Involvement of focal adhesion kinase in hepatocyte growth factor-induced scatter of Madin-Darby canine kidney cells.

Focal adhesion kinase (FAK) has been implicated to play a critical role in integrin-mediated control of cell behavior. However, it is unclear whether FAK also participates in the regulation of growth factor-elicited cellular functions. In this study, we have demonstrated that although overexpression of FAK in Madin-Dardy canine kidney cells did not alter their growth property or ability to form tubules within collagen gel upon hepatocyte growth factor (HGF) stimulation, it apparently enhanced HGF-induced cell scattering. This enhancement was largely because of an increase in the third phase (i.e. cell migration) of cell scattering rather than the first two phases (i.e. cell spreading and cell-cell dissociation). Conversely, the expression of FAK-related nonkinase significantly ( approximately 60%) inhibited HGF-induced cell migration. Moreover, we have found that the effect of FAK on promoting HGF-induced cell motility was greatly dependent on cell-matrix interactions. We showed that HGF treatment selectively increased the expression of integrins alpha(2) and, to a lesser extent, alpha(3) in Madin-Dardy canine kidney cells and that a monoclonal antibody against integrin alpha(2) efficiently blocked HGF-enhanced cell migration on collagen. In our efforts to determine the mechanism by which FAK promotes HGF-induced cell migration, we found that FAK mutants deficient in phosphatidylinositol 3-kinase or p130(Cas) binding failed to promote HGF-induced cell migration. Interestingly, cells expressing a FAK mutant defective in Grb2 binding exhibited a rate of migration approximately 50% lower than that of cells expressing wild type FAK in response to HGF stimulation. Taken together, our results suggest a link between HGF-increased integrin expression, FAK activation, and enhanced cell motility and implicate a role for FAK in the facilitation of growth factor-induced cell motility.

Focal adhesion kinase (FAK) has been implicated to play a critical role in integrin-mediated control of cell behavior. However, it is unclear whether FAK also participates in the regulation of growth factor-elicited cellular functions. In this study, we have demonstrated that although overexpression of FAK in Madin-Dardy canine kidney cells did not alter their growth property or ability to form tubules within collagen gel upon hepatocyte growth factor (HGF) stimulation, it apparently enhanced HGF-induced cell scattering. This enhancement was largely because of an increase in the third phase (i.e. cell migration) of cell scattering rather than the first two phases (i.e. cell spreading and cell-cell dissociation). Conversely, the expression of FAK-related nonkinase significantly (ϳ60%) inhibited HGF-induced cell migration. Moreover, we have found that the effect of FAK on promoting HGF-induced cell motility was greatly dependent on cell-matrix interactions. We showed that HGF treatment selectively increased the expression of integrins ␣ 2 and, to a lesser extent, ␣ 3 in Madin-Dardy canine kidney cells and that a monoclonal antibody against integrin ␣ 2 efficiently blocked HGFenhanced cell migration on collagen. In our efforts to determine the mechanism by which FAK promotes HGFinduced cell migration, we found that FAK mutants deficient in phosphatidylinositol 3-kinase or p130 Cas binding failed to promote HGF-induced cell migration. Interestingly, cells expressing a FAK mutant defective in Grb2 binding exhibited a rate of migration ϳ50% lower than that of cells expressing wild type FAK in response to HGF stimulation. Taken together, our results suggest a link between HGF-increased integrin expression, FAK activation, and enhanced cell motility and implicate a role for FAK in the facilitation of growth factor-induced cell motility.
Focal adhesion kinase (FAK), 1 a 125-kDa cytoplasmic tyrosine kinase localized in focal adhesions, has been implicated to play an important role in regulating integrin-mediated cellular functions, including cell spreading (1,2), cell migration (3,4), cell cycle progression (5,6), and cell survival (7)(8)(9). The ability of FAK to regulate these cellular functions is believed to be dependent on its ability to interact with several intracellular signaling molecules including Src family kinases (10,11), phosphatidylinositol 3-kinase (PI3K; Ref. 12), adapter protein Grb2 (13), and docking protein p130 Cas (14,15). Tyr-397 has been identified as the major site of FAK autophosphorylation (16), and the binding site for the Src homology 2 domains of Src (10,11) and PI3K (17). A FAK mutant deficient only in PI3K binding has recently been introduced by a substitution of Asp-395 with Ala (18). The proline-rich sequence region of FAK (residues 712-718) has been identified as the major binding site for the SH3 domain of p130 Cas (14,19). Upon binding to FAK, Src phosphorylates FAK at Tyr-925 (20), creating a binding site for the complex of Grb2 and Sos. Because Sos functions as a guanine nucleotide exchange factor for Ras (21), it has been proposed that FAK may link integrin-initiated signals to the Ras/ mitogen-activated protein kinase cascades (13,20,22).
Hepatocyte growth factor (HGF), also known as scatter factor, is a multifunctional growth factor that elicits mitogenic, motogenic, and morphogenic activities in various cell types (23). The diverse biological effects of HGF are transmitted through activation of its transmembrane receptor encoded by the c-met proto-oncogene (24,25). The Met receptor is a heterodimer composed of a 45-kDa ␣ chain that remains entirely extracellular and a 145-kDa ␤ chain that traverses the plasma membrane and contains the intracellular tyrosine kinase domain (26 -29). Upon HGF binding, the intrinsic tyrosine kinase of the receptor is activated resulting in autophosphorylation on specific tyrosine residues in the ␤ chain (30,31). The phosphorylated tyrosine residues can then associate with molecules containing Src homology 2 and phosphotyrosine-binding domains that act to transduce extracellular signals to the cell interior (32).
We have previously demonstrated that HGF stimulates tyrosine phosphorylation of FAK and its association with the Grb2-Sos complex, which further contributes to the activation of extracellular signal-regulated kinase (ERK) by HGF in human embryonic kidney 293 cells (33). However, the role of FAK in HGF-elicited cellular functions has not been clarified. In this study, we have found that overexpression of FAK in Madin-Darby canine kidney (MDCK) cells apparently enhanced HGFinduced cell scattering. In contrast, the expression of FAKrelated nonkinase (FRNK) significantly inhibited HGFinduced cell motility. Moreover, our results suggest that, in addition to PI3K and p130 Cas , the Grb2 binding may also contribute to the ability of FAK to promote cell migration. Taken together, these results strongly implicated a role for FAK in the facilitation of growth factor-enhanced cell motility. Technologies, Inc. Collagen type I, fibronectin, and vitronectin were purchased from Collaborative Biomedical Products (Bedford, MA). The MEK inhibitor PD98059 and G418 sulfate were purchased from Calbiochem (San Diego, CA). The monoclonal anti-hemagglutinin (HA) epitope was purchased from Roche Molecular Biochemicals. The rabbit polyclonal anti-integrins ␣ 2 (AB1944), ␣ 3 (AB1948), and ␣ V (AB1930) were purchased from Chemicon (Temecula, CA). The monoclonal anti-FAK (clone 77) and monoclonal anti-phosphotyrosine (PY20) were purchased from Transduction Laboratories (Lexington, KY). The rabbit polyclonal anti-ERK (sc-94) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The rabbit polyclonal anti-phosphoERK was purchased from New England Biolabs, Inc. (Beverly, MA). The monoclonal antibody (5E8) recognizing ␣ 2 integrin was kindly provided by Dr. R. B. Bankert (Roswell Park Cancer Institute, Buffalo, NY) and was described previously (34). The pKH3 expression plasmid encoding HA epitope-tagged FRNK was provided by Dr, J.-L. Guan (Cornell University, Ithaca, NY) and described previously (5).
Cell Lines and Transfections-MDCK II 3B5 cells overexpressing HA epitope-tagged wild type (WT) FAK, D395A, P712A/P715A, and Y925F have been described previously (9) and were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 0.5 mg/ml G418. To generate cells stably expressing FRNK, MDCK cells were grown on 60-mm dishes and transfected essentially as described (9) using LipofectAMINE following the manufacturer's instructions. Clones were selected in G418-containing medium and screened by immunoblotting with anti-HA.
Biological Assays-For cell proliferation assays, 10 5 MDCK cells were plated on 100-mm culture dish in medium containing 10 or 1% serum with or without 1 ng/ml HGF. The medium was changed daily for 6 days. Cell number was measured every 2 days using a hemacytometer.
For tubulogenesis assays, MDCK cells were suspended at a concentration of 5 ϫ 10 4 cells/ml in cold collagen solution as described previously (35). An aliquot (1 ml) of cell suspension was dispensed into a well of 6-well plate and allowed to gel for about 20 min at 37°C before adding 1 ml of medium containing 10% serum and 20 ng/ml HGF. The medium was changed daily. After 8 days, the tubules were photographed using a phase contrast microscope.
For scatter assays, MDCK cells were allowed to grow as discrete colonies by seeding at 2 ϫ 10 3 /60-mm dish. When the majority of colonies contained between 20 -40 cells (60 -72 h after seeding), the medium was replaced by fresh medium containing 5% serum and 10 ng/ml HGF. The effect of HGF on scatter of MDCK cells was evaluated every 3 h for 12 h and photographed under a phase contrast microscope at 100ϫ magnification. To quantify the assay, cells that have lost contact with their neighboring cells and exhibited a fibroblast-like shape in total 50 colonies were enumerated.
For cell spreading assays, MDCK cells were allowed to grow as discrete colonies and then treated with 10 ng/ml HGF as described above. 2 h after HGF stimulation, cells were collected by trypsinization and suspended in serum-free medium at 5 ϫ 10 4 cells/ml. 2 ml of cell suspension was added to 60-mm dish that had been coated with collagen (20 g/dish) and blocked with bovine serum albumin. Cells were allowed to spread for 20 min, and spread cells were scored under a phase contrast microscope. Cells with extended processes and not phase bright were defined as spread cells.
For cell dissociation assays, MDCK cells were allowed to grow as discrete colonies and then treated with 10 ng/ml HGF as described above. 4 h after HGF stimulation, cells were collected in serum-free medium by scraping and passed through a micropipette 30 times under a constant force. The number of cell particles (N p ) containing more than three cells was measured using a hemacytometer. The cell suspension containing cell particles was subsequently subjected to centrifugation and trypsinization for cell number (N c ) measurement. The cell dissociation index was expressed as N p /N c ϫ 100%.
For cell migration assays, MDCK cells were allowed to grow as discrete colonies and then treated with 10 ng/ml HGF as described above. 12 h after HGF stimulation, cells were collected by trpsinization and suspended in serum-free medium at 5 ϫ 10 5 /ml. Migration assays were carried out in a Neuro Probe 48-well chemotaxis chamber (Cabin John, MD). The medium containing collagen or other matrix proteins at 10 g/ml was added to the lower chamber. The lower and upper chambers were separated by a polycarbonate membrane (8-m pore size, Poretics, Livermore, CA). Cells were allowed to migrate for 7 h at 37°C in a humidified atmosphere containing 5% CO 2 . The membrane was fixed in methanol for 10 min and stained with modified Giemsa stain (Sigma) for 1 h. Cells on the upper side of the membrane were removed by cotton swabs. Cells on the lower side of the membrane were counted using a light microscope at 200ϫ magnification. To evaluate the significance of integin ␣ 2 ␤ 1 in cell migration, cells were suspended in serumfree medium containing various amounts of the monoclonal antibody 5E8 before loading them to the chemotaxis chamber. In some experiments, cells were pretreated with PD98059 at the final concentration of 100 M for 1 h before trypsinization and allowed to migrate on collagen in the presence of the inhibitor.

RESULTS
We have previously established stable MDCK cell lines overexpressing HA epitope-tagged FAK, in which the expression level of ectopic FAK is 2-3-fold of endogenous FAK (9). To examine whether FAK plays a role in HGF-elicited cellular functions, cells from MDCK clones overexpressing wild type FAK (WT) and control clones (Neo) resistant to G418 sulfate were subjected to assays for evaluating their proliferation, tubulogenesis, and scatter upon HGF stimulation (Fig. 1). No matter whether HGF was present or not, no differences in the growth curves of control and WT cells were observed in medium with 10 or 1% serum (Fig. 1A). Given that the doubling time of MDCK cells is approximately 12 h, even relatively small differences in the rate of cell proliferation should become apparent after 6 days. Thus, these results indicate that neither FAK overexpression nor HGF stimulates proliferation of MDCK cells under our experimental conditions. In addition, judging from the percentage of tubule formation and the morphology of tubules, FAK overexpression did not appear to affect the ability of MDCK cells to form branch tubules in three-dimensional collagen gels in response to HGF stimulation (Fig. 1B).
However, FAK overexpression in MDCK cells apparently enhanced HGF-induced cell scattering ( Fig. 1C and see below). Three days after sparsely seeding, MDCK cells grew as discrete colonies, each of which contained between 20 and 40 cells. Consistent with previous studies (36,37), HGF induced cell colonies to spread during the first 2-4 h. Subsequently, some cells within each colony began to detach from their neighboring cells and exhibited a shape resembling that of motile fibroblasts. These cells continued to migrate, finally leading to a "scatter" phenomenon. We quantified the HGF-induced cell scattering by counting the number of cells that have lost contact with their neighbors in a total of 50 colonies. As shown in Fig. 1C, FAK overexpression in MDCK cells prominently enhanced the scatter effect of HGF. The quantitative results indicated that the number of scattered cells from WT clones was approximately twice that of control clone 6 or 9 h after HGF stimulation.
As demonstrated in this study (Fig. 1C) and elsewhere (36,37), the HGF-induced cell scattering can be further divided into three phases, namely cell spreading, cell dissociation, and cell migration. To determine which phase was potentially affected by FAK overexpression, control and WT cells were allowed to grow as discrete colonies and then treated with or without HGF for 2, 4, or 12 h before subjecting to cell spreading, dissociation, or migration assays, respectively. As shown in Fig. 2A, in the absence of HGF treatment, the number of WT cells spreading on collagen was ϳ2-fold of Neo cells, supporting a role of FAK in promoting cell spreading on matrix proteins, as described previously (1,2). The HGF treatment induced a similar increase (ϳ10%) in cell spreading for both control and WT cells. This result was consistent with our observation that HGF induced colonies of control and WT cells to spread with a similar extent during first 3 h after HGF stimulation (Fig. 1C). Thus, these results suggest that the effect of HGF on cell spreading is likely to be independent of FAK expression. Next we determined the degree of cell-cell adhesion using cell dissociation assays. As shown in Fig. 2B, 4-h HGF treatment induced a similar increase (ϳ10%) in the degree of cell dissociation for both control and WT cells. However, FAK overexpression did not appear to have any effect on this event. This conclusion was further supported by our observations that the expression and tyrosine phosphorylation of E-cadherine or ␤-catenin, two proteins known to be involved in cell-cell adhesion (38), were not affected by FAK overexpression in MDCK cells (data not shown).
To examine whether FAK participates in the third phase (i.e. cell migration) of HGF-induced cell scattering, cells from control and WT clones were trypsinized 12 h after HGF stimulation and then subjected to cell migration assays using collagen as an attractant (Fig. 2C). In the absence of HGF stimulation, WT cells had exhibited a higher (ϳ5-fold) rate of cell migration in comparison with the control cells. This is consistent with previous studies (4,18) showing that FAK overexpression in Chinese hamster ovary (CHO) cells promoted their migration on fibronectin. Significantly, cells overexpressing FAK dramatically increased their migration on collagen upon HGF stimu-lation, indicating that FAK overexpression synergized with the effect of HGF to promote cell migration. To further confirm the role of FAK in HGF-enhanced cell migration, FRNK, a FAK carboxyl-terminal construct known to function as a dominant negative mutant of FAK (1, 2), was overexpressed in MDCK cells. As shown in Fig. 3, approximately 60% of HGF-induced cell migration was inhibited by FRNK expression. Together, these results indicate that FAK is involved in regulating MDCK cell motility induced by not only collagen but also HGF.
To examine whether the effect of FAK overexpression on promoting HGF-induced cell motility is matrix-dependent, control and WT cells were subjected to cell migration assays with or without using a matrix protein as an attractant (Fig. 4A). Although the HGF treatment generally increased MDCK cell motility, the effect of FAK overexpression on promoting HGFinduced cell motility was much more prominent when a matrix protein, especially collagen, was present. These data indicated that the ability of FAK to promote HGF-induced cell migration is mainly matrix-dependent. It is known that integrin ␣ 2 ␤ 1 is the major receptor for collagen on MDCK cells (39). To examine the involvement of integrin ␣ 2 ␤ 1 in HGF-enhanced cell migration on collagen, a monoclonal antibody 5E8 (34), which recognizes integrin ␣ 2 subunit and prevents integrin ␣ 2 ␤ 1 from its ligand binding, was applied to the migration assays (Fig. 4B). The treatment of WT cells with this blocking antibody, but not a control antibody (monoclonal anti-HA), resulted in a dose-dependent inhibition of HGF-induced migration, indicating that the effect of FAK on promoting HGF-induced cell migration on collagen is mediated by integrin ␣ 2 ␤ 1 . It is noteworthy that at

FIG. 1. Effect of FAK overexpression on MDCK cell proliferation, tubulogenesis, and scattering in response to HGF stimulation.
A, cell proliferation assay. MDCK cells overexpressing FAK (WT) or control cells (Neo) were seeded at 10 5 cells per 10-cm culture dish in medium containing 10% or 1% serum with (ϩ) or without (Ϫ) 1 ng/ml HGF. The number of cells was measured every 2 days. Mean cell counts from three experiments are shown. B, tubulogenesis assay. Control or WT cells were grown in collagen gels covered with a layer of medium containing 10% serum and 20 ng/ml HGF and photographed 8 days later using a phase contrast microscope. Representative fields of tubules from Neo and WT cells are shown, respectively. Magnification, 100ϫ. C, scatter assay. Cell colonies of Neo or WT cells were incubated with 10 ng/ml HGF, and the number of scattered cells in total 50 colonies was measured every 3 h. A cell was judged as a scattered cell when it has lost contact with its neighbors and exhibited a fibroblast-like phenotype. Data (means Ϯ S.E.) are from at least nine independent experiments using three independent clones for each experimental group. The HGF-induced morphological changes of two representative colonies from Neo and WT cells are shown, respectively. Magnification, 100ϫ. the concentration of 5 g/ml, this blocking antibody was able to completely inhibit WT cell migration in the absence of HGF stimulation but reached only a partial (ϳ60%) inhibition in the presence of HGF stimulation. These results suggested that HGF might increase the expression of integrins ␣ 2 ␤ 1 and/or alter their conformation, leading to a less efficiency for this antibody to exert its blocking function. To examine the effect of HGF on integrin expression, control and WT cells were treated with or without HGF for 12 h, and their whole cell lysates were prepared for immunoblotting analysis. As shown in Fig. 4C, HGF induced a marked (2-fold) increase in the expression of integrin ␣ 2 and, to a lesser extent, ␣ 3 . Conversely, the expression of integrin ␣ V (Fig. 4C) or FAK (data not shown) did not appear to be affected by HGF stimulation. Together, these results suggest a link between HGF-increased integrin expression, FAK activation, and enhanced cell motility.
To investigate the mechanisms by which FAK promotes HGF-induced cell migration, stable MDCK cell lines overexpressing FAK mutants including D395A, P712A/P715A, and Y925F, deficient in binding to PI3K, p130 Cas , and Grb2, respectively, were subjected to cell migration assays. The expression levels of these FAK mutants were slightly higher than that of WT FAK (Fig. 5A). In addition, HGF treatment was able to induce an increase in tyrosine phosphorylation of these exogenously expressed FAK proteins (Fig. 5B). Similar to prior experiments using CHO cells (18), FAK D395A and P712A/ P715A mutants failed to promote MDCK cell migration in the absence of HGF stimulation (Fig. 5C). Interestingly, these two FAK mutants also failed to promote HGF-induced cell migra-  (Fig. 5C). These results suggest that direct bindings of both PI3K and p130 Cas are required for FAK to promote cell migration induced by matrix proteins alone or together with HGF stimulation.
Cary et al. (40) reported that the mutation at Tyr-925 had no effect on the ability of FAK to promote CHO cell migration. In contrast, we found that MDCK cells expressing FAK Y925F mutant exhibited a rate of cell migration ϳ50% lower than that of cells expressing WT FAK in the condition with or without HGF treatment (Fig. 5C). These results suggest that the direct Grb2 binding and its downstream signals (e.g. activation of ERKs) may also contribute to FAK-promoted cell migration. In fact, we have previously demonstrated that the tyrosine phosphorylation of FAK increased by HGF stimulation leads to ERK activation (33). To examine the potential role of ERKs in FAK-promoted cell migration, control and WT cells were subjected to cell migration assays in the presence of the selective MEK inhibitor PD98059. At the concentration of 100 M, PD98059 efficiently inhibited both ERK1 and ERK2 (Fig. 6A) and decreased ϳ50% of cell migration promoted by FAK overexpression in WT cells in the presence or absence of HGF treatment (Fig. 6B). Together, these results indicate that the Grb2 binding and subsequent ERK activation contribute partially to the ability of FAK to promote cell migration. DISCUSSION In this study, we have used MDCK cells as a model to examine the effect of FAK overexpression on the biological functions of HGF. First we showed that although FAK overexpression did not affect MDCK cell proliferation or tubulogenesis, it apparently enhanced HGF-induced cell scattering. The scatter response of MDCK cells to HGF stimulation can be FIG. 4. Role of matrix proteins and integrin ␣ 2 ␤ 1 in FAK promotion of HGF-induced cell migration. A, cells from a control clone (Neo) or MDCK cells overexpressing FAK (WT) were incubated with (ϩ) or without (Ϫ) 10 ng/ml HGF for 12 h before subjecting to a migration assay. Cells in the upper half of the chamber were allowed to migrate through a porous membrane toward the lower chamber containing medium alone (Dulbecco's modified Eagle's medium) or supplemented with 10 g/ml of collagen, fibronectin, or vitronectin as an attractant. Mean cell counts are from eight fields and two independent experiments. Relative cell migration was calculated based on the level of WT cell migration on collagen in the absence of HGF stimulation. B, WT cells were incubated with (ϩ) or without (Ϫ) 10 ng/ml HGF for 12 h before subjecting to a migration assay using collagen as an attractant. The upper half of the chamber contained cells and various amounts of monoclonal anti-␣ 2 integrin as indicated. Monoclonal anti-HA (50 g/ ml) was used as a control. Mean cell counts are from eight fields and two independent experiments. Relative cell migration was calculated based on the level of WT cell migration without treatments of HGF and anti-␣ 2 integrin. C, control or WT cells were treated with (ϩ) or without (Ϫ) 10 ng/ml HGF for 12 h and then lyzed. An equal amount (50 g) of whole cell lysates was analyzed by immunoblotting with an antibody against integrin ␣ 2 , ␣ 3 , or ␣ V . To precipitate similar amounts of HA-tagged FAK proteins from various FAK-overexpressed clones, cell lysates were adjusted before incubating with anti-HA. The immunocomplexes were washed and analyzed by immunoblotting with anti-phosphotyrosine (PY) or anti-FAK. IP, immunoprecipitation. C, stable cell clones were incubated with (ϩ) or without (Ϫ) 10 ng/ml HGF for 12 h before subjecting to migration assays using collagen as an attractant. Values (means Ϯ S.E.) are from at least 36 fields and nine experiments using three independent clones for each experimental group. Relative cell migration was calculated based on the level of WT cell migration in the absence of HGF stimulation. visualized first as centrifugal spreading of cell colonies (after 2-4 h) followed by cell dissociation (after 4 -6 h) and subsequent cell migration (from 6 h). Consistent with the role of FAK in promoting integrin-mediated cell migration (3,4), FAK overexpression appeared only to affect the third phase (i.e. cell migration) of HGF-induced cell scattering (Fig. 1C). Furthermore, we showed that FAK overexpression prominently (ϳ6fold) enhanced the effect of HGF on cell migration (Fig. 2C). Conversely, the expression of FRNK, which functions as a dominant negative mutant of FAK, significantly (ϳ60%) inhibited HGF-induced cell migration (Fig. 3B). Taken together, our results strongly suggest a role for FAK in the facilitation of HGF-induced cell motility.
FAK has previously been shown to play a role in promoting the G 1 phase progression of the cell cycle in NIH-3T3 and human foreskin fibroblasts (5,6). However, in this study, we do not observe any differences in the proliferation of MDCK epithelial cells by FAK overexpression. In fact, even the constitutively active form of FAK (CD2-FAK) by anchoring it to the plasma membrane was unable to promote MDCK cell proliferation (7). These results suggest that the function of FAK in cell cycle regulation may be cell type-dependent. Moreover, consistent with previous reports (41,42), we found that HGF at 1 ng/ml did not stimulate the proliferation of MDCK cells (Fig.  1A), although it is capable of stimulating endothelial cell proliferation (43). Surprisingly, a higher concentration (10 ng/ml) of HGF, which induces cells to scatter, appears to reduce the growth rate of MDCK cells (data not shown).
The tubulogenesis is a complicated process involving various cellular activities including cell proliferation, apoptosis, and cell migration (44 -46). Despite the fact that FAK overexpression enhances HGF-induced cell migration, we did not observe any apparent differences in the tubulogenesis of MDCK cells by FAK overexpression (Fig. 1B). A possible explanation for this is that HGF promotion of tubulogenesis involves the coordinate regulation of a number of intracellular signaling pathways such as the STAT (47), PI3K (48), and mitogen-activated protein kinase (48,49) pathways; therefore a single molecule like FAK may not be sufficient to disturb the balance among these signaling pathways. Alternatively, it could be simply because the enforced expression of FAK in MDCK cells was not high enough to manifest its effect on the tubulogenesis. Nevertheless, our results do not exclude the possibility that FAK may play a permissive role for tubulogenesis.
Although a number of growth factors are known to modulate cell motility, HGF is unique because of the intensity with which it stimulates motility and induces epithelial-mesenchymal transition. Several intracellular signaling proteins have been implicated to act downstream of the HGF receptor to mediate scatter response. For example, PI3K and Ras have been shown to be essential for cell dissociation and migration following stimulation of MDCK cells with HGF (36,37,51). Although GTP-bound Ras interacts with PI3K and may contribute to its activation (52,53), recent studies suggested that Ras and PI3K might act on different signal transduction cascades (i.e. Ras/ ERK and PI3K/Rac) to facilitate HGF-induced cell motility (54,55). We found that the overexpression of FAK in MDCK cells is able to increase the extent of the activation of ERK (Fig. 6A) and PI3K (data not shown) upon HGF stimulation, suggesting that the role of FAK in HGF-induced cell motility may be because of its contribution to both pathways. This assumption was also supported by our observation that FAK enhancement of HGF-induced cell motility could be completely inhibited by the PI3K inhibitor LY294002 (data not shown) and partially (ϳ50%) inhibited by the MEK inhibitor PD98059 (Fig. 6B).
We have previously demonstrated that HGF induces a rapid (within 10 min) increase in FAK phosphorylation, which likely results from the activation of Src upon HGF stimulation (33). Importantly, this immediate effect of HGF on FAK activation was independent of integrin-mediated cell adhesion (33). In this report, we show the first time that a prolonged (12 h) stimulation of HGF selectively increases the expression of integrins ␣ 2 and, to a lesser extent, ␣ 3 in MDCK cells (Fig. 4C), which presumably further leads to the activation of FAK upon cell adhesion to collagen. It is therefore likely that the effect of HGF on FAK activation can be through an immediate (integrin-independent) response and a delayed (integrin-dependent) response. Moreover, it has previously been shown that integrin ␣ 2 ␤ 1 is essential for the HGF promotion of tubulogenesis (45).
Here we show that a blocking antibody against integrin ␣ 2 ␤ 1 efficiently inhibits HGF-enhanced cell migration on collagen (Fig. 4B). Together with our observation that HGF induced an increase in integrin ␣ 2 ␤ 1 expression (Fig. 4C), it is possible that de novo synthesis of integrin ␣ 2 ␤ 1 is required for the long term cellular responses to HGF, such as cell scattering and tubulogenesis. Consistent with this, we found that cycloheximide, a translation inhibitor, blocked HGF-induced cell scattering and tubulogenesis (data not shown).
Similar to previous studies using CHO cells (4,18), we showed that overexpression of WT FAK, but not its mutant deficient in PI3K or p130 Cas binding, in MDCK cells promoted their migration on collagen (Fig. 5C). In addition, our results indicated that the bindings of PI3K and p130 Cas were essential for the ability of FAK to promote HGF-induced cell migration. Together with our finding that HGF-induced cell migration on collagen depends on the expression of integrin ␣ 2 ␤ 1 (Fig. 4B), it is likely that PI3K and p130 Cas act downstream of FAK to facilitate the HGF-regulated integrin-mediated cell migration. Furthermore, we showed that MDCK cells expressing the Y925F FAK mutant defective in the Grb2 binding exhibited a rate of migration ϳ50% lower than that of cells expressing WT FAK (Fig. 5C). In addition, the MEK inhibitor PD98059 at 100 M, which efficiently inhibited the activation of ERKs, significantly (ϳ50%) decreased the migration of MDCK cells overexpressing FAK and their control cells in the condition with or without HGF stimulation (Fig. 6). Taken together, our results suggest that, in addition to PI3K and p130 Cas , the Grb2 binding and the ERK signaling pathway may also contribute, at least in part, to the ability of FAK to promote cell migration.
However, Cary et al. (40) has shown previously that CHO cells expressing the Y925F FAK mutant exhibited an increased level of migration comparable with that of cells expressing WT FAK. In addition, they found that ectopically expressed FAK from CHO cells did not bind to the Src homology 2 domain of Grb2 in vitro, nor did it activate ERKs in CHO cells. Based on these observations, they concluded that the Grb2 binding and the ERK signaling pathway are not involved in FAK-mediated cell migration. The discrepancy between their work and ours is likely because of different FAK constructs used in respective studies. In the study done by Cary et al. (40), the exogenous FAK expressed in CHO cells had one HA epitope tag at the amino terminus and one large T-antigen epitope tag at the carboxyl terminus. In our study, the exogenous FAK expressed in MDCK cells was tagged with three repeats of the HA epitope at the amino terminus and no epitope tag at the carboxyl terminus. The results from in vitro binding experiments using these two different FAK constructs transiently expressed in 293 cells indicated that the carboxyl-terminal epitope tag on FAK might disrupt the binding of FAK to Grb2 (40). Therefore, it is possible that the increased cell migration by the expression of WT FAK tagged at both amino and carboxyl termini may have been underestimated, without considering the contribution of the Grb2 binding and its downstream signaling. This may explain why the migration of CHO cells expressing the Y925F FAK mutant is comparable with that of cells expressing WT FAK, as discussed above.
In addition to cell adhesion to matrix proteins, FAK phosphorylation is also stimulated by growth factors including HGF (33), platelet-derived growth factor (56), epidermal growth factor (57), and fibroblast growth factor (58). However, the mechanism of FAK phosphorylation and its functional significance in cellular responses to these growth factors are poorly understood. In this study, we have demonstrated that FAK plays a role in HGF promotion of integrin-mediated cell motility. Because other growth factors are also known to regulate the expression and/or adhesive properties of integrins (50), our results render it possible that FAK may generally participate in the control of growth factor-elicited cell motility.