Vascular endothelial growth factor regulates focal adhesion assembly in human brain microvascular endothelial cells through activation of the focal adhesion kinase and related adhesion focal tyrosine kinase.

Vascular endothelial growth factor (VEGF) plays a significant role in blood-brain barrier breakdown and angiogenesis after brain injury. VEGF-induced endothelial cell migration is a key step in the angiogenic response and is mediated by an accelerated rate of focal adhesion complex assembly and disassembly. In this study, we identified the signaling mechanisms by which VEGF regulates human brain microvascular endothelial cell (HBMEC) integrity and assembly of focal adhesions, complexes comprised of scaffolding and signaling proteins organized by adhesion to the extracellular matrix. We found that VEGF treatment of HBMECs plated on laminin or fibronectin stimulated cytoskeletal organization and increased focal adhesion sites. Pretreating cells with VEGF antibodies or with the specific inhibitor SU-1498, which inhibits Flk-1/KDR receptor phosphorylation, blocked the ability of VEGF to stimulate focal adhesion assembly. VEGF induced the coupling of focal adhesion kinase (FAK) to integrin alphavbeta5 and tyrosine phosphorylation of the cytoskeletal components paxillin and p130cas. Additionally, FAK and related adhesion focal tyrosine kinase (RAFTK)/Pyk2 kinases were tyrosine-phosphorylated by VEGF and found to be important for focal adhesion sites. Overexpression of wild type RAFTK/Pyk2 increased cell spreading and the migration of HBMECs, whereas overexpression of catalytically inactive mutant RAFTK/Pyk2 markedly suppressed HBMEC spreading ( approximately 70%), adhesion ( approximately 82%), and migration ( approximately 65%). Furthermore, blocking of FAK by the dominant-interfering mutant FRNK (FAK-related non-kinase) significantly inhibited HBMEC spreading and migration and also disrupted focal adhesions. Thus, these studies define a mechanism for the regulatory role of VEGF in focal adhesion complex assembly in HBMECs via activation of FAK and RAFTK/Pyk2.

The blood-brain barrier (BBB) 1 is formed by a complex cellular system of endothelial cells, astroglia, pericytes, perivascular macrophages, and basal lamina (1). The microvasculature responds dynamically to the flow of substances, free radical exposure, and cytokine generation associated with focal ischemia. During focal cerebral ischemia, with the fall in tissue oxygenation, microvascular permeability barriers are lost (1)(2)(3)(4), and the expression of VEGF is significantly enhanced in the ischemic core and penumbra of the ischemic brain (5).
Angiogenesis, the formation of new blood capillaries, is an important component of normal physiological processes such as wound healing and development (6). Angiogenesis is also a complex process involving endothelial cell movement and proliferation and endothelial cell-mediated degradation of the extracellular matrix. VEGF is a key regulator of endothelial cell functions (6 -14). Brain endothelial cells are linked to each other by gap, tight, and adherens-type junctions and are linked to the extracellular matrix by a variety of integrin and other adhesion molecules. VEGF activates endothelial cells, in part, by stimulating signal transduction pathways that regulate the enzymatic components of adhesion complexes. VEGF induced the tyrosine phosphorylation of vascular endothelial (VE) cadherins, molecules important in endothelial cell migration (15), and also induced phosphorylation of the tight junction proteins occludin and occludin-1 (16,17). VEGF enhanced the expression of ␣ 1 ␤ 1 and ␣ 2 ␤ 2 integrins, and neutralizing antibodies to ␣ v ␤ 1 integrins blocked growth factor-induced neovascularization (18 -20).
VEGF exhibits high affinity binding to two distinct receptor tyrosine kinases, the Fms-like tyrosine kinase Flt-1 and the Flk-1/KDR (21,22). Flk-1/KDR and not Flt-1 is able to mediate the mitogenic and chemotactic effects of VEGF in endothelial cells. VEGF stimulates the tyrosine phosphorylation of phospholipase C␥, mitogen-activated protein kinase, phosphatidylinositol 3-kinase, FAK, and paxillin in human umbilical vein endothelial cells and of phospholipase C␥, p120 GAP , and NCK in bovine aortic endothelial cells (23)(24)(25)(26). However, the key targets that mediate the diverse biological functions of VEGF in endothelial cells remain incompletely understood for both VEGF receptors.
Attachment of cells to the extracellular matrix (ECM; e.g. laminin, fibronectin, and collagen) is mediated by structures called "focal adhesions," which connect the extracellular matrix with the plasma membrane and the underlying actin cytoskeletal network. Attachment of cells to the ECM results in the clustering of integrin receptors and initiates the recruitment of numerous cytoplasmic proteins to the focal adhesion complex. These proteins include both structural and catalytically active signaling proteins. Signaling through focal adhesions regulates a variety of cellular processes including cell growth, migration, and apoptosis (27). Focal adhesions are dynamic structures, and thus, their formation and breakdown are regulated by many different extracellular stimuli. Clustering of integrin subunits by ECM, growth factor stimulation (platelet-derived growth factor and epidermal growth factor), and signaling through certain G protein-coupled receptors (such as the receptor for lysophosphatidic acid) have been shown to result in focal adhesion formation. Formation of these complexes has been shown to require the activity of the small GTP-binding protein Rho (30,31).
Activation of tyrosine kinases is a prerequisite for focal adhesion assembly because inhibitors of tyrosine kinase activity block cell adhesion and focal adhesion formation (28). Clustering of integrins appears sufficient to recruit FAK and tensin and results in the autophosphorylation and activation of FAK (29). FAK is recruited to clustered integrins by a direct interaction of the N-terminal domain of FAK and the cytoplasmic domain of the ␤-integrin subunits. Subsequent tyrosine phosphorylation of FAK at Tyr 397 (proximal to the kinase domain) creates a binding site for Src, resulting in the formation of a complex consisting of the two tyrosine kinases. The formation of the FAK/Src complex results in the activation of Src and the subsequent activation of downstream signals (28,29). The C terminus of FAK contains binding sites for a number of signaling molecules including phosphatidylinositol 3-kinase, Crk-associated substrate, GRB2, paxillin, and the GTPase regulator associated with FAK. In some cells, the C-terminal domain of FAK is expressed as a separate protein called FRNK (FAKrelated non-kinase), whose overexpression results in inhibition of the rate of cell spreading and cell migration (30,31). These effects can be rescued by overexpression of FAK or Src, indicating that FRNK can act as a biological inhibitor of FAK signaling.
We have previously cloned and characterized the related adhesion focal tyrosine kinase (RAFTK), also known as Pyk2, CAK-␤, and CADTK. RAFTK is a nonreceptor tyrosine kinase that is related to FAK (32)(33)(34)(35)(36)(37)(38)(39). It is implicated in the regulation of ion channel activity, stress responses, cell adhesion/cytoskeletal reorganization, vesicle trafficking, and nerve growth factor signaling (40). RAFTK, which regulates various cellular functions, is emerging as a critical "platform kinase" upon which a number of signaling molecules integrate (41). RAFTK also associates with focal adhesion-like structures, participates in integrin-mediated signaling in megakaryocytes, and is tyrosinephosphorylated following ␤ 1 -integrin or B cell antigen receptor stimulation in B cells (42).
The assembly and disassembly of focal adhesions play a key role in the mechanism by which several extracellular stimuli regulate both cell morphology and movement. The assembly of focal adhesions in brain microvascular endothelial cells was reported (26,43). In this study, we explored the cell signaling proteins that couple VEGF binding to its receptors with focal adhesion assembly in HBMECs.

EXPERIMENTAL PROCEDURES
Materials-Human recombinant VEGF 165 and anti-human VEGF monoclonal antibody were obtained from Genentech, Inc. (San Fran-cisco, CA). Monoclonal antibodies for ␣ v ␤ 5 (P1F6) were used for integrin immunoprecipitation, and rabbit polyclonal anti-␤5 antibodies were used for immunoblotting. Monoclonal antibodies for p130 cas and paxillin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
Cell Culture-HBMECs were purchased from Cell Systems, Inc. (Kirkland, WA). The cells were seeded onto attachment factor-coated culture plates and maintained in CSC complete medium according to the protocol of the manufacturer. The HBMECs formed tubular-like networks on Matrigel and produced the endothelial-specific marker, von Willebrand factor, indicating that these cells maintained the general properties of endothelial cells. Interestingly, the in vitro resistance in both rat brain microvascular endothelial cells and human brain microvascular endothelial cells is very unique, and showed measurements of 1200 -1900 ohm cm 2 as compared with 10 -20 ohm cm 2 in peripheral endothelial cells, in accordance with previous reports (44). During the course of the experiment, the cells were routinely checked for expression of von Willebrand factor. The cells in this study were used until passages 4 -6. HBMECs were maintained in F-10 medium supplemented with 4% fetal bovine serum, 5% penicillin/streptomycin, 1% glutamine, 1% heparin, 0.7% endothelial mitogen, and 15% horse serum, and the cells were grown at 37°C with 5% CO 2 .
Western Blotting-The cells were lysed in kinase lysis buffer (New England Biolabs). The proteins were separated by SDS-PAGE under reducing conditions and transferred onto nitrocellulose membrane (Millipore, Boston, MA). The membranes were blocked with 5% bovine serum albumin in PBS and subsequently incubated with primary antibody for overnight incubation at 4°C. Bound antibodies were detected by horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence (Amersham Biosciences).
Focal Adhesion Assay-HBMECs were seeded onto the indicated substrate-coated coverslips: fibronectin (20 g/ml) and gelatin (0.2%). After 4 h of incubation in serum-free CSC medium, the cells were stimulated with VEGF (20 ng/ml) for the indicated times. The cells were then fixed in 4% paraformaldehyde for 10 min, permeabilized with 0.1% Triton X-100 in PBS for 5 min at room temperature, and treated with blocking buffer (0.1% goat serum and 0.1% bovine serum albumin in PBS). Polyclonal anti-FAK (1:500) antibodies or anti-RAFTK antibodies (1:1000) were generated in our lab as detailed previously (36,38,39) and were used for this study. Monoclonal antibody for vinculin (1:250) or control antibody was added, followed by incubation with goat antirabbit IgG conjugated to fluorescein (1:200). For the labeling of F-actin, Texas Red-phalloidin (Molecular Probes, Eugene, OR) (1:1000) was used. The cells were then washed again and mounted in anti-fading compound (phenylenediamine; 1 mg/ml, 50% glycerol with PBS). The fluorescent signals were detected by optical sectioning using a Leica TCS-NT confocal laser scanning microscope. A minimum of 400 cells/ condition were evaluated for the presence of focal adhesions. The cells that are positive usually have ϳ15-30 plaques/cell. The cells with less than 5-7 plaques/cell were scored as negative. The experiments were repeated a minimum of three times.
Cell Migration Assay and Cell Spreading-The migration assay utilized a modified Boyden chamber and was performed essentially as previously described. The cells were grown to 50% confluency and then either untreated or infected with adenoviral constructs for 2 h as described below. The cells were lifted from the dish using nonenzymatic cell dissociation solution (Sigma), centrifuged for 2 min at 1000 rpm, and resuspended in medium containing 0.5% bovine serum albumin. The cells were then plated on fibronectin-precoated Transwells (Costar Corp.) at a density of 10 5 cells/well. Medium with VEGF (10 ng/ml) was used as a chemoattractant in the lower wells. Cell migration assays were performed for 8 h at 37°C. The cells that migrated through the Transwells were stained with chloromethylfluorescein diacetate dye (Sigma) and counted under a fluorescent microscope. Four different fields were counted for each experiment, and all of the samples were performed in triplicate. For cell spreading, the cells were detached with 1 mM EDTA and washed with PBS, and then 1 ϫ 10 5 cells were plated on fibronectin-coated coverslips in serum-free medium. After the indicated times, the coverslips were fixed, stained, and analyzed by fluorescent microscopy. For quantification of the spreading cells, 150 -200 cells were scored from each of three independent experiments.
Adenovirus Constructs and Infections-Replication-deficient recombinant adenoviruses mediating overexpression have been previously described (45). RAFTK/Pyk2 cDNA fragments encoding both wild type (Pyk2/RAFTK-WT) and a phosphorylation-deficient mutant (Pyk2/ RAFTK-Y 402 ) were used to create replication-defective recombinant adenoviruses with the AdEasy system (Microbix Biosystem Inc.) as directed by the manufacturer. The Tyr 402 residue of RAFTK is the autophosphorylation site that also binds and activates Src. Mutation in this residue blocks autophosphorylation of RAFTK and its binding to Src, leading to the abrogation of Src phosphorylation. Subconfluent HBMEC cultures were infected with recombinant adenoviruses, including ␤-galactosidase expressing adenovirus as a control, for 2 h at a multiplicity of infection of 10, and then the medium was aspirated and replaced with maintenance medium. The lysates for the biochemical analyses were prepared ϳ24 -48 h after infection.
A replication-defective adenovirus encoding GFP-tagged FRNK was also constructed as previously described (46). An adenovirus expressing GFP alone was used to control for the nonspecific effects of adenoviral infection. HBMECs were infected at a multiplicity of infection of 10 for 2 h at 25°C with gentle agitation with replication-defective adenoviruses diluted in medium. The medium was then replaced with virusfree medium, and the cells were cultured for an additional 48 h before assaying for cell migration and focal adhesion assembly.
Data Analysis-The results are expressed as the means Ϯ S.D. Differences among the means were considered significant at p Ͻ 0.05. The data were analyzed using the SigmaStat statistical software package, version 1.0 (Jandel Scientific, San Rafael, CA).

RESULTS
HBMECs as a Model System-We chose to carry out these studies in HBMECs because of their direct application as a model system for the in vivo system of the human BBB and for their high relevance to the studies of ischemia and stroke in humans. HBMECs formed tubular-like networks on Matrigel and had the ability to uptake acetylated low density lipoprotein (47,49) and to produce von Willebrand factor, indicating that these cells maintained the general properties of endothelial cells. HBMECs also produced ␥-glutamyl transpeptidase, a brain endothelial marker. HBMECs maintained the unique characteristic of resistance greater than 1000 -2000 ohm/cm 2 as compared with the 10 -20 ohm/cm 2 characteristic of peripheral endothelial cells. HBMECs constitute the major component of the BBB and are therefore critical in maintaining its structural and functional integrity.
VEGF Induced the Formation of Focal Adhesions-Cell adhesion to ECM is a highly dynamic process involving structures heterogeneous with respect to size, composition, and orientation to actin filaments (50 -52). The largest and tightest structures are usually referred to as focal adhesions, which link the actin cytoskeleton to ECM by integrin receptor complexes. Actin-binding proteins that co-localize with integrins in focal adhesions include actin, talin, vinculin, and tensin. Focal contacts were detected by staining the fixed cells either with the monoclonal antibody to vinculin, a cytoskeletal protein that localizes at focal adhesion contacts (52-55), or with FAK monoclonal antibody (54). To analyze whether VEGF can induce focal adhesion formation in brain microvascular endothelial cells and to determine the specific role of integrins in VEGF-mediated focal adhesion assembly, HBMECs were grown overnight on coverslips coated with fibronectin (20 g/ml) (integrin-dependent) or gelatin (0.2%) (integrin-independent). After 4 h of incubation in serum-free CSC medium, the cells were stimulated with VEGF (20 ng/ml) for 1 or 4 h as indicated. The cells were then fixed and immunostained for FAK-and RAFTK-specific antibodies. Actin stress fibers were also analyzed using Texas Red-phalloidin (Molecular Probes). Clear actin stress fibers that cross the cell body and condense at the cell periphery were observed in HBMECs (Fig. 1, red). No focal adhesion sites were observed when HBMECs were seeded on gelatin (0.2%). Upon VEGF stimulation for 1 h, we observed the dynamic redistribution of FAK with actin in the outer membrane, which was increased upon 4 h of stimulation of the HBMECs with VEGF. When HBMECs were seeded on fibronectin, upon 1 h of treatment with VEGF, a significant co-redistribution of FAK with actin to the outer side of the cells was enhanced (Fig. 1B). After 4 h of stimulation with VEGF, there was a significant increase in the dynamic co-redistribution of FAK and actin with the formation of focal contacts. Similarly, focal contacts immunostained with RAFTK/Pyk2 were detected after their seeding on fibronectin and stimulation with VEGF (Fig. 1C). These results suggest that VEGF plays a role in focal adhesion assembly in HBMECs and that FAK and RAFTK/Pyk2 are important components of these focal adhesion sites.
To further analyze the effects of VEGF on focal adhesion assembly, HBMECs were seeded on fibronectin and were treated with VEGF for 1 and 4 h. Focal adhesion assays were performed and quantitated. The cells were examined for the presence of focal adhesions by sectioning using a Leica TCS-NT confocal laser scanning microscope. A minimum of 400 cells/ condition were evaluated for the presence of focal adhesions. The cells that are positive usually have ϳ15-30 plaques/cell. The cells with less than 5-7 plaques/cell were scored as negative. The experiments were repeated a minimum of three times. As shown in Fig. 2A, VEGF mediated focal adhesion assembly, which peaked at 4 h. To determine whether VEGF signaling is essential for stimulation of focal adhesion assembly, HBMECs were pretreated with antibodies to VEGF or with the inhibitor SU-1498, which inhibits Flk-1/KDR receptor tyrosine phosphorylation and VEGF-mediated downstream effects. As shown in Fig. 2A, preincubation of cells with VEGF monoclonal antibody or with the inhibitor SU-1498 blocked the ability of VEGF to stimulate focal adhesion assembly.
The Effects of Integrins and VEGF Signaling on Focal Adhesion Formation-Next, we examined the ability of HBMECs to assemble focal adhesions when plated on integrin-dependent (e.g. laminin and fibronectin) or integrin-independent (e.g. gelatin) substrates. In the presence of VEGF, HBMECs plated on laminin or fibronectin showed increased focal adhesion sites as compared with control cells (Fig. 2B). In contrast, these changes were significantly reduced in cells plated on gelatin (Fig. 2B) or cells plated on poly-L-lysine (data not shown).
Tyrosine Phosphorylation of FAK and RAFTK/Pyk2 in HB-MECs upon VEGF Stimulation-HBMECs expressed the Flk-1/KDR receptor (Fig. 3). Upon VEGF stimulation, the phosphorylation of Flk-1/KDR was increased (Fig. 3). Both RAFTK/ Pyk2 and FAK were expressed in HBMECs and were tyrosinephosphorylated in a time-dependent manner following VEGF stimulation (Fig. 4, A and B). Interestingly, p130 cas and paxillin were also tyrosine-phosphorylated upon VEGF stimulation (Fig. 4, C and D). Our data indicate that phosphorylation of these VEGF receptors leads to their activation, which may mediate VEGF signaling in HBMECs, and that VEGF-mediated signaling in these cells involves RAFTK, FAK, and the cytoskeletal components p130 cas and paxillin.
VEGF Induced the Formation of a FAK/␣ v ␤ 5 Complex in HBMECs-Ligation of integrin ␣ v ␤ 5 has been shown to be essential for VEGF-induced angiogenesis (19). To elucidate the potential recruitment of FAK to ␣ v ␤ 5 integrin upon VEGF stimulation, lysates of VEGF-stimulated HBMECs were subjected to immunoprecipitation with anti-integrin antibodies. These immunoprecipitates were analyzed by SDS-PAGE and Western blotting for the presence of FAK. As shown in Fig. 4E, VEGF induced a FAK/␣ v ␤ 5 complex in HBMECs in a time-dependent manner and was associated with increased FAK phosphorylation. The expression level of ␤ 5 -integrin is shown in the lower panel using anti-␤5 antibodies.
Effects of Overexpressing RAFTK/Pyk2-WT and Mutant on HBMEC Spreading and Migration-Because RAFTK/Pyk2 is activated in HBMECs upon VEGF stimulation, the effects of increased RAFTK/Pyk2 signaling in HBMECs were examined using recombinant adenoviruses expressing either RAFTK/ Pyk2 WT or a phosphorylation-deficient mutant (RAFTK/Pyk2-Y 402 ) (45). HBMECs were infected with the catalytically inactive RAFTK mutant Tyr 402 and RAFTK wild type, as well as with control adenoviral vector for 2 h. The cells were washed, and the medium was then replaced with virus-free medium. The cells were cultured for an additional 24 -48 h and assayed accordingly. RAFTK WT increased the spreading and migration of HBMECs, whereas the catalytically inactive mutant Tyr 402 markedly suppressed HBMEC spreading (ϳ70%), adhesion (82%, not shown), and migration (ϳ65%) (Fig. 5). This suggests that the physiological function of RAFTK/Pyk2 kinase activity is important in regulating HBMEC motility and migration.
Effects of GFP-FRNK on Focal Adhesion Assembly and Migration-To specifically determine the effects of FAK on focal adhesion assembly and to target FAK-dependent signaling in HBMECs, we employed a replication-defective adenovirus encoding FRNK (46). As seen in Fig. 6, overexpression of FRNK significantly inhibited HBMEC migration and disrupted focal adhesions, whereas the control adenovirus had no effect. The endogenous FAK levels were similar in both the GFP-and GFP-FRNK-overexpressing HBMECs (data not shown).

DISCUSSION
The BBB is involved in the maintenance and protection of the microenvironment of the central nervous system. Although the BBB is immensely important in maintaining brain function, very little is known about its function at the cellular and molecular level. Endothelium of the cerebral blood microvessels, which constitutes the major component of the BBB, controls leukocyte and metastatic cancer cell adhesion and trafficking into the brain parenchyma. Brain microvascular endothelial cells are critical in maintaining the structural and functional integrity of the BBB.
During and after cerebral ischemia, brain vasculature becomes leaky and unstable, and the normally impermeable blood-brain barrier breaks down. Several days after the ische-  4. A and B, FAK, RAFTK/Pyk2 kinases are tyrosine-phosphorylated upon VEGF stimulation of HBMECs. 500 g of total cell lysates obtained from VEGF-treated HBMECs for the indicated times were immunoprecipitated (IP) with either RAFTK (A) or FAK (B) antibodies. The blots were probed with anti-Py-20 or 4G10. The blots were then stripped and reprobed with anti-RAFTK or anti-FAK antibodies. C and D, VEGF stimulated the tyrosine phosphorylation of p130 cas (C) and paxillin (D) in HBMECs. 500 g of total cell lysates were obtained from VEGF-treated HBMECs and were immunoprecipitated with either p130 cas or paxillin antibodies. The blots were probed with 4G10 (phosphotyrosine antibodies), p130 cas , or paxillin antibodies as indicated. E, VEGF induced the assembly of FAK with ␣ v ␤ 5 integrin in HBMECs. 1 mg of total cell lysates was obtained from VEGF-treated HBMECs and immunoprecipitated with anti-␣ v ␤ 5 or control antibody. The immunocomplexes were immunoblotted with anti-FAK or ␤5 antibody as indicated. IgG, control antibody. WB, Western blot. mic insult, endothelial cells begin to proliferate, and angiogenesis occurs. Expression studies have shown that key vascular growth factors, such as VEGF, are regulated during these processes in a complex and coordinated manner. VEGF may play a role in the initial vascular destabilization and subsequent angiogenesis that occur after cerebral ischemia. Indeed, VEGF not only mediated the proliferation of vascular smooth muscle and endothelial cells but also regulated vascular differentiation, regression, migration, and permeability (6 -20). However, the role of VEGF in regulating brain microvascular endothelial cell migration and permeability had not been well elucidated.
The ability of cells to form cell contacts, adhere to the extracellular matrix, change morphology, and migrate is essential for development, wound healing, metastasis, cell survival, and the immune response. These events depend on the binding of integrins to the extracellular matrix and on the assembly of focal adhesions, which are complexes comprised of scaffolding and signaling proteins organized by adhesion to the extracellular matrix. In this study, we found that VEGF regulated the assembly of focal adhesions, as well as the spreading and migration of HBMECs. VEGF induced the formation of focal adhesions in HBMECs when seeded on fibronectin, and pretreatment of cells with VEGF antibodies or with the specific inhibitor SU-1498, which inhibits Flk-1/KDR receptor phosphorylation, blocked the ability of VEGF to stimulate focal adhesion assembly. Furthermore, HBMECs plated on laminin or fibronectin showed increased focal adhesion sites as compared with control cells or cells plated on gelatin.
Adhesive interactions with ECM components play a critical role in regulating intracellular signaling pathways that control cell growth, survival, and differentiation. The integrin family of transmembrane cell surface receptors mediates cell contact with ECM and is responsible for initiating the formation of focal adhesion structures that tether the integrin cytoplasmic tail with the actin cytoskeleton. The identification and characterization of FAK and RAFTK/Pyk2 have provided insights into adhesion signaling. Clustering of integrins leads to the recruitment of FAK and/or RAFTK/Pyk2 to the newly formed focal adhesion sites, activation of FAK-Src and/or RAFTK/Pyk2-Src complexes, and the phosphorylation of a variety of downstream effects (Fig. 7). We found that HBMECs expressed both FAK and RAFTK/Pyk2 kinases, although the expression of FAK was much more abundant than that of RAFTK/Pyk2. Both kinases are important components of these focal adhesion sites, because abrogation of their expression and/or activation significantly inhibited HBMEC focal adhesion assembly, migration, and adhesion.
Our results describe a signal transduction pathway that couples VEGF binding to its receptor with focal adhesion formation and cell migration. VEGF binding to the Flt-1 and Flk-1/KDR receptors induces the NCK, FAK, and RAFTK/Pyk2 complex with subsequent tyrosine phosphorylation of these proteins. This complex is then recruited to focal adhesions, which is followed by the enhancement of cell migration. Overexpression of RAFTK wild type increased the spreading and migration of HBMECs, whereas overexpression of the mutant Y402F markedly suppressed HBMEC spreading (ϳ70%), adhesion (ϳ82%), and migration (ϳ65%). Overexpression of FRNK, which inhibits FAK, also inhibited cell migration and focal adhesion assembly. These results demonstrate the physiological function of FAK and RAFTK/Pyk2 in regulating HBMEC focal adhesion and migration. Interestingly, we observed that FRNK overexpression inhibited the VEGF-induced phosphorylation of RAFTK and FAK (data not shown). These results are in agreement with other published reports (30,46,56,57) showing that FRNK inhibited both RAFTK and FAK phosphorylation in cardiomyocytes and vascular smooth muscle cells. FRNK may interfere with the activation of both FAK and RAFTK and displace both kinases from focal adhesions. Richardson and Parsons (30) showed that FRNK overexpression in chicken embryonic cells inhibited the tyrosine phosphorylation of FAK, tensin, and paxillin. FRNK overexpression delayed focal adhesion formation, but vinculin-positive focal adhesions were still present, and there was no decrease in total FAK, paxillin, or tensin levels. These results suggest that tyrosine phosphorylation of FAK is not required for the maintenance of focal adhesion structure in some cell types. However, embryonic cells derived from FAK-null mice formed dense focal adhesions that demonstrated impaired cell spreading and migration (58), indicating a functional role of FAK in focal adhesions. The structural motifs within the N-terminal region of FAK and RAFTK/Pyk2 may stabilize FAK and RAFTK/Pyk2 associations with integrins within focal adhesions (59). Interestingly, the N-terminal region of FAK bears substantial sequence homology to the filamentous actin-capping proteins ezrin, radixin, and moesin (56).
Localization of RAFTK and FAK in focal adhesions may be required for their activation in response to VEGF. The apparent requirement of both kinases in the focal adhesion assembly of HBMECs may reflect the need to activate signaling pathways downstream of FAK and RAFTK/Pyk2 in concert with pathways that directly induce angiogenesis and the migration of HBMECs. The similarity in function between both RAFTK/ Pyk2 and FAK in this study suggests a close interaction between these kinase family members and the cytoskeleton. Both kinases regulate signaling pathways involved in VEGF-induced focal adhesion assembly in HBMECs.
Recent studies indicate that VEGF promotes integrin-dependent cell biological responses in vivo and in vitro (20, 60 -63), suggesting that the coordination of inputs from the extracellular matrix and growth factors are physiologically important. This study indicates that VEGF induced FAK phosphorylation, leading to the formation of a complex between FAK and ␣ v ␤ 5 in HBMECs. Interestingly, similar results showing that Src mediated the coupling of FAK to integrin ␣ v ␤ 5 in VEGF signaling were reported recently (61). In addition, p130 cas and paxillin were also tyrosine-phosphorylated upon VEGF stimulation of HBMECs. Both p130 cas and paxillin have been previously shown to be downstream of FAK (36,37,38). Thus, activation of p130 cas and paxillin through FAK may coordinate growth factor-dependent integrin signaling during VEGF-mediated focal adhesion assembly. Our recently published studies (47)(48) have shown that VEGF treatment of HBMECs induced changes in cell permeability in a time-dependent manner. These observed changes in HBMEC permeability are characteristic of BBB breakdown. Taken together, the VEGF-mediated effects on focal adhesion assembly and the role of VEGF in modulating HBMEC permeability indicate that it is an important regulator of brain microvascular function and integrity. Future studies will determine the role of Src in mediating the coupling of FAK to ␣ v ␤ 5 integrin in VEGF signaling in HBMECs leading to focal adhesion assembly, as well as the role of FAK and RAFTK/Pyk2 during focal cerebral ischemia and in the breakdown of the BBB following brain injury.