![]()
|
|
||||||||
J. Biol. Chem., Vol. 279, Issue 13, 12001-12004, March 26, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||




¶
||**
From the
Cardiovascular Research Center and ||Departments of Microbiology and Biomedical Engineering and Mellon Prostate Cancer Institute, University of Virginia, Charlottesville, Virginia 22908 and ¶Department of Cell Biology, Scripps Research Institute, La Jolla, California 99037
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
Integrins, as the main receptors that connect the cytoskeleton to the extracellular matrix (ECM),1 have an intimate relationship to force. Integrins transmit mechanical stresses across the plasma membrane; thus, tractional forces developed in the cytoskeleton are conveyed to the ECM through the integrins, as are stresses applied to cells from the ECM (6). Because integrins also regulate signaling pathways (7), they are positioned to transduce physical forces into chemical signals as well.
Current evidence makes a strong case for a bidirectional relationship between integrin-mediated adhesion and mechanical forces. Cells exert tractional forces on the substratum to which they are attached through adhesive structures such as focal adhesions (FA) and focal complexes (FC) (8, 9). Conversely, formation of these structures requires tension generated by actomyosin within the cell and resisted by a substratum of sufficient stiffness (10, 11). These regulatory arrows, pointed in opposite directions, create opportunities for regulatory loops that may guide migration, matrix assembly, and tissue organization.
Our goal is to review what is known about integrins in mechanotransduction. We will consider formation of normal adhesions as a mechanically sensitive system, the responses of cells to externally applied strains, and the responses of cell to fluid flow. Integrins are strongly implicated in all three; thus, possible connections between these systems will be explored and a unifying hypothesis offered.
| Force and Adhesion Formation |
|---|
|
|
|---|
FC are precursors to FA (18), which are larger, elongated structures that form in response to Rho activation. They require increases in cellular contractility (10) mediated by the Rho effector Rho kinase, which induces increased phosphorylation of myosin light chain. The increase in force at adhesion sites results in an increase in the density of integrins, the recruitment of adhesion structural proteins, and the elongation of the adhesion site in the direction of force (16, 19, 20). Rho activation and increased contractility also induces assembly of fibrillar adhesions and fibronectin matrix (2123). Matrix assembly in response to force may therefore contribute to increased integrin density and signaling. Indeed, a ligand density of 1 fmol of RGD ligands/cm2 supported cell spreading, but FA formation required 10-fold more (24), suggesting that tension-dependent matrix assembly may be important. FA produce forces in the neighborhood of a few nanonewtons/µm2 (16, 25) and require surfaces that are sufficiently stiff to support this tension (11). Thus, FA as well as FC both transmit and require tension. Importantly, outcomes are similar whether forces are generated by the cell or applied externally (12, 20).
FA remodeling also occurs in a tension-dependent manner (19, 23, 26, 27). Under some conditions, FA proteins labeled with green fluorescent protein appear to "slide" centripetally, in the direction of tension. This effect is due to increased assembly at the front and increased disassembly at the rear of each adhesion and appears to be tension-dependent because it is prevented by inhibitors of contraction (19). FA are generally linked by actin stress fibers that are in a state of balanced isometric contraction; thus, tension pulls FA toward the center of the cell. In stationary cells, all adhesions consequently move toward the cell center. However, in polarized, migrating cells, adhesions at the front of the cell remain fixed, whereas adhesions at the rear of the cell tend to slide (26), indicating the existence of additional mechanisms that modulate this process.
Cell proliferation has long been recognized to require cell spreading and exertion of force on the ECM, but the mechanisms that mediate effects on growth have been poorly understood (6, 28). Recent work has shown that integrin clustering in FA in response to force generation is required for sustained MAP kinase activation in response to growth factors (29). Artificially clustering integrins with antibodies relieves the requirement for actomyosin-dependent tension, which strongly suggests that the FA itself is the key site of mechanotransduction.
Because migration requires that cells exert directional forces on their surroundings, it is not surprising that mechanical forces also regulate cell migration. Cells grown on flexible substrates show enhanced rates of cell migration compared with rigid substrates (11). On matrices that have boundaries between regions of different elasticities, cells migrate toward areas of higher substrate rigidity, a process termed durotaxis (30). This effect requires expression of focal adhesion kinase (FAK) suggesting that this molecule is involved in sensing forces (31). FAK is well known to respond to integrin clustering and FA formation by autophosphorylation and activation (32). Interestingly, FAK phosphorylation on a specific site, tyrosine 397, is sensitive to the tethering of integrins to a rigid substratum, whereas clustering alone regulates phosphorylation of other sites (33). As FAK is implicated in the regulation of both Rho (34) and Rac (35), FAK activity might control cell movement in durotaxis. When binding of proteins from cell lysates to isolated, detergent-insoluble cytoskeletons was assayed, stretching of the cytoskeletons increased binding of FAK, paxillin, and p130cas (36). Thus, FAK appears to be a key component of the mechanotransduction apparatus, and both FAK and its interacting proteins are candidate transducers. However, the functional importance of FAK tyrosine 397 phosphorylation in durotaxis is unclear as durotaxis was seen in FAK/ cells reconstituted with nonphosphorylatable Y397F FAK (31).
Tyrosine phosphatases have also been implicated in force-dependent adhesion reinforcement. Adhesion reinforcement in response to a restraining force applied to fibronectin-coated beads was inhibited by the tyrosine phosphatase inhibitor phenyl arsine oxide (37). Adhesion reinforcement through integrin
v
3 involves the sequential activation of receptor protein phosphatase
and Fyn (38). The phosphatase activity of SHP-2 enhances the focal complex strengthening by inhibiting FAK-induced phosphorylation and inactivation of the cytoskeletal linker
-actinin (38).
Taken together, the above discussion suggests two likely mechanisms for mechanotransduction within focal adhesions. One is that force applied to integrins results in increased clustering, perhaps as a result of increased actin and myosin recruitment and cytoskeletal assembly (10). Increased fibronectin fibril assembly in response to tension could increase the density of integrin binding sites and therefore contribute to increased receptor density. This mechanism would give rise to increased integrin signaling and is also consistent with the key role of integrin clustering in growth control (29). A second possibility is that tension alters the conformation of certain force-sensitive components of the FA to induce new binding interactions or direct modulation of enzymatic activity. A comparable effect has been demonstrated for fibronectin, providing one mechanism for force-dependent matrix assembly (21). These conformation changes would result in phosphatase-dependent activation of tyrosine kinases such as FAK to mediate mechanotransduction. Identification of such force-dependent interactions is therefore an important goal. These two mechanisms are not mutually exclusive and could function cooperatively.
| Responses of Cells to Externally Applied Strain |
|---|
|
|
|---|
Further evidence for the involvement of integrins comes from effects of strain on focal adhesion signaling. Cyclic strain induces tyrosine phosphorylation of the focal adhesion proteins FAK, paxillin, and p130cas (41, 42), events that appear to be mediated by stretch-induced activation of c-Src (42). The phosphorylation of FAK is critical to stretch-induced extracellular signal-regulated kinase (ERK) and p38 activation (43, 44). Thus, an important subset of mechanotransduction events occurs within focal adhesions.
Some responses to strain depend on the specific ECM to which cells are adhered (Table I). Static biaxial stretch of cardiac fibroblasts activated the MAP kinase pathways, ERK2 and c-Jun N-terminal kinase (JNK1) in an ECM-specific manner (45). ERK2 was stimulated only in cells plated on fibronectin; JNK1 was activated on fibronectin, vitronectin, or laminin; and cells on collagen did not activate any of these MAP kinases. In cardiac myocytes, mechanical stretch also activated p38 MAP kinase (39), and this effect is mediated by a pathway involving integrins, FAK, Src, and Ras (44). In vascular smooth muscle cells, cyclic strain stimulated mitogenesis through integrin
v
3-dependent release of platelet-derived growth factor (46). Cyclic stretch of vascular smooth muscle cells increased apoptosis of cells on collagen I, but not on other matrixes, and this effect was mediated by a pathway involving Rac, p38 MAP kinase, and p53 (40). Smooth muscle contraction was inhibited by brief pretreatment with RGD peptides that bind
v
3 and
5
1 integrins (47). Whether these peptides disrupt existing integrin-ligand bonds in the short time of treatment, act as positive stimulators of integrin signals, or interfere with integrin dynamics is unclear. In many cases, these effects are consistent with known specific signaling properties of individual integrins (7). The ECM/integrin specificity of these responses provides some of the best evidence that integrins participate in mechanotransduction.
|
The simplest explanation for the involvement of integrins in cellular responses to strain is that externally applied deformations modify the balance of forces within focal adhesions and the cytoskeleton. Thus, mechanisms of transduction involving changes in integrin clustering or opening of cryptic sites in FA molecules are similar to events during normal cell adhesion, spreading, and migration. This hypothesis may explain the wide distribution of mechanical sensitivity among many cell types. For instance, the effects of tension on directional FA remodeling (19, 26) may contribute to the alignment of cells in response to strain. It is likely, however, that additional, perhaps cell type-specific mechanisms mediate some responses.
| Responses to Fluid Shear Stress |
|---|
|
|
|---|
|
v
3 and
5
1 to their ligands block the shear-induced vasodilation of coronary arteries in vivo (53). One appealing aspect of this model is that it avoids difficulties posed by the endothelial glycocalyx. The endothelium has an apical glycocalyx a few hundred microns thick that forms a barrier to fluid flow (54, 55). Thus, the apical surface per se experiences negligible shear stress. The proteins that anchor the carbohydrate chains experience shear and therefore could conceivably be mechanosensors, although there are no data to support this idea. Alternatively, surface glycoproteins could transmit forces through the cytoskeleton to integrins or cell-cell adhesion receptors that serve as anchorage points that resist the flow. Thus, the notion that integrins or junctional receptors are shear stress sensors is consistent with the properties of the glycocalyx.
Shear stress stimulates many of the pathways that are regulated by integrin binding to ECM. Both
1-activating antibody and shear stress elicit similar endothelial cell activation of MAP kinase (56) and similar changes in Rho activity (57). Shear stress triggers activation of FAK and c-Src (58, 59), tyrosine phosphorylation of p130cas, and its association with Crk (60), all of which are characteristic of integrin signaling. In several cases, integrin antagonists block shear stress signaling. Antibody to
v
3 blocked shear stress activation of ERK, JNK, and NF
B (61, 62), whereas an anti-
1-blocking antibody prevented activation of sterol regulatory element-binding proteins by shear stress (63). Blocking integrins with an RGD peptide abolishes the antiapoptotic effect of shear stress (64) and the shear stress-induced secretion of basic fibroblast growth factor (65). Shear stress triggered association of integrins with Shc, and this interaction was ECM-specific in that only those integrins that could bind to the available matrix interacted with Shc (66).
Brief treatment of endothelial monolayers with antibodies to the relevant ECM proteins that inhibited new integrin binding to the matrix without disrupting existing adhesions prevented Shc association with integrins (66). This result suggested that dynamic interactions between integrins and the ECM contribute to shear stress signaling and that changes in integrin activation or ligand binding could be involved. Increased binding of the monovalent Fab fragment, WOW-1, that reacts selectively with activated
v
3, directly demonstrated a rapid increase in integrin affinity following shear (57). At later times, endothelial monolayers showed increased staining with antibodies that recognize the ligand-occupied conformations of
1 and
3 integrins. Blocking this de novo occupancy inhibited several downstream responses, including regulation of Rho, Rac, and Cdc42 (57, 67, 68) and the above mentioned interaction of integrins with Shc (66). The Rho family GTPases in turn mediated alignment of the actin and tubulin cytoskeletons and activation of NF
B and its target genes. Integrin activation and downstream signaling is therefore central to the endothelial-specific shear stress response.
Although the apical membrane of ECs is directly exposed to flow, WOW-1 immunostaining was seen only at the basal side of the sheared ECs (57), indicating selective activation of integrins on the abluminal membrane. The restriction of activation to the basal surface may be a general feature of EC behavior rather than a specific consequence of shear stress signaling, as active integrins on the luminal surface of endothelia would be potentially thrombogenic. It is noteworthy, however, that neither
v
3 activation nor the initial increase in integrin occupancy seen by antibody staining showed indications of directionality. By contrast, remodeling of FA by shear is biased for assembly toward the downstream side relative to the direction of flow (52). Activation of Rac and Cdc42 downstream of integrins also shows polarization toward the downstream edge of the cell (67, 68). Flow may therefore affect FA remodeling in a directional manner that is distinct from the overall increases in affinity and ligand binding.
The actual forces exerted on cells by fluid flow in these experiments are low compared with the FA traction forces. If forces were distributed evenly, typical experimental shear stresses of 10 dynes/cm2 (or 1 pN/µm2) are 10005000 times lower than typical traction forces (15 nN/µm2). Cell-matrix adhesions occupy only a fraction (typically on the order of 10%) of the total ventral surface, so if forces are efficiently transmitted, FA would feel
10x larger stresses. Additionally, deformations of the cytoskeleton due to fluid flow are not distributed evenly throughout the cell; small regions show higher than average displacement, indicating the existence of "strain focusing" in which certain regions absorb more of the applied force (69). But even taking these factors into account, flow would exert at most a few percent of the tractional tension from actomyosin. These considerations make it less likely that FA themselves are the sites where shear forces that trigger activation of integrins are initially transduced.
| Hypothesis and Conclusions |
|---|
|
|
|---|
For fluid shear stress, we hypothesize that integrins may be involved in two distinct ways. Integrins in endothelial cells appear to be activated by fluid flow, resulting in a spatially uniform increase in cell-ECM adhesion (57, 66). By contrast, FA dynamics in endothelial cells under shear are directional, preferentially remodeling in the downstream direction (52). As discussed above, FA show tension-dependent remodeling in which they appear to slide toward the center of the cell. Mogilner and colleagues2 hypothesize that even though the total force of shear may be relatively low, this force would add vectorially to the centripetal tension from the actin stress fibers. Thus shear stress would decrease tension at the downstream side of the cell and increase it on the upstream side. Shear may therefore create a significant bias, altering net force balance between upstream and downstream, resulting in directional remodeling of FA through the normal tension-dependent mechanisms. This effect would be distinct from but superimposed on the activation-ligation sequence described above.
Fundamental questions about mechanotransduction within integrin-mediated adhesions remain unresolved. The detailed mechanisms by which forces are converted to chemical signals and cytoskeletal rearrangements are subjects of intense interest. Whatever the answers, we suggest that continued study of integrin dynamics and force-dependent effects are likely to shed light on diverse phenomena such as endothelial responses to shear stress, regulation of vascular tone, cardiac hypertrophy and failure, and regulation of growth and gene expression in many tissues.
| FOOTNOTES |
|---|
These three authors contributed equally and are listed in alphabetical order. ![]()
** To whom correspondence should be addressed. E-mail: maschwartz{at}virginia.edu.
1 The abbreviations used are: ECM, extracellular matrix; FA, focal adhesion(s); FAK, focal adhesion kinase; FC, focal complex(es); N, newton; MAP, mitogen-activated protein; ERK, extracellular signal-regulated kinase; JNK, c-Jun N-terminal kinase; EC, endothelial cell. ![]()
2 C. Civelekoglu-Scholey, A. W. Orr, J. Novak, J. J. Meister, M. A. Schwartz, and A. Mogilner, submitted for publication. ![]()
| ACKNOWLEDGMENTS |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
C. Fung, R. Lock, S. Gao, E. Salas, and J. Debnath Induction of Autophagy during Extracellular Matrix Detachment Promotes Cell Survival Mol. Biol. Cell, March 1, 2008; 19(3): 797 - 806. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. T. Parsons, J. Slack-Davis, R. Tilghman, and W. G. Roberts Focal Adhesion Kinase: Targeting Adhesion Signaling Pathways for Therapeutic Intervention Clin. Cancer Res., February 1, 2008; 14(3): 627 - 632. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Alessandri-Haber, O. A. Dina, E. K. Joseph, D. B. Reichling, and J. D. Levine Interaction of Transient Receptor Potential Vanilloid 4, Integrin, and Src Tyrosine Kinase in Mechanical Hyperalgesia J. Neurosci., January 30, 2008; 28(5): 1046 - 1057. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Schlunck, H. Han, T. Wecker, D. Kampik, T. Meyer-ter-Vehn, and F. Grehn Substrate Rigidity Modulates Cell Matrix Interactions and Protein Expression in Human Trabecular Meshwork Cells Invest. Ophthalmol. Vis. Sci., January 1, 2008; 49(1): 262 - 269. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. Zhang and S. J. Gunst Interactions of Airway Smooth Muscle Cells with Their Tissue Matrix: Implications for Contraction Proceedings of the ATS, January 1, 2008; 5(1): 32 - 39. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. H. Craig, B. Haimovich, and M. D. Basson {alpha}-Actinin-1 phosphorylation modulates pressure-induced colon cancer cell adhesion through regulation of focal adhesion kinase-Src interaction Am J Physiol Cell Physiol, December 1, 2007; 293(6): C1862 - C1874. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. Rizzo Lights, camera, actin! The cytoskeleton takes center stage in mechanotransduction. Focus on "Mapping the dynamics of shear stress-induced structural changes in endothelial cells." Am J Physiol Cell Physiol, December 1, 2007; 293(6): C1771 - C1772. [Full Text] [PDF] |
||||
![]() |
Y. Wang, L. M. McNamara, M. B. Schaffler, and S. Weinbaum A model for the role of integrins in flow induced mechanotransduction in osteocytes PNAS, October 2, 2007; 104(40): 15941 - 15946. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Balasubramanian, A. Ahmed, C.-M. Lo, J. S. K. Sham, and K.-P. Yip Integrin-mediated mechanotransduction in renal vascular smooth muscle cells: activation of calcium sparks Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2007; 293(4): R1586 - R1594. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Kato, K. Nakamura, H. Kudo, Y. H. Tran, Y. Yamamoto, H. Doi, and S. Hirose Characterization of the Column and Autocellular Junctions That Define the Vasculature of Gill Lamellae J. Histochem. Cytochem., September 1, 2007; 55(9): 941 - 953. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. F. Liu, C. M. Nelson, J. L. Tan, and C. S. Chen Cadherins, RhoA, and Rac1 Are Differentially Required for Stretch-Mediated Proliferation in Endothelial Versus Smooth Muscle Cells Circ. Res., August 31, 2007; 101(5): e44 - e52. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Seebach, G. Donnert, R. Kronstein, S. Werth, B. Wojciak-Stothard, D. Falzarano, C. Mrowietz, S. W. Hell, and H.-J. Schnittler Regulation of endothelial barrier function during flow-induced conversion to an arterial phenotype Cardiovasc Res, August 1, 2007; 75(3): 598 - 607. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. J. Zhang, G. A. Truskey, and W. E. Kraus Effect of cyclic stretch on beta1D-integrin expression and activation of FAK and RhoA Am J Physiol Cell Physiol, June 1, 2007; 292(6): C2057 - C2069. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Lamalice, F. Le Boeuf, and J. Huot Endothelial Cell Migration During Angiogenesis Circ. Res., March 30, 2007; 100(6): 782 - 794. [Abstract] [Full Text] [PDF] |
||||
![]() |
Q. Lin, D. Gourdon, C. Sun, N. Holten-Andersen, T. H. Anderson, J. H. Waite, and J. N. Israelachvili Adhesion mechanisms of the mussel foot proteins mfp-1 and mfp-3 PNAS, March 6, 2007; 104(10): 3782 - 3786. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Hu, L. Ji, K. T. Applegate, G. Danuser, and C. M. Waterman-Storer Differential Transmission of Actin Motion Within Focal Adhesions Science, January 5, 2007; 315(5808): 111 - 115. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Feng, X. Lu, J. C. Resendiz, and M. H. Kroll Pathological shear stress directly regulates platelet {alpha}IIbbeta3 signaling Am J Physiol Cell Physiol, December 1, 2006; 291(6): C1346 - C1354. [Abstract] [Full Text] [PDF] |
||||
![]() |
|