Department of Dermatology, Helsinki University Central Hospital, Snelbnanink. 14,S.F-00170 Helsinki 17 and ^Department of Pathology and ^Department of Virology,University of Helsinki, Haartmanink. 3, SF-00290 Helsinki 29, Finland
SUMMARY
Pure adult human keratinocyte cultures were raised from suction-blister roof epidermis andcultured in MCDB-151 medium. In primary culture the epidermal cells rapidly adhered,spread and began to proliferate on collagen-coated growth substrata but not on uncoatedplastic or glass substrata. A fibrillar keratin-specific fluorescence, showing a typical cell-cellarrangement, was seen in all cells in indirect immunofluorescence microscopy, whereas onlysome cells also showed vimentin-specific staining. A fine fibrillar fibronectin-specific surfacestaining was seen at the margin of attaching cells and in marginal cells of spreading cell islands,whereas no fluorescence could be seen in epidermal cells, with antibodies against type IVcollagen or laminin. Interestingly, the marginal cells also showed intracellular fibronectin. Thesynthesis of fibronectin in epidermal cell cultures could also be revealed by metabolic labellingexperiments with P5S]methionine.
In contrast to primary cultures, subcultivated keratinocytes also adhered to uncoated plasticand glass substrata. After subcultivation, keratin and surface fibronectin distribution remainedunaltered but after some subcultivations, most of the cells also showed fibrillar vimentin andexpressed fibronectin intracellularly.
The results show that the suction-blister method provides an easy way to obtain pureepidermal cell cultures without contaminating mesenchymal cells. Our results also suggest adirect role for fibronectin but not for collagen type IV or laminin in adhesion and spreadingof epidermal cells m vitro.
INTRODUCTION
Mouse and human keratinocytes have been widely used to study proliferation andkeratinization of epidermal cells in culture and different techniques have been devel-oped to isolate and establish keratinocyte cultures without dermal contamination(Green, 1979; Prunieras, 1979; Yuspa, Hawley-Nelson, Stanley & Hennings, 1980).For the time being, epidermal cells have been most successfully cultured on lethallyirradiated mouse 3T3 cells as a feeder layer (e.g. see Rheinwald & Green, 1975).Recently, several techniques have also been elaborated to culture keratinocytes withoutfeeder cells, e.g. on collagen- or fibronectin-coated growth substrata (Gilchrest,
• Author for correspondence.
5<D A.-L. Kariniemi, V.-P. Lehto, T. Vartio and I. Virtanen
Nemore & Maciag, 1980; Hawley-Nelson et al. 1980; Liu & Karasek, 1978), or inspecial media with either low pH or low calcium content (Eisinger et al. 1979;Hennings et al. 1980).
Newborn foreskin and whole-skin explants from surgical operations and autopsieshave usually been the sources of human keratinocytes used in cultures. In thesestudies epidermal cell suspensions have been made from enzymically separatedepidermis. However, these methods often result in the loosening of dermal fibroblastsand their appearance in epidermal cell cultures.
The suction-blister method enables separation of the epidermis from the dermisin vivo (Kiistala, 1968). The roofs of the suction blisters are composed of pureepidermal cells lacking any dermal contamination (Kiistala & Mustakallio, 1967). Inthe present study we have used the suction-blister roof epidermis to initiate pureepidermal cell cultures subcultivable up to five times. These cultures show typicalorganization of cytokeratin and deposit fibronectin but not basement membranecollagen, type IV, or the basement membrane glycoprotein, laminin.
MATERIALS AND METHODS
Epidermal cell cultures
Cultures of suction-blister roof keratinocytes. The skin donors were patients and staff from theDepartment of Dermatology, Helsinki University Central Hospital with a mean age of 36 years(in the range from 18 to 60 years). The suction blisters were generated with a special suctiondevice (Kiistala, 1968) on healthy abdominal skin; 24-36 blisters, 5 mm in diameter, from eachvolunteer. The roofs of the blisters were cut off with scissors and incubated in phosphate-buffered saline (PBS) containing 025 % trypsin (Difco 11250) and 0-02 % EDTA, at 37 °C for30 min. Trypsin was inactivated by adding 20 % foetal bovine serum (FBS) and the stratumcorneum sheets were removed. Thereafter, the cells were pelleted by low-speed centrifugationand dispersed in MCDB-151 medium (Peehl & Ham, 1980), supplemented with 10% FBS,hydrocortisone (io/*g/ml, Upjohn), penicillin and streptomycin. Antimycotics were not used.The cells were plated at a density of 1 x io8 to 2 x io6 cells per cm1 on collagen-coated dishesand incubated at 37 °C in water-saturated air containing 5 % CO,. In order to remove deadand non-adherent cells the medium was changed after 24 h and thereafter every 3 days. Forsome experiments the cultures were labelled with L-[35S]methionine (1000 Ci/mmol, Amer-sham, U.K.) at a concentration of 10 fiCi/ml for 48 h in minimal essential medium supple-mented with 0-2 % bovine albumin. The culture medium was centrifuged and then eitherprecipitated with ammonium sulphate (i76mg/ml) for analysis of the secreted proteins, ortreated with agelatin-Sepharose (Engvall & Ruoslahti, 1977) to detect cell-derived fibronectin inthe culture medium. Polyacrylamide gel electrophoresis in the presence of SDS was doneaccording to Laemmli (1970), using 5 % slab gels. For fluorography the gells were processedand exposed to Kodak X-Omat film as described by Bonner & Laskey (1974).
Epidermal cell cultures from whole skin. To separate the epidermis from the dermis, smallpieces of whole skin obtained with a keratotome, were incubated in PBS containing 0-25 %trypsin and 002% EDTA at 37 °C for 2-2-5 h. The epidermis was removed and incubated infresh trypsin—EDTA solution for 5-10 min. The dispersed epidermal cells were cultured as thesuction -blister roof cells.
Subcultivation of epidermal cells. For subcultivation, 2 to 5-week-old confluent cultures wereincubated in 0-25 % trypsin and 002 % EDTA in PBS for 10 min. The detached cells weredispersed in cultured medium and plated on either collagen-coated or uncoated plastic dishes.
Collagen-coated dishes. Sterile collagen suspension was prepared by dissolving calf skincollagen (Calbiochem, La Jolla, U.S.A.) at a concentration of 1 mg/ml in diluted (1:500)
Organization of pure adult human keratinocytes
Fig. i. Keratinocytes isolated from suction-blister roof epidermis and allowed tospread, and cultured on collagen-coated growth substratum for 4 h (A, C) and for 4 days(B), and on glass coverslips for 4 h (D). Note that cells grown on collagen spreadrapidly (A), and form dense cell sheets (B) with typical leading lamella-like structuresin the marginal cells (arrows in c), whereas cells grown on glass fail to spread (D).Human keratinocytes from suction-blister roofs at their third subculture on glasscoverslips for 4 h (E) and for 7 days (F) after trypsinization and plating. Note the lead-ing lamella-like structures on spreading cells in E and the dense cell sheet in F.Phase-contrast: X400.
52 A.-L. Kariniemi, V.-P. Lehto, T. Vartio and I. Virtanen
C
Organization of pure adult human keratinocytes 53
acetic acid. Two to five ml of collagen suspension was dried at 50 °C overnight on culturedishes, which were then washed with PBS before use.
Indirect immunofluorescence microscopy and tight microscopy
For indirect immunofluorescence microscopy (IIF) the cells were grown on either uncoated(subcultures) or collagen-coated (primary culture) coverslips. The coverslips were fixed eitherin methanol at — 20 °C for 10 min or in 3-5 % paraformaldehyde buffered with o-i M-phosphatebuffer (pH 7-2) for 30 min. Rabbit antibodies against vimentin, the subunit protein of fibroblastintermediate filaments (Virtanen et al. 1981a), and against keratin polypeptides, isolated fromhuman plantar epidermis (Keski-Oja, Lehto & Virtanen, 1981; Virtanen et al. 1981a, b), wereprepared and used as described in detail earlier. Rabbit antibodies against fibronectin werekindly provided by Dr A. Vaheri (Hedman, Vaheri & Wartiovaara, 1978; Department ofVirology, University of Helsinki), and FITC- and tetramethyl rhodamine isothiocyanate-(TRITC)-coupled goat anti-human fibronectin antibodies were from Cappel Laboratories(Cochraneville, U.S.A.). Rabbit and sheep antibodies against type IV collagen and lamininhave been described elsewhere in detail (Foidart et al. 1980; Foidart & Reddi, 1980). Theseantibodies were kindly provided by Dr J. M. Foidart (Liege, Belgium). For IIF microscopythe specimens were first exposed to the rabbit or goat antibodies and then to the FITC-coupledsecond antibody. FITC-coupled sheep anti-rabbit immunoglobulin G (IgG) and FITC-coupled rabbit anti-sheep IgG antisera were from Cappel Laboratories. The specimens weremounted in Na-veronal-glycerol buffer (pH 8-5). A Zeiss Universal microscope equipped withan epi-illuminator III RS, phase-contrast optics and filters for FITC fluorescence was used.
RESULTS
Morphology of the cultures
During the first day of culture, 40-80% of the keratinocytes attached on thecollagen substratum as small aggregates that spread within 2-5 days (Fig. IA, C).Thereafter the cells formed colonies that grew to confluency within 2-3 weeks(Fig. IB). If the cells were plated on uncoated plastic or glass substrata, only 10-20%of them adhered without spreading (Fig. 1 D). NO fibroblasts were seen in culturesderived from suction-blister roof epidermis.
Interestingly, keratinocytes from young persons attached and grew better than those
Fig. 2. Keratin (A, B; E-I) and vimentin (c, D; H-J) distribution in dispersed keratin-ocytes, in primary cultures from suction-blister roof epidermis (A-H), and in culturesderived from whole skin (1, j). Note the bright homogeneous keratin-specific'fluores-cence in suspended cells (A, B) and the distinct vimentin-specific fluorescence seenin one of the cells in D.
In primary culture, all cells (E, F; cultured for 5 days) show a bright keratin-specificfluorescence (E, F). The keratin fibrils often appear to align from cell-to-cell (arrowin F). In multilayered epidermal cell sheets the central cells show a more homogeneouskeratin-specific fluorescence in IIF with the typical gap of fluorescence between thecells (arrow in the inset to F). In double IIF with anti-keratin (c) and anti-vimentin(H) antibodies a bright fibrillar keratin-specific staining is seen in all cells, whereasonly one of the cells shows also vimentin-specific staining (arrows in G). Note thatthe vimentin-specific fluorescence does not extend throughout the whole cytoplasmbut is confined to part of the cell (G, H). In cultures derived from thin layers of wholeepidermis both keratin-positive epidermal cells (1) and cells positive only for vimentin0), i.e. fibroblasts (marked with arrows also in 1), are seen in double immunofluores-cence. x 400.
54 A.-L. Karimemi, V.-P. Lehto, T. Vartio and I. Virtanen
from older persons. In the cultures with a high primary attachment rate, the cells weresmall and polygonal, whereas in cultures with a low attachment rate the cells lookedmore spindle-shaped and formed long projections connecting the colonies. In theformer type of cultures the cells grew earlier to confluency than in the latter. Inconfluent cultures grown for 4 weeks about half of the cells were large and partially orfully keratinized. Upon subcultivation, 80-90 % of the cells attached in a few hoursonto both the collagen substratum and uncoated plastic dishes, spread rapidly andbegan to proliferate (Fig. 1 E, F). The keratinocytes were successfully subcultivated upto five times.
Cytoskeletal organization of keratinocyte cultures
Suspended cells obtained from suction-blister epidermis showed a bright keratin-specific fluorescence in IIF (Fig. 2 A, B). Interestingly, in some cells a bright vimentin-specific fluorescence could also be seen (Fig. 2C, D). In primary cultures, a fibrillarkeratin-specific staining could often be seen to be arranged in a cell-cell pattern
Fig. 3. Subcultured epidermal cells (third passage), show an apparently unalteredkeratin distribution in IIF (A). Note the narrow gap in the fluorescence between thekeratin fibres ot adjacent cells (arrows in A). Most cells are also brightly stained forvimentin at this stage of culture (B). Note that the cells now show fibrillar vimentinthroughout their cytoplasm (B). X400.
(Fig. 2E, F), as described earlier for a variety of cultured epithelial cells (Franke et al.1978; Virtanen et al. 1981 a). On the other hand, a more homogeneous keratin-specificfluorescence could be seen in the central cells in cell islands (Fig. 2F, inset). Some ofthe marginal keratin-containing cells also showed a bright fibrillar vimentin-specificstaining in double IIF, although most of the cells were distinctly negative (Fig. 2G, H)and phase-contrast microscopy showed no contamination of fibroblast-like cells. Incontrast to cultured suction-blister keratinocytes, primary cultures from whole-skinpreparations always contained spindle-shaped fibroblastoid cells as well, which couldonly be stained with anti-vimentin antibodies in double IIF (Fig. 21, j). The vimentin-specific staining seen in such contaminating fibroblasts, with a distinct nuclear gap,was distinctly different to that seen in some keratinocytes in cultures derived from
Organization of pure adult human keratinocytes 55A
Fig. 4. Adhering (4 h) keratinocyte3 from primary cultures (A, B) and from the thirdsubculture (c, D), and marginal cells from cell islands (E, primary culture; F, thirdsubculture), surface-stained for fibronectin in I IF. Note the fibrillar staining seenunder one of the cells in B and under both adhering cells in D. Also the marginal cellsin both E and F show distinct fibrillar fibronectin-specific staining. In primarycultures also, the central cells in epidermal cell sheets occasionally show dots and finefibrils of fibronectin (c). A-G, X 700.
suction blisters (cf. Figs. 2H and j). In proliferating keratinocyte subcultures thekeratin-specific fibrillar staining resembled that seen in cells from primary culture(Fig. 3 A). However, after some subcultivations most of the keratinocytes appeared toexpress fibrillar vimentin also (Fig. 3 B).
Pericellular matrix of cultured human keratinocytes
After surface staining, suspended epidermal cells were negative for fibronectin, typeIV collagen and laminin in IIF (data not shown). Instead, both adhering epidermal(Fig. 4A-D) cells and the marginal cells in cell islands of both primary and sub-cultured cultures showed a fibrillar fibronectin-specific fluorescence (Fig. 4E, F). Faintfibronectin-positive fibrils could also occasionally be seen on the surface of central
56 A.-L. Karimemi, V.-P. Lehto, T. Vartio and I. Virtanen
a b c
94-68-
43-
Fig. s. Polypeptides secreted by primary cultures of keratinocytes metabolicallylabelled with [MS]methionine and analysed by gel electrophoresis. a. Ammoniumsulphate precipitation of total polypeptides secreted within 24 h; b, polypeptidesbound to Sepharose 4B alone; c, polypeptides bound to gelatin-Sepharose 4B. Notethe prominent 220 x 10* molecular weight polypcptide both in total secreted polypep-tides (arrowhead in lane a) and bound to gelatin-Sepharose (lane c). Molecular weightsare shown on the left ( x io~').
cells in cell islands (Fig. 4G). However, after both surface and intracellular staining,all cultures were negative in IIF for both type IV collagen and laminin (results notshown). In polyacrylamide gel electrophoresis of metabolically labelled cells severalpolypeptides were revealed in the culture medium of keratinocytes, including a distinctpolypeptide with an apparent molecular weight of 220000 (Fig. 5, lane a). Thispolypeptide could be identified as fibronectin by its binding to gelatin-Sepharose(Fig. 5, lane c). Interestingly, in primary culture only some of the cells, mainly themarginal cells in cell islands, showed fibronectin-specific staining intracellularly(Fig. 6A, B). In double IIF at least some of these cells also appeared to show vimentin-specific staining (Fig. 6c, D). In subcultured keratinocyte cultures, on the other hand,most of the cells also displayed intracellular fibronectin (Fig. 6E) as well as vimentin(see above).
DISCUSSION
In the present study we have shown that pure epidermal cell cultures can beestablished from suction-blister epidermal roofs when cultured on collagen-coatedgrowth substrata.
Organization of pure adult human keratinocytes 57
B
Fig. 6. Intracellular fibronectin in I IF of permeabilized cells from a primary culture(A, B), in double IIF with vimentin in a primary culture (c, D), and in central cellsfrom a culture at third passage (E). Note that only the marginal cells are distinctlyfibronectin-positive in A, B. Some of the fibronectin-positive marginal cells (arrow inc) appear to contain vimentin also, in double IIF (arrow in n). In subcultured kera-tinocytes most of the central cells in cell islands also contain intracellular fibronectin(E). X500.
At the ultrastructural level, the suction blisters appear to be formed at the dermal-epidermal junction area, subepidermally at the space between the basal lamina andthe basal epidermal cells (Kiistala & Mustakallio, 1967). The blister roofs are thereforecomposed of epidermal cells - keratinocytes, melanocytes and Langerhans' cells - butlack dermal cells. In addition, the suction procedure does not appear to damage theepidermal cells: autoradiography of suspended suction-blister roof cells has revealedcells labelled to the same extent as normal epidermis in vivo (Kariniemi, Kousa &Asko-Seljavaara, 1981).
Earlier studies have stressed the importance of basal membrane collagen (type IVcollagen) in the attachment and spreading of human and mouse keratinocytes in vitro(Kleinman, Klebe & Martin, 1981; Kleinman et al. 1978; Murray et al. 1979). An
$ CEL 58
58 A.-L. Kariniemi, V.-P. Lehto, T. Vartio and I. Virtanen
important role has also been assigned recently to laminin (Terranova, Rohrbach &Martin, 1981), a non-collagenous basal lamina glycoprotein (cf. Foidart et al. 1980;Timpl et al. 1979), in the adhesion process of epidermal cells to their growth sub-stratum. Such a role for laminin has also recently been described for other epithelialcells in vitro (Kleinman et al. 1981). The role of fibronectin in the adhesion of keratin-ocytes, on the other hand, has remained controversial (Federgreen & Stenn, 1980;Gilchrest et al. 1980).
Our results show that human keratinocytes in primary culture rapidly adhere andspread on substrata made of acid-soluble collagen (types I and III) but fail to adhereon either uncoated glass or plastic substrata. Interestingly, in our study the spreadingand organizing keratinocytes showed both production and deposition of fibronectinbut not of either type IV collagen or laminin. In IIF the cell surface-associatedfibronectin in primary cultures was mainly located at the edges of cell islands, com-posed of several cells, and intracellular fibronectin was mostly in the marginal cells.These results are in contrast to recent results of Federgreen & Stenn (1980) and clearlyalso suggest a role for fibronectin in epidermal cell attachment and spreading, in linewith results on several other cell types (Vaheri & Mosher, 1978; Yamada & Olden,1978). Gilchrest et al. (1980) have also shown recently that human keratinocytes canbe grown efficiently on fibronectin-coated growth substratum. Interestingly, sub-cultured epidermal cells could also be plated efficiently on glass and plastic substratawithout exogenous collagen. It may be that subcultured keratinocytes are capable ofproducing enough endogenous fibronectin for their spreading (cf. Grinnell & Field,1979) and therefore do not require exogenous macromolecules for adhesion. Sub-cultured epidermal cells did not show detectable deposition of either type IV collagenor laminin.
Fibronectin has been detected both in the dermis and in the dermal-epidermaljunction of the skin (Couchman et al. 1979; Foidart & Yaar, 1981; Fyrand, 1979),whereas both type IV collagen and laminin have been found only in the basal laminaregion of the dermal—epidermal junction (Foidart & Yaar, 1981). However, little isstill known about the origin of these basal lamina components in skin tissue, althoughboth basal epidermal cells and dermal fibroblasts have been suggested to be the sourceof fibronectin in dermal-epidermal junctions (Couchman et al. 1979). Since fibro-blasts in vitro do not produce type IV collagen or laminin (cf. Foidart et al. 1980;Kleinman et al. 1981; Timpl et al. 1979), they could be derived from basal epidermalcells (cf. also Foidart & Yaar, 1981). Unexpectedly, however, human epidermal cellsin our study did not show any deposition of basal lamina components, type IV collagenor laminin. This is in accordance with recent findings, which show that pure epidermalcell cultures do not form ultrastructurally detectable basal lamina structure (Woodley,Regnioz & Prunieras, 1980) and lack typical epidermal glycosaminoglycans (King &Fabiowa, 1980) if cultured in the absence of dermal cell component.
Recently, Breitkreutz, Boukamp, Lueder & Fusenig (1981) and Franke et al.(1979 a) have suggested that primary cultures of epidermal cells contain only pre-keratin as an intermediate filament protein. However, during adaptation to tissue-culture conditions the keratinocytes also appeared to express vimentin-type intermediate
Organization of pure adult human keratinocytes 59
filaments (Breitkreutz et al. 1981; Franke et al. 1979a) and the fibroblast-type ofintermediate filament protein (Franke et al. 19796). We have shown a similar pheno-menon in cultures of human amnion epithelial cells, which also contained cells withboth prekeratin and vimentin during primary culture (Virtanen et al. 1981a). Wesuggested that an increased number of cells containing both vimentin and keratincould be due to a more rapid proliferation of cells initially containing both types ofintermediate filament proteins. This suggestion is also supported by the results of thepresent study, which show that both suspended and primary cultured keratinocyteshomogeneously express keratin but a few cells also express vimentin. In subculturesmost of the epidermal cells appeared to express both types of intermediate filamentproteins.
The suction-blister method is suitable for experimental dermatological studies,because the suction procedure does not cause pain and the blisters do not leave scarson the test subjects. This method has been used widely to study components of blisterfluid in normal skin and in inflammatory skin diseases. Cultured human keratinocytesobtained from blister roof epidermis also appear to be ideal for studies of humanepidermal cell characteristics in vitro and do not require any further treatment toabolish contaminating dermal cells (cf. Gusterson, Edwards, Foster & Neville, 1981;Jensen & Therkelsen, 1981).
The skilful technical assistance of Ms Leena Marjanen, Ms Raili Taavela and Miss PipsaKaipainen is acknowledged. This study was supported by the Finnish Medical ResearchCouncil, the Sigrid Juselius Foundation and the Association of Finnish Life InsuranceCompanies.
REFERENCESBONNER, W. M. & LASKEY, R. A. (1974). A film detection method for tritium labelled proteins
and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46, 83-88.BREITKREUTZ, D., BOUKAMP, P., LUEDEH, M. & FUSENIG, N. E. (1981). Morphological and
biochemical criteria for keratinization in primary and permanent mouse epidermal cellcultures. Front Matrix Biol. 9, 57-82.
COUCHMAN, J. R., GIBSON, W. T., THORN, D., WEAVER, A. C, REES, D. A. & PARISH, W. E.(1979). Fibronectin distribution in epithelial and associated tissues of the rat. Archs Derm.Res. 366, 295-310.
EISINCEH, M., Soo LEE J., HEFTON, J. M., DARZYNKIEWICZ, Z., CHIAO, J. W. & DE HARVEN, E.(1979). Human epidermal cell cultures: Growth and differentiation in the absence of dermalcomponents or medium supplements. Proc. natn. Acad. Sci. U.S.A. 76, 5340-5344.
ENGVALL, E. & RUOSLAHTI, E. (1977). Binding of soluble form of fibroblast surface protein,fibronectin, to collagen. Int. J. Cancer 20, 1-5.
FEDERGREEN, W. & STENN, K. S. (1980). Fibronectin (LETS) does not support epithelial cellspreading. .7. invest. Derm. 75, 261-263.
FOIDART, J. M., BERE, E. J., YAAR, M., RENNARD, S. I., GULLINO, M., MARTIN, G. R. &ICATZ, S. I. (1980). Distribution and immunoelectron microscopic localization of laminin, anoncollagenous basement membrane glycoprotein. Lab. Invest. 42, 336-342.
FOIDART, J. M. & REDDI, A. H. (1980). Immunofluorescent localization of type IV collagenand laminin during endochornal bone differentiation and regulation by pituitary growthhormone. Devi Biol. 75, 130-136.
FOIDART, J. M. & YAAR, M. (1981). Type IV collagen, laminin and fibronectin at the dermo-epidermal junction. Front Matrix Biol. 9, 175-188.
3-3
60 A.-L. Karinietm, V.-P. Lehto, T. Vartio and I. Virtanen
FRANKE, W. W., SCHMID, E., BREITKREUTZ, D.( LODER, M., BOUKAMP, P., FUSENIG, N. E.,OSBORN, M. & WEBER, K. (1979a). Simultaneous expression of two different types of inter-mediate sized filaments in mouse keratinocytes proliferating in vitro. Differentiation 14, 35-50.
FRANKE, W. W., SCHMID, W., WINTER, S., OSBORN, M. & WEBER, K. (19796). Widespreadoccurrence of intermediate-sized filaments of the vimentin-type in cultured cells from diversevertebrates. Expl Cell Res. 123, 25-46.
FRANKE, W. W., WEBER, K., OSBORN, M., SCHMID, E. & FREUDENSTEIN, C. (1978). Antibodyto prekeratin. Decoration of tonofilament-like arrays in various cells of epithelial character.Expl Cell Res. 116, 420-445.
FYRAND, O. (1979). Studies on fibronecrin in the skin. Br. J. Derm. 101, 263-269.GILCHREST, B. A., NEMORE, R. E. & MACIAG, T. (1980). Growth of human keratinocytes on
fibronectin-coated plates. Cell Biol. Int. Rep. 4 (no. 11), 1009-1016.GREEN, H. (1979). The kerarinocyte as differentiated cell type. The Harvey Lectures, series 74,
pp. 101-139. New York: Academic Press.GRINNELL, F. & FIELD, M. K. (1979). Initial adhesion of human fibroblasta in serum-free
medium: possible role of secreted fibronectin. Cell 17, 117-128.GUSTERSON, B. A., EDWARDS, P. A. W., FOSTER, C. S. & NEVILLE, A. M. (1981). The selective
culture of keratinocytes using a cytotoxic antifibroblast monoclonal antibody. Br. J. Derm.105, 273-277.
HAWLEY-NELSON, P., SULLIVAN, J. E., KUNG, M., HENNINGS, H. & YUSPA, S. H. (1980).Optimized conditions for the growth of human epidermal cells in culture. J. invest. Derm.75, 176-182.
HEDMAN, K., VAHERI, A. & WARTIOVAARA, J. (1978). External fibronectin of cultured humanfibroblasts is predominantly a matrix protein. J. Cell Biol. 76, 748-760.
HENNINGS, H., MICHAEL, D., CHENG, C, STEINERT, P., HOLBROOK, K. & YUSPA, S. H. (1980).Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell19, 245-254-
JENSEN, P. K. A. & THERKELESEN, A. J. (1981). Cultivation at low temperature as a measure toprevent contamination with fibroblasts in epithelial cultures from human skin. J. invest.Derm. 77, 210-212.
KARINIEMI, A.-L., KOUSA, M. & ASKO-SELJAVAARA, S. (1981). Dissociation of suction blisterroof epidermis with trypsin and desoxyribonuclease into viable single cells. Acta derm.Stockh. 61, 58-61.
KE8KI-OJA, J., LEHTO, V.-P. & VIRTANEN, I. (1981). Keratin filaments in mouse epithelial cellsare rapidly affected by epidermal growth factor. J. Cell Biol. 90, 537-541.
KIISTALA, U. (1968). Suction blister device for separation of viable epidermis from dermis.J. invest. Derm. 50, 129—137.
KIISTALA, U. & MUSTAKALLIO, K. K. (1967). Dermoepidermal separation with suction. Electronmicroscopic and histochemical study of initial events of blistering of human skin. J. invest.Derm. 48, 466-477.
KING, I. A. & FABIOWA, A. (1980). The dermis is required for the synthesis of extracellularglycosaminoglycans in cultured pig epidermis. Biochim. biophys. Acta 63a, 234-243.
KLEINMAN, H. K., KLEBE, R. J. & MARTIN, G. R. (1981). Role of collagenous matrices in theadhesion and growth of cells. J. Cell Biol. 88, 473-485.
KLEINMAN, H. K., MURRAY, J. C , ERMONA, B., MCGOODWIN, B. S. & MARTIN, G. R. (1978).Connective tissue structure: cell binding to collagen. J'. invest. Derm. 71, 9-11.
LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature, Lond. az7, 680-685.
Liu, S. C. & KARASEK, M. (1978). Isolation and growth of adult human epidermal keratinocytesin cell culture. J. invest. Derm. 71, 157-162.
MURRAY, J. C , STINGL, G., KLEINMAN, H. K., MARTIN, G. R. & KATZ, S. I. (1979). Epidermalcells adhere preferentially to type IV (basement membrane) collagen. J. Cell Biol. 80,197-202.
PEEHL, D. M. & HAM, R. G. (1980). Clonal growth of human keratinocytes with small amountsof dialyzed serum. In Vitro 16, 526-538.
PRUNIERAS, M. (1979). Epidermal cell cultures as models for living epidermis. J. invest. Derm.73. 135-137-
Organization of pure adult human keratinocytes 61
RHEINWALD, J. G. & GREEN, H. (1975). Serial cultivation of strains of human epidermalkeratinocytea: the formation of keratinizing colonies from single cells. Cell 6, 331-344.
TERRANOVA, V. P., ROHRBACH, D. H. & MARTIN, G. R. (1981). Role of laminin in the attach-ments of PAM12 cells to basement membrane collagen. Cell 33, 719-̂ 726.
TIMPL, R., ROHDE, H., GEHRON-ROBEY, P., RENNARD, S., FOIDART, J. M. & MARTIN, G. R.(1979). Laminin - a glycoprotein from basement membranes. J. biol. Chem. 354, 9933-9937-
VAHERI, A. & MOSHER, D. E. (1978). High molecular weight cell surface glycoprotein (fibro-nectin) lost in malignant transformation. Biochim. biopkys. Acta 516, 1-25.
VIRTANEN, I., LEHTO, V.-P., LEHTONEN, E., VARTIO, T., STENMAN, S., KURKI, P., WAGER, O.,SMALL, J. V., DAHL, D. & BADLEY, R. A. (1981a). Expression of intermediate filaments incultured cells. J. Cell Set. 50, 45-63.
VIRTANEN, I., VON KOSKULL, H., LEHTO, V.-P., VARTIO, T. & AULA, P. (19816). Culturedhuman amniotic fluid cells characterized with antibodies against intermediate filaments inindirect immunofluorescence microscopy.^, clin. Invest. 68, 1348-1355.
WOODLEY, D., REGNIOZ, M. & PRUNIERAS, M. (1980). In vitro basal lamina formation mayrequire non-epidermal living substrata. Br. J. Derm. 103, 397-404.
YAMADA, K. M. & OLDEN, K. (1978). Fibronectins - adhesive glycoproteins of cell surface andblood. Nature, Lend. 375, 179-184.
YUSPA, S. H., HAWLEY-NELSON, P., STANLEY, J. R. & HENNINGS, H. (1980). Epidermal cellculture. Transplantn PTOC. 13, 114-122.
