Portal de Periódicos da CAPES

Keratinocytes Influence the Maturation and Organization of the Elastin Network in a Skin Equivalent11The authors declared in writing to have no conflict of interest.

2000; Elsevier BV; Volume: 114; Issue: 2 Linguagem: Inglês

10.1046/j.1523-1747.2000.00885.x

1523-1747

Florence Duplan-Perrat, Odile Damour, Caroline Montrocher, S Peyrol, Guillaume Grenier, Marie‐Paule Jacob, F. Braye,

Skin and Cellular Biology Research

Elastic fibers form a complex network that contributes to the elasticity of connective tissues. Alterations in the elastic fiber network are involved in several disease affecting organs in which compliance of the connective tissue is essential: skin, main vasculature, lung, joints, muscle, and ligament. The aim of our work was to study the deposition, maturation, and organization of elastic fiber components in a dermal equivalent model consisting of collagen-GAG-chitosan seeded with fibroblasts. The influence of keratinocytes was studied in parallel, thus constituting a skin equivalent model. These models were examined by transmission electron microscopy (TEM) and by immunohistochemistry to determine the staining patterns of fibrillin-1 and elastin proteins representative of the microfibrillar framework and of the elastic fibers, respectively. After 2 mo of fibroblast culture in the dermal equivalent, elastin was undetectable, whereas fibrillin-1 staining was weak and microfibrils were infrequently observed by TEM. In the skin equivalent, fibrillin-1 and elastin were detected by immunostaining 15 d after epidermization and TEM revealed the typical structure and organization of the elastic network in the dermis, with elastin deposition on the microfibrillar scaffold. This in vitro skin equivalent model is to our knowledge the first in which elastic fibers have been detected, thus demonstrating the influence of keratinocytes on the maturation and organization of the elastic network. Elastic fibers form a complex network that contributes to the elasticity of connective tissues. Alterations in the elastic fiber network are involved in several disease affecting organs in which compliance of the connective tissue is essential: skin, main vasculature, lung, joints, muscle, and ligament. The aim of our work was to study the deposition, maturation, and organization of elastic fiber components in a dermal equivalent model consisting of collagen-GAG-chitosan seeded with fibroblasts. The influence of keratinocytes was studied in parallel, thus constituting a skin equivalent model. These models were examined by transmission electron microscopy (TEM) and by immunohistochemistry to determine the staining patterns of fibrillin-1 and elastin proteins representative of the microfibrillar framework and of the elastic fibers, respectively. After 2 mo of fibroblast culture in the dermal equivalent, elastin was undetectable, whereas fibrillin-1 staining was weak and microfibrils were infrequently observed by TEM. In the skin equivalent, fibrillin-1 and elastin were detected by immunostaining 15 d after epidermization and TEM revealed the typical structure and organization of the elastic network in the dermis, with elastin deposition on the microfibrillar scaffold. This in vitro skin equivalent model is to our knowledge the first in which elastic fibers have been detected, thus demonstrating the influence of keratinocytes on the maturation and organization of the elastic network. dermal equivalent dermal substrate glycosaminoglycans rough endoplasmic reticulum skin equivalent transmission electron microscopy Elastic fibers form a complex network that is responsible for the elasticity of tissues: skin, arteries, lungs, etc. (Gibson et al., 1989Gibson M.A. Kumaratilake J.S. Cleary E.G. The protein components of the 12-nanometer microfibrils of elastic and non elastic tissues.J Biol Chem. 1989; 264: 4590-4598Abstract Full Text PDF PubMed Google Scholar;Robert, 1994Robert L. Le Vieillissement. Belin, CNRS éditions1994Google Scholar). These fibers play an important role in the extracellular matrix (ECM) as shown by the consequences of their ultrastructural and histopathologic alterations observed in several skin diseases and genetic diseases such as Marfan's syndrome (Christiano and Uitto, 1994Christiano A.M. Uitto J. Molecular pathology of the elastic fibers.J Invest Dermatol. 1994; 103: 53S-57SCrossref PubMed Scopus (49) Google Scholar;Reinhardt et al., 1995Reinhardt D.P. Chalberg S.C. Sakai L.Y. The ultrastructure and function of fibrillin.Ciba Foundation Symposium. 1995; 192: 128-147PubMed Google Scholar). The regular structure of the elastic fibers contributes to the compliant properties of the ECM; in contrast, structural alterations as observed for instance in Marfan's syndrome impair aortic wall tonicity with progressive aortic diameter enlargement and final dissection (Christiano and Uitto, 1994Christiano A.M. Uitto J. Molecular pathology of the elastic fibers.J Invest Dermatol. 1994; 103: 53S-57SCrossref PubMed Scopus (49) Google Scholar;Grimaud and Peyrol, 1994Grimaud J.A. Peyrol S. Inherited Connective Tissue Diseases: Potential Diagnostic Value of Fibrillin Immunodetection in Skin Biopsies of Marfan Patients. vol. 35. Proceedings of the 13th International congress on Electron Microscopy, Applications in Biological Sciences. Physique, Paris1994: 1069-1072Google Scholar;Reinhardt et al., 1995Reinhardt D.P. Chalberg S.C. Sakai L.Y. The ultrastructure and function of fibrillin.Ciba Foundation Symposium. 1995; 192: 128-147PubMed Google Scholar). Morphologic studies of skin have revealed the structural heterogeneity of elastic fibers (Cotta-Pereira et al., 1978Cotta-Pereira G. Guerro-Rodriguez F. Bittencourt-Sampaio S. Oxytalan, elaunin and the elastic fibers in the human skin.J Invest Dermatol. 1978; 66: 143-148Crossref Scopus (239) Google Scholar): (i) fine oxytalan fibers are perpendicularly oriented to the dermal–epidermal junction (DEJ) and contribute to the connection between dermis and epidermis. At the upper reticular border (ii) thicker elaunin fibers are disposed parallel to the epidermis and (iii) large elastic fibers are found between the collagen fiber bundles in the reticular dermis (for review, seeFrances and Robert, 1984Frances C. Robert L. Elastin and elastic fibers in normal and pathologic skin.Int J Dermatol. 1984; 23: 166-179Crossref PubMed Scopus (94) Google Scholar;Robert, 1994Robert L. Le Vieillissement. Belin, CNRS éditions1994Google Scholar). Oxytalan fibers are composed of bundles of parallel microfibrils with a circular 12 nm cross-section; immuno-labeling reveals fibrillin as a microfibrillar component and only small amounts of elastin in spots associated with microfibrils (Sakai et al., 1986Sakai L.Y. Keene D.R. Engvall E. Fibrillin, a new 350-kD glycoprotein, is a component of extracellular microfibrils.J Cell Biol. 1986; 103: 2499-2509Crossref PubMed Scopus (866) Google Scholar;Schwartz and Fleischmajer, 1986Schwartz E. Fleischmajer R. Association of elastin with oxytalan fibers of the dermis and with extracellular microfibrils of cultured skin fibroblasts.J Histochem Cytochem. 1986; 34: 1063-1068Crossref PubMed Scopus (31) Google Scholar;Dahlbäck et al., 1990Dahlbäck K. Ljungquist A. Löfberg H. Dahlbäck B. Engvall E. Sakai L.Y. Fibrillin immunoreactive fibers constitute a unique network in the human dermis: immunohistochemical comparison of the distributions of fibrillin, vitronectin, amyloïd P component, and orcein stainable structures in normal skin and elastosis.J Invest Dermatol. 1990; 94: 284-291Abstract Full Text PDF PubMed Google Scholar). Elaunin fibers are constituted by bare microfibrils and a small amount of covering amorphous elastin, as shown after glutaraldehyde fixation; they exhibit a structure intermediate between oxytalan fibers and elastic fibers that develop in the reticular dermis as a dense network of larger fibers mainly composed of elastin. It is currently hypothesized that these three levels of organization reproduce the ontogenetic pathway of elastic fiber maturation: elastogenesis is initiated with the deposition of microfibrils that form a template for tropoelastin deposition (Christiano and Uitto, 1994Christiano A.M. Uitto J. Molecular pathology of the elastic fibers.J Invest Dermatol. 1994; 103: 53S-57SCrossref PubMed Scopus (49) Google Scholar). Alignment of elastin on microfibrils is stabilized by the formation of intermolecular crosslinks (desmosine and isodesmosine; for review seeFrances and Robert, 1984Frances C. Robert L. Elastin and elastic fibers in normal and pathologic skin.Int J Dermatol. 1984; 23: 166-179Crossref PubMed Scopus (94) Google Scholar). The mechanisms underlying the formation of elastic fibers are not completely understood. Fibroblasts, cultured in monolayer, synthesize tropoelastin, the soluble precursor of elastin (Sephel et al., 1986Sephel G.C. Davidson B.S. Davidson J.M. Elastin production in human skin fibroblast cultures and its decline with age.J Invest Dermatol. 1986; 86: 279-285Crossref PubMed Scopus (89) Google Scholar) and produce microfibrils (Haynes et al., 1997Haynes S.L. Shuttleworth C.A. Kielty C.M. Keratinocytes express fibrillin and assemble microfibrils: implications in dermal matrix organization.Br J Derm. 1997; 137: 17-23Crossref PubMed Scopus (42) Google Scholar). Keratinocytes in culture are also able to synthesize tropoelastin (Kajiya et al., 1997Kajiya H. Tanaka N. Inaumi T. Seyama Y. Tajima S. Ishibashi A. Cultured human keratinocytes express tropoelastin.J Invest Dermatol. 1997; 109: 641-644Crossref PubMed Scopus (24) Google Scholar) and fibrillin (Haynes et al., 1997Haynes S.L. Shuttleworth C.A. Kielty C.M. Keratinocytes express fibrillin and assemble microfibrils: implications in dermal matrix organization.Br J Derm. 1997; 137: 17-23Crossref PubMed Scopus (42) Google Scholar). Nevertheless, bi-dimensional fibroblast cultures do not produce mature elastic fibers.Fleischmajer et al., 1993Fleischmajer R. MacDonald E.D. Contard P. Perlish J.S. Immunochemistry of a keratinocyte-fibroblast co-culture model for reconstruction of human skin.J Histochem Cytochem. 1993; 41: 1359-1366Crossref PubMed Scopus (66) Google Scholar, using a three-dimensional coculture model of fibroblasts and keratinocytes on a nylon mesh, have demonstrated the presence of 10–12 nm microfibrils. These microfibrils, which have a beaded appearance, contain fibrillin; however, no elastin could be detected in this model. The aim of this study was to investigate the maturation and ultrastructural organization of elastic fibers in two three-dimensional models: a dermal equivalent (DE) and a skin equivalent (SE). The DE was obtained by culturing fibroblasts in a porous dermal substrate (DS) made of collagen, glycosaminoglycans, and chitosan (Collombel et al., 1989Collombel C. Damour O. Gagnieu C. Marichy C. Poinsignon F. Biomaterials with a base of collagen, chitosane and glycosaminoglycans, process for preparing them and their application in human medecine. 1989Google Scholar). The SE was obtained by culturing keratinocytes on a DE substrate (Shahabeddin et al., 1990Shahabeddin L. Berthod F. Damour O. Collombel C. Characterization of skin reconstructed on a chitosan-cross-linked collagen-glycosaminoglycan matrix.Skin Pharmacol. 1990; 3: 107-114Crossref PubMed Scopus (92) Google Scholar). These two models have applications in the treatment of extensive burns or chronic wounds, in basic research to study cell–cell interactions (Saintigny et al., 1993Saintigny G. Bonnard M. Damour O. Collombel C. Reconstruction of epidermis on a chitosan-cross-linked collagen-GAG lattice: effect of fibroblasts.Acta Derm Venereol. 1993; 73: 175-180PubMed Google Scholar) or cell–matrix interactions (Sahuc et al., 1996Sahuc F. Nakazawa K. Berthod F. Collombel C. Damour O. Mesenchymal–epithelial interactions regulate gene expression of type VII collagen and kalinin in keratinocytes and dermal-epidermal junction in a skin equivalent model.Wound Rep Reg. 1996; 4: 93-102Crossref PubMed Scopus (40) Google Scholar;Berthod et al., 1997Berthod F. Germain L. Guignard R. et al.Differential expression of collagens XII and XIV in human skin and in reconstructed skin.J Invest Dermatol. 1997; 108: 737-742Crossref PubMed Scopus (69) Google Scholar), and also for pharmacologic tests as an alternative to animal experimentation (Augustin et al., 1997Augustin C. Collombel C. Damour O. Measurement of the photoprotective effect of topically applied sunscreeens using in vitro dermal and skin equivalent.Photochem Photobiol. 1997; 66: 853-859Crossref PubMed Scopus (24) Google Scholar,Augustin et al., 1998Augustin C. Collombel C. Damour O. Use of dermal equivalent and skin equivalent models for in vitro cutaneous irritation testing of cosmetic products: comparaison with in vivo human data.J Toxicol Cut Ocular Toxicol. 1998; 17: 5-17Crossref Scopus (21) Google Scholar;Damour et al., 1998Damour O. Augustin C. Black A.F. Applications of reconstructed skin models in pharmaco-toxicological trials.Med Biol Eng Comput. 1998; 36: 825-832Crossref PubMed Scopus (25) Google Scholar). The distribution of elastic components was evaluated by immunofluorescence using antibodies to fibrillin-1 and elastin, and the ultrastructural organization of the neosynthesized elastic fibers by transmission electron microscopy (TEM). Keratinocytes and fibroblasts were isolated from human foreskin. Fibroblasts were grown in Dulbecco's modified Eagle's medium (DMEM) (Sigma, St Quentin-Fallavier, France) supplemented with 10% calf serum (Hyclone, La Varenne-St Hilaire, France), 4 mM L-glutamine (Sigma), 20 μg gentamicin (Dakota, St Quentin-Fallavier, France) per ml, 100 IU penicillin (Sarbach, Suresnes, France) per ml, and 1 μg amphotericin B (Bristol Myers Squibb, Paris, France) per ml. Keratinocytes were grown in a mixture of DMEM and Ham's F12 (Sigma) (3:1), supplemented with 10% calf serum (Hyclone), 4 mM L-glutamine, 10 ng epidermal growth factor (EGF) (Austral Biologic, San Raton) per ml, 0.12 IU insulin (Lilly, St Cloud, France) per ml, 0.4 μg hydrocortisone (UpJohn, St Quentin en Yuelines, France) per ml, 5 μg triiodo-L-thyronine (Sigma) per ml, 24.3 μg adenine (Sigma) per ml and antibiotics. Fibroblasts from the seventh to ninth passage were seeded at a density of 200 000 cells per cm2 onto DS made of chitosan-cross-linked collagen-GAG matrix (18 mm diameter and 5 mm thick discs) prepared as previously described byCollombel et al., 1989Collombel C. Damour O. Gagnieu C. Marichy C. Poinsignon F. Biomaterials with a base of collagen, chitosane and glycosaminoglycans, process for preparing them and their application in human medecine. 1989Google Scholar. DE were cultured for 15 d in a 5% CO2 atmosphere at 37°C. The medium was supplemented with 50 μg L-ascorbic acid (Sigma) per ml and changed every day. DE were cultured for up to 75 d. Keratinocytes were seeded at a density of 200 000 cells per cm2 on 15-d-old DE. After 7 d of culture in keratinocyte medium, the immersed SE were elevated to the air–liquid interface and cultured for 38 d in a modified keratinocyte medium (DMEM supplemented with 10% calf serum, 4 mmol L-glutamine per liter, 10 ng EGF per ml, 0.12 IU insulin per ml, 0.4 μg hydrocortisone and antibiotics per ml). The medium was supplemented with 50 μg per ml ascorbic acid and changed three times a week. Expression and ultrastructural organization of elastic fibers (elastin and fibrillin) were studied in DE and SE. DE and SE samples were harvested after 30, 45, 60, and 75 d in culture, fibroblast seeding in DE being considered as day 1. At each time point, samples were harvested for histology, immunohistochemistry, and electron microscopy studies. SE and DE were fixed in Bouin fixative and embedded in paraffin. Six micrometer sections were stained with Hematoxylin-Phloxin-Safron (HPS). DE and SE were embedded in OCT Tissue-Tek (Miles, Immunotech, Marseille, France). Six micrometer frozen sections were obtained with a Frigocut 2800 cryotome (Reichert-Jung, Paris, France), air dried, blocked in phosphate buffer saline containing 1% (wt/vol) bovine serum albumin. The antibodies used for this study were directed against human elastin (polyclonal, raised in rabbit, 1:100 dilution, generously given by D. Hartmann, 25011 Novotec, France) and bovine fibrillin-1 (monoclonal, murine, 1:50 dilution, 11C1.3 or MS 231 P1 Interchim, France). Secondary antibodies, either goat antirabbit IgG (1:50 dilution, Sanofi Diagnostics Pasteur, France) or goat antimouse IgG (1:50 dilution, Cedarlane, Canada) labeled with FITC, were mixed with 0.1% Evans Blue to reduce nonspecific staining of the sponge network (Kieny and Mauger, 1984Kieny H.K. Mauger A. Immunofluorescence localization of extracellular matrix componments during muscle morphogenesis.J Exp Zool. 1984; 232: 327-341Crossref Scopus (21) Google Scholar). For controls, the primary antibody was omitted. Multiple sections of each specimen were processed to ensure representative samples. Samples were fixed in 2% glutaraldehyde - 0.1 M NaCacodylate/HCl at pH 7.4 for 1 h at 4°C, postfixed in 1% OsO4-0.15 M Na Cacodylate/HCl pH 7.4 for 1 h at 4°C, dehydrated in graded ethanol, and embedded in Epon. Ultrathin sections were cut with a LKB Ultratome V, contrasted with methanolic uranyl acetate (10 mn) and lead citrate (5 mn) and examined in a JOEL 1200 EX transmission electron microscope. After seeding, the cells migrate into the porous structure, proliferate, and progressively synthesize ECM. After 30 d of culture in DS (Figure 1a), fibroblasts had colonized one-third of the DS. After 60 d (Figure 1c), the colonization was no more than three-quarters of the thickness and ECM neosynthesis by fibroblasts was poor; however, after 15 d, fibroblast proliferation and ECM synthesis was sufficient to fill the porous structure of the DS surface allowing keratinocyte seeding. On day 30 (Figure 1b), we observed on the epidermal part of the skin equivalent the development of a continuous pluristratified and well-differentiated epidermis without penetration of the keratinocytes into the DE. In the dermal part, the fibroblasts continued to migrate and to fill the pores with the neosynthesized ECM. On day 60 (corresponding to 45 d of epidermization) (Figure 1d), the epidermis was still stratified and organized into basal cells, numerous suprabasal layers, and stratum corneum. In the dermal compartment, the fibroblasts had migrated and proliferated, leading to complete colonization of the DS at day 60. Moreover, the porous structure of the DS was completely filled with fibroblasts and neosynthesized ECM. Histologic comparison of DE and SE showed that colonization of DS by fibroblasts and ECM was more extensive in the SE than in the DE (Figure 1). After 30 d of culture, a fine, diffuse, and disorganized expression of fibrillin-1 was observed in DE (Figure 2c). Fibrillin-1 labeling progressively increased with time, though its expression remained low, diffuse, and disorganized (Figure 2e); however, elastin could not be seen even after 75 days of culture (Figure 2d,f). In the SE, fibrillin labeling was more intense and showed various modes of organization in the dermis. Fifteen days after epidermization, fibrillin staining appeared as short fibers perpendicularly anchored to the DEJ. Deeper in the dermis, longer fibers formed a ribbon-like network parallel to the DEJ. In the reticular dermis, the fibrillin positive fibers formed a denser network (Figure 2g,i). In the same samples, elastin labeling was detected (Figure 2h), its intensity increasing with time in culture and its expression following fibroblast colonization of the DS (Figure 2j). After 45 d, very short and regularly disposed elastin deposits appeared closely perpendicularly to the DEJ. Underneath the DEJ, elastin was deposited with a parallel orientation (Figure 2j). Comparative analysis of DE and SE at different times showed that elastin and fibrillin expression and organization seemed to be dependent on the presence of keratinocytes. Collagenic and elastic matrix was more abundant and better organized in SE than in the DE. Indeed at any time (30, 45, 60, and 75 d), in the DE, hollow shaped microfibrils were rare, scattered, and localized close to fibroblasts (Figure 3a). In contrast, in the dermal part of the SE at any time, the neosynthesized ECM formed a framework of collagen fiber bundles containing abundant hollow-shaped microfibrils (Figure 3b,c). In the SE, at 30 d, groups of parallel hollow cross-sectioned microfibrils formed oxytalan fibers perpendicularly oriented to the DEJ basal lamina (Figure 3d,e); further down, microfibril bundles were found within the neosynthesized ECM but elastin was not observed. At 30 d, microfibrils were more abundant and presented the same organization; however, at this time, deposition of elastin, with amorphous aspect, was observed upon microfibril bundles (Figure 3f). After 75 d, amorphous elastin deposition upon microfibrils extended to the deeper part of the dermis (Figure 3g). At this time, from the surface downwards, we found, successively, microfibrils similar to oxytalan fibrils at the DEJ, elastin deposition on microfibrils similar to elaunin fibers, then mature elastic fibers. Previous studies have shown that the porous structure of the collagen-GAG-chitosan DS provides an excellent environment for the growth of fibroblasts (Berthod et al., 1993Berthod F. Hayek D. Damour O. Collombel C. Collagen synthesis by fibroblasts cultured within a collagen sponge.Biomaterials. 1993; 14: 749-754Crossref PubMed Scopus (102) Google Scholar,Berthod et al., 1996Berthod F. Sahuc F. Hayec D. Damour O. Collombel C. Deposition of collagen fibril bundles by long term culture of fibroblasts in a collagen sponge.J Biomed Mater Res. 1996; 32: 87-94Crossref PubMed Scopus (53) Google Scholar). Cells continue to proliferate into the pores and synthesize their own ECM until the pores are completely filled, giving rise to a DE. When keratinocytes are seeded, the quality of the DE permits a good basal adhesion of keratinocytes leading, after 2 wk of culture, to a pluristratified and well-differentiated epidermis (Saintigny et al., 1993Saintigny G. Bonnard M. Damour O. Collombel C. Reconstruction of epidermis on a chitosan-cross-linked collagen-GAG lattice: effect of fibroblasts.Acta Derm Venereol. 1993; 73: 175-180PubMed Google Scholar;Sahuc et al., 1996Sahuc F. Nakazawa K. Berthod F. Collombel C. Damour O. Mesenchymal–epithelial interactions regulate gene expression of type VII collagen and kalinin in keratinocytes and dermal-epidermal junction in a skin equivalent model.Wound Rep Reg. 1996; 4: 93-102Crossref PubMed Scopus (40) Google Scholar). Keratinocyte adhesion is mediated by fibroblast ECM proteins and multiple receptors such as integrins (Mosher et al., 1992Mosher D.F. Sottile J. Wu C. McDonald J.A. Assembly of extra-cellular matrix.Curr Opin Cell Biol. 1992; 4: 810-818Crossref PubMed Scopus (136) Google Scholar;Aumailley and Krieg, 1994Aumailley M. Krieg T. Structure and function of cutaneous extracellular matrix.Eur J Dermatol. 1994; 4: 271-280Google Scholar). The direct contact of fibroblasts and keratinocytes permits the formation of a DEJ by increasing the production and the organization of DEJ components visible by TEM, i.e., anchoring fibrils, continuous lamina densa, and numerous hemidesmosomes (Sahuc et al., 1996Sahuc F. Nakazawa K. Berthod F. Collombel C. Damour O. Mesenchymal–epithelial interactions regulate gene expression of type VII collagen and kalinin in keratinocytes and dermal-epidermal junction in a skin equivalent model.Wound Rep Reg. 1996; 4: 93-102Crossref PubMed Scopus (40) Google Scholar). In the dermal part of the SE, the neosynthesized ECM has an ultrastructural organization in which can be found collagen bundles perpendicularly oriented to each other and also numerous microfibrils (Berthod et al., 1996Berthod F. Sahuc F. Hayec D. Damour O. Collombel C. Deposition of collagen fibril bundles by long term culture of fibroblasts in a collagen sponge.J Biomed Mater Res. 1996; 32: 87-94Crossref PubMed Scopus (53) Google Scholar,Berthod et al., 1997Berthod F. Germain L. Guignard R. et al.Differential expression of collagens XII and XIV in human skin and in reconstructed skin.J Invest Dermatol. 1997; 108: 737-742Crossref PubMed Scopus (69) Google Scholar). After the demonstration of collagen organization close to that found in normal human dermis, the aim of these experiments was to investigate elastic fiber maturation. Histologic analysis confirmed that the keratinocytes influence fibroblast migration, proliferation, and ECM synthesis because, even after 75 d in culture, DS was only partially colonized by fibroblasts in DE whereas SE was completely colonized after only 45 d (30 d of epidermization). The elastic network was analysed by immunofluorescence using antibodies against elastin and fibrillin-1, the major component of elastin-associated microfibrils, showed the influence of keratinocytes on the maturation of the elastic fibers: fibrillin-1 is always detected in both models, but differences could be observed in labeling intensity, network distribution, and organization. In SE, at the DEJ, fibrillin-containing fibers perpendicular to the DEJ could be observed. The synthesis of fibrillin with assembly into microfibrils concomitant with the basal lamina formation, could favor the regular perpendicular distribution of oxytalan fibers at the DEJ observed in our model (Haynes et al., 1997Haynes S.L. Shuttleworth C.A. Kielty C.M. Keratinocytes express fibrillin and assemble microfibrils: implications in dermal matrix organization.Br J Derm. 1997; 137: 17-23Crossref PubMed Scopus (42) Google Scholar;Kielty and Shuttleworth, 1997Kielty C.M. Shuttleworth C.A. Microfibrillar elements of the dermal matrix.Microsc Res Tech. 1997; 38: 413-427Crossref PubMed Scopus (56) Google Scholar). It thus appears that oxytalan and elaunin fibers could represent the connection between deep elastic fibers and DEJ. These results are in agreement with those ofRaghunath et al., 1996Raghunath M. Bächi T. Meuli M. Altermatt S. Gobet R. Bruckner-Tuderman L. Steinmann B. Fibrillin and elastin expression in skin regenerating from cultured keratinocyte autografts: morphogenesis of microfibrils begins at the dermo-epidermal junction and precedes elastic fibre formation.J Invest Dermatol. 1996; 106: 1090-1095Crossref PubMed Scopus (61) Google Scholar who showed, after grafting cultured autologous epithelium in vivo (CAE model), that the microfibrillar scaffold is first generated perpendicularly to the DEJ. Deeper, the deposition of fibrillin-containing microfibrils followed downwards fibroblast migration. In the DE, fibrillin labeling reveals a diffuse network that appears not competent alone for deposition and polymerization of tropoelastin, indeed elastin deposition is not observed in DE despite the presence of fibrillin-1. On the contrary, elastin and elastic fibers are detected in SE. This could be explained by soluble components such as growth factors and cytokines secreted by keratinocytes that are known to stimulate migration, proliferation, and metabolic activities of fibroblasts (Delaporte et al., 1989Delaporte R. Croute F. Vincent C. Bonnefoy J.Y. Robert J. Thivolet J. Nicolas J.F. Interactions kératinocytes × fibroblastes: production par les kératinocytes de facteurs solubles stimulant la prolifération de fibroblastes dermiques humains normaux.Path Biol. 1989; 37: 875-880PubMed Google Scholar), especially elastin expression (Tajima and Izumi, 1996Tajima S. Izumi T. Differential in vitro responses of elastin expression to basic fibroblast growth factor and transforming growth factor β1 in upper, middle and lower dermal fibroblasts.Arch Dermatol Res. 1996; 288: 753-756Crossref PubMed Scopus (14) Google Scholar). It is remarkable that elastin was present only in SE. The expression of elastin in SE is already demonstrated at 30 d but with a lower intensity than fibrillin-1. By TEM elastin is only seen after 45 d, secondary to microfibrils assembly into bundles. At 30 d of culture, elastin expression was restricted to the upper (pseudo-papillary) dermis, whereas microfibrils deposition was yet detectable deeper. These results suggest the following sequence of events: first, microfibrils assembled at the DEJ without elastin; then, elastin is deposited on hollow shaped microfibril bundles beneath DEJ; and last, true elastic fibers combining microfibrillar and amorphous components are formed. Moreover, at the last time studied, elastin network organization confines to normal human skin where oxytalan, elaunin, and elastic fibers can be morphologically distinguished. This sequence of events can be compared with that known about formation of elastic fiber network in skin during ontogenesis, where collagen fibers accumulate rapidly and elastic tissue appears later (Sephel et al., 1987Sephel G.C. Buckley A. Davidson J.M. Developmental Initiation of elastin gene expression by human fetal skin fibroblasts.J Invest Dermatol. 1987; 88: 732-735Abstract Full Text PDF PubMed Google Scholar). All of these results agree with the hypothesis that microfibrils act as a scaffold for elastin deposition, preceding elastin cross-linking (Raghunath et al., 1996Raghunath M. Bächi T. Meuli M. Altermatt S. Gobet R. Bruckner-Tuderman L. Steinmann B. Fibrillin and elastin expression in skin regenerating from cultured keratinocyte autografts: morphogenesis of microfibrils begins at the dermo-epidermal junction and precedes elastic fibre formation.J Invest Dermatol. 1996; 106: 1090-1095Crossref PubMed Scopus (61) Google Scholar). In the in vivo CAE model,Raghunath et al., 1996Raghunath M. Bächi T. Meuli M. Altermatt S. Gobet R. Bruckner-Tuderman L. Steinmann B. Fibrillin and elastin expression in skin regenerating from cultured keratinocyte autografts: morphogenesis of microfibrils begins at the dermo-epidermal junction and precedes elastic fibre formation.J Invest Dermatol. 1996; 106: 1090-1095Crossref PubMed Scopus (61) Google Scholar have shown that elastin first covers fibrillin-containing microfibrils of the reticular dermis, whereas, in our in vitro model, elastin deposition on microfibrils begins in the upper (pseudo-papillary) dermis. Our results suggest that, in our models, morphogenesis of elastic fibers follows fibroblast migration and neosynthesis of other ECM components. The fact that fibroblasts arise from the wound bed in the CAE model and from the top in our in vitro model could explain this difference. In agreement with our results, the influence of keratinocytes on the maturation and the organization of fibrillin microfibrils was demonstrated in an in vitro three-dimensional nylon mesh model (Fleischmajer et al., 1991Fleischmajer R. Contard P. Schwartz E. MacDonald E.D. Jacobs L. Sakai L. Elastin-associated microfibrils (10 nm) in a three-dimensional fibroblast culture.J Invest Dermatol. 1991; 97: 638-643Abstract Full Text PDF PubMed Google Scholar,Fleischmajer et al., 1993Fleischmajer R. MacDonald E.D. Contard P. Perlish J.S. Immunochemistry of a keratinocyte-fibroblast co-culture model for reconstruction of human skin.J Histochem Cytochem. 1993; 41: 1359-1366Crossref PubMed Scopus (66) Google Scholar;Contard et al., 1993Contard P. Bartel R.L. Jacobs I.L. et al.Culturing keratinocytes and fibroblasts in a three-dimensional mesh results in epidermal differentiation and formation of a basal lamina-anchoring one.J Invest Dermatol. 1993; 92: 122-125Google Scholar). When the nylon mesh was seeded with fibroblasts alone, microfibrils were seen by TEM only in proximity to fibroblasts or collagen fibrils in single, randomly arranged units or bundles (Fleischmajer et al., 1991Fleischmajer R. Contard P. Schwartz E. MacDonald E.D. Jacobs L. Sakai L. Elastin-associated microfibrils (10 nm) in a three-dimensional fibroblast culture.J Invest Dermatol. 1991; 97: 638-643Abstract Full Text PDF PubMed Google Scholar). The addition of keratinocytes (5 wk coculture after 4 wk culture of fibroblasts) allowed an organization of fibrillin in a banded distribution in the dermo–epidermal junction and a reticular network otherwise without distribution of fibrillin-1 perpendicular to the DEJ (Fleischmajer et al., 1993Fleischmajer R. MacDonald E.D. Contard P. Perlish J.S. Immunochemistry of a keratinocyte-fibroblast co-culture model for reconstruction of human skin.J Histochem Cytochem. 1993; 41: 1359-1366Crossref PubMed Scopus (66) Google Scholar); however, in this model, amorphous elastin was never detected. These results show that the presence of keratinocytes is not sufficient to trigger elastin deposition on microfibrils. We suggest that the macromolecular components and the porous structure of our DS could trigger the migration of fibroblasts, the oriented deposition of ECM synthesized by these fibroblasts and finally the deposition of elastin on microfibrils and keratinocytes are required for this organization. Collagen and glycosaminoglycans, two components of DS, are known to increase the synthesis of ECM by fibroblasts. The influence of the porous structure of our model on the cellular and ECM organization was also demonstrated in transplantation experiment (Berthod et al., 1999Berthod F. Germain L. Li H. Xu W. Damour O. Auger F.A. Collagen fibril network and elastic system remodeling in a reconstructed skin transplanted on nude mice. 1999Google Scholar) or in clinical assays. Indeed, the porous collagen-GAG-chitosan DS has been used in the treatment of extensive burns, and has provided better aesthetic results (Yannas and Burke, 1980Yannas I. Burke J.F. Design of an artificial skin. I. Basic design principles.J Biomed Mater Res. 1980; 14: 65-81Crossref PubMed Scopus (897) Google Scholar;Damour et al., 1994Damour O. Gueugniaud P.Y. Berthin-Maghit M. Rousselle P. Berthod F. Sahuc F. Collombel C. a dermal substrate made of collagen-GAG-chitosan for deep burn coverage: first clinical use.Clin Mat. 1994; 15: 273-276Crossref PubMed Scopus (61) Google Scholar) and better elasticity (data not shown). In summary, these data suggest that keratinocytes and the collagen-GAG-chitosan porous structure allows the neosynthesis and organization of ECM, as well as influencing the maturation of elastic fibers and their organization similarly to the skin elastin network. DE and SE are useful tools towards a better understanding of elastic network maturation and wound healing. We thank David Hulmes for his advice on the manuscript. This work was supported by grants from Direction de la Recherche et de la Technologie, Center National de la Recherche Scientifique, and Hospices Civils de Lyon. We thank Daniel Hartmann for generously giving elastin antibodies.