were positive for both stains.

2.7 RNA interference

H1975 cells were transfected with control or MYH9 siRNA using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA), following the manufacturer's instructions. The MYH9 siRNA target sequence (cgt tac tac tca ggg ctc atc tac a) was designed and synthesized by Invitrogen using BLOCK-iT™ RNAi Designer. The negative universal control (46–2001) siRNA was purchased from Invitrogen.

2.8 Real-time polymerase chain reaction

The total RNA was extracted by TRIzol reagent (Invitrogen, Carlsbad, CA) and was reverse-transcribed by M-MLV reverse transcriptase (Promega, Madison, WI). The cDNA of COX-2 was analyzed by real-time PCR in a LightCycler 1.0 system (Roche, Indianapolis, IN) using a KAPA SYBR^® FAST qPCR Kit (Kapa Biosystems, Woburn, MA) following the manufacturer's recommendations. The primers used for the amplification of COX-2 were as follows: forward 5′-ttaatgagtaccgcaaacgc-3′ and reverse 5′- accagaagggcaggatacag-3′. The PCR amplification consisted of 40 cycles (95 °C for 2 s, and 60 °C for 30 s) after an initial denaturation step (95 °C for 1 min). Each sample was normalized to the individual GAPDH mRNA content.

3.1 The identification of the cytoskeletal proteins MYH9 and β-actin as EGFR-associating proteins in H1975 cells

To determine whether the EGFR-T790M mutation alters the specificity of EGFR for substrates, certain EGFR-associating proteins were investigated in H1975 cells, which have been widely used to investigate molecular mechanisms of acquired drug resistance. EGFR-interacting partners were isolated by co-immunoprecipitation (co-IP) using an anti-EGFR antibody in cell lysates from H1975 and H322 cells, which are EGFR wt and were used as a control. Three unique bands were detected in lysates prepared from H1975 cells in a Coomassie blue-stained gel (Figure 1A). Two of the bands with a higher molecular weight were identified as MYH9 and the third band was identified as β-actin by in-gel trypsin digestion followed by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) analysis. Hits which had a non-specific interaction with EGFR or an incorrect molecular weight were considered as insignificant hits. To confirm the identity of these proteins, co-IP was performed in H1975 and H358 (EGFR wt) cells followed by Western blotting analysis. Both cell lines expressed similar levels of EGFR, MYH9, and β-actin (Figure 1C). Increased levels of MYH9 and β-actin were detected in the anti-EGFR pull-down products in H1975 cells, whereas background levels of MYH9 were pulled down by the anti-EGFR antibody in H358 cells (Figure 1B). The resistance of EGFR-MYH9 and EGFR-β-actin interactions to a wash with high-salt buffer and the increase of salt concentration in the wash buffer magnifying the binding intensity of MYH9 and β-actin with EGFR suggested that a strong and hydrophobic interaction may exist between EGFR and MYH9 and β-actin.

3.2 EGFR interacts with MYH9 and β-actin in the nuclei of H1975 cells

To understand the role of MYH9 and β-actin in EGFR-involved pathways, we first investigated the location of the EGFR-MYH9 and EGFR-β-actin interactions. Co-IP was conducted to examine EGFR interactions with MYH9 and β-actin in the cytoplasmic and nuclear fractions. An additional EGFR-mutant NSCLC cell line, H1650 (EGFR-Del), was used to determine whether EGFR interactions with MYH9 and β-actin universally occur in EGFR-mutant NSCLC cells. As shown in Figure 2A, neither the nuclear nor the cytoplasmic EGFR showed detectable physical interactions with MYH9 and β-actin in H322, H358, or H1650 cells. The failure to detect EGFR interactions with MYH9 and β-actin in H1650 cells suggests that these interactions may uniquely occur in H1975 cells. Moreover, EGFR appeared to associate with MYH9 and β-actin predominately in the nuclei of H1975 cells, implying that MYH9 and β-actin may be involved in the EGFR nuclear pathway in H1975 cells. To understand why EGFR has different degrees of binding intensity to MYH9 and β-actin in diverse cell lines, the protein levels of all 3 proteins from cytoplasmic and nuclear fractions were examined in these cell lines (Figure 2B). Tubulin, a cytoplasmic marker, was detected in the cytoplasmic fractions, and lamin B, a nuclear marker, was only present in the nuclear fractions, indicating a clean separation of the cytoplasmic and nuclear fractions. The cytoplasmic EGFR levels of H358 and H1650 cells were relatively low compared to those of H322 and H1975 cells. There were no large differences in the protein levels of cytoplasmic MYH9 and β-actin among the 4 cell lines. Notably, relatively higher levels of EGFR were observed in the nuclei of H1975 cells compared to those in H322, H358, and H1650 cells, suggesting that abundant nuclear EGFR may at least partially contribute to the stronger binding intensity of EGFR with MYH9 and β-actin in H1975 cells.

Considering that the EGF ligand initiates EGFR endocytosis and subsequent EGFR accumulation in the nuclei of different cancer cell lines (Lo et al., 2006; Wang and Hung, 2009), we were curious whether EGF treatment could increase the levels of nuclear EGFR and enhance the EGFR interaction with MYH9 and β-actin in H1975 cells. However, the levels of nuclear EGFR in H1975 cells were not significantly altered by the addition of EGF up to 45 min (Figure 2C). Moreover, EGF treatment did not affect the EGFR interaction with MYH9 and β-actin in either H322 or H1975 cells as evidenced by the observation that those interactions did not change in response to EGF (Figure 2D).

3.3 Disrupted cytoskeleton integrity by blebbistatin reduces the EGFR interaction with MYH9 and β-actin

MYH9 is a heavy chain of myosin IIA that interacts with β-actin to maintain cytoskeleton integrity. To determine whether a fully-functioning cytoskeleton is required for EGFR-MYH9 and EGFR-β-actin interactions, blebbistatin was used to disrupt cytoskeletal integrity. Blebbistatin has been identified as a small molecule inhibitor for myosin IIA and has been shown to inhibit actin-activated MgATPase activity of myosin IIA (Straight et al., 2003). Subsequent characterizations revealed its selectivity and high affinity for several classes of myosin II (Limouze et al., 2004). Kinetic analysis indicated that blebbistatin binds to the myosin-ADP-inorganic phosphate complex with high affinity and disrupts the release of phosphate which traps myosin in a state with low actin affinity (Kovacs et al., 2004). In this study, the MTS assay was used first to determine the effects of blebbistatin on the proliferation of H1975 cells. As shown in Figure 3A, blebbistatin exhibited anti-proliferation effects against H1975 cells with an EC₅₀ value of gt; 40 μM. The EC₅₀ value is the effective concentration required for a certain compound to inhibit cell growth by 50%. In addition, a feature of dendritic morphology was observed in H1975 cells in the presence of 10 and 100 μM blebbistatin, suggesting its efficacy in H1975 cells (Figure 3B). As previously described (Kovacs et al., 2004), blebbistatin interrupted the physical interactions between MYH9 and β-actin as determined by the observation that reduced β-actin was detected in the pull-down products of an anti-MYH9 antibody in H1975 cells treated with blebbistatin for 2 h (Figure 3C).An increase of blebbistatin up to 100 μM did not further inhibit the interactions of MYH9 and β-actin. Further studies showed that higher concentrations of blebbistatin impaired not only the MYH9-β-actin interaction but also the EGFR interaction with MYH9 and β-actin, as shown by the decreased levels of MYH9 and β-actin pulled down by an anti-EGFR antibody in cells treated with 100 μM blebbistatin (Figure 3D). It was observed that a 2-hr treatment of blebbistatin did not significantly change the protein levels of EGFR, MYH9, and β-actin in the nuclei (Figure 3E); therefore, we excluded the possibility that the disruption of the EGFR interactions with MYH9 and β-actin by blebbistatin results from reduced levels of these 3 proteins. It is evident that blebbistatin can physically disrupt EGFR interaction with MYH9 and β-actin.

3.4 The down-regulation of MYH9 improves the sensitivity of H1975 cells to CL-387,785

The compound CL-387,785 was used to determine whether EGFR kinase activity is required for nuclear EGFR interactions with MYH9 and β-actin. CL-387,785 is an irreversible inhibitor that covalently binds to the ATP-binding region of EGFR (Discafani et al., 1999). Consistent with previous studies, CL-387,785 effectively inhibited the phosphorylation of EGFR and its downstream targets, ERK and AKT (Kobayashi et al., 2006), whereas gefitinib had a minimal effect on EGFR-mediated pathways in H1975 cells (Figure 4A). Co-IP was performed to examine the interactions among EGFR, MYH9 and, β-actin in H1975 cells treated with 0.1, 1, or 10 μM CL-387,785 for 2 h. CL-387,785 efficiently inhibited EGFR phosphorylation at Tyr1068 in the nuclei of H1975 cells but did not significantly affect the levels of nuclear MYH9 and β-actin after a 2-hr treatment (Figure 4B). EGFR kinase activity was not required for the EGFR interactions with MYH9 and β-actin because nuclear EGFR still associated with residual MYH9 and β-actin, despite EGFR-Tyr1068 being completely inhibited by 1 μM CL-387,785 (Figure 4B, C). Compared to low concentrations of CL-387,785, increasing CL-387,785 to 10 μM led to a further disruption of the EGFR-MYH9 and EGFR-β-actin interactions (Figure 4C). Surprisingly, CL-387,785, which is known as an EGFR kinase inhibitor, also functioned to weaken the interaction between MYH9 and β-actin, similar to blebbistatin (Figure 4D).

Our results indicate physical interactions among EGFR, MYH9, and β-actin in H1975 cells; however, it remains unclear whether these interactions have functional significance. To elucidate this question, the functional significance of MYH9 in EGFR-mediated pathways was determined by ascertaining whether the down-regulation of MYH9 improved the efficacy of CL-387,785 in H1975 cells. Small-interfering RNA (siRNA) was used to knock-down MYH9 expression. Modest down-regulation of MYH9 was observed in H1975 cells 48 h after transfection with 10 nM MYH9 siRNA (Figure 4E). A higher concentration of MYH9 siRNA (50 nM) resulted in further down-regulation of MYH9. We examined the effects of MYH9 down-regulation on the sensitivity of H1975 cells to CL-387,785. H1975 cells were transfected with control or 50 nM MYH9 siRNA. After 24 h, cells were treated with vehicle (DMSO) or various concentrations of CL-387,785 for 72 h. MYH9 siRNA alone only caused minor cytotoxicity (<10%, data not shown). Because the proliferation of vehicle-treated cells in each group (transfected with control or MYH9 siRNA) was set at 100%, the cytotoxicity caused by MYH9 siRNA has been excluded. Treatment with CL-387,785 inhibited the cell proliferation of control siRNA-transfected H1975 cells; this inhibition was more prominent when MYH9 was down-regulated by siRNA treatment (Figure 4F), suggesting that the EGFR interaction with MYH9 is functionally significant and important for the proliferation of H1975 cells.

3.5 The disruption of cytoskeleton integrity by blebbistatin induces higher levels of apoptosis in CL-387,785-treated H1975 cells

Drug combinations have demonstrated clinical advantages in the treatment of different cancers; therefore, we were interested in understanding whether there is a cooperative interaction between an EGFR kinase inhibitor, CL-387,785, and a myosin inhibitor, blebbistatin. The pharmacological interactions between CL-387,785 and blebbistatin were determined by the isobologram and combination index (CI) methods. The data obtained from the MTS analysis were used to perform these analyses by the median-effect principle (CalcuSyn analysis) (Chou and Talalay, 1983). The individual doses of CL-387,785 and blebbistatin to achieve 90% growth inhibition (ED90; effective dose 90), 75% growth inhibition (ED75), and 50% growth inhibition (ED50) were plotted on the x- and y-axes. CI values that fall on the line represent an additive interaction between drugs, whereas CI values below or above the line represent a synergistic or antagonistic interaction, respectively. Figure 5A indicates that the combined treatment of CL-387,785 and blebbistatin synergistically inhibited the proliferation of H1975 cells by 50%, 75%, and 90%. To determine whether apoptosis was triggered synergistically by the combination of CL-387,785 and blebbistatin, the apoptosis signals were examined by flow cytometry analysis 3 days after treatment. Moderate doses of CL-387,785 (0.5 μM) and blebbistatin (20 μM) were selected to examine the combined effects on apoptosis in H1975 cells. The results from the flow cytometry analysis indicated that the combination of CL-387,785 and blebbistatin enhanced the accumulation of apoptotic cells compared to each agent alone (Figure 5B). Taken together, these observations suggest that a synergistic effect exists between CL-387,785 and blebbistatin for the inhibition of cell proliferation and the induction of apoptosis in H1975 cells. Because of the observation that each agent effectively disrupts the interaction of EGFR with MYH9 and β-actin, we next examined whether the cooperation of CL-387,785 and blebbistatin has synergistic effects on the physical interactions of these proteins. The protein interactions from cells treated with either agent or the combination were analyzed by co-IP. Although neither the single treatments nor the combined treatment caused significant changes in levels of MYH9 and β-actin, a synergistic reduction in the EGFR interactions with MYH9 and β-actin was observed in cells treated with the combined treatment compared to those with a single treatment (Figure 5C).

In order to investigate whether the enhanced cytotoxicity of CL-387,785 and blebbistatin results from the suppression of the EGFR nuclear pathway, the levels of COX-2, a transcriptional target of nuclear EGFR, were examined. COX-2 catalyzes the biosynthesis of prostaglandins and is implicated in the invasiveness and angiogenesis of a number of cancers (Greenhough et al., 2009; Lo et al., 2010). A COX-2 inhibitor, celecoxib, has demonstrated anti-tumor effects in glioma xenografts (Nam et al., 2004). Our results showed that CL-387,785 alone dramatically inhibited the protein levels of COX-2, whereas blebbistatin had no effects on levels of COX-2 (Figure 5D). It has been proposed that nuclear EGFR activates the transcription of COX-2; therefore, we sought to understand whether the reduced protein levels of COX-2 resulted from a decrease in transcript levels of COX-2. As shown in Figure 5D, blebbistatin did not effectively inhibit the transcription of COX-2, whereas a tremendous reduction of COX-2 transcript was observed in cells treated with CL-387,785, implying that CL-387,785 functions to suppress the transcriptional activity of nuclear EGFR. Co-treatment with blebbistatin resulted in a further decline in transcript and protein levels of COX-2 in cells treated with CL-387,785 (Figure 5E).