Effects of fibroblasts and basic fibroblast growth factor on facilitation of dermal wound healing by type I collagen matrices Michael G . Marks, Charles Doillon,* and Frederick H. Silvert Biomaterials Center and Department of Pathology, UMDNJ-Robert Wood Johnson Medical School, Piscataway, NJ 08854 Healing of large open dermal wounds is associated with decreased values of the tensile strength even up to 6 months postwounding. Results of previous studies have shown that healing is facilitated in the presence of a type I collagen sponge by promoting deposition of newly synthesized large-diameter collagen fibers parallel to the fibers of the sponge. In this study healing is evaluated in dermal wounds treated with a collagen sponge seeded with fibroblasts or coated with basic fibroblast growth factor (bFGF). Experimental results indicate that the presence of a collagen sponge results in
increased wound tensile strength and increased collagen fiber diameters in the upper dermis 15 days post-wounding in an excisional guinea pig dermal wound model. In comparison, dermal wounds treated with collagen sponges seeded with fibroblasts or coated with bFGF showed increased tensile strengths 15 days postimplantation and increased degree of reepithelialization. These results indicate that fibroblast seeding and bFGF coating in conjunction with a type I collagen sponge matrix facilitate early dermal and epidermal wound healing.
IN TRODUCTION
Replacement of skin damaged as a result of thermal and mechanical trauma or contact with a pathogenic agent has been the subject of intense research interest especially in the hope of development of an “artificial skin.” The consequence of this research has been the development of (a) improved skin grafting technique~l-~ using both human and animal skins: (b) use of biological membranes such as amnion as a burn (c) development of wound dressings composed of and bi~logical’~~-’~ polymers, and (d) development of fibroblast13-16 and epithelial ~ell’~-’l seeding techniques. A variety of growth factors including epidermal growth factor, fibroblast growth factors (acidic and basic), platelet-derived growth factor, transforming growth factors (alpha and beta) and insulin-like growth factor influence the behavior of cells found in the extracellular matrix” and are likely to be ultimately useful as components of “artificial skin.” *Current address: Lab. of Experimental Surgery, Pavillon Des Services, Lava1 University, Cite Universitaire, Quebec, GlK7P4, Canada. tTo whom correspondence should be addressed. Journal of Biomedical Materials Research, Vol. 25, 683-696 (1991) CCC 0021-9304/91/050683-14$4.00 0 1991 John Wiley & Sons, Inc.
MARKS, DOILLON, A N D SILVER
684
Basic fibroblast growth factor (bFGF) stimulates angiogene~is?~,~~ proliferation of k e r a t i n ~ c y t e and s ~ ~is~ associated ~ with the extracellular matrix.28Iodinated bFGF binds to heparan sulfate proteoglycan but not to laminin or type IV collagen and binding is inhibited by hepa~in.2~ Specific interaction sites for bFGF are widely distributed in mammalian tissues.30bFGF levels are elevated in tissue that displays persistent regeneration3' and in Schwann ~ m a . ~Exogenous ' addition of FGF promotes wound healing in a rabbit cornea Acidic fibroblast growth factor (aFGF) in the presence of heparin stimulates a n g i ~ g e n e s i sand ~ ~ -binds ~ ~ to type I and IV c011agens.~~ It has been reported that the mitogenic activity of bovine aFGF is stabilized by preincubation with culture medium containing heparin.38 The objective of this study is to compare the wound healing response of a type I collagen matrix that is coated with bFGF with that observed in the presence of a collagen matrix seeded with fibroblasts. Results presented in this study indicate that type I collagen matrices that are treated once with bFGF before implantation or seeded with fibroblasts in cell culture faciIitate the development of wound tensile strength and reepithelialization when implanted on full-thickness guinea pig excised dermal wounds.
MATERIALS A N D METHODS
Type I collagen preparation Insoluble collagen, obtained from Devro, Inc. (Somerville, NJ) was prepared as described p r e v i o u ~ l y and ~ ~ -was ~ ~ characterized as being typical of type I by sodium dodecyl sulfate polyacrylamide gel electrophoresis and amino acid analysis. Collagen sponge preparation For each collagen sponge preparation, 30 mL of a collagen dispersion was poured into a 100 x 15 mm petri dish (Falcon, model 1029) as described previously.'6*21 The dispersion was spread evenly across the bottom of the dish, and any remaining air bubbles and large pieces of collagen were removed. The petri dishes were covered with lids and placed in a freezer for at least 3 h. The materials were then lyophilized at 0°C for 24 h. After removal from the petri dishes, the sponges were crosslinked using a two-step process involving dehydrothermal treatment followed by exposure to cyanamide. Collagen sponge crosslinking Sponges crosslinked by severe dehydration at 110°C for 3 days and then placed in a bath containing 1 L of 1% w/v aqueous cyanamide for 24 h at
FACILITATION OF DERMAL WOUND HEALING
685
pH 7.2 as described previously.16,21 They were then lyophilized, coated with a layer of silicone rubber, and then sterilized by exposure to a dose of 2.5 Mrad of gamma radiation. Basic FGF was obtained from Amgen, Inc. (Tall Oaks, CA). Dilutions were made to yield 10, 25, and 50-puglmL solutions in a 400-pL volume. Using an Eppendorf pipette with sterile tips, 400-pL was dispersed across the middle third region of a sterilized collagen sponge, an area of about 15 cm2(on the side that was not coated with silicone), under a laminar air flow bench. The bFGF solution was allowed to air dry prior to implantation under a laminar air flow hood. Isolation of fibroblasts Fibroblasts were isolated from autologous guinea pig skin and then placed in a petri dish containing sterile phosphate buffer solution (PBS). Under a laminar-flow hood the skin was scraped on both sides and then minced. It was then soaked in a sterile solution containing collagenase (440 units/mg) and Dulbecco’s Modified Eagle’s Medium (DMEM) and incubated for 24 h at 37°C in 5% C02. Digested pieces of skin were separated and then put into a cell culture flask. The cells were given fresh medium, and allowed to proliferate until they were confluent as described previously.” Cell culture techniques Cells were fed every other day with DMEM containing 10%calf serum and antibiotics. When confluent, cells were split and seeded in new flasks. Cells were passaged for several generations before seeding onto collagen sponges. Sterile collagen sponges, about 50 cm2in surface area, were saturated with DMEM containing 10% calf serum. 1 x lo5 cells were dispersed throughout the sponge and allowed to proliferate for 1 week prior to implantation as described previously.’6 Animal surgery
White female Hartley Albino guinea pigs weighing 400-500 g were obtained from Charles River laboratories (Wilmington, MA). Animals were acclimated to vivarium housing for 1 week prior to surgery and were fed and given water supplemented with vitamin C ad Eibitum. Guinea pigs were anesthetized with intraperitoneal injections of ketamine (100 mg/mL, dose: 35 mg/kg) and xylazine (1 mg/mL, dose: 5 mg/kg) and the backs were then shaved, depilated, and cleansed with a povidine-iodine antiseptic solution. Using sterile procedure, an elliptical full-thickness wound, 10 cm x 3 cm, down to and including the panniculus carnosus, was excised transverse to the animal’s spine.
686
MARKS, DOILLON, AND SILVER
Skin autografts were used as a positive control (Graft) and open wounds (Open) were used as a negative control. Experimental treatment groups included a collagen sponge (Sponge), a collagen sponge with fibroblasts (FIB), or a collagen sponge containing bFGF (FGF). In each case, a sponge coated with silicone adhesive was cut to the size of the wound and clipped to the wound periphery using 11-mm wound clips. During implantation, wounds were inspected and any infected wounds were eliminated from the study. Animals were sacrificed at 15 and 30 days postimplantation by an overdose of sodium pentobarbitol injected intraperitoneally. ,4minimum of three animals per treatment group were used; an average of three skin strips per animal were available for mechanical testing at sacrifice. After sacrifice, the granulation tissue and unwounded tissue on both edges of the wound were cut into parallel strips about 5 mm wide and excised on a tissue plane just above the fascia using a razor blade as previously deSkin strips were then placed in PBS and immediately tested in uniaxial tension on an Instron testing machine (Instron Corporation, model 1122). In addition, a small piece of wound tissue was excised approximately 5 mm from the right wound edge and placed in Carson’s fixative. Tissue specimens were dehydrated using routine methods, paraffin embedded, and stained with hematoxylin and eosin (H&E),and picro-Sirius red for light microscopic analysis.
Histological analysis Tissue sections cut perpendicular to the epithelial surface were stained with H&E and picro-Sirius red. H&E stained sections were examined for the amount of sponge remaining in the granulation tissue, degree of reepithelialization, and arrangement of newly deposited collagen. Picro-Sirius red stained sections were examined for collagen fiber thickness as a function of depth. A Leitz light microscope (Laborlux 12 Pol) was used for histological evaluation.
Biomechanical testing Skin strips were tested on an Instron tester in uniaxial tension. The sample was stretched between the Instron grips until it was slightly taut and positioned with the wound tissue centered. The cross-sectional area of the sample was measured by assuming the sample was a rectangle and measuring the width and thickness of the sample using calipers. The gage length (starting sample length) was set to 10 mm and specimen width and thickness used were 3 cm and 0.5 cm, respectively. The crosshead speed was set to 1 mm/min (10% strain rate), and the chart speed was set to 20 mm/min. Breaking loads were measured directly from the force-displacement graph. Breaking energy was calculated by measuring the area under the forcedisplacement curve using a Hipad digitizer (Houston Instruments) interfaced
FACILITATION OF DERMAL WOUND HEALING
687
with a computer (IBM-PC). Ultimate tensile strength was determined by dividing breaking load by the cross sectional area. Values were reported as means 2 standard error and compared statistically using a one-tailed, unpaired Student’s t test based on at least nine observations per treatment group. A p value less than 0.05 was considered significant. Biomechanical tests were discarded if failure was observed at the grip surface or slippage was observed at the grip-tissue interface. Test results were also discarded if the sample was not uniform or appeared damaged.
RESULTS
Biomechanical measurements The load required to cause failure of dermal wound tissue (Fig. 1) was found at 15 days postimplantation to be lowest for the control open wound (451 ? 80 g), higher for the collagen sponge, bFGF-treated, and fibroblasttreated wounds and the breaking load was highest for the skin graft (2128 k 636 g). Breaking loads for the collagen sponge were significantly greater at 15 days postimplantation compared to the open wounds ( p < 0.05) and those for the skin graft were significantly greater than fibroblast and bFGF-treated wounds. At 30 days postimplantation the breaking load was
3000
T
OPEN s#TJGE 15days
FIB
FFF
GRAFT
30 days
Figure 1. Breaking load of open wounds. Relationship between dermal wound breaking load and treatment for open wounds (OPEN), wounds treated with collagen sponges (SPONGE), fibroblast-seeded collagen sponges (FIB), basic fibroblast growth factor (10 p g ) coated-collagen sponges (bFGF), and skin grafts (GRAFT)at 15 (black box) and 30 (hatched box) days postimplantation. Each data point represents the average of at least nine separate measurements (N = 9).
MARKS, DOILLON, AND SILVER
638
statistically greater for the collagen sponge treated wounds but not for any of the other treatments (Fig. 1)compared to the control open wounds (OPEN). When the breaking load was divided by the initial cross-sectional area to determine the engineering breaking stress (ultimate tensile strength) at 15 days postimplantation the values (Fig. 2) obtained were 0.19 2 0.04 MPa (open wounds), 0.36 +- 0.07 MPa (collagen sponge), 0.73 2 0.1 MPa (fibrobiast and bFGF-treated wounds), and 0.56 2 0.14 MPa (graft). Statistically, wounds treated with bFGF and fibroblasts have higher tensile strengths than wounds treated with collagen sponge or open wounds ( p < 0.05). At 30 days postimplantation ultimate tensile strength of open, sponge-treated, f ibroblast-treated, and bFGF-treated wounds were all about 0.4 MPa. Wounds treated with skin grafts had slightly higher ultimate strengths (Fig. 2); however, they were not significantly different from the values found for the other treatment groups. Breaking energies at 15 days postimplantation (Fig. 3) were computed from the areas under the stress-strain curves and were similar for collagensponge-treated, fibroblast-treated, and bFGF-treated wounds. For skin-graf ttreated wounds the breaking energy was significantly higher than for the other groups. At 30 days postimplantation the breaking energy for sponge treated wounds was higher than that found for the other groups; however, no significant difference existed between breaking energy at 15 and 30 days Fostimplantation for the collagen-sponge-treated group.
T
-8
0.6
c a
0.4
z cn
T
0.2
0.0 OPEN SFCNGE 15days
FIB
FGF
G
W
30days
Figure 2. Ultimate tensile strength of open wounds. Ultimate tensile strength (UTS) as a function of treatment for open wounds (OPEN), wounds treated with collagen sponges (SPONGE), fibroblast-seeded collagen sponges (FIB), bFGF (10 &-coated collagen sponges (FGF), and skin grafts (GRAFT) at 15 (black box) and 30 days (hatched box) postwounding. Each data point represents the average of at least nine separate measurements (N = 9).
FACILITATION OF DERMAL WOUND HEALING
689
800 r
r I5
600
m
400
e
200
c 3 6
a
0
Figure 3. Breaking energy of wound tissue. Relationship between breaking energy (cm’) and treatment for open wounds (OPEN), wounds treated with collagen sponges (SPONGE), fibroblast seeded collagen sponges (FIB), bFGF (10 pg)-coated collagen sponges (FGF), and skin grafts (GRAFT) at 15 (black box) and 30 (hatched box) days postwounding. Each data point represents the average of at least nine separate measurements (N = 9).
Histological analyses The extent of collagen sponge biodegradation was determined based on the staining intensity differences between collagen fibers in the sponge and newly synthesized collagen deposited in dermal tissue by light microscopic analysis (Table I). In 1 out 3 wounds treated with a collagen sponge, fragments were observed at days 15 and 30 postimplantation. Fibroblast-seeded sponges had fragments remaining in 2 out of 3 samples at both 15 and 30 days postimplantation. In wounds treated with bFGF-coated sponges, collagen fragments were observed only at day 15 postimplantation. TABLE I Amount of Collagen Sponge Remaining in Granulation Tissue
Open Graft Sponge Fib FGF4 pg FGF-10 p g FGF-25 /*g
Day 15
Day 30
-
-
-
+ ++ 0 + +
+ ++ 0 0 0
N = 3: 0 = 0/3 samples contained sponge fragments, + = 113 samples contained sponge fragments, ++ = 213 samples contained sponge fragments.
690
MARKS, DOILLON, AND SILVER
Complete epithelialization was observed at both days 15 and 30 for skin grafts. Incomplete reepithelialization was observed by day 30 for fibroblast treated wounds and at days 15 and 30 for bFGF treated wounds (Table 11). Collagen fiber diameters were determined in the upper (750 pm) and lower (1500 pm) dermis (Fig. 4). At day 15 the average fiber diameter in the upper dermis was 1.7 pm for all treatments except for the skin graft for which it was significantly higher (4.4 pm). At day 15, the fiber diameters in the lower dermis for collagen-sponge- and fibroblast-treated wounds were similar (Fig. 4). Collagen-sponge-treated wounds had significantly larger diameters when compared to bFGF-treated wounds; however, they were smaller compared to skin-graf t-treated wounds. At day 30, in the upper dermis, fiber diameters were larger for collagensponge-treated wounds than for fibroblast- and bFGF-treated wounds. Skingrafted wounds at day 30 had fiber diameters similar to those seen in collagen-sponge-treated wounds in the upper dermis. In the lower dermis at day 30, collagen-sponge-treatedwounds had significantly larger fiber diameters compared with wounds treated with fibroblasts or bFGE The network organization of newly synthesized and deposited collagen fibers was observed by polarized light microscopy on histological sections of wound tissue. As reported in Table 111, the moderately organized networks were observed in wounds treated with fibroblast seeded and bFGF treated (10 pg) collagen sponges. DISCUSSION
The purpose of this study is to evaluate the biomechanical response observed when a collagen sponge is placed in a large excisional dermal wound and to investigate the effects of fibroblast seeding and addition of FGE We have previously shown, using this model, that tensile strength of dermal wounds is proportional to the average collagen fiber diameter.43 Results of studies presented in this paper indicate that the presence of a collagen sponge is associated with increased wound breaking loads during the first 15 days post-wounding compared to the control open wound. IncorTABLE I1 Degree of Epithelialization of Excised Dermal Wounds
Open Graft Sponge Fib FGF-4 pg FGF-10 pg FGF-20 pg
Day 15
Day 30
None Complete None None None Incomplete Incomplete
None Complete None Incomplete Incomplete Complete Incomplete
N = 3. None = no epidermal cells present, incomplete = epidermal cells at wound periphery, complete = epidermis intact.
FACILITATION OF DERMAL WOUND HEALING
691
n
E a
Y
FIB
fGF
GRAFT
OPEN SPONGE FIB
FGF
GRAFT
OPEN SQGE (A)
8 n
E
6
3
w
L Q) c Q)
s i
4
2 0
H Eoum
0
twum
(B) Figure 4. Collagen fiber diameters in wound tissue. Collagen fiber diameters in the lower (black box) and upper (hatched box) dermis at 15 (A) and 30 (B) days post-wounding for open wounds (OPEN), wounds treated with collagen sponges (SPONGE),fibroblast seeded collagen sponges (FIB),bFGF (10 &-coated collagen sponges (FGF), and skin grafts (GRAFT). Fiber diameter measurements were made on at least ten observations per histologic section. At least three animals per treatment were studied (N = 30).
poration of fibroblasts or bFGF into the sponge did not increase the breaking load over that observed in the presence of a collagen sponge; however, application of a skin graft increased the breaking load above all other treatments. By 30 days post-wounding the breaking load of the open wound was approximately equal to that observed for wounds subjected to the other treatments.
MARKS, DOILLON, AND SILVER
692
TABLE 111 Collagen Network Organization in Excised Dermal Wounds
Open Graft Sponge Fib FGF-4 PLg FGF-10 Pg FGF-20 pg
Day 15
Day 30
Random Organized Random Slight Slight Slight Slight
Slight Organized Slight Moderate Slight Moderate Slight
N = 3. Random = no organization, slight = regions of parallel aligned fibers, moderate = regions of wavy fibers, organized = organization typical of normal skin.
In comparison, the ultimate tensile strength (UTS) of wounds treated with the collagen sponges containing fibroblasts or bFGF were greater than UTS values observed for open wounds and wounds treated with collagen sponge at day 15 post-wounding. By 30 days postimplantation UTS values for all treatments were similar. A question that arose during the course of these studies concerns the different mechanisms responsible for increased breaking load in the presence of collagen sponge compared to increased tensile strengths observed in the presence of fibroblast-seeded and FGF-coated sponges at 15 days postwounding. Previously we have shown that the presence of type I collagen matrix in an open wound results in increased collagen fiber diameters of the newly deposited Therefore the increased breaking load observed in the presence of the collagen sponge probably reflects the increased collagen fiber diameter observed in the lower dermis (Fig. 4) as well as increased amounts of granulation tissue and increased wound thickness. Increased thickness of wounds treated with collagen sponges would not change the UTS since the breaking load is divided by the cross-sectional area to obtain this number. In comparison, since wounds treated with fibroblast-seeded or FGF-coated collagen sponges did not have increased values of collagen fiber diameters the increased UTS values probably reflect other structural changes. The changes responsible for increased UTS values at day 15 postimplantation appear to be associated with the organization of the collagen network within the wound. It is well known that the organization of collagen within a healing wound changes with increasing time post-wounding. Initially, the collagen is deposited in an amorphous structure. By 28 days post-wounding the collagen fibers are observed as thin wavy structures and by 180 days post-wounding appear to orient parallel to the surface unlike the biaxial orientation seen in normal skin.43Based on these previous observations and examination of the collagen network microstructure under polarized light (Table 111) it appears that the collagen network formed in wounds treated with fibroblast-seeded and bFGF-coated sponges was more mature than the network observed in the presence of the collagen sponge alone. Other experimental studies have shown that the amount of collagen within wound tissue increases with time post-~ounding,4~ suggesting that the increased tensile
FACILITATION OF DERMAL WOUND HEALING
593
strengths of wounds treated with fibroblast-seeded and FGF-coated collagen sponges may reflect a higher collagen concentration within the wound tissue. Increased tensile strength of wounds treated with a fibroblast-seeded collagen sponge is probably a result of the decreased time required for fibroblast migration into the wound area. In addition, the deposition of newly synthesized collagen present within the collagen sponge that occurs during cell cultureI6 can serve as the nucleus for collagen network formation. Increased tensile strength of wounds treated with FGF-coated sponges probably reflects improved angiogenesis within the collagen sponge that is required to stimulate granulation tissue formation and remodeling. It is interesting to note that the tensile strengths of open and all treated wounds approach that of the skin graft by day 30 suggesting that the observed effects of fibroblast seeding, collagen sponge treatment, and bFGF coating from a mechanical point of view occur primarily during the first 15 days post-wounding. Although preliminary data suggest that fibroblast seeding and bFGF coating of collagen sponges may promote reepithelialization (Table 11) further studies are needed to fully evaluate this observation. Results of other studies” suggest that the presence of a fibroblast-seeded collagen sponge stimulates epithelial cell proliferation in cell culture which parallels the results found in this study. In summary, the experimental results suggest that treatment of large excisional dermal wounds with a collagen sponge results in increased breaking loads early in the wound-healing response and is probably associated with the deposition of large-diameter collagen fibers in the lower dermis and increased amounts of granulation tissue. In comparison, wounds treated with collagen sponges containing fibroblasts or FGF exhibit increased tensile strengths and well developed collagen networks earlier than wounds treated with a collagen sponge alone. In the presence of fibroblasts and bFGF reepithelialization appears to be improved. The authors would like to thank Dr. Richard A. Berg for helpful discussions concerning the fibroblast cultures.
References 1. J. Pachence, R. A. Berg, and F. H. Silver, “Collagen: its place in the medical industry,“ Med. Dev. Diagn. lnd., Jan, 49-55 (1987). 2. M. J. Tavis, J.W. Thorton, J. H. Harney, R.T. Danet, E.A. Woodroof, and R. H. Bartlett, “Mechanism of skin graft adherence: Collagen, elastin, and fibrin interactions,“ Surg. Forum, 28, 522-524 (1977). 3. J. L. Hunt, R. Sato, and C. Baxter, “Early tangential excision and immediate mesh autografting of deep dermal hand wounds,” Ann. Surg., 189, 147-1511 (1979). 4. D.M. Morris, G.M. Hall, and E.G. Elias, ”Porcine heterograft dressings for split-thickness graft donor sites,” Surg. Gynecol. Obstet., 149, 893-894 (1979). 5. M.C. Robson, T. H. Krizek, N. Koss, and J. L. Samburg, ‘Amniotic membranes as a temporary wound dressing,” Surg. Gynecol. Obstet., 136,904-906 (1974).
MARKS, DOILLON, AND SILVER
694 6. 7. 8.
9.
10. 11.
12. 13.
14. 15.
16. 17. 18.
19. 20. 21. 22. 23. 24.
P. D. Thomson and D. H. Parks, ”Amnion as a burn dressing,” Burn Wound Coverings, Val. 1, D. L. Wise (ed.), CRC Press, Boca Raton, 1984, pp. 47-53. A.I. Walder, P. D. May, C. P. Bingham, and J. R. Wright, ”Evaluation of synthetic films as wound covers,” Trans. Am. SOC.Artif. Intern. Organs, 15, 29-31 (1969). R.A.B. Wood, R.H.P. Williams, and L.E. Hughes, ”Foam elastomer dressing in the management of open granulating wounds: experience with 250 patients,” Br. J. Surg., 64, 554-557 (1977). M. Chvapil, R. Kronenthal, and W. van Winkle, Jr. ”Medical and surgical application of collagen,” lnt Rev. Connec. Tiss. Res., 6, 1-61 (1973). I.V. Yannas, J. F. Burke, D. P. Orgill, and E.M. Skrabut, “Wound tissue can utilize a polmeric template to synthesize a functional extension of skin,” Science, 215, 174-176 (1982). C. J. Doillon, C.F. Whyne, R.A. Berg, R.M. Olson, and EH. Silver, ”Fibroblast-collagen sponge interactions and the spatial deposition of newly synthesized collagen fibers in vitro and in vivo,” Scanning Electron Microsc., 111, 1313-1320 (1984). C. J. Doillon, C. F. Whyne, S. Brandwein, and E H. Silver, “Collagenbased wound dressings: control of the pore structure and morphology,” J, Biomed. Mater. Res., 20, 1219-1228 (1986). E. Bell, H. P. Ehrlich, S. Sher, C. Merrill, R. Sarber, B. Hull, T. Nakatsuji, D. Church, and D. J. Buttle, ”Development and use of a living skin equivalent,” Plast. Reconstr. Surg., 67, 386- 392 (1981). E. Bell, D. J. Buttle, H.P. Ehrlich, and T. Nakatsuji, “Living tissue formed in vitro and accepted as skin-equivalent tissue of full thickness,” Science, 211,1052-1054 (1981). E. Bell, S. Sher, B. Hull, C. Merrill, S. Rosen, A. Chamson, D. Asselineau, D. Dubortrot, B. Coulomb, C. Lapiere, B. Nusgens, and J. Neveux, ”The reconstitution of living skin,” J. Invest. Dermatol., 81, 2s-10s (1983). C. J. Doillon, F. H. Silver, and R. A. Berg, “Fibroblast cell growth on a porous collagen sponge containing hyaluronic acid and fibronectin,“ Biomaterials, 8, 195-200 (1987). C. Cuono, R. Langdon, and J. McGuire, ”Use of cultured epidermal autografts and dermal allografts as skin replacement after burn injury,” Lancet, 1, 1123-1124 (1986). M. Eisinger, J.-S. Lee, J. M. Hefton, Z. Darzynkiewicz, J.W. Chiao, and E. de Haven, ”Human epidermal cell culture: Growth and differentiation in the absence of dermal components or medium supplements,” Proc. Nut. Acad. Sci. U.S.A., 76, 5340-5344 (1979). M. Eisinger, M. Monden, J. H. Raaf, and J.C. Fortner, “Wound coverage by a sheet of epidermal cells grown from dispersed single cell preparations,“ Surgery, 88, 287-293 (1980). H. Green, 0.Kehinde, and J. Thomas, “Growth of cultured human epidermal cells into multiple epithelia suitable for grafting,” Proc. Nut. Acad. Scr. U.S.A., 76, 5665-5668 (1979). C. J. Doillon, A. J. Wasserman, R. A. Berg, and F. H. Silver, “Behaviour of fibroblasts and epidermal cells cultivated on analogues of extracellular matrix,” Biomaterials, 9, 91-96 (1988). M. B. Sporn and A. 8. Roberts, “Peptide growth factors and inflammation, tissue repair, and cancer,” J. Clin. Invest., 78, 329-332 (1986). J. Folkman and M. Klagsbrun, ’Angiogenic factors,” Science, 235, 442-446 (1987). K. Hanai, Y. Oomura, Y. Kai, K. Nishikawa, N. Shimizu, H. Morita, and R. Plata-Salaman, “Central action of acidic fibroblast growth factor in feeding regulation,” Am. J. Physiol., 256, R217-R223 (1989).
FACILITATION OF DERMAL WOUND HEALING 25. 26. 27.
28.
29.
30.
31. 32. 33. 34. 35. 36.
37. 38. 39.
40.
41.
695
E. O’Keefe, M. L. Chiu, and R. E. Payne, ”Stimulation of growth of keratinocytes by basic fibroblast growth factor,” J. Invest. Dermatol., 90, 767-769 (1988). H.-J. Ristow and T.O. Messmer, ”Basic fibroblast growth factor and insulin-like growth factor I are strong mitogens for cultured mouse keratinocytes,” J. Cell. Physiol., 137, 277-284 (1988). G. D. Shipley, W.W. Keeble, J. E. Hendrickson, R. J. Coffey, Jr., and M. R. Pittlekow, ”Growth of normal human keratinocytes and fibroblasts in serum-free medium is stimulated by acidic and basic fibroblast growth factor,” J. Cell. Physiol., 138, 511-518 (1989). I. Vlodavsky, J. Folkman, R. Sullivan, R. Fridman, R. Ishai-Michaeli, J. Sasse, and M. Klagsburn, ”Endothelial cell-derived basic fibroblast growth factor: Synthesis and deposition into subendothelial extracellular matrix,” Proc. Nat. Acad. Sci. U S A . , 84, 2292-2296 (1987). M. Vigny, M. P. Ollier-Hartmann, M. Lavigne, N. Fayein, J.C. Jaenny, M. Laurent, and Y. Courtois, ”Specific binding of basic fibroblast growth factor to basement membrane-like structures and to purified heparan sulfate proteoglycan of the EHS tumor,” J. Cell. Physiol., 137, 321-328 (1988). D. Ledoux, A. Mereau, M.C. Dauchel, D. Barritault, and J. Couty, ”Distribution of basic fibroblast growth factor binding sites in various tissue membrane preparations from adult guinea pig,” Biochem. Biophys. Res. Commun., 159, 290-296 (1989). J. DiMario, N. Buffinger, S. Yamada, and R.C. Strohman, ”Fibroblast growth factor in the extracellular matrix of dystrophic mouse muscle,” Science, 244, 688-690 (1989). P. R. Murphy, Y. Myai, Y. Sato, R. Sato, M. West, and H.G. Friesen, ”Elevated expression of basic fibroblast growth factor messenger ribonulcleic acid in acoustic neuromas,” Mol. Endocrind., 3, 225-231 (1989). D. Fredj-Reygrobellet, J. Plouet, Ch. Baudouin, F. Bourret, and Ph. Lapalus, ”Effects of aFGF and bFGF on wound healing in rabbit corneas,” Curr. €ye Res., 6, 1205-1209 (1987). R. Montesano, 1.-D. Vassalli, A. Baird, R. Guillemin, and L. Orci, “Basic fibroblast growth factor induces angiogenesis in vitro,” Proc. Nat. Acad. Sci. LISA., 83, 7297-7301 (1986). J. M. Herbert, M.C. Laplace, and J. P. Maffrand, ”Effect of heparin on the angiogenic potency of basic and acidic fibroblast growth factors in the rabbit cornea assay,” lrrt. J. Tiss. Reac., X, 133-139 (1988). J. Sudhalter, J. Folkman, C. M. Svahn, K. Bergendal, and P. A. D’Amore, “Importance of size, sulfation, and anticoagulant activity in the potentiation of acidic fibroblast growth factor by heparin,” 1. Biol. Chem., 264, 6892-6897 (1989). J.A. Thompson, K.D. Anderson, J.M. DiPietro, ].A. Zwiebel, M. Zamette, W. F. Anderson, and T. Maciag, ”Site directed neovessel formation in vivo,” Science, 241, 1349-1352 (1988). S. N. Mueller, K. A. Thomas, J. DiSalvo, and E. M. Levine, ”Stabilization by heparin of acidic fibroblast growth factor,” J. Cell. Physiol., 140, 439-448 (1989). A. J. Wasserman, C. J. Doillon, A.I. Glasgold, Y.P. Kato, D. Christiansen, A. Rizvi, E. Wong, J. Goldstein, and F. H. Silver, “Clinical applications of electron microscopy in the analysis of collagenous biomaterials,” Scanning Microsc., 2, 1635-1646 (1988). K. R. Meade and E H. Silver, “Immunogenicity of collagenous implants,’’ Biomaterials, 11, 176-380 (1990). Y.P. Kato, D.L. Christiansen, R.A. Hahn, S.-J. Shieh, and F.H. Silver, “Mechanical properties of collagen fibres: a comparison of reconstituted and rat tail tendon fibres,” Biomaterials, 10, 38-42 (1989).
MARKS, DOILLON,. AND SILVER
696
K. Weadock, R.M. Olson, and EH. Silver, "Evaluation of collagen crosslinking techniques," Biomater. Med. Dev. Artif. Organs, 11, 293-318 (1983-1984). 43. C. J. Doillon, M.C. Dunn, and F. H. Silver, "Relationship between mechanical properties and collagen structure of closed and open wounds," J. Biomech. Eng., 110, 352-356 (1988). 42.
Received April 12, 1990 Accepted December 26, 1990
increased wound tensile strength and increased collagen fiber diameters in the upper dermis 15 days post-wounding in an excisional guinea pig dermal wound model. In comparison, dermal wounds treated with collagen sponges seeded with fibroblasts or coated with bFGF showed increased tensile strengths 15 days postimplantation and increased degree of reepithelialization. These results indicate that fibroblast seeding and bFGF coating in conjunction with a type I collagen sponge matrix facilitate early dermal and epidermal wound healing.
IN TRODUCTION
Replacement of skin damaged as a result of thermal and mechanical trauma or contact with a pathogenic agent has been the subject of intense research interest especially in the hope of development of an “artificial skin.” The consequence of this research has been the development of (a) improved skin grafting technique~l-~ using both human and animal skins: (b) use of biological membranes such as amnion as a burn (c) development of wound dressings composed of and bi~logical’~~-’~ polymers, and (d) development of fibroblast13-16 and epithelial ~ell’~-’l seeding techniques. A variety of growth factors including epidermal growth factor, fibroblast growth factors (acidic and basic), platelet-derived growth factor, transforming growth factors (alpha and beta) and insulin-like growth factor influence the behavior of cells found in the extracellular matrix” and are likely to be ultimately useful as components of “artificial skin.” *Current address: Lab. of Experimental Surgery, Pavillon Des Services, Lava1 University, Cite Universitaire, Quebec, GlK7P4, Canada. tTo whom correspondence should be addressed. Journal of Biomedical Materials Research, Vol. 25, 683-696 (1991) CCC 0021-9304/91/050683-14$4.00 0 1991 John Wiley & Sons, Inc.
MARKS, DOILLON, A N D SILVER
684
Basic fibroblast growth factor (bFGF) stimulates angiogene~is?~,~~ proliferation of k e r a t i n ~ c y t e and s ~ ~is~ associated ~ with the extracellular matrix.28Iodinated bFGF binds to heparan sulfate proteoglycan but not to laminin or type IV collagen and binding is inhibited by hepa~in.2~ Specific interaction sites for bFGF are widely distributed in mammalian tissues.30bFGF levels are elevated in tissue that displays persistent regeneration3' and in Schwann ~ m a . ~Exogenous ' addition of FGF promotes wound healing in a rabbit cornea Acidic fibroblast growth factor (aFGF) in the presence of heparin stimulates a n g i ~ g e n e s i sand ~ ~ -binds ~ ~ to type I and IV c011agens.~~ It has been reported that the mitogenic activity of bovine aFGF is stabilized by preincubation with culture medium containing heparin.38 The objective of this study is to compare the wound healing response of a type I collagen matrix that is coated with bFGF with that observed in the presence of a collagen matrix seeded with fibroblasts. Results presented in this study indicate that type I collagen matrices that are treated once with bFGF before implantation or seeded with fibroblasts in cell culture faciIitate the development of wound tensile strength and reepithelialization when implanted on full-thickness guinea pig excised dermal wounds.
MATERIALS A N D METHODS
Type I collagen preparation Insoluble collagen, obtained from Devro, Inc. (Somerville, NJ) was prepared as described p r e v i o u ~ l y and ~ ~ -was ~ ~ characterized as being typical of type I by sodium dodecyl sulfate polyacrylamide gel electrophoresis and amino acid analysis. Collagen sponge preparation For each collagen sponge preparation, 30 mL of a collagen dispersion was poured into a 100 x 15 mm petri dish (Falcon, model 1029) as described previously.'6*21 The dispersion was spread evenly across the bottom of the dish, and any remaining air bubbles and large pieces of collagen were removed. The petri dishes were covered with lids and placed in a freezer for at least 3 h. The materials were then lyophilized at 0°C for 24 h. After removal from the petri dishes, the sponges were crosslinked using a two-step process involving dehydrothermal treatment followed by exposure to cyanamide. Collagen sponge crosslinking Sponges crosslinked by severe dehydration at 110°C for 3 days and then placed in a bath containing 1 L of 1% w/v aqueous cyanamide for 24 h at
FACILITATION OF DERMAL WOUND HEALING
685
pH 7.2 as described previously.16,21 They were then lyophilized, coated with a layer of silicone rubber, and then sterilized by exposure to a dose of 2.5 Mrad of gamma radiation. Basic FGF was obtained from Amgen, Inc. (Tall Oaks, CA). Dilutions were made to yield 10, 25, and 50-puglmL solutions in a 400-pL volume. Using an Eppendorf pipette with sterile tips, 400-pL was dispersed across the middle third region of a sterilized collagen sponge, an area of about 15 cm2(on the side that was not coated with silicone), under a laminar air flow bench. The bFGF solution was allowed to air dry prior to implantation under a laminar air flow hood. Isolation of fibroblasts Fibroblasts were isolated from autologous guinea pig skin and then placed in a petri dish containing sterile phosphate buffer solution (PBS). Under a laminar-flow hood the skin was scraped on both sides and then minced. It was then soaked in a sterile solution containing collagenase (440 units/mg) and Dulbecco’s Modified Eagle’s Medium (DMEM) and incubated for 24 h at 37°C in 5% C02. Digested pieces of skin were separated and then put into a cell culture flask. The cells were given fresh medium, and allowed to proliferate until they were confluent as described previously.” Cell culture techniques Cells were fed every other day with DMEM containing 10%calf serum and antibiotics. When confluent, cells were split and seeded in new flasks. Cells were passaged for several generations before seeding onto collagen sponges. Sterile collagen sponges, about 50 cm2in surface area, were saturated with DMEM containing 10% calf serum. 1 x lo5 cells were dispersed throughout the sponge and allowed to proliferate for 1 week prior to implantation as described previously.’6 Animal surgery
White female Hartley Albino guinea pigs weighing 400-500 g were obtained from Charles River laboratories (Wilmington, MA). Animals were acclimated to vivarium housing for 1 week prior to surgery and were fed and given water supplemented with vitamin C ad Eibitum. Guinea pigs were anesthetized with intraperitoneal injections of ketamine (100 mg/mL, dose: 35 mg/kg) and xylazine (1 mg/mL, dose: 5 mg/kg) and the backs were then shaved, depilated, and cleansed with a povidine-iodine antiseptic solution. Using sterile procedure, an elliptical full-thickness wound, 10 cm x 3 cm, down to and including the panniculus carnosus, was excised transverse to the animal’s spine.
686
MARKS, DOILLON, AND SILVER
Skin autografts were used as a positive control (Graft) and open wounds (Open) were used as a negative control. Experimental treatment groups included a collagen sponge (Sponge), a collagen sponge with fibroblasts (FIB), or a collagen sponge containing bFGF (FGF). In each case, a sponge coated with silicone adhesive was cut to the size of the wound and clipped to the wound periphery using 11-mm wound clips. During implantation, wounds were inspected and any infected wounds were eliminated from the study. Animals were sacrificed at 15 and 30 days postimplantation by an overdose of sodium pentobarbitol injected intraperitoneally. ,4minimum of three animals per treatment group were used; an average of three skin strips per animal were available for mechanical testing at sacrifice. After sacrifice, the granulation tissue and unwounded tissue on both edges of the wound were cut into parallel strips about 5 mm wide and excised on a tissue plane just above the fascia using a razor blade as previously deSkin strips were then placed in PBS and immediately tested in uniaxial tension on an Instron testing machine (Instron Corporation, model 1122). In addition, a small piece of wound tissue was excised approximately 5 mm from the right wound edge and placed in Carson’s fixative. Tissue specimens were dehydrated using routine methods, paraffin embedded, and stained with hematoxylin and eosin (H&E),and picro-Sirius red for light microscopic analysis.
Histological analysis Tissue sections cut perpendicular to the epithelial surface were stained with H&E and picro-Sirius red. H&E stained sections were examined for the amount of sponge remaining in the granulation tissue, degree of reepithelialization, and arrangement of newly deposited collagen. Picro-Sirius red stained sections were examined for collagen fiber thickness as a function of depth. A Leitz light microscope (Laborlux 12 Pol) was used for histological evaluation.
Biomechanical testing Skin strips were tested on an Instron tester in uniaxial tension. The sample was stretched between the Instron grips until it was slightly taut and positioned with the wound tissue centered. The cross-sectional area of the sample was measured by assuming the sample was a rectangle and measuring the width and thickness of the sample using calipers. The gage length (starting sample length) was set to 10 mm and specimen width and thickness used were 3 cm and 0.5 cm, respectively. The crosshead speed was set to 1 mm/min (10% strain rate), and the chart speed was set to 20 mm/min. Breaking loads were measured directly from the force-displacement graph. Breaking energy was calculated by measuring the area under the forcedisplacement curve using a Hipad digitizer (Houston Instruments) interfaced
FACILITATION OF DERMAL WOUND HEALING
687
with a computer (IBM-PC). Ultimate tensile strength was determined by dividing breaking load by the cross sectional area. Values were reported as means 2 standard error and compared statistically using a one-tailed, unpaired Student’s t test based on at least nine observations per treatment group. A p value less than 0.05 was considered significant. Biomechanical tests were discarded if failure was observed at the grip surface or slippage was observed at the grip-tissue interface. Test results were also discarded if the sample was not uniform or appeared damaged.
RESULTS
Biomechanical measurements The load required to cause failure of dermal wound tissue (Fig. 1) was found at 15 days postimplantation to be lowest for the control open wound (451 ? 80 g), higher for the collagen sponge, bFGF-treated, and fibroblasttreated wounds and the breaking load was highest for the skin graft (2128 k 636 g). Breaking loads for the collagen sponge were significantly greater at 15 days postimplantation compared to the open wounds ( p < 0.05) and those for the skin graft were significantly greater than fibroblast and bFGF-treated wounds. At 30 days postimplantation the breaking load was
3000
T
OPEN s#TJGE 15days
FIB
FFF
GRAFT
30 days
Figure 1. Breaking load of open wounds. Relationship between dermal wound breaking load and treatment for open wounds (OPEN), wounds treated with collagen sponges (SPONGE), fibroblast-seeded collagen sponges (FIB), basic fibroblast growth factor (10 p g ) coated-collagen sponges (bFGF), and skin grafts (GRAFT)at 15 (black box) and 30 (hatched box) days postimplantation. Each data point represents the average of at least nine separate measurements (N = 9).
MARKS, DOILLON, AND SILVER
638
statistically greater for the collagen sponge treated wounds but not for any of the other treatments (Fig. 1)compared to the control open wounds (OPEN). When the breaking load was divided by the initial cross-sectional area to determine the engineering breaking stress (ultimate tensile strength) at 15 days postimplantation the values (Fig. 2) obtained were 0.19 2 0.04 MPa (open wounds), 0.36 +- 0.07 MPa (collagen sponge), 0.73 2 0.1 MPa (fibrobiast and bFGF-treated wounds), and 0.56 2 0.14 MPa (graft). Statistically, wounds treated with bFGF and fibroblasts have higher tensile strengths than wounds treated with collagen sponge or open wounds ( p < 0.05). At 30 days postimplantation ultimate tensile strength of open, sponge-treated, f ibroblast-treated, and bFGF-treated wounds were all about 0.4 MPa. Wounds treated with skin grafts had slightly higher ultimate strengths (Fig. 2); however, they were not significantly different from the values found for the other treatment groups. Breaking energies at 15 days postimplantation (Fig. 3) were computed from the areas under the stress-strain curves and were similar for collagensponge-treated, fibroblast-treated, and bFGF-treated wounds. For skin-graf ttreated wounds the breaking energy was significantly higher than for the other groups. At 30 days postimplantation the breaking energy for sponge treated wounds was higher than that found for the other groups; however, no significant difference existed between breaking energy at 15 and 30 days Fostimplantation for the collagen-sponge-treated group.
T
-8
0.6
c a
0.4
z cn
T
0.2
0.0 OPEN SFCNGE 15days
FIB
FGF
G
W
30days
Figure 2. Ultimate tensile strength of open wounds. Ultimate tensile strength (UTS) as a function of treatment for open wounds (OPEN), wounds treated with collagen sponges (SPONGE), fibroblast-seeded collagen sponges (FIB), bFGF (10 &-coated collagen sponges (FGF), and skin grafts (GRAFT) at 15 (black box) and 30 days (hatched box) postwounding. Each data point represents the average of at least nine separate measurements (N = 9).
FACILITATION OF DERMAL WOUND HEALING
689
800 r
r I5
600
m
400
e
200
c 3 6
a
0
Figure 3. Breaking energy of wound tissue. Relationship between breaking energy (cm’) and treatment for open wounds (OPEN), wounds treated with collagen sponges (SPONGE), fibroblast seeded collagen sponges (FIB), bFGF (10 pg)-coated collagen sponges (FGF), and skin grafts (GRAFT) at 15 (black box) and 30 (hatched box) days postwounding. Each data point represents the average of at least nine separate measurements (N = 9).
Histological analyses The extent of collagen sponge biodegradation was determined based on the staining intensity differences between collagen fibers in the sponge and newly synthesized collagen deposited in dermal tissue by light microscopic analysis (Table I). In 1 out 3 wounds treated with a collagen sponge, fragments were observed at days 15 and 30 postimplantation. Fibroblast-seeded sponges had fragments remaining in 2 out of 3 samples at both 15 and 30 days postimplantation. In wounds treated with bFGF-coated sponges, collagen fragments were observed only at day 15 postimplantation. TABLE I Amount of Collagen Sponge Remaining in Granulation Tissue
Open Graft Sponge Fib FGF4 pg FGF-10 p g FGF-25 /*g
Day 15
Day 30
-
-
-
+ ++ 0 + +
+ ++ 0 0 0
N = 3: 0 = 0/3 samples contained sponge fragments, + = 113 samples contained sponge fragments, ++ = 213 samples contained sponge fragments.
690
MARKS, DOILLON, AND SILVER
Complete epithelialization was observed at both days 15 and 30 for skin grafts. Incomplete reepithelialization was observed by day 30 for fibroblast treated wounds and at days 15 and 30 for bFGF treated wounds (Table 11). Collagen fiber diameters were determined in the upper (750 pm) and lower (1500 pm) dermis (Fig. 4). At day 15 the average fiber diameter in the upper dermis was 1.7 pm for all treatments except for the skin graft for which it was significantly higher (4.4 pm). At day 15, the fiber diameters in the lower dermis for collagen-sponge- and fibroblast-treated wounds were similar (Fig. 4). Collagen-sponge-treated wounds had significantly larger diameters when compared to bFGF-treated wounds; however, they were smaller compared to skin-graf t-treated wounds. At day 30, in the upper dermis, fiber diameters were larger for collagensponge-treated wounds than for fibroblast- and bFGF-treated wounds. Skingrafted wounds at day 30 had fiber diameters similar to those seen in collagen-sponge-treated wounds in the upper dermis. In the lower dermis at day 30, collagen-sponge-treatedwounds had significantly larger fiber diameters compared with wounds treated with fibroblasts or bFGE The network organization of newly synthesized and deposited collagen fibers was observed by polarized light microscopy on histological sections of wound tissue. As reported in Table 111, the moderately organized networks were observed in wounds treated with fibroblast seeded and bFGF treated (10 pg) collagen sponges. DISCUSSION
The purpose of this study is to evaluate the biomechanical response observed when a collagen sponge is placed in a large excisional dermal wound and to investigate the effects of fibroblast seeding and addition of FGE We have previously shown, using this model, that tensile strength of dermal wounds is proportional to the average collagen fiber diameter.43 Results of studies presented in this paper indicate that the presence of a collagen sponge is associated with increased wound breaking loads during the first 15 days post-wounding compared to the control open wound. IncorTABLE I1 Degree of Epithelialization of Excised Dermal Wounds
Open Graft Sponge Fib FGF-4 pg FGF-10 pg FGF-20 pg
Day 15
Day 30
None Complete None None None Incomplete Incomplete
None Complete None Incomplete Incomplete Complete Incomplete
N = 3. None = no epidermal cells present, incomplete = epidermal cells at wound periphery, complete = epidermis intact.
FACILITATION OF DERMAL WOUND HEALING
691
n
E a
Y
FIB
fGF
GRAFT
OPEN SPONGE FIB
FGF
GRAFT
OPEN SQGE (A)
8 n
E
6
3
w
L Q) c Q)
s i
4
2 0
H Eoum
0
twum
(B) Figure 4. Collagen fiber diameters in wound tissue. Collagen fiber diameters in the lower (black box) and upper (hatched box) dermis at 15 (A) and 30 (B) days post-wounding for open wounds (OPEN), wounds treated with collagen sponges (SPONGE),fibroblast seeded collagen sponges (FIB),bFGF (10 &-coated collagen sponges (FGF), and skin grafts (GRAFT). Fiber diameter measurements were made on at least ten observations per histologic section. At least three animals per treatment were studied (N = 30).
poration of fibroblasts or bFGF into the sponge did not increase the breaking load over that observed in the presence of a collagen sponge; however, application of a skin graft increased the breaking load above all other treatments. By 30 days post-wounding the breaking load of the open wound was approximately equal to that observed for wounds subjected to the other treatments.
MARKS, DOILLON, AND SILVER
692
TABLE 111 Collagen Network Organization in Excised Dermal Wounds
Open Graft Sponge Fib FGF-4 PLg FGF-10 Pg FGF-20 pg
Day 15
Day 30
Random Organized Random Slight Slight Slight Slight
Slight Organized Slight Moderate Slight Moderate Slight
N = 3. Random = no organization, slight = regions of parallel aligned fibers, moderate = regions of wavy fibers, organized = organization typical of normal skin.
In comparison, the ultimate tensile strength (UTS) of wounds treated with the collagen sponges containing fibroblasts or bFGF were greater than UTS values observed for open wounds and wounds treated with collagen sponge at day 15 post-wounding. By 30 days postimplantation UTS values for all treatments were similar. A question that arose during the course of these studies concerns the different mechanisms responsible for increased breaking load in the presence of collagen sponge compared to increased tensile strengths observed in the presence of fibroblast-seeded and FGF-coated sponges at 15 days postwounding. Previously we have shown that the presence of type I collagen matrix in an open wound results in increased collagen fiber diameters of the newly deposited Therefore the increased breaking load observed in the presence of the collagen sponge probably reflects the increased collagen fiber diameter observed in the lower dermis (Fig. 4) as well as increased amounts of granulation tissue and increased wound thickness. Increased thickness of wounds treated with collagen sponges would not change the UTS since the breaking load is divided by the cross-sectional area to obtain this number. In comparison, since wounds treated with fibroblast-seeded or FGF-coated collagen sponges did not have increased values of collagen fiber diameters the increased UTS values probably reflect other structural changes. The changes responsible for increased UTS values at day 15 postimplantation appear to be associated with the organization of the collagen network within the wound. It is well known that the organization of collagen within a healing wound changes with increasing time post-wounding. Initially, the collagen is deposited in an amorphous structure. By 28 days post-wounding the collagen fibers are observed as thin wavy structures and by 180 days post-wounding appear to orient parallel to the surface unlike the biaxial orientation seen in normal skin.43Based on these previous observations and examination of the collagen network microstructure under polarized light (Table 111) it appears that the collagen network formed in wounds treated with fibroblast-seeded and bFGF-coated sponges was more mature than the network observed in the presence of the collagen sponge alone. Other experimental studies have shown that the amount of collagen within wound tissue increases with time post-~ounding,4~ suggesting that the increased tensile
FACILITATION OF DERMAL WOUND HEALING
593
strengths of wounds treated with fibroblast-seeded and FGF-coated collagen sponges may reflect a higher collagen concentration within the wound tissue. Increased tensile strength of wounds treated with a fibroblast-seeded collagen sponge is probably a result of the decreased time required for fibroblast migration into the wound area. In addition, the deposition of newly synthesized collagen present within the collagen sponge that occurs during cell cultureI6 can serve as the nucleus for collagen network formation. Increased tensile strength of wounds treated with FGF-coated sponges probably reflects improved angiogenesis within the collagen sponge that is required to stimulate granulation tissue formation and remodeling. It is interesting to note that the tensile strengths of open and all treated wounds approach that of the skin graft by day 30 suggesting that the observed effects of fibroblast seeding, collagen sponge treatment, and bFGF coating from a mechanical point of view occur primarily during the first 15 days post-wounding. Although preliminary data suggest that fibroblast seeding and bFGF coating of collagen sponges may promote reepithelialization (Table 11) further studies are needed to fully evaluate this observation. Results of other studies” suggest that the presence of a fibroblast-seeded collagen sponge stimulates epithelial cell proliferation in cell culture which parallels the results found in this study. In summary, the experimental results suggest that treatment of large excisional dermal wounds with a collagen sponge results in increased breaking loads early in the wound-healing response and is probably associated with the deposition of large-diameter collagen fibers in the lower dermis and increased amounts of granulation tissue. In comparison, wounds treated with collagen sponges containing fibroblasts or FGF exhibit increased tensile strengths and well developed collagen networks earlier than wounds treated with a collagen sponge alone. In the presence of fibroblasts and bFGF reepithelialization appears to be improved. The authors would like to thank Dr. Richard A. Berg for helpful discussions concerning the fibroblast cultures.
References 1. J. Pachence, R. A. Berg, and F. H. Silver, “Collagen: its place in the medical industry,“ Med. Dev. Diagn. lnd., Jan, 49-55 (1987). 2. M. J. Tavis, J.W. Thorton, J. H. Harney, R.T. Danet, E.A. Woodroof, and R. H. Bartlett, “Mechanism of skin graft adherence: Collagen, elastin, and fibrin interactions,“ Surg. Forum, 28, 522-524 (1977). 3. J. L. Hunt, R. Sato, and C. Baxter, “Early tangential excision and immediate mesh autografting of deep dermal hand wounds,” Ann. Surg., 189, 147-1511 (1979). 4. D.M. Morris, G.M. Hall, and E.G. Elias, ”Porcine heterograft dressings for split-thickness graft donor sites,” Surg. Gynecol. Obstet., 149, 893-894 (1979). 5. M.C. Robson, T. H. Krizek, N. Koss, and J. L. Samburg, ‘Amniotic membranes as a temporary wound dressing,” Surg. Gynecol. Obstet., 136,904-906 (1974).
MARKS, DOILLON, AND SILVER
694 6. 7. 8.
9.
10. 11.
12. 13.
14. 15.
16. 17. 18.
19. 20. 21. 22. 23. 24.
P. D. Thomson and D. H. Parks, ”Amnion as a burn dressing,” Burn Wound Coverings, Val. 1, D. L. Wise (ed.), CRC Press, Boca Raton, 1984, pp. 47-53. A.I. Walder, P. D. May, C. P. Bingham, and J. R. Wright, ”Evaluation of synthetic films as wound covers,” Trans. Am. SOC.Artif. Intern. Organs, 15, 29-31 (1969). R.A.B. Wood, R.H.P. Williams, and L.E. Hughes, ”Foam elastomer dressing in the management of open granulating wounds: experience with 250 patients,” Br. J. Surg., 64, 554-557 (1977). M. Chvapil, R. Kronenthal, and W. van Winkle, Jr. ”Medical and surgical application of collagen,” lnt Rev. Connec. Tiss. Res., 6, 1-61 (1973). I.V. Yannas, J. F. Burke, D. P. Orgill, and E.M. Skrabut, “Wound tissue can utilize a polmeric template to synthesize a functional extension of skin,” Science, 215, 174-176 (1982). C. J. Doillon, C.F. Whyne, R.A. Berg, R.M. Olson, and EH. Silver, ”Fibroblast-collagen sponge interactions and the spatial deposition of newly synthesized collagen fibers in vitro and in vivo,” Scanning Electron Microsc., 111, 1313-1320 (1984). C. J. Doillon, C. F. Whyne, S. Brandwein, and E H. Silver, “Collagenbased wound dressings: control of the pore structure and morphology,” J, Biomed. Mater. Res., 20, 1219-1228 (1986). E. Bell, H. P. Ehrlich, S. Sher, C. Merrill, R. Sarber, B. Hull, T. Nakatsuji, D. Church, and D. J. Buttle, ”Development and use of a living skin equivalent,” Plast. Reconstr. Surg., 67, 386- 392 (1981). E. Bell, D. J. Buttle, H.P. Ehrlich, and T. Nakatsuji, “Living tissue formed in vitro and accepted as skin-equivalent tissue of full thickness,” Science, 211,1052-1054 (1981). E. Bell, S. Sher, B. Hull, C. Merrill, S. Rosen, A. Chamson, D. Asselineau, D. Dubortrot, B. Coulomb, C. Lapiere, B. Nusgens, and J. Neveux, ”The reconstitution of living skin,” J. Invest. Dermatol., 81, 2s-10s (1983). C. J. Doillon, F. H. Silver, and R. A. Berg, “Fibroblast cell growth on a porous collagen sponge containing hyaluronic acid and fibronectin,“ Biomaterials, 8, 195-200 (1987). C. Cuono, R. Langdon, and J. McGuire, ”Use of cultured epidermal autografts and dermal allografts as skin replacement after burn injury,” Lancet, 1, 1123-1124 (1986). M. Eisinger, J.-S. Lee, J. M. Hefton, Z. Darzynkiewicz, J.W. Chiao, and E. de Haven, ”Human epidermal cell culture: Growth and differentiation in the absence of dermal components or medium supplements,” Proc. Nut. Acad. Sci. U.S.A., 76, 5340-5344 (1979). M. Eisinger, M. Monden, J. H. Raaf, and J.C. Fortner, “Wound coverage by a sheet of epidermal cells grown from dispersed single cell preparations,“ Surgery, 88, 287-293 (1980). H. Green, 0.Kehinde, and J. Thomas, “Growth of cultured human epidermal cells into multiple epithelia suitable for grafting,” Proc. Nut. Acad. Scr. U.S.A., 76, 5665-5668 (1979). C. J. Doillon, A. J. Wasserman, R. A. Berg, and F. H. Silver, “Behaviour of fibroblasts and epidermal cells cultivated on analogues of extracellular matrix,” Biomaterials, 9, 91-96 (1988). M. B. Sporn and A. 8. Roberts, “Peptide growth factors and inflammation, tissue repair, and cancer,” J. Clin. Invest., 78, 329-332 (1986). J. Folkman and M. Klagsbrun, ’Angiogenic factors,” Science, 235, 442-446 (1987). K. Hanai, Y. Oomura, Y. Kai, K. Nishikawa, N. Shimizu, H. Morita, and R. Plata-Salaman, “Central action of acidic fibroblast growth factor in feeding regulation,” Am. J. Physiol., 256, R217-R223 (1989).
FACILITATION OF DERMAL WOUND HEALING 25. 26. 27.
28.
29.
30.
31. 32. 33. 34. 35. 36.
37. 38. 39.
40.
41.
695
E. O’Keefe, M. L. Chiu, and R. E. Payne, ”Stimulation of growth of keratinocytes by basic fibroblast growth factor,” J. Invest. Dermatol., 90, 767-769 (1988). H.-J. Ristow and T.O. Messmer, ”Basic fibroblast growth factor and insulin-like growth factor I are strong mitogens for cultured mouse keratinocytes,” J. Cell. Physiol., 137, 277-284 (1988). G. D. Shipley, W.W. Keeble, J. E. Hendrickson, R. J. Coffey, Jr., and M. R. Pittlekow, ”Growth of normal human keratinocytes and fibroblasts in serum-free medium is stimulated by acidic and basic fibroblast growth factor,” J. Cell. Physiol., 138, 511-518 (1989). I. Vlodavsky, J. Folkman, R. Sullivan, R. Fridman, R. Ishai-Michaeli, J. Sasse, and M. Klagsburn, ”Endothelial cell-derived basic fibroblast growth factor: Synthesis and deposition into subendothelial extracellular matrix,” Proc. Nat. Acad. Sci. U S A . , 84, 2292-2296 (1987). M. Vigny, M. P. Ollier-Hartmann, M. Lavigne, N. Fayein, J.C. Jaenny, M. Laurent, and Y. Courtois, ”Specific binding of basic fibroblast growth factor to basement membrane-like structures and to purified heparan sulfate proteoglycan of the EHS tumor,” J. Cell. Physiol., 137, 321-328 (1988). D. Ledoux, A. Mereau, M.C. Dauchel, D. Barritault, and J. Couty, ”Distribution of basic fibroblast growth factor binding sites in various tissue membrane preparations from adult guinea pig,” Biochem. Biophys. Res. Commun., 159, 290-296 (1989). J. DiMario, N. Buffinger, S. Yamada, and R.C. Strohman, ”Fibroblast growth factor in the extracellular matrix of dystrophic mouse muscle,” Science, 244, 688-690 (1989). P. R. Murphy, Y. Myai, Y. Sato, R. Sato, M. West, and H.G. Friesen, ”Elevated expression of basic fibroblast growth factor messenger ribonulcleic acid in acoustic neuromas,” Mol. Endocrind., 3, 225-231 (1989). D. Fredj-Reygrobellet, J. Plouet, Ch. Baudouin, F. Bourret, and Ph. Lapalus, ”Effects of aFGF and bFGF on wound healing in rabbit corneas,” Curr. €ye Res., 6, 1205-1209 (1987). R. Montesano, 1.-D. Vassalli, A. Baird, R. Guillemin, and L. Orci, “Basic fibroblast growth factor induces angiogenesis in vitro,” Proc. Nat. Acad. Sci. LISA., 83, 7297-7301 (1986). J. M. Herbert, M.C. Laplace, and J. P. Maffrand, ”Effect of heparin on the angiogenic potency of basic and acidic fibroblast growth factors in the rabbit cornea assay,” lrrt. J. Tiss. Reac., X, 133-139 (1988). J. Sudhalter, J. Folkman, C. M. Svahn, K. Bergendal, and P. A. D’Amore, “Importance of size, sulfation, and anticoagulant activity in the potentiation of acidic fibroblast growth factor by heparin,” 1. Biol. Chem., 264, 6892-6897 (1989). J.A. Thompson, K.D. Anderson, J.M. DiPietro, ].A. Zwiebel, M. Zamette, W. F. Anderson, and T. Maciag, ”Site directed neovessel formation in vivo,” Science, 241, 1349-1352 (1988). S. N. Mueller, K. A. Thomas, J. DiSalvo, and E. M. Levine, ”Stabilization by heparin of acidic fibroblast growth factor,” J. Cell. Physiol., 140, 439-448 (1989). A. J. Wasserman, C. J. Doillon, A.I. Glasgold, Y.P. Kato, D. Christiansen, A. Rizvi, E. Wong, J. Goldstein, and F. H. Silver, “Clinical applications of electron microscopy in the analysis of collagenous biomaterials,” Scanning Microsc., 2, 1635-1646 (1988). K. R. Meade and E H. Silver, “Immunogenicity of collagenous implants,’’ Biomaterials, 11, 176-380 (1990). Y.P. Kato, D.L. Christiansen, R.A. Hahn, S.-J. Shieh, and F.H. Silver, “Mechanical properties of collagen fibres: a comparison of reconstituted and rat tail tendon fibres,” Biomaterials, 10, 38-42 (1989).
MARKS, DOILLON,. AND SILVER
696
K. Weadock, R.M. Olson, and EH. Silver, "Evaluation of collagen crosslinking techniques," Biomater. Med. Dev. Artif. Organs, 11, 293-318 (1983-1984). 43. C. J. Doillon, M.C. Dunn, and F. H. Silver, "Relationship between mechanical properties and collagen structure of closed and open wounds," J. Biomech. Eng., 110, 352-356 (1988). 42.
Received April 12, 1990 Accepted December 26, 1990
