E XP ER I ME NTAL C E LL RE S E ARCH
] (]]]]) ]]]–]]]
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/yexcr
Research Article
Cellular contractility changes are sufficient to drive epithelial scattering Jacob P. Hoj, John A. Davis, Kendra E. Fullmer, David J. Morrell, Nicholas E. Saguibo, Jeffrey T. Schuler, Kevin J. Tuttle, Marc D.H. Hansenn Physiology and Developmental Biology Brigham Young University, 574 WIDB Provo, UT 84602, USA
article information
abstract
Article Chronology:
Epithelial scattering occurs when cells disassemble cell–cell junctions, allowing individual epithelial
Received 24 May 2013
cells to act in a solitary manner. Epithelial scattering occurs frequently in development, where it
Received in revised form
accompanies epithelial–mesenchymal transitions and is required for individual cells to migrate and
10 April 2014
invade. While migration and invasion have received extensive research focus, how cell–cell junctions
Accepted 14 April 2014
are detached remains poorly understood. An open debate has been whether disruption of cell–cell interactions occurs by remodeling of cell–cell adhesions, increased traction forces through cell
Keywords:
substrate adhesions, or some combination of both processes. Here we seek to examine how changes
Epithelial scattering
in adhesion and contractility are coupled to drive detachment of individual epithelial cells during
Epithelial mesenchymal transition
hepatocyte growth factor (HGF)/scatter factor-induced EMT. We find that HGF signaling does not
Adhesion
alter the strength of cell–cell adhesion between cells in suspension, suggesting that changes in cell–
Cytoskeleton
cell adhesion strength might not accompany epithelial scattering. Instead, cell–substrate adhesion
Contractility
seems to play a bigger role, as cell–substrate adhesions are stronger in cells treated with HGF and
Myosin
since rapid scattering in cells treated with HGF and TGFβ is associated with a dramatic increase in focal adhesions. Increases in the pliability of the substratum, reducing cells ability to generate traction on the substrate, alter cells' ability to scatter. Further consistent with changes in substrate adhesion being required for cell–cell detachment during EMT, scattering is impaired in cells expressing both active and inactive RhoA mutants, though in different ways. In addition to its roles in driving assembly of both stress fibers and focal adhesions, RhoA also generates myosin-based contractility in cells. We therefore sought to examine how RhoA-dependent contractility contributes to cell–cell detachment. Inhibition of Rho kinase or myosin II induces the same effect on cells, namely an inhibition of cell scattering following HGF treatment. Interestingly, restoration of myosin-based contractility in blebbistatin-treated cells results in cell scattering, including global actin rearrangements. Scattering is reminiscent of HGF-induced epithelial scattering without a concomitant increase in cell migration or decrease in adhesion strength. This scattering is dependent on RhoA, as blebbistatin-induced scattering is reduced in cells expressing dominant-negative RhoA mutants. This suggests that induction of myosin-based cellular contractility may be sufficient for cell–cell detachment during epithelial scattering. & 2014 Elsevier Inc. All rights reserved.
n
Corresponding author. Fax: þ801 422 0700. E-mail address: [email protected] (M.D.H. Hansen).
http://dx.doi.org/10.1016/j.yexcr.2014.04.011 0014-4827/& 2014 Elsevier Inc. All rights reserved.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
2
E XP E RI ME N TAL CE L L R ES E ARC H
Introduction Strong cell–cell junctions integrate individual epithelial cells into tissues, permitting coordination of cellular processes. During morphogenetic programs, cell–cell interactions are remodeled in order to allow individual cells to either reposition with respect to the tissue [1]. In some cases, complete ablation of cell–cell adhesions is enacted, which allows cells to fully detach form the epithelium and move independently [2]. Full detachment of individual epithelial cells occurs during epithelial–mesenchymal transitions, where epithelial cells scatter and migrate to distant sites to establish new organs and tissues [3]. It is thought that inappropriate activation of this morphogenetic program drives cancer invasion and metastasis [4]. Scattering of epithelial cells, resulting from rupture of cell–cell adhesions, is accompanied by changes in cell adhesion, motility, and cellular contractility. It has been proposed that cell–cell junctions become actively disassembled, which is required for their subsequent rupture during epithelial scattering. A contrasting view is that scattering cells increase migration and contractility to physically tear cell–cell junctions apart and that this does not require significant molecular changes in cell–cell junction composition or architecture [5]. The requirement for molecular changes in cell–cell adhesions is supported by the observation that cadherin expression changes accompany EMT [6]. Further, dramatic changes in actin organization at cell–cell junctions accompany scattering [7]. A number of actin regulatory proteins are redistributed with respect to cell–cell contacts during scattering, including those proposed to link actin to an adhesion system, like zyxin, vinculin, and α-actinin [7]. Importantly, positively or negatively targeting the function of zyxin prevents or facilitates, respectively, the rupturing of cell–cell contacts [7]. Support for a model where tension forces associated with the traction of cell migration drives scattering comes from observations that detachment of cell–cell adhesions appears to precede many changes in gene transcription and protein modification that are typically associated with EMT events [8]. Contractile forces are associated with epithelial scattering; transcellular networks assembled from actin that was formerly associated with cell–cell contacts become highly decorated with myosin II and are associated with deformations of the plasma membrane [7]. Importantly, the actin regulatory proteins that become redistributed with respect to cell–cell junctions include those (zyxin and VASP) that are known to relocalize to sites of actin-membrane connection in response to tension forces [9]. Further, changes in extracellular matrix type dramatically alter epithelial scattering [10,11], demonstrating that integrin-based adhesions play a fundamental role in epithelial scattering. A model that emphasizes contraction forces as a means of driving scattering suggests that actin remodeling with respect to cell–cell junctions is carried out to focus contractile forces and not to alter the strength of cell–cell interactions. Prior work from our group has focused on actin rearrangements associated with epithelial scattering, including the mechanism and function of zyxin in anchoring actin at site of cell–cell adhesion. Our group has worked under the assumption that scattering requires priming by altering the molecular composition and architecture of cell–cell adhesion systems. Here we perform a series of experiments that are designed to address the contribution of changes in cell–cell adhesion, cell–substrate adhesion, and
] (]]]]) ]]]–]]]
cellular contractility in driving epithelial scattering. We find that cell substrate adhesion changes are more dramatic than cell–cell adhesion changes during early scattering. Further, cellular contractility plays a central role in driving scattering. Surprisingly, forced alterations in cellular contractility in groups of epithelial cells is sufficient to drive detachment and scattering in the absence of a signal for EMT. Based on these results, we argue that detachment of cell–cell adhesions during epithelial scattering can be driven by changes in cellular contractility alone.
Materials and methods Cell culture MDCK cells were cultured in low glucose DMEM supplemented with 10% fetal bovine serum. Cells bearing tetracycline-repressible cassettes were also maintained in 20 ng/ml doxycycline to maintain suppression of cassette genes, the RhoA G14V and T19N mutants. Stimulation of cells with HGF and/or TGFβ (Antigenix America) was performed at a final dose of 25 ng/ml and 0.1 ng/ml, respectively. Blebbistatin and Rho kinase inhibitor VII (Calbiochem) were used at a final concentration of 20 μM and 1.0 μM, respectively. To generate flexible substrates, the protocol reported by Wang and Pelham [12] was adapted. Briefly, glass coverslips were etched with a dried film of sodium hydroxide, silanized with 3-aminopropyltrimethoxysilane, and activated for covalent linkage with 0.5% (v/v) glutaraldehyde. A thin polyacrylamide gel, made with 8% acrylamide and within a range of 0.03–0.26% BIS, was polymerized onto this activated surface. The polyacrylamide gel was activated for covalent linkage of extracellular matrix by addition of 1 mM sulfo-SANPAH and exposure to UV light for 5 min, then removal of remaining sulfoSANPAH solution and overnight incubation with collagen I. Surfaces were washed with PBS and medium prior to seeding with cells.
Cell-based assays Cellular aggregation assays were performed as reported [13]. Briefly, cells were trypsinized and placed in suspension at 250,000 cells per ml. 20 μl drops were placed in the inner lid of a petri dish and then incubated for 0–4 h. Drops were either spread onto a glass slide for analysis directly following culturing or after trituration of the cell suspension, which disrupts weak adhesions with cellular aggregates. The proportion of the total number of cells in small (o10 cells), medium (11–50 cells), and large (450 cells) aggregates was quantified for each condition. For each condition the experiment was performed in triplicate. Cell migration was determined by seeding cells into the wells of a 96 well Oris I (Platypus) cell migration plate and culturing the cells overnight. The stopper was removed and the progression of cell migration into the now open region of each well was monitored by microscopy at the indicated timepoint, generally 0, 6 and 12 h. The size of the open area covered by cells was then determined using SlideBook software.
Live cell imaging and analysis For live cell imaging experiments, collagen-coated Delta-T imaging chambers (Bioptechs) were seeded with 10,000 MDCK cells and cultured overnight, allowing formation of small colonies.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
Prior to imaging, media was changed to HEPES buffered DMEM supplemented with 10% fetal bovine serum. Cells were treated with growth factors and/or inhibitors as appropriate for the experiment. A heated microscope stage using Delta T4 temperature regulation system (Bioptechs) maintained temperature at 37 1C. Multiple positions were imaged for each experiment and each experiment was repeated, with identical parameters, in triplicate. Phase contrast images of cell scattering were taken at 16 magnification using a 10 (0.30 aperture) objective and 1.6 slider, with 30 ms exposure time. Images were recorded every 2 min for 12 h. The first 10 h of each series was used for analysis, with the remaining time demonstrating the continued health of the culture. Quantitative image analysis was performed using SlideBook software. Cell–cell rupture events were quantified as the cumulative number of spaces appearing within colonies of MDCK cells during the first 6 h of imaging. For each colony, the cumulative number of rupture events was normalized to number of cells in the colony to reflect more or less
opportunities for rupture events as a function of colony size. Complete cell detachments were quantified as the cumulative occurrence of individual cells completely breaking contact form the other cells within a colony during the first 6 h of imaging. For each colony analyzed, the cumulative count was normalized to the number of cells in the colony to reflect increased opportunities for such detachments in larger colonies. For statistical analysis, the data was fit to a generalized linear mixed model using R. Because of the skewed distribution and high number of zeros in the count data, the Poisson distribution was used. A quadratic term was added where appropriate. P-values were then generated via a likelihood-ratio test.
Immunofluorescence For immunofluorescence imaging of fixed samples, 10,000 MDCK cells were seeded onto collagen-coated coverslips in 12 well plates and cultured overnight. Cells were treated with growth
With trituration 100%
80%
80%
2 hour culture
2 hour culture
Without trituration 100%
60% 40% 20% 0%
3
] (]]]]) ]]]–]]]
60% 40% 20% 0%
HGF
TGFB
Combo
Control
HGF
TGFB
Large 50+
13.17%
14.18%
20.01%
30.88%
Large 50+
0.00%
0.00%
4.43%
0.00%
Medium11-50
52.10%
24.59%
51.29%
37.50%
Medium11-50
47.84%
49.65%
45.46%
62.48%
Small 1-10
34.73%
61.24%
28.70%
31.62%
Small 1-10
52.16%
50.35%
50.11%
37.52%
Combo
100%
100%
80%
80%
3 hour culture
3 hour culture
Control
60% 40% 20% 0%
Combo
60% 40% 20% 0%
Control
HGF
TGFB
Combo
Control
HGF
TGFB
Large 50+
90.35%
84.43%
96.98%
96.23%
Large 50+
22.38%
7.32%
5.76%
5.39%
Medium11-50
7.72%
6.42%
2.76%
1.98%
Medium11-50
67.24%
75.14%
89.44%
87.78%
Small 1-10
1.93%
9.15%
0.92%
1.79%
Small 1-10 100%
10.38%
17.55%
4.81%
6.84%
*
80%
4 hour culture
4 hour culture
100%
60% 40%
*
20% 0%
80% 60% 40% 20% 0%
Control
HGF
TGFB
Combo
Large 50+
95.26%
50.51%
77.73%
82.01%
Large 50+
Control
HGF
TGFB
Combo
32.53%
37.25%
30.96%
25.80%
Medium11-50
2.88%
36.01%
14.46%
13.93%
Medium11-50
54.23%
47.31%
53.05%
53.04%
Small 1-10
1.86%
13.48%
7.81%
4.06%
Small 1-10
13.23%
15.44%
15.99%
21.16%
Fig. 1 – Cell–cell adhesion changes during epithelial scattering. Distribution of an MDCK cell population into small, medium, and large aggregates after the indicated amount of time in suspension in inverted droplets of medium and following the indicated growth factor treatment. Cells were subjected to analysis with and without trituration, which disrupts weak adhesions between contacting cells. Asterisks indicate instances where a statistically significant difference (Po 0.05) in the proportion of cells occurring in aggregates is observed between a treatment condition and untreated control cells. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
4
E XP E RI ME N TAL CE L L R ES E ARC H
factors and/or inhibitors as appropriate for the individual experiment. Cells were fixed by brief washing in ice cold Ringer's saline and incubated with 3.7% paraformaldehyde on ice. After fixation, cells were rinsed, blocked with PBS buffer containing 0.4% BSA and normal goat serum, and stained using antibodies, DAPI, and/or Alexa fluor-conjugated phalloidin. Antibody staining was performed against myosin IIa (Novus Biologicals) and vinculin (Cytoskeleton), as well as Alexa fluor-conjugated secondary antibodies.
Western blotting Cells were washed with ice cold PBS then harvested and lysed with Buffer A (50 mM Tris, 2% (w/v) SDS, 5% (v/v) glycerol, 1% 2-mercaptoethanol, 5 mM sodium orthovanadate, and 5 mM protease inhibitor cocktail). Equalized concentrations of protein lysates – determined via Bradford assay – were run in triplicate in 8–12% SDS–polyacrylamide gels and transferred to nitrocellulose membranes. The membranes were incubated with anti-phospho-MLC (Bioss) or anti-actin (Sigma-Aldrich) primary antibodies, followed by HRP-conjugated secondary antibodies. Antibody binding was detected using the enhanced chemiluminescence and autoradiography film. Densitometry was performed using the Fluorchem SP imaging system.
Results Changes in cell adhesion during HGF-induced epithelial scattering In order to better understand how cell–cell and cell–substrate adhesion might individually contribute to epithelial scattering
] (]]]]) ]]]–]]]
during EMT, we measured adhesion changes after stimulation of MDCK cells with HGF. MDCK cells are a well-established model system for studying epithelial cell biology [14], and for epithelial scattering specifically [15]. These cells retain highly epithelial characteristics, recapitulating the original tissue organization and function under the correct culture conditions. The response of MDCK cells to HGF has been documented at a high level of resolution [7] and the behavior of scattering cells under various conditions well studied [16]. Also, since the cells originate from normal tissue, normal cellular signaling and mechanics are expected to be retained in this cell line. MDCK cells were cultured in the presence or absence of additional growth factors, then assessed for cell–cell and cell– substrate adhesion using cell based assays. For these experiments, cells were stimulated with growth factors for 24 h in culture prior to analysis in each assay system. Growth factors used were HGF and TGFβ. HGF was used to induce characteristic epithelial scattering and EMT in the cultures of MDCK cells. The addition of TGFβ increases epithelial scattering during HGF induced EMT [17], so the addition of this second factor was used to examine how cells change adhesion properties under conditions where scattering is accelerated beyond what is observed with treatment of HGF alone. We measured cell–cell adhesion by placing cells in suspension in an inverted drop of medium, which results in cells settling into close proximity and the forming aggregates through cell–cell junction formation. The ratio of cells in aggregates of different sizes is tracked over time revealing the cells' propensity to form cell–cell junctions (Fig. 1). When aggregates are triturated prior to analysis, weaker cell–cell junctions are disrupted and only aggregates with strong cell–cell junctions remain allowing a determination of the strength of newly formed cell–cell adhesions. We analyzed aggregate sizes of MDCK cells taken from cultures grown
untreated
TGFβ
HGF
HGF +TGFβ
Fig. 2 – Changes to focal adhesions during epithelial scattering. Phalloidin (red), DAPI (blue), and vinculin antibody (green) staining of MDCK cells treated with indicated growth factors for 2 h. Scale bar ¼30 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
in the presence and absence of both HGF and TGFβ. Cell aggregates were analyzed after 2–4 h in suspension for all conditions. We find that untreated control MDCK cells are found in larger and larger aggregates with increasing amounts of time. By four hours, almost all cells are found in the largest aggregates. Aggregates become increasingly resistant to trituration forces as time progresses, as few large aggregates get disrupted into smaller sizes after four hours in suspension. Interestingly, the pattern of aggregate formation, whether the analysis was done with or without trituration, was similar for cells grown under all culture conditions. The only statistically significant difference is at a single timepoint under single condition for a single treatment, namely a reduction in aggregate size at 4 h in cells treated with HGF alone. We next sought to determine how growth factor treatments alter cell–substrate adhesion of MDCK cells. Cells were stimulated
0.24% bis
5
] (]]]]) ]]]–]]]
to undergo EMT with growth factor treatment and the number and distribution of focal adhesions examined. Cells on coverslips and treated with or without growth factors were fixed and stained with vinculin antibodies, revealing that focal adhesion vastly increased in cells treated with HGF and TGFβ (Fig. 2), suggesting that epithelial scattering is driven largely through cell– substrate adhesions and not changes in cell–cell adhesion.
Changes to cell–substrate adhesion alter scattering behavior during EMT Given the correlation of cell–substrate adhesions with HGFinduced epithelial scattering, we next sought to examine the contribution of cell–substrate adhesion to epithelial scattering. Cell–substrate adhesion plays a role in cellular contractility, cell spreading, and cell migration, all of which are increased during
0.12% bis
0.06% bis
0:00
0:00
0:00
2:00
2:00
2:00
4:00
4:00
4:00
6:00
6:00
6:00
Fig. 3 – Epithelial scattering on pliant substrates. Timelapse images of MDCK cells undergoing HGF-induced epithelial scattering on polyacrylamide surfaces generated with increased flexibility, a result of decreasing amounts of crosslinking with bis acrylamide. Time stamps are in hours:minutes. Scale bar¼ 50 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
6
E XP E RI ME N TAL CE L L R ES E ARC H
] (]]]]) ]]]–]]]
RhoAT19N +dox
RhoA G14V -dox
+dox
-dox
0:00
0:00
0:00
0:00
1:00
1:00
1:00
1:00
2:00
2:00
2:00
2:00
3:00
3:00
3:00
3:00
4:00
4:00
4:00
4:00
Fig. 4 – Epithelial scattering in cells expressing RhoA mutants. Timelapse images of MDCK cells expressing constitutively active (G14V) or dominant negative (T19N) RhoA mutants and undergoing epithelial scattering in response to HGF stimulation. Mutant expression is negatively regulated by the presence of doxycycline (dox). Time stamps are in hours:minutes. Scale bar¼ 50 μm.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
RhoA is required for cell–cell detachment in response to HGF stimulation Given that changes in cell–substrate flexibility properties affect epithelial scattering, we next sought to perturb cell–substrate adhesion systems in another manner and examine the effects on HGF-induced epithelial scattering. To accomplish this we used MDCK cells expressing RhoA mutants. We used MDCK cells that express RhoA G14V and T19N mutants under control of a tetracycline repressor. These cells have been extensively characterized, particularly in terms of how these mutants alter actin organization and cell–cell adhesion [19,20]. Cells were plated onto collagen-coated imaging dishes and cultured in the presence or absence of doxycycline, blocking or permitting mutant expression.
Cells were stimulated with HGF and subjected to time-lapse imaging and quantitative analysis. Imaging (Fig. 4) reveals that in the presence of doxycycline, cells bearing either RhoA mutant undergo epithelial scattering that is typical of normal MDCK cells. In contrast, cells expressing a dominant negative RhoA T19N mutant fail to undergo epithelial scattering. Following HGF stimulation, these cells undergo spreading and increase their migration within the colony, but fail to rupture cell–cell junctions. In experiments imaging HGFinduced scattering in MDCK cells expressing this RhoA mutant, rupturing of cell–cell adhesions was only seen in one or two instances. Meanwhile, cells expressing a constitutively active RhoA G14V mutant undergo scattering that is empirically similar to that observed in control cells; mutant-expressing cells respond to HGF stimulation by undergoing cell spreading that is followed by increased migration and the rupturing of cell–cell adhesions. However, spreading is reduced and ruptures of cell–cell adhesion appear with a significant delay. Quantitative image analysis of scattering by MDCK cells expressing RhoA mutants largely confirm these empirical results (Fig. 5). In blind analysis, we tracked the cumulative number of cells detaching from colonies of MDCK cells stimulated with HGF, then
Detachment events/ cell
1.2
RhoA T19N cells Treatment −dox +dox
0.8
-16
p < 2.2x10
0.4
0.0
0
100
200
300
Time (Minutes)
Detachment events/ cell
epithelial scattering. To assess the roles of migration and spreading, we prepared flexible substrates on which cells could be plated and induced to undergo scattering. On more pliant substrates, cell migration and cell spreading is limited significantly, providing an opportunity to examine how changes in these cellular behaviors contributes to scattering. Glass coverslips were covalently linked with a thin polyacrylamide gel that was in turn covalently decorated with collagen I. Cellular attachment to the extracellular matrix link cells to the glass coverslip, while varying the properties of the polyacrylamide gel permits control the resistance of the substrate to forces applies through cell–substrate adhesions. We prepared substrates of varying degrees of stiffness, seeded cells onto these substrates, and imaged their response to HGF stimulation (Fig. 3). Cells grown on substrates of the maximum stiffness exhibit epithelial scattering behavior that is indistinguishable from that reported when cells are grown directly on glass coverslips coated with collagen I. Cells respond to HGF stimulation by spreading, then rupturing cell–cell adhesions and increasing the rate of cell migration. As the flexibility of the substrate is increased, a result of reducing the ratio of bis-acrylamide in generation of the polyacrylamide gel linking the collagen to the coverslip, cells have more difficulty spreading and few ruptures of cell–cell adhesions are observed. Despite this, cells grown on the most flexible substrates successfully undergo epithelial scattering. Clearly detachment of individual epithelial cells does not require increased cell migration or cell spreading. We then tested whether alterations in substrate flexibility also affected the accelerated scattering we observe with combined HGF and TGFβ treatment. Cells seeded on the least flexible substrates undergo scattering that is accelerated compared to the cells treated with HGF alone, as expected. As the flexibility of the substrate increased, however, the cellular response is apoptosis and not epithelial scattering (data not shown). Membrane blebbing is observed at the cell surface and cells detach from the substrate a short time later. The finding that the effect of TGFβ signaling is different when substrate stiffness is altered is consistent with a recent report that shows that NMuMG cells respond to TGFβ by undergoing EMT on rigid substrates and apoptosis on pliant substrates [18]. It appears that the ability of TGFβ signaling to affect HGF-induced scattering requires a cellular context that includes microenvironmental cues pertaining to substrate stiffness. Unfortunately, the requirement of a stiff substrate for TGFβ to accelerate HGF-induced scattering precludes us from determining whether a more pliable substrate would impact cell–cell in these conditions.
7
] (]]]]) ]]]–]]]
RhoA G14V cells 0.75
Treatment −dox +dox
0.50
p = 0.008294
0.25
0.00
0
100
200
300
Time (Minutes)
Fig. 5 – Quantitative analysis of epithelial scattering in cell expressing RhoA mutants. Plot of the cumulative occurrence of complete detachment of individual cells from larger colonies during HGF-induced epithelial scattering. Lines represent mean values and the ribbons encapsulate the range for 95% confidence in mean determination. The n value, representing the number of colonies analyzed from 3 independent experiments for each condition, was 47 and 84 for cells expressing RhoA T17N in the presence and absence of doxycycline, respectively, and 36 and 46 for cells expressing RhoA G14V in the presence and absence of doxycycline, respectively. Counts were normalized to the number of cells analyzed.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
8
E XP E RI ME N TAL CE L L R ES E ARC H
Blebbistatin
Blebbistatin +HGF
] (]]]]) ]]]–]]]
Rho kinase inh.
Rho kinase inh. +HGF
0:00
0:00
0:00
0:00
2:00
2:00
2:00
2:00
4:00
4:00
4:00
4:00
6:00
6:00
6:00
6:00
Fig. 6 – Blebbistatin and Rho kinase inhibitors induce cell spreading. Timelapse images of MDCK cells responding to blebbistatin or Rho kinase inhibitor treatment, with or without HGF stimulation of epithelial scattering. Time stamps are in hours:minutes. Scale bar¼ 50 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
RhoAT19N +dox
9
] (]]]]) ]]]–]]]
RhoA G14V -dox
+dox
-dox
0:00
0:00
0:00
0:00
0:30
0:30
0:30
0:30
1:00
1:00
1:00
1:00
2:00
2:00
2:00
2:00
3:00
3:00
3:00
3:00
Fig. 7 – Blebbistatin washout induces epithelial scattering independently of RhoA activity. Cellular behavior during recovery from blebbistatin treatment in MDCK cells expressing constitutively active (G14V) or dominant negative (T19N) RhoA mutants under control of a doxycycline (dox) repressible promoter. Time stamps are in hours:minutes. Scale bar¼ 50 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E XP E RI ME N TAL CE L L R ES E ARC H
Inhibition of cellular contractility prevents scattering and induces cell spreading Increasing substrate flexibility likely reduces scattering by reducing the ability of traction forces to be applied through cell– substrate adhesions. While RhoA plays a number of roles in cell motility and cell adhesion, the effect of RhoA mutants on epithelial scattering may also be derived from its role in regulating cellular contractility [22]. RhoA activates Rho kinase, which phosphorylates and activates myosin light chain kinase, increasing cellular contractility by myosin II. Our analysis of HGF-induced epithelial scattering in cells expressing RhoA mutants is consistent with RhoA contributing to scattering by controlling contractility. Dominant negative RhoA would prevent myosin II activation and contractility, thus blocking scattering after spreading. Constitutively active RhoA would not prevent scattering, but the reduced spreading may make it difficult for contractile forces to rupture adhesion, resulting in the observed delay in cell–cell detachment. To analyze the contribution of contractility to epithelial scattering, we examined the effects of blebbistatin and rho kinase inhibitors on MDCK cells (Fig. 6). We then examined how blebbistatin and Rho kinase inhibitor treatment affected epithelial scattering in response to HGF stimulation. Time lapse imaging reveals that cells fail to undergo cell–cell detachment once cell spreading has been completed, a behavior change that is similar to that observed in cells expressing the dominant negative RhoA mutant. We sought to understand the effect of these agents on epithelial scattering by performing quantitative analysis on these experiments. We observed that colonies of MDCK cells treated with HGF increased in area by 91.7% and 113.5% when in the presence of blebbistatin or Rho kinase inhibitor, respectively. This amount of spreading is similar to the reported spreading observed in untreated MDCK cell [7]. In the absence of HGF stimulation, only moderate spreading is observed; cell spreading of colonies of MDCK cells is 19.3% and 14.9% of the original colony when treated with these respective agents in the absence of HGF.
Restoration of cellular contractility in spread cells results in RhoA-independent scattering We reasoned that epithelial scattering might be driven by a cycle of cell spreading and cellular contractility. To test this idea, we examined cellular behavior in MDCK cells recovering from blebbistatin treatment. In order to explore whether any effects are dependent on RhoA activity, we performed this experiment with
MDCK cells expressing RhoA mutants under tetracycline-repressible promoters. MDCK cells were cultured in collagen-coated imaging chambers in the presence or absence of doxycycline, then treated with blebbistatin for 2 h to induce maximal spreading. Upon washout of the blebbistatin from these cultures, colonies of MDCK cells were analyzed using time lapse imaging (Fig. 7). Colonies of MDCK cells grown in the presence of doxycycline and recovering from blebbistatin undergo epithelial scattering that is highly reminiscent of HGF-treated cells. As cells immediately reduce the extent of cell spreading on the substrate, ruptures of cell–cell adhesions are observed. Cells expressing constitutively active RhoA (G14V) respond to blebbistatin recovery by undergoing epithelial scattering in a manner that is indistinguishable from control cells. However, cells expressing dominant negative RhoA (T19N) undergo reduced scattering in this experiment. Quantitative analysis (Fig. 8) shows a marked reduction in both the number of cumulative cell–cell junction rupture events and complete cell detachment events following blebbistatin removal when cells expressing T19N RhoA mutant are compared to non-expressing controls. These data indicate that activation of RhoA is required for epithelial scattering induced by blebbistatin treatment and recovery.
2.5
Cell−cell rupture events/ cell
normalized this to the total number of cells analyzed. The results are almost indistinguishable for both RhoA mutant cell lines in control conditions, that is in the presence of doxycycline. In contrast, dominant negative RhoA (T19N) expression results in an almost complete prevention of detachment events during the time analyzed. Constitutively active RhoA (G14V) expression also blocks scattering relative to controls, though it appears to merely delay detachment of individual cells. Taken together, these results indicate that dynamic changes in RhoA activity are required for HGF-induced epithelial scattering. This is consistent with several previous studies that show a role for RhoA in epithelial mesenchymal transition [21,22].
] (]]]]) ]]]–]]]
Treatment −dox
2.0
+dox
1.5 1.0
p < 2.2x10-16
0.5 0.0 0
100
200
300
Time (Minutes)
Cell−cell rupture events/ cell
10
Treatment
2.0
−dox +dox
1.5 1.0 0.5
p < 6.2x10-13 0.0
0
100
200
300
Time (Minutes)
Fig. 8 – Quantitative analysis of epithelial scattering induced by blebbistatin withdrawal. Plots of the cumulative occurrences of cell–cell junction rupture events or complete cell detachment during epithelial scattering induced by removal of blebbistatin from the culture. Lines represent mean values and the ribbons encapsulate the range for 95% confidence in mean determination. Control cells (black lines, RhoA T19Nþdox) and cells expressing dominant negative RhoA (gray lines, RhoA T19N – dox) are plotted here. The data is taken from analysis of 24 colonies in 3 independent experiments for each condition and normalized to the total number of cells in those colonies.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
Interestingly, we observe that scattering occurs in the absence of increased cell migration. We sought to ensure that the effects of blebbistatin and the rho kinase inhibitor did not have effects on cell migration and cell–cell adhesion that might explain rupturing (Fig. 9). While increased cell migration has been proposed to drive epithelial scattering, including rupturing of cell–cell junctions, we observe that blebbistatin and rho kinase inhibitor both decrease the rate of cell migration. It is therefore possible that restoration of normal migration rates following blebbistatin washout contributed to scattering, though it is hard to see how a normal rate of migration could rupture cell–cell contacts that withstand this normal migration rate in control conditions. We therefore also sought to determine whether a blebbistatin or rho kinase inhibitor would reduce cell–cell adhesion by analyzing aggregation of cells cultured in suspension, as before. After 4 h in blebbistatin or rho kinase inhibitor, cells are observed in large aggregates of cells with adhesions that have been sufficiently strengthened to withstand trituration forces at levels that actually slightly exceed those observed in untreated MDCK cells. In short, we can detect no statistically significant effect on the strength of cell–cell adhesions from blebbistatin or rho kinase inhibitor treatment that might explain how blebbistatin washout leads to epithelial scattering. Though we did not formally and directly show that blebbistatin
11
] (]]]]) ]]]–]]]
washout does not induce changes in cell–cell adhesion, the fact that cells' transition between states with the same strength of adhesion suggests this is unlikely. Together these data suggest that epithelial scattering induced by blebbistatin washout largely results from effects on cellular contractility. We therefore examined actin organization and myosin IIa localization of cells treated with blebbistatin for increasing amounts of time (Fig. 10A). Actin staining at cell–cell junctions rapidly diminishes following blebbistatin treatment and more medial actin cables are observed. When the same staining is performed on cells in a time course following recovery from 2.5 h of blebbistatin treatment (Fig. 10B), the actin cytoskeleton is reorganized into radial contractile networks. Interestingly, these networks are reminiscent of those observed during epithelial scattering of MDCK cells treated with HGF [7]. Does blebbistatin treatment and washout induce epithelial scattering by mimicking molecular events of epithelial scattering induced by physiologically relevant stimuli, like HGF stimulation? We assessed the levels of phosphor-myosin in cell extracts before and after induction of epithelial scattering of MDCK cells by western blot analysis (Fig. 10C). A significant drop in phospho-myosin levels occurs within 15 min of HGF stimulation. Two hours after HGF
Uncovered space (R.U.)
1.2
0.9
0.6
0.3
0
With Trituration 100%
80%
80%
Aggregation (% of cells)
Aggregation (% of cells)
Without Trituration 100%
60% 40%
60% 40% 20%
20% 0%
0% CONTROL
Rho kinase inhibitor
Blebbistatin
Large 50+
0.787
0.960
0.855
0.001
Medium11-50
0.119
0.031
0.118
0.002
Small 1-10
0.093
0.010
0.027
CONTROL
Rho kinase inhibitor
Blebbistatin
Large 50+
0.954
0.998
0.997
Medium11-50
0.043
0.001
Small 1-10
0.004
0.001
Fig. 9 – Effect of blebbistatin and rho kinase inhibitor on cell migration and cell–cell adhesion. (A) Cell migration was determined by evaluating the amount of open area remaining oncovered following coverage by a migrating monolayer after 8 and 16 h (dark and light gray, respectively), relative to the start of the experiment (black). Cell were maintained under controlled conditions or treated of blebbistatin or rho kinase inhibitors. The experiment was performed in the presence and absence of HGF. (B) Distribution of an MDCK cell population into small, medium, and large aggregates after 4 h in suspension in inverted droplets of medium and following treatment with the indicated inhibitor. Cells were subjected to analysis with and without trituration, which disrupts weak adhesions between contacting cells. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
12
E XP E RI ME N TAL CE L L R ES E ARC H
] (]]]]) ]]]–]]]
0:00
0:30
1:00
2:00
3:00
4:00
0:00
0:30
1:00
2:00
3:00 control
blebb.
Actin
P-MLC
Time after HGF stimulation (minutes) 15
30
60
120
Actin
P-MLC
0.2
240
Normalized band intensity (A.U.)
0
0.15 0.1 0.05 0 0
15
30
60
120
240
Time after HGF stimulation (minutes)
Fig. 10 – Actin and myosin II localization in cells undergoing epithelial scattering following blebbistatin recovery (A and B). Phalloidin (green), DAPI (blue), and myosin IIa antibody (red) staining of MDCK cells in a time course following blebbistatin treatment (A) or recovery following 2.5 h of blebbistatin treatment (B). Time stamps are in hours:minutes. Scale bar¼ 30 μm. (C) Western blot detection of phospho-myosin light chain (P-MLC) and actin in cell extracts taken from MDCK cells before and different times after HGF stimulation with quantitative densitometry analysis normalized to actin and (D) western blot detection of actin and phosphor-myosin (P-MLC) in control and blebbistatin-treated (15 min) MDCK cells. stimulation, phospho-myosin levels increase dramatically. Blebbistatin treatment, not surprisingly, reduces phospho-myosin levels compared to untreated cells (Fig. 10D). HGF stimulation clearly induces an
immediate drop and later recovery in myosin contractility, an effect that is likely being recapitulated pharmacologically with blebbistatin treatment and washout.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
Discussion Epithelial cell–cell adhesions integrate individual epithelial cells into epithelial tissues, allowing coordination of cellular activities by individual cells. Epithelial scattering occurs when individual epithelial cells rupture cell–cell junctions to invade and migrate as lone, solitary cells. Cell–cell detachment is accompanied by reported changes in cell–cell adhesion, cell–substrate adhesion, cell motility, and cellular contractility. What has remained unclear is whether cell–cell junctions must be reorganized at the molecular level to allow cell–cell detachment, or whether changes in contractility, motility, and cell–substrate adhesion are sufficient for detachment. Given the coordination of these cellular processes, it has been exceptionally difficult to tease out answers to this question. This report addresses this question directly by assessing the role of cell–substrate adhesion and contractility in epithelial scattering. Interestingly, we did not observe significant changes in cell–cell adhesion properties in assays measuring the ability of cells to form aggregates on suspension after treatment with growth factors that induce epithelial scattering. Though this does appear to be in contrast to findings where HGF-stimulated cells demonstrate loss of E-cadherin from cell–cell junctions [27], other studies argue that the HGF-induced reduction in E-cadherin occurs after cells undergoing scattering [5]. Further, other cell types demonstrate maintenance of cadherins at adhesion sites after HGF stimulation [28]. Further, it is unclear whether cells in suspension would retain their mesenchymal, scattered behavior or revert to a more epithelial phenotype. Certainly we have found that acceleration of HGFinduced epithelial scattering by TGFβ is dependent on adhesion to collagen I and the effects of suspension growth might be expected to eliminate any acceleration effects from TGFβ. Though this experiment does indicate that changes in cell–cell adhesion might not be fundamentally important to epithelial scattering, the aforementioned caveats is a caution against overinterpretation of the results obtained from this experiment alone. We did find that induction of epithelial scattering by HGF dramatically increases the number of cell–substrate adhesions. Acceleration of scattering by additional treatment with TGFβ [17] induces a further increase in the number of focal adhesions. Further implicating cell–substrate adhesions in epithelial scattering were our findings that increasing cell–substrate flexibility and expression of a dominant negative RhoA mutant blocked cell–cell detachment following HGF stimulation. In fact, the effect of RhoA and substrate flexibility indicates that cellular contractility, focused through cell–substrate adhesions, might play the dominant role in epithelial scattering, just as originally proposed by de Rooij et al. [5]. In support of this, we find that rho kinase inhibitors and blebbistatin prevent rupture of cell–cell adhesion following HGF stimulation. A model for epithelial scattering is that cellular contractility relaxes to allow cellular spreading, accompanied by formation of new cell–substrate adhesions at the cell periphery, and that this is followed by a restoration of contractility that then ruptures cell– cell adhesions. We found that reducing than restoring cellular contractility by treatment and then washout of blebbistatin results in epithelial scattering in the absence of growth factor stimulation, a result we found surprising. Interestingly, this scattering occurs independent of any concomitant increase in cell
] (]]]]) ]]]–]]]
13
migration, demonstrating that increased cell migration is not required for scattering. In fact, we find that blebbistatin and rho kinase inhibitor actually reduce the rate of cell migration. Of additional interest, cells undergoing scattering do so without first undergoing significant cell spreading; neither blebbistatin nor rho kinase inhibitor induced cell spreading to the level observed with growth factor treatment-induced epithelial scattering. This idea is generally consistent with studies showing that RhoA is involved in focal adhesion and stress fiber formation in response to HGF stimulation, but not cell spreading [26] and further supports our overall conclusion that cellular contractility can drive scattering in the absence of cell spreading. A final surprise here was that neither blebbistatin nor rho kinase inhibitor generated any reduction in cell–cell adhesion in our suspension aggregation assays. Myosin-based contractility clearly affects cell–cell junction architecture at the molecular level, as specific actin binding proteins are recruited to these cellular sites in response to contractility [23–25]. While we had considered that inhibition of cellular contractility might reduce cell–cell adhesion and thus make cells more prone to scattering upon reintroduction of contractility following blebbistatin washout, the lack of reduced adhesion here shows that this is not the case. Epithelial scattering appears to be generated by the sudden restoration of cellular contractility alone. Consistent with this, phospho-myosin levels in MDCK cells plummet immediately after HGF stimulation and then recover significantly after 2 h, precisely the time at which we begin to see rupturing of cell–cell contacts [7]. Epithelial scattering induced by blebbistatin washout requires RhoA activity, as expression of a dominant negative mutant blocks rupturing and complete detachment of cell–cell adhesions. Of further interest is that epithelial scattering induced by blebbistatin treatment and washout is accompanied by changes in actin organization that are also observed during HGF-induced scattering. This surprising result is consistent with changes in cellular contractility being sufficient to drive rupture of cell–cell adhesions, including driving majority of actin rearrangements associated with this process. Epithelial scattering is a complex cellular process in which changes in cell–cell adhesion, cell–substrate adhesion, and cellular contractility have been reported. Understanding the contribution of each system to epithelial scattering has been a challenge. While our suspicions to date have been that cell–cell adhesion changes are critical to epithelial scattering, our efforts to isolate the specific contributions of cell–substrate adhesion and contractility reported here suggest that epithelial scattering may result from coordination of cellular contractility alone.
Acknowledgments Thanks to members of the Hansen Lab group for their discussion of the manuscript.
references [1] A.I. Ivanov, N.G. Naydenov, Dynamics and regulation of epithelial adherens junctions: recent discoveries and controversies, Int. Rev. Cell. Mol. Biol. 303 (2013) 27–99. [2] J. Lim, J.P. Thiery, Epithelial–mesenchymal transitions: insights from development, Development 139 (2012) 3471–3486.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
14
E XP E RI ME N TAL CE L L R ES E ARC H
[3] S. Dietrich, F. Abou-Rebyeh, H. Brohmann, F. Bladt, E. SonnenbergRiethmacher, T. Yamaai, A. Lumsden, B. Brand-Saberi, C. Birchmeier, The role of SF/HGF and c-Met in the development of skeletal muscle, Development 126 (1999) 1621–1629. [4] Y. Nakaya, G. Sheng, EMT in developmental morphogenesis, Cancer Lett 341 (2013) 9–15. [5] J. de Rooij, A. Kerstens, G. Danuser, M.A. Schwartz, C.M. Waterman-Storer, Integrin-dependent actomyosin contraction regulates epithelial cell scattering, J. Cell Biol. 171 (2005) 153– 164. [6] S. Grotegut, D. von Schweinitz, G. Christofori, F. Lehembre, Hepatocyte growth factor induces cell scattering through MAPK/ Egr-1-mediated upregulation of Snail, EMBO J. 25 (2006) 3534– 3545. [7] R.B. Sperry, N.H. Bishop, J.J. Bramwell, M.N. Brodeur, M.J. Carter, B.T. Fowler, Z.B. Lewis, S.D. Maxfield, D.M. Staley, R.M. Vellinga, M.D.H. Hansen, Zyxin controls migration in epithelial–mesenchymal transition by mediating actin-membrane linkages at cell– cell junctions, J. Cell Physiol. 222 (2010) 612–624. [8] M. Maeda, K.R. Johnson, M.J. Wheelock, Cadherin switching: essential for behavioral but not morphological changes during an epithelium-to-mesenchyme transition, J. Cell Sci. 118 (2005) 873–887. [9] M. Yoshigi, L.M. Hoffman, C.C. Jensen, H.J. Yost, M.C. Beckerle, Mechanical force mobilizes zyxin from focal adhesions to actin filaments and regulates cytoskeletal reinforcement, J. Cell Biol. 171 (2005) 209–215. [10] P. Clark, Modulation of scatter factor/hepatocyte growth factor activity by cell-substratum adhesion, J. Cell Sci. 107 (1994) 1265– 1275. [11] E.E. Sander, S. van Delft, J.P. ten Klooster, T. Reid, R.A. van der Kammen, F. Michiels, J.G. Collard, Matrix-dependent Tiam1/Rac signaling in epithelial cells promotes either cell–cell adhesion or cell migration and is regulated by phosphatidylinositol 3-kinase, J. Cell Biol. 143 (1998) 1385–1398. [12] Y.L. Wang, R.J. Pelham Jr., Preparation of a flexible, porous polyacrylamide substrate for mechanical studies of cultured cells, Methods Enzymol. 298 (1998) 489–496. [13] J.B. Kim, S. Islam, Y.J. Kim, R.S. Prudoff, K.M. Sass, M.J. Wheelock, K.R. Johnson, N-cadherin extracellular repeat 4 mediates epithelial to mesenchymal transition and increased motility, J. Cell Biol. 151 (2000) 1193–1206. [14] J. Behrens, W. Birchmeier, S.L. Goodman, B.A. Imhof, Dissociation of Madin–Darby canine kidney epithelial cells by the monoclonal antibody anti-arc-1: mechanistic aspects and identification of the antigen as a component related to uvomorulin, J. Cell Biol. 101 (1985) 1307–1315. [15] M. Stoker, M. Perryman, An epithelial scatter factor released by embryo fibroblasts, J. Cell Sci. 77 (1985) 209–223.
] (]]]]) ]]]–]]]
[16] O.F. Santos, S.K. Nigam, HGF-induced tubulogenesis and branching of epithelial cells is modulated by extracellular matrix and TGF-beta, Dev. Biol. 160 (1993) 293–302. [17] J.Y. Chung, J.A. Davis, B.D. Price, D.M. Staley, M.V. Wagner, S.L. Warner, D.J. Bearss, M.D.H. Hansen, Competitive enhancement of HGF-induced epithelial scattering by accessory growth factors, Exp. Cell Res. 317 (2011) 307–318. [18] J.L. Leight, M.A. Wozniak, S. Chen, M.L. Lynch, C.S. Chen, Matrix rigidity regulates a switch between TGF-β1-induced apoptosis and epithelial–mesenchymal transition, Mol. Biol. Cell 23 (2012) 781–791. [19] T.S. Jou, W.J. Nelson, Effects of regulated expression of mutant RhoA and Rac1 small GTPases on the development of epithelial (MDCK) cell polarity, J. Cell Biol. 142 (1998) 85–100. [20] T.S. Jou, E.E. Schneeberger, W.J. Nelson, Structural and functional regulation of tight junctions by RhoA and Rac1 small GTPases, J. Cell Biol. 142 (1998) 101–115. [21] L. Fan, A. Sebe, Z. Péterfi, A. Masszi, A.C. Thirone, O.D. Rotstein, H. Nakano, C.A. McCulloch, K. Szászi, I. Mucsi, A. Kapus, Cell contact-dependent regulation of epithelial–myofibroblast transition via the rho-rho kinase-phospho-myosin pathway, Mol. Biol. Cell 18 (2007) 1083–1097. [22] J.D. Berndt, M.R. Clay, T. Langenberg, M.C. Halloran, Rho-kinase and myosin II affect dynamic neural crest cell behaviors during epithelial to mesenchymal transition in vivo, Dev. Biol. 324 (2008) 236–244. [23] Y. Miyake, N. Inoue, K. Nishimura, N. Kinoshita, H. Hosoya, S. Yonemura, Actomyosin tension is required for correct recruitment of adherens junction components and zonula occludens formation, Exp. Cell Res. 312 (2006) 1637–1650. [24] S. Yonemura, Y. Wada, T. Watanabe, A. Ganafuchi, M. Shibita, alpha-Catenin as a tension transducer that induces adherens junction development, Nat. Cell Biol. 12 (2010) 533–542. [25] Q. Le Duc, Q. Shi, I. Blonk, A. Sonnenberg, N. Wang, D. Leckband, J de Rooij, Vinculin potentiates E-cadherin mechanosensing and is recruited to actin-anchored sites within adherens junctions in a myosin II-dependent manner, J. Cell Biol. 189 (2010) 1107–1115. [26] I. Royal, N. Lamarche-Vane, L. Lamorte, K. Kaibuchi, M. Park, Activation of cdc42, rac, PAK, and rho-kinase in response to hepatocyte growth factor differentially regulates epithelial cell colony spreading and dissociation, Mol. Biol. Cell 11 (2000) 1709–1725. [27] H. Miura, K. Nishimura, A. Tsujimura, K. Matsumiya, K. Matsumoto, T. Nakamura, A. Okuyama, Effects of hepatocyte growth factor on E-cadherin-mediated cell–cell adhesion in DU145 prostate cancer cells, Urology 58 (2001) 1064–1069. [28] M. Watanabe, K. Matsumoto, T. Nakamura, M. Takeichi, Effect of hepatocyte growth factor on cadherin-mediated cell–cell adhesion, Cell Struct. Funct. 18 (1993) 117–124.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
] (]]]]) ]]]–]]]
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/yexcr
Research Article
Cellular contractility changes are sufficient to drive epithelial scattering Jacob P. Hoj, John A. Davis, Kendra E. Fullmer, David J. Morrell, Nicholas E. Saguibo, Jeffrey T. Schuler, Kevin J. Tuttle, Marc D.H. Hansenn Physiology and Developmental Biology Brigham Young University, 574 WIDB Provo, UT 84602, USA
article information
abstract
Article Chronology:
Epithelial scattering occurs when cells disassemble cell–cell junctions, allowing individual epithelial
Received 24 May 2013
cells to act in a solitary manner. Epithelial scattering occurs frequently in development, where it
Received in revised form
accompanies epithelial–mesenchymal transitions and is required for individual cells to migrate and
10 April 2014
invade. While migration and invasion have received extensive research focus, how cell–cell junctions
Accepted 14 April 2014
are detached remains poorly understood. An open debate has been whether disruption of cell–cell interactions occurs by remodeling of cell–cell adhesions, increased traction forces through cell
Keywords:
substrate adhesions, or some combination of both processes. Here we seek to examine how changes
Epithelial scattering
in adhesion and contractility are coupled to drive detachment of individual epithelial cells during
Epithelial mesenchymal transition
hepatocyte growth factor (HGF)/scatter factor-induced EMT. We find that HGF signaling does not
Adhesion
alter the strength of cell–cell adhesion between cells in suspension, suggesting that changes in cell–
Cytoskeleton
cell adhesion strength might not accompany epithelial scattering. Instead, cell–substrate adhesion
Contractility
seems to play a bigger role, as cell–substrate adhesions are stronger in cells treated with HGF and
Myosin
since rapid scattering in cells treated with HGF and TGFβ is associated with a dramatic increase in focal adhesions. Increases in the pliability of the substratum, reducing cells ability to generate traction on the substrate, alter cells' ability to scatter. Further consistent with changes in substrate adhesion being required for cell–cell detachment during EMT, scattering is impaired in cells expressing both active and inactive RhoA mutants, though in different ways. In addition to its roles in driving assembly of both stress fibers and focal adhesions, RhoA also generates myosin-based contractility in cells. We therefore sought to examine how RhoA-dependent contractility contributes to cell–cell detachment. Inhibition of Rho kinase or myosin II induces the same effect on cells, namely an inhibition of cell scattering following HGF treatment. Interestingly, restoration of myosin-based contractility in blebbistatin-treated cells results in cell scattering, including global actin rearrangements. Scattering is reminiscent of HGF-induced epithelial scattering without a concomitant increase in cell migration or decrease in adhesion strength. This scattering is dependent on RhoA, as blebbistatin-induced scattering is reduced in cells expressing dominant-negative RhoA mutants. This suggests that induction of myosin-based cellular contractility may be sufficient for cell–cell detachment during epithelial scattering. & 2014 Elsevier Inc. All rights reserved.
n
Corresponding author. Fax: þ801 422 0700. E-mail address: [email protected] (M.D.H. Hansen).
http://dx.doi.org/10.1016/j.yexcr.2014.04.011 0014-4827/& 2014 Elsevier Inc. All rights reserved.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
2
E XP E RI ME N TAL CE L L R ES E ARC H
Introduction Strong cell–cell junctions integrate individual epithelial cells into tissues, permitting coordination of cellular processes. During morphogenetic programs, cell–cell interactions are remodeled in order to allow individual cells to either reposition with respect to the tissue [1]. In some cases, complete ablation of cell–cell adhesions is enacted, which allows cells to fully detach form the epithelium and move independently [2]. Full detachment of individual epithelial cells occurs during epithelial–mesenchymal transitions, where epithelial cells scatter and migrate to distant sites to establish new organs and tissues [3]. It is thought that inappropriate activation of this morphogenetic program drives cancer invasion and metastasis [4]. Scattering of epithelial cells, resulting from rupture of cell–cell adhesions, is accompanied by changes in cell adhesion, motility, and cellular contractility. It has been proposed that cell–cell junctions become actively disassembled, which is required for their subsequent rupture during epithelial scattering. A contrasting view is that scattering cells increase migration and contractility to physically tear cell–cell junctions apart and that this does not require significant molecular changes in cell–cell junction composition or architecture [5]. The requirement for molecular changes in cell–cell adhesions is supported by the observation that cadherin expression changes accompany EMT [6]. Further, dramatic changes in actin organization at cell–cell junctions accompany scattering [7]. A number of actin regulatory proteins are redistributed with respect to cell–cell contacts during scattering, including those proposed to link actin to an adhesion system, like zyxin, vinculin, and α-actinin [7]. Importantly, positively or negatively targeting the function of zyxin prevents or facilitates, respectively, the rupturing of cell–cell contacts [7]. Support for a model where tension forces associated with the traction of cell migration drives scattering comes from observations that detachment of cell–cell adhesions appears to precede many changes in gene transcription and protein modification that are typically associated with EMT events [8]. Contractile forces are associated with epithelial scattering; transcellular networks assembled from actin that was formerly associated with cell–cell contacts become highly decorated with myosin II and are associated with deformations of the plasma membrane [7]. Importantly, the actin regulatory proteins that become redistributed with respect to cell–cell junctions include those (zyxin and VASP) that are known to relocalize to sites of actin-membrane connection in response to tension forces [9]. Further, changes in extracellular matrix type dramatically alter epithelial scattering [10,11], demonstrating that integrin-based adhesions play a fundamental role in epithelial scattering. A model that emphasizes contraction forces as a means of driving scattering suggests that actin remodeling with respect to cell–cell junctions is carried out to focus contractile forces and not to alter the strength of cell–cell interactions. Prior work from our group has focused on actin rearrangements associated with epithelial scattering, including the mechanism and function of zyxin in anchoring actin at site of cell–cell adhesion. Our group has worked under the assumption that scattering requires priming by altering the molecular composition and architecture of cell–cell adhesion systems. Here we perform a series of experiments that are designed to address the contribution of changes in cell–cell adhesion, cell–substrate adhesion, and
] (]]]]) ]]]–]]]
cellular contractility in driving epithelial scattering. We find that cell substrate adhesion changes are more dramatic than cell–cell adhesion changes during early scattering. Further, cellular contractility plays a central role in driving scattering. Surprisingly, forced alterations in cellular contractility in groups of epithelial cells is sufficient to drive detachment and scattering in the absence of a signal for EMT. Based on these results, we argue that detachment of cell–cell adhesions during epithelial scattering can be driven by changes in cellular contractility alone.
Materials and methods Cell culture MDCK cells were cultured in low glucose DMEM supplemented with 10% fetal bovine serum. Cells bearing tetracycline-repressible cassettes were also maintained in 20 ng/ml doxycycline to maintain suppression of cassette genes, the RhoA G14V and T19N mutants. Stimulation of cells with HGF and/or TGFβ (Antigenix America) was performed at a final dose of 25 ng/ml and 0.1 ng/ml, respectively. Blebbistatin and Rho kinase inhibitor VII (Calbiochem) were used at a final concentration of 20 μM and 1.0 μM, respectively. To generate flexible substrates, the protocol reported by Wang and Pelham [12] was adapted. Briefly, glass coverslips were etched with a dried film of sodium hydroxide, silanized with 3-aminopropyltrimethoxysilane, and activated for covalent linkage with 0.5% (v/v) glutaraldehyde. A thin polyacrylamide gel, made with 8% acrylamide and within a range of 0.03–0.26% BIS, was polymerized onto this activated surface. The polyacrylamide gel was activated for covalent linkage of extracellular matrix by addition of 1 mM sulfo-SANPAH and exposure to UV light for 5 min, then removal of remaining sulfoSANPAH solution and overnight incubation with collagen I. Surfaces were washed with PBS and medium prior to seeding with cells.
Cell-based assays Cellular aggregation assays were performed as reported [13]. Briefly, cells were trypsinized and placed in suspension at 250,000 cells per ml. 20 μl drops were placed in the inner lid of a petri dish and then incubated for 0–4 h. Drops were either spread onto a glass slide for analysis directly following culturing or after trituration of the cell suspension, which disrupts weak adhesions with cellular aggregates. The proportion of the total number of cells in small (o10 cells), medium (11–50 cells), and large (450 cells) aggregates was quantified for each condition. For each condition the experiment was performed in triplicate. Cell migration was determined by seeding cells into the wells of a 96 well Oris I (Platypus) cell migration plate and culturing the cells overnight. The stopper was removed and the progression of cell migration into the now open region of each well was monitored by microscopy at the indicated timepoint, generally 0, 6 and 12 h. The size of the open area covered by cells was then determined using SlideBook software.
Live cell imaging and analysis For live cell imaging experiments, collagen-coated Delta-T imaging chambers (Bioptechs) were seeded with 10,000 MDCK cells and cultured overnight, allowing formation of small colonies.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
Prior to imaging, media was changed to HEPES buffered DMEM supplemented with 10% fetal bovine serum. Cells were treated with growth factors and/or inhibitors as appropriate for the experiment. A heated microscope stage using Delta T4 temperature regulation system (Bioptechs) maintained temperature at 37 1C. Multiple positions were imaged for each experiment and each experiment was repeated, with identical parameters, in triplicate. Phase contrast images of cell scattering were taken at 16 magnification using a 10 (0.30 aperture) objective and 1.6 slider, with 30 ms exposure time. Images were recorded every 2 min for 12 h. The first 10 h of each series was used for analysis, with the remaining time demonstrating the continued health of the culture. Quantitative image analysis was performed using SlideBook software. Cell–cell rupture events were quantified as the cumulative number of spaces appearing within colonies of MDCK cells during the first 6 h of imaging. For each colony, the cumulative number of rupture events was normalized to number of cells in the colony to reflect more or less
opportunities for rupture events as a function of colony size. Complete cell detachments were quantified as the cumulative occurrence of individual cells completely breaking contact form the other cells within a colony during the first 6 h of imaging. For each colony analyzed, the cumulative count was normalized to the number of cells in the colony to reflect increased opportunities for such detachments in larger colonies. For statistical analysis, the data was fit to a generalized linear mixed model using R. Because of the skewed distribution and high number of zeros in the count data, the Poisson distribution was used. A quadratic term was added where appropriate. P-values were then generated via a likelihood-ratio test.
Immunofluorescence For immunofluorescence imaging of fixed samples, 10,000 MDCK cells were seeded onto collagen-coated coverslips in 12 well plates and cultured overnight. Cells were treated with growth
With trituration 100%
80%
80%
2 hour culture
2 hour culture
Without trituration 100%
60% 40% 20% 0%
3
] (]]]]) ]]]–]]]
60% 40% 20% 0%
HGF
TGFB
Combo
Control
HGF
TGFB
Large 50+
13.17%
14.18%
20.01%
30.88%
Large 50+
0.00%
0.00%
4.43%
0.00%
Medium11-50
52.10%
24.59%
51.29%
37.50%
Medium11-50
47.84%
49.65%
45.46%
62.48%
Small 1-10
34.73%
61.24%
28.70%
31.62%
Small 1-10
52.16%
50.35%
50.11%
37.52%
Combo
100%
100%
80%
80%
3 hour culture
3 hour culture
Control
60% 40% 20% 0%
Combo
60% 40% 20% 0%
Control
HGF
TGFB
Combo
Control
HGF
TGFB
Large 50+
90.35%
84.43%
96.98%
96.23%
Large 50+
22.38%
7.32%
5.76%
5.39%
Medium11-50
7.72%
6.42%
2.76%
1.98%
Medium11-50
67.24%
75.14%
89.44%
87.78%
Small 1-10
1.93%
9.15%
0.92%
1.79%
Small 1-10 100%
10.38%
17.55%
4.81%
6.84%
*
80%
4 hour culture
4 hour culture
100%
60% 40%
*
20% 0%
80% 60% 40% 20% 0%
Control
HGF
TGFB
Combo
Large 50+
95.26%
50.51%
77.73%
82.01%
Large 50+
Control
HGF
TGFB
Combo
32.53%
37.25%
30.96%
25.80%
Medium11-50
2.88%
36.01%
14.46%
13.93%
Medium11-50
54.23%
47.31%
53.05%
53.04%
Small 1-10
1.86%
13.48%
7.81%
4.06%
Small 1-10
13.23%
15.44%
15.99%
21.16%
Fig. 1 – Cell–cell adhesion changes during epithelial scattering. Distribution of an MDCK cell population into small, medium, and large aggregates after the indicated amount of time in suspension in inverted droplets of medium and following the indicated growth factor treatment. Cells were subjected to analysis with and without trituration, which disrupts weak adhesions between contacting cells. Asterisks indicate instances where a statistically significant difference (Po 0.05) in the proportion of cells occurring in aggregates is observed between a treatment condition and untreated control cells. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
4
E XP E RI ME N TAL CE L L R ES E ARC H
factors and/or inhibitors as appropriate for the individual experiment. Cells were fixed by brief washing in ice cold Ringer's saline and incubated with 3.7% paraformaldehyde on ice. After fixation, cells were rinsed, blocked with PBS buffer containing 0.4% BSA and normal goat serum, and stained using antibodies, DAPI, and/or Alexa fluor-conjugated phalloidin. Antibody staining was performed against myosin IIa (Novus Biologicals) and vinculin (Cytoskeleton), as well as Alexa fluor-conjugated secondary antibodies.
Western blotting Cells were washed with ice cold PBS then harvested and lysed with Buffer A (50 mM Tris, 2% (w/v) SDS, 5% (v/v) glycerol, 1% 2-mercaptoethanol, 5 mM sodium orthovanadate, and 5 mM protease inhibitor cocktail). Equalized concentrations of protein lysates – determined via Bradford assay – were run in triplicate in 8–12% SDS–polyacrylamide gels and transferred to nitrocellulose membranes. The membranes were incubated with anti-phospho-MLC (Bioss) or anti-actin (Sigma-Aldrich) primary antibodies, followed by HRP-conjugated secondary antibodies. Antibody binding was detected using the enhanced chemiluminescence and autoradiography film. Densitometry was performed using the Fluorchem SP imaging system.
Results Changes in cell adhesion during HGF-induced epithelial scattering In order to better understand how cell–cell and cell–substrate adhesion might individually contribute to epithelial scattering
] (]]]]) ]]]–]]]
during EMT, we measured adhesion changes after stimulation of MDCK cells with HGF. MDCK cells are a well-established model system for studying epithelial cell biology [14], and for epithelial scattering specifically [15]. These cells retain highly epithelial characteristics, recapitulating the original tissue organization and function under the correct culture conditions. The response of MDCK cells to HGF has been documented at a high level of resolution [7] and the behavior of scattering cells under various conditions well studied [16]. Also, since the cells originate from normal tissue, normal cellular signaling and mechanics are expected to be retained in this cell line. MDCK cells were cultured in the presence or absence of additional growth factors, then assessed for cell–cell and cell– substrate adhesion using cell based assays. For these experiments, cells were stimulated with growth factors for 24 h in culture prior to analysis in each assay system. Growth factors used were HGF and TGFβ. HGF was used to induce characteristic epithelial scattering and EMT in the cultures of MDCK cells. The addition of TGFβ increases epithelial scattering during HGF induced EMT [17], so the addition of this second factor was used to examine how cells change adhesion properties under conditions where scattering is accelerated beyond what is observed with treatment of HGF alone. We measured cell–cell adhesion by placing cells in suspension in an inverted drop of medium, which results in cells settling into close proximity and the forming aggregates through cell–cell junction formation. The ratio of cells in aggregates of different sizes is tracked over time revealing the cells' propensity to form cell–cell junctions (Fig. 1). When aggregates are triturated prior to analysis, weaker cell–cell junctions are disrupted and only aggregates with strong cell–cell junctions remain allowing a determination of the strength of newly formed cell–cell adhesions. We analyzed aggregate sizes of MDCK cells taken from cultures grown
untreated
TGFβ
HGF
HGF +TGFβ
Fig. 2 – Changes to focal adhesions during epithelial scattering. Phalloidin (red), DAPI (blue), and vinculin antibody (green) staining of MDCK cells treated with indicated growth factors for 2 h. Scale bar ¼30 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
in the presence and absence of both HGF and TGFβ. Cell aggregates were analyzed after 2–4 h in suspension for all conditions. We find that untreated control MDCK cells are found in larger and larger aggregates with increasing amounts of time. By four hours, almost all cells are found in the largest aggregates. Aggregates become increasingly resistant to trituration forces as time progresses, as few large aggregates get disrupted into smaller sizes after four hours in suspension. Interestingly, the pattern of aggregate formation, whether the analysis was done with or without trituration, was similar for cells grown under all culture conditions. The only statistically significant difference is at a single timepoint under single condition for a single treatment, namely a reduction in aggregate size at 4 h in cells treated with HGF alone. We next sought to determine how growth factor treatments alter cell–substrate adhesion of MDCK cells. Cells were stimulated
0.24% bis
5
] (]]]]) ]]]–]]]
to undergo EMT with growth factor treatment and the number and distribution of focal adhesions examined. Cells on coverslips and treated with or without growth factors were fixed and stained with vinculin antibodies, revealing that focal adhesion vastly increased in cells treated with HGF and TGFβ (Fig. 2), suggesting that epithelial scattering is driven largely through cell– substrate adhesions and not changes in cell–cell adhesion.
Changes to cell–substrate adhesion alter scattering behavior during EMT Given the correlation of cell–substrate adhesions with HGFinduced epithelial scattering, we next sought to examine the contribution of cell–substrate adhesion to epithelial scattering. Cell–substrate adhesion plays a role in cellular contractility, cell spreading, and cell migration, all of which are increased during
0.12% bis
0.06% bis
0:00
0:00
0:00
2:00
2:00
2:00
4:00
4:00
4:00
6:00
6:00
6:00
Fig. 3 – Epithelial scattering on pliant substrates. Timelapse images of MDCK cells undergoing HGF-induced epithelial scattering on polyacrylamide surfaces generated with increased flexibility, a result of decreasing amounts of crosslinking with bis acrylamide. Time stamps are in hours:minutes. Scale bar¼ 50 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
6
E XP E RI ME N TAL CE L L R ES E ARC H
] (]]]]) ]]]–]]]
RhoAT19N +dox
RhoA G14V -dox
+dox
-dox
0:00
0:00
0:00
0:00
1:00
1:00
1:00
1:00
2:00
2:00
2:00
2:00
3:00
3:00
3:00
3:00
4:00
4:00
4:00
4:00
Fig. 4 – Epithelial scattering in cells expressing RhoA mutants. Timelapse images of MDCK cells expressing constitutively active (G14V) or dominant negative (T19N) RhoA mutants and undergoing epithelial scattering in response to HGF stimulation. Mutant expression is negatively regulated by the presence of doxycycline (dox). Time stamps are in hours:minutes. Scale bar¼ 50 μm.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
RhoA is required for cell–cell detachment in response to HGF stimulation Given that changes in cell–substrate flexibility properties affect epithelial scattering, we next sought to perturb cell–substrate adhesion systems in another manner and examine the effects on HGF-induced epithelial scattering. To accomplish this we used MDCK cells expressing RhoA mutants. We used MDCK cells that express RhoA G14V and T19N mutants under control of a tetracycline repressor. These cells have been extensively characterized, particularly in terms of how these mutants alter actin organization and cell–cell adhesion [19,20]. Cells were plated onto collagen-coated imaging dishes and cultured in the presence or absence of doxycycline, blocking or permitting mutant expression.
Cells were stimulated with HGF and subjected to time-lapse imaging and quantitative analysis. Imaging (Fig. 4) reveals that in the presence of doxycycline, cells bearing either RhoA mutant undergo epithelial scattering that is typical of normal MDCK cells. In contrast, cells expressing a dominant negative RhoA T19N mutant fail to undergo epithelial scattering. Following HGF stimulation, these cells undergo spreading and increase their migration within the colony, but fail to rupture cell–cell junctions. In experiments imaging HGFinduced scattering in MDCK cells expressing this RhoA mutant, rupturing of cell–cell adhesions was only seen in one or two instances. Meanwhile, cells expressing a constitutively active RhoA G14V mutant undergo scattering that is empirically similar to that observed in control cells; mutant-expressing cells respond to HGF stimulation by undergoing cell spreading that is followed by increased migration and the rupturing of cell–cell adhesions. However, spreading is reduced and ruptures of cell–cell adhesion appear with a significant delay. Quantitative image analysis of scattering by MDCK cells expressing RhoA mutants largely confirm these empirical results (Fig. 5). In blind analysis, we tracked the cumulative number of cells detaching from colonies of MDCK cells stimulated with HGF, then
Detachment events/ cell
1.2
RhoA T19N cells Treatment −dox +dox
0.8
-16
p < 2.2x10
0.4
0.0
0
100
200
300
Time (Minutes)
Detachment events/ cell
epithelial scattering. To assess the roles of migration and spreading, we prepared flexible substrates on which cells could be plated and induced to undergo scattering. On more pliant substrates, cell migration and cell spreading is limited significantly, providing an opportunity to examine how changes in these cellular behaviors contributes to scattering. Glass coverslips were covalently linked with a thin polyacrylamide gel that was in turn covalently decorated with collagen I. Cellular attachment to the extracellular matrix link cells to the glass coverslip, while varying the properties of the polyacrylamide gel permits control the resistance of the substrate to forces applies through cell–substrate adhesions. We prepared substrates of varying degrees of stiffness, seeded cells onto these substrates, and imaged their response to HGF stimulation (Fig. 3). Cells grown on substrates of the maximum stiffness exhibit epithelial scattering behavior that is indistinguishable from that reported when cells are grown directly on glass coverslips coated with collagen I. Cells respond to HGF stimulation by spreading, then rupturing cell–cell adhesions and increasing the rate of cell migration. As the flexibility of the substrate is increased, a result of reducing the ratio of bis-acrylamide in generation of the polyacrylamide gel linking the collagen to the coverslip, cells have more difficulty spreading and few ruptures of cell–cell adhesions are observed. Despite this, cells grown on the most flexible substrates successfully undergo epithelial scattering. Clearly detachment of individual epithelial cells does not require increased cell migration or cell spreading. We then tested whether alterations in substrate flexibility also affected the accelerated scattering we observe with combined HGF and TGFβ treatment. Cells seeded on the least flexible substrates undergo scattering that is accelerated compared to the cells treated with HGF alone, as expected. As the flexibility of the substrate increased, however, the cellular response is apoptosis and not epithelial scattering (data not shown). Membrane blebbing is observed at the cell surface and cells detach from the substrate a short time later. The finding that the effect of TGFβ signaling is different when substrate stiffness is altered is consistent with a recent report that shows that NMuMG cells respond to TGFβ by undergoing EMT on rigid substrates and apoptosis on pliant substrates [18]. It appears that the ability of TGFβ signaling to affect HGF-induced scattering requires a cellular context that includes microenvironmental cues pertaining to substrate stiffness. Unfortunately, the requirement of a stiff substrate for TGFβ to accelerate HGF-induced scattering precludes us from determining whether a more pliable substrate would impact cell–cell in these conditions.
7
] (]]]]) ]]]–]]]
RhoA G14V cells 0.75
Treatment −dox +dox
0.50
p = 0.008294
0.25
0.00
0
100
200
300
Time (Minutes)
Fig. 5 – Quantitative analysis of epithelial scattering in cell expressing RhoA mutants. Plot of the cumulative occurrence of complete detachment of individual cells from larger colonies during HGF-induced epithelial scattering. Lines represent mean values and the ribbons encapsulate the range for 95% confidence in mean determination. The n value, representing the number of colonies analyzed from 3 independent experiments for each condition, was 47 and 84 for cells expressing RhoA T17N in the presence and absence of doxycycline, respectively, and 36 and 46 for cells expressing RhoA G14V in the presence and absence of doxycycline, respectively. Counts were normalized to the number of cells analyzed.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
8
E XP E RI ME N TAL CE L L R ES E ARC H
Blebbistatin
Blebbistatin +HGF
] (]]]]) ]]]–]]]
Rho kinase inh.
Rho kinase inh. +HGF
0:00
0:00
0:00
0:00
2:00
2:00
2:00
2:00
4:00
4:00
4:00
4:00
6:00
6:00
6:00
6:00
Fig. 6 – Blebbistatin and Rho kinase inhibitors induce cell spreading. Timelapse images of MDCK cells responding to blebbistatin or Rho kinase inhibitor treatment, with or without HGF stimulation of epithelial scattering. Time stamps are in hours:minutes. Scale bar¼ 50 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
RhoAT19N +dox
9
] (]]]]) ]]]–]]]
RhoA G14V -dox
+dox
-dox
0:00
0:00
0:00
0:00
0:30
0:30
0:30
0:30
1:00
1:00
1:00
1:00
2:00
2:00
2:00
2:00
3:00
3:00
3:00
3:00
Fig. 7 – Blebbistatin washout induces epithelial scattering independently of RhoA activity. Cellular behavior during recovery from blebbistatin treatment in MDCK cells expressing constitutively active (G14V) or dominant negative (T19N) RhoA mutants under control of a doxycycline (dox) repressible promoter. Time stamps are in hours:minutes. Scale bar¼ 50 μm. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E XP E RI ME N TAL CE L L R ES E ARC H
Inhibition of cellular contractility prevents scattering and induces cell spreading Increasing substrate flexibility likely reduces scattering by reducing the ability of traction forces to be applied through cell– substrate adhesions. While RhoA plays a number of roles in cell motility and cell adhesion, the effect of RhoA mutants on epithelial scattering may also be derived from its role in regulating cellular contractility [22]. RhoA activates Rho kinase, which phosphorylates and activates myosin light chain kinase, increasing cellular contractility by myosin II. Our analysis of HGF-induced epithelial scattering in cells expressing RhoA mutants is consistent with RhoA contributing to scattering by controlling contractility. Dominant negative RhoA would prevent myosin II activation and contractility, thus blocking scattering after spreading. Constitutively active RhoA would not prevent scattering, but the reduced spreading may make it difficult for contractile forces to rupture adhesion, resulting in the observed delay in cell–cell detachment. To analyze the contribution of contractility to epithelial scattering, we examined the effects of blebbistatin and rho kinase inhibitors on MDCK cells (Fig. 6). We then examined how blebbistatin and Rho kinase inhibitor treatment affected epithelial scattering in response to HGF stimulation. Time lapse imaging reveals that cells fail to undergo cell–cell detachment once cell spreading has been completed, a behavior change that is similar to that observed in cells expressing the dominant negative RhoA mutant. We sought to understand the effect of these agents on epithelial scattering by performing quantitative analysis on these experiments. We observed that colonies of MDCK cells treated with HGF increased in area by 91.7% and 113.5% when in the presence of blebbistatin or Rho kinase inhibitor, respectively. This amount of spreading is similar to the reported spreading observed in untreated MDCK cell [7]. In the absence of HGF stimulation, only moderate spreading is observed; cell spreading of colonies of MDCK cells is 19.3% and 14.9% of the original colony when treated with these respective agents in the absence of HGF.
Restoration of cellular contractility in spread cells results in RhoA-independent scattering We reasoned that epithelial scattering might be driven by a cycle of cell spreading and cellular contractility. To test this idea, we examined cellular behavior in MDCK cells recovering from blebbistatin treatment. In order to explore whether any effects are dependent on RhoA activity, we performed this experiment with
MDCK cells expressing RhoA mutants under tetracycline-repressible promoters. MDCK cells were cultured in collagen-coated imaging chambers in the presence or absence of doxycycline, then treated with blebbistatin for 2 h to induce maximal spreading. Upon washout of the blebbistatin from these cultures, colonies of MDCK cells were analyzed using time lapse imaging (Fig. 7). Colonies of MDCK cells grown in the presence of doxycycline and recovering from blebbistatin undergo epithelial scattering that is highly reminiscent of HGF-treated cells. As cells immediately reduce the extent of cell spreading on the substrate, ruptures of cell–cell adhesions are observed. Cells expressing constitutively active RhoA (G14V) respond to blebbistatin recovery by undergoing epithelial scattering in a manner that is indistinguishable from control cells. However, cells expressing dominant negative RhoA (T19N) undergo reduced scattering in this experiment. Quantitative analysis (Fig. 8) shows a marked reduction in both the number of cumulative cell–cell junction rupture events and complete cell detachment events following blebbistatin removal when cells expressing T19N RhoA mutant are compared to non-expressing controls. These data indicate that activation of RhoA is required for epithelial scattering induced by blebbistatin treatment and recovery.
2.5
Cell−cell rupture events/ cell
normalized this to the total number of cells analyzed. The results are almost indistinguishable for both RhoA mutant cell lines in control conditions, that is in the presence of doxycycline. In contrast, dominant negative RhoA (T19N) expression results in an almost complete prevention of detachment events during the time analyzed. Constitutively active RhoA (G14V) expression also blocks scattering relative to controls, though it appears to merely delay detachment of individual cells. Taken together, these results indicate that dynamic changes in RhoA activity are required for HGF-induced epithelial scattering. This is consistent with several previous studies that show a role for RhoA in epithelial mesenchymal transition [21,22].
] (]]]]) ]]]–]]]
Treatment −dox
2.0
+dox
1.5 1.0
p < 2.2x10-16
0.5 0.0 0
100
200
300
Time (Minutes)
Cell−cell rupture events/ cell
10
Treatment
2.0
−dox +dox
1.5 1.0 0.5
p < 6.2x10-13 0.0
0
100
200
300
Time (Minutes)
Fig. 8 – Quantitative analysis of epithelial scattering induced by blebbistatin withdrawal. Plots of the cumulative occurrences of cell–cell junction rupture events or complete cell detachment during epithelial scattering induced by removal of blebbistatin from the culture. Lines represent mean values and the ribbons encapsulate the range for 95% confidence in mean determination. Control cells (black lines, RhoA T19Nþdox) and cells expressing dominant negative RhoA (gray lines, RhoA T19N – dox) are plotted here. The data is taken from analysis of 24 colonies in 3 independent experiments for each condition and normalized to the total number of cells in those colonies.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
Interestingly, we observe that scattering occurs in the absence of increased cell migration. We sought to ensure that the effects of blebbistatin and the rho kinase inhibitor did not have effects on cell migration and cell–cell adhesion that might explain rupturing (Fig. 9). While increased cell migration has been proposed to drive epithelial scattering, including rupturing of cell–cell junctions, we observe that blebbistatin and rho kinase inhibitor both decrease the rate of cell migration. It is therefore possible that restoration of normal migration rates following blebbistatin washout contributed to scattering, though it is hard to see how a normal rate of migration could rupture cell–cell contacts that withstand this normal migration rate in control conditions. We therefore also sought to determine whether a blebbistatin or rho kinase inhibitor would reduce cell–cell adhesion by analyzing aggregation of cells cultured in suspension, as before. After 4 h in blebbistatin or rho kinase inhibitor, cells are observed in large aggregates of cells with adhesions that have been sufficiently strengthened to withstand trituration forces at levels that actually slightly exceed those observed in untreated MDCK cells. In short, we can detect no statistically significant effect on the strength of cell–cell adhesions from blebbistatin or rho kinase inhibitor treatment that might explain how blebbistatin washout leads to epithelial scattering. Though we did not formally and directly show that blebbistatin
11
] (]]]]) ]]]–]]]
washout does not induce changes in cell–cell adhesion, the fact that cells' transition between states with the same strength of adhesion suggests this is unlikely. Together these data suggest that epithelial scattering induced by blebbistatin washout largely results from effects on cellular contractility. We therefore examined actin organization and myosin IIa localization of cells treated with blebbistatin for increasing amounts of time (Fig. 10A). Actin staining at cell–cell junctions rapidly diminishes following blebbistatin treatment and more medial actin cables are observed. When the same staining is performed on cells in a time course following recovery from 2.5 h of blebbistatin treatment (Fig. 10B), the actin cytoskeleton is reorganized into radial contractile networks. Interestingly, these networks are reminiscent of those observed during epithelial scattering of MDCK cells treated with HGF [7]. Does blebbistatin treatment and washout induce epithelial scattering by mimicking molecular events of epithelial scattering induced by physiologically relevant stimuli, like HGF stimulation? We assessed the levels of phosphor-myosin in cell extracts before and after induction of epithelial scattering of MDCK cells by western blot analysis (Fig. 10C). A significant drop in phospho-myosin levels occurs within 15 min of HGF stimulation. Two hours after HGF
Uncovered space (R.U.)
1.2
0.9
0.6
0.3
0
With Trituration 100%
80%
80%
Aggregation (% of cells)
Aggregation (% of cells)
Without Trituration 100%
60% 40%
60% 40% 20%
20% 0%
0% CONTROL
Rho kinase inhibitor
Blebbistatin
Large 50+
0.787
0.960
0.855
0.001
Medium11-50
0.119
0.031
0.118
0.002
Small 1-10
0.093
0.010
0.027
CONTROL
Rho kinase inhibitor
Blebbistatin
Large 50+
0.954
0.998
0.997
Medium11-50
0.043
0.001
Small 1-10
0.004
0.001
Fig. 9 – Effect of blebbistatin and rho kinase inhibitor on cell migration and cell–cell adhesion. (A) Cell migration was determined by evaluating the amount of open area remaining oncovered following coverage by a migrating monolayer after 8 and 16 h (dark and light gray, respectively), relative to the start of the experiment (black). Cell were maintained under controlled conditions or treated of blebbistatin or rho kinase inhibitors. The experiment was performed in the presence and absence of HGF. (B) Distribution of an MDCK cell population into small, medium, and large aggregates after 4 h in suspension in inverted droplets of medium and following treatment with the indicated inhibitor. Cells were subjected to analysis with and without trituration, which disrupts weak adhesions between contacting cells. Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
12
E XP E RI ME N TAL CE L L R ES E ARC H
] (]]]]) ]]]–]]]
0:00
0:30
1:00
2:00
3:00
4:00
0:00
0:30
1:00
2:00
3:00 control
blebb.
Actin
P-MLC
Time after HGF stimulation (minutes) 15
30
60
120
Actin
P-MLC
0.2
240
Normalized band intensity (A.U.)
0
0.15 0.1 0.05 0 0
15
30
60
120
240
Time after HGF stimulation (minutes)
Fig. 10 – Actin and myosin II localization in cells undergoing epithelial scattering following blebbistatin recovery (A and B). Phalloidin (green), DAPI (blue), and myosin IIa antibody (red) staining of MDCK cells in a time course following blebbistatin treatment (A) or recovery following 2.5 h of blebbistatin treatment (B). Time stamps are in hours:minutes. Scale bar¼ 30 μm. (C) Western blot detection of phospho-myosin light chain (P-MLC) and actin in cell extracts taken from MDCK cells before and different times after HGF stimulation with quantitative densitometry analysis normalized to actin and (D) western blot detection of actin and phosphor-myosin (P-MLC) in control and blebbistatin-treated (15 min) MDCK cells. stimulation, phospho-myosin levels increase dramatically. Blebbistatin treatment, not surprisingly, reduces phospho-myosin levels compared to untreated cells (Fig. 10D). HGF stimulation clearly induces an
immediate drop and later recovery in myosin contractility, an effect that is likely being recapitulated pharmacologically with blebbistatin treatment and washout.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
E X PE R IM EN TA L C ELL R E S EA RC H
Discussion Epithelial cell–cell adhesions integrate individual epithelial cells into epithelial tissues, allowing coordination of cellular activities by individual cells. Epithelial scattering occurs when individual epithelial cells rupture cell–cell junctions to invade and migrate as lone, solitary cells. Cell–cell detachment is accompanied by reported changes in cell–cell adhesion, cell–substrate adhesion, cell motility, and cellular contractility. What has remained unclear is whether cell–cell junctions must be reorganized at the molecular level to allow cell–cell detachment, or whether changes in contractility, motility, and cell–substrate adhesion are sufficient for detachment. Given the coordination of these cellular processes, it has been exceptionally difficult to tease out answers to this question. This report addresses this question directly by assessing the role of cell–substrate adhesion and contractility in epithelial scattering. Interestingly, we did not observe significant changes in cell–cell adhesion properties in assays measuring the ability of cells to form aggregates on suspension after treatment with growth factors that induce epithelial scattering. Though this does appear to be in contrast to findings where HGF-stimulated cells demonstrate loss of E-cadherin from cell–cell junctions [27], other studies argue that the HGF-induced reduction in E-cadherin occurs after cells undergoing scattering [5]. Further, other cell types demonstrate maintenance of cadherins at adhesion sites after HGF stimulation [28]. Further, it is unclear whether cells in suspension would retain their mesenchymal, scattered behavior or revert to a more epithelial phenotype. Certainly we have found that acceleration of HGFinduced epithelial scattering by TGFβ is dependent on adhesion to collagen I and the effects of suspension growth might be expected to eliminate any acceleration effects from TGFβ. Though this experiment does indicate that changes in cell–cell adhesion might not be fundamentally important to epithelial scattering, the aforementioned caveats is a caution against overinterpretation of the results obtained from this experiment alone. We did find that induction of epithelial scattering by HGF dramatically increases the number of cell–substrate adhesions. Acceleration of scattering by additional treatment with TGFβ [17] induces a further increase in the number of focal adhesions. Further implicating cell–substrate adhesions in epithelial scattering were our findings that increasing cell–substrate flexibility and expression of a dominant negative RhoA mutant blocked cell–cell detachment following HGF stimulation. In fact, the effect of RhoA and substrate flexibility indicates that cellular contractility, focused through cell–substrate adhesions, might play the dominant role in epithelial scattering, just as originally proposed by de Rooij et al. [5]. In support of this, we find that rho kinase inhibitors and blebbistatin prevent rupture of cell–cell adhesion following HGF stimulation. A model for epithelial scattering is that cellular contractility relaxes to allow cellular spreading, accompanied by formation of new cell–substrate adhesions at the cell periphery, and that this is followed by a restoration of contractility that then ruptures cell– cell adhesions. We found that reducing than restoring cellular contractility by treatment and then washout of blebbistatin results in epithelial scattering in the absence of growth factor stimulation, a result we found surprising. Interestingly, this scattering occurs independent of any concomitant increase in cell
] (]]]]) ]]]–]]]
13
migration, demonstrating that increased cell migration is not required for scattering. In fact, we find that blebbistatin and rho kinase inhibitor actually reduce the rate of cell migration. Of additional interest, cells undergoing scattering do so without first undergoing significant cell spreading; neither blebbistatin nor rho kinase inhibitor induced cell spreading to the level observed with growth factor treatment-induced epithelial scattering. This idea is generally consistent with studies showing that RhoA is involved in focal adhesion and stress fiber formation in response to HGF stimulation, but not cell spreading [26] and further supports our overall conclusion that cellular contractility can drive scattering in the absence of cell spreading. A final surprise here was that neither blebbistatin nor rho kinase inhibitor generated any reduction in cell–cell adhesion in our suspension aggregation assays. Myosin-based contractility clearly affects cell–cell junction architecture at the molecular level, as specific actin binding proteins are recruited to these cellular sites in response to contractility [23–25]. While we had considered that inhibition of cellular contractility might reduce cell–cell adhesion and thus make cells more prone to scattering upon reintroduction of contractility following blebbistatin washout, the lack of reduced adhesion here shows that this is not the case. Epithelial scattering appears to be generated by the sudden restoration of cellular contractility alone. Consistent with this, phospho-myosin levels in MDCK cells plummet immediately after HGF stimulation and then recover significantly after 2 h, precisely the time at which we begin to see rupturing of cell–cell contacts [7]. Epithelial scattering induced by blebbistatin washout requires RhoA activity, as expression of a dominant negative mutant blocks rupturing and complete detachment of cell–cell adhesions. Of further interest is that epithelial scattering induced by blebbistatin treatment and washout is accompanied by changes in actin organization that are also observed during HGF-induced scattering. This surprising result is consistent with changes in cellular contractility being sufficient to drive rupture of cell–cell adhesions, including driving majority of actin rearrangements associated with this process. Epithelial scattering is a complex cellular process in which changes in cell–cell adhesion, cell–substrate adhesion, and cellular contractility have been reported. Understanding the contribution of each system to epithelial scattering has been a challenge. While our suspicions to date have been that cell–cell adhesion changes are critical to epithelial scattering, our efforts to isolate the specific contributions of cell–substrate adhesion and contractility reported here suggest that epithelial scattering may result from coordination of cellular contractility alone.
Acknowledgments Thanks to members of the Hansen Lab group for their discussion of the manuscript.
references [1] A.I. Ivanov, N.G. Naydenov, Dynamics and regulation of epithelial adherens junctions: recent discoveries and controversies, Int. Rev. Cell. Mol. Biol. 303 (2013) 27–99. [2] J. Lim, J.P. Thiery, Epithelial–mesenchymal transitions: insights from development, Development 139 (2012) 3471–3486.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
14
E XP E RI ME N TAL CE L L R ES E ARC H
[3] S. Dietrich, F. Abou-Rebyeh, H. Brohmann, F. Bladt, E. SonnenbergRiethmacher, T. Yamaai, A. Lumsden, B. Brand-Saberi, C. Birchmeier, The role of SF/HGF and c-Met in the development of skeletal muscle, Development 126 (1999) 1621–1629. [4] Y. Nakaya, G. Sheng, EMT in developmental morphogenesis, Cancer Lett 341 (2013) 9–15. [5] J. de Rooij, A. Kerstens, G. Danuser, M.A. Schwartz, C.M. Waterman-Storer, Integrin-dependent actomyosin contraction regulates epithelial cell scattering, J. Cell Biol. 171 (2005) 153– 164. [6] S. Grotegut, D. von Schweinitz, G. Christofori, F. Lehembre, Hepatocyte growth factor induces cell scattering through MAPK/ Egr-1-mediated upregulation of Snail, EMBO J. 25 (2006) 3534– 3545. [7] R.B. Sperry, N.H. Bishop, J.J. Bramwell, M.N. Brodeur, M.J. Carter, B.T. Fowler, Z.B. Lewis, S.D. Maxfield, D.M. Staley, R.M. Vellinga, M.D.H. Hansen, Zyxin controls migration in epithelial–mesenchymal transition by mediating actin-membrane linkages at cell– cell junctions, J. Cell Physiol. 222 (2010) 612–624. [8] M. Maeda, K.R. Johnson, M.J. Wheelock, Cadherin switching: essential for behavioral but not morphological changes during an epithelium-to-mesenchyme transition, J. Cell Sci. 118 (2005) 873–887. [9] M. Yoshigi, L.M. Hoffman, C.C. Jensen, H.J. Yost, M.C. Beckerle, Mechanical force mobilizes zyxin from focal adhesions to actin filaments and regulates cytoskeletal reinforcement, J. Cell Biol. 171 (2005) 209–215. [10] P. Clark, Modulation of scatter factor/hepatocyte growth factor activity by cell-substratum adhesion, J. Cell Sci. 107 (1994) 1265– 1275. [11] E.E. Sander, S. van Delft, J.P. ten Klooster, T. Reid, R.A. van der Kammen, F. Michiels, J.G. Collard, Matrix-dependent Tiam1/Rac signaling in epithelial cells promotes either cell–cell adhesion or cell migration and is regulated by phosphatidylinositol 3-kinase, J. Cell Biol. 143 (1998) 1385–1398. [12] Y.L. Wang, R.J. Pelham Jr., Preparation of a flexible, porous polyacrylamide substrate for mechanical studies of cultured cells, Methods Enzymol. 298 (1998) 489–496. [13] J.B. Kim, S. Islam, Y.J. Kim, R.S. Prudoff, K.M. Sass, M.J. Wheelock, K.R. Johnson, N-cadherin extracellular repeat 4 mediates epithelial to mesenchymal transition and increased motility, J. Cell Biol. 151 (2000) 1193–1206. [14] J. Behrens, W. Birchmeier, S.L. Goodman, B.A. Imhof, Dissociation of Madin–Darby canine kidney epithelial cells by the monoclonal antibody anti-arc-1: mechanistic aspects and identification of the antigen as a component related to uvomorulin, J. Cell Biol. 101 (1985) 1307–1315. [15] M. Stoker, M. Perryman, An epithelial scatter factor released by embryo fibroblasts, J. Cell Sci. 77 (1985) 209–223.
] (]]]]) ]]]–]]]
[16] O.F. Santos, S.K. Nigam, HGF-induced tubulogenesis and branching of epithelial cells is modulated by extracellular matrix and TGF-beta, Dev. Biol. 160 (1993) 293–302. [17] J.Y. Chung, J.A. Davis, B.D. Price, D.M. Staley, M.V. Wagner, S.L. Warner, D.J. Bearss, M.D.H. Hansen, Competitive enhancement of HGF-induced epithelial scattering by accessory growth factors, Exp. Cell Res. 317 (2011) 307–318. [18] J.L. Leight, M.A. Wozniak, S. Chen, M.L. Lynch, C.S. Chen, Matrix rigidity regulates a switch between TGF-β1-induced apoptosis and epithelial–mesenchymal transition, Mol. Biol. Cell 23 (2012) 781–791. [19] T.S. Jou, W.J. Nelson, Effects of regulated expression of mutant RhoA and Rac1 small GTPases on the development of epithelial (MDCK) cell polarity, J. Cell Biol. 142 (1998) 85–100. [20] T.S. Jou, E.E. Schneeberger, W.J. Nelson, Structural and functional regulation of tight junctions by RhoA and Rac1 small GTPases, J. Cell Biol. 142 (1998) 101–115. [21] L. Fan, A. Sebe, Z. Péterfi, A. Masszi, A.C. Thirone, O.D. Rotstein, H. Nakano, C.A. McCulloch, K. Szászi, I. Mucsi, A. Kapus, Cell contact-dependent regulation of epithelial–myofibroblast transition via the rho-rho kinase-phospho-myosin pathway, Mol. Biol. Cell 18 (2007) 1083–1097. [22] J.D. Berndt, M.R. Clay, T. Langenberg, M.C. Halloran, Rho-kinase and myosin II affect dynamic neural crest cell behaviors during epithelial to mesenchymal transition in vivo, Dev. Biol. 324 (2008) 236–244. [23] Y. Miyake, N. Inoue, K. Nishimura, N. Kinoshita, H. Hosoya, S. Yonemura, Actomyosin tension is required for correct recruitment of adherens junction components and zonula occludens formation, Exp. Cell Res. 312 (2006) 1637–1650. [24] S. Yonemura, Y. Wada, T. Watanabe, A. Ganafuchi, M. Shibita, alpha-Catenin as a tension transducer that induces adherens junction development, Nat. Cell Biol. 12 (2010) 533–542. [25] Q. Le Duc, Q. Shi, I. Blonk, A. Sonnenberg, N. Wang, D. Leckband, J de Rooij, Vinculin potentiates E-cadherin mechanosensing and is recruited to actin-anchored sites within adherens junctions in a myosin II-dependent manner, J. Cell Biol. 189 (2010) 1107–1115. [26] I. Royal, N. Lamarche-Vane, L. Lamorte, K. Kaibuchi, M. Park, Activation of cdc42, rac, PAK, and rho-kinase in response to hepatocyte growth factor differentially regulates epithelial cell colony spreading and dissociation, Mol. Biol. Cell 11 (2000) 1709–1725. [27] H. Miura, K. Nishimura, A. Tsujimura, K. Matsumiya, K. Matsumoto, T. Nakamura, A. Okuyama, Effects of hepatocyte growth factor on E-cadherin-mediated cell–cell adhesion in DU145 prostate cancer cells, Urology 58 (2001) 1064–1069. [28] M. Watanabe, K. Matsumoto, T. Nakamura, M. Takeichi, Effect of hepatocyte growth factor on cadherin-mediated cell–cell adhesion, Cell Struct. Funct. 18 (1993) 117–124.
Please cite this article as: J.P. Hoj, et al., Cellular contractility changes are sufficient to drive epithelial scattering, Exp Cell Res (2014), http://dx.doi.org/10.1016/j.yexcr.2014.04.011
