Materials and methods

Materials

Recombinant platelet-derived growth factor BB (PDGF-BB) was a gift from C. Hart (ZymoGenetics, Seattle, WA). Fibronectin and thrombin were provided by G. Marx (New York Blood Center, New York, NY). Fibronectin was 99% pure and intact by silver-stained PAGE. Bovine serum albumin (BSA; fatty-acid-free, fraction V) was obtained from Miles (Kankakee, IL). The chymotrysin digestion fragment of fibronectin (FN-120) was obtained from Gibco (Rockville, MD). Fibronectin peptide CS1 (EILDVPST) and CS1i (i=inactive, EILEVPST) (^{Humphries et al, 1987; Komoriya et al, 1991}) were purchased from Peninsula Laboratories Inc. (Belmont, CA). Fibronectin eptide I (YEKPGSPPREVVPRPRPGV) and scrambled peptide I (PEVVPPPVGPRPKRGRYSE) (^{McCarthy et al, 1988, 1990}); peptide II (KNNQKSEPLIGRKKT) and scrambled peptide II (LQEPKISTKNKGNRK) (^{McCarthy et al, 1988, 1990}); peptide III (YRVRVTPKEK TGPMKE) and scrambled peptide III (MKGKVTREVRYPTKPE) (^{Iida et al, 1992}); peptide IV (SPPRRARVT), a LL for RR substituted peptide IV (SPPLLARVT), and scrambled peptide IV (RVPRSTPAR) (^{Mooradian et al, 1993}); peptide V (WQPPRARI) and scrambled peptide V (RQAPRIPW) (^{Mooradian et al, 1993}); and CS5 (GEEIQIGHIPREDIDYHLYP) (^{Humphries et al, 1987; Mould et al, 1991}) were synthesized by SynPep (Dublin, CA). The purity of all peptides was higher than 97%.

Human fibrinogen was obtained from Calbiochem (San Diego, CA). To prevent fibrinolysis, plasminogen was removed from the fibrinogen by passage through a lysine–Sepharose 4B column (Pharmacia, Piscataway, NJ) (^{Deutsch and Mertz, 1970}). To remove any contaminating fibronectin, plasminogen-depleted fibrinogen was further passed through a gelatin–Sepharose 4B column (Pharmacia) (^{Engvall and Ruoslahti, 1977}). The removal of fibronectin was confirmed by SDS–PAGE and ELISA. Clottability after treatment with lysine–Sepharose 4B and gelatin–Sepharose 4B was over 90%. Tissue culture flasks were from Costar (Cambridge, MA), and 24-well tissue culture plates for all migration assays were from Becton-Dickinson (Lincoln Park, NJ).

Anti-3 antibody (clone P1B5) (^{Wayner and Carter, 1987; Wayner et al, 1988}) was purchased from Gibco BRL. Antibodies against 4 (clone P4G9) (^{Wayner et al, 1989; Massia and Hubbell, 1992}), HLA (mAb 1280) and vinculin (mAb 1624) were from Chemicon (Temecula, CA). Rabbit anti-fibronectin was obtained from Dako (Carpinteria, CA). Rabbit anti-fibrinogen and alkaline-phosphatase-conjugated goat anti-rabbit antibody were purchased from Cappel/Organon Teknika (Durham, NC). All experiments on animals and human tissue were approved by the compliance office at SUNY Stony Brook.

Porcine cutaneous wound model

Full-thickness excisional wounds were made by an 8-mm circular punch into the paravertebral skin of Minipigs. The wounds were dressed with Tegaderm, a polyurethane occlusive dressing, and harvested 1 and 3 d later as previously described (^{Welch et al, 1990}).

Specimens from all wound sites were fixed in formalin, paraffin embedded, sectioned at 5 m, and stained with Masson trichrome to delineate morphologic alterations. For immunohistochemistry the deparaffinized sections were first incubated with 0.4% pepsin in 0.1 M HCl (^{Folkvord et al, 1989}). Nonspecific antibody binding was blocked with 2% skim milk. Sections were incubated with the primary antibody (antifibronectin or antifibrinogen, diluted 1:1000 in phosphate-buffered saline (PBS)) overnight at 4°C, followed by a 1.5-h incubation with alkaline-phosphatase-conjugated anti-rabbit immunoglobulin (diluted 1:600 in PBS) at room temperature. Sections were then covered with Fast Red (BioGenex, San Ramon, CA) until red stain developed and counterstained with Mayer hematoxylin (Sigma, St. Louis, MO).

Cell culture

Primary cultures of human adult dermal fibroblasts (a gift from Marcia Simon, Living Skin Bank, SUNY at Stony Brook) were cultured in Dulbecco's modified Eagle's medium (DMEM, Life Technologies) containing 42 mM sodium bicarbonate, 100 U/mL penicillin, 100 g/mL streptomycin, and 10% fetal bovine serum (HyClone, Logan, UT) at 37°C and 5% CO₂/95% air in an humidified atmosphere. The cells were used between passages 4 and 12. HT-1080 cells were acquired from American Type Culture Collection (Rockville, MD) and also cultured in DMEM containing 42 mM sodium bicarbonate, 100 U/mL penicillin, 100 g/mL streptomycin, and 10% fetal bovine serum at 37°C and 5% CO₂/95% air in an humidified atmosphere.

Attachment assay

Each well of 24-well tissue culture plates (Becton-Dickinson) was coated with 120 nM intact fibronectin, FN-120, or one of several recombinant fibronectin fragments (^{Barkalow and Schwarzbauer, 1991, 1994}). Each chain of the fibronectin dimer contains one cell and one Hep-binding domain; therefore, because domain equivalency was desired, the molecular weight of a single chain (250,000 kDa) was used to calculate molar concentrations of fibronectin. Aliquots (450 L) of these proteins in DMEM were added to the appropriate wells. After 2 h of incubation at 37°C in 5% CO₂, plates were dried overnight at room temperature under sterile conditions. The next day plates were washed once with PBS and incubated with 2% BSA for 1 h at 37°C to block nonspecific binding sites.

The assay for measuring fibroblast attachment to matrix proteins was performed essentially as described (^{Gailit et al, 1993}) except that the cell number added to plates was lowered to 10,000 cells per well. In all experiments cells were allowed to attach for 60 min at 37°C before unattached cells were washed away and attached cells were fixed with 2% glutaraldehyde. After fixation, attached cells were air dried at room temperature and then 100 L of 0.1% crystal violet in 0.2 M boric acid, pH 9, was added to each well and the microtiter plate was shaken at 600 rpm on a plate mixer for 20 min. (The staining solution was prepared fresh from a stock solution of 5% crystal violet in 20% methanol.) Excess stain was removed by three washes with water. Stained cells were again air-dried before the crystal violet was solubilized by adding 100 L of 10% acetic acid to each well and then shaking the plate at 600 rpm for 20 min. Absorbance at 590 nm was measured with a dual-wavelength microtiter plate reader (THERMOmax, Molecular Devices, Menlo Park, CA) and that reading was corrected for light scattering by subtraction of the absorbance at 450 nm.

Cytoimmunochemistry of adherent cells

After cells were allowed to adhere to either 120 nM FN-120 or intact fibronectin for various time periods, they were fixed in 10% formaldehyde in PBS and then permeabilized with 0.4% Triton in PBS. Visualization of actin fibers and focal contacts was accomplished by incubating cells first with mouse monoclonal antivinculin antibody (Chemicon) followed by and FITC-conjugated sheep anti-mouse IgG in conjunction with TRITC-conjugated phalloidin as previously described (^{Chrzanowska-Wodnicka and Burridge, 1996}). Stained specimens were observed and photographed using a Nikon Microphot FXA epifluorescence microscope equipped with a Nikon FX-35DX 35-mm camera.

Migration assays

Outmigration and transmigration assays were performed as previously described (^{Greiling and Clark, 1997}).

Outmigration assay

Fibroblast cultures harvested at 80% confluence were washed twice with DMEM+2% BSA and resuspended at a concentration of 110⁶ cells/mL. Fibroblasts were mixed with neutralized collagen (Vitrogen 100, Celtrix Laboratories, Santa Clara, CA), 2% BSA, 30 ng/mL PDGF-BB, and 30 g/mL fibronectin in DMEM. A total of 600 mL of 110⁵ cells suspended in approximately 2 mg/mL collagen was added to the wells of a 24-well tissue culture plate, pre-coated with 2% BSA. After 2 h of incubation, gels were gently detached from the plastic surface to allow contraction. A quantity of 0.5 mL of DMEM with 2% BSA and 30 ng/mL PDGF-BB was added to each well, and gels were incubated overnight at 37°C in 5% CO₂. Another set of plates was coated with proteins as described under "Attachment assay." After washing the plate three times with PBS, contracted collagen gel constructs were attached to the coated plates. DMEM, 2% BSA, and 30 ng/mL PDGF-BB were added to assay plates so that the medium was level with the top of the collagen gel. Plates were incubated at 37°C in 5% CO₂ and 100% humidity for 24 h.

Transmigration assay

For three-dimensional transmigration assays, dried fibrin fibril-coated dishes were washed once with PBS, and contracted collagen gels were placed on the surface. Fibrinogen, fibronectin, and PDGF-BB were mixed with DMEM at a final concentrations of 300 g/mL, 30 g/mL, and 30 ng/mL, respectively, and then 1.0 U/mL thrombin was added and the clotting solution cast around the collagen gels. When needed, other materials such synthetic peptides were added to the mixture. The collagen gels were surrounded with the fibrinogen solution so that the fibrin gel was level with the top of the collagen gel. Plates were incubated at 37°C in 5% CO₂ and 100% humidity for 24 h.

Cell quantification for outmigration and transmigration assays

The number of migrated cells was quantified under a Nikon inverted-phase microscope by visually counting identifiable cell nuclei located outside of the contracted collagen gel in the fibrin gel (transmigration assay) or on the matrix (outmigration assay). Within a given experiment each condition was run in triplicate and meansSD were calculated. All experiments were repeated at least three times. Previous studies (^{Greiling and Clark, 1997}) demonstrated that cell proliferation did not contribute to cell number after a 24-h incubation.

Videomicroscopy migration assay

Plates were coated with proteins, washed with PBS, and incubated with 2% BSA for 1 h at 37°C to block nonspecific binding sites. After the plates were washed three times with PBS, 40,000 fibroblasts in 3 mL of CO₂-independent medium (Life Technologies) plus the appropriate biologic modifier were added. Cell movement was recorded using a Nikon Diaphot-TMD inverted microscope, equipped with an acrylic incubator housing and a Nikon NP-2 incubator for temperature control, coupled to a Dage-MTI CCD-72S camera and a JVC BR-9050U time lapse video cassette recorder. Plates were incubated at 37°C for 24 h. Images from a randomly selected field were recorded on videotape during the incubation period. Selected individual frames were transferred to a dedicated Macintosh Quadra 800 computer using Adobe PhotoShop. Aldus Superpaint and NIH Image software are applied to measure the distances traveled by individual cells (m nuclear displacement/24 h).

Agarose droplet assay

Cell migration on fibronectin and FN-120 was also quantified using the agarose droplet method of^{Varani et al (1978)}. Briefly, Immulon I 96-well plastic plates (Dynatech Laboratories, Chantilly, VA) were coated with fibronectin or FN-120 at various concentrations by drying overnight. The plates were then blocked with 2% BSA, rinsed with PBS, and dried. Cells were harvested and resuspended in serum-free growth medium containing 0.2% low-gelling-temperature agarose (FMC BioProducts, Rockland, ME) at a concentration of 3.310⁷ cells/mL. Droplets (1.5 L) of this cell suspension were carefully placed onto the Immulon I-pre-coated plates. After the agarose droplets were gelled by incubation at 4°C for 5 min, serum-free growth medium with 2% BSA was added with care to avoid detachment of the droplets. Cells were cultured at 37°C in 5% CO₂ for 24 h and then fixed with 2% glutaraldehyde, stained with crystal violet, destained in water, and air-dried. The outer perimeter of the radially migrating cells was digitized using a dissecting microscope connected to a computer. Using NIH Image software the total areas of cell outmigration was calculated. The baseline area of each agarose droplet was also quantified and subtracted from the outmigration area to assess the net area of outmigration.

Fibronectin reduction and alkylation

Fibronectin was reconstituted in 3 mL of 8 M urea, 0.5 M Tris-HCl, pH 8.3, and reduced by incubation for 30 min at room temperature with 40 mg of dithioerythritol (Sigma). Next 240 mg iodoacetic acid was added to alkylate the reduced fibronectin and the sample was incubated for 30 min at room temperature (^{Skorstengaard et al, 1982}). The solution was extensively dialyzed against sterile PBS. The monomeric state of reduced and alkylated fibronectin was confirmed on 6% SDS–PAGE run in nonreducing buffer.

Recombinant fibronectin fragments

III1-6, CHep, and Hep were expressed in bacteria as fusions with maltose-binding protein (MBP) as previously described (^{Aguirre et al, 1994}). Plasmids encoding MBP-CHepV and MBP-HepV were constructed using the vector pMALcRI (New England Biolabs, Beverly, MA). The alternatively spliced V region nucleotide sequence of fibronectin contains three closely spaced arginine codons of the sequence AGG/A (amino acid positions 2160, 2162, and 2164). tRNAs that recognize these codons are very rare in Escherichia coli. To improve translation of V-containing fusion proteins, these three codons were changed to CGA/T codons using degenerate oligonucleotides in combination with primers homologous to flanking fibronectin sequences for PCR amplification of the V region. The resulting PCR product extended from a BamHI site at position 6092 to a XbaI site engineered at position 6860 (^{Aguirre et al, 1994}). Its sequence was confirmed using Sequenase (US Biochemical, Cleveland, OH). This fragment was then inserted into pMALcRI-CHep and pMALcRI-Hep to yield pMALcRI-CHepV and pMALcRI-HepV, respectively. Plasmids were used to transform E. coli strain TB1. Fusion protein expression was induced and lysates were prepared as suggested by New England Biolabs.

Preparation of deletions and mutations within the Hep domain (RR/TM, III13a, -III13b14a, -III14b) have been previously described (^{Barkalow and Schwarzbauer, 1994}). RR/TM changes a pair of arginine residues to threonine and methionine; this mutation ablates heparin binding. –13a removes residues 7–41 of repeat III13. –13b14a deletes residues 44–90 of III13 and 1–39 of III14. –14b removes amino acids 82–90 of III14 and 1–4 of III15. Restriction fragments containing each of these mutations were inserted into pMALcRI-CHep or pMALcRI-Hep plasmids as indicated. Fusion proteins were expressed in E. coli TB1.

MBP-III1-6, MBP-CHep, MBP-CHep (RR/TM), and deletion mutants were purified by amylose resin affinity chromatography. MBP-CHepV, MBP-HepV, and MBP-Hep fusion proteins were purified by precipitation with 35% ammonium sulfate followed by heparin–agarose affinity chromatography (^{Aguirre et al, 1994}). Protein purity was checked by SDS–PAGE.

Results

During tissue healing, fibrin/fibronectin bridges appear to connect periwound stroma to the fibrin-laden wound directly before fibroblast invasion

In tissue repair there is a 3-d delay between injury and fibroblast invasion of the fibrin clot-filled wound (^{Clark, 1996}). In porcine cutaneous wounds, fibrinogen and fibronectin accumulated in the periwound stroma on the third day after injury forming a conduit between collagenous fibrous septae in the subcutaneous connective tissue and the fibrin clot in the wound space (Figure 1). One-day wounds (Figure 1a,c,e) showed a well-defined boundary between the fibrin clot in the wound and the underlying subcutaneous tissue. Furthermore, the fibrous septae within the subcutaneous tissue showed tightly packed collagen bundles without obvious cell infiltrate or fluid extravasation. In contrast, 3-d wounds (Figure 1b,d,f) showed extravasated fluid as well as fibronectin (Figure 1b) and fibrin (Figure 1d) intercalated between adipocytes in proximity of the wound space and between collagen bundles in proximal fibrous septae. Because fibroblasts do not invade the fibrin clot in a wound until 4 d after injury (^{McClain et al, 1996}), the presence of the fibrin/fibronectin bridges between the collagenous dermis and fibrin clot may be critical for the cell influx. Our recent in vitro observations that fibronectin is required for human dermal fibroblast transmigration from a collagenous matrix into a fibrin clot supports this contention (^{Greiling and Clark, 1997}).

Figure 1.

Photomicrographs of day 1 (A,C,E) and day 3 (B,D,F) porcine wounds stained immunohistochemically for the detection of fibronectin (A,B) and fibrinogen (C,D). Adjacent tissue section were incubated with nonimmune IgG antibodies for controls (E,F). Arrows, border between the wound (above) and the subcutaneous tissue (below). At day 1, FN (A) and fibrinogen (C) are only detectable in the wound area (area above the arrows, red) but not in the subcutaneous tissue and fibrous septae. In contrast, at day 3 fibronectin (B) and fibrinogen (D) are both found in the wound space as well as in the subcutaneous tissue. (E) and (F) are negative controls for day 1 (E) and day 3 (F) in which the first antibody was substituted by nonimmune rabbit IgG. Bar, 500 m.

Full figure and legend (212K)

The RGD containing FN-120 fragment of fibronectin supports initial adult dermal fibroblast attachment and spreading but does not sustain it

The carboxy-terminal half of fibronectin contains three major domains that can mediate cell–fibronectin interactions. To determine the contributions of the cell (C or FN-120), heparin (Hep), and IIICS (V) domains to attachment and migration of human adult dermal fibroblasts, a variety of proteolytic and recombinant fragments were used (Figure 2). These proteins contained the cell-binding domain alone (FN-120), the heparin-binding domain alone (Hep), or combinations of two or more domains (CHep, HepV, CHepV). The cell-binding domain is abbreviated C when it was generated as a recombinant protein and FN-120 when it was generated from proteolytic cleavage of plasma fibronectin. CHep-RR/TM is equivalent of CHep but with a pair of Arg residues mutated to Thr and Met in repeat III13. These mutations ablate heparin binding activity (^{Barkalow and Schwarzbauer, 1991}). III1-6 contained none of the cell interactive domains and acted as a negative control.

Figure 2.

A diagram showing the domain organization and modular structure of the plasma fibronectin dimer is shown at the top. Fibronectin fragments and recombinant proteins used in this study are aligned with corresponding fibronectin repeats in the intact molecule. Small rectangles, fibronectin type I repeats; ovals, fibronectin type II repeats; squares, fibronectin type III repeats (^{Yamada and Clark, 1996}). FN-120 is a proteolytic fragment of fibronectin. III1-6, CHepV, HepV, CHep, and Hep were expressed in bacteria as fusions with MBP (^{Barkalow and Schwarzbauer, 1991; Aguirre et al, 1994}). V, an alternatively spliced variant which contains the IIICS region.

Full figure and legend (40K)

As a first step toward understanding the role of these domains in fibroblast motility, we investigated their ability to mediate fibroblast attachment in an 1-h assay. Human adult dermal fibroblasts had optimal attachment for intact fibronectin when coated on plastic at a concentration of 30 g/mL (120 nM) (Figure 3) (dose response was performed from 1 to 100 g/mL but is not shown). Surfaces coated with 120 nM CHepV or CHep-RR/TM had the same adhesivity for fibroblasts as intact fibronectin. CHep promoted somewhat less adhesive activity (Figure 3). By itself the Hep was non adhesive for fibroblasts. FN-120 at 15 g/mL (120 nM) promoted 70% attachment (Figure 3) and at 30 g/mL (400 nM) promoted 90% attachment (not shown). Spreading and focal contact formation of fibroblasts on FN-120 was similar to intact fibronectin at 4 h (Figure 4a and Figure 4b,c and Figure 4d, respectively) and at 8 h (not shown). By 12 h, however, fibroblast spreading and focal contact formation on FN-120 became less robust (Figure 4e,f) than on intact fibronectin (Figure 4g,h). By 24 h fibroblasts plated on FN-120 began to collapse into a spindle-shaped morphology with less distinct actin bundles and a relative paucity of focal contacts (Figure 4i,j). In contrast, fibroblast on intact fibronectin appeared more spread at 24 h than at 4 h (Figure 4k,l). Experiments with primary fibroblasts plated on the recombinant CHepV protein demonstrated that actin bundle arrays and focal contact formation was indistinguishable from those formed in primary fibroblast plated on intact fibronectin at 2, 4, 8, 12, and 24 h (data not shown). Thus, the relatively good initial fibroblast attachment and spreading on FN-120 dissipated over a 24-h incubation, whereas attachment and spreading on intact FN was sustained.

Figure 3.

Fibroblast attachment to fibronectin, 120-kDa fibronectin fragment containing the RGD cell-binding domain (FN-120), and recombinant fibronectin moieties. CHep-RR/TM contains the RGD cell-binding and heparin domains with mutations of the RR pair in repeat III13 to Thr-Met as previously described (^{Barkalow and Schwarzbauer, 1991}). Proteins were coated on the assay plates by drying 120 nM solutions overnight at room temperature. Plates were blocked with 2% BSA and washed, and then 10,000 fibroblasts per well were added. After incubation for 1 h, attached cells were fixed with 2% glutaraldehyde and stained with 0.1% crystal violet. Absorbance of solubilized crystal violet was measured at 590 nm with a dual-wavelength, microtiter plate reader (THERMOmax, Molecular DevicesCA). Experimental conditions were run in triplicate and data are displayed as meansSD.

Full figure and legend (14K)

Figure 4.

Fibroblast spreading and focal contact formation on FN-120 (A and B,E and F,I and J) and intact fibronectin (C and D,G and H,K and L). Fibroblasts were incubated in protein-coated Lab-Tec chambers with DMEM and 2% albumin for 4 (A–D), 12 (E–H), or 24 (I–L) h at 37°C and then fixed, permeabilized, and stained for vinculin with mouse monoclonal antibody and secondary FITC-anti-mouse IgG (A,C,E,G,I,K) and for actin with TRITC-phalloidin (B,D,F,H,J,L). Bar, 5 m.

Full figure and legend (183K)

FN-120 does not support adult human fibroblast migration

Next we investigated whether PDGF-stimulated adult human fibroblasts could migrate from a contracted collagen gel (a dermal organotypic construct) over surfaces coated with either intact fibronectin or FN-120. Despite supporting nearly equivalent initial attachment and spreading compared to intact fibronectin (Figure 3, Figure 4), FN-120 did not support fibroblast migration at any concentration between 1 and 400 nM (400 nM shown in Figure 5). In contrast, robust migration was observed on intact fibronectin (Figure 5). Reduction and alkylation of fibronectin dimer to the monomeric state did not affect the ability of fibronectin to support either attachment or PDGF-BB-stimulated migration (data not shown).

Figure 5.

Fibroblast migration on intact fibronectin and FN-120. An organotypic construct of human dermal fibroblasts in collagen gels containing 30 g/mL fibronectin was pasted on tissue culture dishes coated with fibronectin or FN-120 as previously described (^{Greiling and Clark, 1997}). Cultures were incubated for 24 h in medium containing 30 ng/mL PDGF-BB. The nuclei of cells that had migrated out of the collagen gels were counted visually. Each histogram represents the meanSD of cells migrating out of triplicate constructs.

Full figure and legend (21K)

To further explore whether primary adult human fibroblasts could move on FN-120, cells were plated directly onto FN-120-coated plastic wells and observed with videomicroscopy for 24 h at 37°C (Table I). Whereas intact fibronectin supported baseline fibroblast motility, FN-120 did not. Neither phorbol 12-myristate 13-acetate nor PDGF-BB was able to stimulate substantial cell movement on FN-120 whereas both increased fibroblast migration on intact fibronectin. Additionally, normal adult human mesenchymal cell migration was studied using the agarose droplet assay (^{Varani et al, 1978}). None of seven strains of adult primary human dermal fibroblasts nor four strains of adult primary human smooth muscle cells could migrate out of an agarose droplet over a surface coated with FN-120 in the absence or presence of PDGF-BB (data not shown). In contrast, all strains of adult fibroblasts and smooth muscle cells showed robust movement on intact fibronectin using this assay. To determine whether the FN-120 preparation used in our assays could support the movement of any cell type, we investigated the motility of HT-1080 fibrosarcoma cells and human neonatal fibroblasts. Both immortilized fibrosarcoma cells and neonatal fibroblasts moved approximately the same distance over FN-120 as they did over intact fibronectin (data not shown). These results are in concert with those previously reported by^{Nagai et al (1991)} and by^{Akiyama et al (1989)}, respectively.

Table I - Fibroblast movement on intact fibronectin and FN-120.

Full table

Optimal fibroblast migration on fibronectin requires the cell and heparin-binding domains and the variably spliced IIICS domain

The Hep, composed of the 12, 13, and 14 type III repeats adjacent to the carboxy-terminus of FN-120 (Figure 2), was examined for its contribution to adult human fibroblast motility. Fibroblast migration on plates coated with Hep was less than 20% of the migration observed on intact fibronectin (Figure 6). Coating plates with a combination of FN-120 fragments and recombinant Hep protein, which together presented the cell- and heparin-binding domains in a noncontiguous array, supported fibroblast migration to approximately 45% of the maximum level observed with intact fibronectin (Figure 6). Recombinant CHep protein, which contains contiguous cell- and heparin-binding domains in one molecule (Figure 2), supported migration to approximately 55% of the maximum (Figure 6). Thus, similar levels of migration occurred whether the cell- and heparin-binding domains were contiguous or noncontiguous. In contrast, neither FN-120 fragment alone, nor the control recombinant III1-6 protein without any known cell-binding sites (Figure 2), allowed any migration (Figure 6).

Figure 6.

Fibroblast migration on intact fibronectin (FN), fragments of fibronectin containing the RGD cell-binding domain (FN-120), and recombinant RGD cell-binding (C), heparin (Hep), and IIICS domains (V) of fibronectin. Assay plates were coated with 4–400 nM FN-120 or recombinant proteins as described under Methods. The coating efficiency of all proteins was essentially the same as judged by the bicinchoninic acid protein assay (^{Tuszynski and Murphy, 1990}). After blocking with 2% BSA, organotypic constructs of fibroblasts in a collagen gel containing fibronectin and PDGF-BB were pasted onto the protein-coated assay plates. DMEM containing 2% BSA and 30 ng/mL PDGF-BB was added so that the fluid level was even with the top of the collagen gel. Fibroblasts that had migrated over the protein-coated surface after a 24-h incubation were quantified by visual counting. All experimental conditions were run in triplicate, and data are displayed as meansSD. All data were normalized to maximum migration of intact fibronectin.

Full figure and legend (31K)

To test the role of the IIICS segment (V), which contains cell-binding sites for 41 integrin, recombinant HepV and CHepV were used for migration assays. As shown in Figure 6, HepV supported 40% maximal migration, whereas HepV plus FN-120 supported migration comparable to that of intact fibronectin. Essentially the same result was obtained with recombinant CHepV. In summary, maximal levels of PDGF-stimulated fibroblast migration were only observed when the RGD cell-binding domain, the heparin-binding domain, and the IIICS domain, were available on a surface, regardless of whether the domains were in a contiguous or noncontiguous array.

Heparin-binding activity is required for fibroblast migration

Involvement of the heparin-binding domain raises the possibility that interactions with glycosaminoglycans might be important for adult dermal fibroblast migration on fibronectin. To examine this possibility, uncharged amino acids were substituted for the arginine-pair (RR) in repeat III13 of CHep (CHep-RR/TM). Heparin and chondroitin sulfate interactions were eliminated by this substitution as previously described (^{Barkalow and Schwarzbauer, 1991}) and adult dermal fibroblast migration was completely abolished compared to approximately 50% maximal migration on nonmutated recombinant protein CHep and maximal migration on intact fibronectin (not shown). In contrast to its inability to support migration, CHep-RR/TM supported fibroblast attachment at levels comparable to intact fibronectin (Figure 3).

To test whether other short sequences within the heparin domain were required for cell mobility, deletion mutants of the domain were constructed. Hep-13a, Hep-13b14a, and Hep-14b were missing selected sequences of amino acids in repeats III13 and III14 (^{Barkalow and Schwarzbauer, 1991}). Hep and its mutants supported little migration by themselves at concentrations up to 400 nM (Table II). FN-120 in combination with Hep, however, supported migration just short of 50% that observed on intact fibronectin, whereas FN-120 addition to the Hep deletion mutants supported little, if any, migration (Table 2). Interestingly, the Hep-13a mutant, which deletes a segment including the RR pair, supported the least migration. These results taken together clearly indicate that multiple regions within the heparin domain are required for optimal adult dermal fibroblast migration and that the RR pair in repeat III13 is particularly important.

Table II - Fibroblast migration on Hep and its deletion mutants.

Full table

Maximal fibroblast migration on fibronectin requires 41

Involvement of the variable IIICS domain, as well as the heparin domain, raises the possibility that 41 might be important for adult human fibroblast migration on fibronectin. Whereas 51 and v3 both bind fibronectin at the RGD cell-binding site (^{Ruoslahti, 1988}), 41 binds fibronectin at several additional sites: the carboxy-terminus of the heparin-binding domain (^{Mould and Humphries, 1991}) and the CS1 (^{Komoriya et al, 1991}) and CS5 (^{Mould et al, 1991}) peptide sequences in the IIICS domain. We previously showed that adult human fibroblasts expressed functional 41 (^{Gailit et al, 1993}). As demonstrated by Figure 7, monoclonal antibody P4G9 against the integrin 4 subunit inhibited fibroblast migration on intact fibronectin approximately 50%, whereas anti-HLA monoclonal antibody HLA and P1B5 to 3 subunit (not shown) had no effect. P4G9 also inhibited migration of adult human fibroblast on CHep about 50%.

Figure 7.

P4G9 monoclonal antibody to 4 inhibits fibroblast migration on intact fibronectin (FN) and the recombinant protein containing the cell and heparin binding domains (CHep). The outmigration assay as described in the legend to Figure 6 was allowed to proceed for 24 h in the presence or absence of monoclonal antibodies and then the migrated cells were visually counted. Experimental conditions were run in triplicate, and data are displayed as meansSD. All data were normalized to maximum migration on intact fibronectin in the absence of antibodies. Controls included mouse monoclonal anti-HLA and mouse monoclonal P1B5 to 3-integrin subunit neither of which inhibited movement (data not shown).

Full figure and legend (12K)

Biologically active synthetic peptides constructed from sequence arrays within the heparin- and IIICS-binding domains block fibroblast migration on fibronectin

To further define which heparin and IIICS subdomains are involved in adult dermal fibroblast migration on fibronectin, synthetic peptides were manufactured from sequences known to interact with cells in repeats III12, III13, and III14 (peptides III, IV, V, I, and II, respectively) and the IIICS segment (CS1 and CS5) (Figure 8). When added to the outmigration assay, peptides III, IV, V, I, and II inhibited the migration of fibroblasts over fibronectin in a concentration-dependent manner (Figure 9). CS1 had a slight enhancing effect at higher concentrations, whereas CS5 did not influence migration at tested concentrations. A scrambled peptide of CS1 (CS1i) used as a control did not influence migration. To determine whether the observed effect of synthetic peptides on the migration was specific for fibronectin, collagen-coated assay plates were substituted for fibronectin coated plates. Peptides III, IV, and V had no effect on migration over collagen, whereas peptides I and II had an inhibitory effect in a concentration-dependent manner. CS1 enhanced the migration at higher concentrations, whereas CS1i did not influence the activity.

Figure 8.

Localization of peptide sequences within the heparin and IIICS domains.III, IV, V, I, II, CS1, and CS5 indicate sites of sequence arrays used to model synthetic peptides. CS/H, HS, and 41 indicate putative binding sites for chondroitin sulfate, heparin, heparan sulfate, and the integrin receptor 41, respectively.

Full figure and legend (11K)

Figure 9.

Synthetic peptides III, IV, and V from the heparin-binding domain inhibit fibroblast migration on fibronectin, but not on collagen substratum. Assay plates were coated with either 120 nM fibronectin (open triangles) or collagen (closed triangles) as described under Methods. Synthetic peptides (III, IV, V, I, II, CS1, CS5, and control peptide CS1i) were added at the concentrations shown to assay medium (DMEM, 2% BSA) in the presence of 30 ng/mL PDGF-BB. Fibroblasts that had migrated over the protein-coated surface after a 24-h incubation were quantified by visual counting. All data were normalized to maximum migration of intact fibronectin without addition of synthetic peptides (closed circle). No migration was observed on fibronectin without PDGF-BB (not shown). All experimental conditions were run in triplicate, and data are displayed as meansSD.

Full figure and legend (32K)

The ability of peptides I and II to inhibit adult human fibroblast migration on both fibronectin and collagen raises the possibly of nonspecificity, especially because peptides I–V have a net positive charge. Therefore, we constructed scrambled controls of peptides I–V. Furthermore, because peptide IV (SPPRRARVT) contains the heparin-binding RR pair, LL was substituted for RR (SPPLLARVT) in an additional peptide IV control. As shown in Figure 10, peptides I–V (Figure 10, open squares) inhibited adult human fibroblast migration on intact fibronectin, whereas scrambled controls (Figure 10, closed squares) did not. Interestingly, the peptide IV derivative carrying the LL substitution for RR (Figure 10, crossed triangle) did not inhibit except modestly at the highest concentration. These data taken together indicate that sequence is important for function of these peptides.

Figure 10.

Synthetic peptides from the heparin-binding domain, but not related scrambled peptides, inhibit fibroblast migration on fibronectin. Assay plates were coated with 120 nM fibronectin. Synthetic peptides III, IV, V, I, and II (open squares), scrambled variants (closed squares), or a LL substitution for RR in peptide IV (crossed triangle) were added at the concentrations shown to assay medium (DMEM, 2% BSA) in the present of 30 ng/mL PDGF-BB. Fibroblasts that had migrated over fibronectin coated surface after a 24-h incubation were quantified by visual counting. All data were normalized to maximum migration of intact fibronectin without addition of synthetic peptides (closed circles). No migration was observed on fibronectin without PDGF-BB (not shown). All experimental conditions were run in triplicate, and data are displayed as meansSD.

Full figure and legend (32K)

The apparent lack of inhibitory effect of CS1 and CS5 on fibroblast movement on fibronectin (Figure 9) was surprising because the presence of the IIICS domain (V) was required for maximal migration (Figure 6). Therefore, CS1 and CS5 synthetic peptides were added to the migration assay together. As shown in Figure 11, when present concomitantly, CS1 and CS5 inhibited migration in a concentration-dependent fashion.

Figure 11.

CS1 and CS5 peptides conjointly inhibit fibroblast migration on fibronectin. (Top) In the presence of 100 M CS1 cell-binding peptide, the CS5 cell-binding peptide progressively inhibits fibroblast migration on fibronectin with concentrations up to 100 M (closed triangles). By itself 100 M CS1 had no effect of migration (closed circle). (Bottom) Likewise, in the presence of 100 M CS5 cell-binding peptide, the CS1 cell-binding peptide progressively inhibits migration with concentrations up to 100 M (closed triangles). By itself 100 M CS5 had no effect of migration (closed circle).

Full figure and legend (18K)

Synthetic peptides block fibroblast invasive migration into fibrin gels

To determine whether the results obtained from the two-dimensional outmigration assay related to more complex, three-dimensional transmigration, the same synthetic peptides were added to the fibrin/fibronectin gel in the transmigration assay as previously described (^{Greiling and Clark, 1997}). All peptides, except peptide III, gave essentially the same results (data not shown). Surprisingly, peptide III had no effect on fibroblast transmigration from collagen matrix into a three-dimensional array of fibrin and fibronectin, which is in contrast to its ability to completely inhibit migration over a two-dimensional fibronectin coat (Figure 10). To determine whether the interaction of fibrin with fibronectin might have led to the apparent discrepancy between transmigration and outmigration, assay plates were coated with fibronectin in the presence of dried fibrin fibrils. Migration over two-dimensional dried fibrin/fibronectin fibrils was also not influenced by peptide III (data not shown). Therefore, interactions between fibrin and fibronectin may have abrogated the ability of peptide III to inhibit migration.

Discussion

The data presented in this report demonstrate that three fibronectin molecular domains, the RGDS cell-binding, the heparin-binding, and the alternatively spliced IIICS domains, are necessary and sufficient for optimal migration of primary human fibroblasts on fibronectin. In fact, multiple strains of normal human fibroblasts and smooth muscle cells failed to migrate on FN-120, a fragment containing the classic RGDS cell-binding site and the PHSRN synergy site. This is in marked contrasted to the ability of certain human fibrosarcoma cell lines (HT 1080) to migrate on FN-120 and on even smaller subfragments of FN-120 (^{Akiyama et al, 1995}) or rat NR6 cells ability to migrate of RGD nanoscale clusters (^{Maheshwari et al, 2000}). Likewise involuting marginal zone cells from a Xenopus embryo migrate on fusion proteins containing the RGD and synergy site sequences PHSRN (^{Ramos and DeSimone, 1996}). The studies of HT 1080 (^{Akiyama et al, 1989; Nagai et al, 1991; Akiyama et al, 1995}), NR6 cells (^{Maheshwari et al, 2000}), and involuting marginal zone cell (^{Ramos and DeSimone, 1996}) migration on RGD clusters or RGD-containing fibronectin fragments, in fact, have led some investigators to extrapolate that fibroblasts, in general, only need the fibronectin cell-binding domain for migratory activity (^{Ramos and DeSimone, 1996}).

FN-120 supported initial attachment of adult human fibroblast, albeit suboptimally compared to intact fibronectin or recombinant CHepV, but could not support sustained spreading. The inability of fibronectin fragments containing only the RGD cell-binding site to support actin stress fiber and focal contact formation was first reported in human embryonic fibroblasts 15 y ago (^{Woods et al, 1986}). In that report Woods et al demonstrated that the heparin domain was required for such cytoskeletal organization. More recently, investigators using fibroblast cell lines (MG-63, Rat1, and NIH 3T3 cells) found that actin bundles and focal contacts were not well formed on FN-120 or a recombinant fibronectin containing the 7–11 type III repeats (called FN 7–11) even at 2 h (^{Bloom et al, 1999}). Our data with primary human adults dermal fibroblasts differ from both of these studies in that actin stress fibers and focal contacts are formed when cells are plated on FN-120 for up to 8 hr, but this fibronectin fragment containing the RGD cell-binding domain alone cannot sustain cytoskeletal organization.

Previously we demonstrated that the RGDS cell-binding peptide within FN-120, and that 51 and v3 integrins, which bind FN-120 through the RGDS site, were required for fibroblast migration over fibronectin (^{Greiling and Clark, 1997}). These data taken together with the results presented here predict that the FN-120 fragment is necessary, but not sufficient, for the movement of normal adult human fibroblasts over fibronectin. Thus, other fibronectin domains that promote attachment and/or migration of other cells were examined for their ability to support adult human fibroblast migration.

Hep interacts with at least two distinct classes of cell surface receptors: 41 integrin receptors and heparan sulfate and chondroitin sulfate glycosaminoglycans found on cell surface proteoglycan receptors (^{McCarthy et al, 1990; Barkalow and Schwarzbauer, 1991, 1994; Drake et al, 1992; Iida et al, 1992; Woods et al, 1993}). Recombinant Hep protein alone supported only 20% maximal migration, but CHep or Hep plus FN-120 supported migration to approximately 55% maximum. Thus, half-maximal migration occurred whether the RGD cell- and heparin-binding domains were contiguous or noncontiguous. Similarly, heparin-binding repeat III13 plus cell-binding repeats III7–11 were found necessary and sufficient for actin bundle and focal contact formation regardless of their spatial relationship to each other (^{Bloom et al, 1999}).

Deletion mutants within the Hep domain indicate that repeats III13 and III14 were both involved in adult human fibroblast migration. These two repeats have also been implicated in formation of focal contacts. Repeat III13 in conjunction with a recombinant cell-binding domain was able to induce actin bundles and focal contacts in fibroblast cell lines (^{Bloom et al, 1999}). A peptide (WQPPRARI) from repeat III14 was shown to promote focal attachment formation in primary human embryonic fibroblasts (^{Woods et al, 1993}). In vitro binding assays have shown that both repeats III13 and III14 contribute to heparin-binding activity (^{Ingham et al, 1993}). The discontinuity of residues involved in heparin binding has been confirmed by structural modeling studies (^{Busby et al, 1995}). Therefore, it is not surprising that multiple sites within the heparin-binding domain participate in this process.

At least five peptide sequences have been implicated in cell interactions with the heparin domain (^{McCarthy et al, 1988; Iida et al, 1992}). Peptide III from repeat III12 is functional in melanoma cell binding to fibronectin (^{Iida et al, 1992}). Peptide IV from repeat III13 promotes rabbit corneal epithelial cell attachment and spreading (^{Mooradian et al, 1993}). Peptide V from repeat III14 promotes multiple cell activities including rabbit corneal epithelial cell attachment, spreading, and migration (^{Mooradian et al, 1993}); focal contact formation in human embryonic fibroblasts (^{Woods et al, 1993}); human keratinocyte attachment and spreading (^{Wilke et al, 1993}); endothelial cell spreading and migration (^{Huebsch et al, 1995}); and rabbit synovial fibroblast expression of MMPs (^{Huhtala et al, 1995}). Peptides I and II from repeat III14 also mediate attachment and spreading of melanoma cells (^{McCarthy et al, 1988, 1990}), human keratinocytes (^{Wilke et al, 1991}) and rabbit corneal epithelial cells (^{Mooradian et al, 1992}).

All of these peptides inhibited primary fibroblast migration on fibronectin, whereas the scrambled peptides had no effect. Surprising to us, peptides I and II also inhibited fibroblast migration on type 1 collagen. These two peptides, however, can bind chondroitin sulfate proteoglycans (^{Iida et al, 1992}), and a CD44-related chondroitin sulfate proteoglycan is required for melanoma cell motility on type 1 collagen (^{Faassen et al, 1993}). The need for intact heparin domain type III repeats 12–14 and the almost complete inhibition of migration by a variety of heparin domain peptides suggest that a high degree of cooperativity must occur within the heparin domain for proper receptor interaction. The three-dimensional structure of the heparin domain, in fact, predicts that this might be so (^{Sharma et al, 1999}).

In addition to the cell- and heparin-binding domains, the variable IIICS is also required for maximal fibroblast migration. Within this segment are CS1 and CS5 peptide sequences that bind 41 integrin (^{Komoriya et al, 1991; Mould et al, 1991}). Recombinant protein containing cell, heparin, and IIICS domains supported migration comparable to intact fibronectin. Furthermore, maximal fibroblast migration was observed regardless of whether cell, heparin, and IIICS domains were contiguous. These data taken together indicate that multiple receptors interacting with multiple sites on fibronectin are needed for adult human fibroblast migration.

Previously, we demonstrated that both 51 and v3 are required for human fibroblast migration (^{Greiling and Clark, 1997}). The ability of P4G9 anti-4 antibody to inhibit human dermal fibroblast migration on fibronectin, albeit weakly, indicates that 41 is also necessary for optimal human fibroblast motility. These data are in concert with observations that activation of the 4 cytoplasmic tail is required for rhabdomyosarcoma cell movement (^{Chan et al, 1992}). Other cell lines, such as melanoma cells (^{Humphries et al, 1986; Mould et al, 1994}) and murine S180 cells transfected with 41 (^{Beauvais et al, 1995}), have the capacity to migrate on isolated IIICS domain using 41 integrin. 41 binds IIICS at two sites, CS1 (^{Komoriya et al, 1991}) and CS5 (^{Mould et al, 1991}), and more weakly binds the heparin domain near its carboxy-terminal end (^{Mould and Humphries, 1991}). The ability of anti-41 antibodies to partially inhibit fibroblast migration on fragments lacking IIICS suggests that the heparin-binding site for 41 may be functional. That CS1 and CS5 synthetic peptides added singly fail to block fibroblast migration suggests that 41 may use any one of its various binding sites on fibronectin alternatively to promote movement. Taken together, our data demonstrate that a minimum of three integrin receptors are needed for maximal adult human fibroblast migration on fibronectin.

The heparin domain requirement raises the possibility that one or more proteoglycan receptor(s) may be needed for adult human fibroblast motility. Two such receptors have been identified: a dermatan sulfate CD44 proteoglycan (^{Clark et al, 2003}) and syndecan-4 (Lin and Clark, manuscript in preparation). Thus, the movement of normal human mesenchymal cells is highly restrained on fibronectin necessitating the presence of three distinct functional domains and the proper complement of expressed and activated cell surface receptors. Furthermore, deposits of fibronectin may be requisite for movement of mesenchymal cells in the wounded animal.

This antimigratory characteristic of normal connective tissue cells may explain, in part, why stromal cells are relatively stationary in the absence of architectural boundaries such as basement membranes. Furthermore, the high stringency for mesenchymal cell migration on fibronectin may explain why tissue cells fail to repopulate chronic wounds which abound with proteases (^{Wysocki et al, 1993; Grinnell and Zhu, 1994}) and demonstrate fragmentation and loss of tissue fibronectin (^{Grinnell et al, 1992; Herrick et al, 1992}). Loss of migratory constraint on fibronectin can occur in embryogenesis, as demonstrated by involuting marginal zone cells (^{Ramos and DeSimone, 1996}), and in malignancy as evident with fibrosarcoma (^{Akiyama et al, 1989, 1995; Nagai et al, 1991}), squamous cell carcinoma (^{Nagai et al, 1991}), and melanoma cells (^{Humphries et al, 1986; Mould et al, 1994}). Because fibronectin is widely expressed during embryogenesis and widely deposited at sites of cancer, unconstrained cells would have a road to move on.

References

References

1.	Aguirre KM, McCormick RJ & Schwarzbauer JE. Fibronectin self-association is mediated by complementary sites within the amino-terminal one-third of the molecule. J Biol Chem (1994) 0 27863–27868. | ChemPort |
2.	Akiyama SK, Aota S & Yamada KM. Function and receptor specificity of a minimal 20 kilodalton cell adhesive fragment of fibronectin. Cell Adhes Commun (1995) 3: 13–25. | PubMed | ISI | ChemPort |
3.	Akiyama SK, Yamada SS, Chen W-T & Yamada KM. Analysis of fibronectin receptor function with monoclonal antibodies: Roles in cell adhesion, migration, matrix assembly, and cytoskeletal organization. J Cell Biol (1989) 109: 863–875. | Article | PubMed | ISI | ChemPort |
4.	Barkalow FJ & Schwarzbauer JE. Interactions between fibronectin and chondroitin sulfate are modulated by molecular context. J Biol Chem (1994) 269: 3957–3962. | PubMed | ISI | ChemPort |
5.	Barkalow FJ & Schwarzbauer JE. Localization of the major heparin-binding site in fibronectin. J Biol Chem (1991) 266: 7812–7818. | PubMed | ISI | ChemPort |
6.	Beauvais A, Erickson CA, Goins T, Craig SE, Humphries MJ, Thiery JP & Dufour S. Changes in the fibronectin-specific integrin expression pattern modify the migratory behavior of sarcoma S180 cells in vitro and in the embryonic environment. J Cell Biol (1995) 128: 699–713. | Article | PubMed | ISI | ChemPort |
7.	Bloom L, Ingham KC & Hynes RO. Fibronectin regulates assembly of actin filaments and focal contacts in cultured cells via the heparin-binding site in repeat III13. Mol Biol Cell (1999) 10: 1521–1536. | PubMed | ISI | ChemPort |
8.	Busby TF, Argraves WS, Brew SA, Pechik I, Gilliland GL & Ingham KC. Heparin binding by fibronectin module III-13 involves six discontinuous basic residues brought together to form a cationic cradle. J Biol Chem (1995) 270: 18558–18562. | Article | PubMed | ISI | ChemPort |
9.	Chan BM, Kassner PD, Schiro JA, Byers R, Kupper TS & Hemler ME. Distinct cellular functions mediated by different VLA integrin subunit cytoplasmic domains. Cell (1992) 68: 1051–1060. | Article | PubMed | ISI | ChemPort |
10.	Chrzanowska-Wodnicka M & Burridge K. Rho-stimulated contractility drives the formation of stress fibers and focal adhesions. J Cell Biol (1996) 133: 1403–1415. | Article | PubMed | ChemPort |
11.	Clark RAF. Wound repair: Overview and general considerations. In: Clark R (edMolecular and Cellular Biology of Wound Repair (1996) New York: Plenum Press p3–50.
12.	Clark RAF, Lin F, Greiling D & Couchman JR. Fibroblast invasive migration into fibronectin/fibrin gels requires dermatan sulfate-CD44 proteoglycan. J Invest Dermatol (2003) in press.
13.	Deutsch DG & Mertz ET. Plasminogen: Purification from human plasma by affinity chromatography. Science (1970) 170: 1095–1096. | PubMed | ISI | ChemPort |
14.	Drake SL, Klein DJ, Mickelson DJ, Oegema TR, Furcht LT & McCarthy JB. Cell surface phosphatidylinositol-anchored heparan sulfate proteoglycan initiates mouse melanoma cell adhesion to a fibronectin-derived, heparin-binding synthetic peptide. J Cell Biol (1992) 117: 1331–1341. | Article | PubMed | ISI | ChemPort |
15.	Engvall E & Ruoslahti E. Binding of soluble form of fibroblast surface protein, fibronectin, to collagen. Int J Cancer (1977) 20: 1–5. | PubMed | ISI | ChemPort |
16.	Faassen AE, Mooradian DL, Tranquillo RT, Dickinson RB, Letourneau PC, Oegema TR & McCarthy JB. Cell surface CD44-related chondroitin sulfate proteoglycan is required for transforming growth factor--stimulated mouse melanoma cell motility and invasive behavior on type I collagen. J Cell Sci (1993) 105: 501–511. | PubMed | ISI | ChemPort |
17.	Folkvord JM, Viders D, Coleman-Smith A & Clark RAF. Optimization of immunohistochemical techniques to detect extracellular matrix proteins in fixed skin specimens. J Histochem Cytochem (1989) 37: 105–113. | PubMed | ISI | ChemPort |
18.	Gailit J, Pierschbacher M & Clark RAF. Expression of functional 41 by human dermal fibroblasts. J Invest Dermatol (1993) 100: 323–328. | Article | PubMed | ISI | ChemPort |
19.	Gailit J, Xu J, Bueller H & Clark RAF. Platelet-derived growth factor and inflammatory cytokines have differential effects on the expression of integrins 11 and 51 by human dermal fibroblasts in vitro. J Cell Physiol (1996) 169: 281–289. | Article | PubMed | ISI | ChemPort |
20.	Greiling D & Clark RAF. Fibronectin provides a conduit for fibroblast transmigration from collagenous stroma into fibrin clot provisional matrix. J Cell Sci (1997) 110: 861–870. | PubMed | ISI | ChemPort |
21.	Grinnell F, Ho CH & Wysocki AJ. Degradation of fibronectin and vitronectin in chronic wound fluid: Analysis by cell blotting, immunoblotting, and cell adhesion assays. J Invest Dermatol (1992) 98: 410–416. | Article | PubMed | ISI | ChemPort |
22.	Grinnell F & Zhu M. Identification of neutrophil elastase as the proteinase in burn wound fluid responsible for degradation of fibronectin. J Invest Dermatol (1994) 103: 155–161. | Article | PubMed | ISI | ChemPort |
23.	Guan JL & Hynes RO. Lymphoid cells recognize an alternatively spliced segment of fibronectin via the integrin receptor 41. Cell (1990) 60: 53–61. | Article | PubMed | ISI | ChemPort |
24.	Hautanen A, Gailit J, Mann DM & Ruoslahti E. Effects of modifications of the RGD sequence and its context on recognition by the fibronectin receptor. J Biol Chem (1989) 264: 1437–1442. | PubMed | ISI | ChemPort |
25.	Herrick SE, Sloan P, McGurk M, Freak L, McCollum CN & Ferguson MWJ. Sequential changes in histologic pattern and extracellular matrix deposition during the healing of chronic venous ulcers. Am J Pathol (1992) 141: 1085–1095. | PubMed | ISI | ChemPort |
26.	Huebsch JC, McCarthy JB, Diglio CA & Mooradian DL. Endothelial cell interactions with synthetic peptides from the carboxyl-terminal heparin-binding domains of fibronectin. Circ Res (1995) 77: 43–53. | PubMed | ISI | ChemPort |
27.	Huhtala P, Humphries MJ, McCarthy JB, Tremble PM, Werb Z & Damsky CH. Cooperative signaling by 51 and 41 integrins regulates metalloproteinase gene expression in fibroblasts adhering to fibronectin. J Cell Biol (1995) 129: 867–879. | Article | PubMed | ISI | ChemPort |
28.	Humphries MJ, Akiyama S, Komoriya A, Olden K & Yamada KM. Identification of an alternatively spliced site in human plasma fibronectin that mediates cell type-specific interaction. J Cell Biol (1986) 103: 2637–2647. | Article | PubMed | ISI | ChemPort |
29.	Humphries MJ, Akiyama SK, Komoriya A, Olden K & Yamada KM. Identification of two distinct regions of the type III connecting segment of human plasma fibronectin that promote cell type-specific adhesion. J Biol Chem (1987) 262: 6886–6892. | PubMed | ISI | ChemPort |
30.	Hynes RO. Integrins: Versatility, modulation, and signaling in cell adhesion. Cell (1992) 69: 11–25. | Article | PubMed | ISI | ChemPort |
31.	Iida J, Skubitz AP, Furcht LT, Wayner EA & McCarthy JB. Coordinate role for cell surface chondroitin sulfate proteoglycan and 41 integrin in mediating melanoma cell adhesion to fibronectin. J Cell Biol (1992) 118: 431–444. | Article | PubMed | ISI | ChemPort |
32.	Ingham KC, Brew SA, Migliorini MM & Busby TF. Binding of heparin by type III domains and peptides from the carboxy terminal hep-2 region of fibronectin. Biochemistry (1993) 32: 12548–12553. | Article | PubMed | ISI | ChemPort |
33.	Komoriya A, Green LJ, Mervic M, Yamada SS, Yamada KM & Humphries MJ. The minimal essential sequence for a major cell type-specific adhesion site (CS 1) within the alternatively spliced type III connecting segment domain of fibronectin is leucine-aspartic acid-valine. J Biol Chem (1991) 266: 15075–15079. | PubMed | ISI | ChemPort |
34.	Maheshwari G, Brown G, Lauffenburger DA, Wells A & Griffith LG. Cell adhesion and motility depend on nanoscale RGD clusterin. J Cell Sci (2000g) 113: 1677–1686. | PubMed | ISI | ChemPort |
35.	Massia SP & Hubbell JA. Vascular endothelial cell adhesion and spreading promoted by the peptide REDV of the IIICS region of plasma fibronectin is mediated by the integrin 41. J Biol Chem (1992) 267: 14019–14026. | PubMed | ISI | ChemPort |
36.	McCarthy JB, Chelberg MK, Mickelson DJ & Furcht LT. Localization and chemical synthesis of fibronectin peptides with melanoma adhesion and heparin binding activities. Biochemistry (1988) 27: 1380–1388. | Article | PubMed | ISI | ChemPort |
37.	McCarthy JB, Chelberg MK, Mickelson DJ & Furcht LT. RGD-independent cell adhesion to the carboxy-terminal heparin binding fragment of fibronectin involves heparin-dependent and -independent activities. J Cell Biol (1990) 110: 777–787. | Article | PubMed | ISI | ChemPort |
38.	McClain SA, Simon M & Jones E et al. Mesenchymal cell activation is the rate-limiting step of granulation tissue induction. Am J Pathol (1996) 149: 1257–1270. | PubMed | ISI | ChemPort |
39.	Mooradian DL, McCarthy JB, Cameron JD, Skubitz APN & Furcht LT. Rabbit corneal epithelial cells adhere to two distinct heparin-binding synthetic peptides derived from fibronectin. Invest Ophthalmol Vis Sci (1992) 33: 3034–3040. | PubMed | ISI | ChemPort |
40.	Mooradian DL, McCarthy JB, Skubitz APN, Cameron JD & Furcht LT. Characterization of FN-C/H-V, a novel synthetic peptide from fibronectin that promotes rabbit corneal epithelial cell adhesion, spreading, and motility. Invest Ophthalmol Vis Sci (1993) 34: 153–164. | PubMed | ISI | ChemPort |
41.	Mould AP, Askari JA, Craig SU, Garratt AN, Clements J & Humphries MJ. Integrin 41-mediated melanoma cell adhesion and migration on vascular cell adhesion molecule-1 (VCAM-1) and the alternatively spliced IIICS region of fibronectin. J Biol Chem (1994) 269: 27224–27230. | PubMed | ISI | ChemPort |
42.	Mould AP & Humphries MJ. Identification of a novel recognition for the integrin 41 in the COOH-terminal heparin-binding domain of fibronectin. EMBO J (1991) 10: 4089–4095. | PubMed | ISI | ChemPort |
43.	Mould AP, Komoriya A, Yamada KM & Humphries MJ. The CS5 peptide is a second site in the IIICS region of fibronectin recognized by the integrin 41. J Biol Chem (1991) 266: 3579–3585. | PubMed | ISI | ChemPort |
44.	Nagai T, Yamakawa N, Aota S-I, Yamada SS, Akiyama SK, Olden K & Yamada KM. Monoclonal antibody characterization of two distant sites required for function of the central cell-binding domain of fibronectin in cell adhesion, cell migration, and matrix assembly. J Cell Biol (1991) 114: 1295–1305. | Article | PubMed | ISI | ChemPort |
45.	Obara M, Knag MS & Yamada KM. Site-directed mutagenesis of the cell-binding domain of fibronectin: Separable, synergistic sites mediate adhesive function. Cell (1988) 53: 649–657. | Article | PubMed | ISI | ChemPort |
46.	Pierschbacher MD & Ruoslahti E. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature (1984) 309: 30–33. | Article | PubMed | ISI | ChemPort |
47.	Pytela R, Pierschbacher MD & Ruoslahti E. Identification and isolation of a 140 kD cell surface glycoprotein with properties expected of a fibronectin receptor. Cell (1985) 40: 191–198. | Article | PubMed | ISI | ChemPort |
48.	Ramos JW & DeSimone DW. Xenopus embryonic cell adhesion to fibronectin: Position-specific activation of RGD/synergy site-dependent migratory behavior at gastrulation. J Cell Biol (1996) 134: 227–240. | Article | PubMed | ISI | ChemPort |
49.	Ruoslahti E. Fibronectin and its receptors. Annu Rev Biochem (1988) 57: 375–423. | Article | PubMed | ISI | ChemPort |
50.	Ruoslahti E & Pierschacher MD. New perspectives in cell adhesion: RGD and integrins. Science (1987) 238: 291–297.
51.	Sharma A, Askari JA, Humphries MJ, Jones EY & Stuart DI. Crystal structure of a heparin- and integrin-binding segment of human fibronectin. EMBO J (1999) 18: 1468–1479. | Article | PubMed | ISI | ChemPort |
52.	Skorstengaard K, Thogersen HC, Vibe-Pedersen K, Petersen TE & Magnusson S. Purification of twelve cyanogen bromide fragments from bovine plasma fibronectin and the amino acid sequence of eight of them: Overlap evidence aligning two plasmic fragmenys, internal homology in gelatin-binding region and phosphorylation site near C terminus. Eur J Biochem (1982) 128: 605–623. | PubMed | ISI | ChemPort |
53.	Tuszynski GP & Murphy A. Spectrophotometric quantitation of anchorage-dependent cell numbers using the bicinchoninic acid protein assay reagent. Anal Biochem (1990) 184: 189–191. | Article | PubMed | ISI | ChemPort |
54.	Varani J, Orr W & Ward PA. Comparison of the migration patterns of normal and malignant cells in two assay systems. Am J Pathol (1978) 90: 159–172. | PubMed | ISI | ChemPort |
55.	Wayner EA & Carter WG. Identification of multiple cell adhesion receptors for type VI collagen and fibronectin in human fibrosarcoma cells possessing unique and common subunits. J Cell Biol (1987) 105: 1873–1884. | Article | PubMed | ISI | ChemPort |
56.	Wayner EA, Carter WG, Piotrowicz R & Kunicki TJ. The function of multiple extracellular matrix receptors (ECMRs) in mediating cell adhesion to ECM. Preparation of monoclonal antibodies to the fibronectin receptor that specifically inhibit cell adhesion to fibronectin and react with platelet glycoproteins Ic/IIa. J Cell Biol (1988) 107: 1881–1891. | Article | PubMed | ISI | ChemPort |
57.	Wayner EA, Garcia-Pardo A, Humphries MJ, McDonald JA & Carter WG. Identification and characterization of the T lymphocyte adhesion receptor for an alternative cell attachment domain (CS-1) in plasma fibronectin. J Cell Biol (1989) 109: 1321–1330. | Article | PubMed | ISI | ChemPort |
58.	Wayner EA & Kovach NL. Activation-dependent recognition by hematopoietic cells of the LDV sequence in the V region of fibronectin. J Cell Biol (1992) 116: 489–497. | Article | PubMed | ISI | ChemPort |
59.	Welch MP, Odland GF & Clark RAF. Temporal relationships of F-actin bundle formation, collagen and fibronectin matrix assembly, and fibronectin receptor expression to wound contraction. J Cell Biol (1990) 110: 133–145. | Article | PubMed | ISI | ChemPort |
60.	Wilke MS, Skubitz APN, Furcht LT & McCarthy JB. Human keratinocytes adhere to two distinct heparin binding synthetic peptides derived from fibronectin. J Invest Dermatol (1991) 97: 573–579. | Article | PubMed | ISI | ChemPort |
61.	Wilke MS, Vespa J, Skubitz APN, Furcht LT & McCarthy JB. Human keratinocytes adhere to and spread on synthetic peptide FN-C/H-V derived from fibronectin. J Invest Dermatol (1993) 101: 43–48. | Article | PubMed | ISI | ChemPort |
62.	Woods A, Couchman JR & Hook M. Adhesion and cytoskeletal organization of fibroblasts in response to fibronectin fragments. EMBO J (1986) 5: 665–670. | PubMed | ISI | ChemPort |
63.	Woods A, McCarthy JB, Furcht LT & Couchman JR. A synthetic peptide from the COOH-terminal heparin binding domain of fibronectin promotes focal adhesion formation. Mol Biol Cell (1993) 4: 605–613. | PubMed | ISI | ChemPort |
64.	Wysocki AB, Staiano-Coico L & Grinnell F. Wound fluid from chronic leg ulcers contains elevated levels of metalloproteinases MMP-2 and MMP-9. J Invest Dermatol (1993) 101: 64–68. | Article | PubMed | ISI | ChemPort |
65.	Xu J & Clark RAF. Extracellular matrix alters PDGF regulation of fibroblast integrins. J Cell Biol (1996) 132: 239–249. | Article | PubMed | ISI | ChemPort |
66.	Yamada KN. Fibronectin and other cell interactive glycoproteins. In: Hay ED (edCell Biology of Extracellular Matrix (1991) New York: Plenum Press 111–145.
67.	Yamada KM & Clark RAF. Provisional matrix. In: Clark R (edMolecular and Cellular Biology of Wound Repair (1996) New York: Plenum Press p 51–93.
68.	Yamada KM & Kennedy DW. Dualistic nature of adhesive protein function: Fibronectin and its biologically active peptide fragments can autoinhibit fibronectin function. J Cell Biol (1984) 99: 29–36. | Article | PubMed | ISI | ChemPort |

Acknowledgments

This work was supported by grants from the NIA (AG 101143) and the NIAMS (AR 42987) to R.A.F.C. and from the NCI (CA 44627) to J.E.S. D.G. was supported during this work by a Dermatology Foundation postdoctoral fellowship. We also acknowledge the technical assistance of Jennifer Luczak and Anne Marie Cumiskey.