AJP - GI Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Am J Physiol Gastrointest Liver Physiol 287: G886-G898, 2004. First published June 10, 2004; doi:10.1152/ajpgi.00065.2004
0193-1857/04 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
287/4/G886    most recent
00065.2004v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (8)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hopkins, A. M.
Right arrow Articles by Nusrat, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hopkins, A. M.
Right arrow Articles by Nusrat, A.

MUCOSAL BIOLOGY

Epithelial cell spreading induced by hepatocyte growth factor influences paxillin protein synthesis and posttranslational modification

Ann M. Hopkins,1,* Matthias Bruewer,1,2,* G. Thomas Brown,1 A’Drian A. Pineda,1 Julie J. Ha,1 L. Matthew Winfree,1 Shaun V. Walsh,1 Brian A. Babbin,1 and Asma Nusrat1

1Epithelial Pathobiology Research Unit, Department of Pathology and Laboratory Medicine, Emory University, Atlanta, Georgia 30322; and 2Department of General Surgery, University of Muenster, 48149 Muenster, Germany

Submitted 9 February 2004 ; accepted in final form 3 June 2004


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Superficial wounds in the gastrointestinal tract rapidly reseal by coordinated epithelial cell migration facilitated by cytokines such as hepatocyte growth factor (HGF)/scatter factor released in the wound vicinity. However, the mechanisms by which HGF promotes physiological and pathophysiologic epithelial migration are incompletely understood. Using in vitro models of polarized T84 and Caco-2 intestinal epithelia, we report that HGF promoted epithelial spreading and RhoA GTPase activation in a time-dependent manner. Inducible expression of enhanced green fluorescent protein-tagged dominant-negative RhoA significantly attenuated HGF-induced spreading. HGF expanded a zone of partially flattened cells behind the wound edge containing basal F-actin fibers aligned in the direction of spreading. Concomitantly, plaques positive for the focal adhesion protein paxillin were enhanced. HGF induced an increase in the translation of paxillin and, to a lesser extent, {beta}1-integrin. This was independent of cell-matrix adhesion through {beta}1-integrin. Subcellular fractionation revealed increased cosedimentation of paxillin with plasma membrane-containing fractions following HGF stimulation, without corresponding enhancements in paxillin coassociation with {beta}1 integrin or actin. Tyrosine phosphorylation of paxillin was reduced by HGF and was sensitive to the Src kinase inhibitor PP2. With these taken together, we propose that HGF upregulates a free cytosolic pool of paxillin that is unaffiliated with either the cytoskeleton or focal cell-matrix contacts. Thus early spreading responses to HGF may partly relate to increased paxillin availability for incorporation into, and turnover within, dynamic cytoskeletal/membrane complexes whose rapid and transient adhesion to the matrix drives migration.

epithelium; cell migration; actin; focal complexes

THE INTESTINE IS LINED by a monolayer of epithelial cells that serves as the first barrier between the intestinal lumen and the internal milieu. In disease states such as idiopathic inflammatory bowel disease and ischemic colitis, this barrier is disrupted by mucosal wounds, which then serve as conduits for lumenal bacteria and other antigens to access deeper tissues. Small defects in the epithelium efficiently recover by cohesive migration of epithelial sheets (8, 35, 58). This form of migration contrasts fundamentally with that of hematopoetic cells or fibroblasts, where cells migrate individually instead of as a sheet. Cohesively migrating epithelial cells must maintain cell-cell contacts to prevent new gaps in the epithelial barrier from forming and to ensure rapid coverage of denuded surfaces. Epithelial migration has been extensively studied using both in vivo and in vitro models (5, 9, 23, 27, 40, 41, 78, 79).

Migrating epithelial cells undergo a remarkable transition in cell shape, changing from erect columnar cells to flattened, elongated cells. Dynamic modifications in the actin cytoskeleton mediate such changes in epithelial cell shape and cell movement (8, 41). Epithelial cells at the wound edge extrude thin F-actin-rich projections termed lamellipodia and filopodia, which are regulated by the small GTPases Rac and Cdc42, respectively (24, 34, 39, 50, 63). These projections transiently affiliate with the extracellular matrix (ECM) at focal adhesion complex sites (3). Such interactions are dynamic and play a key role in linking the intracellular environment to the ECM, providing a bidirectional sensing mechanism during cell migration (13, 25, 26, 62, 68).

Epithelial injury and active inflammation are associated with release of cytokines into the milieu, which, in turn, influences cell migration. One such cytokine, which is secreted by subepithelial mesenchymal cells and acts in a paracrine fashion on epithelial cells, is termed hepatocyte growth factor (HGF)/scatter factor. HGF is a pluripotent cytokine that has potent motogenic effects on epithelial cells (12, 37, 53, 67, 69). Secreted in an inactive form and activated by proteolytic cleavage, HGF binds specific receptors exhibiting tyrosine kinase activity (15, 54–56, 64). Engagement of the main HGF receptor c-MET (38) regulates cell motility, mitosis, and tubule formation/branching (6, 29, 40) in different human tissues and organ systems (10, 21, 66, 74). Thus HGF is important for gastrointestinal wound healing, neural growth, and development (1, 30, 52), which is best illustrated by the fact that HGF knockout mice die in utero of liver and placental developmental defects (61, 71). However, HGF has also been implicated in the pathophysiological growth and metastasis of several human tumors (7, 16, 31, 33, 36). Therefore, a better understanding of the mechanisms by which HGF influences coordinated epithelial migration is central to manipulating both the physiological and pathophysiological consequences of its activity.

Our study demonstrates that HGF enhances coordinated epithelial cell spreading at 6 h and expands a partially flattened cell population behind the leading edge. Activation of the small GTPase RhoA is important at this time, and expression of dominant-negative RhoA significantly attenuates the HGF-induced cell-spreading response. HGF also promotes aggregation of the focal adhesion protein paxillin and enhances the protein synthesis of paxillin and, to a lesser extent, {beta}1-integrin. This is independent of adhesion to the cell matrix through {beta}1-integrin. Although paxillin protein levels increase both in whole cell lysates and plasma membrane-containing fractions in response to HGF, paxillin coassociation with either {beta}1-integrin or actin is not enhanced by HGF. Moreover, HGF diminishes Src kinase-mediated tyrosine phosphorylation of paxillin, which may limit paxillin accumulation or function at focal cell-matrix contact sites. Our results suggest that HGF induces a "free" cytosolic pool of paxillin whose availability may facilitate rapid and dynamic turnover of transient focal cell-matrix contacts during the early spreading response to HGF.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Cell culture. T84 colonic epithelial cells (American Type Culture Collection, Manassas, VA) were grown in 1:1 DMEM and Ham’s F-12 medium supplemented with 15 mM HEPES (pH 7.5), 14 mM NaHCO3, 2 mM L-glutamine, 40 µg/ml penicillin, 8 µg/ml ampicillin, 90 µg/ml streptomycin, and 6% newborn calf serum as previously described (28). Caco-2 colonic epithelial cells were grown in high glucose DMEM supplemented with 15 mM HEPES (pH 7.4), 2 mM L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, 1% nonessential amino acids, and 10% fetal calf serum. Cells were subcultured every 6–8 days with 0.1% trypsin and 1.0 mM EDTA in Ca2+- and Mg2+-free PBS.

Antibodies, cytokines, and chemicals. Anti-paxillin (Zymed Laboratories, San Francisco, CA), actin (Sigma, St. Louis, MO), {beta}1-integrin [MAb13, kind gift of Dr. K. Yamada, National Institutes of Health (NIH)] and Y118-phospho-paxillin (Cell Signaling Technologies, Beverly, MA) were used for immunolocalization, Western blot analysis, and immunoprecipitation. Recombinant human HGF was a kind gift of Genentech (P. Godowski, San Francisco, CA). It has previously been shown (40) that a concentration of 300 ng/ml has an optimal effect in T84 epithelial cells. Protease and phosphatase inhibitor cocktails and alkaline phosphatase assay kits were purchased from Sigma, and Triton X-100 (TX-100) from Roche Applied Sciences (Indianapolis, IN).

Wounding of epithelial monolayers. In confluent T84 intestinal epithelial cells grown on 24-well plates, reproducible linear wounds were made using a sterile stainless-steel wounding device consisting of a metal tube 1.5 cm in length (external diameter: 0.5 mm, beveled tip internal diameter: 0.25 mm) attached to a vacuum flask. Monolayers were then incubated basolaterally with HGF (300 ng/ml) or serum-free medium alone and allowed to migrate for 6 h. An inverted Zeiss microscope (Axiovert 35 M; Zeiss, Oberkochen, Germany) and a video camera (Hamamatsu Photonic Systems, Bridgewater, NJ) were used to photograph wounds at 0 and 6 h, and wound sizes were measured using Scion Image (Scion, Frederick, MD). For biochemical experiments, epithelial cells on 5- or 45-cm2 permeable supports were wounded using a wounding comb with numerous parallel ports 1 mm apart to enrich for spreading/migrating cells.

GTPase activation assays. To test for RhoA activation in T84 cells on HGF treatment, a rhotekin-binding assay was performed as previously described (46). Briefly, confluent T84 monolayers on 5-cm2 collagen-coated inserts were wounded, allowed to migrate for up to 6 h with or without HGF, and harvested into lysis buffer (25 mM HEPES, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 10 mM MgCl2, 1 mM EDTA, 2% glycerol, and protease inhibitor cocktail). After centrifugation to remove cell debris, lysates were rotated for 45 min/4°C with a 40-µl slurry of a Rhotekin-GST fusion protein coupled to agarose beads (Upstate Biotechnology, Lake Placid, NY). GTP{gamma}S was used as a positive control for Rho activation. Beads were collected by centrifugation, washed three times with lysis buffer, and resuspended in reducing sample buffer. After being boiled, equivalent protein concentrations of pull-down (active GTPase) and whole cell lysate (total GTPase) were separated by SDS-PAGE on 15% Tris·HCl gels, transferred to nitrocellulose, and Western blot analyzed with monoclonal anti-RhoA (Santa Cruz Biotechnology, Santa Cruz, CA). The activation status of Rac and Cdc42 was similarly tested using a GST fusion protein of the Rac/Cdc42 effector PAK-1 conjugated to agarose beads (Upstate Biotechnology). Activated Rac/Cdc42 from pull-down experiments and total Rac/Cdc42 from whole cell lysates were detected by SDS-PAGE and Western blot analysis for Rac or Cdc42 (Transduction Laboratories, Lexington, KY).

Generation of stable, inducible cell lines expressing dominant negative RhoA. The GeneSwitch system (Invitrogen, Carlsbad, CA) was used to generate an inducible epithelial cell line expressing dominant negative (DN)-RhoA. Caco-2 cells, which have many structural and functional features in common with T84 cells, were used because they are significantly more responsive to transfection than T84s. The system uses a regulatory plasmid (pSwitch) encoding a fusion protein (GAL4-DBD/hPR-LBD/p65-AD), which is activated by a synthetic steroid, mifepristone. Binding of mifepristone to the transcriptional fusion protein results in an activating conformational change inducing interaction of the fusion protein with the GAL4 binding site on both the regulatory and expression plasmid (pGene/V5-His B). The resulting positive feedback loop regulates expression of the transcriptional element as well as expression of DN-RhoA. The DN-RhoA mutant sequence was amplified from a pU-myc-RhoA-N19 construct (gift of M. Symons) using the following primers: forward: 5'-cggaattcaatggctgccatccggaagaaactgg-3'; reverse: 5'-gcggatccggacaagacaaggcaaccag-3'. Amplified sequences were digested with EcoR1 5' and BamH1 and ligated into an enhanced green fluorescent protein (pEGFP) plasmid (Clontech, BD Biosciences, NJ). The fusion sequence was digested from the FpEGFP plasmid using EcoR1 and NotI and ligated into the GeneSwitch inducible plasmid pGene/V5-His B. Plasmids were sequenced for confirmation. The regulatory plasmid was transfected into subconfluent Caco-2 cell monolayers (30–40% confluent) using Fugene 6 and selected in media containing 100 µg/ml hygromycin. Secondary transfections with the plasmid containing DN-RhoA-EGFP were then performed as above and selected with zeocin (200 ng/ml). For wounding experiments, cells were plated on 24-well plates as described above, and EGFP-DN-RhoA expression was induced in test wells using mifepristone (10 nM/24 h). Mutant induction was verified visually before each experiment by GFP positivity under an inverted fluorescence microscope. Mifepristone was also added during subsequent incubations with HGF to maintain induction. Noninduced controls (containing the GFP-DN-RhoA plasmid but not expressing it) were never exposed to mifepristone. Negative control cells not transfected with EGFP-DN-RhoA were exposed to mifepristone to exclude any effects of the latter on normal cells.

Immunofluorescence. Wounded T84 monolayers migrating for 6 h with or without HGF were washed in Hanks’ balanced salt solution (HBSS). To visualize F-actin, monolayers were fixed in 3.7% paraformaldehyde, permeabilized in 0.5% TX-100, and incubated with Alexa Fluor 568 phalloidin (Molecular Probes, Eugene, OR) for 40 min at room temperature (RT). Paxillin and {beta}1-integrin were highlighted after fixing in 100% ethanol at –20°C for 20 min. Monolayers were then incubated in 5% normal goat serum in HBSS (1 h/RT), primary antibodies (1 h/RT), and appropriate fluoresceinated secondary antibodies (2.5 µg/ml; 1 h/RT, Jackson Immunoresearch Laboratories, West Grove, PA). Cell nuclei were highlighted with To-Pro-3-iodide (1 µg/ml; Molecular Probes). Monolayers were mounted in 1:1:0.01 (vol/vol/vol) PBS/glycerol/p-phenylenediamine and analyzed on an LSM 510 confocal microscope (Carl Zeiss Microimaging, Thornwood, NY).

Cell isolation and inhibitor experiments. To examine the involvement of Src family tyrosine kinases in HGF-mediated phosphorylation of paxillin, wounded T84 monolayers on 5-cm2 permeable supports were exposed to the Src inhibitor PP2 (10 µM) for 6 h in the presence or absence of HGF during migration. PP3 (10 µM) was used as a negative control. Monolayers were washed in HBSS and scraped into TX-100 extraction buffer (in mM: 100 KCl, 3 NaCl, 3.5 MgCl2, and 10 HEPES, pH 7.4) containing 1% TX-100 with protease and phosphatase inhibitor cocktails (Sigma). Cell lysates were subjected to low-speed centrifugation (1,500 g·5 min–1·4°C–1) and equivalent protein concentrations in the postnuclear lysate from control and HGF-treated monolayers subjected to SDS-PAGE and immunoblotted for paxillin, phospho-paxillin, and {beta}1-integrin as described (40). For {beta}1-integrin blocking experiments, wounded T84 monolayers on 5-cm2 permeable supports were either preexposed to 5 µg/ml of the functionally inhibitory {beta}1-integrin antibody MAb13 for 2 h before HGF/control medium application for a subsequent 6 h (Pre); or coexposed to MAb13 and HGF/control medium during a 6-h migration period (Co). As a negative control, identical monolayers were coincubated for 6 h with HGF/control medium and a 1:50 dilution of normal rat serum. Postnuclear lysates were prepared and Western blot analyzed as above.

Subcellular fractionation. T84 monolayers on 45-cm2 inserts were wounded as previously described and allowed to migrate in the presence or absence of HGF for 6 h. Cells were harvested into an extraction buffer (as previously described but lacking TX-100). Cells were then disrupted by nitrogen cavitation (250 lbs.·in.–2·15 min–1·4°C–1) as described previously (22). Unbroken nuclei and cellular debris were removed from the resultant lysate by low-speed centrifugation (1,500 g·5 min–1·4°C–1). The supernatant was subjected to isopycnic sucrose density sedimentation on continuous 20–55% (wt/wt) linear sucrose gradients in a Beckman SW28 swinging bucket rotor (23,500 rpm·3 h–1·4°C–1) as previously described (22). After ultracentrifugation, up to 30 sequential fractions were collected and analyzed for protein content by bicinchoninic acid assay (Pierce, Rockford, IL) and alkaline phosphatase activity by measuring alkaline hydrolysis of p-nitrophenyl phosphate (Sigma). Actin, {beta}1-integrin, and paxillin distribution in the fractions were then assessed by Western blot analysis as detailed in HGF enhances paxillin association with plasma membranes.

Immunoblotting and immunoprecipitation. Whole cell lysate and subcellular fractionation samples were denatured by boiling in Laemmelli sample buffer containing 20 mM dithiothreitol, subjected to SDS-PAGE, and transferred to nitrocellulose. Transferred proteins were probed with antibodies to paxillin, Y118-phospho-paxillin, actin, or {beta}1-integrin and visualized by enhanced chemiluminescence (ECL; Amersham Pharmacia, Piscataway, NJ) using previously described methods (20, 42). Paxillin and {beta}1-integrin were immunoprecipitated from either postnuclear lysates or diluted plasma membrane fractions (from subcellular fractionations) by incubation with the relevant primary antibody (5 µg·3 h–1·4°C–1) followed by protein G-Sepharose (Amersham Pharmacia; 3 h/4°C). Immunoprecipitated proteins were subjected to SDS-PAGE, transferred to nitrocellulose, probed with antibodies as above and visualized by ECL.

Statistics. Results are expressed as means ± SE throughout. Paired or unpaired Student’s t-tests were used to compare results from different experimental procedures. Statistical significance was assumed at P values <0.05. Individual experiments were performed with two to six replicates, and each experiment was performed independently at least three times.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 GRANTS
 REFERENCES
 
Cohesive migration of epithelial sheets is facilitated by HGF in a RhoA-dependent fashion. Coordinated migration of epithelial sheets is crucial for efficient sealing of superficial wounds in the gastrointestinal tract in a process termed "restitution." For cells to migrate effectively, they must first flatten out and spread into a wound. Several bioactive mediators are known to influence this process. To further our studies on the effects of one such mediator, recombinant human HGF, on cell migration, we used the human intestinal epithelial cell lines T84 and Caco-2 as related model systems. First, we show that HGF promotes wound closure in T84 monolayers in a time-dependent manner (Fig. 1A). Migratory responses to HGF in Caco-2 cells were similar to those in T84 cells (data not shown). At the 6-h time point used throughout the rest of the study, HGF-stimulated migration was quantitatively analogous in both cell types.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 1. Hepatocyte growth factor (HGF) enhances epithelial restitution in a Rho-dependent fashion. A: confluent T84 intestinal epithelial monolayers were wounded and allowed to migrate for up to 6 h in the presence or absence of HGF (300 ng/ml) in serum-free medium. Although both control and HGF-treated cells migrated progressively into the wound over time, wound closure was significantly enhanced by HGF at 4 and 6 h (*P < 0.05 and **P < 0.01, respectively, by 2-tailed unpaired Student’s t-test). Data are expressed as means ± SE from 3 separate experiments. B: activation status of Rho family GTPases was tested in T84 monolayers incubated for 0, 2, 4, and 6 h postspreading in the presence or absence of HGF. RhoA activity peaked at 6 h, corresponding with a dip in the activity of Rac1 and Cdc42. GTP{gamma}S (100 µM) was used as a positive control to verify GTPase activation. C: effect of dominant-negative (DN) RhoA expression on basal and HGF-stimulated spreading was examined at 6 h in Caco-2 cells expressing enhanced green fluorescent protein (EGFP)-DN-RhoA under inducible control of the synthetic steroid mifepristone. In noninduced (NI) cells not expressing EGFP-DN-RhoA, HGF significantly increased spreading over control levels (***P <0.001 by 2-tailed unpaired Student’s t-test). When EGFP-DN-RhoA expression was induced (I) with mifepristone, wound closure in control monolayers was slightly reduced, but there was a significant attenuation of HGF-stimulated wound closure (***P < 0.001 by 2-tailed unpaired Student’s t-test). Mifepristone itself was without effect on cell migration in nontransfected cells treated with vehicle or HGF.

 
The most significant effects of HGF on T84 cell spreading were observed 4 and 6 h postwounding, after which time wound closure began to plateau. In parallel with the timing of maximal spreading responses in T84 cells, there was also a peak in activity of the small GTPase RhoA in HGF-treated samples at 6 h (Fig. 1B). At earlier time points, RhoA activity in control samples was similar to, or sometimes exceeded that of, HGF samples. This may reflect rapid and dynamic cycling of GTPases between active and inactive states. However, we consistently observed peak induction of RhoA activity in response to HGF after 6 h migration.

In contrast to RhoA, activity levels of the closely related GTPases Rac1 and Cdc42 in HGF-treated T84 cells peaked at early time points but were lower than those in control samples by 6 h. The nonhydrolyzable analog GTP{gamma}S was used as a positive control for GTPase activation, and levels of total (active + inactive) GTPase were determined by Western blot analysis of whole cell lysates. Total levels of RhoA, Rac1, and Cdc42 were similar between all experimental conditions (data not shown).

Because we observed increased RhoA activity following 6 h incubation with HGF, we further analzyed the role of this GTPase in HGF-stimulated spreading by expressing DN-RhoA in an inducible system. As shown in Fig. 1C, a Caco-2 cell line stably expressing EGFP-tagged DN-RhoA under the control of a mifepristone-inducible promoter was wounded in a scratch assay and allowed to spread into the wounds for 6 h in the presence or absence of HGF. Wound closure in monolayers in which DN-RhoA expression was not induced (NI) was compared with that in monolayers induced (I) to express DN-RhoA. Successful mutant induction was confirmed by a positive GFP signal in induced monolayers (data not shown) and the appearance of a higher molecular weight species representing EGFP-DN-RhoA in addition to native RhoA in whole cell lysates blotted for RhoA (data not shown). In samples not induced to express DN-RhoA, HGF-treated monolayers spread significantly more than controls by 6 h (P < 0.001 by unpaired Student’s t-test). Induction of DN-RhoA expression inhibited basal cell migration by 24% compared with that in NI controls. However, induction of DN-RhoA in HGF-treated samples significantly attenuated epithelial spreading compared with that in NI samples treated with HGF (46% reduction; P < 0.001 by unpaired Student’s t-test). These effects were specifically due to mifepristone-induced expression of the EGFP-DN-RhoA construct, because mifepristone itself had no effect on the motility of nontransfected cells treated with vehicle or HGF (compare gray bars in Fig. 1C to NI control and NI-HGF). Our results indicate that although RhoA is involved in basal spreading of control cells, enhancement of its activation by HGF was important for HGF-mediated enhancements in cell spreading.

HGF enhances reorganization of F-actin and focal adhesion proteins during spreading. Because RhoA activity governs F-actin stress fiber formation (50) and stress fibers, in turn, anchor cells to the matrix and influence migration, we therefore examined the influence of HGF on stress fiber organization in migrating epithelial sheets. As seen in Fig. 2A, F-actin (highlighted by fluorescent phalloidin) was organized in discrete basal structures of migrating T84 cells. We observed broadening of a flattened zone of cells behind the leading edge of T84 migrating epithelial sheets exposed to HGF. Similar results were observed in Caco-2 cells (data not shown). This led us to define three distinct morphological zones in migrating epithelial cells, summarized pictorially in Fig. 2B. Closest to the wound edge (zone I), cells have a flattened-out morphology and display prominent cellular extensions referred to as lamellipodia and filopodia, which are also a characteristic of nonepithelial cell types (25, 34). Adjacent to zone I is a region of partially-flattened cells extending 3–4 cell diameters back from the wound edge (zone II). Behind zone II is zone III, where epithelial cells are tall, columnar, and polarized similar to nonmigrating or stationary cells. Exposure to HGF induced a marked expansion in the proportion of partially flattened (or zone II) epithelial cells oriented in the direction of migration, as shown by F-actin staining and imaging in the xz plane (Fig. 2C). These cells made an important contribution toward covering denuded surfaces, facilitating wound closure. Zones I and III were morphologically similar between control and HGF-exposed cells.



View larger version (55K):
[in this window]
[in a new window]
  		Fig. 2. HGF augments a zone of F-actin organization in spreading epithelial sheets. A: F-actin organization in T84 epithelial monolayers migrating for 6 h in the presence or absence of HGF was highlighted by Alexa Fluor 568 phalloidin staining and en face confocal microscopy. Lamellipodia were observed in flattened cells at the leading edge of both control and HGF-treated monolayers. HGF induced expansion of a zone of partially flattened cells behind the leading edge, containing basal F-actin fibers that effectively covered denuded areas by stretching into the wound. (Scale bar = 50 µm). B: schematic representation of the 3 major morphologic zones observed in epithelial cells migrating as a cohesive sheet into a wounded area: 1) a zone of completely flattened cells at the leading edge (zone I); 2) an intermediate zone of partially-flattened cells (zone II); and 3) normal tall columnar cells furthest removed from the wound (zone III). C: control (CON) and HGF-treated monolayers migrating for 6 h were phalloidin-labeled to highlight F-actin and imaged in the reconstructed x-z plane. HGF-treated cells showed an expanded proportion of partially flattened zone II cells stretching into the wound.

 
Because the membrane of migrating cells must dynamically anchor to the cell matrix during cell motility, we next concentrated on potential links between F-actin stress fibers and the cell membrane. One key link is formed by the actin-binding protein paxillin, which binds in a cytoplasmic complex to F-actin and putatively binds to {beta}1-integrin in the cell membrane (70). Localization of paxillin at focal cell matrix contact sites, termed focal adhesions, is critically important in migration. We therefore analyzed the distribution of paxillin and {beta}1-integrin in a cohesively migrating epithelium and determined the influence of HGF on this distribution. As shown in en face images (Fig. 3A, a–d), paxillin (green) was distributed in discrete focal plaques at the leading edge of lamellipodia and filopodia (arrows) in zone I of both control (a) and HGF-incubated (c) cells. HGF incubation induced aggregation of paxillin-positive plaques in zone II cells (c; arrowheads). Some paxillin was also observed at the base of actively migrating and spreading cells in zones I and II, respectively. Basal paxillin-positive plaques observed in tall columnar cells of zone III were not dramatically different in control vs. HGF-treated cells. To determine the spatial distribution of paxillin relative to the actin cytoskeleton, cells were double labeled for paxillin and F-actin (b, d; red). Paxillin-labeled plaques were identified along the tips of basal F-actin filaments. HGF treatment induced not only an expansion in the size of zone II, but also accentuated basal paxillin-positive plaques in this zone (c, d, arrowheads).



View larger version (48K):
[in this window]
[in a new window]
  		Fig. 3. HGF alters the localization of paxillin and {beta}1-integrin in spreading epithelial sheets. A: wounded T84 epithelial monolayers spreading for 6 h in the presence or absence of HGF were double-labeled for paxillin (green) and F-actin (red) and imaged by confocal microscopy in the en face plane. Paxillin was seen in plaques capping F-actin stress fibers extending into the wound (arrows) of both control (a, b) and HGF-incubated (c, d) cells. HGF induced an increase in the density of paxillin-positive plaques in zone II cells behind the leading edge (arrowheads). B: spreading monolayers treated as above were double-labeled for {beta}1-integrin (green) and F-actin (red). Distinct pools of {beta}1-integrin were observed at the leading edge (arrows) of zone I cells and in the lateral membrane of zone II cells (arrowheads) in control (e, f) and HGF-incubated (g, h) monolayers. (Scale bar = 10 µm).

 
Focal complex proteins such as paxillin signal to the ECM via transmembrane integrins. In particular, {beta}1-integrin plays an important role in mediating migration of T84 epithelial cells (27). In our model of epithelial spreading (Fig. 3B, e–h), {beta}1-integrin was observed to localize in thin, veil-like structures at the leading edge of control migrating cells (e; arrow) and also in their basal membranes (e; arrowhead). There were no major differences in {beta}1-integrin localization at the leading edge (g; arrow) or in basal membranous structures (g; arrowhead) of epithelial cells following HGF treatment.

HGF upregulates paxillin and {beta}1-integrin protein synthesis. Given increases in both number and density of paxillin-based plaques in spreading epithelial cells exposed to HGF, we next examined whether paxillin protein levels were altered in whole cell lysates prepared from migrating T84 monolayers treated with control medium vs. HGF for 6 h. As shown in Fig. 4A, there was a large increase in abundance of the 68-kDa species corresponding to paxillin in HGF-exposed T84 monolayers. {beta}1-Integrin protein levels were also upregulated in HGF-treated T84 cells relative to controls, but less consistently and to a lesser extent than those of paxillin. Total actin protein levels were stable in both control and HGF-treated T84 samples. Similar results were observed in Caco-2 cells (data not shown).



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 4. HGF enhances paxillin (Pax) and {beta}1-integrin protein levels in spreading epithelial sheets. A: whole cell lysates prepared from wounded T84 epithelial monolayers spreading for 6 h in the presence or absence of HGF were separated by SDS-PAGE and Western blot analysis for paxillin, {beta}1-integrin, and actin. Enhanced protein levels of paxillin were consistently observed in T84 samples exposed to HGF. {beta}1-Integrin protein levels were occasionally increased on HGF treatment, but there was no effect of HGF on actin protein levels. B: wounded T84 monolayers spreading for 6 h in the presence of HGF or control medium were coincubated with the translation inhibitor cycloheximide (CHX; 10–20 µM), and whole cell lysates were Western blot analyzed for Pax, {beta}1-integrin, and actin. Pax was sensitive to CHX in both control and HGF-treated cells, indicating the presence of a translation-sensitive pool of protein. {beta}1-Integrin and actin protein levels, respectively, were somewhat sensitive to and completely insensitive to CHX treatment.

 
A number of potential explanations could account for enhancements in paxillin and {beta}1-integrin levels. A positive effect of HGF on cell division within the 6-h time frame was excluded, because the doubling time for T84 cells is in the order of days rather than hours (data not shown). Therefore, we next investigated whether enhanced translation could account for the increase (Fig. 4B). In the presence of the translation inhibitor cycloheximide (10–20 µM), we observed a decrease in paxillin protein levels both in control migrating T84 monolayers and to a greater extent in those exposed to HGF. The existence of a cycloheximide-sensitive pool of paxillin suggests rapid translational upregulation in response to wounding itself and implicates paxillin translation as an event in HGF-induced cell spreading. {beta}1-Integrin levels in HGF-treated monolayers were also somewhat decreased by cycloheximide treatment, suggesting translational regulation to a lesser extent than with paxillin. In contrast, actin protein levels were unaltered by cycloheximide, implying that its redistribution rather than new protein synthesis may be important in HGF-induced cell spreading.

Enhanced paxillin protein synthesis by HGF is independent of enhancements in cell migration. To investigate the relationship between increased paxillin translation and engagement of its potential binding partner {beta}1-integrin with the matrix after HGF exposure, we tested whether observed increases in paxillin could be blocked by MAb13, a functionally inhibitory anti-{beta}1-integrin antibody that inhibits T84 cell migration (27). Thus Western blot analysis was used to compare protein levels in cells migrating in the presence or absence of HGF (Fig. 5). Neither preincubation of wounded T84 monolayers with anti-{beta}1 (MAb13, Pre) before HGF exposure nor coincubation of wounded monolayers with HGF and anti-{beta}1 (MAb13, Co) affected HGF-mediated upregulation of paxillin protein levels. T84 monolayers coincubated with normal rat serum and HGF (rat serum, Co) as a negative control also showed similar upregulation of paxillin. Inhibition of migration with antibodies to {beta}1-integrin did not substantially affect total protein levels of {beta}1-integrin and actin over that in control and HGF-treated monolayers not exposed to blocking antibody.



View larger version (61K):
[in this window]
[in a new window]
  		Fig. 5. HGF enhances the translation of paxillin independent of migration. Cell migration was blocked by either preincubation (Pre) or coincubation (Co) of migrating epithelial monolayers with the functionally inhibitory {beta}1-integrin antibody MAb13 in the presence of HGF or control medium to test whether functional interactions between {beta}1-integrin and the cell matrix are required for HGF-mediated upregulation of paxillin protein levels. Monolayers were coincubated with normal rat serum and HGF or control medium as a negative control for normal spreading. MAb13, whether coincubated with or preincubated before HGF, did not diminish the ability of HGF to enhance paxillin protein levels over those in negative control samples. Similarly, blocking antibodies did not have any effect on small and inconsistent upregulations in {beta}1-integrin protein levels observed in the presence of HGF. Actin protein levels were completely unaltered by any blocking treatment or by exposure to HGF alone.

 
HGF enhances paxillin association with plasma membranes. Having determined that paxillin translation is upregulated in migrating or nonmigrating epithelial monolayers in response to HGF, we questioned whether there was an increase in paxillin affiliation with the plasma membrane. Potential assembly of actin-paxillin-{beta}1 complexes could be an important mechanism accounting for increased migration. We used a subcellular fractionation approach to isolate plasma membranes from spreading cells and to define the affiliation of {beta}1-integrin, paxillin, and actin with this pool. Control or HGF-exposed T84 cells were lysed and subjected to isopycnic sucrose density gradient centrifugation on 20–55% linear gradients. The profiles of a plasma membrane marker (alkaline phosphatase) and the total protein concentrations in successive gradient fractions are plotted in Fig. 6A. The linear sucrose density profile in the two gradients is also shown in Fig. 6A. Exposure to HGF did not influence recovery of plasma membranes, which were observed as a unimodal peak of activity for the brush border enzyme alkaline phosphatase in ~33–41% sucrose. Total protein concentrations were similar between control and HGF-exposed samples, with soluble cytosolic proteins mostly recovered at the top of the gradient after fractionation.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 6. Subcellular fractionation profiles from epithelial monolayers spreading in the presence or absence of HGF. A: epithelial cells spreading for 6 h in the presence or absence of HGF were subjected to subcellular fractionation on isopycnic sucrose density gradients. Sequential fractions and the pellet (P) obtained from control vs. HGF gradients were then analyzed for sucrose density, alkaline phosphatase (Alk. phos.) activity, and protein concentration. No differences in sucrose density profiles were detected between control and HGF gradients. Maximal activity of the membrane-bound enzyme alkaline phosphatase, used as a marker for plasma membranes, was comparable in both control and HGF gradients as a unimodal peak in approximately fractions 12–15 (~33–41% sucrose). Similarly, HGF did not influence the protein concentration profiles from sequential fractions. B: representative immunoblots of paxillin, {beta}1-integrin, and actin in gradient fractions from control and HGF-incubated monolayers are shown here. Two main pools of paxillin were detected in control gradients: a high-density pool near the top of the gradient that coincided with the bulk of cellular protein and one between 33–41% sucrose, which coincided with peak membrane alkaline phosphatase activity. Paxillin levels in both pools were substantially increased in gradients prepared from HGF-incubated monolayers. In both control and HGF gradients, {beta}1-integrin was mainly recovered from higher-density fractions, with a small percentage in plasma membrane-containing fractions. HGF treatment was associated with variable increases in {beta}1 recovery from both compartments, the plasma membrane compartment in particular. The profile of actin recovered from control and HGF gradients was wider than those for paxillin or {beta}1-integrin and stretched continuously from the top of the gradients into the plasma membrane-containing fractions. No substantial differences were observed between actin profiles in control and HGF gradients.

 
Western blot analysis of fractions separated by SDS-PAGE (Fig. 6B) revealed essentially two pools of paxillin in control gradients. Whereas the majority of paxillin was identified in low-density fractions from the top of the gradient containing the largest pool of cellular proteins, a small but defined pool of paxillin cosedimented with plasma membranes (fractions ~15–19). Both pools were substantially increased in response to HGF treatment. A small but distinct fraction of {beta}1-integrin cosedimented with plasma membrane-containing fractions. Small increases were observed following HGF treatment. HGF did not exert any significant or reproducible effect on the plasma membrane-associated profile of actin in subcellular gradients. The majority of actin sedimented in light-density fractions at the top of the gradients, with a small pool entering the plasma membrane peak in samples from both control and HGF-treated cells.

HGF does not increase paxillin and {beta}1-integrin coassociation in plasma membrane fractions. Because HGF treatment enriched paxillin and {beta}1-integrin protein levels in epithelial plasma membrane-containing fractions, we investigated whether this reflected enhanced association with each other. {beta}1-integrin was immunoprecipitated from plasma membrane-containing fractions and whole cell lysates of control and HGF-treated T84 cells, and immunoprecipitates were probed for paxillin, {beta}1-integrin, and actin (Fig. 7). Increased coassociation of paxillin and {beta}1-integrin was observed in response to HGF treatment when {beta}1 was immunoprecipitated directly from whole cell lysates. Only low levels of paxillin cosedimented with {beta}1-integrin when {beta}1-integrin was immunoprecipitated directly from plasma membrane-containing fractions. Thus we did not detect any differences between paxillin and {beta}1-integrin coassociation in this plasma membrane compartment following HGF treatment.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 7. HGF does not increase paxillin/{beta}1-integrin coassociation in plasma membrane fractions of spreading cells. {beta}1-integrin and paxillin were immunoprecipitated (IP) from whole cell lysates or plasma membrane-containing fractions prepared from monolayers spreading in the presence or absence of HGF for 6 h. {beta}1-Integrin immunoprecipitates immunoblotted (IB) with the same antibody revealed increased {beta}1-integrin levels in whole cell lysates and plasma membrane fractions following HGF treatment. Paxillin was detected in {beta}1-integrin immunoprecipitates from whole cell lysates, with a small increase in the presence of HGF. Actin was also contained within the complex immunoprecipitated by {beta}1-integrin from both control and HGF whole cell lysates. When paxillin was immunoprecipitated from whole cell lysates or plasma membrane fractions, no {beta}1-integrin was detected in either immunoprecipitates. Increased levels of paxillin were observed in paxillin immunoprecipitates from HGF-treated monolayers relative to controls. Levels of actin recovered from paxillin immunoprecipitates of control vs. HGF-treated cells were not different under any condition, although as expected, the amount of actin that coprecipitated with paxillin was greatest in whole cell lysate samples.

 
Paxillin was also immunoprecipitated from whole cell lysates or plasma membrane fractions of control vs. HGF-treated T84 cells. Cosedimentation of {beta}1-integrin with paxillin was below detectable limits in all paxillin immunoprecipitates. However, increased levels of paxillin were detected in paxillin immunoprecipitates from either whole cell lysates or plasma membrane fractions in response to HGF. This is consistent with increased paxillin levels observed in Western blot analyses of whole cell lysates prepared from HGF-treated samples (Fig. 4). Increased paxillin-actin coprecipitation was not observed in HGF-exposed whole cell lysates or plasma membrane fractions. Taken together, our data support that HGF-mediated enhancements in paxillin protein levels reflect the enlargement of a cytosolic pool of paxillin, which is unaffiliated with {beta}1-integrin in the membrane.

Src phosphorylation of paxillin is reduced by HGF. Because there is evidence that tyrosine phosphorylation of paxillin influences its localization and function at focal adhesions (2, 47), we reasoned that the tyrosine phosphorylation status of paxillin might delineate between two cellular pools, one currently engaged in adhesion at focal cell matrix contact sites and one in the cytosol. In light of the fact that paxillin was originally identified as a tyrosine phosphorylated substrate of oncogenic Src (14), we incorporated the specific Src inhibitor PP2 into our experimental design to test for the involvement of Src in HGF-stimulated phosphorylation events. As shown in Fig. 8A, high basal levels of Y118-phosphorylated paxillin in control T84 whole cell lysates were reduced by coincubation with PP2 (10 µM) but not its negative control inhibitor PP3 (10 µM). Substantially less Y118-phospho-paxillin was observed in HGF-treated samples, which was also sensitive to PP2 but not PP3. We observed an identical pattern for Y31-paxillin phosphorylation in T84 cells (data not shown). HGF evoked an increase in paxillin protein levels at the same time point, and actin levels were stable across all conditions. This hints further at the possibility of a free cytosolic pool of paxillin following 6-h stimulation of epithelial cell spreading with HGF. In fact, we found >80% of total cellular paxillin in T84 cells to reside in a biochemical compartment soluble in TX-100, implying a lack of affiliation with large cytoskeletal or membrane complexes (data not shown). In time-course experiments up to 6 h postwounding, levels of paxillin in both Triton-soluble and -insoluble compartments of HGF-incubated cells exceeded those in control cells (data not shown), but Triton-insoluble paxillin putatively affiliated with cytoskeletal/membrane complexes comprised only a tiny percentage of total paxillin. We also quantitated by densitometry the effect of HGF on protein levels of paxillin, Y118-phosphorylated paxillin, and actin in whole cell lysates from control and HGF-treated samples. In Fig. 8B, levels of each protein in HGF-treated samples are presented as percentages of those in control samples. For statistical analysis, paired Student’s t-tests were used to compare HGF vs. control samples from each experiment (n = 4–7 experiments). Paxillin levels were increased in HGF-treated samples relative to controls, but there was a significant decrease in Y118-phosphorylated paxillin levels (P < 0.01). The effect of HGF on Y118-paxillin phosphorylation was also visible at earlier time points of migration (data not shown). We did not observe any changes in the threonine-phosphorylation profile of paxillin immunoprecipitates from control or HGF-treated T84 cells (data not shown). There were no significant differences in actin protein levels between HGF and control samples.



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 8. Src family tyrosine kinases are involved in basal and HGF-stimulated spreading. A: to examine whether paxillin was phosphorylated by Src kinases in response to HGF, wounded T84 monolayers were allowed to spread for 6 h in the presence of HGF/control medium and the Src family tyrosine kinase inhibitor PP2 (10 µM) or its negative control inhibitor PP3 (10 µM). High levels of Y118-phosphorylated paxillin were detected in whole cell lysates of control monolayers, which was reduced substantially by incubation with PP2 but not PP3. Y118-phosphorylated paxillin was much less in HGF-treated samples relative to controls. This too was reduced by exposure to PP2 but not to PP3. Total paxillin protein levels in HGF-treated samples were greater than those in controls and unaffected by either inhibitor. Total actin protein was unchanged across all conditions. B: effect of HGF on whole cell lysate levels of paxillin, Y-118-phosophorylated paxillin, and actin was quantitated by densitometry. Results are plotted as the %total protein in HGF relative to control samples (n = 4–7 independent experiments; **P < 0.01 by paired Student’s t-test).

 

   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
GRANTS
 REFERENCES
 
We have analyzed migration of a cohesive sheet of epithelial cells and addressed mechanisms by which HGF exerts its motogenic properties on these cells. An early time point (6 h) was chosen in our study to focus on early cell spreading events induced by HGF, which precede better-characterized responses of cell dissociation/scattering followed by tube formation or morphogenesis after prolonged HGF activity (52, 67). Furthermore, our reductionist model of epithelial cell migration and wound closure took advantage of the slow doubling time of T84 intestinal epithelial cells to exclude any proliferative effects of HGF on wound-closure events. We are confident that remodeling of epithelial cell shape rather than cell division is responsible for the effects we observed over the time courses used. This is based on previous studies (41) in which T84 cells subjected to irradiation (to maximally reduce proliferation) displayed short-term wound closure rates identical to those of nonirradiated control cells.

Coordinated migration of T84 and Caco-2 human colonic epithelial cells into a scratch wound was similarly increased in a time-dependent fashion in our model in response to HGF. Because cell motility depends on F-actin cytoskeletal rearrangements induced by Rho family GTPases (17, 24, 39, 49, 57), we analyzed the activation status of RhoA, Rac1, and Cdc42 in T84 cells over the same time periods and found RhoA activation to be at its maximum at the time of maximal HGF-induced cell spreading (6 h). Furthermore, inducible expression of dominant-negative RhoA in Caco-2 cells significantly attenuated HGF-stimulated epithelial cell spreading but had a comparatively smaller effect on basal epithelial spreading in the absence of HGF. Thus HGF must exert a profound effect on RhoA activation during its enhancement of spreading, because induction of DN-RhoA in the ~70% of cells in our line stably expressing the mutant transgene at high levels was sufficient to attenuate HGF-induced cell spreading back to that of control levels. This concurs with observations in Madin-Darby canine kidney cells (MDCK) epithelial cells that early responses to HGF involve RhoA activation but that later responses (8–16 h posttreatment) feature reduced RhoA activity and the disassembly of characteristic stress fibers that might otherwise promote matrix anchorage and impede motility (48). Indeed, it has recently been reported (73) that site-specific inactivation of RhoA at cellular protrusions can be achieved by targeting the ubiquitin ligase Smurf1 to Cdc42/Rac1/PAR6 effector complexes containing PKC-{omega}. Thus early activation of RhoA appears to be important for the initial spreading response to HGF, but temporal inactivation likely sustains the subsequent scattering response. In support of this, it has been noted that a transient loss of RhoA activity is necessary to downregulate integrin-based adhesive events that allow leukocyte migration (75).

Consistent with our observations of early RhoA activation in response to HGF, we observed morphological differences in F-actin stress fiber profiles between epithelial cell spreading in the presence or absence of HGF. In addition to enhanced parallel orientation of actin fibers spreading into wounds, HGF induced dramatic expansion of an area of partially flattened cells immediately behind the leading edge. This led us to define three distinct planar zones in migrating epithelial cells: zone I, comprising completely flattened cells at the leading edge; zone II, encompassing partially flattened cells behind the edge; and zone III, designating normal tall columnar cells distant from the wound. The ability of HGF to increase the dimensions of zone II and to activate RhoA could undoubtedly facilitate the coverage of denuded areas and promote wound closure. Interestingly, although RhoA activity promotes cell spreading in response to HGF in our model, it may limit cell spreading at later stages of HGF-induced morphogenesis. Accordingly, it has been observed that the RhoA effector Rho kinase is necessary to limit excessive extension of cytoplasmic protrusions in a three-dimensional MDCK cyst model in response to HGF-conditioned media (77).

Another mechanism whereby Rho may play an important role in HGF-mediated enhancement of migration is via regulation of focal adhesion complexes, sites where cell membranes interface with the matrix through integrin-based proteinaceous complexes (11). RhoA is known to promote focal adhesion formation, and inhibition of Rho activity decreases focal adhesion complexes in spreading cells (57, 65). In particular, paxillin is a key scaffolding protein in focal adhesion complexes (3, 4, 62), and increased paxillin phosphorylation and affiliation with focal adhesion kinase has been documented in renal epithelial cells following incubation with HGF (26). In our intestinal epithelial model, we noted significant aggregation of paxillin-positive plaques in zone II cells of HGF-treated monolayers in addition to a normal distribution at the tips of actin filaments in zone I of control cells. Most interestingly, this was accompanied by upregulation of paxillin, and, to a lesser extent, {beta}1-integrin, protein levels in a cycloheximide-sensitive manner. This indicates stimulation of translation during the early spreading response to HGF. In a similar scenario, enhanced paxillin protein levels have been reported in response to the growth factor heregulin in a number of cancer cell lines, as well as in metastatic breast cancer biopsies (72). Thus combined evidence suggests that paxillin upregulation in response to growth factors correlates with a promigratory phenotype, although invasive potential has previously been correlated only with the tyrosine phosphorylation status of paxillin rather than its total protein levels (51).

To test whether paxillin upregulation in our T84 model was specifically dependent on the promigratory effects of HGF (via its potential binding partner {beta}1-integrin), we blocked migration using an antibody to {beta}1-integrin (27). However, paxillin protein levels were still increased in response to HGF treatment. Furthermore, although HGF increased the amount of paxillin, and {beta}1-integrin to a lesser extent, in epithelial plasma membrane compartments following subcellular fractionation, there was no increase in the coassociation of paxillin with either {beta}1-integrin or actin in these plasma membrane compartments. Thus it is intriguing to speculate that an early consequence of HGF-induced epithelial cell spreading is augmentation of a cytosolic pool of paxillin, which may be available to facilitate dynamic assembly of focal complexes and their rapid turnover during cell spreading or migration. This is supported by our observations of reduced Y118-phosphorylated paxillin recovered from HGF lysates relative to controls in our model, particularly because tyrosine phosphorylated paxillin may preferentially associate with focal adhesion sites to facilitate the assembly of multiprotein scaffolding complexes (2, 60, 70). However, paxillin phosphorylation and its contribution to migration is a contentious issue. Multiple reports (32, 43, 45) have shown enhanced tyrosine phosphorylation of paxillin in response to HGF or other growth factors. Paxillin phosphorylation by Src kinases seems to be a prerequisite for paxillin/ERK associations that drive morphogenesis (19). Furthermore, a phosphorylation-deficient mutant of paxillin has been reported to retard random motility of rat bladder epithelial cells (44), but with identical methodology, the invasiveness of rat hepatoma epithelial cells was reportedly enhanced by phosphorylation-deficient paxillin mutants (76). It has been speculated that the discrepancies highlight differential roles for tyrosine-phosphorylated paxillin in random vs. stimulated migration (59). Our observations that HGF-stimulated reductions in tyrosine-phosphorylated paxillin coincide with enhanced epithelial spreading support this model. It is particularly interesting that such decreases in tyrosine-phosphorylated paxillin occur against a background of increased paxillin protein synthesis in response to HGF. However, caution must be exercised in interpreting the effects of growth factors on motility, because spatial and temporal events ultimately govern responsiveness in individual systems. HGF personifies the complexity of such biological systems by exerting distinct temporal effects; beginning with spreading, proceeding to scattering, dissociation, and terminating with morphogenesis. With this in mind, HGF has been shown to have no effect on tyrosine phosphorylation of paxillin during a late scattering response in HT-29 colonic epithelial cells (18).

To conclude, we have reported a novel upregulation of paxillin translation concomitant with a reduction in paxillin tyrosine phosphorylation in an early response to Rho-dependent epithelial cell spreading induced by HGF. A large pool of paxillin localizes behind rather than at the leading edge of spreading cells and is not complexed with the actin cytoskeleton or the membrane. Thus we propose that HGF promotes the formation of a free cytosolic pool of paxillin that dynamically feeds rapidly assembling focal complexes as a mechanism of enhancing epithelial cell spreading.


   GRANTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
REFERENCES
 
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant DK-55679 (to A. Nusrat), Crohn’s and Colitis Foundation of America (Research Fellowship Award; to A. M. Hopkins), the German Research Foundation (Deutsche Forschungsgemeinschaft Br 2093/1–1; to M. Bruewer), NIDDK training Grant T32DK-07771 (to J. J. Ha), and the American Digestive Health Foundation (Fellowship/Faculty transition award; to S. V. Walsh).


   ACKNOWLEDGMENTS
 
The authors thank Dr. Susan Voss for cell culture assistance, Dr. Kenneth Yamada (National Institutes of Health, Bethesda, MD) for MAb13 {beta}1-integrin antibody, Dr. Marc Symons (Picowar Institute, Manhasset, NY) for supplying the RhoA N19 construct, and Drs. Charles Parkos and Markus Utech for critical reading of the manuscript.


   FOOTNOTES
 

Address for reprint requests and other correspondence: A. Nusrat, Dept. of Pathology and Laboratory Medicine, Emory Univ., Rm. 105E, Whitehead Research Bldg., 615 Michael St., Atlanta, GA 30322 (E-mail: anusrat{at}emory.edu)

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

* A. M. Hopkins and M. Bruewer contributed equally to this work. Back


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Birchmeier C and Gherardi E. Developmental roles of HGF/SF and its receptor, the c-Met tyrosine kinase. Trends Cell Biol 8: 404–410, 1998.[CrossRef][ISI][Medline]
  2. Cattelino A, Cairo S, Malanchini B, and de Curtis I. Preferential localization of tyrosine-phosphorylated paxillin in focal adhesions. Cell Adhes Commun 4: 457–467, 1997.[ISI][Medline]
  3. Critchley DR. Focal adhesions–the cytoskeletal connection. Current Op Cell Biol 12: 133–139, 2000.
  4. Critchley DR, Holt MR, Barry ST, Priddle H, Hemmings L, and Norman J. Integrin-mediated cell adhesion: the cytoskeletal connection. Biochem Soc Symp 65: 79–99, 1999.[Medline]
  5. Crosson CE, Klyce SD, and Beuerman RW. Epithelial wound closure in the rabbit cornea. A biphasic process. Invest Ophthalmol Vis Sci 27: 464–473., 1986.[Abstract]
  6. Dignass AU, Lynch-Devaney K, and Podolsky DK. Hepatocyte growth factor/scatter factor modulates intestinal epithelial cell proliferation and migration. Biochem Biophys Res Commun 202: 701–709, 1994.[CrossRef][ISI][Medline]
  7. Elliott BE, Hung WL, Boag AH, and Tuck AB. The role of hepatocyte growth factor (scatter factor) in epithelial-mesenchymal transition and breast cancer. Can J Physiol Pharmacol 80: 91–102, 2002.[CrossRef][ISI][Medline]
  8. Feil W. Repair of rabbit duodenal mucosa after acid injury in vivo and in vitro. Gateroenterology 92: 1973–1986, 1987.
  9. Fenteany G, Janmey PA, and Stossel TP. Signaling pathways and cell mechanics involved in wound closure by epithelial cell sheets. Curr Biol 10: 831–838, 2000.[CrossRef][ISI][Medline]
  10. Galimi F, Brizzi MF, and Comoglio PM. The hepatocyte growth factor and its receptor. Stem Cells 11, Suppl 2: 22–30, 1993.
  11. Geiger B, Bershadsky A, Pankov R, and Yamada KM. Transmembrane crosstalk between the extracellular matrix–cytoskeleton crosstalk. Nat Rev Mol Cell Biol 2: 793–805, 2001.[CrossRef][ISI][Medline]
  12. Gherardi E and Stoker M. Hepatocytes and scatter factor. Nature 346: 228, 1990.[Medline]
  13. Giancotti FG. Integrin signaling. Science 285: 1028–1032, 1999.[Abstract/Free Full Text]
  14. Glenney JR Jr and Zokas L. Novel tyrosine kinase substrates from Rous sarcoma virus-transformed cells are present in the membrane skeleton. J Cell Biol 108: 2401–2408, 1989.[Abstract/Free Full Text]
  15. Goldberg ID and Rosen EM. Hepatocyte growth factor-scatter factor (HGF-SF) and the c-met receptor. Introduction Exs 65: 13–15, 1993.
  16. Haddad R, Lipson KE, and Webb CP. Hepatocyte growth factor expression in human cancer and therapy with specific inhibitors. Anticancer Res 21: 4243–4252, 2001.[ISI][Medline]
  17. Hall A, Paterson HF, Adamson P, and Ridley AJ. Cellular responses regulated by rho-related small GTP-binding proteins. Philos Trans R Soc Lond B Biol Sci 340: 267–271, 1993.[ISI][Medline]
  18. Herrera R. Modulation of hepatocyte growth factor-induced scattering of HT29 colon carcinoma cells. Involvement of the MAPK pathway. J Cell Sci 111: 1039–1049, 1998.[Abstract]
  19. Ishibe S, Joly D, Zhu X, and Cantley LG. Phosphorylation-dependent paxillin-ERK association mediates hepatocyte growth factor-stimulated epithelial morphogenesis. Mol Cell 12: 1275–1285, 2003.[CrossRef][ISI][Medline]
  20. Jesaitis LA and Goodenough DA. Molecular characterization and tissue distribution of ZO-2, a tight junction protein homologous to ZO-1 and the Drosophila discs-large tumor suppressor protein. J Cell Biol 124: 949–961, 1994.[Abstract/Free Full Text]
  21. Jiang WG and Hiscox S. Hepatocyte growth factor/scatter factor, a cytokine playing multiple and converse roles. Histol Histopathol 12: 537–555, 1997.[ISI][Medline]
  22. Kaoutzani P, Parkos CA, Delp-Archer C, and Madara JL. Isolation of plasma membrane fractions from the intestinal epithelial model T84. Am J Physiol Cell Physiol 264: C1327–C1335, 1993.[Abstract/Free Full Text]
  23. Kim JS, McKinnis VS, Nawrocki A, and White SR. Stimulation of migration and wound repair of guinea-pig airway epithelial cells in response to epidermal growth factor. Am J Respir Cell Mol Biol 18: 66–74, 1998.[Abstract/Free Full Text]
  24. Kozma R, Ahmed S, Best A, and Lim L. The Ras-related protein Cdc42Hs and bradykinin promote formation of peripheral actin microspikes and filopodia in Swiss 3T3 fibroblasts. Mol Cell Biol 15: 1942–1952, 1995.[Abstract]
  25. Lauffenburger DA and Horwitz AF. Cell migration: a physically integrated molecular process. Cell 84: 359–369, 1996.[CrossRef][ISI][Medline]
  26. Liu ZX, Yu CF, Nickel C, Thomas S, and Cantley LG. Hepatocyte growth factor induces ERK-dependent paxillin phosphorylation and regulates paxillin-focal adhesion kinase association. J Biol Chem 277: 10452–10458, 2002.[Abstract/Free Full Text]
  27. Lotz MM, Nusrat A, Madara JL, Ezzell R, Wewer UM, and Mercurio AM. Intestinal epithelial restitution. Involvement of specific laminin isoforms and integrin laminin receptors in wound closure of a transformed model epithelium. Am J Pathol 150: 747–760, 1997.[Abstract]
  28. Madara JL, Stafford J, Dharmsathaphorn K, and Carlson S. Structural analysis of a human intestinal epithelial cell line. Gastroenterology 92: 1133–1145, 1987.[ISI][Medline]
  29. Maffe A and Comoglio PM. HGF controls branched morphogenesis in tubular glands. Eur J Morphol 36, Suppl: 74–81, 1998.
  30. Maina F and Klein R. Hepatocyte growth factor, a versatile signal for developing neurons. Nat Neurosci 2: 213–217, 1999.[CrossRef][ISI][Medline]
  31. Matsumoto K, Date K, Ohmichi H, and Nakamura T. Hepatocyte growth factor in lung morphogenesis and tumor invasion: role as a mediator in epithelium-mesenchyme and tumor-stroma interactions. Cancer Chemother Pharmacol 38: S42–S47, 1996.[Medline]
  32. Maulik G, Kijima T, Ma PC, Ghosh SK, Lin J, Shapiro GI, Schaefer E, Tibaldi E, Johnson BE, and Salgia R. Modulation of the c-Met/hepatocyte growth factor pathway in small cell lung cancer. Clin Cancer Res 8: 620–627, 2002.[Abstract/Free Full Text]
  33. Michalopoulos GK. HGF in liver regeneration and tumor promotion. Prog Clin Biol Res 391: 179–185, 1995.[Medline]
  34. Mitchison TJ. Actin-based cell motility and cell locomotion. Cell 84: 371–379, 1996.[CrossRef][ISI][Medline]
  35. Moore R, Carlson S, and Madara JL. Rapid barrier restitution in an in vitro model of intestinal epithelial injury. Lab Invest 60: 237–244, 1989.[ISI]
  36. Nabeshima K, Shimao Y, Inoue T, Itoh H, Kataoka H, and Koono M. Hepatocyte growth factor/scatter factor induces not only scattering but also cohort migration of human colorectal-adenocarcinoma cells. Int J Cancer 78: 750–759, 1998.[CrossRef][ISI][Medline]
  37. Nakamura T, Nawa K, and Ichihara A. Partial purification and characterization of hepatocyte growth factor from serum of hepatectomized rats. Biochem Biophys Res Commun 122: 1450–1459, 1984.[CrossRef][ISI][Medline]
  38. Naldini L, Vigna E, Narsimhan RP, Gaudino G, Zarnegar R, Michalopoulos GK, and Comoglio PM. Hepatocyte growth factor (HGF) stimulates the tyrosine kinase activity of the receptor encoded by the proto-oncogene c-MET. Oncogene 6: 501–504., 1991.[ISI][Medline]
  39. Nobes CD and Hall A. Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81: 53–62, 1995.[CrossRef][ISI][Medline]
  40. Nusrat A. Hepatocyte growth factor/scatter factor effects on epithelia. Regulation of intercellular junctions in transformed and nontransformed cell lines, basolateral polarization of c-met receptor in transformed and natural intestinal epithelia, and induction of rapid wound repair in a transformed model epithelium. J Clin Invest 93: 2056–2065, 1994.[ISI][Medline]
  41. Nusrat A. Intestinal epithelial restitution. Characterization of a cell culture model and mapping of cytoskeletal elements in migrating cells. J Clin Invest 89: 1501–1511, 1992.[ISI][Medline]
  42. Nusrat A. Tight junctions are membrane microdomains. J Cell Sci 113: 1771–1781, 2000.[Abstract]
  43. Parr C, Davies G, Nakamura T, Matsumoto K, Mason MD, and Jiang WG. The HGF/SF-induc