0022-202X/81/7703-0283$02.00/ 0 THE JOUHNAL OF INVESTIGATIVE DEHMATOLOGY, 77:283-286,1981 Vol. 77, No.3 Copyright © 1981 by The Williams & Wilkins Co. P"':l1.led ill U.S.A. Basal Lamina Components in Experimentally Induced Skin Blisters

OLLI SAKSELA, M.D., KARl ALITALO, M.D., URPO KnSTALA, M.D., AND ANTTI V AHERI, M.D. Departmen.t of Virology, University of Helsinki, Helsinki, Finlan.d (OS, KA, and A Vi , Department of Dermatology, Central MilitalY Hospital, Helsinki, Finland (UK)

The distribution of basement membrane glycopro­ followed techniques described earlier [10]. Antibodies to lalllinin and teins, fibronectin, laminin, and type IV collagen was collagen types IV (antigen obtained from human placenta), I and II as studied in experimentally induced skin blisters in which well as to procollagen types I and III were those used in earlier studies the epidermis is separated from the dermis through the [11; kindly provided by Dr. Rupert Timpl]. The specificities of the lamina lucida part of basal lamina. Fibronectin was antisera and lack of their cross-reactions between the vaJ'ious antigens have been published in om previous reports [4,12-14]. For the blocking found surrounding the blister cavity and in a primary tests 50 fLg / ml of the pmified proteins were incubated with the respec­ covering formed on the bottom of the blister. Neither tive antibodies or antisera at the working dilutions overnight in the laminin nor type IV collagen were incorporated into this cold and clarified by centrifugation (12,BOO x g for 15 min). Control clot-like matrix. The layer of laminin found in normal sections were treated with nonilllmune rabbit serum. epidermal-dermal junction was found to be cleaved by the blistering method, whereas all type IV collagen an­ Immunofluorescence Stainin.g tigenicity was confined to the base of the blister presum­ Unfixed frozen sections from blisters of various ages were dried and ably due to its location in the lamina densa part of the treated with antibodies (10-30 f,tg/ ml, diluted in phosphate-buffered basal lamina. No interstitial collagens were found in the saline) for 30 min at room temperat.me, washed in saline and labeled separated epidermis in accordance with previous ultra­ with fluorescein-isothiocyanate-conjugated sheep anti-rabbit. globulin structural evidence. These studies support the proposed (FITC) (Wellcome, Beckenham, England; diluted 1:20 in PBS) for 30 role of laminin in the adhesion of epidermal cells to min at room telllperatme. After 3 washes each for 10 min with saline, basement membrane collagen. the coverslips were mounted wet in a solution containing equal volumes of glycerol and Veronal-buffered (100 mM, pH B.6) 0.15 M sodium chloride.

Ultrastructurally, the epidermal-dermal junction can be di­ Microscopy vided into 4 layers: (1) the basal cell plasma membrane, (II) the lamina lucida, (III) the lamina densa, and (IV) the subbasal Stained sections were examined with a Zeiss Universal microscope equipped with a II RS epi-iUuminator (Carl Zeiss, Oberkochen, Ger­ lamina fibrous elements, anchoring fibrils, dermal microfibril many) and a high-pressure mercury lamp (HPO, 200 W) . The filters bundles, and collagen fibers [for review see references 1, 2]. used for specifi c FITC fluorescence were excitation filters BP 455-490, Known proteins in the basal lamina of the junction are type IV dichroic mirror FT 510, and barrier ftlters LP 520. All sections were collagen, laminin, fibronectin, and possibly type V collagen [3- tested for autofluorescence by the intense green emission light (545 7]. The junction serves to anchor epidermis to the dermis and nm) of the mercury lamp for fluorescence excitation together with acts as a barrier to the migration of cells and some large excitation filter PB 546/ 10, dichroic mirror FT 5BO, and barrier ftiter molecules across the junction. It obviously also plays an impor­ LP 590. Autofluorescence was intensely red and could be distinguished tant role in the differentiation of epidermal keratinocytes. The from the green FITC fluorescence detected with blue emission light. role of the component glycoproteins in the ultrastructw'e and Slight nonspecific fluorescence was observed in the granular layer of epidermis in some specimens (see reference 15). To minimize autoflu­ function of the epidermal-dermal junction is not known. In the orescence, the specimens were stained and photographed without delay. present work we have investigated the distribution of 3 glyco­ proteins of the epidermal-dermal junction in experimentally RESULTS induced suction blisters of adult skin causing detachment of the epidermis-dermis along the lamina lucida [8]. The results con­ The morphology of the induced blister is shown in Fig 1, in firm that laminin is a component of the lamina lucida while which. the clot-like matrix material on the bottom of the blister is also seen. type IV collagen may be confined to the lamina densa compo­ nent of the basal lamina. The results also point to a role for Fibronectin fibronectin as a temporary matrix in wound healing prior to re­ epithelization. Staining for fibronectin was prominent on the inner aspects of the blister cavity (Fig. 2, a, b). In the earliest samples taken MATERIALS AND METHODS after induction of the blisters, fibronectin was ab'eady found in The induction of skin blisters in healthy young volunteer men was a network structw'e covering the denuded base of the blister carried out as described elsewhere [B,9]. Blisters were cut off after local (Fig 2 a). This clot-like structure could be recovered in an intact anesthesia within 1 Ill' after their induction, frozen with liquid nitrogen, form by lifting it from the blister base. Soluble fibronectin along and cut into slices. In some experiments, the samples were taken during with lX2-macroglobulin and other plasma proteins were detected the healing process 1, 2, or 3 days after blistering. in the fluid inside the blister cavity (data not shown). The blocked antifibronectin serum gave no staining of the sections. Antibodies The preparation of antiserum against human plasma fibronectin Laminin In sections of intact adult skin laminin was stained mainly in Manuscript received October 27, 19BO; accepted for publication Feb­ the dermal vessel walls and in the dermal basement membrane rUllJ'Y 23, 19B1. (Fig 2 c). Weak but define staining was also observed in the This work was supported by the Finnish Cancer Foundation, and epidermal part of the disrupted basement membrane (Fig 2 d). the Association of Finnish Life Assurance Companies. Reprint requests to: Dr. OUi Saksela, Department of Vi.rology, Uni­ Laminin staining could not be blocked with purified fibronectin. versity of Helsinki, Haartmaninkatu, 3, SF-00290 Helsinki 29, Finland. Type IV collagen Abbreviations: FITC: fluorescein-isothiocyanate-conjugated sheep anti-rabbit Absence of staining for type IV collagen in the roof of the globulin induced blisters is shown in Fig 2 f. Type IV collagen could

283 284 SAKSELA ET AL Vol. 77, No.3 anyhow be found in the basal part of the cleaved basement membrane (Fig 2 e). Interstitial Collagens Staining for type III pro collagen was obtained in the upper dermis (Fig 2 g) and staining for collagen type I in all parts of dermis (now shown), but no staining for type II collagen was observed either in intact skin or in the blisters (not shown). No major differences were seen in the staining patterns during the healing process.

DISCUSSION In the present paper we demonstrated that suction blistering of the skin, thought to cleave the epidermal-dermal junction

FIG 1. Hematoxylin-eosin-stained section of a blister on abdominal skin of a young adult male volunteer (reduced from x 50).

FIG 2. Suction blisters in indirect immunofluorescence analysis for connective tissue glycoproteins. The figures on the left show the blister base (reduced from x 100) and the figures on the right show the blister roofs at higher magnification (reduced from x 250). The sections were stained with antiserum to fibronectin (1:100 dilution, a, b) and antibodies to laminin (10 /-Lg/ml, c, d), type IV collagen (30 /-Lg/ml, e, f) and type III procollagen (30 /-Lg/ml, g, h). Note that while the antibodies react mostly with the upper portions of skin biochemical evidence does not demonstrate such a gradient of type III procoUagen in skin (for references, see 16) . Sept. 1981 BASAL LAMINA COMPONENTS 285 along its weakest link, cleaves the basal lamina along a laminin­ nantly on the plasma membrane of the basal epidermal cell and containing zone. On the other hand, type IV collagen antigen­ in the adjacent lamina lucida zone and also in the subbasal icity remains confined to the basal lamina structure on the lamina area of the epidermal-dermal junction. These findings bottom of the blisters. Upon accumulation of interstitial tissue fit well to the present results, which show a strong fluorescence fluid, fibronectin is deposited in a lattice-like clot matrix along both on the epidermal and dermal sides of the basement mem­ the base of the blister cavity. brane. Fibronectin has been supposed to play no role in me­ Some laminin antigenicity was observed in the roof of the diating attachment of epithelial cells to basement membrane blisters, in close contact to the surface of the basal cells. collagen [21]. An interesting feature in our induced skin blisters Laminin is known to bind to type IV basement membrane is the rapid dormation of a covering on the bottom of the blister collagen in vitro [17]. On the other hand, epidermal cells, and cavity. In the present immunofluorescence study this matrix specifically the basal ones, adhere better to substrata of type was found to stain heavily with antifibronectin. Since the struc­ IV collagen than to interstitial collagen in vitro [18,19]. This ture is present immediately after blister induction, it is most points to a role for laminin in mediating the adhesion of basal probably not synthesized either by epidermal or dermal cells epidermal cell layer to the underlying basal lamina collagen but is a result of transudation of plasma-derived fibronectin. A type IV, as also concluded by Terranova, Rohrback, and Martin simila.r "prima.ry matrix" containing fibronectin has also been [17]. described in other models of wound healing [22] and connective Since fibronectin is present in large amounts of plasma, it is tissue matrix remodeling [23]. expected to diffuse into the blister fluid. As the soluble and Although separated epidermal cells of the blister remain matrix forms of fibronectin have not been distinguished anti­ viable for several days in vivo, new basement membrane colla­ genically, it is not possible to define the origin of fibronectin gen was not seen to be formed beneath them. This may be localized by immunofluorescence in the basement membrane explained by the finding of Briggaman, Dalldorf, and Wheeler lining of the blister cavity. However, earlier immunoelectron [24] that epidermal cells require an underlying connective tissue microscopy studies [20] have localized fibronectin predomi- matrix to express the functions of basal cells in basal lamina 286 SAKSELA ET AL Vol. 77, No.3 formation. Terminal differentiation in the epidermis of the chern J 183:331-337, 1979 blister roof is also indicated by accumulation of tonofIlaments 11. Alitalo K: Production of both basement membrane and interstitial procollagens by fibroblastic WI-38 cells fi'om human embryonic' in the basal cells as seen in ultrastructural studies. We did not lung. Biochem Biophys Res Commun 93:873-880, 1980 determine whether during the healing process the migrating 12. Nowack H, Gay S, Wick G, Becker U, Timpl R: Preparation and epidermal cells lay down type IV collagen as they apparently use in immunohistology of antibodies specific for type I and type do in tissue culture [7] but are currently characterizing the III collagen and procollagen. J Immunol Meth 12:117-124, 1976 13. Timpl R, Mrutin GR, Bruckner 1', Wick G, Wiedemann H: Nature connective tissue glycoproteins produced by pure populations of the collagenous protein in a tumor basement membrane. Eur of epidermal keratinocytes obtained from the blister roofs when J Biochem 84 :4 3-52, 1978 grown on fibroblastic 3T3 cell feeder layers [Alitalo et aI, 14. Vaheri A, Kurkinen M, Lehto V-I', Linder E, Timpl R: Codistri­ submitted]. bution of pericellular matrix proteins in cultured fibroblasts and loss in transformation: fibronectin and procollagen. Proc Natl We thank Dr. Rupert Timpl for laminin and collagen antibodies, Dr. Acad Sci USA 75:4944-4948, 1978 John Couchman for critical reading of the manuscript, and Ms. Tarja 15. Beutner EH, Jablonska S, Jarzabek, Chorzelska M, Marviejowska Heinio and Ms. Piiivi Lehtovuori for technical assistance. E, Rzesa G, Chorzelski TP: Studies in immunodermatology. CI. IF Studies of autoantibodies to the stratum corneum and of in vivo fixed IgG in stratum corneum of psoriatic lesions. Int Arch REFERENCES Allergy Appl Immunol 48:301-323, 1975 1. Briggaman RA, Wheeler Jr CE: The epidermal-dermal junction. J 16. Timpl R: Antibodies to procollagens and collagens. Meth Enzymol, Invest Dermatol 65:71-84, 1975 in press. 2. Daroczy J, Feldman J, Kiraly K: Human epidermal basal lamina; 17. Terranova VP, Rohrbach DH, Martin GR: The role of laminin in its structUJ'e, connections and function, Biochemistry and Pa­ epidermal cell attachment to basement membrane collagen. Cold thology of Basement Membranes. Edited by AM Robert, R Spring Harbor Symposium on the Biology of the Vascular Cell , Boniface, LR Creteil. 1977, pp 208-233 in press. 3. Yaiota H, Foidart J-M, Katz SI: Localization of the collagenous 18. Murray JC, Stingl G, Kleinman HK, Martin GR, Katz SI: Epider­ component in skin basement membrane. J Invest Dermatol 70: mal cells adhere preferentially to type IV basement membrane 191-193, 1978 collagen. J Cell Bioi 80:107-202, 1979 4. Timpl R, Rohde H, Robey PG, Rennard SI, Foidart J-M, Martin 19. Stanley JR, Foidru·t J-M, Murray JC, Mal·tin GR, Katz SI: The GR: Laminin-A glycoprotein from basement membranes. J Bioi epidermal cell which selectively adheres to a collagen substrate Chern 254:9933-9937, 1979 is the basal cell. J Invest Dermatol 74:54-58, 1980 5. Stenn KS, Madri JA, Roll FJ: Migrating epidermis produces AB2 20. Couchman JR, Gibson WT, Thom D, Weawer AC, Rees DA, Parish collagen and requires continual collagen synthesis for movement. WE: Fibronectin distribution in epithelial and associated tissues Nature 277:229-232, 1979 of the rat. Arch Dermatol Res 266:295-310, 1979 6. Stenman S, Vaheri A: Distribution of a major connective tissue 21. Kleinman HK, Hewitt AT, Grotendorst GR, Martin GR, Munay protein, fibronectin, in normal human tissues. J Exp Med 147: JC, Rohrbach DH, Terranova VP, Rennal'd SI, Val'ner HH, 1054-1064, 1978 Wilkes CM: Role of matrix components in adhesion and growth 7. Hintner H, Fritsch PO, Foidart J -M, Stingl G, Schuler G, Katz S1: of cells, Current Reseru'ch Trends in Craniofacial Development. Expression of basement membrane zone antigens at the dermo­ Edited by RM Pratt. Elsevier North-Holland, N.Y. 1980, in press epibolic junction in organ cultures of human skin. J Invest 22. Kurkinen M, Vaheri A, Roberts PJ, Stenman S: Sequential ap­ Dermatol 74:200-204, 1980 pearance of fibronectin and collagen in experimental granulation 8. Kiistala U, Mustakallio KK: Dermo-epidermal separation with tissue. Lab Invest 43:47-51, 1980 suction. Electron microscopic and histochemical study of initial 23. Weiss RE, Reddi AH: Synthesis and localization of fibronectin events of blistering on human skin. J Invest Dermatol 48:466- during collagenous matrix-mesenchymal cell interaction and dif­ 477, 1967 ferentiation of cartilage and bone in vivo. Proc Nat! Acad Sci 9. Kiistala U~ Suction blister device for separation of viable epidermis USA 77:2074-2978, 1980 from dermis. J Invest Dermatol 50:129-137, 1968 24. Briggaman RA, DalldorfPG, Wheeler Jr. CE: Formation and origin 10. Vuento M, Vaheri A: Purification of fibronectin from human plasma of basal lamina and anchoring fibrils in adult hUJTIan skin. J Cell by affinity chromatography under nondenatUJ'ing conditions. Bio- Bioi 51:384-395, 1971