Portal de Periódicos da CAPES

Epidermal Tight Junctions: ZO-1 and Occludin are Expressed in Mature, Developing, and Affected Skin and In Vitro Differentiating Keratinocytes

2001; Elsevier BV; Volume: 117; Issue: 5 Linguagem: Inglês

10.1046/j.0022-202x.2001.01493.x

1523-1747

Kati Pummi, Maria Malminen, Heikki Aho, Seija-Liisa Karvonen, Juha Peltonen, Sirkku Peltonen,

Oral Health Pathology and Treatment

This study demonstrates the presence of tight junction antigens in adult and developing human epidermis. Indirect immunofluorescence labeling and immunoelectron microscopy with antibodies to ZO-1 and occludin localized tight junction components ZO-1 and occludin to a narrow zone of the granular cells of adult epidermis. Double immunolabeling for tight junction components with adherens junction or desmosome proteins suggested that occludin is more specific for tight junctions than ZO-1, which may also be associated with adherens junctions. In developing skin, tight junctions interconnected the peridermal cells, and after the fetal stratification localized to the granular cell layer. Immunolabeling of psoriasis, lichen planus, and ichthyosis vulgaris, representing aberrant differentiation of the epidermis, showed that these conditions were associated with relocation of ZO-1 and occludin to the spinous cells. Cultures of epidermal keratinocytes, which offer a useful model for the formation of cellular contacts, revealed that tight junction components, ZO-1 and occludin, displayed a marked degree of colocalization relatively late during the process when the fusion zone had assumed a linear appearance. This suggests that the formation of adherens junctions and desmosomes precedes that of tight junctions. We speculate that the epidermal barrier, isolating the human body from the external environment, is in part formed by tight junctions of stratum granulosum. This study demonstrates the presence of tight junction antigens in adult and developing human epidermis. Indirect immunofluorescence labeling and immunoelectron microscopy with antibodies to ZO-1 and occludin localized tight junction components ZO-1 and occludin to a narrow zone of the granular cells of adult epidermis. Double immunolabeling for tight junction components with adherens junction or desmosome proteins suggested that occludin is more specific for tight junctions than ZO-1, which may also be associated with adherens junctions. In developing skin, tight junctions interconnected the peridermal cells, and after the fetal stratification localized to the granular cell layer. Immunolabeling of psoriasis, lichen planus, and ichthyosis vulgaris, representing aberrant differentiation of the epidermis, showed that these conditions were associated with relocation of ZO-1 and occludin to the spinous cells. Cultures of epidermal keratinocytes, which offer a useful model for the formation of cellular contacts, revealed that tight junction components, ZO-1 and occludin, displayed a marked degree of colocalization relatively late during the process when the fusion zone had assumed a linear appearance. This suggests that the formation of adherens junctions and desmosomes precedes that of tight junctions. We speculate that the epidermal barrier, isolating the human body from the external environment, is in part formed by tight junctions of stratum granulosum. estimated gestational age indirect immunofluorescence Tris-buffered saline Simple epithelium, such as intestinal lining, is composed of one layer of epithelial cells. Each cell of simple epithelia forms five distinct types of junctions between neighboring cells or the underlying basement membrane (Alberts et al., 1994Alberts B. Bray D. Lewis J. Raff M. Roberts K. Watson J.D. Cell junctions.in: Molecular Biology of the Cell. 3rd edn. Garland Publishing, New York1994: 950-963Google Scholar). These cellular connections include tight junctions, adherens junctions, desmosomes, gap junctions, and hemidesmosomes. In simple epithelium, the cellular junctions can be identified by electron microscopy on the basis of their characteristic ultrastructural appearance. In contrast, the identification of selected cellular junctions in stratified epithelium has been more complicated; ultrastructural analysis readily identifies desmosomes and hemidesmosomes in all epithelia. The epidermal adherens junctions were characterized as late as 1993–96, however, as electron microscopy analysis alone was not sufficient to identify these junctions. Recognition of molecular markers for adherens junctions and the immunoelectron microscopy approach were the keys in the characterization of epidermal adherens junctions (Kaiser et al., 1993Kaiser H.W. Ness W. Jungblut I. Briggman R.A. Kreysel H.W. O'Keefe E.J. Adherens junctions: demonstration in human epidermis.J Invest Dermatol. 1993; 100: 180-185Abstract Full Text PDF PubMed Google Scholar;Haftek et al., 1996Haftek M. Hansen M.U. Kaiser H.W. Kreysel H.W. Schmitt D. Interkeratinocyte adherens junctions: immunocytochemical visualization of cell-cell junctional structures, distinct from desmosomes, in human epidermis.J Invest Dermatol. 1996; 106: 498-504Crossref PubMed Scopus (43) Google Scholar). To date, ultrastructural analyses have not conclusively identified tight junctions in human epidermis (see below). In simple epithelia, tight junctions form a paracellular permeability barrier regulating the movement of water, solutes, and immune cells. Furthermore, tight junctions segregate cell surface membrane proteins and lipids into the apical and the basolateral membrane domains (for reviews seeStevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita and Furuse, 1999Tsukita S. Furuse M. Occludin and claudins in tight-junction strands: leading or supporting players?.Trends Cell Biol. 1999; 9: 268-273Abstract Full Text Full Text PDF PubMed Scopus (494) Google Scholar,Tsukita and Furuse, 2000Tsukita S. Furuse M. Pores in the wall: claudins constitute tight junction strands containing aqueous pores.J Cell Biol. 2000; 149: 13-16Crossref PubMed Scopus (385) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). In, for example, the intestinal epithelium, tight junctions (zonula occludens) and adherens junctions (zonula adherens) form the most apical component of the lateral junctional complex. Below these two junctions are the spot-like contacts formed by desmosomes and gap junctions (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar). The epidermis represents a stratified epithelium, and it is composed of four cellular layers: basal cell, spinous cell, granular cell, and cornified layers. Unlike in simple epithelium, where each cell expresses all known types of cellular junctions, cells of different layers of the stratified epidermis express only selected types of cellular junctions. Desmosomes and adherens juctions are present in basal, spinous, and granular cell layers, whereas gap junctions are mainly located in the spinous cell layer of the epidermis (Burge, 1994Burge S. Review cohesion in epidermis.Br J Derm. 1994; 131: 153-159Crossref PubMed Scopus (19) Google Scholar;Salomon et al., 1994Salomon D. Masgrau E. Vischer S. et al.Topography of mammalian connexins in human skin.J Invest Dermatol. 1994; 103: 240-247Crossref PubMed Scopus (121) Google Scholar;Garrod et al., 1996Garrod D. Chidgey M. North A. Desmosomes differentiation, development, dynamics and disease.Curr Opin Cell Biol. 1996; 5: 670-678Crossref Scopus (124) Google Scholar;Green and Jones, 1996Green K.J. Jones J.C. Desmosomes and hemidesmosomes: structure and function of molecular components.FASEB J. 1996; 10: 871-881Crossref PubMed Scopus (289) Google Scholar). Hemidesmosomes connect the basal cells to the dermal-epidermal basement membrane (Green and Jones, 1996Green K.J. Jones J.C. Desmosomes and hemidesmosomes: structure and function of molecular components.FASEB J. 1996; 10: 871-881Crossref PubMed Scopus (289) Google Scholar). The presence of tight junctions in epidermis has remained unclear. Tight-junction-like structures were demonstrated in the granular cell layer of human epidermis in the early 1970s by electron microscopy (Hashimoto, 1971Hashimoto K. Intercellular spaces of the human epidermis as demonstrated with lanthanum.J Invest Dermatol. 1971; 57: 17-31Crossref PubMed Scopus (94) Google Scholar).Elias et al., 1977Elias P.M. McNutt N.S. Friend D.S. Membrane alterations during cornification of mammalian squamous epithelia: a freeze-fracture, tracer, and thin-section study.Anat Rec. 1977; 189: 577-594Crossref PubMed Scopus (174) Google Scholar used the freeze fracture method to study human and mouse stratified epithelia and concluded that tight-junctional elements were either absent or fragmentary. Thus, the epidermal diffusion barrier, which separates the human body from the outer environment, has been thought to be formed not by the tight junctions but by the epidermal lipids, especially the lipid bilayers present in the cornified layer (for review seeWertz, 2000Wertz P.W. Lipids and barrier function of the skin.Acta Derm Venereol Suppl.(Stockh. 2000; 208: 7-11Crossref PubMed Google Scholar). In simple epithelia, tight junctions are formed by series of spot-like contacts or “kisses” that in freeze fracture electron microscopy appear as rows of intramembrane particles that form long branching fibrils circumscribing the cell. The number of fibrils correlates with the tightness of the barrier, which can be investigated by measuring the transcellular electrical resistance (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar;Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita and Furuse, 1999Tsukita S. Furuse M. Occludin and claudins in tight-junction strands: leading or supporting players?.Trends Cell Biol. 1999; 9: 268-273Abstract Full Text Full Text PDF PubMed Scopus (494) Google Scholar,Tsukita and Furuse, 2000Tsukita S. Furuse M. Pores in the wall: claudins constitute tight junction strands containing aqueous pores.J Cell Biol. 2000; 149: 13-16Crossref PubMed Scopus (385) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). Occludin is a transmembrane protein, which is localized exclusively in tight junctions. Occludin has four hydrophobic transmembrane helixes, and both NH2 and COOH terminals are located in the cytoplasmic side of the plasma membrane. The family of claudins are also transmembrane proteins of tight junctions (Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita and Furuse, 1999Tsukita S. Furuse M. Occludin and claudins in tight-junction strands: leading or supporting players?.Trends Cell Biol. 1999; 9: 268-273Abstract Full Text Full Text PDF PubMed Scopus (494) Google Scholar,Tsukita and Furuse, 2000Tsukita S. Furuse M. Pores in the wall: claudins constitute tight junction strands containing aqueous pores.J Cell Biol. 2000; 149: 13-16Crossref PubMed Scopus (385) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). ZO-1 is found in cytoplasmic plaques of tight junctions but is found also in adherens type junctions in cells that lack tight junctions, e.g. fibroblasts and cardiac myocytes (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar;Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). ZO-1 is a member of a MAGUK (membrane-associated guanylate kinase homolog) protein family. It is thought that ZO-1 is involved in creating the proper organization of proteins within the tight junctional plaque (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar;Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). Other cytoplasmic plaque proteins include ZO-2, ZO-3, cingulin, symplekin, and 7H6 (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar;Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). Tight junctional proteins are connected to actin cytoskeleton (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar;Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar;Wittchen et al., 1999Wittchen E.S. Haskins J. Stevenson B.R. Protein interactions at the tight junction.J Biol Chem. 1999; 274: 35179-35185Crossref PubMed Scopus (396) Google Scholar). The tight junction components occludin, ZO-1, and ZO-2, were recently studied in the epidermis of adult rodent skin (Morita et al., 1998Morita K. Itoh M. Saitou M. et al.Subcellular distribution of tight junction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.J Invest Dermatol. 1998; 110: 862-866Crossref PubMed Scopus (107) Google Scholar). Occludin was concentrated at the cell-cell borders only in the most superficial granular cell layers, whereas ZO-1 and ZO-2 were detected also in the uppermost spinous cell layers (Morita et al., 1998Morita K. Itoh M. Saitou M. et al.Subcellular distribution of tight junction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.J Invest Dermatol. 1998; 110: 862-866Crossref PubMed Scopus (107) Google Scholar). This was also the first study to provide clear evidence showing occludin in tight junction strands in mammalian epidermis (Morita et al., 1998Morita K. Itoh M. Saitou M. et al.Subcellular distribution of tight junction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.J Invest Dermatol. 1998; 110: 862-866Crossref PubMed Scopus (107) Google Scholar). Tight junctions have also been studied in cultured human epidermal and gingival keratinocytes using freeze fracture and transmission electron microscopy (Kitajima et al., 1983Kitajima Y. Eguchi K. Ohno T. Mori S. Yaoita H. Tight junctions of human keratinocytes in primary culture: a freeze-fracture study.J Ultrastruct Res. 1983; 82: 309-313Crossref PubMed Scopus (24) Google Scholar;Meyle et al., 1999Meyle J. Gültig K. Rascher G. Wolburg H. Transepithelial electrical resistance and tight junctions of human gingival keratinocytes.J Periodont Res. 1999; 34: 214-222Crossref PubMed Scopus (30) Google Scholar). In our study, the expression of occludin and ZO-1 was elucidated in normal adult and developing human epidermis. Furthermore, localization of ZO-1 and occludin was investigated in pathologic skin conditions known to affect the stratum granulosum, namely psoriasis vulgaris, lichen planus, and ichthyosis vulgaris. Finally, the dynamics of ZO-1 and occludin expression were studied in cultured keratinocytes, which were induced to differentiate by elevating the extracellular calcium (Ca2+) concentration. Skin samples were obtained from plastic surgery operations carried out for cosmetic reasons from 10 healthy persons (aged 20–67 y) at the University Hospital of Turku, Finland. These samples were used for keratinocyte cultures, indirect immunofluorescence (IIF), Western blotting, and transmission electron and immunoelectron microscopies. Skin samples from 13 normal fetuses aged 8, 10, 10, 11, 11, 12, 12, 13, 13, 15, 16, 17, and 21 gestational weeks were obtained from the Department of Obstetrics and Gynecology, University Hospital of Turku, Finland. Fetal skin samples were used for IIF and transmission electron and immunoelectron microscopies. Paraffin-embedded punch biopsy samples of two psoriasis vulgaris, two lichen planus, and two ichthyosis vulgaris were obtained from the Department of Pathology, University Hospital of Turku, Finland, and were used for avidin-biotin immunolabeling. Primary cultures of normal human keratinocytes were established from adult skin samples by a modification of the method ofBoyce and Ham, 1985Boyce S.T. Ham R.G. Cultivation frozen storage, and clonal growth of normal human keratinocytes in serum free media.J Tissue Cult Meth. 1985; 9: 83-93Crossref Scopus (251) Google Scholar. Keratinocytes were maintained in serum-free low Ca2+ (<0.1 mM) keratinocyte growth medium (Clonetics, San Diego, CA) or Defined Keratinocyte-SFM (10744 Gibco BRL Life Technologies, Grand Island, NY). For experimentation, identical groups of Petri dishes or glass coverslips were seeded, and the medium was changed either to the low Ca2+ medium or to the medium containing 1.8 mM Ca2+. Cells were fixed for IIF and harvested for Western blotting after 0, 1, 2, 4, and 24 h incubation. Frozen sections of fetal and adult skin cut on silanated glass slides, and keratinocytes grown on glass coverslips, were fixed with 100% methyl alcohol at -20°C for 10 min. The primary antibodies were diluted in 1% bovin serum albumin (BSA) with phosphate-buffered saline (PBS) and incubated on the samples at 4°C for 20 h. After several washes in PBS, the samples were incubated with secondary antibodies at 20°C for 1 h. In control immunoreactions, primary antibodies were replaced with 1% BSA-PBS. Primary and secondary antibodies used in double labelings were tested for cross-reactions. Formalin-fixed and paraffin-embedded specimens were immunolabeled with avidin-biotin method using Histostain-Plus Kit (Zymed Laboratories, San Francisco, CA) according to the protocol supplied with the kit by the manufacturer. 3.3′-Diaminobenzidine tetrahydrochloride (DAB-Plus Kit, Zymed Laboratories) was used to visualize the antibody localization. Slides were counterstained with Mayer's hematoxylin (Oy Reagena, Kuopio, Finland). Confocal laser scanning microscopy was carried out using a Leica TCS SP spectral confocal laser scanning microscope equipped with an air-cooled argon-krypton ion-laser system (Leica Microsystems Heidelberg, Heidelberg, Germany) and Leica TCS NT software (Version 1.6.551). Final images were saved in tagged image file format (TIFF). The following antibodies were used: mouse monoclonal antibodies to human ZO-1 (33–9100), occludin (33–1500), and E-cadherin (13–1700), polyclonal rabbit antibodies to human ZO-1 (61–7300) and occludin (71–1500), all from Zymed Laboratories; and mouse monoclonal antibody to human desmoplakin I and II (881 147, Boehringer Mannheim Biochemica, Mannheim, Germany). Secondary antibodies for IIF were tetramethylrhodamine isothiocyanate (TRITC) conjugated swine antirabbit (R0156) and rabbit antimouse IgG (R0270). In double labelings, TRITC-conjugated swine antirabbit IgG was mixed with fluorescein isothiocyanate (FITC) conjugated F(ab′)2 fragment of goat antimouse immunoglobulins (F0479, all from Dako, Glostrup, Denmark), or Alexa FluorTM 488 conjugated goat antimouse IgG (Molecular Probes, Eugene, OR). Cells were rinsed with PBS supplemented with protease inhibitors (Complete, Mini, EDTA-free Protease Inhibitors Cocktail Tablets, 1 tablet per 10 ml, Boehringer Mannheim) and then extracted with RIPA buffer (1% Igepal CA-630, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate in PBS) supplemented with protease inhibitors (see above). Cells were scraped from the dishes and incubated on ice for 30 min. Cell lysate was then centrifuged at 15,000g at 4°C for 10 min. Protein concentrations of soluble fraction were detected with DC Protein Assay (Bio-Rad Laboratories, Hercules, CA). Nine micrograms of each preparation were loaded on sodium dodecyl sulfate polyacrylamide gel (6% gel for detecting ZO-1 and 10% for occludin). After electrophoresis, proteins were transferred to Immobilon-P filter (Millipore, Bedford, MA) and immunolabeled with ZO-1 and occludin antibodies in 3% BSA/PBS + 0.5% Tween-20 at 4°C for 20 h. Peroxidase-linked sheep-antimouse (NA 931) and donkey-antirabbit (NA 934) (Amersham International, Little Chalfont, Buckinghamshire, U.K.) were used as secondary antibodies in Western blot analysis. Proteins were detected with ECL (Amersham Life Sciences, Little Chalfont, U.K.) and the filter was exposed to autoradiographic film (Eastman Kodak, New York, NY). 1 × 1 mm pieces of normal and fetal skin were fixed with 5% glutaraldehyde in 0.16 M s-collidin buffer, pH 7.4, and postfixed with 2% OsO4 and 3% K-ferrocyanide (1:1) for 2 h. Samples were dehydrated in graded ethanol series and embedded in Epon 812. Ultra-thin sections were cut on coated copper grids and stained with uranyl acetate and lead citrate. For immunoelectron microscopy, 1 × 1 mm pieces of normal and fetal skin were fixed with freshly prepared 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, for 4 h and washed in phosphate buffer for 12 h at 4°C. Tissue samples were then dehydrated in 50% and 70% ethanol, and infiltrated with L.R. White resin (London Resin, Berkshire, U.K.). Polymerization was performed in a 50°C oven for 24 h. The ultra-thin sections were cut on coated nickel grids. To block unspecific binding, the sections were incubated on goat serum diluted in 1% BSA-Tris-buffered saline (TBS) for 30 min. The primary antibodies were diluted in 1% BSA-TBS and incubated on sections for 20 h at 4°C. The samples were then washed several times in 1% BSA-TBS. Secondary antibody, goat-antimouse IgG + IgM coupled to 18 nm gold particles (Jackson ImmunoResearch Laboratories, West Grove, PA), was diluted in 1% BSA-TBS and incubated on samples for 1 h at 20°C. After several washes in TBS, the samples were postfixed with 2% glutaraldehyde in TBS and counterstained with uranyl acetate and lead citrate. The sections were examined and photographed using a Jeol 1200SX electron microscope. Mouse monoclonal and rabbit polyclonal antibodies raised against ZO-1 and occludin were first characterized using Western transfer analysis of normal human epidermis and cultured epidermal keratinocytes. The results revealed a characteristic 225 kDa band representative of ZO-1, and a series of bands of ≈65 kDa previously reported for occludin (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar). Western analysis also demonstrated that the expression levels of ZO-1 or occludin were essentially the same in keratinocytes cultured in low or high Ca2+ concentration Figure 1. Monoclonal antibodies to ZO-1 and occludin were used for immunolabeling of frozen sections of normal human skin. Confocal laser scanning microscopy demonstrated similar, but not identical, labeling patterns for ZO-1 and occludin. ZO-1 was mainly localized to the granular cell layer, and some labeling was present in the uppermost spinous cell layers Figure 2a. Occludin-specific antibodies resulted in a positive immunoreaction that was restricted to the granular cell layer Figure 2b. High magnification revealed that the labeling pattern for both ZO-1 and occludin was punctuate Figure 2c, d. In order to evaluate the spatial relationship of the cell-cell junction proteins, double immunolabelings were carried out using antibodies recognizing components of tight junctions (ZO-1 and occludin) and adherens junctions (E-cadherin) or desmosomes (desmoplakin) Figure 3 a, b, c, d, e, f, g and h. In general, the labeling signals for proteins of different junctions were separate. On selected locations, however, the immunosignals in red and green were mixed (yellow) indicating close spatial relationship of the proteins. Specifically, immunosignals for ZO-1 (red)/E-cadherin (green) Figure 3a, e and occludin (red)/E-cadherin (green) Figure 3c, g were mixed in lateral cell-cell contact zones in the granular and the uppermost spinous cell layers. Immunosignals for ZO-1 (red)/desmoplakin (green) Figure 3b, f and occludin (red)/desmoplakin (green) Figure 3d, h displayed a very low degree of colocalization. At 8 wk of estimated gestational age (EGA), the epidermis consisted of basal and peridermal cell layers. Already at this stage, ZO-1 and occludin were both expressed in the epidermis. More specifically, tight junction proteins were found in intercellular junctions between the peridermal cells Figure 4a, d. At 13 wk of EGA, the epidermis consisted of three cell layers: the basal, intermediate, and peridermal layers. Also at this stage, ZO-1 and occludin were found in cell-cell junctions of the peridermal cells Figure 4b, e. At 21 wk of EGA, peridermal cells had been shed and the epidermis resembled mature tissue consisting of four compartments: the basal, spinous, granular, and cornified layers. Positive immunosignal for tight junction proteins was located to the granular cell layer Figure 4c, f. Transmission electron microscopy revealed cell-cell contacts with typical morphology of tight junctions in the peridermal cell layer of developing skin Figure 5a, b and c. These intercellular contacts were best visualized at the apical pole of the peridermal cells. In postembedding immunoelectron microscopy, these intercellular junctions of peridermal cells showed immunolabeling for ZO-1 and occludin Figure 5f, g. When ultrathin sections of adult skin were immunolabeled for ZO-1 and occludin, the gold particles were detected at the granular cell layer, often in the vicinity of desmosomes Figure 5d, e. Mouse monoclonal antibodies were used to investigate the expression of tight junction proteins ZO-1 and occludin in cultured epidermal keratinocytes. Induction of keratinocyte differentiation revealed marked changes in the expression and subcellular localization of tight junction proteins. In undifferentiated keratinocytes, cultured in low Ca2+ concentration (<0.1 mM), the immunoreaction for ZO-1 and occludin was faint Figure 6a, d. Keratinocytes were then induced to differentiate using elevated (1.8 mM) extracellular Ca2+ concentration. Under these conditions, keratinocytes begin to form intercellular contacts such as desmosomes and adherens junctions (Jones and Goldman, 1985Jones J.C.R. Goldman R.D. Intermediate filaments and the initiation of desmosome assembly.J Cell Biol. 1985; 101: 506-517Crossref PubMed Scopus (101) Google Scholar;O'Keefe et al., 1987O'Keefe E.J. Briggaman R.A. Herman B. Calcium-induced assembly of adherens junctions in keratinocytes.J Cell Biol. 1987; 105: 807-817Crossref PubMed Scopus (104) Google Scholar). After a 4 h incubation in high Ca2+ concentration, ZO-1 and occludin were detected in cell-cell contact sites as distinct rows of dots having resemblance to the adhesion zipper structure described byVasioukhin et al., 2000Vasioukhin V. Bauer C. Yin M. Fuchs E. Directed actin polymerization is the driving force for epithelial cell-cell adhesion.Cell. 2000; 100: 209-219Abstract Full Text Full Text PDF PubMed Scopus (912) Google Scholar Figure 6b, e. In general, the immunosignal for occludin was detected on fewer cell contacts than ZO-1 at this stage of cellular differentiation. After 24 h incubation of the cultures in high Ca2+ concentration, ZO-1 and occludin immunosignals formed continuous linear zones between adjacent cells Figure 6c, f. In order to investigate the spatial relationship of tight junction proteins versus adherens junction and desmosomal proteins in cultured keratinocytes, double immunolabelings were carried out. Immunosignals for ZO-1 (red) and E-cadherin (green) were mainly separate in early (4 h) and late (24 h) phases of keratinocyte differentiation Figure 7a, d. In analogy, immunosignals for ZO-1 (red) and desmoplakin (green) were mostly separate Figure 7b, e. Furthermore, the subcellular distribution of the two tight junction proteins ZO-1 (red) and occludin (green) was compared. In the early phase of differentiation (4 h) immunosignals for ZO-1 and occludin were mainly separate, and ZO-1 seemed to be markedly more abundant than occludin Figure 7c. As the end result (24 h), immunosignals for tight junction proteins ZO-1 and occludin were seen in yellow color, indicating close spatial association of the proteins Figure 7f. The avidin-biotin method revealed positive labeling for ZO-1 and occludin in the thin zone of granular cells of normal epidermis (Figure 8a, b; see also Figure 2a, b). In order to evaluate the expression of tight junction proteins in pathologic conditions affecting the granular layer, paraffin-embedded samples representing psoriasis, lichen planus, and ichthyosis vulgaris were immunolabeled for ZO-1 and occludin. The results revealed marked alterations of ZO-1 and occludin expression in the diseases studied compared to normal skin. In fully developed psoriatic plaque, the ZO-1 epitopes were detected in several layers of thickened stratum spinosum Figure 8c. Occludin epitopes were also detected in a broader zone in the upper epidermis Figure 8d, although the number of intensely labeled cell layers was somewhat smaller than that for ZO-1. In lichen planus Figure 8g, h, the distribution of ZO-1 and occludin resembled that in psoriasis: the upper and middle parts of the spinous cell layer displayed an immunosignal for ZO-1 and occludin. Ichthyosis (dominant vulgaris type), in which the whole skin area is clinically affected, ZO-1 and occludin were detected in the uppermost spinous cell layer as well as in the cell layers immediately below the cornified layer Figure 8e, f. Blood vessel endothelium, hair follicle epithelium (external root sheath), apical cell membrane and inner intercellular membranes of sweat ducts, and the perineurium were labeled in both normal and pathologic skin samples. Among epithelia, the epidermis has a very challenging function: to separate tissues from the surrounding dry atmosphere. In addition, the epidermis protects the underlying tissues against various chemical irritants. The epidermal barrier function has been considered to be constituted by a mixture of lipids in the intercellular spaces of the stratum corneum (for reviews seeKitson and Thewalt, 2000Kitson N. Thewalt J.L. Hypothesis: the epidermal permeability barrier is a porous medium.Acta Derm Venereol Suppl.(Stockh. 2000; 208: 12-15Crossref PubMed Google Scholar;Wertz, 2000Wertz P.W. Lipids and barrier function of the skin.Acta Derm Venereol Suppl.(Stockh. 2000; 208: 7-11Crossref PubMed Google Scholar). The presence of tight junctions, another potential barrier component in human epidermis – has more or less been ignored, and the final identification of these junctions has awaited conclusive evidence. In this study, ultrastructural analysis combined with immunodetection identified the presence of selected marker proteins for tight junctions in the granular cell layer of human epidermis. Specifically, ZO-1 and occludin were localized to a narrow zone restricted to the granular cell layers. These results are in good agreement with a previous study on mouse skin byMorita et al., 1998Morita K. Itoh M. Saitou M. et al.Subcellular distribution of tight junction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.J Invest Dermatol. 1998; 110: 862-866Crossref PubMed Scopus (107) Google Scholar. These results also support the study ofHashimoto, 1971Hashimoto K. Intercellular spaces of the human epidermis as demonstrated with lanthanum.J Invest Dermatol. 1971; 57: 17-31Crossref PubMed Scopus (94) Google Scholar who recognized intercellular junctions morphologically fulfilling the criteria of tight junctions in the granular cell layer of human epidermis. Only a subset of those junctions were impermeable for lanthanum, however, and were thus considered “true” tight junctions. These junctions were located at the distal ends of the granular cells, the location corresponding well to our findings, which show the most intense immunolabeling for ZO-1 and occludin at the periphery of granular cells. In high magnification, the immunosignal of ZO-1 and occludin appeared as closely located dots. This labeling pattern was analogous to that seen for desmosomal and adherens junction proteins. Double immunolabeling for tight junction components with desmosomal or adherens junction proteins, however, showed that the immunosignals for components of different junction types were mostly separate. At light microscopy level, a close spatial relationship of the components of the different cellular junctions was noted, especially at the lateral meeting points of the neighboring granular cells belonging to the same cell layer. This may be due to the concentration of tight junction proteins and other cell junction components on these sites. Furthermore, as shown in immunoelectron microscopy, tight junction antigens were located in the vicinity of desmosomes, explaining partial colocalization of various cell junction proteins at the light microscopy level. In nonepithelial cells, such as fibroblasts and cardiac myocytes, ZO-1 is colocalized with cadherins and catenins at adherens junctions (Anderson and Van Itallie, 1995Anderson J.M. Van Itallie C.M. Tight junctions and the molecular basis for regulation of paracellular permeability.Am J Physiol. 1995; 269: G467-G475PubMed Google Scholar;Stevenson and Keon, 1998Stevenson B.R. Keon B.H. The tight junction: morphology to molecules.Annu Rev Cell Dev Biol. 1998; 14: 89-109Crossref PubMed Scopus (212) Google Scholar;Tsukita et al., 1999Tsukita S. Furuse M. Itoh M. Structural and signalling molecules come together at tight junctions.Curr Opin Cell Biol. 1999; 11: 628-633Crossref PubMed Scopus (261) Google Scholar). It has also previously been speculated that, in mouse, ZO-1 is associated with adherens junctions in the spinous cell layer during the differentiation of keratinocytes (Morita et al., 1998Morita K. Itoh M. Saitou M. et al.Subcellular distribution of tight junction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.J Invest Dermatol. 1998; 110: 862-866Crossref PubMed Scopus (107) Google Scholar). In our study, in addition to being localized to the granular cell layer, ZO-1 was also detected in short stretches in upper spinous cell layers whereas occludin was found exclusively in granular cells. Thus, the results of this study suggest that ZO-1 may, in part, be associated with adherens junctions also in human epidermis. The results of this study demonstrated tight junctions also in the human embryonic and fetal epidermis as evaluated by IIF and transmission electron and immunoelectron microscopies. In early embryonic epidermis, tight junctions were located in the peridermal cell layer. The periderm is the most superficial layer of the developing skin and is surrounded by amniotic fluid (Holbrook and Odland, 1975Holbrook K.A. Odland G.F. The fine structure of developing human epidermis: light, scanning, and transmission electron microscopy of the periderm.J Invest Dermatol. 1975; 65: 16-38Crossref PubMed Scopus (135) Google Scholar;Holbrook, 1997Holbrook K.A. Development of human skin.Retinoids. 1997; 13: 47-53Google Scholar). Our results are in good agreement with a previous study on mouse skin demonstrating occludin at cell-cell borders of peridermal cells (Morita et al., 1998Morita K. Itoh M. Saitou M. et al.Subcellular distribution of tight junction-associated proteins (occludin, ZO-1, ZO-2) in rodent skin.J Invest Dermatol. 1998; 110: 862-866Crossref PubMed Scopus (107) Google Scholar). The embryonic epidermis differs markedly from the adult tissue: the primitive epidermis is an immature epithelium in a moist environment, lacking the lipid barrier of the stratum corneum. Thus, it is feasible to speculate that tight junctions may form a diffusion barrier in embryonic epidermis. The periderm is shed later during development and the epidermis is differentiated into basal, spinous, granular, and cornified layers (Holbrook and Odland, 1975Holbrook K.A. Odland G.F. The fine structure of developing human epidermis: light, scanning, and transmission electron microscopy of the periderm.J Invest Dermatol. 1975; 65: 16-38Crossref PubMed Scopus (135) Google Scholar;Holbrook, 1997Holbrook K.A. Development of human skin.Retinoids. 1997; 13: 47-53Google Scholar). At this developmental stage, the expression of ZO-1 and occludin was seen in the fetal granular cell layer, a finding resembling that of adult epidermis. Expression of ZO-1 and occludin was investigated in selected pathologic skin conditions, namely psoriasis, lichen planus, and ichthyosis vulgaris. All these diseases are known to have hyperkeratosis and alterations in the stratum granulosum. The results showed that ZO-1 and occludin were relocated to the upper parts of the thickened spinous layer. It thus seems that localizations of tight junction components are susceptible to the aberrations in keratinocyte differentiation. Expression of ZO-1 and occludin was not related to the presence of the typical keratohyalin granules of the granular cell layer, however. This notion is based on the finding that the labeling patterns for ZO-1 and occludin are similar in lichen planus, in which the granular cell layer is prominent, and in psoriasis and ichthyosis vulgaris, which lack granular cells. The speed of keratinization did not seem to have clear impact on the expression of ZO-1 or occludin either, as both rapid parakeratosis (psoriasis) and the other two dermatoses with low keratinization showed similar reaction patterns for ZO-1 and occludin. Keratinocyte cultures serve as a well-documented model of cellular differentiation and formation of intercellular junctions. When human keratinocytes are cultured in medium containing low Ca2+ concentration, they remain undifferentiated and do not form intercellular contacts (Jones and Goldman, 1985Jones J.C.R. Goldman R.D. Intermediate filaments and the initiation of desmosome assembly.J Cell Biol. 1985; 101: 506-517Crossref PubMed Scopus (101) Google Scholar;O'Keefe et al., 1987O'Keefe E.J. Briggaman R.A. Herman B. Calcium-induced assembly of adherens junctions in keratinocytes.J Cell Biol. 1987; 105: 807-817Crossref PubMed Scopus (104) Google Scholar). Shortly after the Ca2+ concentration is elevated to 1.8 mM, keratinocytes begin to differentiate and form cell-cell contacts. Formation of intercellular junctions is an active and dynamic process driven by actin filament polymerization (Vasioukhin et al., 2000Vasioukhin V. Bauer C. Yin M. Fuchs E. Directed actin polymerization is the driving force for epithelial cell-cell adhesion.Cell. 2000; 100: 209-219Abstract Full Text Full Text PDF PubMed Scopus (912) Google Scholar). Ca2+ activates formation of filopodia, which embed into the neighboring cells. This process is associated with clustering of several adherens junction proteins at sites of intercellular contacts. These concentrations form two distinct and highly organized rows of dots, called “adhesion zippers” by Vasioukhin and coworkers. In their study, the number of these adhesion dots increased by several fold within a few hours, and merged to form a single row by 20 h. During this process first desmosomes appear between adjoining cells (Vasioukhin et al., 2000Vasioukhin V. Bauer C. Yin M. Fuchs E. Directed actin polymerization is the driving force for epithelial cell-cell adhesion.Cell. 2000; 100: 209-219Abstract Full Text Full Text PDF PubMed Scopus (912) Google Scholar). The dynamics of tight junction components in human epidermal keratinocyte cultures have not been reported earlier. In this study, immunolabeling for ZO-1 revealed short streaks in the cell-cell contact zones after incubation in elevated Ca2+ for 4 h. It is of interest to note that immunosignals for two different tight junction components, ZO-1 and occludin, were mainly separate during the formation of intercellular junctions, and colocalized only relatively late during this process. Conclusively the results of this study demonstrate the presence of tight junction antigens in the thin granular cell layer of the human epidermis. These findings place tight junctional elements immediately underneath the well-established lipid barrier located in the outermost cornified layer of the epidermis. We thus propose that the epidermal barrier might in part be formed by tight junctional complexes of the stratum granulosum. This work was supported by the following grants: #13338 from Turku University Central Hospital; #H1039 and #K44734 from Oulu University Hospital; #9A056 from the Medical Research Fund of Tampere University Hospital; Turku University Foundation; Finnish Society of Dermatology; and Cancer Society of Finland.