MUC1 Drives c-Met–Dependent Migration and Scattering

MUC1 is a heterodimeric O-glycosylated transmembrane mucin highly expressed in ductal carcinomas, including carcinomas of the breast, lung, pancreas, colon, and ovary (1–5). MUC1 expression has been shown to correlate with increased severity of disease and metastatic progression in gastric, prostate, breast, and pancreatic cancer in addition to thymic epithelial tumors (6–11) and MUC1 cellular mislocalization has been shown to correlate with disease progression in non–small cell lung cancer (12). In addition, MUC1 expression is increased specifically in metastatic colorectal cancer (13), highlighting the importance of MUC1 for cancer progression and metastasis.

MUC1 has been implicated as a driver of metastatic progression, via specific activities of both the extracellular domain and its intracellular domain (MUC1-C; refs. 14–16). The extracellular domain of MUC1 has been shown to increase endothelial cell adhesion and promote intravasation and extravasation through interactions with ICAM-1 and galectin-3 (17, 18). In addition, localization of MUC1 can alter integrin-dependent binding and cell adhesion (19). Furthermore, MUC1 can drive invasion and migration via the upregulation of matrix metalloproteinase 13 (MMP13), a protein that aids in cell invasion (20), and through the induction of epithelial to mesenchymal transition in pancreatic cancer cells (21). These analyses show a role for MUC1 in promoting cell migration, invasion, and metastasis. In addition, mouse mammary tumor virus (MMTV)-MUC1 transgenic models of breast cancer displayed lung metastasis that was dependent upon expression of the cytoplasmic portion of MUC1 (22).

MUC1-C has been shown to interact with a number of oncogenic proteins, promoting either their membrane function or activity in the nucleus (23–28). One such interaction is via the epidermal growth factor receptor (EGFR), wherein MUC1 expression promotes the oncogenic properties of EGFR by prolonging the activity of EGFR and promoting its nuclear function (29, 30). Studies in the WAP-TGFα model of EGFR-dependent breast cancer showed that loss of Muc1 expression suppresses tumor formation by greater than 60% and significantly blocks early hyperplasia (note: mouse Muc1 and human MUC1 are denoted as such). In addition, loss of Muc1 expression completely abrogates metastatic progression in this model (31). Interactions between MUC1 and EGFR reduce the ligand-dependent degradation of EGFR, and alter EGFR trafficking from the lysosome, resulting in preferential recycling to the plasma membrane (29). In addition, MUC1 coexpression promotes retrotranslocation of EGFR to the nucleus, where interaction with MUC1 promotes the binding of EGFR to the CCND1 promoter (30). MUC1-C has also been shown to interact with β-catenin and p120-catenin to promote their translocation to the nucleus and their activity as transcriptional cofactors (16, 32). Overall, MUC1 promotes the intracellular localization and activity of a number of proto-oncogenes, including EGFR, FGFR, PDGFRβ, β-catenin, p120 catenin, Src, estrogen receptor, p53, HSP70, and HSP90 (14–16, 29, 30, 32–35).

In previous studies, we have generated a MUC1 “decoy” peptide to block protein–protein interactions between MUC1 and EGFR and MUC1 and β-catenin. This was accomplished by synthesizing a 15-amino acid region of MUC1-C that was previously shown to be required for interactions between these proteins in tandem with a cell penetrating peptide (CPP; refs. 36–39). The CPP allows adjacent peptide sequences to be taken up into cells across the plasma membrane, where interactions with endogenous intracellular proteins can occur. The MUC1 peptide was termed protein transduction domain MUC1 inhibitory peptide (PMIP), and in vitro studies showed that treatment of breast cancer cells with PMIP resulted in an inhibition of interaction between MUC1 and EGFR as well as an inhibition of the colocalization between MUC1 and β-catenin (39). PMIP significantly inhibited the growth and invasion of breast cancer cell lines in vivo and the initiation and progression of mammary gland tumors in the MMTV-pyMT transgenic model. In animals treated with PMIP, analysis of remaining mammary glands and tumors revealed a reduction in MUC1 expression after treatment. In addition, PMIP significantly suppressed the ability of primary breast tumors to form secondary metastasis in an MDA-MB-231 orthotopic model of breast cancer (39). Subsequent to this work, Klinge and colleagues, reported PMIP treatment of lung cancer cells resulted in a decrease in proliferation, decreased estrogen receptor α (ERα)-dependent gene transcription, and altered subcellular localization of MUC1, Erα, and ERβ (40).

In the current study, we investigated the mechanism by which MUC1 promotes metastatic progression and whether PMIP could inhibit this phenotype. Analysis of MUC1-induced migration in Matrigel revealed an induction of both migration and cell scattering. Using microarray technology, we identified c-Met mRNA as being significantly downregulated by PMIP. Further characterization of c-Met regulation showed a role for c-Met in driving MUC1 and EGFR-dependent migration and scattering.

BT20 breast cancer cells were treated for 1 hour with 50 μmol/L PMIP, 50 μmol/L control peptide, or peptide vehicle (PBS) and RNA was collected after 24 hours. Six CodeLink Human Whole Genome Bioarrays were hybridized, and data analysis was conducted by the University of Arizona Genomics Facility Core. The False Discovery Rate method of statistical significance was used to interpret the data (41), in addition to the GeneSpring program by Agilent.

RNA was extracted from cells using the RNeasy Mini Kit (Qiagen). The Superscript III First-Strand Synthesis System for RT-PCR was used to generate cDNA (Invitrogen). Polymerase chain reaction was conducted using Crimson Taq DNA Polymerase (New England Biolabs) and the following gene-specific primers: MET: sense 5′-ACTCCCCCTGAAAACCAAAGCC-3′, antisense 5′-GGCTTACACTTCGGGCACTTAC-3′ and ACTB: sense 5′-AAGAGAGGCATCCTCACCCT-3′, antisense 5′-TACATGGCTGGGGTGTTGAA-3′ (42). Cycling conditions were: MET: denaturing at 95°C for 30 seconds, annealing at 62°C for 30 seconds, and extension at 68°C for 30 seconds; ACTB: denaturing at 95°C for 30 seconds, annealing at 58°C for 30 seconds, and extension at 68°C for 30 seconds. The PCR products were separated on a 1% agarose gel. Gels were imaged and densitometry conducted using CareStream Molecular Imaging Software (v 5.0.4.46).

RNA was extracted from cells using the RNeasy Mini Kit (Qiagen) and cDNA was generated through the use of Transcriptor First Strand cDNA Synthesis Kit (Roche). Quantitative PCR (qPCR) was conducted using the following primers synthesized by Operon: MET: 5′-AAATGTGCATGAAGCAGGAA-3′ and 5′-TCTCTGAATTAGAGCGATGTTGA-3′; ALAS1: 5′-GCCTCTGCAGTCCTCAGC-3′ and 5′-AACAACACTCTCCATGTTCAGG-3′. Cycling conditions were: initial denaturing at 95°C for 2 minutes, 55 cycles of: denaturing at 95°C for 10 seconds, annealing at 62–52°C for 10 seconds (temperature was reduced 0.5°C per cycle), and extension at 72°C for 5 seconds, final extension at 72°C for 2 minutes, and a final cooling to 37°C for 2 minutes. Products were detected using Roche Universal Probe Library gene-specific probes. PCR was conducted on a Lightcycler 2.0 (Roche).

MDA-MB-231 cells were acquired from the American Type Culture Collection (ATCC) and grown in RPMI 1640 (Cellgro) with 10% FBS (PAA Laboratories) and 1% Pen/Strep (Cellgro). Cells were selected and grown with 1 mg/mL G418 (Omega Scientific, GN-04). BT20 cells were acquired from ATCC and grown in Eagle's minimal essential medium (EMEM) (ATCC) with 10% FBS and 1% Pen/Strep. All cells were grown in 5% CO₂.

MDA-MB-231 cells were transfected with pCDNA3 plasmid containing the full length MUC1 cDNA, which was a kind gift from MA Hollingsworth (University of Nebraska Medical Center). Lipofectamine 2000 (Invitrogen) was used to transfect constructs into cells according to the manufacturer's specifications.

Human recombinant EGF was from Fisher. SU11274 was from EMD Chemicals. Human recombinant hepatocyte growth factor (HGF) was from R&D Systems. PMIP was produced by GenScript (sequence: NH₂-YARAAARQARAPYEKVSAGNGGSSLS-COOH) and reconstituted in PBS at a concentration of 1 mmol/L. MUC-1 antibody was obtained from Thermo Scientific (AB-5), and β-actin antibody was obtained from Sigma (AC-15).

Cells were lysed in cell lysis buffer (20 mmol/L Tris, pH 7.5, 150 mmol/L NaCl, 1% NP40, 5 mmol/L EDTA, pH 8.0, and 10 mmol/L sodium fluoride, 2 mmol/L sodium orthovanadate, 50 μmol/L ammonium molybdate, and Complete (Roche) protease inhibitors). Cell lysates were vortexed briefly, centrifuged, and the supernatant stored at −80°C. Protein concentrations were determined by BCA assay (Pierce). Proteins were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membrane (Millipore). Membrane was blocked in 3% nonfat milk in PBS [0.1% Tween20 (Fisher Scientific)] or 3% bovine serum albumin in TBS [0.1% Tween20 (Fisher Scientific)] and immunoblotted. The membrane was then treated with Super Signal West Pico Chemiluminscent Substrate (Pierce), visualized on Full Speed Blue film (American Digital Imaging & Medical Products, Inc.), and developed with a Konica SRX-101A. Blot images were manipulated as follows: images were converted to black and white and brightness was adjusted using Adobe Photoshop Elements Version 9.0.

MDA-MB-231 cells were seeded onto Matrigel at a density of 6.0 × 10³ cells per 8-well chamber and incubated in RPMI with 10%FBS, 1% Pen/Strep, and 1 μg/mL G418. Cells were treated with 10 ng/mL EGF, 20 ng/mL HGF, both ligands, or were left untreated. Media and treatment were changed every other day. To count single scattered cells, cells less than 50 pixels in diameter were considered a single cell. We conducted a 2-tailed student t test to define statistical significance. Images were manipulated as follows: all images were converted to black and white, shadows were lightened, highlights were darkened, and brightness and contrast were adjusted using Adobe Photoshop Elements Version 9.0.

MDA-MB-231 cells and BT20 cells were seeded and allowed to grow to complete confluence. A scratch was introduced and media was replaced with complete media containing either 10 ng/mL EGF, 50 μmol/L PMIP, 5 μmol/L SU11274 or no treatment. Cells were observed at 6 hours (MDA-MB-231) or 24 hours (BT20) after scratching. Image J software was used to analyze the area of the scratch, and we conducted a 2-tailed student t test to determine statistical significance. Images were manipulated as follows: all images were converted to black and white, shadows were lightened, highlights were darkened, and brightness and contrast were adjusted using Adobe Photoshop Elements Version 9.0.

MUC1 promotes EGFR-dependent transformation in the WAP-TGFα model of breast cancer, inhibits ligand-dependent degradation of EGFR and promotes EGFR-dependent cyclin D1 expression (29–31). The cell penetrating peptide to the MUC1-cytoplasmic domain, PMIP, blocks MUC1 and EGFR interactions and inhibits tumor progression in both the MMTV-pyMT transgenic model of breast cancer and the scid-MDA-MB-231 orthotopic model of breast cancer (39). In addition to blocking tumor progression, PMIP treatment results in a reduction of metastasis in the scid-MDA-MB-231 mouse model (39). To determine whether inhibition of metastasis is because of decreased cell motility upon PMIP treatment, we conducted an in vitro wound-healing assay. For this assay, parental BT20 breast cancer cells or MDA-MB-231 breast cancer cells transfected with either a CMV-MUC1 overexpression construct (hereafter referred to as MDA-MB-231-MUC1) or the CMV empty vector (MDA-MB-231-control) were used (Fig. 1N). Cells were seeded on Collagen IV-coated tissue culture plastic, allowed to grow to complete confluence and then a scratch was introduced into the monolayer. Complete media alone (Fig. 1G and H), complete media plus 10 ng/mL EGF (Fig. 1A–D and I–J), or complete media plus 10 ng/mL EGF and 50 μmol/L PMIP (Fig. 1 Eand F and K and L) was added to cells. Activation of phosphotyrosine by EGF was observed in EGF-treated groups, although PMIP treatment reduced total pEGFR/EGFR levels (Fig. 1P) as was previously shown (39). Cells were observed at either 6 hours (MDA-MB-231) or 24 hours (BT20) after treatment to assess motility into the scratch. Six hours after treatment, it was found that MDA-MB-231-MUC1 cells filled the scratch 23% more than MDA-MB-231-control cells (P = 0.002), and PMIP treatment reduced this behavior back to baseline (P = 0.0004; Fig. 1M). Twenty-four hours after treatment of BT20 cells, PMIP treatment reduced EGF-dependent migration by 28% (P = 0.017; Fig. 1O). This indicates that MUC1 is a driver of cell motility, and that PMIP blocks this activity. These data indicate that PMIP functions to block EGF-dependent migration.

To analyze a potential mechanism by which PMIP reduces metastasis, we conducted a Codelink Whole Human Genome Bioarray in BT20 breast cancer cells. Cells were treated with PMIP, cell penetrating peptide (PTD-4) or vehicle (PBS) for 24 hours, RNA was collected and hybridized to the array (Fig. 2A). We observed differential regulation greater than 3-fold in 97 genes affecting a variety of processes with a majority of the downregulated genes involved in cell cycle, cell proliferation, and signal transduction (Fig. 2A and B, and Supplementary Fig. S1). Among the downregulated genes was MET, which was downregulated 3-fold with PMIP treatment. To verify that MET expression was regulated by PMIP treatment, we conducted semiquantitative RT-PCR on BT20 cells treated with 10 or 20 μmol/L PMIP, compared MET expression to untreated cells, and found a dose-dependent reduction of MET expression in response to PMIP treatment (Fig. 2C). In addition, we conducted qRT-PCR on BT20 cells treated with 50 μmol/L PMIP, the same concentration that was used in the microarray study. We compared MET expression in these cells to cells treated with vehicle (PBS) and found a significant reduction in MET expression upon PMIP treatment (Fig. 2D). Next, the effect of PMIP treatment on MET expression was evaluated by RT-PCR in MCF10A cells, a nontransformed breast epithelial cell line (Fig. 2E). While MET is expressed in MCF10A, no significant reduction was found upon PMIP treatment, indicating that the effect of PMIP may be tumor specific. These data indicate that one potential mechanism by which PMIP blocks metastasis is through the downregulation of MET.

PMIP is hypothesized to act by blocking MUC1-dependent functions in the cell. To determine whether MUC1-dependent migration relied on c-Met activity, we conducted an in vitro wound-healing assay on Collagen Type IV and Fibronectin extracellular matrix substrates with the c-Met kinase inhibitor, SU11274. MDA-MB-231-control and MDA-MB-231-MUC1 cells were seeded onto Collagen Type IV or Fibronectin and scratched as described in Fig. 1. To determine the effect of c-Met activity on MUC1-dependent migration, cells were either treated with dimethyl sulfoxide (DMSO) vehicle and 10 ng/mL EGF (Fig. 3A–D) or supplemented with 5 μmol/L SU11274 (Fig. 3E and F). As seen before, MDA-MB-231-MUC1 cells exhibited greater wound healing capacity than MDA-MB-231-control cells (Fig. 3D vs. 3B). Addition of SU11274 significantly decreased scratch filling in MUC1-overexpressing cells (Fig. 3F vs. 3D), showing that MUC1-driven cell motility requires activation of c-Met. On Collagen Type IV, MUC1 increased migration by 19% (P = 0.002), which was repressed by c-Met inhibitor back to baseline levels (P = 0.0002). On Fibronectin, MUC1 increased migration by 36% (P = 0.032), which again, was repressed by c-Met inhibitor to control levels (P = 0.033; Fig. 3G and H). In addition to evaluating migration in MDA-MB-231 cells, migration was also assessed in BT20 cells which were treated with either complete media (Fig. 3I and J), complete media plus 10 ng/mL EGF (Fig. 3K and L), or complete media plus 10 ng/mL EGF and 5 μmol/L c-Met inhibitor (Fig. 3M and N). Analysis of c-Met activity revealed an induction of phospho-Met in response to EGF compared with control (Fig. 3P middle lane vs. left lane), whereas this induction is blocked by the c-Met kinase inhibitor (Fig. 3P right lane vs. middle lane). Note that treatment with either EGF or the c-Met kinase inhibitor does not alter total MUC1 levels. When evaluated in BT20 cells (Fig. 3I–N), EGF-stimulated migration (on both Collagen IV and Fibronectin) was again blocked by Met inhibitor. As seen in MDA-MB-231 cells, the c-Met kinase inhibitor blocks wound healing in BT20 cells. On Collagen IV, the c-Met kinase inhibitor reduced migration of BT20 cells by 28% (P = 0.013; data not shown), and on Fibronectin, migration was reduced by 27% (P = 0.0001; Fig. 3O).

One of the hallmarks of c-Met activity is its ability to induce cell scattering. To evaluate the effects of MUC1 and EGF on cell scattering, cells were plated in Matrigel and evaluated for single-cell scattering away from primary colonies. For this assay, MDA-MB-231 cells were seeded in Matrigel and grown for 13 days in the presence of 10% serum. During this time, cells formed colonies and dispersed from the periphery of the colonies through the surrounding matrix (Fig. 4, arrowheads). To determine the effect of MUC1 expression on scattering in this assay, MDA-MB-231-MUC1 cells or MDA-MB-231-control cells were seeded in Matrigel and single cells were counted in the matrix surrounding the cell colonies. We observed that MDA-MB-231-MUC1 exhibited greater cell scattering than MDA-MB-231-control cells (Fig. 4B and D vs. 4A and C, arrowheads, and quantified in Fig. 4E). We also noticed that colonies of MDA-MB-231-MUC1 cells occasionally had branches invading into the surrounding Matrigel matrix, whereas we did not see these branches in MDA-MB-231 control cells. (Fig. 4B and D vs. 4A and C, arrows). These results show that MUC1 promotes cell scattering in breast cancer cells grown on Matrigel.

To evaluate the role of EGFR activity on MUC1-dependent cell scattering, we treated MDA-MB-231 cells with 10 ng/mL human recombinant EGF to stimulate EGFR and evaluated scattering as described above. We found that EGF-treated MDA-MB-231-MUC1 cells exhibited more cell scattering than MDA-MB-231-MUC1 cells grown in the absence of EGF (Fig. 4C and D vs. 4A and B, arrowheads, and quantified in Fig. 4E). We also observed that EGF-treated MDA-MB-231 control cell colonies possessed invasive branching similar to those seen in MDA-MB-231-MUC1 cells without EGF treatment, and the invasive phenotype of the cell colonies was most apparent in cells both overexpressing MUC1 and treated with EGF (Fig. 4C and D vs. 4A and B, arrows). These results indicate that EGFR activation promotes scattering similar to that seen with MUC1 expression in 3-dimensional culture.

Our observation that MUC1 and EGF induced cell branching in addition to cell scattering prompted us to evaluate the branching phenotype. We treated MDA-MB-231-MUC1 and MDA-MB-231 control cells in Matrigel with no exogenous ligand (Fig. 5A and E), 10 ng/mL EGF (Fig. 5B and F), 20 ng/mL HGF, the hepatocyte growth factor and the ligand for c-Met (Fig. 5C and G), or both 20 ng/mL HGF and 10 ng/mL EGF (Fig. 5D and H). MDA-MB-231-MUC1 cells were found to form branches in the presence of EGF or HGF (Fig. 5F and G, arrows). While MUC1 expression was required for the branching phenotype induced by either ligand alone, both ligands together could synergize and form branches in the absence of MUC1 (Fig. 5D, arrows). These data indicate that MUC1 and EGFR together promote cell scattering and branching, a phenotype that is enhanced by the c-Met ligand, HGF.

To explore whether the MUC1 and EGF-driven scattering effects observed in 3-dimensional culture were because of c-Met activity, we treated MDA-MB-231-MUC1 cells with 10 ng/mL EGF and a selective c-Met kinase inhibitor, SU11274, or DMSO vehicle and cells were imaged after 8 days of treatment (Figure 6A). In MDA-MB-231-MUC1 cells treated with EGF, scattering (Fig. 6B, arrowheads), and branching (Fig. 6B, arrows) were observed as shown in Figs. 4A and 5. Alternatively, the c-Met kinase inhibitor completely blocked the MUC1 and EGF-induced phenotype (Fig. 6C). Importantly, SU11274 does not affect proliferation in MDA-MB-231-MUC1 or MDA-MB-231 control cells (Supplementary Fig. S2). These results indicated that MUC1 and EGFR-dependent cell scattering is dependent upon c-Met kinase activity.

In the current study, we have identified c-Met as an important mediator of MUC1-dependent migration, scattering, and branching of breast cancer cells. Microarray analysis of genes with altered expression in response to PMIP treatment identified a number of potential metastatic mediators, including c-Met. In BT20 breast cancer cells RT-PCR verified that PMIP downregulates c-Met expression in a dose-dependent fashion. Further evaluation of c-Met determined that MUC1 expression resulted in an increase in cell migration on Type IV collagen, and this increase in migration can be inhibited by the selective c-Met inhibitor, SU11274. Furthermore, when placed in Matrigel, cell scattering and branching is increased dramatically upon MUC1 overexpression and EGF treatment, and this phenotype is also blocked by treatment with the c-Met kinase inhibitor.

Previously, MUC1 was shown to promote cell motility and invasion through upregulation of MMP13 and through the induction of epithelial-to-mesenchymal transition (EMT) in pancreatic cell lines (20, 21). The breast cancer cells we used in this study, MDA-MB-231 have a mesenchymal phenotype that is not dependent on MUC1. The second cell line used, BT20, have an epithelial phenotype, and this was not altered by either MUC1 downregulation or PMIP treatment (data not shown). It is possible that the EMT phenotype previously identified is specific to pancreatic cell lines. In the microarray analysis, we did not find that MMP13 transcription was altered following PMIP treatment (data not shown). Alternatively, we did find LAMA5 was downregulated by PMIP treatment (Supplementary Fig. S1), and LAMA5 has been shown to be involved in cell migration (43). Whether LAMA5 is a direct or indirect target of MUC1 will be the subject of future studies. Overall, we initially identified 97 genes significantly upregulated/downregulated at the level of the microarray, and further analyzed 6 targets by qRT-PCR. Of these 6, only MET and LAMA5 were confirmed by qRT-PCR in 3 or more replicates (data not shown).

Our data indicate that MUC1 expression can drive c-Met–dependent scattering and branching. These data are in opposition to a previous study in which MUC1 was shown to interact with and be phosphorylated by Met (34). In that study, the authors observed that MUC1 expression reduced MMP1 and decreased cell invasion in pancreatic cells (44). Whether this is a tissue-type specific response, or whether c-Met negatively regulates MUC1 is unclear. Of note, all of our c-Met–dependent migration required EGF, which may be an important difference between our experiments and those in the previous study. In the current study, we have shown that MUC1 promotes the expression of c-Met, whereas EGF promotes the activation of c-Met. Together, MUC1 expression and EGFR activation would result in the activated expression of c-Met, driving metastatic progression. We previously showed that PMIP treatment increases EGF-dependent degradation of EGFR (39). This data, in combination with the current study, may indicate that PMIP blocks metastatic progression by blocking both c-Met expression and its activation via EGFR.

Overall, the body of literature indicates that both MUC1 and c-Met promote metastatic phenotypes (45–48). A recent study by Matsubara and colleagues that show that MUC1, EGFR, and c-Met are coordinately upregulated in aggressive lung cancer patient samples (49). This human data correlates with both our current study demonstrating that MUC1 expression can promote c-Met activity and with our previous study showing that MUC1 promotes EGFR protein stabilization (29). We are currently investigating the mechanism by which MUC1 and/or EGFR promote increased MET expression.

Using metastatic breast cancer cells, we showed that MUC1 promotes cell motility using an in vitro wound-healing assay. This behavior is reversed upon treatment with PMIP, illustrating that MUC1 promotion of metastatic behavior can be targeted therapeutically. This is the first demonstration of the therapeutic targeting of MUC1 activity as it relates specifically to metastasis. Recently, Raina and colleagues demonstrated another decoy peptide, GO-201, that is designed to block MUC1 multimerization and nuclear translocation. Results from that study indicated a loss of tumor growth, but no effects on metastatic potential were reported (50). Interestingly, Zhang et al., recently found that PMIP blocks smoke-induced EMT in airway epithelial cells, another mechanism by which PMIP may be capable of blocking metastasis (51).

Previous work regarding the role of MUC1 in promoting metastatic progression indicated that the extracellular domain of MUC1 was required for its function in this process (52). None of our studies separated the effects of the extracellular domain of MUC1 from MUC1-C per se, in that we overexpressed the intact protein. While PMIP is designed to block protein–protein interactions between MUC1 and EGFR and/or MUC1 and β-catenin, we cannot rule out the possibility that PMIP also affects the localization of the MUC1 or EGFR as a whole and, therefore, alters the interactions of MUC1 with proteins at its extracellular domain.

Downregulation of c-Met by PMIP showed a reduction in an important oncogene and metastasis promoter (53). In addition, c-Met is known to play a role in resistance of EGFR-positive cancers to EGFR-directed therapies, such as specific tyrosine kinase inhibitors and monoclonal antibodies. Bean and colleagues assessed c-Met expression in EGFR-expressing tumors that are resistant to EGFR-targeted therapies and compared them to EGFR-expressing tumors that had never been exposed to therapies. In their study, they found that 21% of the tumors with acquired resistance overexpressed c-Met, whereas only 3% of the unexposed tumors overexpressed c-Met (54). This study implicated a c-Met–dependent mechanism for acquiring resistance to EGFR-targeted therapies. Conversely, it has also been seen that EGFR activation can contribute to resistance to c-Met–directed therapies. Bachleitner–Hofmann and colleagues examined the effect of a c-Met kinase inhibitor on c-Met–overexpressing gastric cancer cell lines. They found that inhibition of c-Met prevented receptor tyrosine kinase cross-talk; specifically, EGFR and HER3 downstream signaling was reduced upon c-Met kinase inhibition. However, they found that stimulating cells with EGF effectively circumvented c-Met kinase inhibition by activating downstream signaling pathways, such as mitogen-activated protein kinase and phosphoinositide 3-kinase, describing one mechanism by which c-Met kinase inhibition can be thwarted (55). Therefore, it may prove that targeting the activity of both of these receptors may prevent this resistance from occurring.

J.A. Schroeder is also the chief scientific officer of Arizona Cancer Therapeutics and has Ownership Interest (including patents). No potential conflicts of interest were disclosed by the other authors.

Conception and design: T.M. Horm, B.G. Bitler, J.A. Schroeder

Development of methodology: T.M. Horm, J.A. Schroeder

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T.M. Horm, B.G. Bitler, D.M. Broka, J.M. Louderbough, J.A. Schroeder

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T.M. Horm, D.M. Broka, J.A. Schroeder

Writing, review, and/or revision of the manuscript: T.M. Horm, B.G. Bitler, J.A. Schroeder

Study supervision: J.A. Schroeder

The authors thank Matthew Hart and Hsin-Yuan Su for critical reading of this manuscript, and they are grateful to M.A. Hollingsworth for the MUC1 expression vector. CodeLink microarrays were provided by a grant from the University of Arizona Genomics Facility Core.

This work was supported by the University of Arizona Biochemistry and Molecular Biology Training Grant (T32GM008659) to T.M. Horm, Cancer Biology Training Grant (T32CA009213) to B.G. Bitler, the Arizona Cancer Center (NIH P30CA023074) to J.A. Schroeder, and the National Institutes of Health (NIH R01CA102113) to J.A. Schroeder.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.