Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jun 2.
Published in final edited form as: J Invest Dermatol. 2015 Jul 2;135(11):2852–2861. doi: 10.1038/jid.2015.251

Filamin A Mediates Wound Closure by Promoting Elastic Deformation and Maintenance of Tension in the Collagen Matrix

Hamid Mohammadi 1, Vanessa I Pinto 1, Yongqiang Wang 1, Boris Hinz 1, Paul A Janmey 2, Christopher A McCulloch 1
PMCID: PMC4890573  NIHMSID: NIHMS789604  PMID: 26134946

Abstract

Cell-mediated remodeling and wound closure are critical for efficient wound healing, but the contribution of actin-binding proteins to contraction of the extracellular matrix is not defined. We examined the role of filamin A (FLNa), an actin filament cross-linking protein, in wound contraction and maintenance of matrix tension. Conditional deletion of FLNa in fibroblasts in mice was associated with ~ 4 day delay of full-thickness skin wound contraction compared with wild-type (WT) mice. We modeled the healing wound matrix using cultured fibroblasts plated on grid-supported collagen gels that create lateral boundaries, which are analogues to wound margins. In contrast to WT cells, FLNa knockdown (KD) cells could not completely maintain tension when matrix compaction was resisted by boundaries, which manifested as relaxed matrix tension. Similarly, WT cells on cross-linked collagen, which requires higher levels of sustained tension, exhibited approximately fivefold larger deformation fields and approximately twofold greater fiber alignment compared with FLNa KD cells. Maintenance of boundary-resisted tension markedly influenced the elongation of cell extensions: in WT cells, the number (~50%) and length (~300%) of cell extensions were greater than FLNa KD cells. We conclude that FLNa is required for wound contraction, in part by enabling elastic deformation and maintenance of tension in the matrix.

INTRODUCTION

The physical interactions of fibroblasts with collagen critically regulate many cellular processes including mechanosensation (Janmey and McCulloch, 2007; Mohammadi and McCulloch, 2013), metastatic invasion (Alexander and Friedl, 2012), and wound healing (Wong and Gurtner, 2012). Wound closure depends on the ability of fibroblasts to interact with collagen and, through contractile forces, close the margins of healing wounds (Tomasek et al., 2002; Grinnell and Petroll, 2010). Failure to close wounds and restore tissue structure and function are associated with diabetic ulcers (Clark et al., 2007) and non-healing wounds of squamous-cell carcinoma (Trent et al., 2003). Despite the importance of cell-generated tension in wound closure, the mechanisms by which fibroblasts maintain and transmit tension in collagen matrices are not defined, a reflection of the spatiotemporal complexity of wounds and multiple complex signaling systems (Werner and Grose, 2003; Hinz, 2007).

Collagen-based models are widely used to study interfaces between fibroblasts and matrix proteins in wound contraction (Carlson and Longaker, 2004; Dallon and Ehrlich, 2008; Grinnell and Petroll, 2010). One critical process in wound healing involves cell adhesion to collagen by integrins (Koivisto et al., 2014) and discoidin domain receptors (Leitinger, 2011). These adhesions mechanically link collagen and actin filaments and create signaling platforms that couple mechanical inputs to matrix remodeling (Jaalouk and Lammerding, 2009). The interactions of actin filaments with actin-binding proteins at adhesions are critical for generating cell tension and regulating collagen remodeling by contraction (Ross et al., 2013). Filamin A (FLNa) is an actin filament cross-linking protein enriched in cell adhesions (Kim and McCulloch, 2011). It mechanically stabilizes cells by organizing subcortical actin filaments into orthogonal networks that resist mechanical force–induced cell distortion (Flanagan et al., 2001; Nakamura et al., 2007). FLNa is also involved in cell-induced deformation of elastic hydrogels (Byfield et al., 2009) and is associated with remodeling of non-cross-linked collagen networks exhibiting inelastic behavior (Gehler et al., 2009). Currently, the role of FLNa in cell-induced collagen remodeling and maintenance of matrix tension in wound closure is not defined.

For analyzing the role of FLNa in skin wound healing, we examined fibroblast-conditional FLNa-null mice. For defining the role of FLNa in maintaining matrix tension, we knocked down FLNa in fibroblasts and plated these cells on a collagen gel system (Mohammadi et al., 2014) that models the structural heterogeneity of healing wounds. We show that FLNa provides instructional cues that regulate maintenance of tension in collagen during remodeling, thereby linking molecule-scale FLNa function to cell-induced contraction and wound closure.

RESULTS

Delayed wound closure in FLNa-deficient mice

As global deletion of FLNa is embryonic lethal in mice (Feng et al., 2006), we created conditional knock-outs (CKO) in which the floxed FLNa allele was deleted in cells expressing type I collagen (Pinto et al., 2014). PCR analysis of genomic DNA showed targeted excision of the floxed FLNa allele (Figure 1a). At 10 weeks of age, male wild-type (WT) mice weighed 27.3 ± 0.5 g and male FLNa CKO mice weighed 25.8 ± 1.2 g (P > 0.2). We examined wound closure after creating circular (6 mm in diameter) full-thickness wounds in dorsal skin. By 8 days, the wounds closed by 84% in WT and by 58% in FLNa CKO mice (Figure 1b). Wound contraction was delayed by ~ 4 days in FLNa CKO mice compared with WT (Figure 1c). Sections stained for picro-sirius red or Masson’s trichrome showed prominent collagen fibers in the periphery of healing sites at 8 days after wounding (Figure 1d). Quantification of collagen staining in the wound periphery showed a marked reduction (~25%) in the staining intensity in wounds from the FLNa CKO mice compared with WT mice (Figure 1e; P < 0.05) and more abundant myofibroblasts in the healing wounds of WT mice (Supplementary Figure S1 online).

Figure 1. Contraction kinetics of filamin A (FLNa) wild-type (WT) and FLNa conditional knock-out (CKO) mice.

Figure 1

Open in a new tab

(a) The genotypes of FLNa WT mice and fibroblast-conditional FLNa mice were confirmed by PCR of tail-clip genomic DNA to determine targeted excision of the floxed FLNa allele of 200–300 bp compared with WT mice FLNa amplicon of 100 bp. (b) Images of full-thickness wounds in the skin of the back of FLNa WT and FLNa CKO mice 8 days after wounding. Bar=1 mm. (c) Measurement of wound area in WT and FLNa-deficient mice at 1, 4, 5, 7, and 8 days after wounding using image analysis. (d) Histology of wound tissue and the surrounding skin of FLNa WT and FLNa CKO mice 8 days after wounding. Bar= 25 µm. (e) Quantification of collagen staining in the wound periphery. Data are mean ± SEM. *P < 0.05 using unpaired Student’s t-tests.

Resistance to cell-induced deformation in laterally supported, non-cross-linked collagen

Fibroblasts (3T3) with WT levels of FLNa expression or cells with FLNa knockdown (by short hairpin RNA) were used (Figure 2a). We quantified cell-induced deformation of collagen (1 mg ml−1) by tracking displacements of micro-beads as fiduciary markers. In non-cross-linked collagen, the compaction rate accelerated within 2 hours after initial cell attachment to gels. Because of resistance from supporting grids, adherent cells could not locally compact collagen gels, and the compaction rate decreased to zero in ~ 4 hours. FLNa WT cells (Figure 2b) maintained matrix tension because of resistance from physical boundaries, which manifested as positive bead velocities at cell peripheries (Figure 2c). In contrast, FLNa KD cells (Figure 2d) could not maintain matrix tension when compaction was resisted by physical boundaries. Tension in the matrix was partly relaxed (manifest as negative values for bead velocities; Figure 2e). Although the deformation field created by WT and FLNa KD cells extended to, and was resisted by, the boundaries (Figure 2f), the total bead displacements at cell peripheries were approximately twofold higher in FLNa WT cells than FLNa KD cells on non-cross-linked gels (Figure 2g; P < 0.0001). Thus, FLNa is not required for initial compaction of non-cross-linked collagen matrices but is necessary for maintenance of matrix tension when compaction was extended to, and resisted by, the physical boundaries. As wound closure involves the coordinated activity of several different mechanisms (Grinnell, 2003), we examined cell migration in vitro and found a small but a significant reduction in FLNa KD cells (Supplementary Figure S2 online).

Figure 2. Dynamics of deformation fields produced by filamin A (FLNa) wild-type (WT) and FLNa KD cells in grid-supported non-cross-linked collagen matrices.

Figure 2

Open in a new tab

(a) Immunoblot analysis of FLNa WT and FLNa knockdown (KD) NIH 3T3 cells. β-Actin was used as a loading control. The deformation rate by adherent cells was measured through the average velocity of beads in the cell periphery area for FLNa WT cells (b and c) and FLNa KD cells (d and e) adherent on non-cross-linked collagen gels. The maximum extent of deformation field from cell centroid (f) and the total bead displacements at the cell periphery (g) created by FLNa WT and FLNa KD cells adherent on non-cross-linked collagen matrices. Data are reported as mean ± SEM. *P < 0.01 using unpaired Student’s t-tests. Grid size is 200 µm× 200 µm. Bar=20 µm.

Cell-induced matrix deformation in cross-linked collagen matrices

As the mechanical behavior of collagen affects tension in the network (Landau and Lifshitz, 1987), we employed cyclic loading and unloading indentations to estimate reversible deformation (maintenance of tension in the network) and irreversible deformation (relaxation of tension in the network; Achilli et al., 2012). Non-cross-linked collagen matrices were subjected to large-strain compressive indentations (50 µm at rates of 1 µm min−1; Figure 3a), which are characteristic of cell-induced matrix remodeling (Mohammadi et al., 2015); they exhibited ~ 32% irreversible deformation. Irreversible deformation was reduced when collagen was cross-linked, indicating that cross-linked gels exhibit highly elastic behavior, which manifest as less than fivefold reduction in irreversible deformation (Figure 3b; P < 0.0001). Moreover, cross-linked collagen gels exhibited approximately twofold (P < 0.001) greater forces at maximum indentations compared with non-cross-linked collagen networks (Figure 3c).

Figure 3. Measurement of mechanical behavior of collagen matrices and the impact of cross-linking on cell-induced matrix deformation.

Figure 3

Open in a new tab

Cyclic indentation of gels of 1 mg ml−1 (a) collagen concentrations at indentation rates of 1 µm min−1. (b) The irreversible deformation, the point where unloading function intersects x-axis and the (c) maximum supported force at maximum indentation depth (50 µm). Dynamics of matrix deformation produced by filamin A (FLNa) wild-type (WT) (d and e) and FLNa knockdown (KD) (f and g) cells in grid-supported cross-linked collagen matrices. The maximum extent of deformation field from the cell centroid (h) and the total bead displacements at the cell periphery (i) created by FLNa WT and FLNa KD cells adherent on cross-linked and cross-linked collagen matrices. Bar = 20 µm. Data are reported as mean ± SEM. P-values were calculated with unpaired Student’s t-tests (*P < 0.01; **P < 0.0001).

We examined the role of FLNa in elastic deformation of collagen matrices. WT or FLNa KD cells were incubated on covalently cross-linked collagen supported by grids. These matrices exhibit very limited inelastic behavior. WT cells deformed the matrix at cell peripheries (Figure 3d) and maintained cell-generated matrix tension, similar to WT cells adherent to non-cross-linked gels. However, maximum bead velocities were approximately twofold lower (P < 0.001; Figure 3e) compared with bead velocities in non-cross-linked matrices. Bead velocity and matrix deformation by FLNa KD cells (Figure 3f) were much lower (approximately sixfold, P < 0.0001, Figure 3g) in cross-linked collagen compared with FLNa KD cells adherent to non-cross-linked collagen.

The deformation field created by FLNa KD cells in cross-linked collagen did not extend to, or was resisted by, the grids. This observation contrasts with cell-induced matrix deformation by FLNa KD cells adherent to non-cross-linked collagen gels or FLNa WT cells on either type of collagen (Figure 3h), in which deformation extended to, and was resisted by, the grids. Bead displacements at the cell periphery were higher with FLNa WT cells than FLNa KD cells on cross-linked collagen (Figure 3i). Thus, FLNa is not required for irreversible (inelastic) cell-induced matrix deformation but is important for reversible (elastic) deformation of cross-linked collagen.

FLNa is involved in local remodeling of cross-linked collagen

As collagen deformation (measured by bead displacement) reports global matrix remodeling and is influenced by FLNa and collagen cross-linking, we examined the role of FLNa in collagen remodeling by measuring local fiber alignment in cross-linked and non-cross-linked collagen. Fast Fourier Transform analysis indicated no directional preference in fiber orientation at time 0, which manifested as low alignment (< 0.4; Figure 4g). After ~ 2 hours of initial cell attachment to gels, and when cell-induced matrix deformation is resisted by the grids, reorganization and remodeling of collagen fibers by FLNa WT and KD cells (Figure 4b and e, respectively) resulted in approximately threefold higher alignment (P < 0.001, Figure 4g). At ~ 6 hours after initial cell attachment to gels, WT cells continued remodeling of collagen, which resulted in an ~40% (P < 0.01) increased alignment (Figure 4c). In contrast, FLNa KD cells did not maintain fiber alignment in the network and the alignment index decreased (by ~35% at ~ 6 hours; Figure 4g; P < 0.01). In WT cells on cross-linked collagen, alignment increased approximately twofold (P < 0.001; Figure 4i) at ~ 2 hours after initial attachment of cells to collagen (Figure 4h). Alignment did not change after ~ 6 hours from initial cell attachment (P > 0.1; Figure 4j). In contrast, FLNa KD cells did not remodel cross-linked collagen, which resulted in nearly similar alignment indices at 0 hour (Figure 4k), 2 hours (Figure 4l), and 6 hours (Figure 4m) after initial attachment (P > 0.5; Figure 4n).

Figure 4. Cell-induced remodeling of non-cross-linked and cross-linked collagen matrices.

Figure 4

Open in a new tab

Representative images of collagen fibers in non-cross-linked and cross-linked collagen gels of 1 mg ml−1 collagen concentration after remodeling of the network by filamin A (FLNa) wild-type (WT) cells (a–c) and (h–j) and FLNa knockdown (KD) cells (d–f) and (k–m), respectively. Pixel intensity in Fast Fourier Transform (FFT) analysis of at least 30 images of collagen fibers at the tip of cell extensions was summed along a straight line from the center to the edge of the image at different angles and this provided information about collagen directionality and alignment. Alignment index was quantified by measurement of area under the intensity curve within ± 10° of the peak (g and n). Bar = 15 µm. Data are reported as mean ± SEM. P-value was calculated with unpaired Student’s t-tests (*P < 0.01, **P < 0.0001).

FLNa is required for boundary-triggered elongation of cell extensions

Lateral resistance to cell-induced deformation in collagen triggers elongation of cell extensions (Mohammadi et al., 2014). As FLNa KD cells could not maintain matrix tension when cell-generated deformations are resisted by grids, we suggest that the formation/elongation of cell extensions is influenced by FLNa expression. Accordingly, we measured the number and length of cell extensions when cell-induced deformation was resisted by grids and when matrix tension may be relaxed (~2 hours and ~ 6 hours after initial cell attachment to gels, respectively). FLNa WT cells adherent to non-cross-linked collagen were well spread and formed few (~4 per cell) but relatively long extensions (~22 µm) at ~ 2 hours after initial attachment (Figure 5a and b). The number (~50% increase; P < 0.001) and length of cell extensions (approximately threefold increase; P < 0.0001) were increased at ~ 6 hours (Figure 5c) after initial attachment. Similar to WT cells (P > 0.1), FLNa KD cells formed few, relatively long cell extensions at ~ 2 hours after initial adherence to gels (Figure 5d and e). In contrast to WT cells, FLNa KD cells were unable to form and elongate new cell extensions in the next 4 hours (Figure 5f). Consequently, there were fewer (by ~ 50%; P < 0.001) and shorter (approximately threefold; P < 0.0001) cell extensions than WT cells at ~ 6 hours after initial cell attachment to gels (Figure 5h and g).

Figure 5. Mechanosensation of lateral physical boundaries and formation and elongation of cell extensions.

Figure 5

Open in a new tab

Representative images of morphology of filamin A (FLNa) wild-type (WT) cells (a–c) and FLNa knockdown (KD) cells (d–f) adherent to non-cross-linked collagen matrices at 1, 2, and 6 hours after initial attachment. Mean length (g) and mean number (h) of cell extensions were measured at each observation time. Effect of cross-linking of collagen matrices on formation of cell extensions by FLNa WT (i–k) and FLNa KD cells adherent to cross-linked collagen matrices (l–n). Mean length (o) and mean number (p) of cell extensions per cell at 1, 2, and 6 hours after initial attachment of cells to the collagen gel. Data are reported as mean ± SEM. P-value was calculated with unpaired Student’s t-tests (*P < 0.01, **P < 0.0001). Bar =15 µm.

Collagen cross-linking resulted in more rapid formation and elongation of cell extensions by FLNa WT cells. Within ~ 1 hour after initial adherence of cells to cross-linked collagen (Figure 5i), WT cells formed long extensions whose length was approximately twofold (P < 0.001) longer compared with adherent cells on non-cross-linked collagen (Figure 5j). Similarly, the number of cell extensions was approximately twofold higher in FLNa WT cells on non-cross-linked gels (P < 0.001). In contrast to WT cells, collagen cross-linking markedly influenced the formation of cell extensions in FLNa KD cells (Figure 5l, m, and n). The mean number and length of cell extensions were reduced (two to fourfold, P < 0.0001) at each of 1, 2, and 6 hours (Figure 5p and o). Although there was no morphological difference between FLNa WT and FLNa KD cells at ~ 2 hours after cell attachment to non-cross-linked matrices, the morphology of FLNa WT and FLNa KD cells was quite distinct from cells attached to cross-linked collagen.

Interactions of FLNa with actin and stabilization of actin network impact cell extensions

We extended the study of the role of FLNa in formation of cell extensions by transfecting FLNa KD cells with dsRed-FLNa constructs (actin-binding domain-deleted (FLNa ABDD) or full-length FLNa (FLNa-Rescue)). FLNa ABDD cells adherent to non-cross-linked collagen gels formed few (~4) and relatively long (> 20 µm) cell extensions within 2 hours after initial attachment, which was similar to FLNa-Rescue cells (Figure 6a and b). However, at ~ 6 hours after attachment to gels, FLNa-Rescue cells formed approximately threefold (P < 0.0001) longer and approximately twofold (P < 0.001) more cell extensions than FLNa ABDD cells (Figure 6a and b). FLNa ABDD cells could not elongate and form new cell extensions after the critical time interval when cell-induced matrix deformation was resisted by grids. Consequently, the length and number of cell extensions were unchanged (P > 0.1; Figure 6a and b). In contrast to FLNa-Rescue cells, FLNa ABDD cells did not continue remodeling collagen after initial remodeling, which occurs within ~ 2 hours after cell attachment to collagen. Accordingly, there was reduced alignment (~50%; P < 0.001) compared with FLNa-Rescue cells at ~ 6 hours after attachment (Figure 6c). Although FLNa-Rescue cells exhibited greater collagen fiber alignment, FLNa ABDD cells did not maintain alignment, manifest as reduced alignment at 6 hours.

Figure 6. Interactions of filamin A (FLNa) with actin and impact on focal adhesions.

Figure 6

Open in a new tab

Morphology of adherent FLNa actin-binding domain-deleted (ABDD) and full-length FLNa (FLNa-rescue cells) was characterized by measurement of the length (a) and number (b) of cell extensions. (c) Collagen fiber alignment and directionality were quantified using Fast Fourier Transform analysis. Mean length (d) and the number (e) of cell extensions per cell of FLNa knockdown (KD), jasplakinolide-treated FLNa KD and FLNa WT cells. Total size of focal adhesions was measured by imaging focal adhesions stained for vinculin (f). Adherent FLNa WT and KD cells were subjected to 4, 8, or 16 washes at 0 hour (g), ~ 2 hours (h), and ~ 6 hours (i) after initial attachment. Data are representative of at least three independent samples per group. Data are mean ± SEM. P-value was calculated with unpaired Student’s t-tests (*P < 0.01).

We examined how stabilizing actin filaments with jasplakinolide (Bubb et al., 2000) affects cell extension formation. Jasplakinolide-treated FLNa KD cells formed longer (~120%; P < 0.0001) and more (~60%; P < 0.001) cell extensions than vehicle-treated controls, although the number and length of cell extensions were less than FLNa WT cells. FLNa WT cells formed ~30% longer (Figure 6d; P < 0.05) and ~20% (Figure 6e; P < 0.05) more cell extensions compared with jasplakinolide-treated FLNa KD cells 6 hours after initial cell attachment.

To examine the effect of FLNa on focal adhesions, we measured focal adhesion size at 0, 2, and 6 hours after initial cell attachment (Rajshankar et al., 2012). Immunostaining of the focal adhesion protein vinculin in cells adherent to non-cross-linked or cross-linked (Supplementary Figure S3 online) collagen was quantified and normalized to total cell area. In non-cross-linked collagen matrices, focal adhesion size was approximately threefold higher (P < 0.0001) in FLNa WT and KD cells at ~ 2 hours after initial cell adherence (Figure 6f). However, there was no difference of normalized area of focal adhesions in FLNa WT and KD cells at ~ 6 hours (Figure 6f) compared with ~ 2 hours (P > 0.05). FLNa WT cells exhibited larger adhesions (~30%, P < 0.01) than FLNa KD cells at ~ 6 hours after initial cell attachment (Figure 6f).

We estimated the strength of adhesions by applying shear forces at 0, 2, or 6 hours after initial cell attachment to collagen. The numbers of attached FLNa WT and KD cells were reduced with increasing numbers of washes (Figure 6g, h, and i). After 16 washes, there was no difference in normalized numbers of attached FLNa WT and KD cells immediately after washing (Figure 6g; P > 0.09). There were only small differences of normalized number of attached cells after 16 washes between FLNa WT and KD cells at ~ 2 hours after initial cell attachment (Figure 6h; ~ 25%, P < 0.01). Normalized numbers of attached FLNa WT cells were ~ 100% more than KD cells subjected to 16 washes at ~ 6 hours after initial cell attachment (Figure 6i; P < 0.0001). Notably, there was no significant difference in the numbers of attached FLNa KD cells at ~ 2 hours and ~ 6 hours after initial cell attachment to the gels (P > 0.08). Thus, interaction of FLNa with actin is required for prolonged collagen remodeling and formation of cell extensions, whereas the impact of FLNa on the size and strength of adhesions during the critical time interval is minimal.

DISCUSSION

Our major finding is that FLNa has a critical role in skin wound healing by modulating wound contraction, a process that involves remodeling of the extracellular matrix by resident fibroblasts. We used an in vitro wound contraction model to examine the role of FLNa in matrix compaction and remodeling. Current in vitro wound assays (Carlson and Longaker, 2004; Dallon and Ehrlich, 2008; Grinnell and Petroll, 2010) for studying cell–matrix interactions are limited by their exclusion of physical boundaries (i.e., the margins of the healing wound), as the margins of healing wounds affect matrix compaction and remodeling by adjacent cells (Clark, 1989). As resistance to cell-induced matrix remodeling by physical boundaries markedly influences cell behavior (Mohammadi et al., 2014), we used a model in which nylon grids provide defined physical boundaries in the plane of the supported collagen matrix. With this model we found that FLNa-induced stabilization of subcortical actin filament networks is required for reversible (i.e., elastic) deformation and remodeling of the collagen matrices that require higher levels of sustained tension in the network. Further, we found that FLNa is required for maintenance of tension when cell-induced matrix deformation is resisted by marginal physical boundaries, which in turn markedly influenced boundary-triggered elongation of cell extensions. However, FLNa is not required for irreversible (i.e., inelastic) deformation and remodeling of collagen, which requires much lower levels of sustained tension.

In early phases of wound healing, abundant, randomly oriented and poorly cross-linked collagen fibers are produced (Clark, 1989). The elastic and inelastic mechanical behaviors of non-cross-linked collagen matrices contribute, respectively, to the initial reversible and irreversible matrix deformations mediated by cells (Billiar, 2011). These deformations occur before cells detect resistance from lateral physical boundaries. We found that, although FLNa was not required for cell-induced inelastic remodeling and deformation of native collagen, matrix remodeling and deformation by FLNa KD cells in cross-linked collagen were markedly reduced. Indeed, the mechanical properties of collagen subjected to slow indentation (at rates similar to cell-induced matrix remodeling; Mohammadi et al., 2015) showed that cross-linking of collagen gels strongly reduced inelastic behavior. Although collagen cross-linking markedly reduces the inelastic behavior of collagen, non-cross-linked collagen gels also exhibit inelastic and elastic behaviors. Accordingly, the initial compaction of non-cross-linked collagen by adherent cells arises from elastic (reversible) and inelastic (irreversible) matrix deformations, which reduces matrix deformation by FLNa KD cells. This observation agrees with a previous report on the importance of FLNa in remodeling of collagen matrices at high collagen concentrations (Gehler et al., 2009), which exhibit less inelastic behavior (Mohammadi et al., 2015).

Our data demonstrate that FLNa helps maintain matrix tension when cell-induced deformations in the matrix extend to, and are resisted by, physical boundaries. As matrix resistance feeds back and influences cytoskeletal filament organization (Discher et al., 2005) and intracellular tension (Wang et al., 1993), we suggest that reduced cytoskeletal tension in FLNa KD cells mediates matrix relaxation. These observations support the notion that cross-linked actin networks can maintain high levels of tension when subjected to external forces (Ferrer et al., 2008) and that FLNa is important for the maintenance of reversible, but not irreversible, matrix tension.

We found that cross-linking collagen enabled rapid elongation and formation of cell extensions in FLNa WT cells. We attribute this observation to transmission of cell-induced deformation to the boundaries and resistance from the boundaries. Indeed, cross-linked fibrillar networks facilitate transmission of applied external forces over relatively long distances (Head et al., 2003), in part due to reduced dissipation of stress in collagen networks (Landau and Lifshitz, 1987; Timoshenko and Gere, 2012). We found that the elastic matrix deformation was reduced in FLNa-depleted cells and that boundary-triggered formation and elongation of extensions was blocked. These data agree with a requirement for FLNa in response to stiffness gradients in cells plated on collagen-coated, linearly elastic polyacrylamide gels (Byfield et al., 2009). Cross-linked fibrillar networks likely facilitate transmission of applied external forces over relatively long distances (Head et al., 2003), which may be in part due to reduced stress dissipation in the network (Landau and Lifshitz, 1987; Timoshenko and Gere, 2012). We found that elastic deformation of the matrix was reduced in cells depleted of FLNa and that boundary-triggered formation and elongation of cells extensions was blocked. Indeed, our data show that FLNa KD cells can sense physical boundaries when plated on non-cross-linked collagen, which exhibits inelastic behavior.

Cross-linking of collagen inhibits inelastic behavior and enhances matrix stiffness (Miron-Mendoza et al., 2010). As cross-linking of collagen resulted in more rapid formation and elongation of cell extensions by FLNa WT cells, it is unclear whether this response is caused by inhibition of inelastic behavior, increased stiffness, or alignment of collagen fibers parallel to the nascent cell extensions. An alternative approach to enhance matrix stiffness and limit cell-induced matrix compaction is to use higher collagen concentrations (5 mg ml−1). To eliminate the fibrous nature of the matrix, we used relatively soft polyacrylamide hydrogels and found that the formation and elongation of cell extensions by FLNa WT cells on collagen were reduced in 5 mg ml−1 collagen and polyacrylamide hydrogels compared with cells on 1 mg ml−1 collagen (Supplementary Figure S4 online). Thus, the generation of cell extensions in grid-supported collagen may depend on transmission of cell-induced matrix compaction to, and the consequent resistance by, physical boundaries and the fibrous nature of the matrix instead of local collagen stiffness.

Interactions between FLNa and other cytoskeletal and adhesion proteins regulate many signaling pathways and functions (Kim and McCulloch, 2011; Razinia et al., 2012). For example, depletion of FLNa reduces the adhesion strength of focal adhesions (Lynch et al., 2011) and linkages between cytoskeletal proteins (Baldassarre et al., 2009; Kim et al., 2010a; 2010b). We examined cells expressing FLNa but without the actin-binding domain, a mutant construct that preserves interaction of FLNa with other adaptor proteins in the cytoplasm but not with actin filaments. We found that the FLNa–actin filament interaction is important for elastic deformation and remodeling of collagen networks and for maintenance of tension when cell-induced deformation in the matrix is resisted by physical boundaries. Further, cells expressing FLNa without the actin-binding domain could not elongate cell extensions. Consistent with this notion, jasplakinolide “rescued” the phenotype of FLNa KD cells by restoring formation of long extensions. Notably, jasplakinolide regulates spontaneous nucleation of actin monomers and actin polymerization and stabilizes actin filaments by inhibiting filament disassembly (Bubb et al., 2000).

We found that FLNa KD cells did not exhibit alterations of adhesion size, in contrast to earlier data showing that depletion of FLNa reduced focal adhesion size in cells plated on glass or collagen-coated linear elastic substrates (Byfield et al., 2009; Lynch and Sheetz, 2011). We attribute this difference to the mechanics of underlying substrates that are very different from collagen (Silver, 2006; Billiar, 2011; Achilli et al., 2012), which exhibits both elastic and inelastic behaviors. Collectively, our data highlight the importance of FLNa-mediated stabilization of actin networks in boundary-triggered formation of cell extensions.

Our observations inform underlying mechanisms by which connective tissue cells sense and respond to the complex mechanics of collagen (Carey et al., 2012; Shin et al., 2013; Bordeleau et al., 2014). Cell extension formation is related to the increased ability of migrating cells to remodel the matrix (Chen and Wang, 1999; Wolf et al., 2007; Yilmaz and Christofori, 2009). Accordingly, our data linking FLNa to the formation of cell extensions may explain in part how cells navigate the matrix and probe remote sites in connective tissues. An improved understanding of how environmental factors (e.g., elastic and inelastic behaviors of collagen, presence of physical boundaries) could help define how cells navigate heterogeneous matrices.

MATERIALS AND METHODS

Detailed materials and methods are provided in the Supplementary Materials and Methods online.

FLNa-conditional mice and skin wounds

We used C57BL/6 male mice homozygous for the targeted FLNa mutation; male WT mice of the same strain served as controls. The genotypes of the fibroblast-conditional FLNa mice and the WT mice were confirmed by PCR of tail-clip genomic DNA as described (Pinto et al., 2014). For skin wounding, mice were anesthetized with methoxyfluorane, and the skin of the back was prepared for aseptic surgery. Paired, 6 mm diameter, full-thickness wounds were created through the skin of 10–12 week-old mice. Wounds were left exposed, and the extent of wound closure was quantified by measuring the wound area at 1, 4, 5, 7, and 8 days after wounding. The wound tissue and the surrounding skin were collected at day 8 after wounding, and these tissues were fixed in formalin, processed, and sectioned. All procedures performed were approved by the University of Toronto, animal care committee.

Cell culture and transfection

NIH 3T3 fibroblasts that constitutively express FLNa or cells transfected with FLNa short hairpin RNA (FLNa KD) were prepared as described previously (Shifrin et al., 2009). Cells were grown in DMEM and cultured in a medium supplemented with 10% fetal bovine serum and antibiotics (146 U ml−1 penicillin G, 50 µg ml−1 gentamicin, and 0.25 µg ml−1 amphotericin). FLNa KD cells were grown in the presence of puromycin (1 µg ml−1) to maintain knockdowns. Transfections of FLNa KD cells with short hairpin-resistant dsRed-FLNa WT or dsRed-FLNa without the FLNa actin-binding domain (amino acids 44–264) were used to generate cells with FLNa expression restored (FLNa-Rescue) or with FLNa in which the actin-binding domain of FLNa was deleted (FLNa ABDD). Cells were transfected with PolyJet transfection reagent (FroggaBio, Toronto, Ontario, Canada) according to the manufacturer’s instructions.

In vitro wound contraction model

Collagen gels with embedded nylon mesh sheets and with square openings (200 µm wide and fiber diameters of ~ 100 µm) were prepared from pepsin-treated, bovine dermal type I collagen (6.0 mg ml−1; ~ 97% type I collagen; Advanced BioMatrix, San Diego, CA) as described earlier (Mohammadi et al., 2014). Nylon mesh sheets were used to create physical boundaries that modeled matrix microenvironments surrounding a single cell and that resist cell-induced matrix contraction. Nylon sheets were cut into 2 cm× 2 cm pieces with ~ 4,500 grids of 200-µm width. The same number of cells was plated so that the ratio of cells to grids was approximately 1:1. We only analyzed those cells that were located in the middle of grids.

Statistical analysis

All continuous variable data are reported as mean ± SEM. Pair-wise comparisons for statistical significance were computed with unpaired Student’s t-test at a significance level of P < 0.05.

Supplementary Material

supplement

Acknowledgments

HM gratefully acknowledges financial support from a CIHR Strategic Training Fellow, STP-53877. The research was supported by CIHR operating grant to CAM (MOP-11106).

Abbreviations

CKO

conditional knock-out

FLNa

filamin A

FLNa ABDD

actin-binding domain-deleted

FLNa-Rescue

full-length FLNa

WT

wild type

Footnotes

CONFLICT OF INTEREST

The authors state no conflict of interest.

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at http://www.nature.com/jid

REFERENCES

  1. Achilli M, Meghezi S, Mantovani D. On the viscoelastic properties of collagen-gel-based lattices under cyclic loading: applications for vascular tissue engineering. Macromol Mater Eng. 2012;297:724–734. [Google Scholar]
  2. Alexander S, Friedl P. Cancer invasion and resistance: interconnected processes of disease progression and therapy failure. Trends Mol Med. 2012;18:13–26. doi: 10.1016/j.molmed.2011.11.003. [DOI] [PubMed] [Google Scholar]
  3. Baldassarre M, Razinia Z, Burande CF, et al. Filamins regulate cell spreading and initiation of cell migration. PLoS One. 2009;4:e7830. doi: 10.1371/journal.pone.0007830. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Billiar KL. The mechanical environment of cells in collagen gel models. In: Gefen, editor. Cellular and Biomolecular Mechanics and Mechanobiology. Berlin, Germany: Springer; 2011. pp. 201–245. [Google Scholar]
  5. Bordeleau F, Alcoser TA, Reinhart-King CA. Physical biology in cancer. 5. The rocky road of metastasis: the role of cytoskeletal mechanics in cell migratory response to 3D matrix topography. Am J Physiol Cell Physiol. 2014;306:C110–C120. doi: 10.1152/ajpcell.00283.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bubb MR, Spector I, Beyer BB, et al. Effects of jasplakinolide on the kinetics of actin polymerization. An explanation for certain in vivo observations. J Biol Chem. 2000;275:5163–5170. doi: 10.1074/jbc.275.7.5163. [DOI] [PubMed] [Google Scholar]
  7. Byfield FJ, Wen Q, Levental I, et al. Absence of filamin A prevents cells from responding to stiffness gradients on gels coated with collagen but not fibronectin. Biophys J. 2009;96:5095–5102. doi: 10.1016/j.bpj.2009.03.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Carey SP, Kraning-Rush CM, Williams RM, et al. Biophysical control of invasive tumor cell behavior by extracellular matrix microarchitecture. Biomaterials. 2012;33:4157–4165. doi: 10.1016/j.biomaterials.2012.02.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carlson MA, Longaker MT. The fibroblast-populated collagen matrix as a model of wound healing: a review of the evidence. Wound Repair Regen. 2004;12:134–147. doi: 10.1111/j.1067-1927.2004.012208.x. [DOI] [PubMed] [Google Scholar]
  10. Chen WT, Wang JY. Specialized surface protrusions of invasive cells, invadopodia and lamellipodia, have differential MT1-MMP, MMP-2, and TIMP-2 localization. Ann NY Acad Sci. 1999;878:361–371. doi: 10.1111/j.1749-6632.1999.tb07695.x. [DOI] [PubMed] [Google Scholar]
  11. Clark RA. Wound repair. Curr Opin Cell Biol. 1989;1:1000–1008. doi: 10.1016/0955-0674(89)90072-0. [DOI] [PubMed] [Google Scholar]
  12. Clark RA, Ghosh K, Tonnesen MG. Tissue engineering for cutaneous wounds. J Investig Dermatol Symp Proc. 2007;127:1018–1029. doi: 10.1038/sj.jid.5700715. [DOI] [PubMed] [Google Scholar]
  13. Dallon JC, Ehrlich HP. A review of fibroblast-populated collagen lattices. Wound Repair Regen. 2008;16:472–479. doi: 10.1111/j.1524-475X.2008.00392.x. [DOI] [PubMed] [Google Scholar]
  14. Discher DE, Janmey P, Wang YL. Tissue cells feel and respond to the stiffness of their substrate. Science. 2005;310:1139–1143. doi: 10.1126/science.1116995. [DOI] [PubMed] [Google Scholar]
  15. Feng Y, Chen MH, Moskowitz IP, et al. Filamin A (FLNA) is required for cell-cell contact in vascular development and cardiac morphogenesis. Proc Natl Acad Sci USA. 2006;103:19836–19841. doi: 10.1073/pnas.0609628104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ferrer JM, Lee H, Chen J, et al. Measuring molecular rupture forces between single actin filaments and actin-binding proteins. Proc Natl Acad Sci USA. 2008;105:9221–9226. doi: 10.1073/pnas.0706124105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Flanagan LA, Chou J, Falet H, et al. Filamin A, the Arp2/3 complex, and the morphology and function of cortical actin filaments in human melanoma cells. J Cell Biol. 2001;155:511–517. doi: 10.1083/jcb.200105148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gehler S, Baldassarre M, Lad Y, et al. Filamin A-beta1 integrin complex tunes epithelial cell response to matrix tension. Mol Biol Cell. 2009;20:3224–3238. doi: 10.1091/mbc.E08-12-1186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Grinnell F. Fibroblast biology in three-dimensional collagen matrices. Trends Cell Biol. 2003;13:264–269. doi: 10.1016/s0962-8924(03)00057-6. [DOI] [PubMed] [Google Scholar]
  20. Grinnell F, Petroll WM. Cell motility and mechanics in three-dimensional collagen matrices. Annu Rev Cell Dev Biol. 2010;26:335–361. doi: 10.1146/annurev.cellbio.042308.113318. [DOI] [PubMed] [Google Scholar]
  21. Head DA, Levine AJ, MacKintosh FC. Deformation of cross-linked semiflexible polymer networks. Phys Rev Lett. 2003;91:108102. doi: 10.1103/PhysRevLett.91.108102. [DOI] [PubMed] [Google Scholar]
  22. Hinz B. Formation and function of the myofibroblast during tissue repair. J Investig Dermatol Symp Proc. 2007;127:526–537. doi: 10.1038/sj.jid.5700613. [DOI] [PubMed] [Google Scholar]
  23. Jaalouk DE, Lammerding J. Mechanotransduction gone awry. Nat Rev Mol Cell Biol. 2009;10:63–73. doi: 10.1038/nrm2597. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Janmey PA, McCulloch CA. Cell mechanics: integrating cell responses to mechanical stimuli. Annu Rev Biomed Eng. 2007;9:1–34. doi: 10.1146/annurev.bioeng.9.060906.151927. [DOI] [PubMed] [Google Scholar]
  25. Kim H, McCulloch CA. Filamin A mediates interactions between cytoskeletal proteins that control cell adhesion. FEBS Lett. 2011;585:18–22. doi: 10.1016/j.febslet.2010.11.033. [DOI] [PubMed] [Google Scholar]
  26. Kim H, Nakamura F, Lee W, et al. Regulation of cell adhesion to collagen via beta1 integrins is dependent on interactions of filamin A with vimentin and protein kinase C epsilon. Exp Cell Res. 2010a;316:1829–1844. doi: 10.1016/j.yexcr.2010.02.007. [DOI] [PubMed] [Google Scholar]
  27. Kim H, Nakamura F, Lee W, et al. Filamin A is required for vimentin-mediated cell adhesion and spreading. Am J Physiol Cell Physiol. 2010b;298:C221–C236. doi: 10.1152/ajpcell.00323.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Koivisto L, Heino J, Hakkinen L, et al. Integrins in Wound Healing. Adv Wound Care. 2014;3:762–783. doi: 10.1089/wound.2013.0436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Landau LD, Lifshitz EM. Theory of Elasticity. New York, NY, USA: Pergamon; 1987. pp. 13–30. [Google Scholar]
  30. Leitinger B. Transmembrane collagen receptors. Annu Rev Cell Dev Biol. 2011;27:265–290. doi: 10.1146/annurev-cellbio-092910-154013. [DOI] [PubMed] [Google Scholar]
  31. Lynch CD, Gauthier NC, Biais N, et al. Filamin depletion blocks endoplasmic spreading and destabilizes force-bearing adhesions. Mol Biol Cell. 2011;22:1263–1273. doi: 10.1091/mbc.E10-08-0661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lynch CD, Sheetz MP. Cellular mechanotransduction: filamin A strains to regulate motility. Curr Biol. 2011;21:R916–R918. doi: 10.1016/j.cub.2011.10.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Miron-Mendoza M, Seemann J, Grinnell F. The differential regulation of cell motile activity through matrix stiffness and porosity in three dimensional collagen matrices. Biomaterials. 2010;31:6425–6435. doi: 10.1016/j.biomaterials.2010.04.064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mohammadi H, Arora PD, Simmons CA, et al. Inelastic behaviour of collagen networks in cell-matrix interactions and mechanosensation. J R Soc Interface. 2015;12:20141074. doi: 10.1098/rsif.2014.1074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Mohammadi H, Janmey PA, McCulloch CA. Lateral boundary mechanosensing by adherent cells in a collagen gel system. Biomaterials. 2014;35:1138–1149. doi: 10.1016/j.biomaterials.2013.10.059. [DOI] [PubMed] [Google Scholar]
  36. Mohammadi H, McCulloch CA. Impact of elastic and inelastic substrate behaviors on mechanosensation. Soft Matter. 2013;10:408–420. doi: 10.1039/c3sm52729h. [DOI] [PubMed] [Google Scholar]
  37. Nakamura F, Osborn TM, Hartemink CA, et al. Structural basis of filamin A functions. J Cell Biol. 2007;179:1011–1025. doi: 10.1083/jcb.200707073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Pinto VI, Senini VW, Wang Y, et al. Filamin A protects cells against force-induced apoptosis by stabilizing talin- and vinculin-containing cell adhesions. FASEB J. 2014;28:453–463. doi: 10.1096/fj.13-233759. [DOI] [PubMed] [Google Scholar]
  39. Rajshankar D, Downey GP, McCulloch CA. IL-1beta enhances cell adhesion to degraded fibronectin. FASEB J. 2012;26:4429–4444. doi: 10.1096/fj.12-207381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Razinia Z, Makela T, Ylanne J, et al. Filamins in mechanosensing and signaling. Annu Rev Biophys. 2012;41:227–246. doi: 10.1146/annurev-biophys-050511-102252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Ross TD, Coon BG, Yun S, et al. Integrins in mechanotransduction. Curr Opin Cell Biol. 2013;25:613–618. doi: 10.1016/j.ceb.2013.05.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Shifrin Y, Arora PD, Ohta Y, et al. The role of FilGAP-filamin A interactions in mechanoprotection. Mol Biol Cell. 2009;20:1269–1279. doi: 10.1091/mbc.E08-08-0872. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Shin Y, Kim H, Han S, et al. Extracellular matrix heterogeneity regulates three-dimensional morphologies of breast adenocarcinoma cell invasion. Adv Health Mater. 2013;2:790–794. doi: 10.1002/adhm.201200320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Silver FH. Mechanosensing and Mechanochemical Transduction in Extracellular Matrix. Berlin, Germany: Springer; 2006. p. 292. [Google Scholar]
  45. Timoshenko SP, Gere JM. Theory of Elastic Stability. New York, NY, USA: Courier Dover Publications; 2012. p. 560. [Google Scholar]
  46. Tomasek JJ, Gabbiani G, Hinz B, et al. Myofibroblasts and mechanoregulation of connective tissue remodelling. Nat Rev Mol Cell Biol. 2002;3:349–363. doi: 10.1038/nrm809. [DOI] [PubMed] [Google Scholar]
  47. Trent JT, Federman D, Kirsner RS. Skin and wound biopsy: when, why, and how. Adv Skin Wound Care. 2003;16:372–375. doi: 10.1097/00129334-200312000-00018. [DOI] [PubMed] [Google Scholar]
  48. Wang N, Butler JP, Ingber DE. Mechanotransduction across the cell surface and through the cytoskeleton. Science. 1993;260:1124–1127. doi: 10.1126/science.7684161. [DOI] [PubMed] [Google Scholar]
  49. Werner S, Grose R. Regulation of wound healing by growth factors and cytokines. Physiol Rev. 2003;83:835–870. doi: 10.1152/physrev.2003.83.3.835. [DOI] [PubMed] [Google Scholar]
  50. Wolf K, Wu YI, Liu Y, et al. Multi-step pericellular proteolysis controls the transition from individual to collective cancer cell invasion. Nat Cell Biol. 2007;9:893–904. doi: 10.1038/ncb1616. [DOI] [PubMed] [Google Scholar]
  51. Wong VW, Gurtner GC. Tissue engineering for the management of chronic wounds: current concepts and future perspectives. Exp Dermatol. 2012;21:729–734. doi: 10.1111/j.1600-0625.2012.01542.x. [DOI] [PubMed] [Google Scholar]
  52. Yilmaz M, Christofori G. EMT, the cytoskeleton, and cancer cell invasion. Cancer Metastasis Rev. 2009;28:15–33. doi: 10.1007/s10555-008-9169-0. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS789604-supplement-supplement.pdf (30.5MB, pdf)

RESOURCES