Skip to main content
NCBI home page
Molecular and Cellular Biology logoLink to Molecular and Cellular Biology
. 2014 Sep;34(18):3486–3499. doi: 10.1128/MCB.00694-14

Cellular Migration and Invasion Uncoupled: Increased Migration Is Not an Inexorable Consequence of Epithelial-to-Mesenchymal Transition

Daneen Schaeffer a,b, Jason A Somarelli a,b, Gabi Hanna c, Gregory M Palmer c, Mariano A Garcia-Blanco a,b,d,e,✉
PMCID: PMC4135620  PMID: 25002532

Abstract

Metastatic dissemination requires carcinoma cells to detach from the primary tumor and invade through the basement membrane. To acquire these characteristics, epithelial tumor cells undergo epithelial-to-mesenchymal transitions (EMT), whereby cells lose polarity and E-cadherin-mediated cell-cell adhesion. Post-EMT cells have also been shown, or assumed, to be more migratory; however, there have been contradictory reports on an immortalized human mammary epithelial cell line (HMLE) that underwent EMT. In the context of carcinoma-associated EMT, it is not yet clear whether the change in migration and invasion must be positively correlated during EMT or whether enhanced migration is a necessary consequence of having undergone EMT. Here, we report that pre-EMT rat prostate cancer (PC) and HMLE cells are more migratory than their post-EMT counterparts. To determine a mechanism for increased epithelial cell migration, gene expression analysis was performed and revealed an increase in epidermal growth factor receptor (EGFR) expression in pre-EMT cells. Indeed, inhibition of EGFR in PC epithelial cells slowed migration. Importantly, while post-EMT PC and HMLE cell lines are less migratory, both remain invasive in vitro and, for PC cells, in vivo. Our study demonstrates that enhanced migration is not a phenotypic requirement of EMT, and migration and invasion can be uncoupled during carcinoma-associated EMT.

INTRODUCTION

During metastatic carcinoma progression, epithelial tumor cells acquire the ability to invade through the basement membrane by degrading through the extracellular matrix (ECM), intravasate into the circulatory system, extravasate into a distant site, and proliferate (1–4). The process by which carcinoma cells acquire the ability to invade is termed the epithelial-to-mesenchymal transition (EMT) (5). EMT is a normal program observed in mesoderm development and neural crest formation (6) as well as wound healing (7); however, carcinoma cells can coopt this series of biological processes to initiate the metastatic cascade (8). Carcinoma-associated EMT is characterized by the loss of several key epithelial phenotypes, including cobblestone-like morphology, tight cell-cell junctions, apical-basal polarity, and expression of epithelial biomarkers (9). In many cases, the increased invasive phenotype of EMT has been positively correlated with increased migration, with both thought of as hallmarks of EMT (10).

Similar to EMT, cellular migration occurs during development, normal physiology, including wound healing, and carcinoma progression (11). Migration is defined as the movement of cells from one area to another, with the type and speed of migration dependent on the extent of cell-cell and cell-matrix adhesion (12). Indeed, the tight cell-cell junctions of pre-EMT cells have been proposed to inhibit single-cell migration (13). A hallmark of EMT is increased invasion, which is defined as movement of cells through tissue and, in carcinoma-associated EMT, the basement membrane (14). Therefore, while invasion is dependent on migration, invasion requires different mechanics.

Studies of EMT in vitro have demonstrated that some post-EMT cells are more migratory and invasive than their pre-EMT counterparts (15–20). EMT can be induced by microenvironmental signals that ultimately result in transcriptional repression of E-cadherin (CDH1), which functions as an adherens junction protein and biomarker of epithelial cells (21). Of the transcription factors capable of repressing E-cadherin (22), and thus inducing EMT, exogenous expression of TWIST1, a basic helix-loop-helix (bHLH) protein, in an immortalized human mammary epithelial cell line (HMLE) is sufficient to increase migration, whereas knockdown of TWIST1 in the 4T1 mouse mammary carcinoma cell line decreased metastasis (i.e., invasion) (23). The positive correlation between increased migration and invasion has also been inferred in a gene expression study wherein motility genes were upregulated in invasive carcinoma cells (24). This and other in vitro data in the field of carcinoma-associated EMT have causally linked both increased migration and invasion to EMT. Two recent studies, however, have shown that the same HMLE-TWIST1 cell line that was reported to be highly migratory is less migratory than control epithelial HMLE cells (HMLE-vector) (25, 26). It is not yet known why this difference in TWIST1-specific cell migration exists.

In light of these contradictory findings, we designed a study to address whether undergoing EMT always results in increased migration and to understand the relationship between migration and invasion following EMT. Here, we demonstrate that acquisition of a mesenchymal cell state is not a prerequisite of a more migratory phenotype and that migration and invasion can act discordantly during carcinoma-associated EMT in vitro and in vivo. In addition to the HMLE cell line, we have also demonstrated these findings in a pre-EMT Dunning rat prostate adenocarcinoma cell line in vivo, which suggests that uncoupled migration and invasion are neither TWIST1 nor tissue specific. Unexpectedly, we found that the pre-EMT cell lines exhibited decreased cell-cell adhesion, which may contribute to increased epithelial cell migration. Gene expression analysis indicated an increase in epidermal growth factor receptor (EGFR) expression in the pre-EMT cell lines. Given that EGFR is a known modulator of cell migration, we treated the pre-EMT prostate cancer cell line with an EGFR inhibitor and observed a decreased migratory phenotype. Therefore, we propose that decreased cell-cell adhesion and EGFR activation are drivers of cellular migration in epithelial cells.

MATERIALS AND METHODS

Cell lines, plasmids, and stable transfections.

DT and AT3 cells (27) (kindly provided by W. McKeehan) and the HMLE cell lines (28) (kindly provided by R. Weinberg) were cultured as previously described. The pBabe-puro plasmid was obtained from Addgene (plasmid 1764) and was originally created by R. Weinberg (23).

Scratch wound assays.

Confluent monolayers were seeded in uncoated and collagen I-coated (Gibco; rat tail), collagen IV-coated (BD BioCoat; mouse), and fibronectin-coated (BD BioCoat; human) polystyrene 24-well tissue culture plates. After 16 h, a linear wound was generated using a sterile 20-μl pipette tip. Immediately after wounding, cells were imaged for 8 h by using a 6.4× objective on an inverted Olympus IX 71 epifluorescence microscope with a DP70 digital camera and analyzed with ImageJ (version 1.46r) (see Fig. 1B and D and 5B). For Movies S1 and S4 in the supplemental material, wounded cell monolayers were placed on an environment-controlled Zeiss Axiovert 200M microscope equipped with an Orca ER camera and imaged every 10 min for 12 h using a 20× bright-field objective. Time-lapse settings were controlled by MetaMorph (version 7.7.5; Molecular Devices). The images were compiled, and movies were created using ImageJ.

FIG 1.

FIG 1

Open in a new tab

Post-EMT AT3 cells migrate slower than pre-EMT DT cells in scratch wound assays. (A and B) Confluent monolayers grown on uncoated and coated polystyrene tissue culture dishes were artificially wounded, and wound closure was observed for 8 h. Wound closure speed was measured as the distance between the edges of the wound over 8 h. (C) Proliferation assays were performed using WST1, and absorbance was measured over 6 days. (D) Scratch wound closure was also measured on PET membranes of transwell inserts. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.01; ***, P < 0.0001.

FIG 5.

FIG 5

Open in a new tab

Post-EMT HMLE-TWIST1 cells migrate slower than epithelial HMLE-vector cells in vitro. (A and B) Confluent monolayers were artificially wounded and observed for 8 h. Wound closure speed was calculated as described for Fig. 1B. (C to E) As described for Fig. 3A to C. n = 144 cells for both HMLE-vector and HMLE-TWIST1. (F) As described for Fig. 1C. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.01; ***, P < 0.0001.

To determine the effects of EGFR signaling on cell migration, DT cells were serum deprived for 2 h, treated with 0.02% ethanol (vehicle) or 300 nM EGFR inhibitor AG1478 (Sigma) in 1% fetal bovine serum (FBS), and seeded into uncoated tissue culture plates. After 16 h, confluent monolayers were wounded, and fresh medium containing vehicle or AG1478 was added. Wound closure was imaged over 5 h and analyzed as described above (n = 6 replicates for DT and AT3 cells and n = 8 for HMLE cell lines on polystyrene and coated plates and for EGFR inhibition; n = 4 for DT and AT3 cells on polyethylene terephthalate [PET]).

WST1 proliferation assays.

For each cell line, 1,000 cells per well were plated in a 96-well plate. At each time point, 10 μl of the WST1 reagent (Roche) was added and incubated for 30 min. Plates were read at 450 nm using a microplate reader (BioTek; Synergy H1). Absorbance was normalized to time zero.

“In monolayer” migration assays.

Assays were performed and images were acquired as described previously (25) except that cells were imaged on an environment-controlled Zeiss Axiovert microscope 12 h after seeding, and time-lapse settings were controlled by MetaMorph. The images were compiled, and movies were created using Imaris (version 7.6; Bitplane). The following fields and numbers of cells were used to quantify individual cell migration and path length: n = 4 fields and 68 cells for DT; n = 4 fields and 106 cells for AT3; n = 12 fields and 144 cells for HMLE-vector; n = 6 fields and 144 cells for HMLE-TWIST1.

Modified Boyden chamber transwell assays.

Growth factor-containing medium was added to the lower chambers and 50,000 cells were added to the upper chambers of 24-well transwell plates (BD Biosciences) in growth factor-free or growth factor-containing medium. After 24 h, nonmigratory cells on the upper side of the inserts were removed. Migratory cells attached to the lower side of the inserts were fixed with 4% paraformaldehyde (PFA) for 15 min, permeabilized with 0.2% Triton X-100–1× phosphate-buffered saline (PBS) for 30 min, and Hoechst stained for 10 min. The inserts were washed with 1× PBS and imaged using a 4× objective on an inverted Olympus IX 71 epifluorescence microscope and analyzed with ImageJ (n = 4 replicates for the DT and AT3 cells; n = 3 for HMLE cells).

Matrigel invasion assays.

Matrigel invasion assays were performed using Matrigel-coated 24-well transwell plates (BD Biosciences) as described above. Migration-dependent invasion was calculated as the ratio of the percentage of cells that migrated through the Matrigel-coated membrane divided by the percentage of cells that migrated through the uncoated membrane. Migration-independent invasion was calculated as the percentage of cells that migrated through the Matrigel-coated membrane.

Cell-matrix adhesion assays.

Each well of an uncoated or ECM-coated 24-well plate was seeded with 50,000 cells. After a 1-h incubation, unbound cells were removed and adherent cells were fixed, permeabilized, and stained with Hoechst dye as described above. Cells were washed with 1× PBS and imaged using the 4× objective on an inverted Olympus IX 71 epifluorescence microscope (n = 4 replicates for all cell lines).

Cell-cell adhesion assays.

Assays were performed as described previously (29) except that cells were fixed in 2.5% PFA. Cells were imaged using the 20× bright-field objective on an inverted Olympus IX 71 epifluorescence microscope. Single cells, doublets, and cell clusters were counted as events (n = 4 replicates for all cell lines).

Subcutaneous tumor cell injection and tail vein injection.

DT and AT3 cells harbored a fluorescent reporter, pRIIIcI2 (30), and for DT cells, the pBabe-puro plasmid. DT (3 × 105) and AT3 (5 × 105) cells were injected subcutaneously into the right flank of 16 male syngeneic Copenhagen rats (Charles River, Wilmington, MA; 100 g, 5 to 6 weeks old; n = 8 rats for both cell lines) on different days. Rats injected with DT and AT3 cells were sacrificed at days 36 and 90 postinjection, respectively.

For tail vein injections, DT-pRIIIcI2-pBabe-puro (2 × 106) and naive AT3 (1 × 106) cells were injected into the tail vein of 13 nude rats (Charles River; n = 6 rats for DT cells and 7 for AT3 cells) on different days. Rats injected with DT and AT3 cells were sacrificed 20 and 14 days postinjection, respectively. All animal procedures were performed according to the Duke University Division of Laboratory Animal Resources (DLAR) and the Institutional Animal Care and Use Committee (IACUC) guidelines.

Window chamber implantation.

Assays were performed as described previously (31). A total of 7 × 103 DT and AT3 cells harboring a fluorescent reporter, pGint (32), were injected into the fascia. The window chamber was imaged every hour from three to 10 h postinjection using the 10× objective of a Zeiss Axio Observer D1 microscope equipped with a DVC 1500M camera. ImageJ was used to measure the x and y coordinates of cells and fixed points in the vasculature in pixels, and a calibration factor of 1.96 was used to convert the coordinates to microns (127.52 pixels = 250 μm). The following calculation was used to calibrate the x and y coordinates of cells to the respective coordinates of a fixed point in the vasculature: cell At=3x−vasculaturet=3x=calibrated cell At=3x, where x is the x coordinate of cell A, t = 3 is the first time point (3 h) postinjection of cell A, and “vasculature” is a fixed point in the vascular to be used as a reference. This calculation was also applied to the y coordinate for each cell at all time points.

To calculate speed, the following equation was used:

cell A speedt=4=(calibrated cell At=4x−calibrated cell At=3x)2+(calibrated cell At=4y−calibrated cell At=3y)2/1 h

where t = 3 is 3 h and t = 4 is 4 h postinjection. This calculation was applied to every cell at all time points by subtracting the previous hour. For average cell speed, the speeds of individual cells at all time points were averaged.

Antibodies, immunoblot analysis, and immunofluorescence.

Total cell lysate preparation, protein separation and transfer, and immunoblotting were performed as previously described (27). The following primary antibodies were used at the described dilution: β-actin (Santa Cruz; sc-47778) at 1:5,000, E-cadherin (BD Transduction; 610181) at 1:1,000, PTB (Intronn) (33) at 1:4,000, and vimentin (AbD Serotec; MCA862) at 1:500. For chemiluminescence-based detection, 5% milk-1× PBS was used as blocking buffer and to dilute primary and secondary antibodies. The SuperSignal West Pico Chemi substrate (Pierce) was used for detection, and membranes were washed between incubations with 1× PBS and exposed to Amersham Hyperfilm MP (GE Healthcare). Film was processed on a medical film processor (SRX-101A; Konica Minolta).

For immunofluorescence, cells were fixed, permeabilized, and stained with Hoechst dye as described above. Cells were blocked with 5% bovine serum albumin (BSA)-1× PBS for 30 min and incubated with primary antibodies for 1 h. The following primary antibodies were used at the described dilution: cellular fibronectin (Sigma; F6140) at 1:500, E-cadherin at 1:50, smooth muscle actin (Abcam; ab5694) at 1:100, TWIST1 (Abnova; H00007291-M01) at 1:50, and vimentin at 1:100. Cells were incubated in a 1:2,000 dilution of Alexa Fluor 647 secondary antibodies (goat anti-mouse or goat anti-rabbit antibody) for 1 h in the dark and then in a 1:2,000 dilution of Hoechst (Sigma) for 5 min in the dark. Cells were washed with 1× PBS between incubations and imaged on an inverted Olympus IX 71 epifluorescence microscope.

GSEA.

Publically available microarray data on HMLE-vector and HMLE-TWIST1 cells (34) was robust multichip average (RMA) normalized using RMAExpress (version 1.0.5) and compared to the “c2.all” curated gene sets in gene set enrichment analysis (GSEA) (35). Statistical significance in GSEA is expressed as an enrichment score (ES), which relies on a weighted Kolmogorov-Smirnoff metric, and a false discovery rate (FDR), which accounts for multiple hypothesis testing by computing random permutations of phenotype labels. The FDR was determined by using 1,000 gene set permutations to randomly assign phenotype labels to samples.

RNA isolation and reverse transcription quantitative PCR.

Total RNA was isolated using the ReliaPrep RNA cell miniprep system (Promega) according to the manufacturer's protocol. The Improm-II reverse transcription (RT) system (Promega) was used to synthesize random hexamer-primed cDNA from 1 μg of total RNA. Each 25-μl quantitative PCR (qPCR) mixture contained 1× Power SYBR green master mix (Life Technologies), 1 μl of a 1:5 dilution of cDNA, and 300 nM each primer. Rat-specific qPCR primers were designed using the Universal ProbeLibrary (Roche). The sequences of the forward and reverse EGFR primers are 5′ TGCACCATCGACGTCTACAT 3′ and 5′ AACTTTGGGCGGCTATCAG 3′, respectively. U1 snRNA was used for normalization, and the sequences of the forward and reverse primers are 5′ GGGAGATACCATGATCACGAAGGT 3′ and 5′ ATGCAGTCGAGTTTCCCACA 3′, respectively. Reactions were loaded into a 96-well plate and ran on the StepOnePlus real-time PCR system (Life Technologies). For relative quantitation, the comparative cycle threshold (ΔΔCT) method was used, and the fold change relative to DT cells was calculated by calculating the log base 2 of the ΔΔCT.

Statistical analyses.

JMP software (version 10.0) was used for statistical analyses. For comparisons of nonnormally distributed data, a nonparametric Wilcoxon two-sample test was used (see Fig. 10B). For multiple comparisons of normally distributed data, a one-way analysis of variance with Tukey's post hoc correction was performed (see Fig. 1C, 5F, and 9E). For all other comparisons of normally distributed data, Student's t test was used.

FIG 10.

FIG 10

Open in a new tab

Weakly migratory post-EMT AT3 cells are invasive in vivo. (A) A total of 3 × 105 and 5 × 105 DT and AT3 cells, respectively, were injected subcutaneously into the flank of male syngeneic Copenhagen rats. Tumors were weighed and measured 36 days postinjection. n = 8 rats for both DT and AT3 cells. (B) A total of 2 × 106 and 1 × 106 DT and AT3 cells were injected via the tail vein into male nude rats, respectively. Lungs were harvested and examined for macroscopic nodules 20 and 14 days postinjection with DT and AT3 cells, respectively. n = 6 rats for DT cells and n = 7 rats for AT3 cells. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.001.

FIG 9.

FIG 9

Open in a new tab

Pre-EMT DT cells migrate faster than post-EMT AT3 cells in vivo. (A) Dorsal skin fold window chamber in a nude rat. Top, posterior; bottom, anterior. (B and C) A total of 7 × 103 EGFP-labeled cells were injected into window chambers and imaged every hour, 3 to 10 h postinjection. Representative cell pairs are indicated with red asterisks. (D and E) The speed of individual cells was measured as the distance traveled over 10 h (D) and every hour (E). n = 13 cells for DT and AT3. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.01; ***, P < 0.0001.

RESULTS

A post-EMT Dunning rat prostate adenocarcinoma cell line, AT3, is less migratory than its epithelial counterpart in vitro.

Because enhanced migration has been shown to occur during EMT (36), we hypothesized that the post-EMT Dunning rat prostate carcinoma cell line, AT3, would be more migratory than the pre-EMT DT cell line. The epithelial DT cell line was derived from a spontaneous, well-differentiated, androgen-sensitive Dunning R-3327 rat prostate tumor, and the mesenchymal AT3 cell line was derived by in vivo passaging of DT cells in castrated male rats (37–41). Surprisingly, we found that AT3 cells migrated 8.8-fold slower than DT cells in a scratch wound assay on a standard uncoated polystyrene tissue culture dish (Fig. 1A and B). Furthermore, while DT cells formed a confluent monolayer, cells were able to migrate as single cells. In contrast, AT3 cells exhibited a spreading-like migratory phenotype, which suggested passive adhesion and cell spreading (42) (see Movie S1 in the supplemental material). Importantly, the slower migration of AT3 cells was not due to a decrease in proliferation (Fig. 1C).

To test whether or not the decreased migratory phenotype of AT3 cells was substrate specific, we performed scratch wound assays on polystyrene tissue culture dishes coated with several extracellular matrix (ECM) components, including collagen I, collagen IV, and fibronectin. Similar to migration on polystyrene, AT3 cells migrated 1.9- to 3.4-fold slower than DT cells on each substrate (Fig. 1B), suggesting that the slower migration of AT3 cells is not substrate specific. In addition, we tested migration on polyethylene terephthalate (PET) membranes, which are the substrates for migration and invasion transwell assays. Much the same as on other substrates, AT3 cells migrated 3.7-fold slower than DT cells (Fig. 1D).

Given the somewhat unexpected nature of these data, we reexamined the morphology and biomarker expression of DT and AT3 cell lines to confirm their epithelial and mesenchymal cell states. As expected, DT cells formed a cobblestone-like monolayer, whereas AT3 cells exhibited a rounded and spindle-like morphology. Consistent with being a pre-EMT line, DT cells expressed E-cadherin (CDH1) (Fig. 2A), particularly at cell-cell junctions, whereas AT3 cells exhibited only background staining (Fig. 2B). In contrast, AT3 cells expressed vimentin (VIM) (Fig. 2A and C). Based on this biomarker expression and that shown by others (27, 30, 38, 43–45), we conclude that our DT and AT3 cell lines are indeed epithelial and mesenchymal, respectively.

FIG 2.

FIG 2

Open in a new tab

Pre-EMT DT cells express E-cadherin at cell-cell junctions. (A to C) Immunoblot analysis of whole-cell lysates and immunofluorescence using antivimentin and anti-E-cadherin antibodies. (B) DT and AT3 cells were plated on uncoated and collagen I-, collagen IV-, and fibronectin-coated tissue culture plates. Scale bars represent 50 μm.

To assess individual cell migration in an unperturbed/intact monolayer, an “in monolayer” migration assay was used. Briefly, 5% of DT or AT3 cells were labeled with whole-cell tracking dye and plated with the same unlabeled cell type. Once confluent, the monolayer was imaged over 12 h, and individual cell speed and distance traveled were calculated. Similar to the scratch wound assays, AT3 cells migrated 1.9-fold slower (Fig. 3A and B; see also Movies S2 and S3 in the supplemental material) and traversed a 1.4-fold-shorter path length than DT cells (Fig. 3C). Both the DT and AT3 cells traveled at approximately half the speed in the “in monolayer” assay than in the scratch wound assay. These in vitro migration assays demonstrate that, although AT3 cells have undergone EMT, they are not as migratory as the pre-EMT DT cell line. While enhanced migration has been demonstrated in some post-EMT models, we propose that EMT and increased migration can be uncoupled.

FIG 3.

FIG 3

Open in a new tab

Post-EMT AT3 cells migrate slower than pre-EMT DT cells in monolayer. (A) Confluent monolayers containing cells labeled with cellular dye in a 1:20 mixture with unlabeled cells were observed every 10 min for 12 h. Centroids of labeled cells are represented as gray circles, and individual cell tracks were created in Imaris. (B and C) The speed (B) and path length (C) of individual cells were measured in Imaris. n = 68 cells for DT and 106 cells for AT3. The error bars shown represent the standard errors of the means. *, P ≤ 0.05.

Forced expression of TWIST1 in HMLE cells reduces cell migration in vitro.

The decreased migratory phenotype of post-EMT AT3 cells described above was similar to the observations of others in which HMLE cells that were driven to undergo EMT via ectopic expression of TWIST1 migrated slower than HMLE-vector cells (25, 26). We also wanted to test the migratory phenotype of the HMLE cell lines, but because HMLE-vector cells can undergo spontaneous EMT, we first checked biomarker expression to confirm that our HMLE-vector and HMLE-TWIST1 cell lines represented pre- and post-EMT cell lines, respectively. HMLE-vector cells expressed E-cadherin (Fig. 4A), whereas HMLE-TWIST1 cells expressed Twist1 (Fig. 4B), vimentin (Fig. 4A and B), and smooth muscle actin (Fig. 4B). Interestingly, HMLE-vector cells exhibited more cellular fibronectin staining than HMLE-TWIST1 cells. (Fig. 4B), which is consistent with what has been observed in intestinal (46, 47) and mammary (48) epithelial cells. The expression of these markers, particularly E-cadherin in HMLE-vector cells and vimentin in HMLE-TWIST1 cells, indicates that these cell lines were pre- and post-EMT.

FIG 4.

FIG 4

Open in a new tab

Post-EMT HMLE-TWIST1 cells express mesenchymal markers. (A) Immunoblot analysis of whole-cell lysates using antivimentin and anti-E-cadherin antibodies. (B) Immunofluorescence using anti-TWIST1, antivimentin, anti-smooth muscle actin, and anti-cellular fibronectin antibodies. Scale bars represent 50 μm.

Subsequent to characterizing biomarker expression in the HMLE cell lines, we tested their migratory phenotypes in scratch wound and “in monolayer” migration assays. Similar to what had been observed by others, HMLE-TWIST1 cells migrated 1.8-fold slower than HMLE-vector cells in both assays (Fig. 5A to D; see also Movies S4 to S6 in the supplemental material). In addition to migrating slower, HMLE-TWIST1 cells traversed a path length that was 1.6-fold shorter than that of HMLE-vector cells (Fig. 5E). Of note, the decreased migratory phenotype of HMLE-TWIST1 cells was not due to a decrease in proliferation (Fig. 5F). Additionally, passage number cannot explain the conflicting reports of cell migration in the HMLE cell lines, because both short-term and long-term passages were used. Therefore, the HMLE-vector and HMLE-TWIST1 cell lines represent a second example of decreased cell migration following EMT, further suggesting that increased migration is not an inexorable consequence of EMT.

Weakly migratory post-EMT cells exhibit more cell-cell adhesion than their pre-EMT counterparts.

To better understand the mechanics of the decreased migratory phenotype of post-EMT cells, we performed a cell-cell adhesion assay (29). Briefly, a tissue culture plate was precoated with 0.5% BSA to prevent cell-matrix attachment, and a single-cell suspension was added and agitated for 3 h. Interestingly, both AT3 and HMLE-TWIST1 cells formed more doublets and large cell clusters than DT and HMLE-vector, respectively, which largely remained as a single-cell suspension (Fig. 6A and B). This effect was particularly strong with post-EMT AT3 cells in which 49% of the cells remained as single cells compared to 90% of the DT cells (Fig. 6A). Of note, clusters of post-EMT cells were larger than the clusters formed in pre-EMT cell lines (data not shown). While these assays were done using biochemical disassociation (i.e., trypsin), similar results were obtained using mechanical disassociation (i.e., cell scraping [data not shown]). Indeed, the increased cell-cell contacts of AT3 cells were also observed in time-lapse microscopy of scratch wound assays, with AT3 cells maintaining cell-cell contacts as they extended projections into the void of the scratch wound (see Movie S1 in the supplemental material). Because AT3 and HMLE-TWIST1 cells expressed very low levels of the adherens junction protein E-cadherin (Fig. 2A and B and 4A), we suggest that these mesenchymal cell lines exhibit E-cadherin-independent cell-cell adhesion (see Discussion).

FIG 6.

FIG 6

Open in a new tab

Post-EMT cells exhibit increased cell-cell adhesion compared to their epithelial counterparts in vitro. (A and B) Single-cell suspensions were plated on BSA-coated tissue culture dishes and agitated for 3 h. Singlets, doublets, and clusters (three or more cells) were counted as events. Inset, higher magnification showing cell size. (C and D) Single-cell suspensions were plated on uncoated or ECM-coated tissue culture plates for 1 h. Unbound cells were removed, and bound cells were Hoechst stained and counted. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.01; ***, P < 0.0001.

While the extensive cell-cell adhesion displayed by these cell lines may partially explain the slowed migratory phenotypes of AT3 and HMLE-TWIST1 cells, the possibility exists that other mechanisms contribute to the decreased mesenchymal cell migration. To determine whether the slowed migratory phenotype of these cell lines also correlated with cell-matrix adhesion, DT, AT3, HMLE-vector, and HMLE-TWIST1 cells were seeded on tissue culture plates and coated with collagen I, collagen IV, and fibronectin. After 1 h, unbound cells were removed, and the cells that bound to the plate were Hoechst stained and quantified. We observed that AT3 cells were more adherent than DT cells to uncoated and collagen IV-coated plates (Fig. 6C), HMLE-vector cells were more adherent than HMLE-TWIST1 cells to uncoated and collagen IV-coated plates, and HMLE-TWIST1 cells were more adherent to fibronectin (Fig. 6D). Given the lack of a correlation between cell-matrix adhesion and migratory capacity, which was shown in Fig. 1B, we propose that cell-matrix adhesion does not drive the migration differences that we observe for these cell lines.

Despite being less migratory, AT3 and HMLE-TWIST1 cells are more invasive than their pre-EMT counterparts in vitro.

To compare migration and invasion in the four cell lines described here, we tested these models in modified Boyden chamber transwell assays. In this assay, cells migrate through an uncoated transwell insert or degrade through Matrigel to migrate through a Matrigel-coated PET membrane of a transwell insert. Migration-dependent invasion has been classically calculated as the ratio of cells that have migrated through Matrigel to the number of cells that have migrated through an uncoated membrane. In contrast to the conventional migration assays described above, the AT3 cell line was 3.5-fold more migratory than DT cells in this transwell assay (Fig. 7A) and exhibited only slightly more migration-dependent invasion than DT cells (Fig. 7B). Interestingly, despite the fact that DT cells were less migratory when passing through the PET membranes, the DT cells were more migratory than AT3 cells in scratch wound assays performed on PET membranes (Fig. 1D). Therefore, we propose that the difference between the transwell migration assay and the other in vitro migration assays is due to the mechanics of migrating through a porous membrane (see Discussion). When invasion was calculated independently of migration (only the percentage of cells that migrate through the Matrigel-coated membrane), 6.7-fold more AT3 cells were capable of migrating through the Matrigel-coated membrane (Fig. 7C). Because the rate of migration differed from invasion by 3.2-fold, the amount of migration-dependent invasion in vitro is diminished, which is likely an underestimate of invasion potential.

FIG 7.

FIG 7

Open in a new tab

Post-EMT cells migrate and invade differentially in a transwell assay. (A to F) Cells were plated on uncoated (A, B, D, and E) and Matrigel-coated (B, C, E, and F) transwell membranes. (A and D) Percent migration was calculated as the number of cells that migrated through uncoated membranes. (B and E) The ratio of migration through Matrigel-coated membranes to migration through uncoated membranes was calculated. (C and F) Migration through Matrigel was calculated as the number of cells that migrated through Matrigel-coated membranes. (D to F) HMLE cells were plated in normal growth medium (nonchemotactic) or in the absence of growth factors and induced to migrate toward growth factor-containing medium (chemotactic). Migration was normalized to DT or HMLE-vector. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.01; ***, P < 0.0001.

Because HMLE-TWIST1 cells were less migratory than HMLE-vector cells in scratch wound and “in monolayer” migration assays, we also tested their migration and invasion using transwell assays. HMLE-TWIST1 cells were originally reported to be more migratory toward growth factor-containing media than HMLE-vector cells in a transwell assay (23). Therefore, transwell assays were performed under chemotactic conditions (growth factor-free medium in the upper chamber and growth factor-containing medium in the lower chamber) and nonchemotactic conditions (growth factor-containing medium in the upper and lower chambers). Similar to the scratch wound and “in monolayer” migration assays, HMLE-TWIST1 cells were 3-fold less migratory than HMLE-vector cells under both conditions (Fig. 7D). While less migratory, HMLE-TWIST1 cells exhibited 8.8- and 14.5-fold more migration-dependent invasion than HMLE-vector cells under nonchemotactic and chemotactic conditions, respectively (Fig. 7E). When calculated independently of migration, HMLE-TWIST1 cells were more invasive (2.6-fold [nonchemotactic] and 4.8-fold [chemotactic]) (Fig. 7F), which demonstrates that their limited migratory capacity is still sufficient for invasion. Taken together, the data obtained from multiple in vitro migration and invasion assays suggest that the AT3 and HMLE-TWIST1 cell lines represent post-EMT models wherein migration and invasion are uncoupled.

Inhibition of EGFR decreases the migratory phenotype of DT cells in vitro.

To elucidate the mechanism of increased epithelial cell migration, we used gene set enrichment analysis (GSEA) (35) to compare publically available microarray data on HMLE cell lines (34) to curated gene sets in GSEA. Of the gene sets significantly enriched (false discovery rate [FDR] of <25%) in HMLE-vector cells, seven were involved in epidermal growth factor receptor (EGFR) signaling, which mediates cell migration, adhesion, and proliferation (49–51). We also observed 97-fold-higher levels of EGFR mRNA in DT cells than in AT3 cells (Fig. 8A). Therefore, we tested the effect of an EGFR inhibitor, AG1478, on DT cell migration in a scratch wound assay. Importantly, we observed a 1.9-fold decrease in wound closure upon treatment with AG1478 (Fig. 8B) that is independent of cell proliferation (Fig. 8C). Therefore, we propose that EGFR signaling contributes to the migratory phenotype of epithelial DT cells.

FIG 8.

FIG 8

Open in a new tab

Inhibition of EGFR reduces cell migration in pre-EMT DT cells. (A) Transcript levels of EGFR were quantified by qPCR. (B) Confluent monolayers of DT cells treated with ethanol (vehicle) or an EGFR inhibitor (AG1478) were artificially wounded, and wound closure was observed for 6 h. Wound closure speed was measured as described for Fig. 1B. (C) Proliferation assays were performed using WST1, and absorbance was read after wound closure in Fig. 1B. The error bars shown represent the standard errors of the means. *, P ≤ 0.05; **, P ≤ 0.01.

AT3 cells are less migratory but more invasive than DT cells in vivo.

All the above-given data were based on well-established, but in vitro, assays. To determine the migratory ability of DT and AT3 cells in vivo, we examined their behavior when injected into a dorsal skin fold window chamber (31). Briefly, a transparent window chamber was implanted in the dorsal skin fold of a nude rat by excising a circular piece of dorsal anterior skin, superficially injecting EGFP-labeled DT or AT3 cells into the fascia, and applying a glass coverslip over the injection site (Fig. 9A). The window chamber was continuously imaged over 10 h with the rat remaining under anesthesia on the microscope stage. Average cell speed over the course of the experiment, and at each time point, was calculated. In concordance with our observations using in vitro migration assays, AT3 cells migrated 2.3-fold slower than DT cells (Fig. 9B to E; see also Movies S7 and S8 in the supplemental material). Interestingly, both cell lines migrated significantly slower in vivo than in vitro, with AT3 cells migrating 2.1-fold slower and DT cells migrating 10-fold slower in vivo.

We compared the tumorigenic and metastatic behavior of DT and AT3 cells in vivo to determine whether their decreased migratory phenotype led to a decrease in tumorigenicity and/or metastasis. To assay tumorigenicity, male syngeneic Copenhagen rats were injected subcutaneously in the flank with 3 × 105 and 5 × 105 DT and AT3 cells, respectively. AT3 cells formed palpable tumors by day 7, whereas DT cells were indolent, with zero of six rats developing palpable tumors 90 days postinjection (Fig. 10A). The difference in tumorigenicity is not due to cell number, because we have previously injected 3 × 105 AT3 cells and obtained palpable tumors by day 7 (data not shown). To measure invasion in vivo, 2 × 106 and 1 × 106 DT and AT3 cells, respectively, were injected into the tail vein of male nude rats. Two weeks postinjection, rats injected with AT3 cells had lung nodules, whereas the lungs of rats injected with DT cells remained tumor-free for up to 20 days postinjection (Fig. 10B). While tail vein injection can report only on extravasation, a form of invasion, but not intravasation, we have also independently shown AT3 cells to be metastatic in a subcutaneous rat model (30). Therefore, the AT3 model suggests that an increased migratory phenotype is not required for tumor formation or metastatic colonization and that migration and invasion are not always positively correlated during EMT. Nonetheless, we cannot exclude the possibility that the lack metastatic colonization by DT cells is unambiguously attributable to their low invasiveness since they do not form primary tumors subcutaneously.

DISCUSSION

In the present study, we addressed the relationship between EMT, migration, and invasion using several in vitro and in vivo migration and invasion assays in two models of EMT. Surprisingly, our data demonstrate that cell migration actually decreased in our post-EMT cell lines. These data indicate that (i) migration does not always increase in post-EMT cells compared to that in pre-EMT cells and (ii) migration and invasion can be uncoupled.

The mechanism responsible for decreased mesenchymal cell migration in AT3 and HMLE-TWIST1 cells may be, in part, due to the increased cell-cell adhesion we observed. Migration speeds are likely to be enhanced by moderate increases in cell-cell adhesion but reduced by extremely strong cell-cell interactions (e.g., as those observed in polarized epithelium) (13, 14, 52). It is also possible that different types of cell-cell adhesion differentially affect migration rates. The epithelial adherens junction protein E-cadherin is strongly downregulated in our post-EMT cells, and therefore it is likely that other cell-cell adhesion proteins are responsible for the cell-cell contacts in AT3 and HMLE-TWIST1 cells. While the role of other cadherins in maintaining these contacts in our post-EMT cells has not been tested, OB-cadherin (CDH2) and N-cadherin (CDH11) are 2.6-fold and 14.04-fold upregulated in HMLE-TWIST1 cells, respectively (34). Additionally, using GSEA to compare microarray data from the HMLE cell lines to curated gene sets, we found significant enrichment of the REACTOME_ADHERENS_JUNCTIONS_INTERACTIONS gene set in HMLE-TWIST1 cells (FDR = 14.3%) (see Fig. S1 in the supplemental material), with transcripts encoding cadherins 2, 4, 18, 10, and 11 contributing to core enrichment. This replacement of E-cadherin by one or both of these cadherins has been observed in mesenchymal cells (53, 54) and has been proposed to play a role in collective or stream-like migration (55). Our observations are fully consistent with these findings.

Given that inhibition of EGFR in pre-EMT DT cells reduced migration, we propose that the mechanism of increased epithelial migration was due to increased EGFR expression and/or activation. Indeed, EGFR inhibition has been shown to reduce cell migration and wound healing in corneal (56) and intestinal (57) tissues, as well as in keratinocytes and non-small cell lung cancer (NSCLC) (58) cells. Similar to the results that we obtained with our epithelial carcinoma cell line, Lauand et al. reported that inhibition of EGFR reduced migration but not the proliferation rate of epithelial NSCLC cells (58). Therefore, we propose that decreased cell-cell adhesion and increased EGFR activation likely contribute to the increased migratory phenotype of our pre-EMT cells.

Cell migration can be measured using several complementary approaches in vitro and in vivo. We have shown that pre-EMT cells are more migratory than post-EMT cells in vivo, as well as in scratch wound assays on biological substrates and in “in monolayer” migration assays in vitro. In contrast, epithelial DT cells are less migratory than mesenchymal AT3 cells in a transwell assay. Because the scratch wound and “in monolayer” migration assays are in agreement with what we observe in vivo, we conclude that our pre-EMT cells are more migratory than their post-EMT counterparts in biologically relevant contexts. The discrepancy that we observed using the transwell assay underscores the importance of using multiple assays to measure migration.

While several studies have demonstrated that increased migration is often a phenotypic consequence of EMT (10), most used different cell systems, and therefore our work does not directly conflict with these findings. It is surprising, however, that the same HMLE-TWIST1 cell line was shown to be more migratory than HMLE-vector cells in one study (23), while a second and third group (25, 26), and now our work, showed the opposite. In the first study, migration was measured exclusively in transwell assays (23), whereas in the second and third studies, migration was measured using “in monolayer” migration assays (25, 26). To obtain the most comprehensive view of migration during EMT, we used these and other assays. The use of different assays cannot fully explain the difference in mesenchymal cell migration since HMLE-TWIST1 cells migrated slower than HMLE-vector cells in both “in monolayer” and transwell assays in our hands. However, we cannot exclude the possibility that the level of TWIST1 expression differed between studies.

Besides the AT3 and HMLE-TWIST1 cell lines, there are three additional examples in which epithelial cells exhibit a more migratory phenotype than mesenchymal cells. First, it is widely appreciated that epithelial keratinocytes migrate using rapid cell gliding (59), whereas mesenchymal fibroblasts migrate at low speed (60, 61). Indeed we have recently shown that keratinocytes migrating from skin explants retain expression of epithelial markers and are unlikely to undergo EMT (27). Second, the epithelial mouse breast cancer cell line 4T1 is more migratory than several mesenchymal breast cancer cell lines (62). Third, forced expression of epithelium-specific miR-200 family members in the mesenchymal mouse mammary tumor cell line 4TO7 induced partial MET and increased cell migration (63).

The traditional view of EMT involves an increased migratory phenotype. While the Dunning and HMLE cell lines likely represent unique examples of altered cell migration following EMT, the work presented here indicates that increased migration should not be considered a pathognomonic characteristic of EMT. Overall, these data and that of others (25, 26) suggest a reevaluation of the current phenotypes associated with pre- and post-EMT cells.

Supplementary Material

Supplemental material
supp_34_18_3486__index.html (2.5KB, html)

ACKNOWLEDGMENTS

This work was supported by a National Institutes of Health (NIH) R01 grant (5R01-CA127727) to M.A.G.-B. D.S. acknowledges an NIH T32 training grant (5T32-CA009111), an NIH F32 NRSA postdoctoral fellowship (F32-CA165482), and a Department of Defense Prostate Cancer Research Program postdoctoral training award (PC121324). J.A.S. acknowledges an American Cancer Society Postdoctoral Fellowship (PF-11-036-01-DDC).

We thank Andrew Armstrong and Shelton Bradrick for helpful discussions, Nicholas Barrows for assistance with statistics, and Kathryn Ware for her input on EGFR signaling. We acknowledge the Duke University Light Microscopy Core Facility and the Duke University Surgical and Optical Imaging Core Facility, shared resources of the Duke Cancer Institute.

We declare no conflicts of interest.

Footnotes

Published ahead of print 7 July 2014

Supplemental material for this article may be found at http://dx.doi.org/10.1128/MCB.00694-14.

REFERENCES

  • 1.Chambers AF, Naumov GN, Varghese HJ, Nadkarni KV, MacDonald IC, Groom AC. 2001. Critical steps in hematogenous metastasis: an overview. Surg. Oncol. Clin. N. Am. 10:243–255, vii [PubMed] [Google Scholar]
  • 2.Folkman J. 1992. The role of angiogenesis in tumor growth. Semin. Cancer Biol. 3:65–71 [PubMed] [Google Scholar]
  • 3.Fidler IJ. 1999. Critical determinants of cancer metastasis: rationale for therapy. Cancer Chemother. Pharmacol. 43(Suppl):S3–S10 [DOI] [PubMed] [Google Scholar]
  • 4.Woodhouse EC, Chuaqui RF, Liotta LA. 1997. General mechanisms of metastasis. Cancer 80:1529–1537 [DOI] [PubMed] [Google Scholar]
  • 5.Polyak K, Weinberg RA. 2009. Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat. Rev. Cancer 9:265–273. 10.1038/nrc2620 [DOI] [PubMed] [Google Scholar]
  • 6.Acloque H, Adams MS, Fishwick K, Bronner-Fraser M, Nieto MA. 2009. Epithelial-mesenchymal transitions: the importance of changing cell state in development and disease. J. Clin. Invest. 119:1438–1449. 10.1172/JCI38019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Nakamura M, Tokura Y. 2011. Epithelial-mesenchymal transition in the skin. J. Dermatol. Sci. 61:7–13. 10.1016/j.jdermsci.2010.11.015 [DOI] [PubMed] [Google Scholar]
  • 8.De Craene B, Berx G. 2013. Regulatory networks defining EMT during cancer initiation and progression. Nat. Rev. Cancer. 13:97–110. 10.1038/nrc3447 [DOI] [PubMed] [Google Scholar]
  • 9.Kalluri R, Weinberg RA. 2009. The basics of epithelial-mesenchymal transition. J. Clin. Invest. 119:1420–1428. 10.1172/JCI39104 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Tiwari N, Gheldof A, Tatari M, Christofori G. 2012. EMT as the ultimate survival mechanism of cancer cells. Semin. Cancer Biol. 22:194–207. 10.1016/j.semcancer.2012.02.013 [DOI] [PubMed] [Google Scholar]
  • 11.Horwitz R, Webb D. 2003. Cell migration. Curr. Biol. 13:R756–R759. 10.1016/j.cub.2003.09.014 [DOI] [PubMed] [Google Scholar]
  • 12.Friedl P, Wolf K. 2010. Plasticity of cell migration: a multiscale tuning model. J. Cell Biol. 188:11–19. 10.1083/jcb.200909003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Kedrin D, van Rheenen J, Hernandez L, Condeelis J, Segall JE. 2007. Cell motility and cytoskeletal regulation in invasion and metastasis. J. Mammary Gland Biol. Neoplasia 12:143–152. 10.1007/s10911-007-9046-4 [DOI] [PubMed] [Google Scholar]
  • 14.Friedl P, Gilmour D. 2009. Collective cell migration in morphogenesis, regeneration and cancer. Nat. Rev. Mol. Cell Biol. 10:445–457. 10.1038/nrm2720 [DOI] [PubMed] [Google Scholar]
  • 15.Kwok WK, Ling MT, Lee TW, Lau TC, Zhou C, Zhang X, Chua CW, Chan KW, Chan FL, Glackin C, Wong YC, Wang X. 2005. Up-regulation of TWIST in prostate cancer and its implication as a therapeutic target. Cancer Res. 65:5153–5162. 10.1158/0008-5472.CAN-04-3785 [DOI] [PubMed] [Google Scholar]
  • 16.Cheng GZ, Chan J, Wang Q, Zhang W, Sun CD, Wang LH. 2007. Twist transcriptionally up-regulates AKT2 in breast cancer cells leading to increased migration, invasion, and resistance to paclitaxel. Cancer Res. 67:1979–1987. 10.1158/0008-5472.CAN-06-1479 [DOI] [PubMed] [Google Scholar]
  • 17.Yang Z, Zhang X, Gang H, Li X, Li Z, Wang T, Han J, Luo T, Wen F, Wu X. 2007. Up-regulation of gastric cancer cell invasion by Twist is accompanied by N-cadherin and fibronectin expression. Biochem. Biophys. Res. Commun. 358:925–930. 10.1016/j.bbrc.2007.05.023 [DOI] [PubMed] [Google Scholar]
  • 18.Matsuo N, Shiraha H, Fujikawa T, Takaoka N, Ueda N, Tanaka S, Nishina S, Nakanishi Y, Uemura M, Takaki A, Nakamura S, Kobayashi Y, Nouso K, Yagi T, Yamamoto K. 2009. Twist expression promotes migration and invasion in hepatocellular carcinoma. BMC Cancer 9:240. 10.1186/1471-2407-9-240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Shiota M, Yokomizo A, Itsumi M, Uchiumi T, Tada Y, Song Y, Kashiwagi E, Masubuchi D, Naito S. 2011. Twist1 and Y-box-binding protein-1 promote malignant potential in bladder cancer cells. BJU Int. 108:E142–E149. 10.1111/j.1464-410X.2010.09810.x [DOI] [PubMed] [Google Scholar]
  • 20.Nakashima H, Hashimoto N, Aoyama D, Kohnoh T, Sakamoto K, Kusunose M, Imaizumi K, Takeyama Y, Sato M, Kawabe T, Hasegawa Y. 2012. Involvement of the transcription factor twist in phenotype alteration through epithelial-mesenchymal transition in lung cancer cells. Mol. Carcinog. 51:400–410. 10.1002/mc.20802 [DOI] [PubMed] [Google Scholar]
  • 21.Bussemakers MJ, van Bokhoven A, Mees SG, Kemler R, Schalken JA. 1993. Molecular cloning and characterization of the human E-cadherin cDNA. Mol. Biol. Rep. 17:123–128. 10.1007/BF00996219 [DOI] [PubMed] [Google Scholar]
  • 22.Vesuna F, van Diest P, Chen JH, Raman V. 2008. Twist is a transcriptional repressor of E-cadherin gene expression in breast cancer. Biochem. Biophys. Res. Commun. 367:235–241. 10.1016/j.bbrc.2007.11.151 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Yang J, Mani SA, Donaher JL, Ramaswamy S, Itzykson RA, Come C, Savagner P, Gitelman I, Richardson A, Weinberg RA. 2004. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 117:927–939. 10.1016/j.cell.2004.06.006 [DOI] [PubMed] [Google Scholar]
  • 24.Wang W, Goswami S, Lapidus K, Wells AL, Wyckoff JB, Sahai E, Singer RH, Segall JE, Condeelis JS. 2004. Identification and testing of a gene expression signature of invasive carcinoma cells within primary mammary tumors. Cancer Res. 64:8585–8594. 10.1158/0008-5472.CAN-04-1136 [DOI] [PubMed] [Google Scholar]
  • 25.Shapiro IM, Cheng AW, Flytzanis NC, Balsamo M, Condeelis JS, Oktay MH, Burge CB, Gertler FB. 2011. An EMT-driven alternative splicing program occurs in human breast cancer and modulates cellular phenotype. PLoS Genet. 7:e1002218. 10.1371/journal.pgen.1002218 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Kim HD, Meyer AS, Wagner JP, Alford SK, Wells A, Gertler FB, Lauffenburger DA. 2011. Signaling network state predicts twist-mediated effects on breast cell migration across diverse growth factor contexts. Mol. Cell. Proteomics 10:M111.008433. 10.1074/mcp.M111.008433 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Somarelli JA, Schaeffer D, Bosma R, Bonano VI, Sohn JW, Kemeny G, Ettyreddy A, Garcia-Blanco MA. 2013. Fluorescence-based alternative splicing reporters for the study of epithelial plasticity in vivo. RNA 19:116–127. 10.1261/rna.035097.112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Elenbaas B, Spirio L, Koerner F, Fleming MD, Zimonjic DB, Donaher JL, Popescu NC, Hahn WC, Weinberg RA. 2001. Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells. Genes Dev. 15:50–65. 10.1101/gad.828901 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Takeichi M. 1977. Functional correlation between cell adhesive properties and some cell surface proteins. J. Cell Biol. 75:464–474. 10.1083/jcb.75.2.464 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Oltean S, Sorg BS, Albrecht T, Bonano VI, Brazas RM, Dewhirst MW, Garcia-Blanco MA. 2006. Alternative inclusion of fibroblast growth factor receptor 2 exon IIIc in Dunning prostate tumors reveals unexpected epithelial mesenchymal plasticity. Proc. Natl. Acad. Sci. U. S. A. 103:14116–14121. 10.1073/pnas.0603090103 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Palmer GM, Fontanella AN, Shan S, Hanna G, Zhang G, Fraser CL, Dewhirst MW. 2011. In vivo optical molecular imaging and analysis in mice using dorsal window chamber models applied to hypoxia, vasculature and fluorescent reporters. Nat. Protoc. 6:1355–1366. 10.1038/nprot.2011.349 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Bonano VI, Oltean S, Garcia-Blanco MA. 2007. A protocol for imaging alternative splicing regulation in vivo using fluorescence reporters in transgenic mice. Nat. Protoc. 2:2166–2181. 10.1038/nprot.2007.292 [DOI] [PubMed] [Google Scholar]
  • 33.Wagner EJ, Carstens RP, Garcia-Blanco MA. 1999. A novel isoform ratio switch of the polypyrimidine tract binding protein. Electrophoresis 20:1082–1086. [DOI] [PubMed] [Google Scholar]
  • 34.Taube JH, Herschkowitz JI, Komurov K, Zhou AY, Gupta S, Yang J, Hartwell K, Onder TT, Gupta PB, Evans KW, Hollier BG, Ram PT, Lander ES, Rosen JM, Weinberg RA, Mani SA. 2010. Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc. Natl. Acad. Sci. U. S. A. 107:15449–15454. 10.1073/pnas.1004900107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. 2005. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U. S. A. 102:15545–15550. 10.1073/pnas.0506580102 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Thiery JP, Sleeman JP. 2006. Complex networks orchestrate epithelial-mesenchymal transitions. Nat. Rev. Mol. Cell Biol. 7:131–142. 10.1038/nrm1835 [DOI] [PubMed] [Google Scholar]
  • 37.Feng S, Wang F, Matsubara A, Kan M, McKeehan WL. 1997. Fibroblast growth factor receptor 2 limits and receptor 1 accelerates tumorigenicity of prostate epithelial cells. Cancer Res. 57:5369–5378 [PubMed] [Google Scholar]
  • 38.Yan G, Fukabori Y, McBride G, Nikolaropolous S, McKeehan WL. 1993. Exon switching and activation of stromal and embryonic fibroblast growth factor (FGF)-FGF receptor genes in prostate epithelial cells accompany stromal independence and malignancy. Mol. Cell. Biol. 13:4513–4522 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Dunning WF. 1963. Prostate cancer in the rat. Natl. Cancer Inst. Monogr. 12:351–369 [PubMed] [Google Scholar]
  • 40.Landstrom M, Damber JE, Bergh A. 1994. Prostatic tumor regrowth after initially successful castration therapy may be related to a decreased apoptotic cell death rate. Cancer Res. 54:4281–4284 [PubMed] [Google Scholar]
  • 41.Lubaroff DM, Canfield L, Reynolds CW. 1980. The Dunning tumors. Prog. Clin. Biol. Res. 37:243–263 [PubMed] [Google Scholar]
  • 42.McGrath JL. 2007. Cell spreading: the power to simplify. Curr. Biol. 17:R357–R358. 10.1016/j.cub.2007.03.057 [DOI] [PubMed] [Google Scholar]
  • 43.Wu X, Jin C, Wang F, Yu C, McKeehan WL. 2003. Stromal cell heterogeneity in fibroblast growth factor-mediated stromal-epithelial cell cross-talk in premalignant prostate tumors. Cancer Res. 63:4936–4944 [PubMed] [Google Scholar]
  • 44.Tennant TR, Kim H, Sokoloff M, Rinker-Schaeffer CW. 2000. The Dunning model. Prostate 43:295–302. [DOI] [PubMed] [Google Scholar]
  • 45.Oltean S, Febbo PG, Garcia-Blanco MA. 2008. Dunning rat prostate adenocarcinomas and alternative splicing reporters: powerful tools to study epithelial plasticity in prostate tumors in vivo. Clin. Exp. Metastasis 25:611–619. 10.1007/s10585-008-9186-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Kolachala VL, Bajaj R, Wang L, Yan Y, Ritzenthaler JD, Gewirtz AT, Roman J, Merlin D, Sitaraman SV. 2007. Epithelial-derived fibronectin expression, signaling, and function in intestinal inflammation. J. Biol. Chem. 282:32965–32973. 10.1074/jbc.M704388200 [DOI] [PubMed] [Google Scholar]
  • 47.Quaroni A, Isselbacher KJ, Ruoslahti E. 1978. Fibronectin synthesis by epithelial crypt cells of rat small intestine. Proc. Natl. Acad. Sci. U. S. A. 75:5548–5552. 10.1073/pnas.75.11.5548 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Taylor-Papadimitriou J, Burchell J, Hurst J. 1981. Production of fibronectin by normal and malignant human mammary epithelial cells. Cancer Res. 41:2491–2500 [PubMed] [Google Scholar]
  • 49.Chen P, Gupta K, Wells A. 1994. Cell movement elicited by epidermal growth factor receptor requires kinase and autophosphorylation but is separable from mitogenesis. J. Cell Biol. 124:547–555. 10.1083/jcb.124.4.547 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Xie H, Pallero MA, Gupta K, Chang P, Ware MF, Witke W, Kwiatkowski DJ, Lauffenburger DA, Murphy-Ullrich JE, Wells A. 1998. EGF receptor regulation of cell motility: EGF induces disassembly of focal adhesions independently of the motility-associated PLCgamma signaling pathway. J. Cell Sci. 111(Part 5):615–624 [DOI] [PubMed] [Google Scholar]
  • 51.Chen P, Xie H, Wells A. 1996. Mitogenic signaling from the EGF receptor is attenuated by a phospholipase C-gamma/protein kinase C feedback mechanism. Mol. Biol. Cell 7:871–881. 10.1091/mbc.7.6.871 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Vitorino P, Meyer T. 2008. Modular control of endothelial sheet migration. Genes Dev. 22:3268–3281. 10.1101/gad.1725808 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Armstrong AJ, Marengo MS, Oltean S, Kemeny G, Bitting RL, Turnbull JD, Herold CI, Marcom PK, George DJ, Garcia-Blanco MA. 2011. Circulating tumor cells from patients with advanced prostate and breast cancer display both epithelial and mesenchymal markers. Mol. Cancer Res. 9:997–1007. 10.1158/1541-7786.MCR-10-0490 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Tomita K, van Bokhoven A, van Leenders GJ, Ruijter ET, Jansen CF, Bussemakers MJ, Schalken JA. 2000. Cadherin switching in human prostate cancer progression. Cancer Res. 60:3650–3654 [PubMed] [Google Scholar]
  • 55.Theveneau E, Mayor R. 2012. Cadherins in collective cell migration of mesenchymal cells. Curr. Opin. Cell Biol. 24:677–684. 10.1016/j.ceb.2012.08.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Zieske JD, Takahashi H, Hutcheon AE, Dalbone AC. 2000. Activation of epidermal growth factor receptor during corneal epithelial migration. Invest. Ophthalmol. Vis. Sci. 41:1346–1355 [PubMed] [Google Scholar]
  • 57.Polk DB. 1998. Epidermal growth factor receptor-stimulated intestinal epithelial cell migration requires phospholipase C activity. Gastroenterology 114:493–502. 10.1016/S0016-5085(98)70532-3 [DOI] [PubMed] [Google Scholar]
  • 58.Lauand C, Rezende-Teixeira P, Cortez BA, Niero EL, Machado-Santelli GM. 2013. Independent of ErbB1 gene copy number, EGF stimulates migration but is not associated with cell proliferation in non-small cell lung cancer. Cancer Cell Int. 13:38. 10.1186/1475-2867-13-38 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Keren K, Pincus Z, Allen GM, Barnhart EL, Marriott G, Mogilner A, Theriot JA. 2008. Mechanism of shape determination in motile cells. Nature 453:475–480. 10.1038/nature06952 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Friedl P, Zanker KS, Brocker EB. 1998. Cell migration strategies in 3-D extracellular matrix: differences in morphology, cell matrix interactions, and integrin function. Microsc. Res. Tech. 43:369–378. [DOI] [PubMed] [Google Scholar]
  • 61.Pankova K, Rosel D, Novotny M, Brabek J. 2010. The molecular mechanisms of transition between mesenchymal and amoeboid invasiveness in tumor cells. Cell. Mol. Life Sci. 67:63–71. 10.1007/s00018-009-0132-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Lou Y, Preobrazhenska O, auf dem Keller U, Sutcliffe M, Barclay L, McDonald PC, Roskelley C, Overall CM, Dedhar S. 2008. Epithelial-mesenchymal transition (EMT) is not sufficient for spontaneous murine breast cancer metastasis. Dev. Dyn. 237:2755–2768. 10.1002/dvdy.21658 [DOI] [PubMed] [Google Scholar]
  • 63.Dykxhoorn DM, Wu Y, Xie H, Yu F, Lal A, Petrocca F, Martinvalet D, Song E, Lim B, Lieberman J. 2009. miR-200 enhances mouse breast cancer cell colonization to form distant metastases. PLoS One 4:e7181. 10.1371/journal.pone.0007181 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental material
supp_34_18_3486__index.html (2.5KB, html)
MCB.00694-14_zmb999100574so1.pdf (216.8KB, pdf)
MCB.00694-14_zmb999100574so2.pdf (68.8KB, pdf)
Download video file (6.3MB, avi)
Download video file (2.1MB, mov)
Download video file (1.7MB, mov)
Download video file (6.9MB, avi)
Download video file (1.8MB, mov)
Download video file (2MB, mov)
Download video file (277.1KB, avi)
Download video file (249.7KB, avi)

Articles from Molecular and Cellular Biology are provided here courtesy of Taylor & Francis

RESOURCES