Physical forces including pressure, strain, and shear can be converted into intracellular signals that regulate diverse aspects of cell biology. Exposure to increased extracellular pressure stimulates colon cancer cell adhesion by a β1-integrin-dependent mechanism that requires an intact cytoskeleton and activation of focal adhesion kinase (FAK) and Src. α-Actinin facilitates focal adhesion formation and physically links integrin-associated focal adhesion complexes with the cytoskeleton. We therefore hypothesized that α-actinin may be necessary for the mechanical response pathway that mediates pressure-stimulated cell adhesion. We reduced α-actinin-1 and α-actinin-4 expression with isoform-specific small interfering (si)RNA. Silencing of α-actinin-1, but not α-actinin-4, blocked pressure-stimulated cell adhesion in human SW620, HT-29, and Caco-2 colon cancer cell lines. Cell exposure to increased extracellular pressure stimulated α-actinin-1 tyrosine phosphorylation and α-actinin-1 interaction with FAK and/or Src, and enhanced FAK phosphorylation at residues Y397 and Y576. The requirement for α-actinin-1 phosphorylation in the pressure response was investigated by expressing the α-actinin-1 tyrosine phosphorylation mutant Y12F in the colon cancer cells. Expression of Y12F blocked pressure-mediated adhesion and inhibited the pressure-induced association of α-actinin-1 with FAK and Src, as well as FAK activation. Furthermore, siRNA-mediated reduction of α-actinin-1 eliminated the pressure-induced association of α-actinin-1 and Src with β1-integrin receptor, as well as FAK-Src complex formation. These results suggest that α-actinin-1 phosphorylation at Y12 plays a crucial role in pressure-activated cell adhesion and mechanotransduction by facilitating Src recruitment to β1-integrin, and consequently the association of FAK with Src, to enhance FAK phosphorylation.
- mechanical signaling
cell signaling in response to environmental stimuli typically requires the specific interaction of integrin receptors with extracellular matrix (ECM) ligands. Yet, physical forces including pressure, strain, and shear are able to affect integrin-mediated intracellular signaling and serve to regulate a diverse array of biological responses (1, 16, 31, 49). The manner in which the mechanical stimuli are translated into cellular responses is currently only poorly understood (24). Stretch-activated ion channels, mechanosensitive membrane-associated enzymes, distortion of cytoskeletal filaments, and integrin-ECM interactions all have been shown to contribute to the relay of mechanical signals in various cell models (2, 15, 20, 24, 59). Although mechanosensitive signaling components continue to be identified, the delineation of specific mechanisms by which these molecules influence cell behavior requires further investigation.
Metastatic progression is dependent on the adhesion of tumor cells to distant tissues. We have previously reported that a pathophysiologically relevant (15 mmHg) increase in extracellular pressure stimulates colon cancer cell adhesion to matrix proteins, endothelial cell monolayers, and surgical wounds in vivo by a β1-integrin-dependent mechanism that requires activation of both focal adhesion kinase (FAK) and Src (4, 50, 52). Cancer cells may be exposed to increased extracellular pressure during vascular and lymphatic transit, in the tumor microenvironment, or iatrogenically through surgical manipulation, laparoscopic insufflation, and postoperative bowel edema (10, 30, 35, 37, 58). Cells from colon cancer lines, murine colonic adenocarcinomas, primary human colon cancers, head and neck squamous cell cancers, and breast adenocarcinomas all display similar pressure-mediated phenomena (4, 8, 9, 50, 52). Pharmacological perturbation of the cell cytoskeleton with actin and microtubule disrupting agents inhibits the effect of pressure on cell adhesion and FAK phosphorylation without affecting Src activation (51). This observation suggests that an intact cytoskeletal network is necessary for the relay of some pressure-stimulated signals as well as the modulation of integrin-dependent cell adhesion. Still, a definite link between the cytoskeleton and associated pressure-activated signaling remains vague. The investigation of cytoskeletal-associated molecules specifically involved in the physical tethering of actin filaments to transmembrane adhesion receptors may provide insight into the underlying mechanisms governing this pressure-induced mechanical response.
α-Actinin is a ubiquitously expressed actin-cross-linking protein localized primarily along cytoskeletal filaments and at focal adhesion plaques in nonmuscle cells (26, 28). In addition to interacting with multiple focal adhesion proteins, the nonmuscle/cytoskeletal isoform of α-actinin associates with several families of adhesion receptors and is able to bind the cytoplasmic domain of β1-integrin (5, 39, 40). α-Actinin was found to be essential for integrin-cytoskeletal linkage and focal adhesion formation in Swiss 3T3 cells (43), and adhesion plaques devoid of α-actinin lacked stability and displayed higher turnover rates in Chinese hamster ovary cells (55). α-Actinin is a FAK substrate; FAK-dependent phosphorylation of tyrosine-12 (Y12) on α-actinin reduces its binding affinity for actin filaments, which may serve to regulate focal adhesion maturation and turnover (23, 55). The spatial and biochemical positioning of α-actinin as well as the importance of FAK activation in pressure-stimulated adhesion suggest a potential role for α-actinin in this process.
We therefore hypothesized that the physical tethering of actin filaments to integrin-associated focal adhesion complexes is critical for mediation of the effects of extracellular pressure on cancer cell adhesion and that α-actinin is a necessary component in this mechanical response pathway. To test this hypothesis, we selectively uncoupled the link between the actin cytoskeleton and focal adhesion complex by specific reduction of α-actinin protein levels with small interfering (si)RNA and assessed the impact on pressure-stimulated adhesion and signaling in colon cancer cells. We further addressed the potential role of α-actinin protein-protein interactions and phosphorylation in modulating this process by coimmunoprecipitation and Western blot analysis of α-actinin wild-type and phosphorylation mutant cell transfectants. Human SW620 cells were used as the primary model for this investigation, and key results were confirmed in additional colon cancer cell lines and in primary cells isolated directly from surgically resected human colon cancers.
MATERIALS AND METHODS
SW620, Caco-2, and HT-29 colon cancer cells were cultured as previously described (4). Single-cell suspensions of primary human colonocytes were isolated from resected tumors by mincing and collagenase digestion (11). Isolated colonocyte viability was >90% as determined by Trypan blue exclusion. Human tumor use was approved by the Wayne State University Human Investigation Committee.
Pressure was controlled using an airtight Lucite box with an inlet valve for gas application and an outlet valve connected to a manometer (4). The box was prewarmed to 37°C to prevent internal temperature and pressure fluctuations. Temperature was maintained within ±2°C and pressure within ±1.5 mmHg of desired levels.
Cell adhesion assay.
Suspended cells were allowed to adhere to collagen I-coated six-well plates (105 cells/well) for 30 min at 37°C under ambient and increased pressure (+15 mmHg) conditions. After 30 min, nonadherent cells were gently washed away with warm phosphate-buffered saline (PBS), adherent cells were fixed in formalin and stained with hematoxylin, and each well was counted in 20 or more random high-power fields using an inverted microscope (4). In parallel studies, SW620 cells transfected with green fluorescent protein (GFP)-tagged constructs were seeded at 106 cells/well and similarly treated and fixed, and then adherent cells were counted using a fluorescence microscope.
Cells were transfected with 50 nM double-stranded siRNA directed toward the mRNA target 5′-CACAGAUCGAGAACAUCGAAG-3′ for α-actinin-1 and toward 5′-CCACAUCAGCUGGAAGGAUGGUCUU-3′ for α-actinin-4 to inhibit isoform-specific α-actinin expression. Total α-actinin expression was reduced using an α-actinin-specific siRNA SMARTpool designed by Dharmacon (Lafayette, CO). FAK expression was inhibited by using siRNAs targeting the 5′-AAGCAUGUGGCCUGCUAUGGA-3′ segment of FAK mRNA. A Dharmacon siCONTROL nontargeting siRNA 1 (siNT) sequence was used as a control. Cotransfection with Dharmacon's siGLO (6-FAM) was used to determine siRNA transfection efficiency. SiRNA sequences were selected using the web-based DEQOR siRNA design program (http://cluster-1.mpi-cbg.de/Deqor/deqor.html). Specific sequences were chosen on the basis of sequence length, GC-content, nucleotide composition, and cross-silencing capability (2-nt mismatch default) (17). The selected siRNA duplexes were synthesized by Dharmacon. SiRNAs were introduced with Oligofectamine according to the manufacturer's protocol (Invitrogen, Carlsbad, CA). In parallel studies, SW620 cells were transfected with plasmids expressing amino-terminal His6-tagged wild-type (aac1 human; GenBank accession no. X15804) and Y12F mutant α-actinin-1, amino-terminal GFP-tagged wild-type and Y12F mutant α-actinin-1 (23), and carboxy-terminal hemagglutinin (HA)-tagged wild-type FAK (gift from Dr. David Schlaepfer) using Lipofectamine 2000 (Invitrogen) as recommended by the vendor. Cell transfectants were used for adhesion and signaling experiments after 48 h.
Immunoprecipitation and Western blotting.
Suspended (nonadherent) cells were subjected to ambient or increased pressure for 30 min in bacteriologic plastic dishes pretreated with 1% heat-inactivated bovine serum albumin to block cell adhesion. Cells were collected by centrifugation and lysed in buffer containing 50 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 1% deoxycholic acid, 0.1% SDS, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 1 mM sodium vanadate, 50 mM NaF, 10 mM sodium pyrophosphate, 2 μg/ml aprotinin, and 2 μg/ml leupeptin. Cell lysate protein concentrations were determined using a BCA protein assay kit (Pierce, Rockford, IL). Equal amounts of protein were resolved by SDS-PAGE and transferred to Hybond ECL nitrocellulose membrane (Amersham Pharmacia Biotech, Piscataway, NJ). Mouse monoclonal antibodies to phosphotyrosine (4G10), total Src, clone L4A1 (Cell Signaling, Beverly, MA), total FAK, clone 4.47 (Upstate, Temecula, CA), β1-integrin, clone 6S6 (Chemicon, Temecula, CA), GAPDH (Biodesign International, Saco, MN), total α-actinin, α-actinin-1 (US Biological, Swampscott, MA), and α-actinin-4 (Alexis, Lausen, Switzerland) and rabbit polyclonal antibodies to phosphorylated FAK (Y397 and Y576; BD Transduction Laboratories, San Diego, CA), phosphorylated Src (Y416), and β-actin (Cell Signaling) coupled with appropriate horseradish peroxidase-conjugated secondary antibodies (Cell Signaling) were used for immunodetection of blotted proteins. Bands were detected with enhanced chemiluminescence (Amersham) and analyzed with a Kodak Image Station 440CF (Perkin Elmer, Boston, MA).
Immunoprecipitation studies were conducted using 200–400 μg of protein per sample from the previously described experimental cell lysates. Protein samples were incubated with anti-His (Qiagen, Valencia, CA), anti-HA.11 (Covance, Berkeley, CA), anti-GFP (JL-8; Clontech, Mountain View, CA), or FAK, β1-integrin, β-actin, and Src monoclonal antibodies for 1 h, followed by an overnight incubation with preblocked protein A/G PLUS-agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C. Samples were washed with cell lysis buffer, eluted with Laemmli's loading buffer, and then resolved by SDS-PAGE and transferred to nitrocellulose membrane.
Soluble and particulate cell fractions were prepared after cells were collected in extraction buffer containing 20 mM HEPES, 5 mM EGTA, 5 mM sodium pyrophosphate, 1 mM MgCl2, and protease inhibitors. Samples were homogenized and centrifuged at 100,000 gfor 60 min, and the soluble (cytosol) fraction was removed. The crude membrane particulate fraction was washed with extraction buffer, resuspended in buffer containing 20 mM HEPES, 150 mM NaCl, 1 mM MgCl2, and 1% Triton X-100, and then centrifuged to remove insoluble materials. Equal protein aliquots of cytosolic and membrane fractions were resolved by SDS-PAGE, transferred to nitrocellulose, and probed for the proteins of interest. Equal protein loading was verified by blotting for Rho-GDI (Cell Signaling) in cytosol fractions and E-cadherin (H-108; Santa Cruz Biotechnology) in particulate membrane fractions.
Apoptosis assay and flow cytometric analysis.
After exposure to experimental conditions, cells were washed twice with PBS and resuspended in 1× binding buffer (BD Biosciences, San Diego, CA). Cells were then incubated with Annexin V-phycoerythrin (PE) and 7-amino-actinomycin (7-AAD; BD Biosciences) for 15 min at room temperature. Cells were diluted in 1× binding buffer and analyzed with flow cytometry within 1 h. Cell viability was determined by gating on either GFP- or 6-FAM-positive cell populations and assessing for Annexin V-PE and 7-AAD staining using a FACSCalibur flow cytometer (BD Biosciences) and WinMDI software.
Fluorescence confocal microscopy.
After exposure to ambient or increased pressure conditions, cells transfected with GFP-α-actinin-1 were washed with PBS and fixed with 4% paraformaldehyde for 10 min at room temperature. Cell suspensions were mounted on ProbeOn Plus glass slides (Fisher Scientific) and coverslipped using Geltol mounting medium (Immunon, Pittsburgh, PA). GFP-α-actinin-1 spatial distribution was visualized on an LSM 510 confocal microscope (Carl Zeiss, Jena, Germany) equipped with a ×40 objective and LSM software version 3.2 (Carl Zeiss).
All data are means ± SE. Statistical analysis was performed using either paired Student's t-test or the Wilcoxon matched-pairs signed-ranks test as appropriate. Sidak's correction was used for multiple comparisons (47). A 95% confidence interval was set a priori as the desired level of statistical significance.
α-Actinin is required for pressure-activated colon cancer cell adhesion.
As previously reported, a 30-min exposure to increased extracellular pressure (+15 mmHg) stimulates human SW620 colon cancer cell adhesion to collagen I by ∼20–40% (4). Pooled α-actinin-specific siRNAs, providing a 64 ± 6% reduction in protein expression (Fig. 1A), were used for preliminary assessment of the role of α-actinin in pressure-stimulated cell adhesion. SW620 cells transfected with a control nontargeting siRNA sequence (siNT) displayed the expected increase (34 ± 4%; n = 5; P < 0.01) in adhesion to collagen I under elevated pressure. α-Actinin siRNA (siACTN) transfectants exhibited a slight, statistically insignificant decrease (10 ± 6%; n = 5) in cell adhesion under ambient conditions. However, the reduction in α-actinin expression completely blocked the increase in adhesion typically accompanying exposure to elevated pressure (Fig. 1B).
α-Actinin-1, but not α-actinin-4, is required for pressure-activated colon cancer cell adhesion.
Four α-actinin isoforms have been identified in humans, of which only α-actinin-1 (ACTN1) and -4 (ACTN4) are found in colonic epithelial cells (19, 42). We therefore sought to assess whether pressure-stimulated cell adhesion involved a specific α-actinin isoform. Multiple isoform-specific siRNA sequences were screened and transfection conditions were optimized to achieve a 50–70% knockdown of α-actinin-1 and -4 protein in SW620 cells (Fig. 1C). A similar knockdown efficiency was achieved with Caco-2 and HT-29 cells (data not shown). Cells transfected with siNT served as a control group for each population. SiRNA transfection efficiency was assessed by cotransfection of SW620 cells with siGLO (6-FAM) fluorescent siRNAs. Flow cytometric analysis of the cotransfectants determined ∼90% of the cells to be 6-FAM+ (Fig. 1E, left). Under elevated pressure conditions, the SW620 siNT transfectants displayed a 33 ± 7% increase (n = 3; P < 0.05) in cell adhesion. SiACTN4 SW620 cell transfectants exhibited a 39 ± 9% increase (n = 3; P < 0.05) in adhesion under elevated pressure, similar to that of the control cells. In contrast, siRNA-mediated reduction of α-actinin-1 prevented the pressure effect (n = 3). Identical results were obtained with both Caco-2 and HT-29 cells (Fig. 1D). As an additional control experiment, we assessed the effect of α-actinin-1 reduction on SW620 cell viability under ambient and increased pressure conditions. After exposure to experimental conditions, cells were stained with Annexin V-PE as an early marker of apoptosis and 7-AAD as a marker of necrosis. Less than 3% of cells under any condition were positive for either Annexin V-PE or 7-AAD staining (Fig. 1E, right; n = 3).
α-Actinin-1 tyrosine phosphorylation is enhanced under elevated pressure conditions.
FAK-dependent phosphorylation of α-actinin affects its affinity for actin and regulates α-actinin accumulation at focal adhesion plaques (23, 32, 60). We therefore sought to assess the phosphorylation status of α-actinin-1 in suspended SW620 cells under ambient and increased pressure conditions. Because neither a phospho-specific α-actinin-1 antibody nor an α-actinin antibody sufficient for immunoprecipitation is commercially available (23), cell lysates were immunoprecipitated with a monoclonal antibody to phosphotyrosine and then probed for α-actinin-1. Tyrosine phosphorylation of α-actinin-1 in SW620 cells was increased by 64 ± 14% (n = 10; P < 0.01, Fig. 2A) in response to elevated pressure. Importantly, we observed a similar 49 ± 7% increase (n = 6; P < 0.04, Fig. 2A) in the amount of tyrosine-phosphorylated α-actinin-1 in primary human colon cancer cells isolated directly from surgical specimens.
Izaguirre et al. (23) reported that the Y12 residue on the amino terminus of α-actinin is the predominant site of phosphorylation in vanadate-treated adherent cells. To assess whether this was also true in cells exposed to increased extracellular pressure, we transfected SW620 cells with a His-tagged α-actinin construct containing a tyrosine-to-phenylalanine substitution (Y12F). After exposure of the cell transfectants to increased pressure, recombinant proteins were immunoprecipitated with an anti-His antibody and probed with antibodies for both α-actinin-1 and phosphotyrosine (Fig. 2B). Confirming our observations of native α-actinin-1, wild-type His-α-actinin cell transfectants displayed a 68 ± 7% pressure-stimulated increase (n = 6; P < 0.04) in α-actinin phosphorylation. In sharp contrast, the Y12F His-α-actinin mutants exhibited almost a complete block of phosphorylation. These results suggest that pressure-activated α-actinin-1 phosphorylation occurs primarily at Y12.
α-Actinin-1 Y12 phosphorylation is required for pressure-stimulated SW620 cell adhesion.
We next sought to determine whether the phosphorylation status of α-actinin-1 is important in its role in pressure-stimulated cell adhesion. SW620 cells were transfected with GFP-tagged wild-type α-actinin-1, GFP-Y12F mutant α-actinin-1, or GFP-vector alone. Plasmid transfection efficiency was typically ∼10% with either the wild-type or mutant constructs (Fig. 2E, left). Transfectants were isolated by fluorescence-activated cell-sorting (FACS) based on their GFP tag, and recombinant α-actinin-1 expression was determined by Western blotting. Both GFP-wild-type and -Y12F α-actinin-1 expression were approximately four times that of endogenous α-actinin-1 (Fig. 2C; n = 3). Next, we assessed the effect of pressure on the adhesion of each group of transfected cells (Fig. 2D). As expected, cells transfected with GFP-vector alone displayed a 41 ± 7% increase (n = 4; P < 0.05) in adhesion to collagen I following exposure to pressure. Cells expressing GFP-wild-type α-actinin-1 similarly exhibited a 38 ± 8% increase (n = 4; P < 0.05) in pressure-stimulated adhesion, whereas basal levels decreased by 10 ± 3% (n = 4; differences not statistically significant). In contrast, the pressure effect was completely blocked in the GFP-Y12F α-actinin-1 transfectants, and basal cell adhesion was diminished by 25 ± 5% (n = 4; P < 0.05) compared with GFP-vector controls (Fig. 2D). The effect of recombinant GFP-wild-type and -Y12F α-actinin-1 expression on SW620 cell viability was assessed as before by staining with Annexin V-PE and 7-AAD after exposure to ambient or increased pressure conditions. Similarly, <4% of cells under any condition displayed signs of apoptosis and/or necrosis (Fig. 2E, right; n = 3).
α-Actinin-1 interaction with FAK and Src is increased under elevated pressure conditions and is Y12 phosphorylation- dependent.
Previous work has established the functional necessity of FAK and Src in the pressure-stimulated adhesion response (50). We therefore sought to assess the relationship between these molecules and α-actinin-1 under elevated pressure conditions. To determine whether α-actinin-1 phosphorylation in response to pressure requires FAK, we transfected SW620 cells with FAK-specific siRNA (Fig. 3A). A 60 ± 4% reduction in total FAK protein completely inhibited pressure-induced α-actinin-1 phosphorylation (Fig. 3B; n = 6; P < 0.04). Next, we assessed α-actinin-1 interaction with both FAK and Src. SW620 cells were exposed to either ambient or increased pressure conditions for 30 min, at which time the cells were lysed and the protein lysates were used for the coimmunoprecipitation of α-actinin-1 with FAK and Src. Under elevated pressure conditions, we found an 81 ± 17% increase (n = 6; P < 0.04) in α-actinin-1 association with FAK compared with ambient pressure conditions (Fig. 3C). Similarly, we observed a 62 ± 8% increase (n = 6; P < 0.04) in α-actinin-1 coimmunoprecipitation with Src in response to increased extracellular pressure (Fig. 3E).
α-Actinin-1 phosphorylation appears to play a regulatory role in pressure-stimulated cell adhesion. Since α-actinin phosphorylation has been reported to reduce α-actinin localization to focal adhesions and along actin filaments, we postulated that the functional necessity of Y12 phosphorylation in our pressure model might reflect augmentation of protein-protein interactions that require phosphorylation at this site. We therefore sought to assess the ability of GFP-Y12F α-actinin-1 to associate with FAK (Fig. 3D) and Src (Fig. 3F) under elevated pressure conditions compared with GFP-wild-type α-actinin-1. Consistent with our previous observations after immunoprecipitation of FAK, coimmunoprecipitation of GFP-wild-type α-actinin-1 with FAK increased by 34 ± 9% (n = 6; P < 0.04) in lysates from cells exposed to increased pressure. However, the α-actinin-1 Y12F substitution reduced basal interaction by 17 ± 12% and blocked the pressure-induced association between α-actinin and FAK (n = 6; differences not significant). Likewise, coimmunoprecipitation of GFP-wild-type α-actinin-1 with Src increased by 36 ± 7% (n = 6; P < 0.04) under elevated pressure conditions. Strikingly, the α-actinin-1 Y12F substitution reduced basal interaction by 63 ± 14% and blocked the pressure-stimulated association between α-actinin-1 and Src (n = 6; differences not significant).
SiRNA-dependent silencing of α-actinin-1 inhibits pressure-stimulated phosphorylation of FAK, but not Src.
We have previously reported that exposure of SW620 cells to increased pressure stimulates the phosphorylation of FAK at Y397 and Y576, and of Src at Y416, and that both FAK and Src activation are required for pressure-stimulated adhesion (50). Therefore, we now sought to determine the impact of siRNA-dependent silencing of α-actinin-1 on FAK Y397/Y576 and Src Y416 phosphorylation (Fig. 4, A–C). SiNT-treated SW620 cells displayed a 32 ± 6% increase (n = 10; P < 0.01) in FAK Y397 phosphorylation under elevated pressure conditions. Transfection with siACTN1 caused a slight, but not statistically significant, increase in basal Y397 phosphorylation. However, exposure to pressure did not elicit any further increase in phosphorylation (n = 10). Similar results were obtained for FAK phosphorylation at Y576. SiNT-treated cells displayed a 23 ± 4% increase (n = 10; P < 0.01) in pressure-mediated induction of FAK Y576 phosphorylation compared with ambient pressure controls. α-Actinin-1 siRNA minimally affected basal Y576 phosphorylation but blocked any increase in Y576 phosphorylation due to pressure (n = 10). Pressure-induced Src Y416 phosphorylation in SW620 cell transfectants was similarly increased by 28 ± 7% in the siNT-transfected cells (n = 10; P < 0.03). However, in contrast to observations on FAK phosphorylation, Src Y416 phosphorylation was insensitive to the reduction in α-actinin-1 expression brought about by siACTN1 (29 ± 8% increase in adhesion; n = 10; P < 0.03).
Expression of Y12F α-actinin-1 inhibits pressure-stimulated phosphorylation of FAK Y397 and Y576.
We sought to determine whether Y12 phosphorylation-dependent association of α-actinin-1 with FAK is necessary for pressure-induced FAK phosphorylation. SW620 cells were cotransfected with plasmids containing HA-tagged FAK and either wild-type or Y12F α-actinin-1. After exposure to increased extracellular pressure, HA-FAK was immunoprecipitated with an anti-HA antibody and probed for FAK Y397 and Y576 phosphorylation by Western blotting (Fig. 4, D and E). Cells cotransfected with wild-type α-actinin-1 displayed 26 ± 5 and 30 ± 6% increases in FAK Y397 and Y576 phosphorylation, respectively (n = 6; P < 0.04 for each). In contrast, cotransfection with Y12F α-actinin-1 reduced basal FAK phosphorylation at Y397 and Y576 by 17 ± 6 and 22 ± 13%, respectively, and blocked the pressure effect. Thus FAK phosphorylation of α-actinin-1 at Y12 may serve to stabilize or potentiate FAK phosphorylation at Y397 and Y576 under pressure.
α-Actinin-1 interaction with β1-integrin is increased under elevated pressure conditions and is necessary for pressure-mediated colocalization of Src.
β1-Integrin has previously been established as an essential element in pressure-mediated cell adhesion and is known to directly associate with α-actinin-1 (4, 39). Therefore, we sought to assess the interaction of α-actinin-1 with β1-integrin after exposure to pressure. We observed a 44 ± 9% (n = 9; P < 0.03) increase in α-actinin-1 coimmunoprecipitation with β1-integrin in response to increased extracellular pressure (Fig. 5A). These results suggest that pressure induces α-actinin-1 localization to β1-integrin-associated focal adhesion complexes. Similarly, we observed a 48 ± 7% increase (n = 6; P < 0.04) in FAK and a 42 ± 5% increase (n = 6; P < 0.04) in Src coimmunoprecipitation with β1-integrin under elevated pressure conditions. The necessity of α-actinin-1 in the association of FAK and Src with β1-integrin was assessed by siRNA-mediated reduction of α-actinin-1. As expected, after cell transfection with siACTN1, basal α-actinin-1 association with β1-integrin was reduced by 51 ± 5% (n = 6; P < 0.04). Pressure-induced FAK interaction with β1-integrin was unaffected by α-actinin-1 reduction (n = 6; differences not significant), whereas Src pressure-stimulated interaction with β1-integrin was reduced by 64 ± 6% (n = 6; P < 0.05) in cells treated with α-actinin-1 siRNA. These results suggest α-actinin-1 may be a crucial scaffolding protein for the pressure-mediated recruitment of Src to β1-integrin-associated focal adhesion complexes.
To further delineate the mechanism by which mutant Y12F α-actinin-1 expression is able to disrupt pressure-stimulated cell adhesion, we additionally assessed competitive interaction of recombinant Y12F-α-actinin-1 versus endogenous α-actinin-1 with β1-integrin. A homogenous population of FACS-sorted GFP-tagged Y12F α-actinin-1 transfectants were exposed to either ambient or increased pressure conditions, and cell lysates were evaluated for β1-integrin coimmunoprecipitation with GFP-α-actinin-1 (Y12F) versus endogenous α-actinin-1 (Fig. 5C, right). GFP-α-actinin-1 (Y12F) association with β1-integrin was 21 ± 8% higher (n = 3) than that of endogenous α-actinin-1 under ambient conditions and increased by 12 ± 9% (n = 3) under elevated pressure. Compared with the typical 40 ± 7% increase (n = 3; P < 0.05) in β1-integrin-α-actinin-1 interaction exhibited by untransfected cells under pressure (Fig. 5C, left), β1-integrin association with endogenous α-actinin-1 in the GFP-α-actinin-1 (Y12F)-transfected cells was significantly decreased by 64 ± 13% (n = 3; P < 0.05) under elevated pressure. These results suggest that overexpression of mutant α-actinin-1 may block the function of the endogenous protein by competitively inhibiting its pressure-mediated interaction with β1-integrin.
SiRNA-mediated silencing of α-actinin-1 disrupts pressure-induced FAK-Src interaction.
FAK and Src regulate the turnover of focal adhesions and promote cell motility (57). In normal cells, multiple integrin-regulated linkages exist to activate FAK or Src (34). We therefore investigated the effect of α-actinin-1 reduction on Src coimmunoprecipitation with FAK. SiNT-treated cells displayed a 36 ± 4% increase (n = 6; P < 0.04) in FAK-Src interaction under elevated pressure conditions (Fig. 5B). α-Actinin-1 reduction by siRNA disrupted pressure-mediated formation of the FAK-Src complex (n = 6). These findings are further supportive of a model in which α-actinin-1 acts as an integrin-associated adapter to regulate the association of FAK and Src under pressure.
α-Actinin-1, FAK, and Src localize to the cell membrane under elevated pressure.
Finally, we sought to assess pressure-mediated changes in α-actinin-1 intracellular localization to further clarify the location of α-actinin-1, β1-integrin, FAK, and Src interaction. Fluorescence confocal microscopy was used to depict changes in spatial distribution of GFP-α-actinin-1 (Fig. 6A). Under ambient conditions, GFP-α-actinin-1 appears more diffuse and is evenly distributed between the cytoplasm and membrane. However, GFP-α-actinin-1 distributes more heavily along the cell membrane following exposure to increased pressure. Because of the limited resolution attainable using suspended cells, changes in FAK and Src localization were too subtle to visually quantify. We therefore assessed pressure-mediated changes in α-actinin-1, FAK, and Src subcellular location by cell fractionation. Soluble and particulate cell fractions were separated by high-speed centrifugation, and membrane fractions were evaluated for total amounts of the respective proteins (Fig. 6B). Elevated pressure caused a 31 ± 7% increase (n = 4; P < 0.03) in α-actinin-1, a 24 ± 5% increase (n = 4; P < 0.05) in FAK and a 28 ± 9% increase (n = 4; P < 0.04) in Src in the membrane fraction. Because cytoskeletal proteins typically were included in the membrane fraction, α-actinin-1 association with β-actin was additionally assessed (Fig. 6C). Increased pressure caused a 60 ± 8% decrease (n = 6; P < 0.04) in α-actinin-1-β-actin association, suggesting α-actinin-1 localization to the membrane is independent of β-actin. Next, because of the ubiquitous distribution of β1-integrin in both the cell cytoplasm and membrane, after cell fractionation we evaluated β1-integrin interaction with α-actinin-1 in both the cytosolic and membrane fractions (Fig. 6D). Although a greater quantity of β1-integrin was found in the membrane cellular fraction, the relative amount of α-actinin-1 associated with β1-integrin between fractions was similar under ambient conditions. Interestingly, only the membrane fraction displayed a significant increase (47 ± 9%; n = 4; P < 0.02) in β1-integrin-α-actinin-1 interaction under elevated pressure. In summary, these findings further support the role of α-actinin-1 in pressure-stimulated mechanotransduction as a pivotal membrane-associated focal adhesion complex adapter protein.
The stimulation of cancer cell adhesion by extracellular pressure is of potential physiological relevance to our understanding of tumor metastasis and is of scientific interest as a paradigm for mechanotransduction and inside-out signaling. Our data indicate that α-actinin-1, but not α-actinin-4, is a critical regulator of cell adhesion and a coupler of β1-integrin-FAK-Src complexes in both colon cancer cell lines and in primary human colon cancer cells exposed to pressure. In addition, the phosphorylation of α-actinin-1 at Y12 in response to extracellular pressure appears to facilitate the association of phosphorylated α-actinin-1 with both FAK and Src and potentiates subsequent focal adhesion complex signaling and regulation of integrin-mediated adhesion.
Although the magnitude of pressure-induced protein phosphorylation and protein-protein interactions in these studies may appear relatively modest, the effects are highly statistically significant. We (46, 53, 56) and others (6, 44, 45) have described pathologically relevant changes in cell adhesion, migration, and proliferation due to signal inductions of similar magnitude both in vitro and in vivo. The effect of extracellular pressure on murine colon cancer cell adhesion results in increased adhesion to murine surgical wounds that is not only statistically significant (53) but also of an order of magnitude that yields important differences in tumor-free survival (52).
α-Actinin has previously been implicated in the regulation of focal adhesions in other cells under ambient pressure. Using chromophore-assisted laser inactivation of GFP-α-actinin, Rajfur et al. (43) reported that loss of α-actinin disrupted the link between actin filaments and focal adhesion-associated integrins and subsequently caused stress fiber retraction. Laukaitis et al. (27) additionally reported that focal adhesions devoid of α-actinin were less stable and turned over more rapidly in Chinese hamster ovary cells migrating on fibronectin. We used an alternative siRNA-mediated approach to modulate the expression of α-actinin. Our observations suggest that the physical presence of a phosphorylation-competent α-actinin-1 within the focal adhesion complex is required for pressure-induced modulation of integrin adhesion receptors. Interestingly, reduction of α-actinin-1 only minimally affected basal adhesion to collagen I, suggesting that the availability of phosphorylated α-actinin-1 becomes rate limiting only when cells are exposed to collagen I under pressure. Likewise, this idea may explain why only a partial knockdown of α-actinin-1 is capable of completely inhibiting pressure-stimulated adhesion and associated downstream signaling. Ingber's “cellular tensegrity” model, in which physical distortion of the cell cytoskeleton is proposed to transfer mechanical loads through actin-associated molecules and initiate intracellular signaling, is consistent with the possibility that α-actinin-1 is at least initially required for a continuous link and mechanotransduction between the actin cytoskeleton and integrins (21, 22). This concept is further supported by our previous observation that pharmacological perturbation of the cell cytoskeleton inhibits pressure-stimulated cell adhesion and FAK activation (51).
Apart from the structural role of α-actinin-1 as a tethering molecule, its ability to function as a scaffolding protein may be paramount in its participation in pressure-induced signaling. Diverse signaling molecules including phosphatidylinositol 3-kinase (PI3K), ERK, protein kinase N (PKN), and MEKK1 have been reported to interact with α-actinin, emphasizing the potential importance of this molecule in diverse cell functions (7, 29, 36, 48).
After exposure to increased extracellular pressure, SW620 cells and primary human colon cancer cells displayed a significant increase in α-actinin-1 tyrosine phosphorylation, whereas expression of the Y12F α-actinin-1 mutant in these cells inhibited the pressure-induced cell adhesion. Both lines of evidence suggest that α-actinin-1 Y12 phosphorylation is an important positive regulatory event in pressure-activated cell signaling. These observations seem to contrast with a previous study of adherent fibroblasts, under ambient conditions, deficient in the FAK phosphatase SHP-2. In such fibroblasts, excessive FAK-mediated α-actinin phosphorylation prevented the recruitment of α-actinin to focal adhesions and subsequent focal adhesion maturation (55). The apparent discrepancy may reflect differences between epithelial cells and fibroblasts, implications of various levels of α-actinin-1 Y12 phosphorylation, distinctions between focal adhesion assembly in suspended cells beginning to adhere and those already adherent, or the consequences of extracellular pressure itself.
The pressure to which the cells are subjected may also differentially affect the relationship between FAK and α-actinin-1. Whereas initial studies determined that α-actinin-1 is a FAK substrate (23), recent data (60) and our present data suggest that α-actinin-1 may also affect the state of FAK activation under both ambient and pressure conditions. Coexpression of wild-type α-actinin-1 and the phosphatase PTP 1B in COS-7 cells under ambient pressure disrupted FAK-Src interaction. Overexpression of Y12F α-actinin-1 mutant in the same cells did not affect FAK-Src interaction. In marked contrast, in the present study we have shown that exposure of SW620 cells to increased pressure enhanced protein-protein interactions between α-actinin-1 and FAK, whereas reduction of α-actinin-1 by specific siRNA blocked pressure-induced FAK phosphorylation. Furthermore, the Y12F mutant prevented pressure-stimulated α-actinin-1 association with FAK and Src as well as FAK Y397 and Y576 phosphorylation. It is therefore possible that the scaffolding function of α-actinin-1 is differentially regulated under ambient pressure compared with increased pressure conditions, which in turn may determine the signaling components recruited to focal adhesions.
Izaguirre et al. (23) previously established that the Y12F mutant α-actinin used in this study binds actin with affinity similar to that of wild-type unphosphorylated α-actinin. The inhibition of pressure-stimulated cell adhesion by transfection with the Y12F mutant therefore suggests that α-actinin-1 Y12 phosphorylation is the determining factor in this process. The observed reduction in α-actinin-1 interaction with β-actin under pressure further supports the concept that the Y12-dependent scaffolding function of α-actinin-1 is independent of its association with actin. However, it still remains plausible that actin tethering to β1-integrin via α-actinin-1 is crucial for the initial relay of mechanical signals. The subsequent reduction in total α-actinin-1-actin interaction may be exemplative of focal adhesion remodeling as well as the redistribution of α-actinin-1 molecules participating in actin cross-linking. Three actin-binding sites have been identified over two tandem calponin homology domains on the amino terminus of α-actinin, of which only the first is affected by phosphorylation (3, 25, 38). Parallel studies have suggested the two remaining actin-binding sites may be regulated through phosphoinositide binding (12–14). Thus it is possible that these latter two actin-binding sites are adequate for actin association and integrin-cytoskeleton tethering independent of the phosphorylation-sensitive Y12 region and may account for the residual α-actinin-1-actin association observed under elevated pressure.
FAK Y397 phosphorylation is known to facilitate FAK interaction with Src, which in turn mediates the phosphorylation of FAK at Y407, Y576/577, Y861, and Y925 (33, 41). Phosphorylation of these residues maximizes FAK kinase activity and creates additional binding sites for several signaling and adapter proteins, including paxillin, which is also required for the effect of pressure on adhesion (18, 54). An active FAK-Src complex may drive cell proliferation, motility, and focal adhesion formation and turnover (34). However, less is known about how the formation of this complex is regulated. Cell exposure to increased pressure significantly enhanced FAK and Src interaction, suggesting a regulatory role for this complex in governing pressure-stimulated cell adhesion. Although phosphorylated α-actinin-1 could form a ternary complex with FAK and Src, the β1-integrin coimmunoprecipitation data suggest that integrin-bound phosphorylated α-actinin-1 primarily serves as an adapter protein for Src. The α-actinin-1-dependent recruitment of Src to FAK-associated focal adhesions enhances FAK-Src interaction and, consequently, FAK activation. The observation that expression of Y12F α-actinin-1 in SW620 cells impaired the recruitment of Src to β1-integrin and, consequently, the formation of FAK-Src complexes under increased pressure conditions supports this model.
In summary, pressure-stimulated colon cancer cell adhesion to collagen I not only requires α-actinin-1 but seems critically dependent on α-actinin-1 Y12 phosphorylation. We did not investigate cell-cell adhesion in this study. However, the same pressure-activated pathway appears to stimulate colon cancer cell adhesion to other matrix proteins, to endothelial cells, and to surgical wounds in a murine model (50, 53). These findings therefore suggest the possibility that α-actinin-1 may prove a useful therapeutic target in the inhibition of tumor cell implantation and metastasis. In addition, this study further elucidates the coregulatory relationship among α-actinin-1, FAK, and Src and identifies α-actinin-1 as a critical link between focal adhesion complex and cytoskeletal signaling in the mechanosensory pathway modulating cellular response to fluctuations in extracellular pressure.
This work was supported in part by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01 DK06771 (to M. D. Basson) and a Veterans Affairs Merit Review (to M. D. Basson).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society