If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Molecular Biology, Princeton University, Princeton, New Jersey, USADepartment of Surgery, Robert Wood Johnson Medical School, New Brunswick, New Jersey, USA
Environmental signals from the extracellular matrix (ECM) are transmitted by cell surface receptors that connect to the actin cytoskeleton and to multiple intracellular signaling pathways. To dissect how the ECM regulates cell functions, we are using a three-dimensional (3D) fibrin–fibronectin matrix, resembling the wound provisional matrix. Fibroblasts adhere to fibronectin in this matrix via concomitant engagement of α5β1 integrin receptors and syndecan-4, a transmembrane proteoglycan. An adhesive phenotype is developed with actin stress fibers and activation of focal adhesion kinase (FAK) and Rho GTPase. Lack of syndecan-4 engagement, as occurs in the presence of the ECM protein tenascin-C, promotes a motile phenotype; FAK and Rho signaling are downregulated and filopodia are extended. Fibronectin matrices have distinct effects on two other receptors: α4β1 and αvβ3 integrins. Although α4β1 does not naturally support strong cell interactions with a fibrin–fibronectin matrix, binding is dramatically enhanced by proteolytic cleavage of fibronectin. Conversely, activity of αvβ3 is stimulated by multimeric fibronectin fibrils showing that the organization of fibronectin differentially affects integrin functions. Thus, deposition of additional ECM components, expression of co-receptors for ECM, cleavage of adhesive proteins, and the architecture of the ECM microenvironment are different mechanisms for modulating cell responses to fibronectin matrix.
2D
two-dimensional
3D
three dimensional
ECM
extracellular matrix
FAK
focal adhesion kinase
Introduction
Wound repair is a dynamic process requiring orchestrated cell movements, ECM deposition and degradation, and changes in cell signaling and gene expression. The adhesive ECM protein fibronectin is present throughout sites of tissue injury, being deposited initially during blood coagulation as a major component of the provisional matrix, then released by neutrophils and macrophages, and subsequently assembled by fibroblasts and endothelial cells to form granulation tissue (
). Modulation of fibronectin expression and deposition provides temporal and spatial cues to direct cell adhesive activities as repair progresses. Fibronectin mediates cell adhesion primarily through heterodimeric integrin receptors including α5β1, which binds to the arg-gly-asp (RGD) and adjacent sequences in the central cell binding domain and α4β1, which recognizes the CS1 site within the alternatively spliced V region (
), are also found in the wound bed and counterbalance the activities of fibronectin and other adhesive proteins in vitro. Tenascin-C is a large multidomain ECM protein with binding sites for cell receptors, fibronectin, and glycosaminoglycans. In cell culture, tenascin-C induces loss of focal adhesions and prevents cell adhesion and spreading on fibronectin (
Our work has focused on modulation of fibronectin interactions with cells throughout the repair process. We have developed three-dimensional (3D) matrix models that resemble the fibrin–fibronectin provisional matrix and the fibrillar fibronectin matrix found in granulation tissue. We are using these models to study the complex interplay between adhesion, adhesion modulation, and matrix turnover during repair and have identified specific molecular changes and intracellular signaling pathways that play critical regulatory roles. Some of these results are described in this review of our work.
Provisional Matrix Model
At sites of tissue damage, the blood coagulation cascade deposits a covalently crosslinked fibrin–fibronectin provisional matrix, which prevents further blood loss and acts as a framework for cell adhesion and migration during repair. Polymerization of the fibrin–fibronectin matrix can be recapitulated in vitro using purified components (
). To examine cell responses to a 3D fibrin–fibronectin matrix, we use immunofluorescence microscopy of cells plated on top of the fibrin–fibronectin matrix to determine protein localization and cell shape. Cell lysates are analyzed for changes in intracellular signaling. This 3D matrix also provides a system for dissecting the regulation of matrix contraction. Cells included during the polymerization reaction become embedded in the 3D matrix and cell contractility is measured by size changes of the matrix over time to give the % contraction (
) and is found near the wound edge where it contacts fibronectin and other provisional matrix components. Tenascin-C has been shown to antagonize fibronectin function in vitro (
), which prompted us to examine its effects on fibroblast morphology in response to a 3D fibrin–fibronectin matrix. We observed dramatic changes in the presence of tenascin-C. Cell spreading and actin stress fiber formation were inhibited; cells remained rounded and extended actin-rich filopodia. In our model system, tenascin-C regulation of Rho GTPase activity was responsible for these changes in the cytoskeleton. The Rho family of GTPases controls actin organization: activation of family members RhoA, Rac, and Cdc42 promotes stress fiber, lamellipodia, and filopodia formation, respectively (
). This implies that tenascin-C-mediated control of the cytoskeleton occurs via inhibition of stress fiber formation, as opposed to stimulation of filopodia or destabilization of actin, as has since been reported in other models (
). Tenascin-C also inhibited focal adhesion assembly in cells on a 3D fibrin-fibronectin matrix by suppressing the activation of focal adhesion kinase (FAK) (
The phenotype induced by tenascin-C is indicative of a migratory cell. Both FAK and RhoA play vital roles in cell migration, regulating the cell adhesion, traction, and contractility required for cell movement (
). By modulating their activities, tenascin-C may promote fibroblast migration from undamaged tissue at the wound edge into the wound bed. This hypothesis is supported by the phenotype of tenascin-C null mice, which although normal in development, viability, and fertility, show some severe defects in tissue repair. These mice exhibit compromised cell migration and proliferation in corneal wounds and in induced glomerulonephritis and dermatitis (
), we examined whether tenascin-C had any effect on matrix contraction. The inclusion of tenascin-C in a fibrin–fibronectin matrix inhibited fibroblast-mediated contraction due to defects in signaling via both RhoA GTPase and FAK-mediated pathways. Activation of either alone partially relieved the effects of tenascin-C, but to regain optimal contractile ability required activation of both FAK and RhoA simultaneously (
). These data suggest that tenascin-C expression in the wound bed up to the point of wound closure prevents premature wound contraction. Downregulation of tenascin-C expression and clearance from the wound would then allow contraction to proceed indicating that tenascin-C exerts temporal control over cell behavior during wound healing.
ECM Communication Through Cell Surface Receptors
Fibroblast binding to fibronectin within the 3D provisional matrix requires α5β1 integrins (
). Evidence that a heparan sulfate proteoglycan also participates in fibroblast interactions was obtained when we compromised heparan sulfate proteoglycan function by competition with soluble heparin or treatment of cells with heparitinase. Both treatments induced a phenotype similar to tenascin-C. α5β1 integrin acts together with the cell surface heparan sulfate proteoglycan syndecan-4 for cell spreading and signaling in fibroblasts on fibronectin-coated surfaces (
), but is transiently upregulated upon injury. Increased expression has been observed in the dermis adjacent to the injury, in migratory keratinocytes at the edges of the wound, as well as throughout the granulation tissue on endothelial cells and fibroblasts (
Syndecan-4-null mouse embryo fibroblasts were used to assess its role in fibroblast response to the 3D fibrin-fibronectin matrix. Null cells had a distinct morphology and impaired contraction of the provisional matrix compared to wild-type fibroblasts. Like wild-type cells cultured in the presence of tenascin-C, cells lacking syndecan-4 extended actin-rich filopodia (
). These data show that the coordinated interaction of fibronectin with syndecan-4 and α5β1 integrins is essential for effective cell response to the provisional matrix. The mechanisms of syndecan-4 action identified using model systems are consistent with results from mice with targeted deletion of syndecan-4. Like tenascin-C-null mice, syndecan-4-deficient mice develop normally but show defects in wound healing. Repair of excisional skin wounds is delayed in the first 7 days: re-epithelialization is retarded, there is reduced accumulation of granulation tissue, and no wound contraction (
In our experiments, cell phenotype in the presence of tenascin-C resembled that of cells with compromised syndecan-4 expression or function. Furthermore, tenascin-C binds to the HepII domain in fibronectin (
). These observations suggested that tenascin-C acts by inhibiting syndecan-4–fibronectin interactions. In support of this model, we found that, in contrast to wild-type fibroblasts, syndecan-4-null cells were able to spread and organize the actin cytoskeleton on a fibrin–fibronectin matrix containing tenascin-C. This shows that cell response to tenascin-C requires expression of syndecan-4 (
). Conversely, overexpression of syndecan-4 circumvented the effects of tenascin-C and restored fibronectin-mediated cell spreading, matrix contraction (
). Therefore, the effects of tenascin-C on cell functions result from blockade of fibronectin signaling through syndecan-4. During wound repair, changes in tenascin-C and/or syndecan-4 expression may act as a regulatory switch to modulate cell–fibronectin interactions and determine the levels of cell growth and matrix contraction.
These results suggest the following model to explain certain cell behaviors during the early stages of repair (Figure 1). Fibroblasts emigrate to the wound bed from surrounding tissue. In areas adjacent to the injury, cells encounter the fibrin–fibronectin provisional matrix (
). α5β1 integrin binds to fibronectin but the presence of tenascin-C induces a motile cell phenotype by blocking syndecan-4 binding to fibronectin, thus suppressing Rho GTPase and FAK activities (Figure 1a, left). These signals promote cell migration and prevent inappropriate matrix contraction at the extremes of the wound. As cells move further into the wound bed, tenascin-C levels are reduced. Fibronectin interactions with α5β1 integrin and syndecan-4 stimulate cell spreading and formation of stress fibers and focal adhesions via RhoA and FAK activation and promote enhanced cell contractility in order to close the wound (Figure 1b, left).
Figure 1Modulation of cell interactions with fibronectin matrices. (a) Adhesion, spreading, and contraction mediated by α5β1 integrin and syndecan-4 are reduced with deposition of tenascin-C into a 3D fibrin–fibronectin matrix (left). By blocking syndecan-4 binding to fibronectin, tenascin-C downregulates signaling through FAK and Rho GTPase. Cells expressing α4β1 integrin show reduced adhesion and are unable to spread on or contract a 3D fibrin–fibronectin matrix when the fibronectin is intact (right). (b) Conditions that support cell adhesion, spreading, and matrix contraction include cells that express α5β1 and syndecan-4 interacting with a 3D fibrin–fibronectin matrix (left) or α4β1-expressing cells interacting with 3D fibrin-fragmented fibronectin matrix (right). Dashed arrows indicate multiple steps between receptors and cell activities.
α4β1 Integrin Interactions with Fibronectin in a Provisional Matrix
Many blood cells express the α4β1 integrin receptor for fibronectin. For example, α4β1-positive neutrophils rapidly emigrate to wound sites and are one of the first cells to arrive (
). However, these cells also express α5β1 integrin. To focus specifically on α4β1 activities, we used CHO cell transfectants expressing α4β1 as their fibronectin receptor (CHOα4) and compared their behaviors to CHO cells expressing α5β1 (CHOα5). The RGD and synergy sequences comprise the binding site for α5β1 on fibronectin while α4β1 binds to the CS1 segment of the alternatively spliced V region (
). This site is found in roughly half of plasma fibronectin subunits whereas fibroblasts synthesize fibronectin with more than 90% V+ subunits. Therefore, as the provisional matrix is replaced by granulation tissue, the number of potential binding sites for α4β1-positive cells increases.
Using the provisional matrix model, we compared CHOα4 and CHOα5 cell adhesion and matrix contraction. As previously reported, CHOα5 cells attach to, spread on, and contract fibrin–fibronectin 3D matrices (
). Both cell lines attached and spread on two-dimensional (2D) substrates of fibronectin showing that the receptors are functional. α4β1 belongs to a class of integrins that requires activation for full functional activity (
). Activation naturally occurs through interactions with cytoplasmic or transmembrane proteins, many of which have not been fully determined. Activation can also be achieved with exogenous treatments that act directly on the integrin extracellular domain. These include addition of Mn2+ or binding of activating antibodies that lock the ectodomain in an active form (
). Furthermore, matrix contraction by activated CHOα4 cells was equivalent to CHOα5 cell contraction. Thus, with appropriate stimulation of α4β1 activity, the integrin is able to function similarly to α5β1.
As healing progresses, the provisional matrix is remodeled, exposing α4β1-positive cells to fibronectin fragments in the wound. Proteolysis of purified fibronectin has been shown to generate sites for integrin binding in vitro (
), which prompted us to test whether α4β1 interactions with fibronectin were modulated during provisional matrix turnover. Fibronectin fragments generated by chymotrypsin digestion were incorporated into the fibrin 3D matrix. Biochemical analyses showed that fibronectin fragments were crosslinked to fibrin by factor XIIIa during the polymerization reaction. Interestingly, CHOα4 cells embedded in the fibrin-fragmented fibronectin matrix were capable of matrix contraction equivalent to that of CHOα5 cells (
). Furthermore, contraction was ablated by function-blocking anti-α4β1 antibodies showing dependence on α4β1–fibronectin interactions. Fragmentation of fibronectin had no stimulatory effect on CHOα5 cell contraction. Thus, the absence of functional interactions between α4β1 and fibronectin in a provisional matrix can be completely reversed by fragmentation of fibronectin in the matrix. In addition to previously identified functions for fibronectin fragments that include competitive inhibitor activities and induction of cell migration and gene expression (
), our results add regulation of cell adhesion and contractility. Furthermore, we suggest that matrix turnover, in addition to matrix composition, provides spatial and temporal control of cell functions during wound repair. A provisional matrix containing intact fibronectin could limit the extent of α4β1-mediated adhesion and matrix contraction (Figure 1a, right) but as fibronectin is degraded, for example, as the provisional matrix is replaced by granulation tissue, fibronectin fragments bind more avidly to α4β1, providing an adhesive environment (Figure 1b, right). The stimulatory effect of fibronectin fragmentation suggests that the CS1 binding site for α4β1 is cryptic but its exposure is enhanced during matrix turnover, perhaps through conformational changes in fibronectin domains or other provisional matrix components that otherwise block cell adhesive sites.
Fibronectin Matrix Deposition and the Microenvironment
One of the primary roles of fibroblasts in wound repair is deposition of new matrix to rebuild damaged tissue. What are the factors within the wound that contribute to fibronectin matrix deposition and its regulation? We examined the assembly of fibronectin fibrils in the provisional matrix model and observed that tenascin-C downregulated fibronectin deposition in a RhoA-dependent manner (
). In fitting with our data, chemically induced dermatitis in tenascin-C-null mice leads to disorganized ECM and ear thickening, suggesting deregulated matrix synthesis in the absence of tenascin-C. Conversely, these null mice have reduced fibronectin in dermal and corneal wounds compared to normal mice (reviewed in
). Apparently, the in vivo role of tenascin-C in ECM assembly is both tissue- and injury-specific.
The level of ECM production may depend on the local microenvironment as we have found using a 3D matrix model to study regulation of matrix assembly. This model is derived from highly confluent fibroblasts that have accumulated a dense fibrillar matrix. Extraction to remove the cells leaves behind a 3D fibrillar network composed primarily of fibronectin (
). Cells are then plated onto this 3D matrix and a combination of biochemical and microscopic approaches are used to follow fibronectin assembly. Comparison to the standard monolayer culture system routinely used to dissect fibronectin matrix assembly has allowed us to show a stimulatory effect of a 3D environment on this process.
Using a variety of fibroblast cell lines that produce their own fibronectin and CHO cells that require exogenous fibronectin to carry out matrix assembly, we showed that fibronectin fibrils accumulate more rapidly, at lower cell density, and with lower amounts of fibronectin when cells are in a 3D matrix environment than when on a fibronectin-coated 2D surface (
). The effect is not due to stimulation of fibronectin synthesis. Immunofluorescence after 6 hours of assembly showed that the fibrils were longer and denser, extending between neighboring cells. In contrast, over the same period on 2D substrates, fibronectin fibrils were short and mostly extended from the cell surface to an adjacent area on the substrate rather than to neighboring cells.
The architecture of the 3D matrix plays a crucial role in stimulating assembly (
). Disrupting the 3D organization in a variety of ways including compressing the 3D matrix from 9 μm down to <3 μm drastically reduced the stimulation of fibril formation. Thus, the architecture and the three-dimensionality of the microenvironment are important to stimulate new matrix assembly.
Of major significance was our finding that a 3D microenvironment stimulates matrix assembly by integrins that would not normally assemble fibronectin fibrils (
). For either of these cell lines, plating on the 3D matrix was sufficient to induce fibronectin assembly in the absence of any exogenous integrin activators. These results indicate that this 3D microenvironment converts the integrins to an activated state. It is possible that interactions with the 3D matrix stimulate formation of novel intracellular complexes that are able to convert the integrins into an active form, thus stimulating assembly. Whatever the mechanism, it raises the interesting possibility that cells encountering an appropriate 3D matrix in vivo will be induced to assemble additional ECM. Upregulation of ECM assembly may also contribute to increased cell–cell cohesion in vivo as recently demonstrated using a hanging drop cell culture system (
). During wound repair, then, it is critical to control the extent of matrix protein synthesis and cell receptor activity in order to rebuild damaged tissue while at the same time limiting the amount of new matrix assembly. If not properly controlled, excess ECM deposition or overactive integrin receptors would most likely delay wound closure and force cells within the wound bed into a fibroproliferative state.
Conclusion
Our studies of 3D fibronectin matrices and their effects on cell activities have highlighted several ways in which cell interactions with fibronectin matrix can be modulated during repair. Changes in ECM architecture or composition, such as deposition of tenascin-C, downregulate fibronectin signals and change cells from stationary to motile phenotypes with requisite modulation of signaling through FAK, Rho, and probably other intracellular pathways. Accumulation of a fibronectin-rich fibrillar matrix has stimulatory effects on further matrix deposition while generation of fibronectin fragments during matrix turnover can have inhibitory or stimulatory effects on cell–matrix interactions. In these examples, the receptor repertoire plays a key role, not only requiring specific integrins but also co-receptors such as syndecan-4. An issue that was not highlighted here is the contributions of signals from other extracellular molecules, such as growth factors or cytokines, to the effects of ECM. This signal integration adds another layer of regulation onto the complexities of cell–ECM interactions that drive the wound healing process.
Conflict of Interest
The authors state no conflict of interest.
ACKNOWLEDGMENTS
This research is funded by grants from the NIH.
REFERENCES
Bass M.D.
Humphries M.J.
Cytoplasmic interactions of syndecan-4 orchestrate adhesion receptor and growth factor receptor signalling.
52622-5&id=gr1.jpg){kind=link}