British Journai of Dermatology (1992) 127. 365-371,
Peripheral anchorage of dermal equivalents C . A . L O P E Z VALLE.*t F.A.AUGER.*$ P.ROMPRE.*§ VERONIQUH BOUVARD* AND LUCIE GERMAIN*i: *Skin Culture iMboratory. Saint-Sacrement Hospital, Quebec. Quebec. Canada. CIS 4L8 •f Department of Microbioioffij. Lava! University. Sainte-Foy. Queiwc, Canada. GIK 7F4 ^Department of Surgery. iMval University, Sainte-i-oy. Quebec, Canada, GIK 7P4 >$ Department of Chemical F.ngineering. Uiva! University. Sainle-Foii, Quebec. Canada. GIK 7P4 Accepted for publication 1 7 April 1992
Summary
Human fibroblasts can induce collagen gel contraction with different kinetics depending on the number ofcells and on the collagen concentration within this lattice, which has been considered as a dermal equivalent. Skin equivalent is a combined culture of dermo-epidermal layers which may be of therapeutic value in the treatment of burn patients. However, the current production ofthe dermal equivalent component gives results that present many drawbacks for their eventual clinical use as a first step in obtaining a skin equivalent. These include: (i Ifinalsurfaces which are very small; less than 20% ofthe initial size Iii) excessive thickness which may hamper successful graft take (iii) fibroblasts that do not have an arrangement comparable with normal dermal tissue. We propose, as a solution to these problems, the utilization of a 5-mm-wide fibre-glass filter ring peripherally attached to the surface ofthe Petri dishes to prevent inordinate contraction while the fibroblasts reorganize the collagen gel. Using this technique the initial surface was preserved and the dermal equivalent contracted only in thickness. Histological analysis of these anchored equivalents confirmed an alignment of fibroblasts and collagen fibres resembling normal derma! tissue. We consider this method useful in the development of dermo-epidermal sheets for clinical purposes.
The various techniques developed for the culture of epidermal sheets have allowed an important breakthrough in the treatment of large burn wounds.' ' However these grafts, although they provide adequate and permanent coverage, are incomplete, because they lack the dermal component of skin. 'I'he absence of dermis may be an important drawback because the dermal layer adds mechanical properties such as constant tension, shock resistance, elasticity and extensibility '^ to the protective barrier quality of the epidermis. Histological analysis of transplanted epidermal grafts has shown that 'neodermis' formation under these grafts is not present during the first year after grafting.'' This delay of at least 1 year in tissue reorganization may be obviated by grafting dermo-epidermal sheets, thus eventually reducing the scar tissue formation stage. The utilization of collagen gel as a matrix for growing cells was first described in 195f>.'' Since then, modifications of this culture technique have been suggested to allow cell growth in a 3-dimensional environment. Correspondenee: Dr F.Auger. Laboratoire de recherche des gr;inds brules. Hopital du Saint-Sucrement. IOSO chemin .Sainte-Foy, Quebec. Canada. t]lS4L«,
which could be experimentally altered to resemble natural and pathological conditions.^ In 1979 a method was reported of using collagen to recreate a dermis-Iike structure in vitro.'' Subsequently, the culture of skin equivalents (SE) has been extensively described.** " The SE consists of two components: the dermal equivalent (DE) which is prepared first, and the epidermal equivalent which is recreated by applying a suspension of keratinocytes on to the DE. The critical step in the SE technique is the DE formation, which is obtained by mixing fibroblasts with a collagen solution. This solution will gel after a few minutes, according to the physicochemical conditions.'-'" The incorporated fibroblasts actively reorganize the collagen fibres and induce a dramatic contraction."'" The final suriace ofthe DE is inversely proportional to the number of cells, and directly proportional to the collagen concentration.' Thus, some authors have proposed that a judicious ratio of cell to collagen concentration in the DE may be selected in order to limit the amount of contraction. However, our own experience has proved otherwise, i.e. that DE contraction is quite severe. Furthermore, the addition of keratinocytes to obtain a SE will always further reduce the final surface in a concentration-
365
C.LOPEZ-VALLE et al
dependent manner.' •*"' ^ All keratinocyte concentrations giving histologically acceptable SF lead to an important additional contraction (final surface < 20% of the initial surface).''* The mechanisms by which the keratinocytes induce such a contraction are not completely understood, although these cells can interact with native collagen fibrils and reorganize the network in the absence of fibroblasts'' independently from serum and fibronectin.'^ Thus the production of a clinically useful hilayered SE presents a formidable challenge. Therefore, we have developed a new technique for obtaining DE with a minimal level of contraction. The addition to the culture dishes of a peripheral ring made of filter material enabled a firm anchorage of these DE. The contraction phenomenon induced by the fibroblasts contained in the collagen gel reduced the thickness but not the DE surface. Furthermore, histological analysis of these anchored dermal equivalents (ADE) demonstrated a cellularfibrillar arrangement much closer to normal skin histhan any previous technique.
Methods Medium preparation
A combination of Dulbecco-Vogt modification of Eagle s medium (DME) with Ham's F12 in a 3:1 proportion (Flow Labs, Mississauga. Ont.. Canada) was used as the base medium. To each litre of medium 24-3 mg of adenine (Sigma Chemical, St Louis. MO. U.S.A.) was added. The medium was sterilized through 0-22-pm filters. Millipak 40 (Millipore. Bedford. MA. U.S.A.) and brought to pH 7 3 5-7-45. Before use this medium was supplemented with 5 /ig/ml bovine crystallized insulin (Sigma). 5 /ig/ml human transferrin (Sigma). 0-4 /ig/ml hydrocortisone (Calbiochem. CA. U.S.A.). 1 0 " ' " M cholera toxin {Schwarz/Mann. Cleveland. OH. U.S.A.). 10% fetal calf serum (Sigma). 10 ng/ml human EGF (Chiron Corp.. Emeryville. CA. U.S.A.). 100 IU/ml penicillin G (Sigma) and 100 /^g/ml streptomycin sulphate (Sigma). This complete medium formulation is based on conditions developed for serial cultivation of keratino-
were inoculated in 75 cm- Falcon tissue culture flasks (Becton Dickinson, Mississauga. Ont.. Canada) and were kept at 37°C in an 8% CO, atmosphere. Cells were resuspended just before confluence with a solution of 0 0 1 % FDTA and ()-()5% trypsin in PBS. A sample was taken to determine the number of cells obtained. The suspension was centrifuged at 300 g for 10 min and a final count performed before bringing the suspension, in DME-HAM, to the desired concentration. Collagen solution
Acid-soluble type I collagen, from calfskin (Sigma type III) was dissolved overnight at 4°C in 1:1000 acetic acid solution to a final concentration of 7 • 11 mg/ml or 4 - 2 2 mg/ml. Preparation of anchoring Petri dishes
Two 5 mm-wide rings were cut from a glass microfibre filter 934-AH (Whatman. Maidstone. U.K.). These were glued together and attached to the bottom of 81-cm^ square bacteriological Petri dishes using epoxy cement. The dishes were then resterilized in ethylene oxide. Control fioating dermal equivalents were prepared in standard bacteriological Petri dishes.' Preparation of dermal equivalents
Fibroblast-populated collagen gels were prepared in 50ml centrifugation tubes (Sarstedt. Ville St-Laurent. QC. Canada) with various collagen and cell ratios. These mixtures were poured into S1 -cm^ LabTek bacteriological dishes, with or without the fibre-glass filter ring. Each dish contained 5-6 ml of x2-7 concentrated DME medium with penicillin and streptomycin. 3 7 ml of fetal calf serum. 9 5 ml of collagen solution. 0-1 5 ml of 0 1 N NaOH. and 1 ml of fibroblasts suspended in x 1 DMEHAM. Dishes were incubated at 37°C in an 8% CO2/92% air atmosphere. Gels were completely set after 10 min. Complete medium was used for media changes. Media changes were initiated on the third day after DE preparation and subsequently performed each third day. Evahiation of dermal equivalent dimensions
Fibroblast cultures
Fibroblast cultures were established from skin obtained at the time of reductive breast surgery. The cells were seeded at 1 x lO*" per flask and propagated in the previously described supplemented medium. These cells
To calculate dermal equivalent dimensions, all Petri dishes were placed on a glass surface 7 cm above an Xray viewbox, A piece of black cardboard was used on the viewbox's screen to obtain indirect transillumination. Square fioating dermal equivalents (FDE) showed slight
PERIPHERAL ANCHORAGE OF DERMAL EQUIVALENTS
differences in side length. Photographs of each equivalent were taken daily. The standardized pictures produced were read with planimetric scales to obtain the daily surface of each equivalent. Peripherally anchored dermal equivalent surfaces did not change. All results are given as a percentage of the initial surface. Biopsies of dermal equivalents At least two punch biopsies of 1 cm diam. were taken from different sites on each dermal equivalent. Biopsies were fixed in Bouin's solution and embedded in paraffin in a standard manner, followed by staining with haematoxylin and eosin for cell analysis. Masson's trichrome was employed for collagen pattern visualization.
367
started more or less rapidly, depending on cell number and collagen concentration. The initial contraction speed varied in relation to cell number. Gels with 18,500 fibroblasts/cm- showed a reduction to 4% of the original surface area at 24 h while those containing 245 cells/cm~ showed a mean reduction to 85% of the original surface area during the same period. By the fourth day most of the contraction process had occurred, and the original surface area was reduced to 2 and 45%. respectively, in the previously described conditions. No further contraction was observed after the sixteenth day in any experimental condition, and the mean final surfaces were less than 1 and 3 5% for the above mentioned cell numbers. The results showing the relationship between fioating DE surfaces and various cell concentrations are summarized in Figure 1.
The thickness measurements of all dermal equivalents were performed by optical microscopy of histoiogical slides. A Leitz Wetzlar (Germany) eyepiece with cross reticulation (TO mm/100 divisions) was employed. All dermal equivalents showed thickness variations. Therefore, we performed five measurements on each equivalent and the mean result with the SD is included in Table 1.
As was shown previously.' the optical properties of collagen gels change during contraction. They become progressively more opaque and eventually whitishyellow. Furthermore, the consistency of dermal equivalents varied depending on cell numbers and collagen concentration. Increase of the collagen concentration was accompanied by a reduction in the contraction rate and by an augmentation of the final surface. Figure 2 shows an example of the influence of collagen concentration.
Results
Peripherally anchored dermal equivalent formation
Floating dermal equivalent formation
Changes in cell morphology were simiiar to those observed in floating dermal equivalent formation as observed by phase-contrast microscopy. After gelation. cells were evenly sprinkled throughout the DE. Refringent fibres appeared within 24 h connecting ceils in a circular pattern. Our observations suggest that daughter
Thickness measurements
Under phase-contrast microscopy, round, trypsinized fibroblasts began to show cell processes a few minutes after gel setting. Two hours later the majority of cells became stellate. After a short time-lag, contraction
Table 1. Thickness of relHted FDR and ADE in
FDE ADE FDE ADE
Collagen concentration (mg/ml)
18,000
211
1499 ± 1 6 8
4 22 4 22
Fibroblasts/cm1235
1 J45±9I 30 7 ±58 Nl) NI)
617
555
1169 ±109 i26±65 ND ND
782±I17 286 ±J6 ND ND
"Expressed as mean±SD. ND = not determined. I-'DK were Ihicker than ADI' in related experimental conditions, Riindom analysis showed thai Ihe anchoring method induced formation of thinner dermal equivalents than the floating method. Initial volume was 20 ml and initial surface was 81 cm* for both PDE and ADE,
ibS
C.LOPEZ-VALLE et aL
SOi Fibroblasts numbers -m-•-0-*-*-
Collagen concentration O 4,22 mg/ml • 2,11 mg/ml
70 -
1,5X10^ 5,0X10'' 4.5X10'' 3,6X10'* 2.0X10'*
60 50 -
E
0 a
40 -
,i
^^*—A—
ace
£ 30 . 20 . 10 -
H
10 -
12 Days
16
Days
Figure 1. Influence of fibrohlasts numbers on final surface, 'I'hc initial contraction speed varied in relation to cell numbers. By the fourth day mosi ofthe contraction process had occurred. No further contraction was observed after the sixteenth day in any experimental amdition. Collagen concentration was 2'11 mg/ml.
Figure 2. Influence of collagen concentration on final surface. Increase ofthe collagen concentration was accompanied by a reduction in the contraction rale and by an augmentation of the final surface. Total libroblast number was constant, 4'5 x 10'' (initial concentration of 555 cells/cm-). Bars represent SEM from three replicates.
cells formed bundles in this original arrangement, and this increasing number of fibroblasts began outward migration in all directions (Fig. 3). At the final stage, cells were arranged in a complex three-dimensional circular array, with most cells parallel to the surface ofthe Petri dish. No surface reduction was noted when the fibreglass filter ring was used. Anchored dermal equivalents with a collagen concentration of 2 1 1 mg/mi and containing 18.500 fibroblasts/cm' had a tendency to tear in their central portion. This phenomenon disappeared when the collagen concentration was raised to 4-22 mg/ml for the same cell number. A progressive opacity and solidification ofthe gel were also observed macroscopically. according to culture time.
2. Fibroblasts within the FDE were orientated perpendicular to tissue length, whereas those present at the FDE surface were parallel to tissue length (Fig. 4b). 3. Collagen fibres were orientated perpendicular to the .surface (Fig. 4b). This histological feature was not observed when the fibrobiast numbers were < 617 cells/ cm^. 4. The surface of floating dermal equivalents containing more than 1235 cells/cm^ showed many folds. These folds were filled with colonies of fibroblasts (Fig. 4al.
Floating dermal equivalents (PDF) Histology of fioating dermal equivalents synthesized with different cell numbers and collagen concentrations presented some common characteristics. 1. A fibrobiast multilayer enveloped both surfaces of floating equivalents. This phenomenon did not occur when the cell number was lower than 444 fibroblasts/
Anchored dermal equivalents (ADF.) Histology of peripherally anchored dermal equivalents differed from the floating equivalents. 1. All experimental conditions showed an absence ofa fibrobiast multilayer envelope around the matrix. 2. Fibroblasts were orientated longitudinally to tissue length (Fig. 5b). 3. Collagen fibre orientation was mostly longitudinal in alignment (Fig. 5b). 4. No folds were observed. A microscopic study of the fibre-glass filter ring in the ADE showed penetration of the collagen into the interstices of this material. The anchoring phenomenon seemed to be related to its entrapment in the porous
PERIPHERAL ANCHORAGE OF DERMAL EQUIVALENTS
569
some related experimental conditions are presented in Table 1.
Discussion
Figure 3. Living peripherally anchored dermal equivalent conlaininn I S x l O ' ' fibrohlasts at 48 h in culture (18,000 cells/cm^). Cells arranged in a circular pattern. Daughter cells began migrulion outward in all directions. View from below using phase contrasl microscopy ( x 1 J 5).
Structure, rather than to any cellular binding, as very lew cells were seen within the fibre-glass ring (Fig. 5c). Stability of peripherally anchored equivalents A set of experiments was carried out to evaluate the stability of these ADK after their detachment from the tilter ring. A contraction of approximately 10% occurred within 5 days after detachment in dermal equivalents with 18.500 libroblasts/cnr (Fig. 61. Thickness measurements Fifty-two fioating dermal equivalents were compared with 44 anchored dermal equivalents. Random analysis showed a mean thickness of 673 fim for FDE. with a range of 151-1560 }im. Mean thickness for ADE was 548 ^m with a range of 100-496 ^m. Comparison of
Kigure 4. (a| Histological appearance of full thickness FDE showing an envelope of fibroblasts mainly on the surface, and one Ibid tilled by u colony of fibroblasts (FC). Initial collagen concentration was 2-1 I mg/ ml and fibrobiast number was l8.5()O/cm' (Haematoxylin and eosin. Taken at x 20 and enlarged photographically to x 4().| (b) FnlargemenE of the indicated portion showing fibroblasts (arrowheads) and collagen Iibres orienlated perpendicular to tissue length, (Taken at x 80 and enlarged photographically to x 160,)
This study was initiated to evaluate whether a peripheral anchoring technique might solve the inherent deficiencies in the usual methods of recreating a DE in vitro. The results obtained with the FDE technique illustrate these pitfalls. The FDE technique has invariably produced iinal surfaces that were so small that they precluded their use as clinically significant grafts for burn therapy. Lowering fibrobiast numbers in these collagen lattices lessened this shrinkage phenomenon.' '"'•''' However, the resultant DE was then so poorly organized that it was very fragile. Furthermore, the histological resemblance to normal dermis was lost. One proposed solution to circumvent the shrinkage problem is the use of much larger tissue culture tlasks. There are. however, practical limits to this approach, and the initial surface would have to be extremely large. For example, to obtain a S x 8 cm graft, considering a conservative 9 5% shrinkage level, an initial culture dish of 100 X 1 60 cm would be necessary. These conditions are impractical and preclude the possibility of any scaling up of operations. In addition, if a solution of keratinocytes is then 'plated' on these DE to obtain SE. a further reduction of the final surface will ensue which will greatly compound the problem.'"*''' The resultant FDE are relatively thick, and this may have significant effects on the process of angiogenesis which normally occurs after transplantation. As showed by Yannas. using cell-free polymeric membranes, this ingrowth process could take between 10 and 46 days after application, depending mainly on thickness, chemical composition and type of polymerization.-"' -' Interference with vaseularization lessens the viability of the
370
C.LOPEZ-VALLE et al
Figure 5. (al Histological appearance of full thickness ADE. Comparison with h'igurt 4|a) shows that the thicicncss of the ADI-; is less for the siime conditions (initial cell and collagen concentrations). Initial collagen concentration was 2-\\ mg/ml and fibroblast number was 18,5()()/cm-. (Haematoxylin and eosin. Taken at x 20 and enlarged photographically to x40.) (b) Fnlargement of tbe indicated portion of Figure 5(a) showing fibroblasts (arrowheads) and collagen fibres orientated longitudinally in relation to tissue length, (Taken at x 4U and enlarged photographically to x 160.) (c| Histology of the libre-giass ring (F) interface witb the ADE (arrowbead), Haematoxylin and eosin staining. (Taken at X 20 and enlarged photographically to x40,)
(b)
overlying epidermis and results in a lower take level. However, this was not noted in previously described experiments using SE.^'"' but it is possible that the small size of these grafts permitted their lateral nourishment. Lastly, our histological analysis of these FDE has shown that their cellular and fibrillar arrangement, mostly perpendicular to the surface, is not comparable with normal human dermis.''* It seems logical to strive for DE with histological features as close as possible to the normal dermis.
80
24
Figure ft. Comparative final surfaces of floating and anchored dermal equivalents. I'ioating dermal equivalent surface: 0-74%. anchored dermal equivalent surface after freeing: 87-8%, Collagen concentration 2 1 I mg/ml, fibroblasts numbers: 18.S(H)/cm^. Arrow indicates the freeing day. Bars represent SHM from tbree replicates.
Our results show that the anchoring of the dermal equivalent is a practical solution to the drawbacks of the EDE. In their studies of tissue morphogenesis. Harris et al. described a related technique for stabilizing collagen gels devoid of incorporated cells. Eibroblasts were subsequently added to the surface of these collagen gels. The objective of these studies was the analysis of the generation of fibrillar patterns created by various mechanical instabilities.-*^ '" whereas we wished to obtain large and stable DE. We have shown that the peripheral anchoring technique can achieve such results without complicated manipulations of the culture conditions. Anchorage transformed the contraction process into a unidimensional rather than a three-dimensional phenomenon. The only axis of shrinkage involved the thickness of the ADK because it was the only axis along which the fibroblasts could act freely. As previously discussed, we fee! that a thinner DE may facilitate graft take and the survival of the keratinocytes contained in the SE. This cultured SE has to be nourished by osmosis during the first few days after grafting, as are standard grafts." The histological features of the ADE are also much closer to normal dermis.'** The longitudinal arrangement of cells and fibrils in ADE is quite remarkable. Our observations suggest that the attachment of the ADE to the filter ring is a fibrillar entrapment phenomenon rather than a cellular binding mechanism. Thus, it is possible that any porous material will permit efficacious anchoring of dermal equivalents. The choice of an optimal material would be dictated by its biocompatibility. by its ease of use and its resistance to sterilization, allowing repeated utilization.
PERIPHERAL ANCHORAGE OF DERMAL EQUIVALENTS
We helicve that the ADE model presented in this report is a first step in the production of new skin suhstitutes for treatment of burn patients by tissue culture methods. The addition of other molecules in the initial mixture could be helpful in obtaining a more complex and structured dermal equivalent. Such an anchoring technique could be readily amenable to large-scale production of skin equivalent for wound coverage of burns.
Acknowledgments This work was supported by Grant No. MA-1022 i from the Medical Research Council of Canada. Award B-2 committee 01A from the Fonds FCAR (Quebec) to C.A.Lopez Valle and Award PCJS 1 from Natural Sciences and Engineering Research Council of Canada to P.Rompre, and Fellowships from the Fonds de la Recherche en Sante du Quebec (FRSQ) to F.A.Auger and L.Germain, We thank Dr George Nascimento from Chiron Corporation for graciously supplying F^GF, Celine D.Fugere for assistance with histology and jo-Anne Masse for help with technical work.
References 1 lioykin IVJr, Molnar ]A, Burn scar and skin equivalents. In: Wound Healing, Biochemica! i'-^ Clinical Aspects iCohen IK, Diegelmann RF, Lindblad WJ. eds|. Philadelphia; WB Saunders Company. 1992; 525-40, 2 Gallico GG, O'Connor NE. Conipton CC et at. Permanent coverage of large burn wounds with autologous cultured human epithelium. New/^fiif// MeiJ 1984; J11:44X-51, ? S/irmai |A. 'Ihe organization ofthe dermis. In: Advances in Biology of Skin, Vol, X, The Dermis. (Montagna W. Bentley JP, Dobson RL, eds|. New York: Appleton-Century-Crofts. 19f»S; 12, 4 Daly CH, Biotnechanical properties of dermis, / Invest IX'rwatol 1982: 79 (SuppI, 1): 17-20, 5 Compton CC. Gill ]M, Bradford HA et ul Skin regenerated from cultured epitheliiil autografts on full-thickness burn wounds from fidays to 5 years after grafting, !Aib !m'est 1989; 60: M)0-\2, fi Fhrmann RL. (ley GO. The growth of cells on a iransparent gel of reconstituted rat-tail collagen, / Natl Cancer !nsl 1956; 16: 1575403. 7 Bell E, Ivarsson B, Merrill C, Production ofa tissue-like structure by contraction of collagen lattices by human fibroblasts of different proliferative potential in vitro. Proc Natl Acad Sci USA 1979: 76: 1274-8. 8 Bell E. EhHich HP. Buttle DJ. Nakatsuji T, Living tissue formed in vitro and accepted as skin-equivalent tissue of full thickness. Science 1981:211: 1052-4, 9 Bell, E, Ehrlich HP. Sher S ct a!. Development and use ofa living skin equivalent. Wast Reconstr Surg 1981; 67: 586-92.
371
10 Bell E, Sher S. Hull B ci al. The reconstitution of living skin, / !nvest Dermatol 1985: 81 tSuppl,): 2-10, 11 Coulomb B. Saiiig P. Bell F et al. A new method for studying epidermaJization ifi vilro, Br I Dermatol 1986; 114:91-101, 12 Elsdale T. Bard |. Collagen substrata for studies on cell behaviour, / Cell Biol 1972: 54:626-37, 1 i Wood CC. Keofh MK, The formation of fibrils from collagen .solutions. Biochem j 1960: 75: 588-98, 14 Rompre P, Auger FA. Clermain L el al Influence of initial collagen and cellular concentrations on the final surface area of dermal and skin equivalents: a Box-Behnken analysis. In Vitro Cell Dev Bio! 1990: 2 6 : 9 8 3 - 9 0 . 15 Souren |M. Ponec M. van Wijk R, Contraction of collagen by human fibroblasts and keratinocytes. In Vilro Cell Dcv iiio! I 98S: 25: 1039-45. 16 Schafer IA, Shapiro A, Kovach M el al The interaction of human papillary and reticular hbroblasts and human keratinocytes in the contraction of three-dimensional Hoating collagen lattices, Exp CeU Res 1989: 18 3: 112-25. 1 7 l.iliie |H, MacCalhim DK. |epsen A, (irowth ofslralified squamous epithelium on reconstituted extracellular matrices: long-term culture. / Invesl Dermatol 1988: 90: 100-9, 18 Rowling P|F. Raxworthy M|, Wood H[ et al. The infiuence of keratinocyte .seeding and keratinocyte growth medium on dermal equivalent contraction. / Invest Dermato! 1989: 92: 510. 19 lx!ver WF, Schaumburg-Lever G |Eds), Histopathology of the SIcin. 6th edn, Philadelphia: JB Lippincott Company. 1983; 29, 20 Rheinwald |G. Green H. Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinining colonies from single cells, Celt 1975: 6: 331-44. 21 Green H, Kehinde (), Thomiis J, Growth of cultured human epidermal cells into multiple epilhelia suitable for grafting, Proc Nail Acad Sci USA 1979: 76: S66S-8. 22 Banks-Sfhlegel S. Green H, Formation of epidermis by serially cultivated human epidermal celis transplanted as an epithelium to athymic mice. Transplantation 1980: 29: 508-1 I, 25 O'Connor NE, Mulliken |B, Biuiks-SchlegelScIu/.C.rafting of burns with cultured epithelium prepared from autologous epidermal ceils. iMmrt 1981: i: 75-8. 24, Beil [i. Sher S, Hull B, The living skin-equivalent as (i structural and immunological model in skin grafting. Scan E!ectron Microsc 1984: (Pt4|; I9S7-62, 25 Yannas IV, Burke JF, Design ofan artificial skin, I, Basic design principles. / Biomed Mater Res 1980: 14: 6S-81. 26 Yannas IV, Burke JF, Orgill DP, Skrabut EM, Wound tissue can utilize a polymeric template to synthesize a fundioiial extension of skin, SVimc 1982; 215: 174-6, 2 7 Yannas IV, Lee E, Orgill DP et al Synthesis and characterization of a model extracellular matrix that induces partial regeneration of adult mammalian skin, Proc Nat! Acad Sci USA 1989:86: 933-7. 28 Harris AK, Stopak D. Wild P, Fibrobiast traction as a mechanism for collagen morphogenesis. Nature 1981: 290: 249-51, 29 Stopak D. Harris AK. Connective tissue morphogenesis by fibrobiast traction, 1. Tissue culture observations, Dev Biol IS82; 90: JHJ-98, M) [larris AK. Stopak D. Wamer P, Generation of spatially periodic patterns by a mechanical instability: A mechanical alternative to the Turing model, / Emhryot Exp Morphot 1984; 80: 1-20, 3 1 Smith JW, Ashton S| (Fdsl Grabb and Smiths Plastic Surgery. 4th edn. Boston; Little. Brown and Company, 1991: 27,
Peripheral anchorage of dermal equivalents C . A . L O P E Z VALLE.*t F.A.AUGER.*$ P.ROMPRE.*§ VERONIQUH BOUVARD* AND LUCIE GERMAIN*i: *Skin Culture iMboratory. Saint-Sacrement Hospital, Quebec. Quebec. Canada. CIS 4L8 •f Department of Microbioioffij. Lava! University. Sainte-Foy. Queiwc, Canada. GIK 7F4 ^Department of Surgery. iMval University, Sainte-i-oy. Quebec, Canada, GIK 7P4 >$ Department of Chemical F.ngineering. Uiva! University. Sainle-Foii, Quebec. Canada. GIK 7P4 Accepted for publication 1 7 April 1992
Summary
Human fibroblasts can induce collagen gel contraction with different kinetics depending on the number ofcells and on the collagen concentration within this lattice, which has been considered as a dermal equivalent. Skin equivalent is a combined culture of dermo-epidermal layers which may be of therapeutic value in the treatment of burn patients. However, the current production ofthe dermal equivalent component gives results that present many drawbacks for their eventual clinical use as a first step in obtaining a skin equivalent. These include: (i Ifinalsurfaces which are very small; less than 20% ofthe initial size Iii) excessive thickness which may hamper successful graft take (iii) fibroblasts that do not have an arrangement comparable with normal dermal tissue. We propose, as a solution to these problems, the utilization of a 5-mm-wide fibre-glass filter ring peripherally attached to the surface ofthe Petri dishes to prevent inordinate contraction while the fibroblasts reorganize the collagen gel. Using this technique the initial surface was preserved and the dermal equivalent contracted only in thickness. Histological analysis of these anchored equivalents confirmed an alignment of fibroblasts and collagen fibres resembling normal derma! tissue. We consider this method useful in the development of dermo-epidermal sheets for clinical purposes.
The various techniques developed for the culture of epidermal sheets have allowed an important breakthrough in the treatment of large burn wounds.' ' However these grafts, although they provide adequate and permanent coverage, are incomplete, because they lack the dermal component of skin. 'I'he absence of dermis may be an important drawback because the dermal layer adds mechanical properties such as constant tension, shock resistance, elasticity and extensibility '^ to the protective barrier quality of the epidermis. Histological analysis of transplanted epidermal grafts has shown that 'neodermis' formation under these grafts is not present during the first year after grafting.'' This delay of at least 1 year in tissue reorganization may be obviated by grafting dermo-epidermal sheets, thus eventually reducing the scar tissue formation stage. The utilization of collagen gel as a matrix for growing cells was first described in 195f>.'' Since then, modifications of this culture technique have been suggested to allow cell growth in a 3-dimensional environment. Correspondenee: Dr F.Auger. Laboratoire de recherche des gr;inds brules. Hopital du Saint-Sucrement. IOSO chemin .Sainte-Foy, Quebec. Canada. t]lS4L«,
which could be experimentally altered to resemble natural and pathological conditions.^ In 1979 a method was reported of using collagen to recreate a dermis-Iike structure in vitro.'' Subsequently, the culture of skin equivalents (SE) has been extensively described.** " The SE consists of two components: the dermal equivalent (DE) which is prepared first, and the epidermal equivalent which is recreated by applying a suspension of keratinocytes on to the DE. The critical step in the SE technique is the DE formation, which is obtained by mixing fibroblasts with a collagen solution. This solution will gel after a few minutes, according to the physicochemical conditions.'-'" The incorporated fibroblasts actively reorganize the collagen fibres and induce a dramatic contraction."'" The final suriace ofthe DE is inversely proportional to the number of cells, and directly proportional to the collagen concentration.' Thus, some authors have proposed that a judicious ratio of cell to collagen concentration in the DE may be selected in order to limit the amount of contraction. However, our own experience has proved otherwise, i.e. that DE contraction is quite severe. Furthermore, the addition of keratinocytes to obtain a SE will always further reduce the final surface in a concentration-
365
C.LOPEZ-VALLE et al
dependent manner.' •*"' ^ All keratinocyte concentrations giving histologically acceptable SF lead to an important additional contraction (final surface < 20% of the initial surface).''* The mechanisms by which the keratinocytes induce such a contraction are not completely understood, although these cells can interact with native collagen fibrils and reorganize the network in the absence of fibroblasts'' independently from serum and fibronectin.'^ Thus the production of a clinically useful hilayered SE presents a formidable challenge. Therefore, we have developed a new technique for obtaining DE with a minimal level of contraction. The addition to the culture dishes of a peripheral ring made of filter material enabled a firm anchorage of these DE. The contraction phenomenon induced by the fibroblasts contained in the collagen gel reduced the thickness but not the DE surface. Furthermore, histological analysis of these anchored dermal equivalents (ADE) demonstrated a cellularfibrillar arrangement much closer to normal skin histhan any previous technique.
Methods Medium preparation
A combination of Dulbecco-Vogt modification of Eagle s medium (DME) with Ham's F12 in a 3:1 proportion (Flow Labs, Mississauga. Ont.. Canada) was used as the base medium. To each litre of medium 24-3 mg of adenine (Sigma Chemical, St Louis. MO. U.S.A.) was added. The medium was sterilized through 0-22-pm filters. Millipak 40 (Millipore. Bedford. MA. U.S.A.) and brought to pH 7 3 5-7-45. Before use this medium was supplemented with 5 /ig/ml bovine crystallized insulin (Sigma). 5 /ig/ml human transferrin (Sigma). 0-4 /ig/ml hydrocortisone (Calbiochem. CA. U.S.A.). 1 0 " ' " M cholera toxin {Schwarz/Mann. Cleveland. OH. U.S.A.). 10% fetal calf serum (Sigma). 10 ng/ml human EGF (Chiron Corp.. Emeryville. CA. U.S.A.). 100 IU/ml penicillin G (Sigma) and 100 /^g/ml streptomycin sulphate (Sigma). This complete medium formulation is based on conditions developed for serial cultivation of keratino-
were inoculated in 75 cm- Falcon tissue culture flasks (Becton Dickinson, Mississauga. Ont.. Canada) and were kept at 37°C in an 8% CO, atmosphere. Cells were resuspended just before confluence with a solution of 0 0 1 % FDTA and ()-()5% trypsin in PBS. A sample was taken to determine the number of cells obtained. The suspension was centrifuged at 300 g for 10 min and a final count performed before bringing the suspension, in DME-HAM, to the desired concentration. Collagen solution
Acid-soluble type I collagen, from calfskin (Sigma type III) was dissolved overnight at 4°C in 1:1000 acetic acid solution to a final concentration of 7 • 11 mg/ml or 4 - 2 2 mg/ml. Preparation of anchoring Petri dishes
Two 5 mm-wide rings were cut from a glass microfibre filter 934-AH (Whatman. Maidstone. U.K.). These were glued together and attached to the bottom of 81-cm^ square bacteriological Petri dishes using epoxy cement. The dishes were then resterilized in ethylene oxide. Control fioating dermal equivalents were prepared in standard bacteriological Petri dishes.' Preparation of dermal equivalents
Fibroblast-populated collagen gels were prepared in 50ml centrifugation tubes (Sarstedt. Ville St-Laurent. QC. Canada) with various collagen and cell ratios. These mixtures were poured into S1 -cm^ LabTek bacteriological dishes, with or without the fibre-glass filter ring. Each dish contained 5-6 ml of x2-7 concentrated DME medium with penicillin and streptomycin. 3 7 ml of fetal calf serum. 9 5 ml of collagen solution. 0-1 5 ml of 0 1 N NaOH. and 1 ml of fibroblasts suspended in x 1 DMEHAM. Dishes were incubated at 37°C in an 8% CO2/92% air atmosphere. Gels were completely set after 10 min. Complete medium was used for media changes. Media changes were initiated on the third day after DE preparation and subsequently performed each third day. Evahiation of dermal equivalent dimensions
Fibroblast cultures
Fibroblast cultures were established from skin obtained at the time of reductive breast surgery. The cells were seeded at 1 x lO*" per flask and propagated in the previously described supplemented medium. These cells
To calculate dermal equivalent dimensions, all Petri dishes were placed on a glass surface 7 cm above an Xray viewbox, A piece of black cardboard was used on the viewbox's screen to obtain indirect transillumination. Square fioating dermal equivalents (FDE) showed slight
PERIPHERAL ANCHORAGE OF DERMAL EQUIVALENTS
differences in side length. Photographs of each equivalent were taken daily. The standardized pictures produced were read with planimetric scales to obtain the daily surface of each equivalent. Peripherally anchored dermal equivalent surfaces did not change. All results are given as a percentage of the initial surface. Biopsies of dermal equivalents At least two punch biopsies of 1 cm diam. were taken from different sites on each dermal equivalent. Biopsies were fixed in Bouin's solution and embedded in paraffin in a standard manner, followed by staining with haematoxylin and eosin for cell analysis. Masson's trichrome was employed for collagen pattern visualization.
367
started more or less rapidly, depending on cell number and collagen concentration. The initial contraction speed varied in relation to cell number. Gels with 18,500 fibroblasts/cm- showed a reduction to 4% of the original surface area at 24 h while those containing 245 cells/cm~ showed a mean reduction to 85% of the original surface area during the same period. By the fourth day most of the contraction process had occurred, and the original surface area was reduced to 2 and 45%. respectively, in the previously described conditions. No further contraction was observed after the sixteenth day in any experimental condition, and the mean final surfaces were less than 1 and 3 5% for the above mentioned cell numbers. The results showing the relationship between fioating DE surfaces and various cell concentrations are summarized in Figure 1.
The thickness measurements of all dermal equivalents were performed by optical microscopy of histoiogical slides. A Leitz Wetzlar (Germany) eyepiece with cross reticulation (TO mm/100 divisions) was employed. All dermal equivalents showed thickness variations. Therefore, we performed five measurements on each equivalent and the mean result with the SD is included in Table 1.
As was shown previously.' the optical properties of collagen gels change during contraction. They become progressively more opaque and eventually whitishyellow. Furthermore, the consistency of dermal equivalents varied depending on cell numbers and collagen concentration. Increase of the collagen concentration was accompanied by a reduction in the contraction rate and by an augmentation of the final surface. Figure 2 shows an example of the influence of collagen concentration.
Results
Peripherally anchored dermal equivalent formation
Floating dermal equivalent formation
Changes in cell morphology were simiiar to those observed in floating dermal equivalent formation as observed by phase-contrast microscopy. After gelation. cells were evenly sprinkled throughout the DE. Refringent fibres appeared within 24 h connecting ceils in a circular pattern. Our observations suggest that daughter
Thickness measurements
Under phase-contrast microscopy, round, trypsinized fibroblasts began to show cell processes a few minutes after gel setting. Two hours later the majority of cells became stellate. After a short time-lag, contraction
Table 1. Thickness of relHted FDR and ADE in
FDE ADE FDE ADE
Collagen concentration (mg/ml)
18,000
211
1499 ± 1 6 8
4 22 4 22
Fibroblasts/cm1235
1 J45±9I 30 7 ±58 Nl) NI)
617
555
1169 ±109 i26±65 ND ND
782±I17 286 ±J6 ND ND
"Expressed as mean±SD. ND = not determined. I-'DK were Ihicker than ADI' in related experimental conditions, Riindom analysis showed thai Ihe anchoring method induced formation of thinner dermal equivalents than the floating method. Initial volume was 20 ml and initial surface was 81 cm* for both PDE and ADE,
ibS
C.LOPEZ-VALLE et aL
SOi Fibroblasts numbers -m-•-0-*-*-
Collagen concentration O 4,22 mg/ml • 2,11 mg/ml
70 -
1,5X10^ 5,0X10'' 4.5X10'' 3,6X10'* 2.0X10'*
60 50 -
E
0 a
40 -
,i
^^*—A—
ace
£ 30 . 20 . 10 -
H
10 -
12 Days
16
Days
Figure 1. Influence of fibrohlasts numbers on final surface, 'I'hc initial contraction speed varied in relation to cell numbers. By the fourth day mosi ofthe contraction process had occurred. No further contraction was observed after the sixteenth day in any experimental amdition. Collagen concentration was 2'11 mg/ml.
Figure 2. Influence of collagen concentration on final surface. Increase ofthe collagen concentration was accompanied by a reduction in the contraction rale and by an augmentation of the final surface. Total libroblast number was constant, 4'5 x 10'' (initial concentration of 555 cells/cm-). Bars represent SEM from three replicates.
cells formed bundles in this original arrangement, and this increasing number of fibroblasts began outward migration in all directions (Fig. 3). At the final stage, cells were arranged in a complex three-dimensional circular array, with most cells parallel to the surface ofthe Petri dish. No surface reduction was noted when the fibreglass filter ring was used. Anchored dermal equivalents with a collagen concentration of 2 1 1 mg/mi and containing 18.500 fibroblasts/cm' had a tendency to tear in their central portion. This phenomenon disappeared when the collagen concentration was raised to 4-22 mg/ml for the same cell number. A progressive opacity and solidification ofthe gel were also observed macroscopically. according to culture time.
2. Fibroblasts within the FDE were orientated perpendicular to tissue length, whereas those present at the FDE surface were parallel to tissue length (Fig. 4b). 3. Collagen fibres were orientated perpendicular to the .surface (Fig. 4b). This histological feature was not observed when the fibrobiast numbers were < 617 cells/ cm^. 4. The surface of floating dermal equivalents containing more than 1235 cells/cm^ showed many folds. These folds were filled with colonies of fibroblasts (Fig. 4al.
Floating dermal equivalents (PDF) Histology of fioating dermal equivalents synthesized with different cell numbers and collagen concentrations presented some common characteristics. 1. A fibrobiast multilayer enveloped both surfaces of floating equivalents. This phenomenon did not occur when the cell number was lower than 444 fibroblasts/
Anchored dermal equivalents (ADF.) Histology of peripherally anchored dermal equivalents differed from the floating equivalents. 1. All experimental conditions showed an absence ofa fibrobiast multilayer envelope around the matrix. 2. Fibroblasts were orientated longitudinally to tissue length (Fig. 5b). 3. Collagen fibre orientation was mostly longitudinal in alignment (Fig. 5b). 4. No folds were observed. A microscopic study of the fibre-glass filter ring in the ADE showed penetration of the collagen into the interstices of this material. The anchoring phenomenon seemed to be related to its entrapment in the porous
PERIPHERAL ANCHORAGE OF DERMAL EQUIVALENTS
569
some related experimental conditions are presented in Table 1.
Discussion
Figure 3. Living peripherally anchored dermal equivalent conlaininn I S x l O ' ' fibrohlasts at 48 h in culture (18,000 cells/cm^). Cells arranged in a circular pattern. Daughter cells began migrulion outward in all directions. View from below using phase contrasl microscopy ( x 1 J 5).
Structure, rather than to any cellular binding, as very lew cells were seen within the fibre-glass ring (Fig. 5c). Stability of peripherally anchored equivalents A set of experiments was carried out to evaluate the stability of these ADK after their detachment from the tilter ring. A contraction of approximately 10% occurred within 5 days after detachment in dermal equivalents with 18.500 libroblasts/cnr (Fig. 61. Thickness measurements Fifty-two fioating dermal equivalents were compared with 44 anchored dermal equivalents. Random analysis showed a mean thickness of 673 fim for FDE. with a range of 151-1560 }im. Mean thickness for ADE was 548 ^m with a range of 100-496 ^m. Comparison of
Kigure 4. (a| Histological appearance of full thickness FDE showing an envelope of fibroblasts mainly on the surface, and one Ibid tilled by u colony of fibroblasts (FC). Initial collagen concentration was 2-1 I mg/ ml and fibrobiast number was l8.5()O/cm' (Haematoxylin and eosin. Taken at x 20 and enlarged photographically to x 4().| (b) FnlargemenE of the indicated portion showing fibroblasts (arrowheads) and collagen Iibres orienlated perpendicular to tissue length, (Taken at x 80 and enlarged photographically to x 160,)
This study was initiated to evaluate whether a peripheral anchoring technique might solve the inherent deficiencies in the usual methods of recreating a DE in vitro. The results obtained with the FDE technique illustrate these pitfalls. The FDE technique has invariably produced iinal surfaces that were so small that they precluded their use as clinically significant grafts for burn therapy. Lowering fibrobiast numbers in these collagen lattices lessened this shrinkage phenomenon.' '"'•''' However, the resultant DE was then so poorly organized that it was very fragile. Furthermore, the histological resemblance to normal dermis was lost. One proposed solution to circumvent the shrinkage problem is the use of much larger tissue culture tlasks. There are. however, practical limits to this approach, and the initial surface would have to be extremely large. For example, to obtain a S x 8 cm graft, considering a conservative 9 5% shrinkage level, an initial culture dish of 100 X 1 60 cm would be necessary. These conditions are impractical and preclude the possibility of any scaling up of operations. In addition, if a solution of keratinocytes is then 'plated' on these DE to obtain SE. a further reduction of the final surface will ensue which will greatly compound the problem.'"*''' The resultant FDE are relatively thick, and this may have significant effects on the process of angiogenesis which normally occurs after transplantation. As showed by Yannas. using cell-free polymeric membranes, this ingrowth process could take between 10 and 46 days after application, depending mainly on thickness, chemical composition and type of polymerization.-"' -' Interference with vaseularization lessens the viability of the
370
C.LOPEZ-VALLE et al
Figure 5. (al Histological appearance of full thickness ADE. Comparison with h'igurt 4|a) shows that the thicicncss of the ADI-; is less for the siime conditions (initial cell and collagen concentrations). Initial collagen concentration was 2-\\ mg/ml and fibroblast number was 18,5()()/cm-. (Haematoxylin and eosin. Taken at x 20 and enlarged photographically to x40.) (b) Fnlargement of tbe indicated portion of Figure 5(a) showing fibroblasts (arrowheads) and collagen fibres orientated longitudinally in relation to tissue length, (Taken at x 4U and enlarged photographically to x 160.) (c| Histology of the libre-giass ring (F) interface witb the ADE (arrowbead), Haematoxylin and eosin staining. (Taken at X 20 and enlarged photographically to x40,)
(b)
overlying epidermis and results in a lower take level. However, this was not noted in previously described experiments using SE.^'"' but it is possible that the small size of these grafts permitted their lateral nourishment. Lastly, our histological analysis of these FDE has shown that their cellular and fibrillar arrangement, mostly perpendicular to the surface, is not comparable with normal human dermis.''* It seems logical to strive for DE with histological features as close as possible to the normal dermis.
80
24
Figure ft. Comparative final surfaces of floating and anchored dermal equivalents. I'ioating dermal equivalent surface: 0-74%. anchored dermal equivalent surface after freeing: 87-8%, Collagen concentration 2 1 I mg/ml, fibroblasts numbers: 18.S(H)/cm^. Arrow indicates the freeing day. Bars represent SHM from tbree replicates.
Our results show that the anchoring of the dermal equivalent is a practical solution to the drawbacks of the EDE. In their studies of tissue morphogenesis. Harris et al. described a related technique for stabilizing collagen gels devoid of incorporated cells. Eibroblasts were subsequently added to the surface of these collagen gels. The objective of these studies was the analysis of the generation of fibrillar patterns created by various mechanical instabilities.-*^ '" whereas we wished to obtain large and stable DE. We have shown that the peripheral anchoring technique can achieve such results without complicated manipulations of the culture conditions. Anchorage transformed the contraction process into a unidimensional rather than a three-dimensional phenomenon. The only axis of shrinkage involved the thickness of the ADK because it was the only axis along which the fibroblasts could act freely. As previously discussed, we fee! that a thinner DE may facilitate graft take and the survival of the keratinocytes contained in the SE. This cultured SE has to be nourished by osmosis during the first few days after grafting, as are standard grafts." The histological features of the ADE are also much closer to normal dermis.'** The longitudinal arrangement of cells and fibrils in ADE is quite remarkable. Our observations suggest that the attachment of the ADE to the filter ring is a fibrillar entrapment phenomenon rather than a cellular binding mechanism. Thus, it is possible that any porous material will permit efficacious anchoring of dermal equivalents. The choice of an optimal material would be dictated by its biocompatibility. by its ease of use and its resistance to sterilization, allowing repeated utilization.
PERIPHERAL ANCHORAGE OF DERMAL EQUIVALENTS
We helicve that the ADE model presented in this report is a first step in the production of new skin suhstitutes for treatment of burn patients by tissue culture methods. The addition of other molecules in the initial mixture could be helpful in obtaining a more complex and structured dermal equivalent. Such an anchoring technique could be readily amenable to large-scale production of skin equivalent for wound coverage of burns.
Acknowledgments This work was supported by Grant No. MA-1022 i from the Medical Research Council of Canada. Award B-2 committee 01A from the Fonds FCAR (Quebec) to C.A.Lopez Valle and Award PCJS 1 from Natural Sciences and Engineering Research Council of Canada to P.Rompre, and Fellowships from the Fonds de la Recherche en Sante du Quebec (FRSQ) to F.A.Auger and L.Germain, We thank Dr George Nascimento from Chiron Corporation for graciously supplying F^GF, Celine D.Fugere for assistance with histology and jo-Anne Masse for help with technical work.
References 1 lioykin IVJr, Molnar ]A, Burn scar and skin equivalents. In: Wound Healing, Biochemica! i'-^ Clinical Aspects iCohen IK, Diegelmann RF, Lindblad WJ. eds|. Philadelphia; WB Saunders Company. 1992; 525-40, 2 Gallico GG, O'Connor NE. Conipton CC et at. Permanent coverage of large burn wounds with autologous cultured human epithelium. New/^fiif// MeiJ 1984; J11:44X-51, ? S/irmai |A. 'Ihe organization ofthe dermis. In: Advances in Biology of Skin, Vol, X, The Dermis. (Montagna W. Bentley JP, Dobson RL, eds|. New York: Appleton-Century-Crofts. 19f»S; 12, 4 Daly CH, Biotnechanical properties of dermis, / Invest IX'rwatol 1982: 79 (SuppI, 1): 17-20, 5 Compton CC. Gill ]M, Bradford HA et ul Skin regenerated from cultured epitheliiil autografts on full-thickness burn wounds from fidays to 5 years after grafting, !Aib !m'est 1989; 60: M)0-\2, fi Fhrmann RL. (ley GO. The growth of cells on a iransparent gel of reconstituted rat-tail collagen, / Natl Cancer !nsl 1956; 16: 1575403. 7 Bell E, Ivarsson B, Merrill C, Production ofa tissue-like structure by contraction of collagen lattices by human fibroblasts of different proliferative potential in vitro. Proc Natl Acad Sci USA 1979: 76: 1274-8. 8 Bell E. EhHich HP. Buttle DJ. Nakatsuji T, Living tissue formed in vitro and accepted as skin-equivalent tissue of full thickness. Science 1981:211: 1052-4, 9 Bell, E, Ehrlich HP. Sher S ct a!. Development and use ofa living skin equivalent. Wast Reconstr Surg 1981; 67: 586-92.
371
10 Bell E, Sher S. Hull B ci al. The reconstitution of living skin, / !nvest Dermatol 1985: 81 tSuppl,): 2-10, 11 Coulomb B. Saiiig P. Bell F et al. A new method for studying epidermaJization ifi vilro, Br I Dermatol 1986; 114:91-101, 12 Elsdale T. Bard |. Collagen substrata for studies on cell behaviour, / Cell Biol 1972: 54:626-37, 1 i Wood CC. Keofh MK, The formation of fibrils from collagen .solutions. Biochem j 1960: 75: 588-98, 14 Rompre P, Auger FA. Clermain L el al Influence of initial collagen and cellular concentrations on the final surface area of dermal and skin equivalents: a Box-Behnken analysis. In Vitro Cell Dev Bio! 1990: 2 6 : 9 8 3 - 9 0 . 15 Souren |M. Ponec M. van Wijk R, Contraction of collagen by human fibroblasts and keratinocytes. In Vilro Cell Dcv iiio! I 98S: 25: 1039-45. 16 Schafer IA, Shapiro A, Kovach M el al The interaction of human papillary and reticular hbroblasts and human keratinocytes in the contraction of three-dimensional Hoating collagen lattices, Exp CeU Res 1989: 18 3: 112-25. 1 7 l.iliie |H, MacCalhim DK. |epsen A, (irowth ofslralified squamous epithelium on reconstituted extracellular matrices: long-term culture. / Invesl Dermatol 1988: 90: 100-9, 18 Rowling P|F. Raxworthy M|, Wood H[ et al. The infiuence of keratinocyte .seeding and keratinocyte growth medium on dermal equivalent contraction. / Invest Dermato! 1989: 92: 510. 19 lx!ver WF, Schaumburg-Lever G |Eds), Histopathology of the SIcin. 6th edn, Philadelphia: JB Lippincott Company. 1983; 29, 20 Rheinwald |G. Green H. Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinining colonies from single cells, Celt 1975: 6: 331-44. 21 Green H, Kehinde (), Thomiis J, Growth of cultured human epidermal cells into multiple epilhelia suitable for grafting, Proc Nail Acad Sci USA 1979: 76: S66S-8. 22 Banks-Sfhlegel S. Green H, Formation of epidermis by serially cultivated human epidermal celis transplanted as an epithelium to athymic mice. Transplantation 1980: 29: 508-1 I, 25 O'Connor NE, Mulliken |B, Biuiks-SchlegelScIu/.C.rafting of burns with cultured epithelium prepared from autologous epidermal ceils. iMmrt 1981: i: 75-8. 24, Beil [i. Sher S, Hull B, The living skin-equivalent as (i structural and immunological model in skin grafting. Scan E!ectron Microsc 1984: (Pt4|; I9S7-62, 25 Yannas IV, Burke JF, Design ofan artificial skin, I, Basic design principles. / Biomed Mater Res 1980: 14: 6S-81. 26 Yannas IV, Burke JF, Orgill DP, Skrabut EM, Wound tissue can utilize a polymeric template to synthesize a fundioiial extension of skin, SVimc 1982; 215: 174-6, 2 7 Yannas IV, Lee E, Orgill DP et al Synthesis and characterization of a model extracellular matrix that induces partial regeneration of adult mammalian skin, Proc Nat! Acad Sci USA 1989:86: 933-7. 28 Harris AK, Stopak D. Wild P, Fibrobiast traction as a mechanism for collagen morphogenesis. Nature 1981: 290: 249-51, 29 Stopak D. Harris AK. Connective tissue morphogenesis by fibrobiast traction, 1. Tissue culture observations, Dev Biol IS82; 90: JHJ-98, M) [larris AK. Stopak D. Wamer P, Generation of spatially periodic patterns by a mechanical instability: A mechanical alternative to the Turing model, / Emhryot Exp Morphot 1984; 80: 1-20, 3 1 Smith JW, Ashton S| (Fdsl Grabb and Smiths Plastic Surgery. 4th edn. Boston; Little. Brown and Company, 1991: 27,
