Gene expression profiles in psoriasis: analysis of impact of body site location and clinical severity.

BACKGROUND: Psoriasis is characterized by symmetry of plaques and modulation of multiple genes within those plaques. OBJECTIVES: We compared gene expression profiles of plaques of psoriasis at different anatomical sites for both symmetrical and asymmetrical disease to ascertain whether the same genes were expressed. METHODS: Gene expression profiles were analysed in biopsies from lesional and uninvolved skin from two groups of patients with either predominantly symmetrical or truncal plaques of psoriasis vulgaris, and from normal skin of healthy volunteers. Genomic analyses were performed using cDNA array and kinetically monitored reverse transcriptase-initiated polymerase chain reaction (kRT-PCR) approaches. A cluster of genes upregulated in involved psoriasis skin as compared with normal skin was identified using each of these two technologies. RESULTS: Clustering of patients based on their gene expression profile did not reveal any correlation with family history of psoriasis, age at onset or association of psoriasis with arthritis. There was no difference in gene expression profile between the type (symmetrical vs. truncal) or location (left vs. right side of body) of psoriatic plaques. Gene expression profiles of involved psoriatic skin analysed by kRT-PCR analysis did correlate with both global (Psoriasis Area and Severity Index) and local (erythema, desquamation and plaque elevation) clinical severity. CONCLUSIONS: These results indicate that it may be feasible to analyse the molecular effects of pharmacological agents on psoriatic skin in 'minizone' protocols, that the obtained data can be correlated with clinical severity and that plaques of psoriasis in the same individual express the same genes.

Ultraviolet B induces hyperproliferation and modification of epidermal differentiation in normal human skin grafted on to nude mice.

BACKGROUND: For ethical and technical reasons, the in vivo biological effects of ultraviolet (UV) radiation on skin are difficult to study in human volunteers. The use of human skin grafted on to nude mice may circumvent this difficulty. OBJECTIVES: To investigate the effects of a single moderate UVB exposure on human skin grafted on to nude mice. METHODS: Modifications of epidermal differentiation markers and patterns of keratin expression were assessed from 24 h to 14 days after a physiological UVB irradiation characterized by the induction of sunburn cells. RESULTS: During the first 48 h postexposure, involucrin, loricrin, transglutaminase type I, filaggrin and keratin K2e expression were altered together with the formation of abnormal horny layers. Constitutive keratin K14 was increased while keratin K10 expression was delayed. Newly synthesized keratins K6, K16, K17 and K19 were induced in parallel with an increase in the epidermal proliferation rate. A progressive normalization of both keratinocyte proliferation and differentiation took place during the following days, reaching completion within 2 weeks. CONCLUSIONS: Exposure of human skin to a UVB dose corresponding to a mild sunburn reaction induces epidermal hyperproliferation and alterations of several constitutive differentiation markers, as well as a drastic modification in the pattern of epidermal keratins. Although these modifications were shown to be progressively reversed in a single exposure model, the data also suggest that subsequent UV exposures occurring during the recovery period may lead to potentially deleterious long-term consequences, such as photoageing and photocarcinogenesis. Grafted human skin appeared to be an attractive and promising model for investigating the biological consequences of UVB radiation in vivo.

Localisation of members of the notch system and the differentiation of vibrissa hair follicles: receptors, ligands, and fringe modulators.

Hair vibrissa follicle morphogenesis involves several cell segregation phases, in the dermis as well as in the epidermis. The expression of Notch-related genes, which are well established mediators of multiple cell segregation events in Drosophila development, was studied by in situ hybridisation during embryonic mouse vibrissa follicle morphogenesis and the first adult hair cycle. The results show that two receptors, Notch1 and -2, three ligands, Delta1, Serrate1, and -2, and the three Fringe regulators, Lunatic, Manic, and Radical, are expressed in different locations and morphogenetic stages. First, the appearance of hair vibrissa primordia involves the expression of complementary patterns of Notch2, Delta1, and Lunatic Fringe in the dermis and of Notch1, Serrate2, and Lunatic Fringe in the epidermis. Second, this expression pattern is no longer found after stage 3 in the dermis. Meanwhile, in the epidermis, the expression of Notch1, Serrate2, and Lunatic Fringe before the formation of the placode may be involved in determining two populations of epidermal cells in the developing follicle. Third, complementary expression patterns for Notch1, Manic, and Lunatic Fringe, as well as Serrate1 and -2 as previously shown (Powell et al., 1998), are progressively established from stage 4 of embryonic development both in the outer root sheath and in the hair matrix. These patterns are consistent with the one found in the adult anagen phase. During the hair vibrissa cycle, Notch1 and Manic Fringe display temporal and spatial changes of expression, suggesting that they may intervene as modulators of trichocyte activities.

Regulation of vascular endothelial growth factor expression in human keratinocytes by retinoids.

Vascular endothelial growth factor (VEGF) is overexpressed in hyperproliferative diseases, such as psoriasis and cancers, which are characterized by increased angiogenesis. Experimentally, VEGF overexpression can be induced by the treatment of cell cultures and biological tissues with phorbol esters, such as 12-O-tetradecanoylphorbol-13-acetate (TPA). Using normal human keratinocytes in conventional cultures and skin grafted onto nude mice in vivo, we show that retinoids can inhibit TPA-mediated VEGF gene induction at the transcriptional level. Because retinoids are biologically active either by interacting with the nuclear retinoic acid receptors or by interfering with the activator protein 1 (AP1) transcription factor, we studied the effect of the retinoic acid derivative CD 2409, which exhibits strong anti-AP1 activity but does not bind to the known retinoic acid receptors in vitro. The results demonstrate that the inhibition of VEGF expression by retinoids only depends on their anti-AP1 activity and does not require gene transactivation via retinoic acid response elements. Because the VEGF promoter contains four potential AP1 binding sites, we used different promoter constructs to identify the functional site responsible for TPA induction and retinoid inhibition. This site turned out to be localized at position -621 of the 5' flanking region of the VEGF gene.

Wound healing of human skin transplanted onto the nude mouse after a superficial excisional injury: human dermal reconstruction is achieved in several steps by two different fibroblast subpopulations.

It has been established that human skin grafted onto the nude mouse is able to regenerate after being subjected to a full-thickness wound. In the present work, we sought to determine the cells involved in the connective tissue repair process following superficial wounding. Two months after transplantation, superficial wounds were made at the center of the graft using mechanical dermabrasion. At various times thereafter, ranging from 2 days to 6 weeks, healing grafts were harvested and processed for immunohistological study with species-specific and cross-reacting antibodies directed against human or mouse antigens. The grafted human skin regenerated according to the following series of events. First, the human dermis underneath the scab became devoid of human fibroblasts while the surrounding human dermis preserved its own characteristics. The TUNEL reaction on earlyphase healing wounds indicated that apoptosis occurred steadily within this area and could be the mechanism by which cells disappeared. Moreover, cell death was reduced when the wound was covered with an occlusive dressing. The human dermis beneath the wound was then invaded by mouse cells which deposited type I collagen on the human extracellular matrix and produced mouse granulation tissue at the surface above it. Human keratinocytes migrated over the mouse granulation tissue to reconstruct the epidermis. Eventually, the mouse granulation tissue was progressively invaded by human fibroblasts, which formed a human neodermis. The overall process appeared to depend upon several successive epithelial-mesenchymal interactions, which were not species-specific. This suggests that myofibroblasts arise from a specific subpopulation of fibroblasts, probably located at the interface between the dermis and adipose tissue, and that the granulation tissue is eventually remodeled by another population of fibroblasts present in the human dermis.

Pharmacology of adapalene.

Adapalene, a synthetic retinoid, is a new drug proposed for the treatment of acne patients. Studies on the in vitro and in vivo pharmacology of adapalene have shown that it is very active on cell and tissue proliferation and differentiation. Furthermore, adapalene has anti-inflammatory potential as determined by its anti-AP1 activity. Adapalene interacts selectively with the nuclear receptors RAR beta and RAR gamma, and its activity on proliferation and differentiation can be blocked by a RAR beta-gamma antagonist. Because RAR beta is not expressed in human keratinocytes, the effect of adapalene on the major cell type of the epidermis is certainly mediated by its interaction with RAR gamma. The unique pharmacological properties of adapalene may explain why, when compared to tretinoin, it has an improved therapeutic ratio due to its better tolerance.

Dermal cysts of the rhino mouse develop into unopened sebaceous glands.

The rhino mouse (hr(rh)hr(rh)) is a mutant strain characterized by a wrinkled and hairless skin with epidermal utriculi (pseudocomedones) and dermal cysts. The epidermal cysts have been extensively studied. The present work focused on the dermal cysts. By electron microscopy it was found that they appear on day 20 after birth and that they originate from a pool of undifferentiated epithelial cells of the deepest part of the initial follicular unit. Progressively, the number of cells in these islets increased and a central cavity was formed. Peripheral cells differentiated into sebocyte-like cells and outer root sheath cells. Staining with Oil Red O solution indicated accumulation of lipid material in the central cavity. The dermal cysts of the adult rhino mouse were isolated and purified in several steps including enzyme digestion, centrifugation, and separation on Nylex sieves. The integrity of the isolated cysts was confirmed by histology and electron microscopy. Study of their keratin polypeptide pattern by gel electrophoresis indicated that they express the mouse keratins 5, 14, 6 and 17. Neutral lipid analysis of the dermal cyst contents showed that they were mainly composed of cholesterol esters, wax esters, lipid fractions which migrate between triglycerides and cholesterol esters but very small amounts of triglycerides, cholesterol and ceramides. In conclusion, the present results demonstrate that dermal cysts of the rhino mouse have strong similarities with sebaceous glands and outer root sheath cells. These structures can easily be isolated and could therefore serve as a 'closed sebaceous gland' model to study the physiology or differentiation of the sebaceous gland, or the effects of pharmacological agents.

Vla and alpha 6 beta 4 integrin expression in neuroendocrine carcinomas of the skin (their xenografts on nude mice and a corresponding primary culture).

Immunohistological expression of VLA1-5 and alpha 6 beta 4 integrins have been studied in 21 cases of primary neuroendocrine carcinomas of the skin (NECS), three xenografts on nude mice and one NECS cell culture. The phenotypic properties of NECS cells were largely maintained in NECS grafted on athymic nude-mice and in the corresponding cell line. Our results indicate that alpha 1 beta 1 and to a lesser extent alpha 3 beta 1, alpha 5 beta 1 are the main integrins expressed in NECS. In addition, VLA2, 4 and alpha 6 beta 4 are heterogeneously expressed in the same group of tumors and very sparsely present. These data suggest that like neuroblastoma and primitive peripheral neuroectodermal tumor (pPNET) the absence or the heterogeneous distribution of such integrins is correlated with the aggressive behaviour of NECS although long-term follow-up was not available for our cases. On the other hand, the alpha 1 expression could be regarded as a novel marker for differential diagnosis between NECS (alpha 1+) and pPNET (alpha 1-). The alpha 1 beta 1, alpha 2 beta 1, alpha 3 beta 1, alpha 5 beta 1 heterodimers in the 21 NECS studied showed an uniform pericellular staining of both the peripheral cells and central cells of the tumor islands. The predominant expression of alpha 1 beta 1 is consistent with the hypothesis of a primitive epithelial totipotential origin in NECS.

Induction of tolerance by administration of hapten-immunoglobulin conjugates is associated with decreased IL-2 and IL-4 production.

Intravenous administration of trinitrophenyl-modified isologous immunoglobulin-induced nonresponsiveness to subsequent epicutaneous painting of sensitizing doses of trinitrochlorobenzene. Isologous immunoglobulin with various degrees of trinitrophenyl substitution (11.2, 14.3, 27 and 47.3) prevented sensitization. The suppression of contact hypersensitivity was dependent on the dose of tolerogen and was hapten specific. Tolerance was inducible in mice of the strains CBA (H-2k), C57BL/6 (H-2b), and DBA/2 (H-2d) but not in Balb/C (H-2d) mice, suggesting that this trait maps outside the murine major histocompatibility complex. Tolerance induced by trinitrophenyl-modified immunoglobulin was associated with decreased hapten-induced proliferation of draining lymph-node cells. Unlike in other models of tolerance in which a decreased interleukin-2 to interleukin-4 ratio can be observed, administration of tolerizing trinitrophenylated immunoglobulin was associated with deficient hapten-induced release of both interleukin-2 and interleukin-4.

Lipid composition and barrier function of human skin after grafting onto athymic nude mice.

The barrier properties of human epidermis grafted for 1-3 months onto nude mice are compared with normal human skin. Beside penetration studies with tritiated water and measurements of transepidermal water loss (TEWL), we analyzed the epidermal lipids by high-performance thin layer chromatography and evaluated the ultrastructure of the intercorneocyte lipid arrangement by freeze fracture electron microscopy (FFEM). The permeability of human skin for tritiated water and the TEWL exhibit no significant changes after grafting onto nude mice. FFEM analysis showed that grafted epidermis has the same morphological pattern as normal human epidermis. Regular desmosomes and lamellar lipid structures are present. Grafting did not qualitatively affect the lipid composition of human epidermis. Ceramides which contribute largely to the barrier function, have the same distribution profile.

Improvement of epidermal differentiation and barrier function in reconstructed human skin after grafting onto athymic nude mice.

To determine whether epidermis reconstructed in vitro at the air-liquid interface on de-epidermized dermis has the capacity to normalize the expression of differentiation-specific markers, its lipid composition and stratum corneum barrier properties, human skin equivalents were transplanted onto athymic nude mice and investigated at different stages ranging from 1 to 4 months after grafting. Indirect immunofluorescence with species- or non-species-specific antibodies revealed that as early as 1 month after transplantation keratinization, and involucrin, loricrin and transglutaminase patterns were normalized. Human melanocytes were observed in the basal layer of the pigmented graft. As revealed by high-performance thin-layer chromatography and transmission electron microscopy after ruthenium tetroxide fixation, the lipid profile and the intracellular lamellar organization were similar to those found in natural epidermis. Transepidermal water loss measurements and penetration studies showed that the barrier properties of the reconstructed epidermis after transplantation were comparable to those of normal human skin.

Migration of Langerhans cells into human epidermis of "reconstructed" skin, normal skin, or healing skin, after grafting onto the nude mouse.

Human skin equivalents composed of keratinocytes cultured on a lattice constituted of human fibroblasts embedded in type I collagen were grafted onto the nude mouse. It is demonstrated, by indirect immunofluorescence and electron microscopy, that, after grafting, mouse Langerhans cells migrate into the human epidermis. Human Langerhans cells are not present in this system. In split-thickness human skin grafts, at long periods (5 and 12 months) after transplantation, a progressive migration of murine Ia(+) cells in the human epidermis and the presence of human Langerhans cells were shown by indirect immunofluorescence. Creation of a wound at the center of the grafted human skin and identification of the Langerhans cell origin shows a repopulation with human Langerhans cells provided the injury was performed early (2 months) after grafting. Injury at a later stage (5 months) resulted in presence of both human and murine Langerhans cells. These observations show 1) that, after grafting of "reconstructed" human skin or of split-thickness human skin onto nude mice, mouse Langerhans cells migrate into the grafted human epidermis; and 2) that the Langerhans cells repopulating a healing grafted epidermis devoid of Langerhans cells derived from the non-injured surrounding epidermis. The present work thus shows that besides bone marrow, lymph nodes, or/and spleen, surrounding cutaneous regions can also serve as sources of Langerhans cells.

Peripherin, a neuronal intermediate protein, is stably expressed by neuroendocrine carcinomas of the skin, their xenograft on nude mice, and the corresponding primary cultures.

The histogenesis of neuroendocrine carcinomas of the skin is still controversial. To determine the degree of neural differentiation of these neoplasias, we studied the expression of intermediate filament proteins in tumoral tissues. Expressions of peripherin, the neurofilament protein NF-L, vimentin, and cytokeratin 8 were analyzed by immunohistochemical methods on 12 human primary tumors and 3 tumor xenografts on nude mice. Peripherin was detected in 10 primary tumors by immunofluorescence. The protein and the corresponding messenger RNA were identified by two-dimensional gel electrophoresis and Northern analysis in extracts of an immunofluorescence-negative tumor. Peripherin, NF-L, and cytokeratin 8 were detected in tumoral cells, whereas vimentin was found exclusively in the stroma. The histological and ultrastructural properties of the original cells of neuroendocrine carcinomas of the skin, as well as coexpression of peripherin, cytokeratin 8, and neurofilament polypeptides, were preserved in tumor xenografts and their primary cultures in vitro. These results bring new elements to the knowledge of the biology of neuroendocrine carcinomas of the skin and indicate that peripherin constitutes a marker for tumor identification.

Migration of Langerhans cells into the epidermis of human skin grafted into nude mice.

In a previous study, it was demonstrated that human Langerhans cells (LC) are preserved in human skin grafted onto a nude mouse. Moreover, although it was observed that mouse LC of the host invade skin grafts from allogeneic mouse or rat, they do not penetrate in human skin grafts. In most of the human skin equivalent systems produced in vitro, LC appear to be lost. The present study was designed to investigate whether the mouse LC will repopulate a human skin equivalent. For this purpose, two different systems of skin equivalent have been grafted into the nude mouse. They were composed of human keratinocytes deposited on dead human dermis, or on lattice composed of human fibroblasts embedded in type I collagen. At different times after grafting, the presence of LC in the transplants was assayed either by indirect immunofluorescence or by electron microscopy. Indirect immunofluorescence was performed on frozen sections or on epidermal sheets with anti-Ia, anti-HLA-DR, or OKT6 antibodies. It was observed that, at 2 months after grafting, Ia(+) HLA-DR(-) OKT6(-) cells are present in grafted human epidermis. Moreover, LC with typical Birbeck granules are also detected by electron microscopy. It could be concluded, from this study, that mouse LC can repopulate human epidermis devoid of human LC.

A new marker of epidermal differentiation associated with the membrane coating granules: characterization and applications to pathology.

A murine monoclonal antibody, BC12, was obtained after immunization against suprabasal human keratinocytes. In the epidermis of normal human skin, the antigen recognized by BC12 (BC12 antigen) is located at the apex of keratinocytes in the upper stratum spinosum and stratum granulosum but is absent in other layers. The BC12 antigen is also present in hair follicles. Immunoblotting performed on keratinocyte subpopulations confirmed the presence of the BC12 antigen in differentiated keratinocytes only. Two-dimensional immunoblotting showed that the BC12 antigen corresponds to a set of polypeptides with an apparent molecular weight of approximately 33kD. In keratinocyte cultures, the antigen is present only in stratified areas. The distribution of the BC12 antigen, as studied by indirect immunofluorescence and immunoelectron microscopy, and its presence in certain subcellular fractions of epidermal cells suggest that it is a component of membrane coating granules (MCGs) or that it is associated with these structures. Strikingly, in psoriasis, eczema and many other diseases, the BC12 antibody does not label the epidermis, but vessels in dermal papillae. The BC12 antibody may thus be a useful tool in the study of keratinocyte differentiation and MCG physiology, and, also, in pathology.

The role of fibroblasts in dermal vascularization and remodeling of reconstructed human skin after transplantation onto the nude mouse.

The vascularization and the dermal remodeling of two different types of human skin reconstructed "in vitro" and grafted onto the nude mouse were studied. They were composed of human keratinocytes grown either on a human acellular deepidermized dermis (DED), or on a lattice composed of human fibroblasts embedded in bovine type I collagen, a living dermal equivalent (LDE). At different stages after grafting, the transplants were harvested and processed for an immunohistological study with species-specific and non-species-specific antibodies. At one month after grafting, the two types of grafted dermis contained blood vessels whose vascular basement membranes were labeled with a mouse-specific anti-type IV collagen antibody. With an antibody specific for human type IV collagen, a constant labeling of the vascular basement membrane was only observed in the LDE containing fibroblasts. In the DED, a constant association of the mouse endothelial cells with human type IV collagen was observed at early stages after grafting. At later stages, the human type IV collagen progressively disappeared. On the other hand, the dermal-epidermal junction underneath the human epidermis contained human type IV collagen in the two types of reconstructed skin. Labeling with the species-specific antibodies directed against human or murine type I collagen showed that the ratio murine type I collagen versus human type I collagen increased with time, suggesting that the DED is progressively invaded by mouse fibroblasts that produce the mouse collagen. On the other hand, in the LDE, the preexisting bovine type I collagen became progressively undetectable while both human type I collagen and elastic fibers were deposited by numerous human fibroblasts. Mouse type I collagen was not detected. Altogether, these observations made by grafting human skin reconstructed "in vitro" onto the nude mouse should be interesting for evaluating the usefulness of grafting a dermal substrate together with the epidermal sheet in the treatment of burns.

Sequence of morphological events during topical application of retinoic acid on the rhino mouse skin.

In a study of the comedolytic action of retinoids on the epidermal pseudocomedones of rhino mouse skin, the earliest changes and the sequence of events induced by the treatment were investigated. Rhino mice were treated topically with all-trans retinoic acid and at various time intervals from day 0 to 3 weeks, histological and ultrastructural studies as well as quantification of mast cells were performed. The earliest changes occurred in the dermis after 2 days of treatment and were characterized by degranulation of mast cells, clumping and association of Langerhans cells and lymphocytes. During the second week of treatment there was hyperplasia, hypergranulosis and an increase in the number of membrane coating granules with disorganization of the horny layer both of the epidermis and the follicular epithelium. These changes in the proliferation and the differentiation of keratinocytes resulted in a comedolytic effect. Associated with these epidermal changes there were changes in the dermis with an increased number of mast cells and a decreased number of Langerhans cells.

Localization and in vivo activity of epidermal transglutaminase.

The plasma membrane-associated transglutaminase is responsible for the formation of a cornified envelope during terminal differentiation of epidermal keratinocytes. We have studied the epidermal distribution of this enzyme ("epidermal transglutaminase") and its activity by fluorescence microscopy in histological sections of normal human skin and human skin grafted onto nude mice. Three different techniques were employed: (i) incubation of skin sections with a monoclonal antibody raised against a purified preparation of the enzyme; (ii) incubation of skin sections with dansylcadaverine, a fluorescent substrate of the enzyme, in the presence of Ca++ ions; and (iii) subcutaneous injection of dansylcadaverine into mice, at least 2 months after grafting. The first technique is supposed to detect all enzyme molecules carrying the epitope that is recognized by the antibody, even when the enzyme is catalytically not active; the second should decorate all sites in which membrane-bound transglutaminase activity is located, and the third detects only sites in which transglutaminase is active in vivo. With the first two techniques a broad band of plasma membrane associated fluorescence, reaching from the middle of the spinous layer to the stratum corneum, was detected in both normal and grafted skin. In vivo enzyme activity, however, was found to be restricted to one, or at most two, cell layers at the interface of the stratum granulosum and stratum corneum and to coincide with the layer in which the antigenicity of involucrin, a natural substrate of epidermal transglutaminase, disappeared.

Use of the chick embryo for pharmacological screening of retinoids.

Retinoic acid and some selected analogs were tested to evaluate their effect on skin morphogenesis and toxicity in the chick embryo. Retinoids dissolved in dimethyl sulfoxide were injected at doses varying from 10 pmol to 10 mumol into the amniotic cavity of 10-day-old chick embryos (n = 20). At 16 days of incubation, the eggs were opened to record the number of dead embryos and the number of embryos presenting club-shaped feathers. A lethal embryotoxic dose (LED50d16) which, at 16 days of incubation, provokes the death of 50% of the embryos and an effective dose (ED50) which induces production of club-shaped feathers in 50% of surviving embryos were then calculated using log-probit analysis. Retinoids could be classified according to their ED50. For example, arotinoid Ro 13-7410 appears approximately 1,000 times more active than all transretinoic acid. However, the analogs which are more active are also more toxic. The assay described in the present study appears to be a simple and useful model for the screening of retinoids.