| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Surgery, John D. Dingell VA Medical Center and Wayne State University, Detroit, Michigan, USA
1Correspondence: Surgical Service, John D. Dingell VA Medical Center, 4646 John R. St., Detroit, Michigan 48201-1932, USA. E-mail: marc.basson{at}va.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: integrin affinity • membrane translocation • p85 subunit • Akt isoforms • physical forces • mechanotransduction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Increased extracellular pressure also stimulates the adhesion of colon cancer cells to wound sites in vivo (3)
, and stimulation of malignant cell adhesion by pressure does not appear limited to colon cancers (4)
. Cancer cells are continuously exposed to such mechanical forces as pressure, strain, and shear in situ during intravascular or lymphatic transit, during surgical manipulation or laparoscopic peritoneal insufflations, and following abdominal surgery, as intraabdominal pressure is increased by intestinal edema. However, the pathways by which such stimuli influence integrin function in cancer cells are not well understood. We have previously demonstrated that pressure increases integrin-mediated colon cancer cell adhesion through inside-out FAK, Src, and cytoskeleton-dependent signaling mechanisms (2
, 5)
. However, how Src and FAK activation lead downstream to the stimulation of malignant cell adhesion has not been elucidated.
Activation of the PI 3-kinase (PI3K)/Akt signaling pathway has been correlated with prostatic metastasis (6)
, colon cancer cell invasion (7)
, and postoperative tumor growth (8)
, while amplification or overexpression of PI3K/Akt pathway elements occurs in several human cancers including ovarian and colonic carcinomas (9
, 10)
. Indeed, PI3K has been implicated as a key signaling molecule for integrin activation and regulation of actin reorganization and cell adhesion (11)
.
Mechanical stimuli other than pressure activate the PI3K/Akt pathway in endothelial and other cells (12
, 13)
. For instance, luminal flow activates PI3K in rat thick ascending limb (14)
, shear activates PI3K in osteoblasts (15)
, and flow-induces monocyte adhesion to HUVEC (16)
, but different physical forces may exert very different effects. In some cellular systems, conversion of integrins to the high affinity state is triggered by PI3K activation (17
, 18)
. Integrin activation may in turn contribute to the adhesive and invasive properties of cancer cells (19)
. We, therefore, hypothesized that activation of PI3K and its downstream signals contribute to the stimulation of colon cancer cell adhesion by extracellular pressure.
PI3K is a heterodimeric protein composed of a catalytic (p110) and a regulatory (p85) subunit. p110 must bind p85 for full activation (20)
. Activated PI3K can stimulate a number of cellular intermediates. Among these, Akt promotes cell survival and enhances tumor cell growth and invasiveness (21
, 22)
. We sought to determine the role of PI3K signaling in pressure-induced colon cancer cell adhesion. We studied the activation of PI3K and its downstream signals and characterized the role of these signals in mediating the effects of pressure. We further investigated the specific Akt isoform responsible for pressure-induced cancer cell adhesion using isoform-specific siRNA. Finally, since FAK and Src are the nonreceptor tyrosine kinases previously shown activated by pressure, we investigated the interaction of FAK and Src with the PI3K pathway and further assessed whether FAK interacts with integrins in a Src-dependent or PI3K-dependent manner to promote adhesion. Our results identify a pressure-activated novel inside-out Src/PI3K/FAK/Akt pathway that actively regulates colon cancer cell adhesion.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture
SW620 cells were maintained and primary human colonocytes were isolated from resected tumors by collagenase digestion as described (1)
. Cell viability was >90% by trypan blue exclusion. These studies were approved by the Human Investigations Committee of Wayne State University.
Pressure regulation
Pressure was regulated using an airtight box prewarmed to 37°C to maintain temperature within ±2°C and pressure within ±1.5 mmHg as described (1)
.
Matrix protein precoating
Bacteriologic plastic was precoated with collagen I (Sigma) at saturating densities of 12.5 µg/ml using an ELISA-based buffer at 4°C as described (1)
.
Cell adhesion
Cells (100,000/well) were allowed to adhere to collagen-coated six-well plates for 30 min at 37°C under ambient and increased pressure (15 mmHg). After 30 min, nonadherent cells were gently washed away with warm phosphate buffered saline. Adherent cells were formalin-fixed, hematoxylin-stained, and counted in 20 or more random high-power fields per well using an Olympus microscope. Experiments were performed in a paired fashion and normalized to control before pooling and data analysis.
Inhibitors
In some studies, cells were pretreated with PP2 (20 µM), LY294002 (20 µM), wortmannin (0.5 µM), LY303511 (20 µM), Akt inhibitor IV (10 µM), or rapamycin (100 nM).
Western blotting
For signaling studies, suspended nonadherent cells were subjected to ambient or increased pressure 30 min in bacteriologic plastic dishes pretreated with heat-inactivated 1% BSA to block nonspecific adhesion, collected by centrifugation, lysed, and subjected to Western blotting as described (2)
for activated Src (Tyr416), Akt (Ser473) and S6K (Thr389), FAK (Tyr 397, -576), p85, and phosphotyrosine. We studied signaling in suspended cells prior to adhesion because adhesion triggers a separate signal cascade that might obscure the events that we wished to study. The signals that we study here necessarily occur prior to adhesion because they regulate it.
Membranes were reprobed with antibodies to total Src, FAK, Akt, and S6 kinase. Bands were detected with ECL plus by a Kodak Image Station 440CF (Perkin Elmer, Boston, MA, USA) within the linear range.
Immunoprecipitation
Immunoprecipitation was performed in lysis buffer without deoxycholic acid or SDS, and with 1 mM NaF to conserve protein–protein interactions. Immunoprecipitations were performed at 4°C using protein G or protein A Sepharose for p85, FAK total protein, or HA-tagged proteins by manufacturers’ recommendations. Immunoprecipitates were washed four times with lysis buffer, boiled with 2x sample buffer, resolved by SDS-PAGE, and processed for Western blot analysis.
Transfection
SW620 cells were transfected with small interfering RNA (siRNA) to reduce FAK or AKT isoforms. FAK siRNA sequence as described (23)
and custom-made siRNA for AKT isoforms and non-targeting control siRNA were obtained from Dharmacon (Lafayette, CO, USA). The siRNA duplexes were introduced with oligofectamine per manufacturer’s protocol. The cells were used after 48 h for adhesion experiments and for Western blot analysis to study the effect of FAK inhibition on PI3K activity and Akt phosphorylation. Akt isoform siRNA-treated cells were used for adhesion and expression studies. In FAK translocation studies, HA-tagged WT-type FAK and HA-tagged FAK Y397F mutant (gifts from Dr. J. L. Guan) were transfected with Lipofectamine Plus and studied after 48 h.
Subcellular fractionation
Soluble and particulate cell fractions were prepared after collecting cells in extraction buffer (20 mM HEPES, pH 7.6, with 5 mM EGTA, 5 mm Na pyrophosphate, 1 mM MgCl2) with protease inhibitors. Samples were homogenized and centrifuged at 100 000 g for 60 min, and the soluble fraction was removed. The crude membrane particulate fraction was washed with extraction buffer and resuspended in 20 mM HEPES, pH 7.6, with 150 mM NaCl, 1 mM MgCl2, and 1% Triton X-100 and centrifuged to remove insoluble materials. Equal protein aliquots of soluble and particulate fractions were resolved by SDS–PAGE, transferred to nitrocellulose, and probed for FAK or p85. Equal protein loading was verified by blotting for GAPDH in soluble fractions and Na+ K+ ATPase in particulate fractions.
ELISA-based PI3K assay
p85 was immunoprecipitated from equal protein cell lysates, and immunoprecipitated PI3K was assayed by competitive ELISA-based assay by manufacturer’s protocol. Briefly, the reaction products were mixed and incubated with a PI (3, 4, 5) P3 detector protein and then added to a PI (3, 4, 5) P3-coated microplate for competitive binding. A peroxidase-linked secondary detection reagent was used to detect PI (3, 4, 5) P3 detector protein binding to the plate, and the amount of PI (3 ,4, 5) P3 produced was calculated from a standard curve prepared from known concentrations of lipid product.
Statistical analysis
Adhesion studies were performed in triplicate wells, and at least three separate studies with similar results were performed prior to results being pooled. Cell counts were normalized to basal adhesion. Images from Western blotting or immunoprecipitation studies were visualized using a Kodak Image Station 440CF (Perkin Elmer, Boston, MA, USA) within the linear range of exposure prior to densitometric quantitation. All data are presented as mean ± SE. Statistical analysis was performed using paired or unpaired t tests as appropriate. A P-value of <0.05 was a priori considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A; n=3; P<0.01) reduced basal adhesion and prevented pressure-induced adhesion. We confirmed that PI3K mediates the effects ofpressure in primary cancer cells isolated directly from human colon cancers. LY294002 lowered basal adhesion compared to primary cells treated with vehicle alone, and blocked pressure-induced primary colon cancer cell adhesion (Fig. 1B
, n=3; P<0.01). Pretreatment of SW620 cells with LY303511, a negative control that does not inhibit PI3K activity, did not block basal or pressure-stimulated adhesion (Fig. 1C
; n=3; P<0.05).
|
Pressure stimulates PI3K activity and p85 subunit tyrosine phosphorylation and translocation
Pressure increased PI3K activity 1.56 ± 0.19-fold (Fig. 2
A; n=6, P<0.001) and increased p85 tyrosine phosphorylation 1.79 ± 0.32-fold (Fig. 2B
, n=6; P<0.05). p85 is normally cytosolic but translocates to cell membranes and the cytoskeleton in response to stimuli (24)
. Pressure increased p85 in a subcellular membrane fraction 1.43 ± 0.09-fold (Fig. 2C
; n=3; P<0.01). Cytoplasmic p85 did not change demonstrably, perhaps because the much greater basal abundance of p85 in the cytosol overshadowed any reduction by translocation.
|
Pressure-induced adhesion requires PI3K-dependent activation of Akt but not S6 kinase
Pressure increased Akt phosphorylation at Ser473 in both primary colon cancer (Fig. 3
A, n=3; P<0.01) and SW620 cells in suspension prior to adhesion (Fig. 3B
; n=6; P<0.05), indicating Akt activation (25)
. Pretreatment with 500 nM wortmannin or 20 µM LY 294002 reduced basal and prevented pressure-induced Akt phosphorylation (Fig. 3B
, P<0.01; n=3). Pretreatment with LY303511 did not block basal or pressure-induced Akt phosphorylation (Fig. 3H
; n=3; P<0.05). More specific Akt inhibition with 10 µM Akt inhibitor IV also markedly reduced basal Akt phosphorylation and blocked pressure-induced Akt activation in SW620 cells (Fig. 3C
; n=3, P<0.01). The Akt inhibitor lowered basal adhesion and abolished pressure-induced adhesion in both primary (Fig. 3D
, n=3; P<0.01) and SW620 colon cancer cells (Fig. 3E
, n=3; P<0.01). These results suggest not only that pressure modulates colon cancer cell adhesion in via PI3K and Akt but also that this pathway is active in primary human colon cancer cells ex vivo.
|
Three Akt isoforms exist (26)
. Isoform-specific Westerns suggested Akt1 and Akt 2 predominate in SW620 cells. Akt1 or Akt2 siRNA lowered Akt1 
50% or Akt2 60% (Fig. 3F
, n=3; P<0.001). Akt2 was not affected by Akt1 siRNA. Akt1 reduction blocked pressure-induced adhesion, but Akt2 reduction had no effect.
To determine whether this apparent isoform specificity might reflect the preponderance of Akt1 in SW620 cells, we studied Caco-2 colonocytes. We first demonstrated that Caco-2 cells express relatively more Akt2. siRNA to Akt1 reduced SW620 Akt1 by 50% and total Akt by 50%, whereas siRNA to Akt1 reduced Caco-2 Akt1 by 80% but total Akt by 30%. In contrast, siRNA to Akt2 also reduced SW620 Akt2 by 50% but reduced total Akt by only 10%, whereas siRNA to Akt2 reduced Caco-2 Akt2 by 80% but total Akt by as much as 70%. Thus, Akt1 siRNA is much more effective in reducing SW620 total Akt than Akt2 siRNA, while this pattern is reversed in Caco-2 cells, indicating a greater preponderance of Akt1 in SW620 cells and of Akt2 in Caco-2 cells. (data not shown) Despite the apparent preponderance of Akt2 in Caco-2 cells, lowering Akt1 in Caco-2 cells prevented pressure-stimulated adhesion, but as in SW620 cells, Akt2 reduction had no effect (Fig. 3G
; n=3; P<0.001).
PI3K also activates S6 kinase (S6K). The mTOR inhibitor rapamycin, an upstream effector of S6K, inhibits murine osteosarcoma metastasis (27)
. However, although both LY294002 (Fig. 4
A, n=3; P<0.001) and rapamycin (Fig. 4B
, n=3; P<0.001) ablated basal S6 kinase phosphorylation, pressure did not activate SW620 S6K, and rapamycin (100 nM) pretreatment did not block pressure-stimulated adhesion. (Fig. 4C
, n=3; P<0.05 for each comparison).
|
PI3K is downstream of Src but upstream of FAK in the pressure response
Since FAK and Src mediate pressure-induced adhesion, we evaluated how PI3K inhibition affects FAK and Src. LY294002 (20 µM) prevented pressure-induced FAK397 phosphorylation (Fig. 5
A; n=5; P<0.05) but did not alter Src-dependent FAK576 phosphorylation (Fig. 5B
; n=4; P<0.05) or Src phosphorylation by pressure (Fig. 5C
; n=6; P<0.05). Thus, pressure activates Src independently of PI3K, but FAK397 phosphorylation in response to pressure requires PI3K.
|
Pressure-induced p85 interaction with FAK requires Src, PI3K, or FAK activation
p85 may associate with FAK by binding to tyrosine phosphorylated residue 397 of FAK (28)
. FAK-PI3K association is stimulated by integrin-mediated astrocytoma adhesion to fibronectin (29)
. Therefore, we investigated whether pressure induces PI3K- FAK interaction in suspended cells prior to adhesion. Pressure increased the amount of FAK coprecipitating with p85 (Fig. 5A
, n=4; P<0.05). Inhibition of Src (Fig. 6
A; n=4) or PI3K (Fig. 6B
; n=4) prevented pressure-induced FAK association with p85, as did FAK reduction by siRNA (Fig. 6C
; n=4), suggesting that pressure-induced interaction of p85 with FAK is regulated by Src and PI3K.
|
Pressure significantly translocates Akt to plasma membrane
Akt translocation to the plasma membrane is essential for activation (30)
. Akt translocated to the membrane fraction in response to pressure (Fig. 6D
, n=3; P<0.05). Akt inhibitor IV reduced basal membrane Akt and prevented pressure-induced translocation of both Akt and FAK (Fig. 6E
, n=3). Pretreatment of SW620 cells with Akt inhibitor IV significantly inhibited basal and prevented pressure-induced FAK397 phosphorylation (Fig. 6F
; n=3; P<0.05).
Pressure-induced Akt activation downstream of PI3K requires FAK
Pressure increased PI3K activity in cells transfected with nontargeting siRNA or with siRNA to reduce FAK (Fig. 7
B, n=4; P<0.05). Significant reduction in FAK protein expression (60%) by FAK siRNA transfection was shown in Fig. 7A
(n=4; P<0.05). This is consistent with the degree in FAK reduction, which we have previously demonstrated to block the effects of increased pressure on adhesion and other pressure-induced signals downstream of FAK (2)
. Pressure also increased Akt phosphorylation in cells transfected with nontargeting siRNA. However, transfection with FAK siRNA blocked pressure-induced Akt phosphorylation without affecting Akt protein (Fig. 7C
; n=4). Thus, pressure-induced Akt activation downstream of PI3K requires FAK.
|
Pressure-induced PI3K/Akt signals require Src
We then asked whether Src inhibition blocks pressure activation of PI3K and Akt. Basal PI3K activity was slightly but significantly increased in cells at ambient pressure pretreated with PP2 (20 µM) (Fig. 7D
, n=4; P<0.05). PP2 prevented PI3K activation by pressure and inhibited both basal and pressure-induced Akt phosphorylation (Fig. 7E
, n=3; P<0.05). Thus, pressure-induced PI3K and Akt activation require Src.
Pressure induces translocation of FAK to plasma membrane
Integrin-mediated adhesion to matrix increases FAK tyrosine phosphorylation and accumulation in focal adhesions (31)
. However, since FAK activation precedes adhesion in this model, we investigated whether pressure activation of FAK is accompanied by FAK movement to focal adhesion sites and increased adhesion. We further asked whether FAK397 phosphorylation is required for FAK translocation by transfecting cells with an HA-tagged FAK Y397F mutant incapable of being tyrosine-397-phosphorylated. Pressure increased FAK in the membrane fraction (Fig. 8
A, n=3; P<0.05). Pressure similarly increased levels of transfected HA-tagged WT FAK protein or the HA-tagged FAK397 mutant in the membrane fraction (Fig. 8B
, n=5; P<0.05).
|
Pressure-induced Src and PI3K activation regulates FAK: integrin ß1 interaction
The N-terminal domain of FAK binds to peptides mimicking the ß1integrin subunit cytoplasmic domain, suggesting that integrins could interact directly with FAK (31)
. We investigated whether pressure induces FAK-ß1 integrin subunit interaction in suspended cells prior to adhesion. Pressure increased the amount of ß1 coprecipitating with FAK, suggesting that FAK complexes with integrin ß1 subunit in response to pressure (Fig. 9
, n>5; P<0.05). Src (Fig. 9A
; n=4) or PI3K (Fig. 9B
; n=4) inhibition prevented pressure-induced FAK-ß1 association.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The present study establishes a role for PI3K and Akt1 in this process and traces a pathway by which PI3K, Akt1, FAK, and Src may interact in response to increased extracellular pressure to stimulate cancer cell adhesion. In particular, we demonstrate that pressure activates PI3K and Akt and that this activation is required for the pressure effect on adhesion. The Akt effect is isozyme-specific, involving only Akt1. Our results delineate a pathway running from Src through PI3K and FAK, which mediates Akt1 activation and translocation of Akt1 and FAK to the membrane, where FAK associates with ß1 integrin heterodimers to modulate integrin-mediated adhesion. Thus, taken together with our previous observations of the role of the cytoskeleton in initial mediation of FAK activation by pressure (4
, 5)
, these results delineate the pressure activation of an inside-out Src-PI3K-FAK-Akt1 pathway, which modulates integrin-binding affinity through interaction of FAK with ß1 integrin (Fig. 10
).
|
Although PI3K may be activated by integrin-induced tyrosine kinase activation (28)
, here PI3K activation in response to pressure precedes the adhesion response and in fact regulates it. PI3K influences basal adhesion of some cancer cells, including human colon carcinoma 201 cells (32)
, consistent with our observation that basal SW620 adhesion is reduced by PI3K inhibition. Forces such as pressure, deformation, and shear stress stimulate PI3K in some (33
34
35)
but not all cells (36)
. Most relevantly, shear stress-induced activation of PI3K promotes platelet adhesion (33)
. In some hematopoietic lineages, conversion of integrins to the high-affinity state is triggered by PI3K (37)
. Affinity modulation of ß1 integrin for fibronectin requires PMA-induced activation of PI3K in marrow-derived mast cells (17)
, and cross-linking of the CD28 Ag on T cells increases ß1-integrin-mediated adhesion to fibronectin via PI3K (38)
. Our results suggest that PI3K activation by pressure modulates integrin affinity in colon cancer cells without chemical agonists.
The requirement for PI3K in adherence to matrix may reflect an alteration in focal adhesion complexes, possibly involving cytoskeletal rearrangements. PI3K regulates actin polymerization (39)
, and pressure-induced cell adhesion requires cytoskeletal integrity (4
, 5)
. Further, our data suggest that Src-dependent PI3K activation causes PI3K to activate cytosolic FAK and initiate its translocation to the cell membrane, where FAK associates with integrins and regulates integrin-mediated adhesion.
Pressure-mediated PI3K activation is accompanied by p85 subunit phosphorylation and association with FAK. The p85 subunit of PI3K is an adapter protein phosphorylated by receptor or Src family kinases (40)
. Association of p85 phosphorylation with PI3K activity may be cell- and stimulus-specific. Some authors suggest that PI3K activation requires p85 tyrosine phosphorylation to relieve p110 inhibition (40)
, modulate PI3K activity (41)
, and alter the SH2 domain binding properties of PI3K (42)
. However, p85 phosphorylation was not associated with PI3K activation in either IGF-stimulated fibroblasts (43)
or mast cells treated with cytokines (44)
. Our results suggest that pressure increases tyrosine phosphorylation of the PI3K p85 subunit in parallel with PI3K activation. The phosphorylation and binding of p85 to tyrosine-phosphorylated proteins may also recruit PI3K from the cytosol to the plasma membrane, where it can contact its activator and its substrate, inositol-containing lipids (24)
. Indeed, we also found that the activation of PI3K by pressure was accompanied by translocation of the p85 subunit to a membrane fraction.
PI3K can stimulate many intermediates, including PKC, S6K, and Akt. Pressure-stimulated adhesion does not require PKC (2)
. Although S6K contributes to adhesion in other cells (45)
, it does not appear to play a role in the stimulation of colon cancer cell adhesion by increased extracellular pressure. Indeed, S6K activation has been reported to be independent of pathways that regulate focal adhesion formation (46)
.
In contrast, our results suggest that Akt activation by PI3K is required for pressure to stimulate adhesion. Akt activation enhances tumor cell invasiveness (22)
. Pressure increased Akt phosphorylation at Ser473 in colon cancer cells, and Akt activation required PI3K. Consistent with our findings, a 10–60 mmHg pressure overload activates myocardial Akt (35)
. PI3K-dependent Akt activation by force has been described in endothelial cells induced by shear (47)
and vascular myocytes by strain (48)
. However, Akt is activated independently of PI3K in response to heat shock and hyperosmotic stress (49)
, ß3-adrenergic receptor activation (50)
, and angiotensin II induced mesangial hypertrophy (51)
. The mechanism of PI3K-independent activation of Akt is unknown, but pressure-induced PI3K activity seems required for Akt phosphorylation in response to pressure.
Akt1 and Akt2 are ubiquitous but Akt3 is found chiefly in brain, heart, and kidney (26)
. Akt1 and Akt2 predominate in SW620 and Caco-2 cells. Although these isoforms are 80% homologous (26)
, only Akt1 mediates pressure-stimulated adhesion.
If pressure stimulation of cancer cell adhesion requires PI3K/Akt activation, how force activates PI3K and Akt is not well understood. Kinases such as FAK, Src, and VEGFR2 may activate PI3K and Akt in response to pressure, shear stress, and mechanical strain (35
, 48
, 52)
. Pressure activates FAK and Src prior to adhesion, and inhibiting FAK or Src inhibits pressure-induced adhesion similarly to inhibiting PI3K, suggesting that these kinases jointly mediate this event.
Src inhibition prevented pressure stimulation of PI3K, consistent with reports that Src kinases activate PI3K in response to other stimuli in other cells (53
, 54)
. Conversely, PI3K inhibition did not prevent Src phosphorylation in response to pressure or Src-dependent FAK576 phosphorylation, suggesting that pressure-induced Src activation may act upstream of PI3K and Akt activation in pressure-induced inside-out signaling.
In some cells, adhesion-induced FAK activation triggers adhesion-dependent PI3K activation (29)
. We studied signaling in suspended cells prior to adhesion. In our model, the effect of pressure on FAK397 phosphorylation prior to adhesion was markedly attenuated by LY294002, suggesting that PI3K regulates pressure-stimulated FAK397 phosphorylation, the first step in FAK activation. This is consistent with reports that FAK activation induced by paracrine or endocrine factors requires PI3K activation in some (55
, 56)
, although not all cells (57
, 58)
.
We also found that FAK associates with the p85 subunit of PI3K. Increased FAK association with the p85 subunit has similarly been demonstrated in glutamate-treated striatal neurons (59)
, and cerebellar granules responding to hypoosmolarity, which may act, to some extent, by mechanotransduction (60)
. Furthermore, v-Crk-induced p85 binding to FAK397 occurs in fibroblasts in suspension (61)
.
The pressure-increased FAK association with p85 in our study was inhibited by blocking PI3K or Src. In contrast, activation of PI3K precedes FAK phosphorylation in TSP-1-treated vascular myocytes, and no direct FAK-p85 interaction was observed in that study (62)
. In another cell type, Src inhibition did not prevent the association of FAK with p85 in glutamate-treated neurons (59)
. Such disparities suggest that the mechanotransduction pathway outlined here differs in the upstream signals that evoke FAK-p85 association.
Akt inhibition not only inhibited pressure-induced Akt translocation but also prevented FAK translocation in response to pressure. Akt mediates the antiapoptotic effects of FAK in some cells (63)
. HMG-CoA reductase inhibitors induce Akt to translocate to endothelial plasma membranes, colocalizing with F-actin-positive, FAK-negative lamellipodia, and filopodia (64)
. In addition to interacting with FAK, Akt could, therefore, also modulate cytoskeletal integrity or rearrangement of the cytoskeletal elements in response to pressure, likely critical for mechanotransduction of the pressure effect (4
, 5)
.
Although Akt inhibitor IV blocked pressure-induced FAK translocation to membrane and FAK 397 phosphorylation, Akt may not regulate FAK through direct physical association. We could not coprecipitate FAK and Akt (data not shown). This is consistent with recent observations by others (65)
. Akt may regulate FAK indirectly by interacting with a copartner protein that can bind to both Akt and FAK to form a ternary complex. Akt activation might be required for Akt binding to such a partner protein, and thus for a ternary association with FAK during translocation of FAK to membrane. Alternatively, Akt might play some other role in the regulation of FAK protein translocation that does not require even indirect complexing with FAK. Such possibilities await further study.
Akt may be activated by various tyrosine kinases in response to different stimuli (35)
. Pressure-induced Akt phosphorylation was inhibited by FAK reduction and Src inhibition, suggesting that FAK and Src are also required for pressure-induced Akt activation. FAK mediates Akt activation in other force models. For instance, FAK is upstream of Akt in pressure-treated myocardium (35)
, HepG2 cells under hyposmotic stress (66)
, and coronary arterioles responding to flow (67)
. In contrast, Akt activation by prolactin was not affected by FAK inhibition in breast cancer cells, illustrating the diversity of signaling in response to different stimuli (68)
.
Notably, Src kinase inhibition inhibited Akt activation in the same prolactin-stimulated model (68)
. Src has also been placed upstream of Akt by molecular manipulation (69
, 70)
. Src inhibition also prevents endothelial Akt activation in response to flow (48)
. However, Src may not always be required for force-induced Akt activation, even in the same cell type. For instance, the Src inhibitor herbimycin A does not prevent Akt phosphorylation by shear stress in HUVEC (47)
.
Our work suggests that after association with p85, FAK translocates to a membrane/cytoskeletal fraction together with p85 and associates there with ß1 integrin heterodimers. Src or PI3K inhibition blocks this interaction, consistent with the concept that Src and PI3K are activated prior to FAK translocation in response to pressure. We have previously reported that pressure-induced adhesion is ß1 integrin-mediated, and that pressure does not change integrin surface expression (2)
. FAK translocation between cytosolic and membrane bound pools is highly regulated by many factors, including tyrosine phosphorylation and actin assembly (71)
. Since PI 3,4,5-trisphosphate can recruit SH2 domain-containing proteins to the cell membrane (72)
, PI3K could contribute directly to FAK translocation in response to pressure. Similar FAK translocation to focal contacts in cardiac myocytes in response to VEGF strengthens myocyte adhesion to the matrix (73)
.
FAK modulation of ß1 integrin binding affinity has been described in pancreatic cancer cells, but by a different and indirect pathway involving ERK (74)
. Pressure-stimulated adhesion is ERK-independent (2)
, and our data raise the possibility of a direct FAK-integrin modulation. The carboxyl-terminal tyrosine residue (Tyr397) of FAK is an important site of FAK phosphorylation that creates a high-affinity binding site recognized by the SH-2 domain of the Src family (75)
, permitting further phosphorylation of FAK by Src and related kinases. However, in vitro studies suggest that the N-terminal domain of FAK binds directly to peptides mimicking the ß1 integrin cytoplasmic domain (31)
. Thus, it may not be surprising that pressure-induced phosphorylation of FAK at Tyrosine 397 was not required for the pressure-induced FAK translocation to plasma membrane.
In summary, our observations suggest that extracellular pressure induces cytoskeletally independent Src activation that initiates a signal cascade involving PI3K, FAK, and Akt1. Cytosolic FAK, p85, and Akt appear to then translocate to the cell membrane where FAK associates with ß1 integrin heterodimers and regulates ß1 integrin heterodimer binding affinity to matrix proteins. Alterations in integrin binding affinity, thereby facilitate colon cancer cell adhesion. Pressure also stimulates the adhesion of cancer cells to endothelial cells (2)
and to surgical wounds in vivo (3)
. Cancer cells from glossal (4)
and ovarian and breast lineages (unpublished data) display similar biology. Thus, a mechanotransduced pathway such as that depicted in Fig. 10
may critically regulate the metastatic potential of malignant cells.
|
ACKNOWLEDGMENTS |
|---|
Received for publication May 24, 2006. Accepted for publication January 18, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
