Rabbit corneal epithelial cells adhere to two distinct heparin-binding ...

Investigative Ophthalmology & Visual Science, Vol. 33, No. 11, October 1992Copyright © Association for Research in Vision and Ophthalmology

Rabbit Corneal Epithelial Cells Adhere to Two DistinctHeparin-Binding Synthetic Peptides

Derived From FibronectinDaniel L. Mooradian,* Jomes B. McCarthy,* J. Douglas Cameron,j- Amy P. N. Skubitz,* and Leo T. Furchr*

Fibronectin plays an important role in corneal reepithelialization during corneal wound healing. In thisstudy, rabbit corneal epithelial (RCE) cell adhesion to fibronectin was further defined using proteolyticfragments of fibronectin and chemically synthesized peptides derived from the amino acid sequence offibronectin. RCE cells adhere to intact fibronectin, the 75 kD fragment containing the RGDS (Arg-Gly-Asp-Ser) cell adhesion-promoting sequence, and the 33/66 kD cell adhesion promoting/heparin-bind-ing fragments of fibronectin. The 75 kD fragment and the 33/66 kD fragments partially inhibited RCEcell adhesion to intact fibronectin, suggesting that these fragments represent distinct sites used by RCEcells to adhere to intact fibronectin. Two chemically synthesized peptides derived from the amino acidsequence of the 33/66 kD fragments of fibronectin, FN-C/H-I (YEKPGSPPREWPRPRPGV) andFN-C/H-III (YRVRVTPKEKTGPMKE), directly promoted the adhesion of RCE cells. As furtherevidence that FN-C/H-I and FN-C/H-III play a role in the adhesion of RCE cells to the 33/66 kDfragments of fibronectin, we have shown that soluble FN-C/H-I and FN-C/H-III inhibited RCE celladhesion on surfaces coated with the 33/66 kD fragments. In addition, polyclonal IgG against FN-C/H-I and FN-C/H-III partially blocked RCE cell adhesion to the 33/66 kD fragments, confirming thatthese sequences represent adhesion-promoting sites within these fragments. In contrast, two previouslydescribed peptides from the 33/66 kD fragments of fibronectin, which promoted the adhesion of avariety of cell types, FN-C/H-II (KNNQKSEPLIGRKKT) and CS-1 (DELPQLVTLPHPNLHG-PEILDVPST), did not support RCE cell adhesion. These data suggest that RCE cell adhesion tofibronectin has a complex molecular basis and is mediated by multiple domains, at least two of whichare located within the car boxy 1-terminal heparin-binding/cell adhesion-promoting domains of fibronec-tin. Invest Ophthalmol Vis Sci 33:3034-3040,1992

The corneal epithelium provides a protective cover-ing for the cornea and is essential for normal cornealfunction. Failure of the cornea to reepithelialize canresult in persistent corneal defects and ulceration ofthe underlying corneal stroma.' While a variety of fac-tors can influence corneal wound healing (growth fac-tors, etc.) there is convincing evidence that fibronec-tin, a glycoprotein found in plasma and extracellularmatrices,2 plays a central role in corneal reepithelial-ization. Fibronectin accumulates with fibrin at sites ofcorneal injury,3 providing a provisional matrix for epi-thelial cell adhesion and migration during wound heal-

From the *Department of Laboratory Medicine and Pathology/Biomedical Engineering Center, and the tDepartment of Ophthal-mology, University of Minnesota, Minneapolis, Minnesota.

This work was supported by National Institutes of Health grantsEY06625 (LTF), AM32660 (LTF), and CA43924 (JBM), and by anAmerican Heart Association (Minnesota Affiliate) PostdoctoralFellowship (DLM). LTF is an Allen-Pardee Professor of Pathology.

Reprint requests: Daniel L. Mooradian, University of Minne-sota, Laboratory Medicine and Pathology, Box 609 UMHC, Minne-apolis, MN 55455.

ing. In addition, fibronectin promotes the adhesion ofcorneal epithelial cells in vitro4 and promotes the che-motactic and haptotactic migration of corneal epithe-lial cells in vitro.4 Finally, the addition of exogenousfibronectin has been shown to promote corneal ree-pithelialization in vitro5 and in vivo6 and has beenused clinically to treat persistent corneal epithelial de-fects.7'8

Plasma fibronectin is a large dimeric glycoproteincomposed of similar, but not identical polypeptidesubunits, termed the A and B chains, that are linkedby disulfide bonds.9 A number of functionally distinctdomains of fibronectin have been identified (Fig. 1).Cell adhesion to fibronectin is complex and involvesthe RGD-dependent cell binding domain located inthe central third of the A and B chains of the fibronec-tin molecule10 as well as the cell-binding/heparin-binding domains located in the 33/66 kD fragmentsfrom the carboxyl-terminal third of the A chain and Bchain of the fibronectin molecule, respectively.11 Pre-vious studies have identified four distinct peptide se-quences within the carboxyl-terminal heparin-bind-

0034

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933389/ on 02/20/2018

No. 11 CORNEAL EPITHELIAL CELL ADHESION TO FIDRONECTIN / Mooradian er al 3035

(B chain) H2N

(A chain) H2N

Fig. 1. Location of synthetic peptides within fibronectin. The33/66 kD fragments of fibronectin are generated by trypsin (T) andcathepsin D (C) digestion of fibronectin. I, weak heparin binding;II, collagen binding (noncovalent); III, free sulfhydryl; IV, RGD-mediated cell adhesion; V, carboxyKterminal strong heparin bind-ing and cell adhesion; VI, free sulfhydryl. The locations of peptidesFN-C/H-I, FN-C/H-II, FN-C/H-III, and CS-1 within the 33 kDfragment also are shown.

ing/cell adhesion-promoting domain that promotethe adhesion of other cell types. These are: FN-C/H-I,12 FN-C/H-II,1314 FN-C/H-III15 and CS-1.16 In thepresent study, these peptides were used to further de-fine rabbit corneal epithelial (RCE) cell adhesion tothe carboxyl-terminal heparin-binding/cell adhesion-promoting domains of fibronectin.

Materials and Methods

Corneal Epithelial Cell Culture

The experiments described in this study conform tothe ARVO resolution on the use of animals in re-search and to the University of Minnesota guidelinesfor animal use. Corneal epithelial cell cultures wereinitiated as previously described.4 Briefly, 3-5 lb NewZealand white rabbits (Birchwood Farms, Red Wing,MN) were killed, and their eyes were enucleated,rinsed with Hank's balanced salt solution (HBSS),and transferred to HBSS containing gentamycin (0.5mg/ml) and HEPES (15 mmol/1). Corneas were ex-cised and cut into 1 -2 mm2 blocks using a sterile razorblade. These blocks were incubated in dispase (1 U/ml, Grade II; Boehringer Mannheim, Indianapolis,IN) for 1 hr at 25°C. The corneal epithelium was sepa-rated from the underlying stroma with forceps andtransferred to 60 mm tissue culture dishes (Costar,Cambridge, MA) containing 0.5 ml of Dulbecco'smodified Eagle's medium (DMEM)/F-12, 10% fetalbovine serum (FBS), 5 Mg/ml insulin, 10 ng/ml epi-dermal growth factor, 0.5 /ig/ml hydrocortisone, and50 Mg/ml gentamycin. Epithelial explants were al-lowed to attach for 24 hr at 37°C before additionalmedium was added. Medium was changed every third

day. Epithelial cells migrate out of the explants andreach confluence within 10-14 days. Only primarycultures were used in these studies.

Isolation of Plasma Fibronectin and Its ProteolyticFragments

Human plasma fibronectin was purified as previ-ously described.17 The 75 kD RGDS-containing celladhesion-promoting fragment, 33/66 kD heparin-binding/cell adhesion-promoting fragments, and the27 kD amino-terminal fragment (Fig. 1) were gener-ated by proteolytic digestion of intact fibronectin andpurified as previously described.''

Peptide Synthesis and Characterization

Synthetic peptides were synthesized at the Micro-chemical Facility (University of Minnesota) using anApplied Biosystems (Foster City, CA) peptide synthe-sizer with previously described modifications of theMerrifield solid-phase method.18 Lyophilized crudepeptides were purified by preparative reverse-phasehigh performance liquid chfomatography (HPLC) ona C-18 column and were eluted with a linear gradientof acetonitrile (0-60%) containing 0.1% trifluoroace-tic acid in water. Purity and composition were veri-fied by analytic HPLC and amino acid analysis priorto use. The sequences and selected properties of thesynthetic peptides used in this study are shown in Ta^ble 1 and Figure 1. Hydropathy indices were calcu-lated using the method of Kyte and Doolittle.19

Peptide Coupling to Ovalbumin

Synthetic peptides were conjugated to ovalbumin(OVA) using l-ethyl-3-(3-dimethylaminopropyl)-car-bodiimide hydrochloride (EDC) (Sigma ChemicalCo., St. Louis, MO) as previously described.13 Theconjugation of adhesion-promoting peptides to car-rier proteins such as albumin has been shown to en-hance their adhesion-promoting activity.20 Briefly, 5mg of each peptide was dissolved in 1 ml of water onice, and 5 mg of OVA was added. Fifty milligrams ofEDC in 150 MI of water was added to the peptide-OVAmixture, and the solution was mixed overnight at 4°Con a rotary shaker. The OVA-conjugated peptidesthen were dialyzed extensively against phosphate buff-ered saline (PBS) to remove excess EDC and any un-coupled peptide (Spectrapore 6, 30 kD cutoff; Spec-trum Medical Industries, Los Angeles, CA). The cou-pling efficiency in each case was 34% and wasdetermined using 125I-labeled peptides of known spe-cific activity and by measuring the radioactivity inpurified peptide-OVA conjugates. Peptide-OVA con-jugates were stored at -80°C.


3036 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / October 1992 Vol. 33

Table 1. Synthetic peptides from the 33 kD fragment of fibronectin

Peptidenomenclature Primary sequence*

Hydropathyindexf

Netcharge%

FN-C/H-I YEKLPGSPPREVVPRPRPGV (residues 1906-1924)FN-C/H-II KNNQKSEPLIGRKJCT (residues 1946-1960)FN-C/H-III YRVRVTPKEKTGPMKE (residues 1721 -1736)CS-1 DELPQLVTLPHPNLHGPEILDVPST (residues 1961-1985, A chain only)

-24.3-29.3-23.7

-9.9

+2+4+3-4

* Sequences shown use the single letter amino acid code (K, lysine; R,arginine; H, histidine; E, gliitamicacid; D, aspartic acid; Q, glutamine; N,asparagine; P, praline; G, glycine; S, serine, T, threonine; V, valine; I, isoleu-cine; L, leucine; Y, tyrosine).

t Hydropathy indices were calculated by the method of Kyte and Doolit-tle.19 Using this method, hydrophilic regions of a protein have a negative

hydropathy index.% The sum of all charged residues where lysine (K) and arginine (R) resi-

dues are positively charged (+1) and glutamic acid (E) and aspartic acid (D)are negatively charged (-1) at neutral pH. Histidine (H) is assumed to beuncharged at neutral pH.

Antibody Production, Purification, andCharacterization

Polyclonal antibodies were generated against FN-C/H-I, FN-C/H-II, and FN-C/H-III coupled to KLH(Keyhole limpet hemocyanin; Sigma Chemical Co.)using carbodiimide as previously described.21 NewZealand white rabbits were immunized on the backby multiple intradermal injections of approximately 2mg per rabbit of peptide/KLH conjugate in completeFreund's adjuvant. Subsequent biweekly boosts weregiven intramuscularly in incomplete Freund's adju-vant. Sera were collected 14 days after the fourth im-munization and tested by ELISA for reactivity againstuncoupled peptide, fibronectin, and other ligands aspreviously described.21 Immiinoglobulin G (IgG) waspurified from pooled immune sera by precipitationwith ammonium sulfate, followed by DEAE anionexchange chromatography. The purity of the IgG wasverified by SDS-PAGE.

Cell Adhesion Assay

Primary RCE cells were labeled with 35S-methio-nine (Tran 35S-label; ICN Radiochemicals, CostaMesa, CA) for 24 hr prior to use in adhesion assays.Purified proteins or OVA-conjugated peptides werediluted in Voller's buffer (see figure legends for spe-cific concentrations), and 50jul aliquots were added tothe wells of a 96-well Immulon I microtiter plate. Pro-teitts/peptides were al'lowed to adsorb overnight at37GC, after which nonspecific binding sites wereblocked using PBS containing 2 mg/ml OVA. Sub-coiiiiiient RCE cells (10-14 days in culture) were har-vested using trypsin:EDTA and resuspended inME3VI/FU2 containing 10% FBS. The cells werecentefoged and. resuspended two times in serum-freeDMEM/F-12 containing 2 mg/ml OVA, counted,and resusptnded in serum-free DMEM/F-12 contain-ing 2 mg/ml OVA at a final cell concentration of 5X 104 cels/ml. One hundred-microliter aliquots ofthis cell suspension were added to protein/peptide-

coated wells and incubated (see figure legends for spe-cific incubation times) at 37°C. Nonadherent cellswere removed by washing three times with PBS con-taining OVA. Adherent cells were solubilized in 100Ail of 0.5 N NaOH/1% sodium dodecyl sulfate, trans-ferred to scintillation vials, and counted in a Beckman(Irvine, CA) 3801 scintillation counter.

Inhibition of RCE Cell Adhesion by Soluble Peptides

RCE cells were preincubated for 60 min at 37°C in200 Mg/ml of soluble OVA-coupled FN-C/H-I, FN-C/H-II, or FN-C/H-III and added to Immulon I plates(Dynatech Labs, Inc., Chantilly, VA) coated with 20Mg/ml of the 33/66 kD fragments of fibronectin. Ad-hesion assays were terminated after 60 min and celladhesion was quantitated as described above.

Inhibition of RCE Cell Adhesion by Polyclonal IgGAgainst FN-C/H-I, FN-C/H-II, and FN-C/H-III

Immulon I plates coated with 20 /ug/ml of the 33/66kD fragments of fibronectin were preincubated for 60min with 500 Mg/ml of purified IgG from normal rab-bits or rabbits immunized with FN-C/H-I, FN-C/H-II, or FN-C/H-III. RCE cells then were added tothe plates, and cell adhesion was quantitated after 60min as described in Materials and Methods.

Results

RCE Cell Adhesion to Fibronectin and FibronectinFragments

Primary RCE cells adhere to fibronectin in a con-centration-dependent manner (Fig. 2). Maximal celladhesion was observed on surfaces coated with 50 Mg/nil of intact fibronectin, although surfaces coated with6 Mg/ml effectively promoted half-maxima! cell adhe-sion. RCE cell adhesion to the 75 kD fragment offibronectin and to the 33/66 kD fragments of fibronec-tin was comparable to cell adhesion to intact fibronec-tin at high coating concentrations (ie, 50 Mg/m0 How-


No. 11 CORNEAL EPITHELIAL CELL ADHESION TO FIBRONECTIN / Moorodion er ol 3037

100 i

80-

60-

40-

20-

0 .1 1 10 100

Coating Concentration (p.g/ml)

Fig. 2. Rabbit corneal epithelial cell adhesion to fibronectin andfragments of fibronectin. Immulon I plates were coated overnightwith intact fibronectin (filled squares); the 75 kD fragment (opensquares); the 33/66 kD fragments (filled circles); or the 27 kD frag-ment of fibronectin (open circles). Cell adhesion was measuredafter I hr, as described in Materials and Methods. Data representthe mean ± one standard deviation with n = 3.

ever, a higher coating concentration of the 33/66 kDfragment was required to promote half-maximal celladhesion (20 Mg/ml). In contrast, RCE cells did notadhere to surfaces coated with the 27 kD fragment offibronectin at any concentration tested (0.1-100 ng/ml), nor did they adhere to uncoated surfaces blockedwith OVA (2 mg/ml).

In competition experiments, the 75 kD fragmentand the 33/66 kD fragments of fibronectin each par-tially inhibited cell adhesion to surfaces coated withintact fibronectin (Fig. 3). RCE cell adhesion to intactfibronectin was inhibited about 50% in the presenceof 500 Mg/ml of the 33/66 kD fragments of fibronectinor the 75 kD fragment of fibronectin. In contrast, celladhesion to intact fibronectin was inhibited approxi-mately 85% in the presence of soluble intact fibronec-tin (500 fig/mY). These findings suggest that domainswithin the 33/66 kD fragments and the 75 kD frag-ment of fibronectin contribute to RCE cell adhesionto intact fibronectin.

RCE Cell Adhesion to Chemically SynthesizedPeptides From the 33/66 kD Fragments ofFibronectin

To further define the nature of the cell adhesion-promoting site or sites within the 33/66 kD fragmentsof fibronectin, we tested the ability of several chemi-cally synthesized peptides derived from the aminoacid sequence of this fragment to promote RCE celladhesion. Peptides FN-C/H-I, FN-C/H-II, and FN-C/

H-IH (Table 1) previously have been shown to pro-mote the adhesion of various cells in culture.12"14

These peptides are cationic (net charge = +2 to +4),hydrophilic (hydropathy index = -23.7 to -29.3),and bind [3H]-heparin. In contrast, peptide CS-1,15

although hydrophilic (hydropathy index = -9.9), isanionic (net charge = -4) and does not bind [3H]-heparin.12

RCE cells adhered to surfaces coated with OVA-coupled FN-C/H-I and FN-C/H-III in a concentra-tion-dependent manner (Fig. 4). Maximal cell adhe-sion was observed on surfaces coated with 10 ju,g/ml ofOVA-coupled FN-C/H-I or FN-C/H-III. Half-maxi-mal cell adhesion was observed on surfaces coatedwith 2 Mg/ml of FN-C/H-I or 5 Mg/ml FN-C/H-III. Incontrast to many other cell types, RCE cell adhesionwas not observed on surfaces coated with OVA-cou-pled FN-C/H-II or OVA-coupled CS-1 or on surfacescoated with OVA-coupled OVA, regardless of thecoating concentration (Fig. 4). This was not due topoor adsorption of FN-C/H-II to the plastic surface orto inactivation of FN-C/H-II by coupling to OVA,because in similar assays, OVA-coupled FN-C/H-IIpromoted the adhesion of other cell types.1314'22

Inhibition of RCE Cell Adhesion to 33/66 kDFragments of Fibronectin by FN-C/H-Iand FN-C/H-III

To establish the relative contributions of FN-C/H-Iand FN-C/H-III to the cell adhesion-promoting activ-ity of the 33/66 kD fragments of fibronectin, we next

100 i

80-

60-

40-

20-

27 kD 75 kD 33/66 kD Fibronectin

Competitor

Fig. 3. Inhibition of RCE cell adhesion to fibronectin by frag-ments of fibronectin. Immulon I plates were coated with fibronec-tin at a concentration that is sufficient to promote half-maximal celladhesion (8 Mg/ml)- RCE cells were preincubated for 60 min with500 Mg/ml of fibronectin, the 75 kD fragment, the 33/66 kD frag-ments, or the 27 kD fragment, then added to fibronectin-coatedplates. Adhesion was measured after 60 min. Data represent themean ± one standard deviation with n = 3.



CD. C

<

CD

o

100

Coating Concentration (|ig/ml)

Fig. 4. Rabbit corneal epithelial cell adhesion to chemically syn-thesized peptides derived from the 33/66 kD fragments offibronec-tin. Immulon I plates were coated overnight with increasing con-centrations of OVA-coupled FN-C/H-I (filled squares); FN-C/H-II(open squares); FN-C/H-III (filled triangles); CS-1 (open triangles);and OVA (filled circles). RCE cell adhesion was measured after 1hr, as described in Materials and Methods. Data represent the mean± one standard deviation with n = 3.

tested the ability of OVA-coupled FN-C/H-I and FN-C/H-III to inhibit RCE cell adhesion to surfacescoated with the 33/66 kD fragments. Soluble OVA-coupled FN-C/H-I or FN-C/H-III at 200 /tg/ml inhib-ited RCE cell adhesion to surfaces coated with the

FN-C/H-II FN-C/H-I FN-C/H-I

Competitor

Fig. 5. Inhibition of RCE cell adhesion to the 33/66 kD fragmentsof fibronectin by synthetic peptides. Immulon I plates were coatedwith the 33/66 kD fragments of fibronectin at a concentration thatis sufficient to promote half-maximal cell adhesion (20 fig/m\).RCE cells were preincubated for 60 min with 200 Mg/ml of OVA-coupled FN-C/H-I, FN-C/H-II, and FN-C/H-III, then added to theplates. Adhesion was measured after 60 min. Data represent themean ± one standard deviation with n = 3.

33/66 kD fragments of fibronectin by 30% (Fig. 5).Little or no inhibition to the 33/66 kD fragments offibronectin was observed in the presence of solubleOVA-coupled FN-C/H-II. These peptides did not in-hibit RCE cell adhesion to intact fibronectin (data notshown).

Inhibition of RCE Cell Adhesion to the 33/66 kDFragments of Fibronectin by Polyclonal IgGAgainst FN-C/H-I or FN-C/H-III

As further evidence for the role of FN-C/H-I andFN-C/H-III in the adhesion-promoting activity of the33/66 kD fragments of fibronectin, we measured theability of polyclonal antibodies against FN-C/H-I orFN-C/H-III to inhibit RCE cell adhesion to the 33/66kD fragments of fibronectin. RCE cell adhesion tosurfaces coated with the 33/66 kD fragments of fibro-nectin was inhibited almost 40% by 500 ng/m\ of poly-clonal rabbit IgG specific for FN-C/H-I or FN-C/H-III (Fig. 6). Cell adhesion to the 33/66 kD fragmentsof fibronectin was not inhibited in the presence ofpolyclonal rabbit IgG specific for FN-C/H-II (Fig. 6)or by normal rabbit IgG (data not shown). None ofthese antibodies inhibited RCE cell adhesion to fibro-nectin (data not shown).

FN-C/H-II FN-C/H-I

Polyclonal IgG

FN-C/H-I

Fig. 6. Inhibition of RCE cell adhesion to the 33/66 kD fragmentsof fibronectin by antibodies against synthetic peptides. Immulon Iplates were coated with the 33/66 kD fragments of fibronectin at aconcentration that is sufficient to promote half-maximal cell adhe-sion (20 Mg/mO- Immulon I plates were preincubated for 60 minwith 500 Mg/ml of normal rabbit IgG (data not shown) or withpurified polyclonal rabbit IgG specific for FN-C/H-I, FN-C/H-II orFN-C/H-III. RCE cells then were added to the plates and adhesionwas measured after 60 min. Data represent the mean ± one stan-dard deviation with n = 3.


No. 11 CORNEAL EPITHELIAL CELL ADHESION TO FIDRONECTIN / Moorodion er ol 3039

Discussion

Fibronectin plays an important role in cornealwound healing. Previous work has shown that theRGDS-containing cell binding domain of fibronectinplays an important role in RCE cell adhesion to fibro-nectin.423 However, corneal epithelial cell adhesionto fibronectin has a complex molecular basis, and inthe present report, we demonstrate that RCE cells alsoadhere to the 33/66 kD carboxyl terminal heparin-binding/cell adhesion-promoting fragments of fibro-nectin that do not contain the RGDS tetrapeptide se-quence. This activity is important in the context ofRCE cell adhesion to fibronectin, as illustrated by theobservation that the 33/66 kD fragments of fibronec-tin inhibited adhesion on surfaces coated with intactfibronectin. These findings are consistent with reportsthat show the 33/66 kD fragments are important celladhesion-promoting domains for various other celltypes,2425 and they suggest that the 33/66 kD frag-ments play an important role in RCE cell adhesion tofibronectin.

FN-C/H-I and FN-C/H-III are chemically synthe-sized peptides from the carboxyl terminal heparin-binding/cell adhesion-promoting domains of fibro-nectin.1314 These hydrophilic and cationic peptidespromoted the adhesion of RCE cells in a concentra-tion-dependent manner. In contrast, FN-C/H-II, alsoa hydrophilic and cationic peptide that has beenshown to promote the adhesion of a variety of celltypes,131422 did not promote RCE cell adhesion, sug-gesting that the activity of FN-C/H-I and FN-C/H-IIIis not due merely to their hydrophilic or cationic na-ture, but rather depends upon the primary structureof these peptides. Furthermore, CS-1, a peptide thatpromotes the adhesion of many cell types,1316 did notpromote the adhesion of RCE cells. These findingsillustrate the cell-specific nature of adhesion to pep-tides from this region of fibronectin. The specific pat-tern of adhesion reflecting an as yet poorly under-stood pattern of receptor expression.

Peptides FN-C/H-I and FN-C/H-III contribute tothe adhesion-promoting activity of the 33/66 kD frag-ments of fibronectin. This is illustrated by studies inwhich soluble FN-C/H-I and FN-C/H-III, as well aspolyclonal antibodies against FN-C/H-I and FN-C/H-III, partially inhibited the adhesion of RCE cells tothe 33/66 kD fragments of fibronectin. The failure ofsoluble peptides and antibodies against these peptidesto completely block RCE cell adhesion to the 33/66kD fragments of fibronectin is consistent with pre-vious reports that describe multiple active domainswithin the 33/66 kD fragments,12"16'22 and suggeststhat additional sites in the 33/66 kD heparin-binding/cell adhesion-promoting domains of fibronectin may

contribute to RCE cell adhesion to these fragments.However, soluble FN-C/H-I, FN-C/H-III, and anti-bodies against these peptides did not effectively blockRCE cell adhesion to fibronectin. This is not surpris-ing given the presence of the RGDS adhesion-pro-moting sequence in intact fibronectin.10

The molecular basis of RCE cell adhesion to thesepeptides remains unknown. A number of reportshave shown that cell surface proteoglycans can play arole in cell adhesion to extracellular matrix glycopro-teins.26"30 A cell surface proteoglycan may be involvedin RCE cell adhesion to FN-C/H-I and FN-C/H-IIIbecause these peptides bind the glycosaminoglycanheparin.12"16 Studies are underway to determine therole of heparan-sulfate proteoglycans in RCE cell ad-hesion to these peptides. We do not know what role, ifany, cell surface integrins31-32 play in RCE cell adhe-sion to FN-C/H-I and FN-C/H-III. However, pre-vious reports have shown that peptide CS-1 promotesthe adhesion of other cell types via interactions withthe a4j8, integrin.33"35 Thus, the failure of CS-1 to pro-mote RCE cell adhesion under the conditions used inthese studies suggests that RCE cells lack the a4/3, in-tegrin or that it is present in a functionally inactivestate.

Given the important role of fibronectin in cornealwound healing in vivo, the elucidation of mecha-nisms mediating RCE cell adhesion to fibronectinmay contribute to our understanding of corneal epi-thelial cell interactions with the extracellular matrixin normal and pathologic processes. In addition, thesequences defined by peptides FN-C/H-I and FN-C/H-III may be useful substitutes for intact fibronectinas topical stimulants of corneal reepithelializationand wound healing in vivo.

Key words: cornea, epithelium, adhesion, motility, fibro-nectin

Acknowledgments

The authors thank Dr. Joji Iida for his critical commentsduring the preparation of this manuscript, and Eric Kjelles-vig, Xiaoling Wang, and Judith Kahm for their excellenttechnical assistance during these studies.

References1. Berman M, Manseau E, and Law M: Ulceration is correlated

with degradation of fibrin and fibronectin at the corneal sur-face. Invest Ophthalmol Vis Sci 24:1358, 1983.

2. Yamada KJV1 and Olden K: Fibronectin: Adhesive glycopro-teins of cell surface and blood. Nature 175:179, 1978.

3. Suda T, Nishida T, Ohashi T, Nakagawa S, and Manabe R:Fibronectin appears at the site of corneal stromal wound inrabbits. Curr Eye Res 1:553, 1981.

4. Cameron JD, Hagen ST, Waterfield RR, and Furcht LT: Ef-fects of matrix proteins on rabbit corneal epithelial cell adhe-sion and migration. Curr Eye Res 7:293, 1988.



5. Nishida T, Nakagawa S, Awata T, Ohashi Y, Watanabe K, andManabe R: Fibronectin promotes epithelial migration of cul-tured rabbit cornea in situ. J Cell Biol 97:1653, 1983a.

6. Nishida T, Nakagawa S, Nishibayashi C, Tanaka H, and Man-abe R: Fibronectin enhancement of corneal epithelial woundhealing of rabbits in vivo. Arch Ophthalmol 102:455, 1984.

7. Nishida T, Ohashi Y, Awata T, and Manabe R: Fibronectin: Anew therapy for corneal trophic ulcer. Arch Ophthalmol101:1046, 1983.

8. Phan T-M, Foster CS, BoruchoffSA, Zagachin LM, and Col-vin RB: Topical fibronectin in the treatment of persistent cor-neal epithelial defects and trophic ulcers. Am J Ophthalmol104:494, 1987.

9. Sekiguchi K, Fukuda M, and Hakomori S-J: Domain structureof hamster plasma fibronectin. J Biol Chem 256:6452, 1981.

10. Pierschbacher MD and Ruoslahti E: The cell attachment activ-ity of fibronectin can be duplicated by small synthetic frag-ments of the molecule. Nature 309:30, 1984.

11. McCarthy JB, Hagen ST, and Furcht LT: Human fibronectincontains distinct adhesion- and motility-promoting domainsfor metastatic melanoma cells. J Cell Biol 102:179, 1986.

12. McCarthy JB, Skubitz APN, Zhao Q, Yi X-Y, Mickelson DJ,Klein DJ, and Furcht LT: RGD-independent cell adhesion tothe carboxyl-terminal heparin-binding fragment of fibronectininvolves heparin-dependent and -independent activities. J CellBiol 110:777, 1990.

13. Haugen PK, McCarthy JB, Skubitz APN, Furcht LT, and Le-tourneau PC: Recognition of the A chain carboxyl-terminalheparin binding region of fibronectin involves multiple sites:Two contiguous sequences act independently to promoteneural cell adhesion. J Cell Biol 111:2733, 1990.

14. McCarthy JB, Chelberg MK, Mickelson DJ, and Furcht LT:Localization and chemical synthesis of fibronectin peptideswith melanoma adhesion and heparin binding activities. Bio-chemistry 27:1380, 1988.

15. Hda J, Skubitz APN, Furcht LT, Wayner E, and McCarthy JB:Coordinate role for cell surface chondroitin sulfate proteogly-can and a4 /?, integrin in mediating melanoma cell adhesion tofibronectin. J Cell Biol 118(2):431, 1992.

16. Humphries MJ, Komoriya A, Akiyama SK, Olden K, and Ya-mada KM: Identification of two distinct regions of the type IIIconnecting segment of human plasma fibronectin that pro-mote cell type specific adhesion. J Biol Chem 262:6886, 1987.

17. Mooradian DL, Weatherbee J, Lucas R, and Furcht LT: Trans-forming growth factor-/?, binds to immobilized fibronectin. JCell Biochem 41:189, 1990.

18. Koliakos GG, Kouzi-Koliakos K, Furcht LT, Reger LA, andTsilibary EC: The binding of heparin to type IV collagen: Do-main specificity with identification of peptide sequences fromthe a 1 (IV) and «2(IV) chain which preferentially bind heparin.J Biol Chem 264:2313, 1989.

19. Kyte J and Doolittle RF: A simple method for displaying thehydropathic character of a protein. J Mol Biol 157:105, 1982.

20. Danilov YN and Juliano RL: (Arg-Gly-Asp)n-albumin conju-

gates as a model substratum for integrin-mediated cell adhe-sion. ExpCell Res 182:186, 1989.

21. Skubitz APN, Charonis AS, Tsilibary EC, and Furcht LT: Lo-calization of a tumor cell adhesion domain of laminin by amonoclonal antibody. Exp Cell Res 173:349, 1987.

22. Wilke MS, Skubitz APN, Furcht LT, and McCarthy JB: Hu-man keratinocytes adhere to two distinct heparin-binding syn-thetic peptides derived from fibronectin. J Invest Dermatol97:573, 1991.

23. Nishida T, Nakagawa S, Watanabe K, Yamada KM, Otari T,and Berman MB: A peptide from fibronectin cell-binding do-main inhibits attachment of epithelial cells. Invest OphthalmolVis Sci 29:1820, 1988.

24. Izzard CS, Radinsky R, and Culp LA: Substratum contacts andcytoskeletal reorganization of BALB/c 3T3 cells on a cell-bind-ing fragment and heparin-binding fragments of plasma fibro-nectin. ExpCell Res 165:320, 1986.

25. Liao NS, St. John J, McCarthy JB, Furcht LT, and Cheung HT:Adhesion of lymphoid cells to the carboxyl terminal heparin-binding domain of fibronectin. ExpCell Res 181:348, 1989.

26. Rapraeger AC and Bernfield M: Heparan sulfate proteoglycansfrom mouse mammary epithelial cells: A putative membraneproteoglycan associates quantitatively with lipid vesicles. J BiolChem 258:3632, 1983.

27. Woods A, Couchman JR, and Hook M: Heparan sulfate pro-teoglycans of rat embryo fibroblasts. J Biol Chem 260:10872,1985.

28. Mugnai G, Lewandowska K, Choi HU, Rosenberg LC, andCulp LA: Ganglioside-dependent adhesion events of humanneuroblastoma cells regulated by the RGDS-dependent fibro-nectin receptor and proteoglycans. Exp Cell Res 175:229,1988.

29. Culp LA, Mugnai G, Lewandoska K, Vallen EA, Kosir MA,and Houmiel KL: Heparan sulfate proteoglycans of ras-trans-formed 3T3 or neuroblastoma cells. Ann NY Acad Sci556:194, 1989.

30. Hook M, Kjellen L, Johansson S, and Robinson J: Cell-surfaceglycosaminoglycans. Annu Rev Biochem 53:847, 1984.

31. Hynes RO: Integrins, a family of cell surface receptors. Cell48:549, 1987.

32. Ruoslahti E and Pierschbacher MD: New perspectives in celladhesion: RGD and integrins. Science 233:491, 1987.

33. Wayner EA, Garcia-Pardo A, Humphries MJ, McDonald JA,and Carter WG: Identification and characterization of the Tlymphocyte adhesion receptor for an alternative cell attach-ment domain (CS-1) in plasma fibronectin. J Cell Biol109:1321, 1989.

34. Guan JL and Hynes RO: Lymphoid cells recognize an alterna-tively spliced segment of fibronectin via the integrin receptora4/?,. Cell 60:53, 1990.

35. Mould AP, Wheldon LA, Komoriyama A, Wayner EA, Ya-mada KM, and Humphries MJ: Affinity chromatographic iso-lation of the melanoma adhesion receptor for the IIIcs region offibronectin and its identification as the integrin aAp,. J BiolChem 265:4020, 1990.


Rabbit corneal epithelial cells adhere to two distinct heparin-binding ...

Documents