Arch Dermatol Res (1992) 284:154-158
9 Springer-Verlag1992
Immunohistological and immunoelectron microscopic identification of TNF in normal human and murine epidermis* G. Koldel, K. Schulze-Osthoff 2, H. Meyer 1, and J. Knop 3 1 Department of Dermatology, University of Miinster, Miinster, Von-Esmarch-Strage 56, W-4400 Mfinster, Federal Republic of Germany 2 Department of Experimental Dermatology, University Miinster, Mfinster, Federal Republic of Germany 3 Department of Dermatology, University of Mainz, Mainz, Federal Republic of Germany Received May 11, 1991
Summary. The presence, distribution and cellular localization of tumour necrosis factor-alpha (TNF~) were investigated in normal human and murine epidermis using immnnohistological and immunoelectron microscopic methods with monoclonal and polyclonal antibodies. The immunostaining revealed an intercellular plasma membrane and cytoplasmic labelling of the epidermal keratinocytes, but no labelling of Langerhans ceils, melanocytes and Merkel cells. Large amounts of TNF~ were regularly found in the sebaceous glands. These findings demonstrate that epidermal keratinocytes and especially sebocytes produce and release TNFe and that this keratinocytederived cytokine may be important for the structural and functional homeostasis of normal epidermis. Key words: TNFc~ - Immunohistochemistry - I m m u noelectron microscopy - Epidermis - Sebaceous glands
immunoelectron microscopic methods with monoclonal and polyclonal antibodies against TNFc~ to analyse the presence, distribution and especially the cellular localization of this cytokine in normal h u m a n and murine epidermis. The results of this analysis establish the production of T N F e by keratinocytes and especially sebocytes, and indicate a functional role of this cytokine in normal epidermal homeostasis.
Material and methods Tissue samples The investigation was performed on normal human skin and normal earskin of female BALB/c mice. The human biopsies were obtained during plastic surgery from the abdominal skin of eight patients (five males, three females; age 28 to 54 years). All tissue biopsies were divided for immunohistochemistry and immunoelectron microscopy and immediately placed in the relevant fixative.
Antibodies T u m o u r necrosis factor-alpha (TNFe) is a 17 k D a polypeptide with potent i m m u n o m o d u l a t o r y properties [16]. This cytokine has, in concert with other soluble mediators, profound effects on the growth, differentiation and function of epidermal keratinocytes, Langerhans cells, and melanocytes, and is thought to be involved in the mediation and regulation of inflammatory skin diseases [5]. Recently, TNFe-specific m R N A has been found by northern blotting in freshly prepared epidermal cells [9] and by in situ hybridization in normal epidermis and epidermal appendages [2]. The presence of the m R N A and the detection of T N F ~ activity in supernatants of epidermal cell suspensions [9] suggest a synthesis and release by normal epidermis, but do not identify the exact cellular localization and source of this cytokine. In this study, we applied immunohistological and
Rabbit polyclonal antibodies raised against recombinant murine and human TNFc~ were kindly provided by G. Adolf (Vienna, Austria). IgG fractions were isolated by protein A sepharose. Anti-human TNFc~ antibodies were additionally affinity-purified by passage over recombinant TNFc~ coupled to sepharose. A mouse monoclonal antibody against TNFe was purchased from Boehringer (Mannheim, FRG).
Immunohistochemistry The biopsies were snap-frozen in liquid nitrogen and stored at - 7 0 ~ Acetone-fixed cryostat sections, about 4 gm thick, were incubated with the monoclonal and polyclonal antibodies against TNFc~. Positive reactions were visualized with a two-step peroxidase technique using goat anti-rabbit IgG and goat anti-mouse IgG (Dianova, Hamburg, FRG), respectively, and aminoethylcarbazole (Sigma Chemicals, St. Louis, Mo., USA) as described in detail elsewhere [10]. The specificity of labelling was controlled by omission of the anti-TNFc~ antibodies or by replacing the anti-TNFc~ IgG by equal amounts of non-immune rabbit or mouse IgG.
Correspondence to: G. Kolde * Presented in part at the 18th annual meeting of the Arbeitsgemeinschaft Dermatologische Forschung (ADF), Mannheim, 9 11 November 1990
Immunoelectron microscopy The tissue samples were placed in Nakane's (periodate-lysine paraformaldehyde) fixative [13] for 2 h at 4 ~ After being washed in
155 0.1 M phosphate buffered solution containing increasing concentrations of sucrose, the tissue was rapidly frozen in liquid nitrogen and serially cryostat-sectioned at 40 gin. The sections were incubated with the monoclonal and polyclonal antibodies against TNF~ for 2-3 h at 4 ~ and challenged with peroxidase-conjugated goat anti-rabbit IgG and goat anti-mouse IgG (Dianova), respectively, for 4 h at 4 ~ Positive reactions were visualized with the 3,3-diaminobenzidine tetrahydrochloride (Sigma) reaction. After immunostaining, the sections were postfixed in half-strength Karnovsky's solution [6] and osmic acid and embedded in Epon 812 as described in detail previously [11]. The ultrathin sections were stained with uranyl acetate and lead citrate and examined with a Philips electron microscope EM 301 using the double condenser system. In some experiments, the 40 gm cryostat sections were permeabilized with 0.1% Triton X-100 before immunostaining to facilitate the intracellular penetration of the primary and secondary antibodies [4]. The specificity of the ultrastructural labelling was checked by the same incubation procedures as for the immunohistochemical stainings.
Results
Staining remarks The monoclonal and polyctonal antibodies against TNF~ gave similar immunohistological and immunoelectron microscopic staining patterns with very low background labelling. The staining intensity was, however, more pronounced with the polyclonal antibody. In the human skin biopsies, there was no difference in labelling with regard to the age and sex of the patients. No labelling was observed in the controls.
Immunohistochemistry Both the human and murine skin showed a weak to moderate staining of the upper epidermal cell layers (Figs. 1 and 2 a). There was an irregular intercellular and diffuse cytoplasmic labelling of the spinous and granular cell layers. Small granular reaction products were also found
between the horny lamellae. The basal epidermis was usually not stained. The hair follicles and sweat gland ducts were characterized by a similar weak to moderate inter- and intracellular labelling of the inner cell layers. The secretory portions of the sweat glands did not react with the antibodies. In contrast, a bright staining was regularly observed in the sebaceous glands (Figs. 1 and 2b). This intercellular and cytoplasmic reaction was most pronounced in the germinative basal and suprabasal sebocytes, and decreased towards the inner cell layers of the sebaceous glands. The dermis contained some positive round-oval and dendritic cells (Figs. i and 2a).
Immunoelectron microscopy The ultrastructural detection of TNFe revealed a weak homogeneous and sometimes granular peroxidase labelling of the cell membrane of the spinous and granular keratinocytes (Fig. 3a, b). Some granular keratinocytes also showed a patchy cytoplasmic reaction which was not restricted to distinct cell organelles. There was no labelling of the Langerhans cells, melanocytes or Merkel cells. In the sebaceous glands, numerous dark deposits were observed in the vacuolized cytoplasm of the basal and suprabasal cells (Fig. 4 a). These deposits often appeared as granular or globular droplets with an increased staining of their surface. Similar reaction products were occasionally found in the intercellular spaces of the upper epidermis and between, but not within, the horny cells. The few positive cells of the dermis were identified as macrophages, dendritic cells and mast cells. After tissue permeabilization with Triton X-100, a positive staining was observed in the cytoplasm of the keratinocytes and, more pronounced, of the sebaceous cells (Figs. 3c and 4b). The intracellular labelling was usually diffusely distributed in the cytoplasm, but in some experiments the endoplasmic membranes and the
Fig. 1. Immunohistological staining of normal murine skin for TNFc~ showing a moderate labelling of the upper epidermis and a strong labelling of the outer portions of the sebaceous glands (arrowheads). (Haemalum counterstaining; x 200)
156
Fig. 2a, b. Immunohistologicalstaining of normal human skin for TNFc~. a Weak to moderate, mainly intercellularlabelling of the upper epidermis and strong labelling of some dermal cells, b High magnificationof a sebaceous gland with intense staining of the basal and suprabasal sebocytes.(Haemalure counterstaining; a x 220, b x 400) Golgi area of the keratinocytes showed a more intense labelling. There was no specific reaction in the endocytotic organelles of the epidermal cells. Discussion This immunomorphological investigation demonstrates small but distinct amounts of TNF~ in normal human and murine epidermis. The presence of this cytokine was proved by highly specific monoclonal and polyclonal antibodies. The light and electron microscopic labelling results confirm and extend the recent immunohistological study of Oxholm et al. [17] who demonstrated a positive staining of normal human epidermis with a polyclonal antibody against TNF~. The proof of TNFe mRNA by northern blotting and in situ hybridization strongly suggests that epidermal TNFe is produced and released by normal epidermal
cells [2, 9]. So far, these investigations have given no information on the exact cellular localization and source of the cytokine. Our immunoelectron microscopic analysis demonstrated a surface labelling of the keratinocytes, but no staining of the Langerhans cells, melanocytes or Merkel cells. The keratinocytes of the epidermis and epidermal appendages exhibited homogeneous and granular peroxidase reaction products on the cell membrane, which may represent in part the 26 kDa transmembraneanchored TNFe [12]. After tissue permeabilization, the cytoplasm of the keratinocytes was also labelled. The latter reaction was occasionally concentrated in the endoplasmic membranes and Golgi area which are known to be involved in the post-translational modification of secreted proteins [18]. This ultrastructural labelling pattern indicates, along with the mRNA findings, that keratinoeytes are the main cellular source of TNF~ in normal epidermis.
157
Fig. 3a, b. Immunoelectron microscopic staining of normal human epidermis for TNFc~. a Globular peroxidase reaction products between the horny lamellae (arrowheads) and cytoplasmic staining of the granular keratinocytes (arrows). b Spinous keratinocytes with moderate staining of the cell membrane (arrowhead). e Diffuse
cytoplasmic labelling and weak membrane staining of keratinocytes after Triton permeabilization of the sections. Note the absent labelling of the intersectioned Langerhans cell with Birbeck granules (arrowhead). (a • 5700, b x9500, e x 17400)
The pattern and intensity of the observed TNFc~ staining showed pronounced regional differences in the various epidermal compartments of normal skin. In accordance with the immunohistological results of Oxholm et al. [17], there was no light or electron microscopic staining of the basal keratinocytes and only a weak to moderate cellular and intercellular staining of the upper cell layers. This staining pattern was also observed in the hair follicles and sweat gland ducts. By contrast, the basal and suprabasal cells of the sebaceous glands demonstrated an intense cytoplasmic labelling with globular deposits. Similar deposits were found in the intercellular space of the upper epidermis and between the horny lamellae, and suggest a secretion of TNF~ by normal sebaceous glands. The differences in detectable TNF~ obviously correlates with the amounts of this cytokine present in the epidermis and may be caused by a different internal regulation and/or external stimulation of the keratinocytes. TNFc~ is known to be induced by a number of stimuli, including the highly potent bacterial lipopolysaccharides and certain cytokines, such as interleukin 1 and 2 and granulocyte-macrophage colony-stimulating factor [1, 7, 16]. In normal skin, the distinct amounts of bacterial
liposaccharides which are usually present in the normal resident flora of the epidermal surface and sebaceous glands [19] may result in TNF~ production by the upper keratinocytes and especially sebocytes. Several in vitro and in vivo experiments have shown that TNF~ affects the structural and functional properties of certain epidermal cells. Most of these effects are known for the keratinocytes in which T N F a causes an inhibition of proliferation, the formation of granular cells, and the expression of cell surface antigens, such as M H C class-II molecules and ICAM-1 [3, 5, 14, 15, 21]. The Langerhans cells are maintained by the cytokine in a viable but functionally immature state [8]. In normal human melanocytes, T N F e inhibits the proliferation and melanogenesis [20]. It is therefore reasonable to assume that the local production of TNFc~ acts, like other keratinocytederived cytokines, in an autocrine and paracrine manner on normal epidermal homeostasis.
References 1. Beutler B, Cerami A (1988) Tumor necrosis, cachexia, shock and inflammation: a common mediator. Annu Rev Biochem 57:505 518
158
Fig. 4a, b. Immunoelectron microscopic staining of basal murine sebocytes for TNF~. a The cytoplasm of the cells contains numerous globular staining products (arrowheads). b After tissue permeabili-
zation, the sebocytes show a diffuse cytoplasmic labelling which is more intense in the Golgi areas (arrowheads). (a x 7200, b x 19200)
2. Boehm KD, Yun JK, Elmets CA (1990) An examination of the expression of interleukin-1 alpha and beta and tumor necrosis factor-alpha genes in normal human skin by in situ hybridization analysis. J Ceil Biol 111: 217a 3. Detmar M, Orfanos CE (1990) Tumor necrosis factor-alpha inhibits cell proliferation and induces class II antigens and cell adhesion molecules in cultured normal human keratinocytes in vitro. Arch Dermatol Res 282:238-245 4. Eldred WD, Zucker C, Karten HJ, Yazulla S (1983) Comparison of fixation and penetration enhancement techniques for use in ultrastructuraI immunocytochemistry. J Histochem Cytochem 31 : 285-292 5. Griffiths CEM, Voorhees J J, Nickoloff BJ (1989) Characterization of intercellular adhesion molecule-1 and HLA-DR expression in normal and inflamed skin: modulation by recombinant gamma interferon and tumor necrosis factor. J Am Acad Dermatol 20:617-629 6. Karnovsky MY (1985) A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J Cell Biol 27: 137A-138A 7. Kasid A, Director EP, Stovroff MC, Lotze MT, Rosenberg SA (1990) Cytokine regulation of tumor necrosis factor-e and (lymphotoxin) - messenger RNA expression in human peripheral blood mononuclear cells. Cancer Res 50:5072-5076 8. Koch F, Heufler Ch, Kfimpgen E, Schneeweiss D, B6ck G, Schuler G (1990) Tumor necrosis factor-e maintains the viability of murine epidermal Langerhans cells in culture, but in contrast to granulocyte/macrophage colony-stimulating factor, without inducing their functional maturation. J Exp. Med 171 : 159 171
11. Kolde G, Knop J (1986) Ultrastructural morphometry of epidermal Langerhans cells: introduction of a simple method for a comprehensive quantitative analysis of the cells. Arch Dermatol Res 278:298-301 12. Kriegler M, Perez C, DeFay K, Albert I, Lu SD (1988) A novel form of TNF/cachectin is a cell surface cytotoxic transmembrane protein: ramifications for the complex physiology of TNF. Cell 53:45 53 13. McLean IW, Nakane PK (1974) Periodate lysine paraformaldehyde fixative: a new fixative for immunoelectron microscopy. J Histochem Cytochem 22:1077-1083 14. Nagano K, Hori K, Nagane T, Sugawara T, Ohishi J, Hayashi H, Watanabe N, Niitsu Y (1990) Effect oftumour necrosis factor in the mouse-tail model of psoriasis. Arch Dermatol Res 282: 459 462 15. Nakamura K, Tamaki (1990) Tumour necrosis factor enhances the interferon-induced class II MHC antigen expression on murine keratinocytes in vivo. Arch Dermatol Res 282:415-417 16. Old LJ (1987) Tumour necrosis factor. Polypeptide mediator network. Nature 326:330-331 17. Oxholm A, Oxholm P, Staberg B, Bendtzen K (1988) Immunohistological detection of interleukin I-like molecules and tumour necrosis factor in human epidermis before and after UVBirradiation in vivo. Br J Dermatol 118:369 376 18. Pfeffer SR, Rothman JE (1987) Biosynthetic protein transport and sorting by the endoplasmic reticulum and Golgi. Annu Rev Biochem 56:829-852 19. Roberts SOB, Highet AS (1986) Normal flora of the skin. In: Rook A, Wilkinson DS, Ebling FJG, Champion RH, Burton JL(eds) Textbook of dermatology. Blackwell Scientific, Oxford, pp 725-730 20. Swope VB, Abdel-Malek Z, Kassem LM, Nordlund JJ (1991) Interleukins lc~and 6 and tumor necrosis factor-c~ are paracrine inhibitors of human melanocyte proliferation and melanogenesis. J Invest Dermatol 96:180 185 21. Symington FW (1989) Lymphotoxin, tumor necrosis factor, and gamma interferon are cytostatic for normal human keratinocytes. J Invest Dermatol 92:798-805
9. K6ck A, Schwarz T, Kirnbauer R, Urbanski A, Perry P, Ansel JC, Luger TA (1990) Human keratinocytes are a source for tumor necrosis factor-c~:evidence for synthesis and release upon stimulation with endotoxin or ultraviolet light. J Exp Med 172: 1609-1614 10. Kolde G, Br6cker EB (1986) Multiple skin tumors of indeterminate cells in an adult. J Am Acad Dermatol 15:591 597
9 Springer-Verlag1992
Immunohistological and immunoelectron microscopic identification of TNF in normal human and murine epidermis* G. Koldel, K. Schulze-Osthoff 2, H. Meyer 1, and J. Knop 3 1 Department of Dermatology, University of Miinster, Miinster, Von-Esmarch-Strage 56, W-4400 Mfinster, Federal Republic of Germany 2 Department of Experimental Dermatology, University Miinster, Mfinster, Federal Republic of Germany 3 Department of Dermatology, University of Mainz, Mainz, Federal Republic of Germany Received May 11, 1991
Summary. The presence, distribution and cellular localization of tumour necrosis factor-alpha (TNF~) were investigated in normal human and murine epidermis using immnnohistological and immunoelectron microscopic methods with monoclonal and polyclonal antibodies. The immunostaining revealed an intercellular plasma membrane and cytoplasmic labelling of the epidermal keratinocytes, but no labelling of Langerhans ceils, melanocytes and Merkel cells. Large amounts of TNF~ were regularly found in the sebaceous glands. These findings demonstrate that epidermal keratinocytes and especially sebocytes produce and release TNFe and that this keratinocytederived cytokine may be important for the structural and functional homeostasis of normal epidermis. Key words: TNFc~ - Immunohistochemistry - I m m u noelectron microscopy - Epidermis - Sebaceous glands
immunoelectron microscopic methods with monoclonal and polyclonal antibodies against TNFc~ to analyse the presence, distribution and especially the cellular localization of this cytokine in normal h u m a n and murine epidermis. The results of this analysis establish the production of T N F e by keratinocytes and especially sebocytes, and indicate a functional role of this cytokine in normal epidermal homeostasis.
Material and methods Tissue samples The investigation was performed on normal human skin and normal earskin of female BALB/c mice. The human biopsies were obtained during plastic surgery from the abdominal skin of eight patients (five males, three females; age 28 to 54 years). All tissue biopsies were divided for immunohistochemistry and immunoelectron microscopy and immediately placed in the relevant fixative.
Antibodies T u m o u r necrosis factor-alpha (TNFe) is a 17 k D a polypeptide with potent i m m u n o m o d u l a t o r y properties [16]. This cytokine has, in concert with other soluble mediators, profound effects on the growth, differentiation and function of epidermal keratinocytes, Langerhans cells, and melanocytes, and is thought to be involved in the mediation and regulation of inflammatory skin diseases [5]. Recently, TNFe-specific m R N A has been found by northern blotting in freshly prepared epidermal cells [9] and by in situ hybridization in normal epidermis and epidermal appendages [2]. The presence of the m R N A and the detection of T N F ~ activity in supernatants of epidermal cell suspensions [9] suggest a synthesis and release by normal epidermis, but do not identify the exact cellular localization and source of this cytokine. In this study, we applied immunohistological and
Rabbit polyclonal antibodies raised against recombinant murine and human TNFc~ were kindly provided by G. Adolf (Vienna, Austria). IgG fractions were isolated by protein A sepharose. Anti-human TNFc~ antibodies were additionally affinity-purified by passage over recombinant TNFc~ coupled to sepharose. A mouse monoclonal antibody against TNFe was purchased from Boehringer (Mannheim, FRG).
Immunohistochemistry The biopsies were snap-frozen in liquid nitrogen and stored at - 7 0 ~ Acetone-fixed cryostat sections, about 4 gm thick, were incubated with the monoclonal and polyclonal antibodies against TNFc~. Positive reactions were visualized with a two-step peroxidase technique using goat anti-rabbit IgG and goat anti-mouse IgG (Dianova, Hamburg, FRG), respectively, and aminoethylcarbazole (Sigma Chemicals, St. Louis, Mo., USA) as described in detail elsewhere [10]. The specificity of labelling was controlled by omission of the anti-TNFc~ antibodies or by replacing the anti-TNFc~ IgG by equal amounts of non-immune rabbit or mouse IgG.
Correspondence to: G. Kolde * Presented in part at the 18th annual meeting of the Arbeitsgemeinschaft Dermatologische Forschung (ADF), Mannheim, 9 11 November 1990
Immunoelectron microscopy The tissue samples were placed in Nakane's (periodate-lysine paraformaldehyde) fixative [13] for 2 h at 4 ~ After being washed in
155 0.1 M phosphate buffered solution containing increasing concentrations of sucrose, the tissue was rapidly frozen in liquid nitrogen and serially cryostat-sectioned at 40 gin. The sections were incubated with the monoclonal and polyclonal antibodies against TNF~ for 2-3 h at 4 ~ and challenged with peroxidase-conjugated goat anti-rabbit IgG and goat anti-mouse IgG (Dianova), respectively, for 4 h at 4 ~ Positive reactions were visualized with the 3,3-diaminobenzidine tetrahydrochloride (Sigma) reaction. After immunostaining, the sections were postfixed in half-strength Karnovsky's solution [6] and osmic acid and embedded in Epon 812 as described in detail previously [11]. The ultrathin sections were stained with uranyl acetate and lead citrate and examined with a Philips electron microscope EM 301 using the double condenser system. In some experiments, the 40 gm cryostat sections were permeabilized with 0.1% Triton X-100 before immunostaining to facilitate the intracellular penetration of the primary and secondary antibodies [4]. The specificity of the ultrastructural labelling was checked by the same incubation procedures as for the immunohistochemical stainings.
Results
Staining remarks The monoclonal and polyctonal antibodies against TNF~ gave similar immunohistological and immunoelectron microscopic staining patterns with very low background labelling. The staining intensity was, however, more pronounced with the polyclonal antibody. In the human skin biopsies, there was no difference in labelling with regard to the age and sex of the patients. No labelling was observed in the controls.
Immunohistochemistry Both the human and murine skin showed a weak to moderate staining of the upper epidermal cell layers (Figs. 1 and 2 a). There was an irregular intercellular and diffuse cytoplasmic labelling of the spinous and granular cell layers. Small granular reaction products were also found
between the horny lamellae. The basal epidermis was usually not stained. The hair follicles and sweat gland ducts were characterized by a similar weak to moderate inter- and intracellular labelling of the inner cell layers. The secretory portions of the sweat glands did not react with the antibodies. In contrast, a bright staining was regularly observed in the sebaceous glands (Figs. 1 and 2b). This intercellular and cytoplasmic reaction was most pronounced in the germinative basal and suprabasal sebocytes, and decreased towards the inner cell layers of the sebaceous glands. The dermis contained some positive round-oval and dendritic cells (Figs. i and 2a).
Immunoelectron microscopy The ultrastructural detection of TNFe revealed a weak homogeneous and sometimes granular peroxidase labelling of the cell membrane of the spinous and granular keratinocytes (Fig. 3a, b). Some granular keratinocytes also showed a patchy cytoplasmic reaction which was not restricted to distinct cell organelles. There was no labelling of the Langerhans cells, melanocytes or Merkel cells. In the sebaceous glands, numerous dark deposits were observed in the vacuolized cytoplasm of the basal and suprabasal cells (Fig. 4 a). These deposits often appeared as granular or globular droplets with an increased staining of their surface. Similar reaction products were occasionally found in the intercellular spaces of the upper epidermis and between, but not within, the horny cells. The few positive cells of the dermis were identified as macrophages, dendritic cells and mast cells. After tissue permeabilization with Triton X-100, a positive staining was observed in the cytoplasm of the keratinocytes and, more pronounced, of the sebaceous cells (Figs. 3c and 4b). The intracellular labelling was usually diffusely distributed in the cytoplasm, but in some experiments the endoplasmic membranes and the
Fig. 1. Immunohistological staining of normal murine skin for TNFc~ showing a moderate labelling of the upper epidermis and a strong labelling of the outer portions of the sebaceous glands (arrowheads). (Haemalum counterstaining; x 200)
156
Fig. 2a, b. Immunohistologicalstaining of normal human skin for TNFc~. a Weak to moderate, mainly intercellularlabelling of the upper epidermis and strong labelling of some dermal cells, b High magnificationof a sebaceous gland with intense staining of the basal and suprabasal sebocytes.(Haemalure counterstaining; a x 220, b x 400) Golgi area of the keratinocytes showed a more intense labelling. There was no specific reaction in the endocytotic organelles of the epidermal cells. Discussion This immunomorphological investigation demonstrates small but distinct amounts of TNF~ in normal human and murine epidermis. The presence of this cytokine was proved by highly specific monoclonal and polyclonal antibodies. The light and electron microscopic labelling results confirm and extend the recent immunohistological study of Oxholm et al. [17] who demonstrated a positive staining of normal human epidermis with a polyclonal antibody against TNF~. The proof of TNFe mRNA by northern blotting and in situ hybridization strongly suggests that epidermal TNFe is produced and released by normal epidermal
cells [2, 9]. So far, these investigations have given no information on the exact cellular localization and source of the cytokine. Our immunoelectron microscopic analysis demonstrated a surface labelling of the keratinocytes, but no staining of the Langerhans cells, melanocytes or Merkel cells. The keratinocytes of the epidermis and epidermal appendages exhibited homogeneous and granular peroxidase reaction products on the cell membrane, which may represent in part the 26 kDa transmembraneanchored TNFe [12]. After tissue permeabilization, the cytoplasm of the keratinocytes was also labelled. The latter reaction was occasionally concentrated in the endoplasmic membranes and Golgi area which are known to be involved in the post-translational modification of secreted proteins [18]. This ultrastructural labelling pattern indicates, along with the mRNA findings, that keratinoeytes are the main cellular source of TNF~ in normal epidermis.
157
Fig. 3a, b. Immunoelectron microscopic staining of normal human epidermis for TNFc~. a Globular peroxidase reaction products between the horny lamellae (arrowheads) and cytoplasmic staining of the granular keratinocytes (arrows). b Spinous keratinocytes with moderate staining of the cell membrane (arrowhead). e Diffuse
cytoplasmic labelling and weak membrane staining of keratinocytes after Triton permeabilization of the sections. Note the absent labelling of the intersectioned Langerhans cell with Birbeck granules (arrowhead). (a • 5700, b x9500, e x 17400)
The pattern and intensity of the observed TNFc~ staining showed pronounced regional differences in the various epidermal compartments of normal skin. In accordance with the immunohistological results of Oxholm et al. [17], there was no light or electron microscopic staining of the basal keratinocytes and only a weak to moderate cellular and intercellular staining of the upper cell layers. This staining pattern was also observed in the hair follicles and sweat gland ducts. By contrast, the basal and suprabasal cells of the sebaceous glands demonstrated an intense cytoplasmic labelling with globular deposits. Similar deposits were found in the intercellular space of the upper epidermis and between the horny lamellae, and suggest a secretion of TNF~ by normal sebaceous glands. The differences in detectable TNF~ obviously correlates with the amounts of this cytokine present in the epidermis and may be caused by a different internal regulation and/or external stimulation of the keratinocytes. TNFc~ is known to be induced by a number of stimuli, including the highly potent bacterial lipopolysaccharides and certain cytokines, such as interleukin 1 and 2 and granulocyte-macrophage colony-stimulating factor [1, 7, 16]. In normal skin, the distinct amounts of bacterial
liposaccharides which are usually present in the normal resident flora of the epidermal surface and sebaceous glands [19] may result in TNF~ production by the upper keratinocytes and especially sebocytes. Several in vitro and in vivo experiments have shown that TNF~ affects the structural and functional properties of certain epidermal cells. Most of these effects are known for the keratinocytes in which T N F a causes an inhibition of proliferation, the formation of granular cells, and the expression of cell surface antigens, such as M H C class-II molecules and ICAM-1 [3, 5, 14, 15, 21]. The Langerhans cells are maintained by the cytokine in a viable but functionally immature state [8]. In normal human melanocytes, T N F e inhibits the proliferation and melanogenesis [20]. It is therefore reasonable to assume that the local production of TNFc~ acts, like other keratinocytederived cytokines, in an autocrine and paracrine manner on normal epidermal homeostasis.
References 1. Beutler B, Cerami A (1988) Tumor necrosis, cachexia, shock and inflammation: a common mediator. Annu Rev Biochem 57:505 518
158
Fig. 4a, b. Immunoelectron microscopic staining of basal murine sebocytes for TNF~. a The cytoplasm of the cells contains numerous globular staining products (arrowheads). b After tissue permeabili-
zation, the sebocytes show a diffuse cytoplasmic labelling which is more intense in the Golgi areas (arrowheads). (a x 7200, b x 19200)
2. Boehm KD, Yun JK, Elmets CA (1990) An examination of the expression of interleukin-1 alpha and beta and tumor necrosis factor-alpha genes in normal human skin by in situ hybridization analysis. J Ceil Biol 111: 217a 3. Detmar M, Orfanos CE (1990) Tumor necrosis factor-alpha inhibits cell proliferation and induces class II antigens and cell adhesion molecules in cultured normal human keratinocytes in vitro. Arch Dermatol Res 282:238-245 4. Eldred WD, Zucker C, Karten HJ, Yazulla S (1983) Comparison of fixation and penetration enhancement techniques for use in ultrastructuraI immunocytochemistry. J Histochem Cytochem 31 : 285-292 5. Griffiths CEM, Voorhees J J, Nickoloff BJ (1989) Characterization of intercellular adhesion molecule-1 and HLA-DR expression in normal and inflamed skin: modulation by recombinant gamma interferon and tumor necrosis factor. J Am Acad Dermatol 20:617-629 6. Karnovsky MY (1985) A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J Cell Biol 27: 137A-138A 7. Kasid A, Director EP, Stovroff MC, Lotze MT, Rosenberg SA (1990) Cytokine regulation of tumor necrosis factor-e and (lymphotoxin) - messenger RNA expression in human peripheral blood mononuclear cells. Cancer Res 50:5072-5076 8. Koch F, Heufler Ch, Kfimpgen E, Schneeweiss D, B6ck G, Schuler G (1990) Tumor necrosis factor-e maintains the viability of murine epidermal Langerhans cells in culture, but in contrast to granulocyte/macrophage colony-stimulating factor, without inducing their functional maturation. J Exp. Med 171 : 159 171
11. Kolde G, Knop J (1986) Ultrastructural morphometry of epidermal Langerhans cells: introduction of a simple method for a comprehensive quantitative analysis of the cells. Arch Dermatol Res 278:298-301 12. Kriegler M, Perez C, DeFay K, Albert I, Lu SD (1988) A novel form of TNF/cachectin is a cell surface cytotoxic transmembrane protein: ramifications for the complex physiology of TNF. Cell 53:45 53 13. McLean IW, Nakane PK (1974) Periodate lysine paraformaldehyde fixative: a new fixative for immunoelectron microscopy. J Histochem Cytochem 22:1077-1083 14. Nagano K, Hori K, Nagane T, Sugawara T, Ohishi J, Hayashi H, Watanabe N, Niitsu Y (1990) Effect oftumour necrosis factor in the mouse-tail model of psoriasis. Arch Dermatol Res 282: 459 462 15. Nakamura K, Tamaki (1990) Tumour necrosis factor enhances the interferon-induced class II MHC antigen expression on murine keratinocytes in vivo. Arch Dermatol Res 282:415-417 16. Old LJ (1987) Tumour necrosis factor. Polypeptide mediator network. Nature 326:330-331 17. Oxholm A, Oxholm P, Staberg B, Bendtzen K (1988) Immunohistological detection of interleukin I-like molecules and tumour necrosis factor in human epidermis before and after UVBirradiation in vivo. Br J Dermatol 118:369 376 18. Pfeffer SR, Rothman JE (1987) Biosynthetic protein transport and sorting by the endoplasmic reticulum and Golgi. Annu Rev Biochem 56:829-852 19. Roberts SOB, Highet AS (1986) Normal flora of the skin. In: Rook A, Wilkinson DS, Ebling FJG, Champion RH, Burton JL(eds) Textbook of dermatology. Blackwell Scientific, Oxford, pp 725-730 20. Swope VB, Abdel-Malek Z, Kassem LM, Nordlund JJ (1991) Interleukins lc~and 6 and tumor necrosis factor-c~ are paracrine inhibitors of human melanocyte proliferation and melanogenesis. J Invest Dermatol 96:180 185 21. Symington FW (1989) Lymphotoxin, tumor necrosis factor, and gamma interferon are cytostatic for normal human keratinocytes. J Invest Dermatol 92:798-805
9. K6ck A, Schwarz T, Kirnbauer R, Urbanski A, Perry P, Ansel JC, Luger TA (1990) Human keratinocytes are a source for tumor necrosis factor-c~:evidence for synthesis and release upon stimulation with endotoxin or ultraviolet light. J Exp Med 172: 1609-1614 10. Kolde G, Br6cker EB (1986) Multiple skin tumors of indeterminate cells in an adult. J Am Acad Dermatol 15:591 597
