Mechanical stretch decreases FAK phosphorylation and reduces cell migration through loss of JIP3-induced JNK phosphorylation in airway epithelial cells

Abstract

JNK is a nonreceptor kinase involved in the early events that signal cell migration after injury. However, the linkage to early signals required to initiate the migration response to JNK has not been defined in airway epithelial cells, which exist in an environment subjected to cyclic mechanical strain (MS). The present studies demonstrate that the JNK/stress-activated protein kinase-associated protein 1 (JSAP1; also termed JNK-interacting protein 3, JIP3), a scaffold factor for MAPK cascades that links JNK activation to focal adhesion kinase (FAK), are both associated and activated following mechanical injury in 16HBE14o− human airway epithelial cells and that both FAK and JIP3 phosphorylation seen after injury are decreased in cells subjected to cyclic MS. Overexpression of either wild-type (WT)-FAK or WT-JIP3 enhanced phosphorylation and kinase activation of JNK and reduced the inhibitory effect of cyclic MS. These results suggest that cyclic MS impairs signaling of cell migration after injury via a pathway that involves FAK-JIP3-JNK.

migration of cells into an injury site is a primary component of repair after injury (13). Epithelial cells in airways have an important role both in defining the physical barrier between the host and external environment and in regulating the response to inflammation in several airways diseases. Airway epithelial cells migrate in response to injury along a basement membrane containing several ECM proteins using cell surface receptors in the integrin family (27, 45). Binding of ECM proteins such as collagen IV, laminin, or fibronectin to appropriate integrins signals to several key pathways through focal adhesion kinase (FAK), a nonreceptor tyrosine kinase residing within the focal adhesion complex (reviewed in Refs. 8, 31). Clustering of integrin receptors at the cell membrane leads to phosphorylation of FAK at Tyr³⁹⁷ with subsequent binding of Src and second-step phosphorylation of FAK at Y576/577 (reviewed in Refs. 8, 16). Overexpression of either ECM proteins or of FAK can promote migration (20), whereas overexpression of negative regulators such as FAK-related nonkinase (FRNK) can inhibit both adhesion and migration (11).

FAK mediates cell migration by promoting membrane protrusion and focal adhesion turnover utilizing several signaling pathways (16, 31), including that mediated by JNK. JNK activation is a key event required in migration of both airway epithelial cells (45) and several other epithelial and nonepithelial cell types (17, 48). JNK may regulate migration via phosphorylation of paxillin in the focal adhesion complex (18) by regulating microtubule assembly by phosphorylation of the microtubule-associated protein 2 (4, 46) and potentially by phosphorylation of c-jun (45), a partner in the activator protein-1 (AP-1) family of transcription factors (48).

Activation of JNK typically occurs as part of a MAPK cascade, a conserved signaling pathway (43) that links diverse external stimuli to nuclear responses (41). However, activation of JNK outside this cascade may occur. The pathway from FAK to JNK includes p130^Cas and Rac1 (23, 37, 39), but the precise mechanism by which JNK is recruited and activated is not clear. One family of JNK binding proteins, the JNK-interacting proteins (JIP), have been suggested as a scaffold factor for MAPK signaling cascades (37, 38). One such JIP, JIP3, also termed JNK/stress-activated protein 1 (JSAP1), binds to several MAPK pathway constituents, including JNK isoforms, and interacts with FAK (37). The JIP3-FAK complex can regulate JNK activation, which can be regulated by Src activation or by cell attachment to fibronectin (38).

Airway epithelial cells exist in an environment that includes cyclic stretch and compression both as a function of respiration and as a result of changes in bronchomotor tone (26, 29). Cyclic mechanical strain (MS) inhibits epithelial cell migration (29, 30), as does inhibition of JNK and other MAPK signaling pathways (45). However, the role of signaling pathways linking focal adhesions to JNK in migration under conditions of cyclic MS is not well-understood. Here, we provide evidence that MS caused an initial increase in FAK phosphorylation that was followed by a sustained decrease in both FAK and JIP3 phosphorylation that contributed to decreased cell migration. Overexpression of either FAK or JIP3 overcomes the inhibition induced by MS. These results suggest that JIP3 and FAK are critical to JNK activation early after injury and that both are regulated by MS in airway epithelial cells.

MATERIALS AND METHODS

Cell culture.

16HBE14o− cells (provided by D. Gruenert, California Pacific Medical Center, San Francisco, CA) were derived from normal human airway epithelial cells using transformation by simian virus 40 (SV40), and these cells maintain the capability to form polarized monolayers and tight junctions and retain other characteristics of differentiated native epithelium (reviewed in Ref. 15). These cells were grown on either collagen IV-coated Silastic membranes (Flexcell International, Hillsborough, NC) or on collagen IV-coated membranes on which laminin-5-rich matrix was deposited by 804G cells (rat bladder carcinoma cells known to lay down a laminin-5 matrix in culture; Ref. 42). 804G cells were grown to confluence for 3 days in MEM and then removed using ammonium hydroxide (11). 16HBE14o− cells have been demonstrated previously to withstand high levels of MS (44). Cyclic strain of 20% with a frequency of 30 cycles/min was applied using a Flexercell FX-4000T tension unit.

Immunoprecipitations and immunoblotting.

16HBE14o− cells grown on either collagen IV or laminin-5 matrix were infected with adenoviruses (Ad) expressing enhanced green fluorescent protein (EGFP; Ref. 9) or Cre-lox (34) as a control or wild-type (WT)-FAK (GFP-FAK; Ref. 3), FRNK (9), FLAG-JIP3 (34), FLAG-ΔJBD (35), or JNK1 or JNK2 [Ref. 34; multiplicity of infection (MOI) of 6–8] as described previously (9). Forty-eight hours later, confluent monolayers were multiply wounded so as to generate a large number of wound edges for each well and exposed to cyclic strain for 2 and 6 h. Cells were lysed with RIPA buffer (150 mM NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Triton X-100, and 50 mM Tris; pH 7.2) containing 1 mM sodium vanadate, 1 mM sodium fluoride, 1 mM PMSF, 5 μg/ml aprotinin, and 5 μg/ml leupeptin at 4°C. The cell lysates (800 μg) were subjected to immunoprecipitation with FAK (BD Transduction Laboratories) and FLAG (Sigma-Aldrich, St. Louis, MO) antibodies, and simultaneously equal amounts of proteins (50 μg) were subjected to electrophoresis. The immunoblotting of the SDS-PAGE-separated protein complex and total protein was done using appropriate antibodies. JIP3, JNK1, JNK2, and GAPDH antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). For immunoblotting experiments to detect phosphorylated (p-) FAK and JNK, 50 μg of the RIPA cell lysates were loaded on gels for SDS-PAGE, transferred onto nitrocellulose membranes, and probed with anti-p-FAK (Tyr³⁹⁷; BioSource, Invitrogen, Carlsbad, CA) and anti-p-JNK1 (Abcam, Cambridge, MA). For detection of expression of WT-GFP-FAK and EGFP-FRNK, blots were probed with GFP antibody (Santa Cruz Biotechnology), whereas JIP3, JNK1, and JNK2 expression was detected using FLAG, hemagglutinin (HA; Covance, Princeton, NJ), and His (Qiagen, Valencia, CA) antibodies, respectively. Secondary antibodies used were either anti-rabbit horseradish peroxidase (HRP) (Amersham Biosciences, Pittsburgh, PA) or anti-mouse HRP (Jackson ImmunoResearch, West Grove, PA).

JNK kinase activation assay.

A nonradioactive JNK assay kit from Cell Signaling Technology was used. Briefly, a c-Jun fusion protein linked to agarose beads was used to pull down JNK enzyme from cell extracts. On addition of kinase buffer and ATP, JNK phosphorylated the c-Jun substrate. Phospho-c-Jun (Ser⁶³) antibody was then used to measure JNK activity by immunoblotting.

Migration.

In all cell migration studies, multiple wounds were scraped through confluent monolayers grown on either collagen IV or laminin-5 matrix using a fine-toothed comb. Approximately seventeen wounds of ∼475 μm were generated in each well, and migration followed for 9 h in serum-free medium as described previously (9). Wound widths were determined in three wounds from each well and averaged for each individual measurement, and measurements were made in one or two independent wells from at least three experiments (n = 3–6).

Statistical analysis.

All values are presented as means ± SE. Statistical analysis was performed using the SigmaStat statistical package (version 3.5; Jandel Scientific, San Rafael, CA) or StatView (version 5.0.1; SAS Institute, Cary, NC). One-way ANOVA was performed for comparisons of multiple treatments for cells grown on a single matrix followed by the Scheffé test to determine individual differences. Two-way ANOVA was performed for comparisons of multiple treatments for cells grown on both matrix proteins followed by the Holm-Sidak method to determine individual differences. Significant differences were determined based on a threshold of P < 0.05.

RESULTS

Cyclic MS attenuates cell migration.

To investigate the role of laminin matrix and the link between signaling from the engaged matrix-receptor-FAK pathway to migration, we compared migration in 16HBE14o− cells grown on either collagen IV or laminin-5 matrix (deposited on collagen IV-coated membranes) with and without cyclic MS (Fig. 1). Cyclic MS slowed migration substantially, and under both static conditions and with cyclic MS, cells grown on laminin-5 migrated more quickly in response to injury compared with cells grown on collagen.

Fig. 1.

Cyclic mechanical strain (MS) attenuates migration of airway epithelial cells, and migration was slower on collagen IV compared with laminin-5. Scrape-wounded human airway epithelial 16HBE14o− cells grown on collagen IV or laminin-5 matrix were exposed to static conditions or cyclic MS for 9 h. Multiple wounds were created in each well, and measurements were made on 3 wounds per well and averaged. A shows representative images of wounds at 0, 3, and 9 h, and B summarizes wound width measurements. Data are expressed as means ± SE, n = 3 independent experiments. *P < 0.05 for laminin static vs. collagen static; #P < 0.05 for MS vs. static conditions on the same matrix.

Cyclic MS has a biphasic effect on FAK phosphorylation.

We then compared phosphorylated FAK levels in 16HBE14o− cells following scrape wounding when cells were grown on either laminin-5 or collagen IV matrix (Fig. 2A). Under static conditions and with either matrix, FAK phosphorylation was transiently elevated ∼4-fold within 30 min of injury, and then the levels subsequently declined (Fig. 2B). When MS was applied, FAK phosphorylation was again transiently elevated by 30 min and increased relative to static cells, but levels were significantly decreased by 6 h compared with static conditions. When we expressed the kinase-inactive form of FAK, FRNK, we blocked the transient increase in FAK phosphorylation at 30 min (Fig. 2C).

Fig. 2.

Cyclic MS modulates focal adhesion kinase (FAK) phosphorylation with a biphasic response. Scrape-wounded 16HBE14o− cells on collagen IV or laminin matrix were exposed to static conditions or cyclic MS for 0.5, 2, or 6 h and then lysed. Representative Western blots probed with phosphorylated (p-) FAK (Tyr³⁹⁷) and FAK antibodies at 2 and 6 h (A); the line graph (B) summarizes the data expressed as means ± SE, n = 3 independent experiments. *P < 0.05 vs. time 0 for each condition; +P < 0.05 for the effect of cyclic MS vs. static. C shows that adenoviral expression of FAK-related nonkinase (FRNK) prevented the transient increase in p-FAK 30 min after wounding.

Overexpression of WT-FAK caused a sustained increase in levels of phosphorylated FAK, whereas overexpression of FRNK attenuated phosphorylation, in 16HBE14o− cells grown on either collagen IV or laminin-5 matrix 2 h (Fig. 3A) or 6 h (Fig. 3B) after injury. (Expression of Ad-WT-FAK resulted in expression of 2 bands because of the additional GFP.) Overexpression of WT-FAK also accelerated migration after injury on either collagen IV (Fig. 3C) or laminin-5 (Fig. 3D) compared with the Ad-EGFP (control Ad-infected) and WT (non-Ad-infected) control cells. As with FAK phosphorylation, overexpressing FRNK attenuated migration in cells grown on either matrix.

Fig. 3.

Cyclic MS inhibits FAK phosphorylation and migration of airway epithelial cells. Scrape-wounded 16HBE14o− cells on collagen IV or laminin matrix, infected with adenovirus (Ad) expressing enhanced green fluorescent protein (EGFP), wild-type (WT)-FAK, or FRNK, were exposed to static and cyclic MS for 2 and 6 h and then lysed. Representative Western blots were probed with p-FAK (Tyr³⁹⁷), FAK, and GAPDH antibodies at 2 h (A) and 6 h (B); the bar graphs summarize the data expressed as means ± SE for each, n = 3 independent experiments. *P < 0.05 vs. static on collagen; #P < 0.05 vs. static on collagen and laminin; +P < 0.05 vs. FRNK on either collagen and laminin. Wound closure of scrape-wounded 16HBE14o− cells expressing WT-FAK and FRNK monitored for 9 h on collagen IV (C) and laminin-5 (D) matrix measurements were made in 1 or 2 wells from 3 independent experiments (n = 3–6). *P < 0.05 for WT-FAK vs. EGFP under either static conditions or conditions of cyclic MS; #P < 0.05 for FRNK vs. EGFP under either static conditions or conditions of cyclic MS; +P < 0.05 for WT-FAK under static conditions vs. cyclic MS. Measurements in WT cells at the same time were indistinguishable from EGFP cells under the same conditions and are omitted from the graph for clarity.

Cyclic MS inhibits association of FAK with JIP3.

We then examined the interaction of FAK and JIP3 in the 16HBE14o− cell line under conditions of cyclic MS. Cells grown on laminin-5 matrix demonstrated enhanced interaction of FAK with JIP3 compared with those grown on collagen IV (Fig. 4A), an interaction that was attenuated by application of cyclic MS. To examine the role of JIP3 on migration, 16HBE14o− cells were infected with Ad expressing either JIP3 or ΔJBD, a truncated form of JIP3 that inhibits JNK (35). JIP3 expression significantly enhanced, whereas ΔJBD expression decreased, cell migration under conditions of cyclic MS (Fig. 4B), compared with the Ad-Cre-lox (control Ad-infected) and WT (non-Ad-infected) control cells.

Fig. 4.

JNK-interacting protein 3 (JIP3) forms a scaffold with FAK and accelerates migration of airway epithelial cells subjected to cyclic MS. A: cell lysates of scrape-wounded 16HBE14o− monolayers on collagen IV (C) or laminin (L) matrix, exposed to static and cyclic MS for 2 or 6 h, were either immunoprecipitated with FAK antibody or equal amounts of proteins were subjected to electrophoresis. Representative Western blots of immunoprecipitation (IP) and equal proteins were probed with JIP3 antibody. The bar graph summarizes the densitometry data expressed as means ± SE, n = 4 independent experiments. *P < 0.05 for cyclic MS on either matrix vs. corresponding static condition; #P < 0.05 for static on laminin vs. static on collagen. B: wound closure of scrape-wounded 16HBE14o− cells on laminin matrix expressing JIP3 and ΔJBD, a truncated form of JIP3 that inhibits JNK, monitored for 9 h. Data are expressed as means ± SE, n = 3 independent experiments (n = 6). *P < 0.05 for ΔJBD vs. Cre-lox in each mechanical stress condition; #P < 0.05 for each group in cyclic MS vs. corresponding static group. Measurements in WT cells at the same time were indistinguishable from Cre-lox cells under the same conditions and are omitted from the graph for clarity. C and D: cell lysates of scrape-wounded 16HBE14o− monolayers on collagen IV and/or laminin matrix, expressing only WT-FAK or FRNK (C) or coexpressing FLAG-JIP3 (D), were immunoprecipitated with FAK or FLAG antibody, respectively, and equal amounts of proteins were subjected to electrophoresis. Representative Western blots were probed with JIP3 or EGFP antibodies (C) or phosphotyrosine or FLAG antibodies. For equal protein determinations, immunoblots (IB) were probed with GAPDH antibody. n = 3 Independent experiments for (C) and (D).

Previous studies demonstrated that JIP3 coprecipitated with FAK phosphorylated at Tyr³⁹⁷ in HeLa cells, since cells expressing FAK mutated at Tyr³⁹⁷ (Y397F-FAK) did not coprecipitate (38). Because we found that overexpression of WT-FAK increased FAK Tyr³⁹⁷ phosphorylation, we tested whether this would increase interactions with JIP3. As shown in Fig. 4C, expression of WT-FAK increased the association with JIP3, whereas expression of FRNK had no effect. Expression of WT-FAK and FRNK was confirmed by probing blots with EGFP antibody.

We then investigated whether JIP3 was phosphorylated by FAK. Coexpression of WT-FAK and JIP3 in 16HBE14o− cells enhanced tyrosine phosphorylation of JIP3 (Fig. 4D). JIP3 overexpression was confirmed by probing blots with a FLAG antibody. These results suggested that phosphorylation of Tyr³⁹⁷ on FAK facilitated the association of FAK with JIP3 and consequently enhanced the tyrosine phosphorylation of JIP3.

Overexpression of WT-FAK and JIP3 accelerates migration by increasing phosphorylation and kinase activation of JNK.

We hypothesized that phosphorylated FAK-JIP3 scaffold activates JNK leading to accelerated migration. Coexpression of WT-FAK and JIP3 in 16HBE14o− cells was sufficient to enhance JNK1 phosphorylation (Fig. 5A) and kinase activation (Fig. 5B) over 2 or 6 h to a similar degree under both static and cyclic MS conditions. Coexpressing JIP3 and FRNK did not elicit significant JNK1 phosphorylation or activation, as expected. When we diminished interactions between FAK and JIP3 by coexpressing FRNK and JIP3, cell migration was significantly inhibited under both static and MS conditions (Fig. 5C) compared with the Ad-Cre-lox (control Ad-infected) and WT (non-Ad-infected) control cells. Conversely, cell migration was significantly enhanced by coexpression of WT-FAK and JIP3.

Fig. 5.

Overexpression of WT-FAK and JIP3 induces activation of JNK1. Scrape-wounded 16HBE14o− monolayers on laminin matrix, coexpressing either WT-FAK or FRNK, and JIP3, were exposed to static conditions or cyclic mechanical stretch for 2 or 6 h, after which cells were lysed. Equal amounts of proteins were subjected to either electrophoresis or immunoprecipitation using the JNK kinase activation assay kit. Representative Western blots were probed for phosphorylated JNK1 (A) or phosphorylated c-Jun (Ser⁶³) detection using the JNK kinase activity assay kit protocol (B). For total proteins, blots were probed with either JNK1 or c-Jun antibodies, respectively. The bar graph summarizes the densitometry data expressed as means ± SE, n = 4 independent experiments. *P < 0.05 for cyclic MS vs. corresponding static condition at 2 and 6 h; #P < 0.05 vs. corresponding WT and corresponding FRNK/JIP3 at 2 and 6 h. C: wound closure of scrape-wounded 16HBE14o− cells on laminin matrix coexpressing WT-FAK and JIP3 or FRNK and JIP3 monitored for 9 h. Data are expressed as means ± SE, n = 3 independent experiments. *P < 0.05 for static FAK/JIP3 or FRNK/JIP3 vs. static Cre-lox; #P < 0.05 for cyclic MS FAK/JIP3 or FRNK/JIP3 vs. cyclic MS Cre-lox; +P < 0.05 for either FAK/JIP3 or FRNK/JIP3 under static condition vs. the same under condition of cyclic MS. Measurements in WT cells at the same time were indistinguishable from Cre-lox cells under the same conditions and are omitted from the graph for clarity.

We then examined whether expression of either JNK isoform would accelerate migration. Overexpression of JNK1 and JNK2 was detected by immunoblotting with HA and His antibody, respectively. Overexpression of constitutively active JNK1 accelerated migration under both static and cyclic MS conditions (Fig. 6), whereas coexpression of constitutively active JNK2 had no effect (data not shown). Under conditions of cyclic MS, expression of JNK1, while accelerating migration, did not overcome completely the inhibitory effect of cyclic MS (Fig. 6).

Fig. 6.

Wound closure of scrape-wounded 16HBE14o− cells on laminin matrix expressing constitutively active JNK1 monitored for 9 h. Data are expressed as means ± SE, n = 3 independent experiments (n = 6). *P < 0.05 for static JNK1 vs. static Cre-lox; #P < 0.05 for cyclic MS JNK1 vs. cyclic MS Cre-lox; +P < 0.05 for MS Cre-lox vs. static Cre-lox. Measurements in WT cells at the same time were indistinguishable from Cre-lox cells under the same conditions and are omitted from the graph for clarity.

Accelerated migration depends on JIP3 binding to JNK1 but not to JNK2.

To confirm that both FAK and JNK resided on the same JIP3 scaffold, we coexpressed WT-FAK or FRNK, JIP3 or ΔJBD, and either JNK1 or JNK2 in 16HBE14o− cells. Cells were examined in both static conditions and following cyclic MS, and immunoprecipitation was done using an anti-FLAG antibody. Under both static conditions and with cyclic MS, JNK1 precipitated with FAK in cells coexpressing JIP3 but not with either FRNK or ΔJBD (Fig. 7A). Interestingly, we found that JIP3 immunoprecipitated with JNK1 (Fig. 7A) but not JNK2 (Fig. 7B). This difference suggested that the phosphorylated FAK-JIP3 scaffold selectively binds to the JNK1 isoform both under static and cyclic MS conditions.

Fig. 7.

JIP3 forms a scaffold for FAK with JNK1 but not JNK2 and is involved in migration of airway epithelial cells. Cell lysates of scrape-wounded 16HBE14o− monolayers on laminin matrix, coexpressing either WT-FAK or FRNK, JIP3 or ΔJBD, or JNK1 (A) or JNK2 (B) adenoviral mutants, were immunoprecipitated with FLAG antibody, and equal amounts of proteins were subjected to electrophoresis. Representative Western blots were probed with JNK1 or JNK2 antibody. For equal protein determinations, immunoblots were probed with GAPDH antibody, n = 3 independent experiments. HA, hemagglutinin.

DISCUSSION

As shown schematically in Fig. 8, wounding initiates interactions between FAK, Src, and integrins that results in the assembly of a scaffold that includes phosphorylated FAK, Src, and JIP3. This scaffold then facilitates the phosphorylation of JIP3 and subsequently JNK. JIP3 and related JIP proteins were originally identified as scaffolding proteins for the JNK cascade by their binding to JNK and related MAPK kinases such as SEK1 and MEKK1 (37, 38). JNK can be localized to focal adhesion contacts (2, 14), and JNK activation is critical to both cytoskeletal reorganization (17, 51) and cell migration (17, 45, 48). Although cyclic MS slows migration of both epithelial (10, 30) and nonepithelial cells (24, 25), how cyclic MS might regulate expression of JNK and signaling to the JNK cascade is unclear. One study suggested that cyclic MS may activate JNK with subsequent activation of AP-1 (19), although the effect on cell migration was not clear. Our study demonstrates for the first time that cyclic MS decreases migration by reduction in JNK phosphorylation. This is associated with decreases in FAK (Tyr³⁹⁷) phosphorylation and decreased association of JIP3 with JNK. Although cyclic MS stimulated an initial increase in FAK phosphorylation, continued MS led to decreased FAK phosphorylation, loss of association with JIP3, and decreased activation of JNK. Restoration of JIP3 binding to JNK1 but not JNK2 accelerated migration under conditions of cyclic MS but not to the degree seen under static conditions. Taken together, our data demonstrate the importance of the FAK-JIP3-JNK1 signaling pathway in the early events of cell migration both under static conditions and under conditions of cyclic MS.

Fig. 8.

Schematic diagram of signaling pathways induced by wounding and how mechanical stretch can disrupt these processes leading to decreased cell migration.

Past studies of the effect of cyclic MS on FAK phosphorylation have led to divergent results. Cyclic MS (10% average strain at 60 cycles/min for 30 min to 4 h) caused increased tyrosine phosphorylation of FAK in endothelial cells (50), but the consequences on migration were not clear. FAK autophosphorylation was enhanced in SW620 colonocytes grown on collagen I matrix and subjected to 15 mmHg increased pressure for 30 min (40) and in neonatal rat cardiomyocytes exposed to 10% static strain for 30 min (33). Interestingly, FAK (Tyr³⁹⁷) phosphorylation increased within 2 min and declined after 60 min of 10% cyclic MS in A549 cells (5) and Caco-2 cells (6) grown on collagen I membranes. In bovine pulmonary artery endothelial cells, 25% cyclic stretch for 15 min upregulated FAK-397 phosphorylation (1). These studies demonstrated increased FAK phosphorylation levels over relatively short time intervals up to 60 min. Another difference between these studies and ours is that our measurements were made in monolayers of wounded cells, which induced an initial transient increase in phosphorylation followed by a later decline (Fig. 2B). The transient increase in FAK autophosphorylation at Tyr³⁹⁷ following wounding is consistent with an increase in integrin-mediated signaling and initiation of cell migration. Under static conditions, this transient peak in phosphorylation was followed by a decrease. When wounded cells were exposed to MS, the peak phosphorylation was higher than under static conditions, but after 6 h the levels of phosphorylation were significantly decreased relative to static conditions. We speculate that the short-term stimulation of FAK phosphorylation may be related to an initial increase in adhesion in response to the mechanical stimulus, and the long-term decrease may be the result of dissociation of the focal adhesion complex due to continuous mechanical disruption.

In contrast, FAK phosphorylation was downregulated, with subsequent deceleration of migration, in 16HBE14o− cells after downregulation of the adhesion molecule catenin α-like 1 (CTNNAL1; Ref. 49). In our study, decreased FAK (Tyr³⁹⁷) phosphorylation was clearly linked to decreased migration, both under static conditions and under conditions of cyclic MS. Although we observed a transient increase in FAK phosphorylation following wounding, levels were decreased after 6 h of MS. This suggests that the duration of cyclic MS may influence the state of FAK phosphorylation, with longer periods leading to less phosphorylation and consequently less migration.

The role of ECM proteins on FAK phosphorylation and migration is not completely clear. In contrast with our results with laminin-5 matrix (Fig. 1), previous studies demonstrated that 16HBE14o− cells migrated at the same rate on collagen IV compared with either laminin-1 or laminin-2 matrix after injury (45). That study also demonstrated that β₁-, α₂-, α₃-, and α₆-integrin subunits were expressed on the surface of 16HBE14o− cells, and antibody blocking studies showed that each of these subunits regulated migration on collagen IV but that only the β₁-subunit regulated migration on laminin-1 or laminin-2. Integrin α₃β₁ binds to laminin-5 with high affinity and facilitates cell migration (reviewed in Ref. 32), so this likely contributes to the enhanced migration in our studies, but we did not verify this with antibody blocking experiments. Overexpressing laminin-5 enhanced migration of A549 cells via increases in FAK phosphorylation (20). Laminin-5 can induce an osteogenic phenotype in human mesenchymal stem cells, a process that requires phosphorylation of FAK (28). Keratinocyte spreading and migration on laminin-5 was substantially inhibited by overexpressing FRNK and was regulated by the interaction of laminin-5 with α₃β₁-integrin, which regulated FAK signaling (7). In the present study, FAK phosphorylation, interaction with JIP3, and migration in 16HBE14o− cells were increased when cells were grown on laminin-5 vs. collagen IV matrix under either static conditions or with cyclic MS. As discussed above, laminin-5 matrix may enhance binding of specific integrins (such as α₃β₁-), and this increased binding would stimulate autophosphorylation of FAK at Tyr³⁹⁷. Increased levels of p-FAK would then enhance association with JIP3 (37, 38). Cyclic MS caused dissociation of this complex, potentially because of mechanical dissociation of the integrins or direct disruption of the complex. However, we did not directly examine integrin interactions in this study. Expression of WT-FAK accelerated, and expression of FRNK attenuated, migration on both laminin-5 and collagen IV, consistent with results seen in other cell types (12, 21, 22, 36, 47). Our results demonstrate that interactions between FAK and JIP3 are essential for efficient cell migration following injury and that cyclic MS disrupts these interactions.

GRANTS

These studies were supported by National Institutes of Health Grants HL-064981 (to C. M. Waters) and HL-080417 (to S. R. White).