Role of Fibronectin in the Healing of Superficial Keratectomies In Vitro

Transcription

1 prr Investigative Ophthalmology & Visual Science, Vol. 30, No. 3, March 1989 Copyright Association for Research in Vision and Ophthalmology Role of Fibronectin in the Healing of Superficial Keratectomies In Vitro Tuyer-Moi M. Phan.t C. Stephen Fosfer.f Luba M. Zagachin,* and Robert B. Colvin* A fibronectin (Fn)-fibrinogen (Fg) surface matrix is not essential for epithelial cell migration in a corneal epithelial scrape wound model, in which the basement membrane is preserved. We have therefore tested whether such a provisional scaffolding becomes more critical in a superficial keratectomy model, when the basement membrane is surgically removed. Exogeneous Fn at 03 mg/ml was added to the medium of organ cultures of rabbit superficial keratectomies. At 48 hr after wounding, the healing rate was 1.12 ± 0.03 mm 2 /hr in control and 1.11 ± 0.03 mm 2 /hr in those cultured with Fn. At 64 hr after wounding, the healing rates were also not significantly different (P > 0.5). Immunofluorescence studies showed that Fn was not detectable on the surface of control but could be depicted under the migrating epithelium. In cultured with Fn, it diffused throughout the entire stroma but did not deposit as a surface matrix. We therefore attempted to obtain formation of a provisional Fn-Fg surface matrix before establishment of the in vitro organ culture by leaving the superficial keratectomies in vivo for 8 hr, 24 hr or 64 hr. At 64 hr after wounding, their healing rate was 0.80 ± 0.04 mm 2 /hr, 0.86 ± 0.04 mm 2 /hr and 0.85 ± 0.06 mm 2 /hr, respectively, which were not significantly different from that of contralateral ex vivo-wounded cultured (0.83 ± 0.04 mm 2 / hr, P > 0.5). Immunofluorescence studies revealed a Fn matrix on the bare surface of in vivo-wounded specimen, which was not detectable on ex vivo-wounded cultured ; there was also no diffuse stromal Fn. In those ex vivo-wounded specimen, however, Fn was present under the migrating epithelium. These data suggest endogeneous production of Fn by corneal cells at the wound edges. We conclude that an exogeneously supplied Fn scaffolding is not essential for wound healing in a superficial keratectomy model: an alternate substrate for epithelial cell migration is endogeneously synthesized by cornea! cells at the migrating edges. Invest Ophthalmol Vis Sci 30: ,1989 Fibronectin (Fn) is a plasma and cellular glycoprotein which mediates fibroblast adhesion to collagen matrices and fibrin substrata. 12 Fn is a prominent component of granulation tissue 3 " 5 and has been thought to play an important role in healing of connective tissue. The role of Fn in epithelial wound healing and cell migration has, however, been more controversial. 6 " 10 In the rabbit cornea, little or no Fn exists in the nor- From the "Immunopathology Unit, Department of Pathology, Massachusetts General Hospital, the fhilles Immunology Laboratory, Massachusetts Eye and Ear Infirmary, and Harvard Medical School, Boston, Massachusetts. Dr. Phan's current address is: Jules Stein Eye Institute, Los Angeles, California. Supported in part by grants CA from the National Institute of Health, Bethesda, Maryland, Fight for Sight Grant-in-aid GA , and by the National Society to Prevent Blindness. Dr. Phan was a recipient of a National Eye Institute Fellowship (F32- EY 05760). Submitted for publication: April 4, 1988; accepted September 7, Reprint requests: Robert B. Colvin, MD, Department of Pathology, Cox 5, Massachusetts General Hospital, Boston, MA mal epithelial basement membrane, but deposits along with fibrin on the denuded basement membrane as a continuous, linear layer shortly after a superficial scrape wound." A similar, more prominent Fn band can also be detected on the bare stromal surface of a corneal superficial keratectomy. 12 In both models, this layer slowly disappears as the wound is reepithelialized. Such a surface matrix is conspicuously absent on which have been wounded ex vivo. 13 We have previously demonstrated that a surface provisional fibronectin-fibrinogen (Fg) scaffolding is not essential for epithelial cell migration in a superficial scrape wound model: its inhibition did not alter the normal wound healing process. 13 In such a model, laminin remains firmly anchored to the underlying lamina densa region of the basement membrane." 12 Since laminin itself is a normal component of the corneal basement membrane and, according to some studies, 89 serves as a preferred attachment glycoprotein for epithelial cells, we have speculated that a provisional Fn-Fg substrate may not be crucial for cell migration when laminin remains intact. In contrast, wounds that destroy or damage the underlying 386

2 No. 3 FIBRONECTIN IN SUPERFICIAL KERATECTOMIES / Phan er al 387 Table 1. In vitro healing rate of rabbit superficial keratectomies: addition of exogeneous Fn Study group Control Fibronectin Control Fibronectin eyes Wound duration 48 hr 48 hr 64 hr 64 hr Healing rale (mm'/hr) ± ± ± 0.03 P value >0.5 >0.5 trephine visualized through a Zeiss operating microscope. Using a #69 Beaver blade, the demarcated epithelium was surgically removed along with its basement membrane and the anterior third of the stroma. The wounded cornea was then excised with its scleral basement membrane may show greater dependence on this temporary scaffolding. We used a rabbit superficial keratectomy model in which the basement membrane has been surgically removed with the anterior stromal third to test this hypothesis. In an organ culture system, the effect of exogeneously added Fn on wound healing rate was measured and the role of an artificially formed Fn-Fg surface matrix in epithelial cell migration determined. Materials and Methods Fibronectin Purification Fn was purified from rabbit plasma by affinity chromatography according to a modified method of Engwall and Ruoslahti. 1 4 Plasma was obtained from 0.38% citrated blood by centrifugation and applied to a lysine-coupled Sepharose 4B affinity column to remove plasminogen. It was then passed through a gelatin-coupled Sepharose 4B affinity column to isolate Fn. After extensive washing with phosphate-buffered saline (PBS, ph 7.2) and equilibration with an isotonic saline solution (ph 7.2), Fn was eluted with 0.15 M NaCl, ph 2.2 and dialyzed against 0.1 M NH4HCO3 and PBS. Electrophoresis of unreduced Fn on a 4 to 12% polyacrylamide slab gel revealed a closely spaced doublet of 440 kd. Upon reduction, Fn migrated as a doublet in the 220 kd region. Purified Fn was added to a culture medium as defined below at a concentration of 0.3 mg/ml. Organ Culture System Excised were rinsed in Ringer's solution and cultured in a defined medium at 37 C as previously described. 15 Protocol A: Addition of Exogeneous Fn Animals and treatment: Sixteen male and female New Zealand albino rabbits weighing 3.0 to 3.5 kg were killed with intravenous sodium pentobarbital at 100 mg/kg. Following enucleation, an 8.5 mm diameter zone was demarcated in the central cornea by a Fig. 1. Immunofluorescence (IF) micrographs of rabbit superficial keratectomies cultured in serum-free medium, stained for Fn and Fg. (A) At 64 hr after wounding, Fn is conspicuously absent on the bare wound surface. (B) A Fg surface substrate is also not detectable. (C) However, Fn is endogenously synthesized under the migrating epithelium at the wound edges. Magnifications XI50.

3 388 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989 Vol. 30 Fig. 2. IF micrographs of rabbit superficial keratectomies cultured in the presence of exogeneous Fn at 0.3 mg/ml, stained for Fn and Fg. (A) At 64 hr after wounding, a discrete Fn surface matrix is not present; there is, however, diffuse increase in fluorescence for Fn throughout the entire stroma. (B) Fg remains undetectable. Magnifications XI50. rim and immediately cultured. In each rabbit, one cornea was cultured in a control medium without serum while the contralateral cornea was cultured in a similar medium with added Fn at 0.3 mg/ml. Experiments were carried out for 48 hr or 64 hr. Animal care conformed to the Arvo Resolution on the Use of Animals in Research. Epithelial defects were stained with 0.5% methylene blue at time 0 and at 48 hr or 64 hr and photograph-documented. The area of the epithelial defect was determined by computerized planimetry and healing rate expressed as square millimeters per hour. The student t-test was used for statistical analysis of differences in healing rates between different treatment groups. Immunojluorescent techniques: At time 48 hr and 64 hr after wounding, rabbit within each study group were frozen in O.C.T. embedding compound (Lab Tech, Miles Laboratories, Naperville, IL). Four micrometer sections were stained with the following: (1) fluoresceinated goat IgG reactive against rabbit Fn" (1:40); and (2) fluoresceinated goat IgG reactive against rabbit Fg (1:40; Cappel Laboratories, Cochranville, PA). Protocol B: Formation of an Artificial Fn-Fg Surface Matrix Animals and treatment: Fifteen male and female New Zealand albino rabbits, weighing 3.0 to 3.5 kg were anesthized with intravenous sodium pentobarbital at 30 mg/kg and with topical 0.5% proparacaine hydrochloride. As described in section A, an 8.5 mm diameter central superficial keratectomy was performed. In each rabbit, the right cornea was wounded in vivo and removed at sacrifice 8 hr, 24 hr or 64 hr later for culture. The left cornea was wounded ex vivo at the time of sacrifice and immediately cultured. A defined culture medium was used as described above. Epithelial defects were stained with 0.5% methylene blue at time 0 and 64 hr and photograph-documented. The healing rate was determined by computerized planimetry and the paired student t-test was used for statistical analysis. At time 64 hr, rabbit within each treatment group were frozen and 4-fim sections stained with antibodies against rabbit Fn and Fg, as described above. Results A: Addition of Exogeneous Fibronectin Healing rate (Table 1): Addition of exogeneous Fn to the culture medium did not accelerate the rate of epithelial cell migration. At 48 hr after wounding, the healing rate of superficial keratectomies cultured in a medium containing Fn at 0.3 mg/ml (1.12 ± 0.03 mm 2 /hr) was not significantly different from that of control specimen (1.11 ± 0.03 mm 2 /hr, P > 0.5). Similarly, no difference could be detected between control and Fn-cultured at 64 hr (P > 0.5). Immunohistologic studies: As in the superficial scrape wound model, 13 a Fn-Fg matrix could not be detected on the bare surface of a rabbit superficial keratectomy which was cultured in serum-free medium for 48 or 64 hr (Fig. 1 A, B). However, Fn can be depicted under the migrating epithelium in such specimens, suggesting active endogeneous Fn synthesis (Fig. 1C). When exogeneous Fn was added to the culture medium, a discrete Fn surface substrate was not present; there was, however, diffuse increase in Fn throughout the entire stroma of cultured

4 No. 3 FIBRONECTIN IN SUPERFICIAL KERATECTOMIES / Phon.er ol 389 Table 2. In vitro healing rate of ex vivo-wounded and in vivo-wounded corneal superficial keratectomies in the rabbit: formation of an artificial Fn-Fg matrix Study group #of eyes Wound duration Healing rate (mm'/hr) P value ex vivo-wounded in vivo-wounded in vivo-wounded in vivo-wounded hr in vitro 8 hr in vivo 56 hr in vitro 24 hr in vivo 40 hr in vitro 64 hr in vivo 0.83 ± ± ± ± 0.06 >0.2 >0.3 >0.4 with exogeneous Fn (Fig. 2A). Fg was not present in Fn-treated (Fig. 2B). B: Formation of an Artificial Fn-Fg Surface Matrix Healing rate (Table 2): Corneas wounded in vivo were left in situ for 8 and 24 hr before organ culture to permit formation of a surface Fn matrix At 64 hr after wounding, the healing rate of in vivo-wounded on which a surface Fn-Fg matrix had been deposited on the bare stroma was 0.80 ± 0.04 mm 2 / hr to 0.86 ± 0.04 mm 2 /hr. This was not significantly different from that of ex vivo-wounded superficial keratectomies on which no Fn-Fg surface matrix was allowed to develop (0.83 ± 0.04 mm 2 /hr, P > 0.5). Immunohistologic studies: Within 8 hr of in vivo wounding, Fn was deposited on the denuded stroma as a thick, irregular band (Fig. 3A), then gradually increased in intensity reaching a maximum at 24 hr (Fig. 3B). By 64 hr after wounding, the Fn matrix had Fig. 4. Rabbit superficial keratectomy wounded and left in vivo for 8 hr before organ culture: Fg also forms a prominent matrix on the bare wound surface. (X230) begun to diminish as the epithelial defect closed. As previously described, 12 Fg also deposited on the wound surface in the same distribution as Fn (Fig. 4). On the contrary, when keratectomies were inflicted ex vivo and immediately cultured in a serum-free medium, no Fn-Fg matrix could be detected on the bare surface (Fig. 5A, B). However, even in those ex vivo-wounded, Fn can be detected at the wound edges and under the migrating epithelium (Fig. 1C), suggesting active endogeneous Fn synthesis. Fg could not be detected either on the bare surface or under the migrating epithelial edges. Discussion These studies demonstrate that the addition of exogeneous fibronectin did not accelerate in vitro epi- Fig. 3. IF micrographs of rabbit superficial keratectomies stained for Fn. (A) Cornea was wounded and left in vivo for 8 hr before organ culture in serum-free medium for 56 hr (total time after wounding is 64 hr): a prominent, band-like Fn surface matrix is present on the denuded stroma. It reaches maximal intensity when left in vivo for 24 hr (B). Magnifications X23O.

5 390 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / March 1989 Vol. 30 Fig. 5. When superficial keratectomies were inflicted ex vivo and immediately cultured for 64 hr, little or no Fn (A) or Fg (B) can be detected on the bare wound surface, although fine fibrillar deposits are seen in the stroma. thelial migration in a superficial keratectomy model. The role of fibronectin in adhesion and spreading of epithelial cells on plastic dishes has been debated. 6 " 8 Our results disagree with those reported by Nishida et al, 16 who showed that exogeneous fibronectin promotes the migration of corneal epithelial cells on stromal collagen in situ. One should note that Nishida used separate corneal blocks of different sizes, with the length of epithelial migration measured along the cut stromal edge after fixation, embedding, sectioning and staining with hematoxylin and eosin. On the contrary, standard 8.5 mm diameter stromal wounds were used in our study, with the area of epithelial defect at different times measured by direct methylene blue staining and computerized planimetry of projected photographs. Since we did not obtain formation of a discrete fibronectin surface matrix through addition of purified fibronectin to the culture medium, such a substrate was created by inflicting a superficial keratectomy and leaving it in vivo for various amounts of time before excision of the wounded cornea for organ culture. The in vitro healing rate of such, however, did not significantly differ from that of contralateral ex vivo-wounded specimen without a Fn-Fg surface matrix. Such a naturally formed scaffolding is therefore not essential for epithelial cell migration and does not enhance the rate of wound closure in this experimental model. These studies do not rule out a role of Fn, even in this model of wound healing, because a granular layer of Fn is formed immediately under the migrating epithelium in the absence of exogeneous Fn (tears, plasma), suggesting endogeneous production of Fn by corneal cells at the wound edges. Such synthesis may in fact provide a sufficient alternate substrate for epithelial cellular migration. One may therefore speculate that, in a wound in which corneal cells at the edge preserve their intrinsic biochemical ability to synthesize their own fibronectin matrix, an exogeneously supplied surface substrate becomes nonessential. It is therefore possible that exogeneous Fn becomes more critical for wound healing in injuries with extensive tissue damage and inflammatory cellular infiltration, such as chemical burns or infections. Cavanagh et al 17 have suggested that active inflammation and tissue damage may profoundly alter the intracellular biochemical machinery, leading to a change in epithelial cellular morphology and to paralysis of both locomotion and mitosis: in such persistent epithelial defects, corneal cells at the edges may not retain their intrinsic ability to synthesize fibronectin endogeneously. We have demonstrated that exogeneous Fn accelerates healing of recurrent or persistent epithelial defects and decreases the incidence of stromal ulceration in a rabbit alkali bum model. 18 In such wounds, Fn is not present under the migrating epithelium; moreover, the surface Fn-Fg matrix which is normally present on all corneal wounds in vivo is rapidly destroyed 60 to 72 hr after wounding. This disintegration corresponds clinically to the phase of secondary epithelial breakdown and persistent epithelial defect; histopathologically, it is associated with lysis of keratocytes, breakdown of the normal stromal collagen structure and a dense inflammatory infiltrate. Key words: fibronectin,fibrinogen,corneal wound healing, epithelial migration References 1. Grinnell F, Feld M, and Minter D: Fibroblast adhesion to fibrinogen and fibrin substrata: Requirement for cold-insoluble globulin (plasmafibronectin).cell 19:517, 1980.

6 No. 3 FIDRONECTIN IN SUPERFICIAL KERATECTOMIES / Phon er ol Klebe RJ: Isolation of a collagen-dependent cell attachment factor. Nature 250:248, Kurkinen M, Vaheri A, Roberts PJ, and Stenma'n S: Sequential appearance of fibronectin and collagen in experimental granulation tissue. Lab Invest 43:47, Grinnell F, Billingham RE, and Burgess L: Distribution of fibronectin during wound healing in vivo. J Invest Derm 76:181, Repesh LA, Fitzgerald TJ, and Furcht LT: Fibronectin involvement in granular tissue and wound healing in rabbits. J Histochem Cytochem 30:351, Federgreen W and Stenn KS: Fibronectin (LETS) does not support epithelial cell spreading. J Invest Dermatol 75:261, Terranova VP, Rohrback DH, and Martin GR: Role of laminin in the attachment of PAM 212 (epithelial) cells to basement membrane collagen. Cell 22:719, Murray TC, Stingl C, Weinman HK, Martin GR, and Katz SI: Epidermal cells adhere preferentially to type IV (basement membrane) collagen. J Cell Biol 80:197, Burrill PH, Bemardini J, KJeinman HK, and Kretchmer N: Effect of serumfibronectinand laminin on adhesion of rabbit intestinal epithelial cells in culture. J Supramolec Struct Cell Biochem 16:385, Gilchrest BA, Calhoun JK, and Maciag T: Attachment and growth of human keratocytes in a serum-free environment. J CellPhysiol 112:197, Fujikawa LS, Foster CS, Harrist TJ, Lanigan JM, and Colvin RB: Fibronectin in healing rabbit corneal wounds. Lab Invest 45:120, Fujikawa LS, Foster CS, Gipson CS, and Colvin RB: Basement membrane components in healing rabbit corneal epithelial wounds. J Cell Biol 98:128, Phan TM, Foster CS, Wasson PJ, Fujikawa LS, Zagachin LM, and Colvin RB: Role offibronectinandfibrinogenin healing of corneal epithelial scrape wounds. Invest Ophthalmol Vis Sci 30:377, Engvall E, Ruoslahti E, and Miller EJ: Affinity of fibronectin to collagens of different genetic types and tofibrinogen.j Exp Med 147:1584, Gipson IK and Anderson RA: Effect of lectinson migration of the corneal epithelium. Invest Ophthalmol Vis Sci 19:341, Nishida T, Nakagawa S, Awata T, Ohashi J, Watanabe K, and Manabe R: Fibronectin promotes epithelial migration of cultured rabbit cornea in situ. J Cell Biol 97:1653, Cavanagh HD, Pihlaja D, Thoft RA, and Dohlman CH: The pathogenesis and treatment of PED. Trans Am Acad Ophthalmol Otolaryngol 81:754, Phan TM, Foster CS, Shaw CE, and Colvin RB: Topical fibronectin in a rabbit alkali burn model of corneal ulceration ARVO abstracts. Invest Ophthalmol Vis Sci 26(Suppl):176, 1985.