The process of corneal wound repair following lacerating injury or refractive surgery generally depends upon mechanical processes such as the migration of activated keratocytes (corneal fibroblasts) into the wound from the surrounding stroma, apposition of the wound edges (wound contraction), and extracellular matrix reorganization (remodeling) (Netto et al., 2005
). These biomechanical mechanisms ultimately control corneal clarity and refractive (visual) outcome (Jester et al., 1999
; Moller-Pedersen et al., 1998a
; Moller-Pedersen et al., 1998b
; Moller-Pedersen et al., 1997
; Petroll et al., 1992
). The small GTPases Rho and Rac are prime candidates for regulating the cytoskeletal and mechanical phenotype of fibroblasts during various phases of wound healing. Activation of Rho by lysophosphatidic acid (LPA) increases contractility on both silicon substrates (Chrzanowska-Wodnicka and Burridge, 1994
; Craig and Johnson, 1996
) and within collagen lattices (Kolodney and Elson, 1993
; Tomasek et al., 1992
), and these responses appear to be mediated by ROCK (Grinnell, 2000
; Parizi et al., 2000
; Tamariz and Grinnel, 2002
). PDGF activates Rac in dermal fibroblasts (Grinnell, 2000
; Sander et al., 1999
), and enhances both cell spreading and migration within 3-D collagen matrices (Andresen et al., 1997
; Grinnell et al., 2006
). PDGF has also been shown to induce local matrix deformation during cell spreading in 3-D culture, as indicated by the inward displacement of microspheres embedded in the ECM (Tamariz and Grinnel, 2002
). In general, the low density and random distribution of embedded beads does not provide a detailed mapping of ECM deformation at the sub-cellular level. Furthermore, cell-induced changes in collagen fibril organization cannot be visualized using this indirect approach.
In this study, we used high resolution time-lapse DIC imaging to directly investigate the dynamic effects of PDGF on the sub-cellular mechanical behavior and local collagen matrix interactions of corneal fibroblasts within 3-D collagen matrices. Addition of PDGF activated Rac and induced dramatic cell spreading, via both the extension of existing pseudopodial processes and the formation of new processes. Cell-induced displacement and realignment of collagen fibrils was also observed during PDGF-induced spreading. In general, relaxation (decompression) of the ECM was observed along the cell body, whereas tractional forces were generated by extending cell processes, as indicated by centripetal displacement of collagen fibrils at the ends of cells. Thus overall, there was a shift in the tractional force distribution from the center to the periphery of corneal fibroblasts in response to Rac activation.
Previous studies have demonstrated that PDGF-induced global reorganization of attached collagen matrices is partially dependent on ROCK (Rhee and Grinnell, 2006
; Tamariz and Grinnel, 2002
). To investigate the role of ROCK on the subcellular pattern of force generation in response to PDGF, we tested the effects of the specific Rho-kinase inhibitor Y-27632. Addition of Y-27632 following PDGF stimulation induced additional cell spreading and elongation. Interestingly, cells also assumed a more convoluted shape with dendritic cell processes, suggestive of a reduction of cellular tension. Overall, there was dramatic relaxation of cell-induced tractional forces following ROCK inhibition. When Y-27632 was washed out by switching the perfusion back to PDGF alone, cell processes again became thicker, and increased tractional forces were observed, particularly at the base of pseudopodial processes. Thus a large portion of the tractional forces observed during Rac-induced spreading were ROCK-dependent.
Various forms of crosstalk between Rho and Rac signaling pathways have been identified, but these can vary substantially depending on cell type and culture conditions (Brzeska et al., 2004
; Burridge and Wennerberg, 2004
; Jilkine et al., 2007
; Pestenjamasp et al., 2006
; Romano et al., 2006
). We did not assess the effect of PDGF on Rho or ROCK activation in the current study. However, a substantial reduction of cellular forces was observed when Y-27632 was added in the absence of PDGF, indicating a significant basal level of ROCK activity under the culture conditions used in this study. Subsequent addition of PDGF induced extensive branching and ruffling of pseudopodia, indicating that induction of cell spreading by PDGF does not require Rho Kinase activation.
Importantly, despite the overall decrease in cellular tractional forces in the presence of Y-27632, small inward displacements of collagen fibrils at the ends of extending pseudopodia were still observed during spreading. To determine whether these small tractional forces were generated through a myosin II dependent mechanism, we used the specific non muscle myosin II inhibitor blebbistatin. A recent study using tractional force microscopy demonstrated that ROCK plays a central role in producing myosin II-based tractional forces in NIH 3T3 fibroblasts (Beningo et al., 2006
). Consistent with these results, we found that ROCK and myosin II had similar effects on corneal fibroblast contractility in 3-D culture. However, despite an overall relaxation of cellular forces, small inward displacements of collagen fibrils continued to be generated at the tips of extending pseudopodia in the presence of blebbistatin, similar to those observed following treatment with Y-27632. Thus a novel finding in our study is that a component of tractional force generation by extending pseudopodia in 3-D matrices appears to be independent of both
ROCK or myosin II. It should be noted that our collagen matrices are much more compliant than the planar elastic substrates used in 2-D tractional force microscopy, and therefore may allow detection of smaller forces on the ECM (such as those required to displace collagen fibrils at the pseudopidal tips). Since PDGF-induced elongation, ruffling and branching of pseudopodia still occurs in the presence of Y-27632 and blebbistatin, it is tempting to speculate that this residual traction may be the result of protrusive forces associated with actin polymerization (Abraham et al., 1999
; Bohnet et al., 2006
; Marcy et al., 2004
; Pollard and Borisy, 2003
; Prass et al., 2006
Sheetz and coworkers recently demonstrated that fibroblasts plated on rigid substrates extend along isolated collagen fibrils placed on the upper surface of the cells, and retract them in a “hand over hand” cycle involving α2β1 integrin and myosin IIB (Meshel et al., 2005
). Fibers remained stationary as lamellipodia extended, whereas centripetal fibril displacement was correlated with lamellipodial retraction. While many aspects of the cell/collagen interactions and GFP-zyxin dynamics observed during Rac-induced spreading in the current study are consistent with this mechanism, collagen fibril displacement was observed during both the extension and retraction phases of pseudopodial movement in our 3-D model. This may by due to the fact that in 2-D culture, the protrusive forces generated during lamellipodial extension can be transmitted to the rigid substrate underneath the cell, without affecting collagen fibrils on top. In contrast all of the forces associated with pseudopodial extension are transmitted to the extracellular collagen matrix in our 3-D model.
An important feature of our experimental model is the ability to directly assess the displacement and realignment of collagen fibrils surrounding cells at high magnification. Thus to further investigate the mechanics of Rac-induced spreading, cellular interactions with individual collagen fibrils at the leading edge were studied. Two main patterns of collagen displacement were observed following PDGF stimulation. First, when a collagen fibril in front of an extending process was aligned nearly parallel to the direction of spreading, the extending process generally engaged the fibril, pulled it into alignment, then continued to spread along it. Second, when collagen fibrils in front of an extending process were aligned more perpendicular to the direction of spreading, the extending process often engaged the first fibril, push past it to engage the second fibril, then pulled the fibrils together. The first pattern of interaction tended to pull collagen fibrils beyond the ends of cells into an alignment parallel with the pseudopodia; whereas, the second pattern resulted in compaction of the collagen fibrils in a direction perpendicular to the extending process. These interactions may underlie, in part, the pattern of collagen organization observed using static reflected light confocal imaging following cell spreading, which includes sprays of collagen aligned parallel to the long axis at the ends of cells, and fibrils aligned more perpendicular to the long axis at the base of pseudopodial processes (Friedl and Brocker, 2000
; Kim et al., 2006
). Overall, our observations indicate that the pattern of pseudopodial extension and of cell-induced fibril displacement and realignment are significantly influenced by the initial organization of the surrounding ECM.