Search

Tight junction (TJ) proteins are known to be involved in proliferation and differentiation. These processes are essential for normal skin wound healing. Here, we investigated the TJ proteins claudin-1 and occludin in ex vivo skin wound healing models and tissue samples of acute and chronic human wounds and observed major differences in localization/expression of these proteins, with chronic wounds often showing a loss of the proteins at the wound margins and/or in the regenerating epidermis. Knockdown experiments in primary human keratinocytes showed that decreased claudin-1 expression resulted in significantly impaired scratch wound healing, with delayed migration and reduced proliferation. Activation of AKT pathway was significantly attenuated after claudin-1 knockdown, and protein levels of extracellular signal-related kinase 1/2 were reduced. For occludin, down-regulation had no impact on wound healing in normal scratch assays, but after subjecting the cells to mechanical stress, which is normally present in wounds, wound healing was impaired. For both proteins we show that most of these actions are independent from the formation of barrier-forming TJ structures, thus demonstrating nonbarrier-related functions of TJ proteins in the skin. However, for claudin-1 effects on scratch wound healing were more pronounced when TJs could form. Together, our findings provide evidence for a role of claudin-1 and occludin in epidermal regeneration with potential clinical importance., (Copyright © 2017 American Society for Investigative Pathology. Published by Elsevier Inc. All rights reserved.)

Background: Adjuvant treatment with interferon (IFN)-α-2a improved disease-free survival (DFS) and showed a trend for improving overall survival (OS) in melanoma. This trial was designed to examine whether PEG-IFN is superior to IFN with regard to distant metastasis-free survival (DMFS), DFS and OS., Patients and Methods: In this multicenter, open-label, prospective randomized phase III trial, patients with resected cutaneous melanoma stage IIA(T3a)-IIIB (AJCC 2002) were randomized to receive PEG-IFN (180 μg subcutaneously 1×/week; 24 months) or IFN α-2a (3MIU subcutaneously 3×/week; 24 months). Randomization was stratified for stage, number of metastatic nodes, age and previous IFN treatment. The primary end point was DMFS; secondary end points were OS, DFS, quality of life (QoL) and tolerability., Results: A total of 909 patients were enrolled (451 PEG-IFN versus 458 IFN). Neither 5-year DMFS [PEG-IFN 61.0% versus IFN 67.3%; hazard ratio (HR) 1.16, P = 0.21] nor 5-year OS (PEG-IFN 73.2% versus IFN 75.2%; HR 1.05, P = 0.70) nor 5-year DFS (PEG-IFN 57.3% versus IFN 60.9%; HR 1.09, P = 0.40) showed significant differences. Subgroup analyses in patients ± ulcerated primaries and of different tumor stages did not find differences in DMFS, OS or DFS between the treatment groups. One hundred and eighteen patients (26.2%) in the PEG-IFN and 61 patients (13.3%) in the IFN population did not receive the full dosage and length of treatment due to adverse events (P < 0.001). Leukopenia and elevation of liver enzymes were more common in the PEG-IFN arm (56% versus 23.5% LCP; 19.1% versus 9.4% AST; 33.0% versus 16.5% ALT). QoL was identical for nearly all domains., Conclusion: PEG-IFN did not improve the outcome over IFN. A higher percentage of patients under PEG-IFN discontinued treatment due to toxicity., Clinical Trialsgov Identifier: NCT00204529., (© The Author 2016. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Follicular penetration has gained increasing interest regarding (i) safety concerns about (environmentally born) xenobiotics available to the hair follicle (HF), e.g. nanomaterials or allergens which should not enter the skin, and (ii) the possibility for non-invasive follicular drug and antigen delivery. However, not much is known about barriers in the HF which have to be surpassed upon uptake and/or penetration into surrounding tissue. Thus, aim of this work was a detailed investigation of this follicular barrier function, as well as particle uptake into the HF of porcine skin which is often used as a model system for human skin for such purposes. We show that follicular tight junctions (TJs) form a continuous barrier from the infundibulum down to the suprabulbar region, complementary to the stratum corneum in the most exposed upper follicular region, but remaining as the only barrier in the less accessible lower follicular regions. In the bulbar region of the HF no TJ barrier was found, demonstrating the importance of freely supplying this hair-forming part with e.g. nutrients or hormones from the dermal microenvironment. Moreover, the dynamic character of the follicular TJ barrier was shown by modulating its permeability using EDTA. After applying polymeric model-nanoparticles (154 nm) to the skin, transmission electron microscopy revealed that the majority of the particles were localized in the upper part of the HF where the double-barrier is present. Only few penetrated deeper, reaching regions where TJs act as the only barrier, and no particles were observed in the bulbar, barrier-less region. Lastly, the equivalent expression and distribution of TJ proteins in human and porcine HF further supports the suitability of porcine skin as a predictive model to study the follicular penetration and further biological effects of dermally applied nanomaterials in humans., (Copyright © 2016 Elsevier GmbH. All rights reserved.)

Tight junctions are important for skin barrier function. The tight junction protein claudin 1 (Cldn-1) has been reported to be down-regulated in nonlesional skin of atopic dermatitis (AD) patients. In contrast, we did not observe a significant down-regulation of Cldn-1 in nonlesional skin of the AD cohort used in this study. However, for the first time, a significant down-regulation of Cldn-1 in the upper and lower epidermal layers of lesional skin was detected. In addition, there was a significant up-regulation of Cldn-4 in nonlesional, but not lesional, AD skin. For occludin, no significant alterations were observed. In an AD-like allergic dermatitis mouse model, Cldn-1 down-regulation in eczema was significantly influenced by dermal inflammation, and significantly correlated with hallmarks of eczema (ie, increased keratinocyte proliferation, altered keratinocyte differentiation, increased epidermal thickness, and impaired barrier function). In human epidermal equivalents, the addition of IL-4, IL-13, and IL-31 resulted in a down-regulation of Cldn-1, and Cldn1 knockdown in keratinocytes resulted in abnormal differentiation. In summary, we provide the first evidence that Cldn-1 and Cldn-4 are differentially involved in AD pathogenesis. Our data suggest a role of Cldn-1 in AD eczema formation triggered by inflammation., (Copyright © 2015 American Society for Investigative Pathology. Published by Elsevier Inc. All rights reserved.)

Tight junctions (TJs) form a selective barrier for ions, water, and macromolecules in simple epithelia. In keratinocytes and epidermis, TJs were shown to be involved in individual barrier functions. The absence of the TJ protein claudin-1 (Cldn1) in mice results in a skin-barrier defect characterized by lethal water loss. However, detailed molecular analyses of the various TJ barriers in keratinocytes and the contribution of distinct TJ proteins are missing. Herein, we discriminate TJ-dependent paracellular resistance from transcellular resistance in cultured keratinocytes using the two-path impedance spectroscopy. We demonstrate that keratinocyte TJs form a barrier for Na(+), Cl(-), and Ca(2+), and contribute to barrier function for water and larger molecules of different size. In addition, knockdown of Cldn1, Cldn4, occludin, and zonula occludens-1 increased paracellular permeabilities for ions and larger molecules, demonstrating that all of these TJ proteins contribute to barrier formation. Remarkably, Cldn1 and Cldn4 are not critical for TJ barrier function for water in submerged keratinocyte cultures. However, Cldn1 influences stratum corneum (SC) proteins important for SC water barrier function, and is crucial for TJ barrier formation for allergen-sized macromolecules.

Upon barrier disturbance, adult CD44 knockout (KO) mice show delayed recovery of epidermal barrier function. This correlates with the loss of apical polarization of lamellar body (LB) secretion. As tight junctions (TJs) are crucial for barrier function and regulate polarized targeting of vesicles, we hypothesized that CD44 regulates TJs and associated cell polarity complexes, which in turn contributes to altered skin barrier function in CD44 KO mice. We show a delay in embryonic barrier formation associated with a loss of apical LB localization in CD44 KO mice, which correlates with alterations in TJ proteins and Par3. Simultaneously, the activity of Rac1, a major regulator of TJ barrier function, was reduced. Importantly, normalization of barrier function at E18.5 coincided with the recovery of these proteins. Tape-stripping experiments revealed that the loss of CD44 also affected TJ proteins upon induced disturbance of the barrier in adult mice. In CD44 KO keratinocytes, cell polarization and TJ barrier function were impaired. An alteration of differentiation markers was also observed, but was less pronounced than alterations of TJ proteins. Taken together, the results reveal an important function for CD44 in the assembly and function of TJs, suggesting their involvement in the skin barrier phenotype of CD44 KO mice.

Tight junctions (TJ) are cell-cell junctions that have proved to form a paracellular barrier for solutes and water between cells of epithelia, including the stratum granulosum of the stratified epithelium of the epidermis of newborn mice. In mice lacking claudin-1, a major barrier-forming TJ component, this barrier was abolished. However, the role of TJ in human skin is controversially discussed as unambiguous data were missing so far. Here, we investigated TJ barrier function in healthy human skin as well as in skin samples from psoriatic lesions which are characterized by an altered localization of TJ proteins. We demonstrate for human skin that occludin- and claudin-1-positive sites in the stratum granulosum form a barrier for extracellular biotin-SH (557Da) and that in psoriatic skin the localization of the barrier and the TJ proteins are altered in parallel., (Copyright 2010 Elsevier GmbH. All rights reserved.)

Psoriasis is an inflammatory skin disease characterized by hyperproliferation of keratinocytes, impaired barrier function, and pronounced infiltration of inflammatory cells. Tight junctions (TJs) are cell-cell junctions that form paracellular barriers for solutes and inflammatory cells. Altered localization of TJ proteins in the epidermis was described in plaque-type psoriasis. Here we show that localization of TJ proteins is already altered in early-stage psoriasis. Occludin, ZO-1, and claudin-4 are found in more layers than in normal epidermis, and claudin-1 and -7 are down-regulated in the basal and in the uppermost layers. In plaque-type psoriasis, the staining patterns of occludin and ZO-1 do not change, whereas the claudins are further down-regulated. Near transmigrating granulocytes, all TJ proteins except for junctional adhesion molecule-A are down-regulated. Treatment of cultured keratinocytes with interleukin-1beta and tumor necrosis factor-alpha, which are present at elevated levels in psoriatic skin, results in an increase of transepithelial resistance at early time points and a decrease at later time points. Injection of interleukin-1beta into an ex vivo skin model leads to an up-regulation of occludin and ZO-1, resembling TJ protein alteration in early psoriasis. Our results show for the first time that alteration of TJ proteins is an early event in psoriasis and is not the consequence of the more profound changes found in plaque-type psoriasis. Our data indicate that cytokines are involved in alterations of TJ proteins observed in psoriasis.

Bacterial translation initiation factor 3 (IF3) is involved in the fidelity of translation initiation at several levels, including start-codon discrimination, mRNA translation, and initiator-tRNA selection. The IF3 C-terminal domain (CTD) is required for binding to the 30S ribosomal subunit. N-terminal domain (NTD) function is less certain, but likely contributes to initiation fidelity. Point mutations in either domain can decrease initiation fidelity, but C-terminal domain mutations may be indirect. Here, the Y75N substitution mutation in the NTD is examined in vitro and in vivo. IF3(Y75N) protein binds 30S subunits normally, but is defective in start-codon discrimination, inhibition of initiation on leaderless mRNA, and initiator-tRNA selection, thereby establishing a direct role for the IF3 NTD in these initiation processes. A model illustrating how IF3 modulates an inherent function of the 30S subunit is discussed.

Tight Junction (TJ) proteins have been shown to exert a barrier function within the skin. Here, we study the fate of TJ proteins during the challenge of the skin by bacterial colonization and infection. We investigated the influence of various exfoliative toxin-negative Staphylococcus strains on TJ, adherens junction (AJ), desmosomal proteins, and actin in a human keratinocyte infection culture and in a porcine skin infection model. We found that the pathogen Staphylococcus aureus downregulates TJ and subsequently AJ and desmosomal proteins, including atypical protein kinase C, an essential player in TJ formation, at the cell-cell borders of keratinocytes in a time and concentration dependent manner. Little changes in protein and RNA levels were seen, indicating redistribution of proteins. In cultured keratinocytes, a reduction of transepithelial resistance was observed. Staphylococcus epidermidis shows only minor effects. All strains induced enhanced expression of occludin and ZO-1 at the beginning of colonization/infection. Thus, we demonstrate that TJ are likely to be involved in skin infection of exfoliative toxin-negative S. aureus. As we did not find a change in actin, and as changes of TJ preceded alterations of AJs and desmosomes, we suggest that S. aureus targets TJ.

In stratified squamous epithelia constituent proteins of tight junctions (TJs) are not restricted to the zonula occludens-related structures of the uppermost living cell layer such as the stratum granulosum of the epidermis but TJ membrane proteins such as occludin and certain members of the claudin family as well as TJ plaque proteins, notably cingulin and protein ZO-1, have also been identified by immunofluorescence and immunoelectron microscopy in more basal layers where they form special cell-cell-connecting structures such as the "lamellated" and the "sandwich" junctions. In the present study, we describe another TJ protein-containing structure, the very small puncta occludentia ("stud junctions"), as the smallest identifiable TJ-like unit that occurs in most, perhaps all strata. We have also determined the specific distributions of TJ proteins in the cell layers of squamous cell metaplasias of the human bronchial tract. Moreover, we show that the occludin-related tetraspanin protein tricellulin-alpha connects and seals the membranes of adjacent "three corner" cell structures of the uppermost layer in keratinocytes growing in culture. We hypothesize the possible occurrence of tricellulin-beta in more basal cell layers of keratinocyte cultures and the general occurrence of different tricellulin splice forms in stratified epithelia in situ, and discuss the possible functions of TJ proteins in stratified epithelia and tumors derived therefrom.

Cutaneous wound healing is a well-coordinated process that includes inflammation, proliferation, and differentiation. Activator protein 1 (AP-1) subunits have been implicated in the regulation of genes important for these processes and have been shown to be involved in wound healing. However, investigation of human healing and non-healing wounds in vivo and ex vivo, and the comparative analysis of several members of the Jun and Fos families are still missing. Here, we show that normal human epidermal wound healing is biphasic. In the first phase all AP-1 subunits investigated, that is c-Jun, Jun B, Jun D, c-Fos, and Fos B are absent from the nuclei at the wound margins/leading edges. This downregulation coincides with that of the gap junction protein connexin 43. Later on, c-Jun, Jun B, Jun D, and c-Fos reappear in the nuclei of the leading edges in a time-dependent manner. In non-healing wounds, a more intensive staining of keratinocytes at the wound margins is often observed. Our findings suggest that coordinated down- and upregulation of the various AP-1 subunits in the course of epidermal wound healing is important for its undisturbed progress, putatively by influencing inflammation and cell-cell communication.

Human Merkel cells were first described by Friedrich S. Merkel in 1875 and named "Tastzellen" (touch cells) assuming a sensory touch function within the skin. Only ultrastructural research revealed their characteristics such as dense-core granules, plasma membrane spines and dendrites as well as a loosely arranged cytoskeleton. Biochemical analysis identified the expression of very specific cytokeratins (most notably CK 20) allowing the immunohistochemical detection of Merkel cells. In humans, they occur within the basal epidermis, being concentrated in eccrine glandular ridges of glabrous skin and in Haarscheiben of hairy skin, within belt-like clusters of hair follicles, and in certain mucosal tissues. Within the human skin, the dense-core granules contain heterogeneously distributed neuropeptides, some of which might work as neurotransmitters through which Merkel cells and their associated nerves exert their classical function as slowly adapting mechanoreceptors type I. This is the case in the Haarscheiben, small sensory organs containing keratinocytes with a special program of differentiation that includes the expression of CK 17 and Ber-EP4. Other peptides may act as growth factors and thus might participate in growth, differentiation and homeostasis of cutaneous structures. It is not yet clear whether the Merkel cell carcinomas, aggressive skin carcinomas, indeed arise from Merkel cells. We summarize and discuss data on the distribution, function and heterogeneity of human Merkel cells in normal and diseased skin.

Although aqueous mistletoe extracts are widely used in complementary cancer therapy, the precise mode of action of their main therapeutic agents, the three mistletoe lectins (MLs), is poorly understood as they act both as cytotoxic agents and as immunomodulators due to their cytokine release by mononuclear cells. Thus, this study aims to investigate both the direct and the indirect effects of MLs on the growth of human melanoma cells in vitro. Proliferation of six human melanoma cell lines under ML treatment and additionally under the influence of cytokines induced by them (TNF-alpha, IL-1, IL-6) was assessed by means of the tetrazolium derived reduction (XTT) assay. Furthermore, ML binding patterns were analysed and correlated with the biological effects. All three MLs inhibited melanoma cell proliferation in a dose-dependent manner starting at very low ML concentrations (0.001-100 ng/ml) with ML-I being the most cytotoxic lectin (significant inhibition of ultra-sensitive cell line MV3 at 1 x 10(-13) ng ML-I/ml). Even if applied in a broad concentration range (0.0001-100 ng/ml) cytokines had no influence on cell proliferation at all. For ML-I, no association between binding intensity and cytotoxicity was observed, while for ML-II and -III an association between binding and toxicity was established. In conclusion, this study emphasises the direct anti-proliferative effect of the mistletoe lectins on melanoma cells with ML-I being superior to MLs-II and -III. The observation of an ultra-sensitivity of one cell line towards ML-I toxicity may serve as an explanation for the therapeutic success in anecdotal case reports and needs further investigations.

In the literature the question of whether a system structurally and functionally related to the barrier function of the tight junctions (TJs) of polarized epithelia exists in the epidermis has been and still is controversially discussed. We have systematically addressed this question in a study of the granular layer of fetal and adult human epidermis, combining different light and electron microscopic methods. We show that the lateral membranes of the cells of the stratum granulosum are connected by an extended subapical complex system integrating desmosomes and TJ structures identified as sites of close membrane-membrane contact and as regions of membrane-to-membrane apposition that in immunoelectron microscopy are positive for TJ marker proteins, notably occludin, indicative of an extended, probably continuous TJ barrier. In addition, we have noted in freeze-fractures of the apical membrane attaching this layer to the basalmost membrane of the stratum corneum an extended system integrating desmosomes with intramembraneous ridge configurations that appear as strands, circles, lariats or complex meshworks showing numerous continuities with the desmosomes. In some regions this system interconnecting desmosomes with curvilinear ridge structures occupies the major part of the plasma membrane. The molecular organizations and possible functional contributions of both structural systems positioned at the border between the living portion of the epidermis and the corneal layer are discussed, in particular in relation to the formation of a stable association between the two layers and of a barrier to the paracellular flow of molecules and particles. It is also discussed whether similar structures occur in other keratinizing stratified squamous epithelia, in squamous metaplasias and in tumors derived from such tissues.

The ring-shaped RNA chaperone Hfq has recently received much attention owing to its multiple roles in RNA metabolism. In this study we have performed a mutational analysis of the Escherichia coli hfq gene, and have studied the effects of amino acid substitutions at several positions in the Hfq protein as well as of C-terminal truncations on its role in phage Qbeta replication, in repression of a target mRNA, and on the stability of the small regulatory RNA DsrA. These functional studies provided insights into the interaction of Hfq with RNA and suggested a role for the C-terminus of Hfq in DsrA stability.

Gap junctions (GJ) are known to be involved in spontaneous wound healing in rodent skin. We analyzed the staining patterns of the GJ proteins Cx26, Cx30, and Cx43 in human cutaneous wound healing and compared ex vivo spontaneous wound healing to non-healing wounds (chronic leg ulcers) and to ex vivo accelerated wound healing after transplantation of cultured keratinocytes. We demonstrate a loss of Cx43 staining at the wound margins during initial wound healing and after transplantation of keratinocytes. In contrast, Cx43 remains present at the margins of most non-healing wounds. We show a subsequent induction of Cx26 and Cx30 near the wound margins in spontaneous wound healing and-even earlier-after the transplantation of keratinocytes. The cells at the wound margins remain negative until the commencement of epidermal regeneration. Cx26/30 are present at the wound margins of most non-healing wounds. Cx stainings are absent in the transplanted keratinocytes during early wound healing, but there is a subsequent induction. Our results suggest that the downregulation of Cx43 is an important event in human wound healing. We discuss the assumption that direct cell-cell communication via GJ contribute to the acceleration of wound healing after the transplantation of keratinocytes.

RNase E is known to affect the turnover of ompA mRNA in a growth rate-dependent manner. Here, we show that this enzyme also plays a role in the temperature-dependent stability of the transcript, thereby maintaining comparable levels of OmpA at 28 and 37 degrees C. An increase in the efficiency of RNase E cleavages at 37 degrees C within the 5(') UTR of the transcript in vitro was found to correlate with a decreased half-life and steady-state level at elevated temperature in vivo. However, measurements of de novo OmpA synthesis and in vitro toeprinting experiments suggest that translation of ompA mRNA is more efficient at 37 degrees C when compared to 28 degrees C. Thus, the enhanced translation apparently counteracts the decreased half-life at elevated temperature. Moreover, we propose that the temperature-dependent inverse correlation between ompA mRNA stability and translation can result from structural changes induced in the 5(') UTR of the transcript.

Tight junctions (TJs), hallmark structures of one-layered epithelia and of endothelia, are of central biological importance as intramembranous "fences" and as hydrophobic "barriers" between lumina represented by liquid- or gas-filled spaces on the one hand and the mesenchymal space on the other. They have long been thought to be absent from stratified epithelia. Recently, however, constitutive TJ proteins and TJ-related structures have also been identified in squamous stratified epithelia, including the epidermis, where they occur in special positions, most prominently in the uppermost living epidermal cell layer, the stratum granulosum. Much to our surprise, however, we have now also discovered several major TJ proteins (claudins 1 and 4, occludin, cingulin, symplekin, protein ZO-1) and TJ-related structures in specific positions of formations of epithelium-derived tissues that lack any lumen and do not border on luminal or body surfaces. Using immunohistochemistry and electron microscopy we have localized TJ proteins and structures in peripheral cells of the Hassall's corpuscles of human and bovine thymi as well as in specific central formations of tumor nests in squamous cell carcinomas, including the so-called "horn pearls". Such structures have even been found in carcinoma metastases. In carcinomas, they often seem to separate certain tumor regions from others or from stroma. The structural significance and the possible functional relevance of the locally restricted synthesis of TJ proteins and of the formations of TJ-related structures are discussed. It is proposed to include the determination of the presence or absence of such proteins and structures in the diagnostic program of tumor pathology.

In contrast to canonical mRNAs, translation of leaderless mRNA has been previously reported to continue in the presence of the antibiotic kasugamycin. Here, we have studied the effect of the antibiotic on determinants known to affect translation of leadered and leaderless mRNAs. Kasugamycin did not affect the Shine-Dalgarno (SD)-anti-SD (aSD) interaction or the function of translation initiation factor 3 (IF3). Thus, the preferential translation of leaderless mRNA in the presence of kasugamycin can neither be attributed to an expanding pool of 30S subunits with a "blocked" aSD nor to a lack of action of IF3, which has been shown to discriminate against translation initiation at 5'-terminal start codons. Using toeprinting, we observed that on leaderless mRNA 70S in contrast to 30S translation initiation complexes are comparatively resistant to the antibiotic. These results taken together with the known preference of 70S ribosomes for 5'-terminal AUGs lend support to the hypothesis that translation of leaderless mRNAs may as well proceed via an alternative initiation pathway accomplished by intact 70S ribosomes.

The occurrence of extended tight junction (TJ) structures, including zonulae occludentes (ZO), and the spatial arrangement of TJ proteins in stratified mammalian epithelia has long been controversially discussed. Therefore, we have systematically examined the localization of TJ proteins in diverse stratified epithelial tissues (e.g., epidermis, heel pad, snout, gingiva, tongue, esophagus, exocervix, vagina, urothelium, cornea) of various species (human, bovine, rodents) as well as in human cell culture lines derived from stratified epithelia, by electron microscopy as well as by immunocytochemistry at both the light and the electron microscopic level, using antibodies to TJ proteins such as occludin, claudins 1 and 4, protein ZO-1, cingulin and symplekin. We have found an unexpected diversity of TJ-related structures of which only those showing colocalization with the most restricted transmembrane TJ marker protein, occludin, are presented here. While in epidermis and urothelium occludin is restricted to the uppermost living cell layer, TJ-related junctions are abundant in the upper third or even in the majority of the suprabasal cell layers in other stratified epithelia. Interfollicular epidermis contains, in the stratum granulosum, extended, probably continuous ZO-like structures which can also be traced at least through the Henle cell layer of hair follicles. Similar apical ZO-like structures have been seen in the upper living cell layers of all other stratified epithelia and cell cultures examined, but in most of them we have noticed, in addition, junctional regions showing relatively broad, ribbon-like membrane contacts which in cross-section often appear pentalaminar, with an electron-dense middle lamella ("lamellated TJs", coniunctiones laminosae). In suprabasal layers of several stratified epithelia we have further observed TJ protein-containing junctions of variable sizes which are characterized by a 10-30-nm dense lamina interposed between the two membranes ("sandwich junctions"; iuncturae structae). Moreover, we have often observed variously sized regions in which the intermembrane distance is rather regularly bridged by short rod-like elements ("cross-bridged cell walls"; parietes transtillati), often in close vicinity of TJ-related structures or desmosomes. The significance of these structures and their possible biological importance are discussed.

Occludin and several proteins of the claudin family have been identiried in simple epithelia and in endothelia as major and structure-determining transmembrane proteins clustered in the barrier-forming tight junctions (TJ), where they are associated with a variety of TJ plaque proteins, including protein ZO-1. To examine whether TJ also occur in the squamous stratified epithelium of the interfollicular human epidermis we have applied several microscopic and biochemical techniques. Using RT-PCR techniques, we have identiried mRNAs encoding protein ZO-1, occludin and claudins 1, 4, 7, 8, 11, 12, and 17 in both tissues, skin and cultured keratinocytes, whereas claudins i and 10 have only been detected in skin tissue. By immunocytochemistry we have localized claudin-1, occludin and protein ZO-1 in distinct plasma membrane structures representing cell-cell attachment zones. While claudin-1 occurs in plasma membranes of all living cell layers, protein ZO-1 is concentrated in or even restricted to the uppermost layers, and occludin is often detected only in the stratum granulosum. Using electron microscopy, typical TJ structures ("kissing points") as well as some other apparently related junctional structures have been detected in the stratum granulosum, interspersed between desmosomes. Modes and patterns of TJ formation have also been studied in experimental model systems, e.g., during wound healing and stratification as well as in keratinocyte cultures during Ca2+-induced stratification. We conclude that the epidermis contains in the stratum granulosum a continuous zonula occludens-equivalent structure with typical TJ morphology and molecular composition, characterized by colocalization of occludin, claudins and TJ plaque proteins. In addition, cell-cell contact structures and certain TJ proteins can also be detected in other epidermal cell layers in specific cell contacts. The pattern of formation and possible functions of epidermal TJ and related structures are discussed.

Drebrin, an actin-binding 70-kDa protein with an unusually slow SDS-PAGE mobility corresponding to approximately 120 kDa, containing a proline-rich, profilin-binding motif, had originally been reported from neuronal cells, but recently has also been found in diverse other kinds of tissues and cell lines. In biochemical analyses of various cells and tissues, employing gel filtration, sucrose gradient centrifugation, immunoprecipitation and -blotting, we have identified distinct states of soluble drebrin: a approximately 4S monomer, an 8S, ca. 217-kDa putative trimer, a 13S and a > 20S oligomer. In the 8S particles only [35S]methionine-labelled drebrin but no other actin-binding protein has been detected in stoichiometric amounts. By immunofluorescence and immunoelectron microscopy, drebrin-positive material often appeared as "granules" up to 400 nm in diameter, in some cell types clustered near the Golgi apparatus or in lamellipodia, particularly at leading edges, or in dense-packed submembranous masses at tips (acropodia) or ruffles of leading edges, in filopodia and at plaques of adhering junctions. We conclude that these drebrin complexes and drebrin-rich structures allow the build-up and maintenance of high local drebrin concentrations in strategic positions for the regulation of actin filament assembly, thereby contributing to cell motility and morphology, in particular local changes of plasticity and the formation of protrusions.

After the completion of skin development, angiogenesis, i.e., the growth of new capillaries from pre-existing blood vessels, is held to occur in the skin only under pathologic conditions. It has long been noted, however, that hair follicle cycling is associated with prominent changes in skin perfusion, that the epithelial hair bulbs of anagen follicles display angiogenic properties, and that the follicular dermal papilla can produce angiogenic factors. Despite these suggestive observations, no formal proof is as yet available for the concept that angiogenesis is a physiologic event that occurs all over the mature mammalian integument whenever hair follicles switch from resting (telogen) to active growth (anagen). This study uses quantitative histomorphometry and double-immunohistologic detection techniques for the demarcation of proliferating endothelial cells, to show that synchronized hair follicle cycling in adolescent C57BL/6 mice is associated with substantial angiogenesis, and that inhibiting angiogenesis in vivo by the intraperitoneal application of a fumagillin derivative retards experimentally induced anagen development in these mice. Thus, angiogenesis is a physiologic event in normal postnatal murine skin, apparently is dictated by the hair follicle, and appears to be required for normal anagen development. Anagen-associated angiogenesis offers an attractive model for identifying the physiologic controls of cutaneous angiogenesis, and an interesting system for screening the effects of potential antiangiogenic drugs in vivo.

Merkel cell carcinoma was first described in 1972 by Toker and is an aggressive neuroendocrine skin tumor with a high metastatic potential. Merkel cell carcinoma is thought to derive from the neuroendocrine (Merkel) cells of the skin, although in contrast to fetal and especially adult Merkel cells, Merkel cell carcinomas express high levels of the Bcl-2 oncoprotein. Bcl-2 is capable of blocking programmed cell death and has been shown to play an important role in normal cell turnover, tumor biology, and chemoresistance. High Bcl-2 expression leading to prolonged survival of cells may therefore be of importance in the biological and clinical characteristics of Merkel cell carcinoma. In a SCID mouse xenotransplantation model for human Merkel cell carcinoma, we investigated the influence of the bcl-2 antisense oligonucleotide G3139 (Genta) on tumor growth in comparison with control oligonucleotides or cisplatin. Bcl-2 antisense treatment, targeting the first six codons of the bcl-2 mRNA, resulted in either a dramatic reduction of tumor growth or complete remission, whereas reverse sequence and two-base mismatch control oligonucleotides or cisplatin had no significant antitumor effects compared with saline-treated controls. Apoptosis was enhanced 2.4-fold in the bcl-2 antisense treated tumors compared with the saline-treated group, and no other treatment showed a comparable increase in apoptosis. Our findings suggest that bcl-2 antisense treatment may be a novel approach to improve treatment outcome of human Merkel cell carcinoma.

Growth factors of the transforming growth factor-beta superfamily are involved in cutaneous wound healing. In this study we analyze the expression of the bone morphogenetic protein-6 (BMP-6) gene, a transforming growth factor-beta related gene, in skin wounds. In normal mouse skin high levels of BMP-6 mRNA and protein are expressed by postmitotic keratinocytes of stratified epidermis until day 6 after birth. BMP-6 expression is strongly reduced in adult epidermis with diminished mitotic activity. After skin injury we found large induction of BMP-6-specific RNA and protein in keratinocytes at the wound edge and keratinocytes of the newly formed epithelium as well as in fibroblast shaped cells in the wound bed. BMP-6-specific RNA was induced within 24 h after injury, whereas significant upregulation of BMP-6 on the protein level was detected only 2-3 d after injury. Protein was confined to outermost suprabasal epidermal layers, whereas BMP-6-specific RNA was distributed throughout all epidermal layers including basal keratinocytes and the leading edge of the migrating keratinocytes. We also detected high levels of BMP-6-specific RNA and protein in chronic human wounds of different etiology. In contrast to the overall distribution pattern of BMP-6-specific RNA, the protein was not detected in keratinocytes directly bordering the wound. In order to test the influence of BMP-6 abundance on the progress of wound healing, we analyzed the wound response of transgenic mice overexpressing BMP-6 in the epidermis. In these mice, reepitheliazation of skin wounds was significantly delayed, suggesting that strict spatial and temporal regulation of BMP-6 expression is necessary not only for formation but also for reestablishment of a fully differentiated epidermis.

Hair follicles are complex organs of the skin, in morphological and ontogenic continuity with the epidermis. We have examined the location of desmosomal cadherins and desmosomal plaque proteins in the hair follicle of adult and fetal human scalp skin by immunohistochemistry and have established a localization "map" of the hair follicle. Using antibodies against the plaque proteins desmoplakin I and II, plakoglobin, and plakophilin 1, we have found that these occur in most, if not all hair follicle desmosomes, whereas plakophilin 2 was absent, except in the basal cells of the outer root sheath, where a weak reactivity was found. By contrast, the desmosomal cadherins were mostly differentially synthesized, displaying a complicated map. While desmocollin Dsc3 was detected in all cell types examined, Dsc1 was detected only in the outer root sheath companion cell layer and the inner root sheath, and Dsc2 showed practically a mutually exclusive presence. Desmoglein Dsg2 was observed in basal cells of the outer root sheath as well as in the central cell layers of the subinfundibular outer rood sheath, matrix cells and trichocytes, in partial overlap with the otherwise different immunopositive reactions of Dsg1 and Dsg3. We have also determined when these proteins are synthesized during fetal hair follicle development. The differential molecular composition of desmosomes is discussed in relation to possible functional differences between the individual cell types.

Few data are available on early regeneration of human epidermis in vivo. We have established a supravital skin organ culture model for epidermal wound healing by setting a central defect (3 mm diameter) in freshly excised skin specimens and culturing under air exposure. Re-epithelialization was followed for up to 7 d by histology and immunohistologic analysis of various markers for differentiation and proliferation. In 12 of 19 cases (63%; 5% fetal calf serum) or six of 21 cases (29%; 2% fetal calf serum), the wounds were re-epithelialized spontaneously after 7 d. After transplantation to the wounds of 1-2 x 10(6) dissociated allogenic cultured epidermal or about 1 x 10(6) autologous outer root sheath keratinocytes, 18 of 21 cases (86%; 5% fetal calf serum) or 17 of 21 cases (81%; 2% fetal calf serum) were healed within the same period. Histologically, early neoepithelium (3 d) was disordered after keratinocyte transplantation, whereas later (7 d) it had gained a more ordered stratification, exhibiting a thin discontinuous granular and a compact horny layer. At this stage, not only hyperproliferative (CK 6) but also, abundantly, maturation-associated cytokeratins (CK 1, CK 10) were detected immunohistochemically. Analyses of regenerated epidermis after transplantation of (i) keratinocytes labeled in vitro with BrdU and (ii) heterosexual keratinocytes by immunohistochemistry and fluorescence in situ hybridization for the Y chromosome, respectively, clearly showed that external keratinocytes are physically integrated into the regenerated epidermis and extendedly contribute to its formation. The data presented here demonstrate improvement and acceleration of epidermal re-epithelialization by transplantation of keratinocytes.

Desmosomes are predominant among the types of plaque-bearing adhering junctions found in human skin. These structures contain a set of desmosomal cadherins and cytoplasmic plaque proteins, the synthesis of which is differentiation dependent. As plakophilin 1, a member of the armadillo gene family, is an important accessory desmosomal plaque protein, we raised several monoclonal antibodies specific for this protein and applied immunohistochemical and immunoblotting procedures to study the distribution of plakophilin 1 in desmosomes in adult and fetal skin, psoriatic epidermis, various epithelial skin tumors, and keratinocyte sheets grown in culture. In epidermis, the spinous layers were prominently immunostained by plakophilin 1 antibodies, whereas the basal cell layer was only weakly stained and the stratum corneum was entirely unstained. The staining observed in psoriatic epidermis was somewhat heterogeneous. In hair follicles, the outer root sheath (ORS) was delineated in its suprabasal cell layers, with variable staining in its upper and lower parts. All basal cells of the ORS remained unstained, as did upper inner root sheath (IRS) and matrix cells of lower bulb. In eccrine sweat glands, the reaction was confined to inner dermal ductal cells, with the acini remaining unstained. The desmosomal immunostaining observed in basal cell carcinomas (BCCs) and squamous cell carcinomas (SCCs) was very heterogeneous: In general, junctions in well-differentiated stratified tumor regions were more intensely stained than sections of poorly differentiated and invasively growing BCCs and SCCs. Plakophilin 1 was also prominent in the desmosomes of keratinocyte sheets grown in culture. The cell type-specific, i.e., differentiation-dependent, distribution of desmosomal plakophilin 1 is discussed in relation both to the stratification of the cutaneous epithelia and to tumor differentiation and growth.

As part of a program designed to characterize human hair keratin genes and their expression, we present the cDNA sequences and deduced amino acid sequences of three type-II hair keratins hHb3, hHb5, and hHb6, which by virtue of their amino acid homologies are the orthologs of the previously described sheep wool keratins, K2.10, K2.12, and K.211 [29]. Amino acid sequences comparisons of these keratins, including the previously characterized human K2.9 ortholog hHb1, show extreme conservation not only in the alpha-helices but also in the aminoterminal and proximal carboxyterminal domains. They also demonstrate higher sequence relationships between hHb1, hHb3, and hHb6 as compared to hHb5, which exhibits chain-specific sequences in both the head and tail domains. In situ hybridization studies using specific 3'-probes for the four type-II hair keratins reveal sequential patterns of gene expression in human anagen follicles. Remarkably the onset of hHb5 mRNA synthesis occurs immediately above a small population of matrix cells at the base of the hair bulb and the trichocytes lining the dermal papilla. hHb5 mRNA synthesis extends upward through the matrix and ends in the lower part of the cortex of the hair shaft. In contrast, both hHb1 and hHb3 mRNA synthesis begins simultaneously in the cortex 10-15 cell layers above the apex of the dermal papilla, thus partially overlapping that of hHb5 but continuing to a point well beyond hHb5 in the upper cortex. Synthesis of hHb6 mRNA starts slightly higher than either hHb1 or hHb3 mRNA and proceeds much farther up into the keratogenous zone of the hair shaft. Our study demonstrates that the differentiation of human hair in terms of hair keratin expression begins much earlier than previously assumed, i.e. in lower matrix cells of the hair bulb. This early phase of hair differentiation is followed by a late cortical phase of terminal differentiation which comprises at least three type-II hair keratins in the zone of elongation and the keratogenous zone of the hair shaft.

The distribution and antigen expression of Merkel cells in mouse skin is as yet ill defined. Since the mouse offers an excellent model for studying the origin and functions of Merkel cells, the Merkel cell distribution as well as the expression of intermediate filament proteins and neuronal markers was characterized in C57 BL/6 mouse skin by immunohistochemistry and electron microscopy. Merkel cells in whisker pads, back, and foot pad skin as identified by staining for neuron-specific enolase-an established neuroendocrine marker--expressed cytokeratins (CK) 8,18, and 20 (i.e., simple-epithelial CKs), but not CKs 4 and 13. Sequential double staining for neuron-specific enolase and CK 20 showed consistent co-expression in Merkel cells, establishing CK 20 as a specific immunocytochemical marker for mouse Merkel cells. The Merkel cells also were immunoreactive for synaptophysin but not for neurofilament proteins, peripherin, S-100 protein, and neural cell adhesion molecule. Using CK 8, 18, and 20 as markers, we detected many Merkel cells in the outer roots sheath of vibrissae hair follicles and in foot pad skin. However, only few Merkel cells were found in back skin. These were restricted to small clusters, localized basally within the Haarscheiben epidermis of tylotrich hair follicles, and formed close contacts to prominent nerve fiber terminals as shown by electron microscopy. In striking contrast to human skin, Merkel cells were never found in the epithelium of pelage hair follicles. Even more strikingly, the density of Haarscheiben-associated Merkel cells changed substantially during the highly synchronized, depilation-induced C 57 BL/6 hair cycle, with a minimum in back skin with all hair follicles in telogen or catagen, and a maximum in back skin with all hair follicles in anagen IV-VI. These observations on the Merkel cell hair cycle-dependent distribution in murine skin point to important differences in Merkel cell functions between humans and mice, and raise intriguing questions as to the role of Merkel cells in hair biology.

We have examined colony-forming ability, localization of colony-forming cells, and in vitro life spans of outer root sheath keratinocytes of different fragments of adult human plucked hair follicles. These were shown by immunohistochemical staining for cytokeratins and integrins to contain a preserved basal cell layer. By microdissection, five fragments of the outer root sheath (B1, B2, B3-1, B3-2, B4) were separated, dispersed by trypsin into single cell suspensions, and grown on human feeder fibroblasts. All fragments gave rise to at least some colonies, but colony-forming ability was mostly marked in the intermediate part (B2) and the lower half of the central part (B3-1); approximately 60% of colony-forming cells of a hair follicle localized to the fragment B3-1 and 28% to the fragment B3-2 (upper half of the central part, including bulge). To compare the in vitro life spans of cells from the various fragments, we subcultured isolated keratinocytes under identical conditions. The longest was found in the fragment B3-2 and the shortest in the fragment B1 (bulb). Moreover, the differentiation state of the native cells and the cells of all cultures were studied during their whole life spans by immunocytochemical analysis of various proliferation and differentiation markers. Surprisingly, keratinocytes of all fragments, as shown by expression of high-molecular-weight cytokeratins and filaggrin, were capable of terminal differentiation. These data indicate that cells with long life spans are localized in central parts of the outer root sheath close to the bulge area and that cells with high colony-forming ability are localized in the lower central parts. The latter are usually removed by plucking and may therefore not represent stem cells but rather cells important for hair growth during a single cycle. Cells with long life spans--also included in plucked hair follicles--may be immediate progeny of stem cells that will be segregated in the bulge area. Finally, our results are important for gene transfer and stem cell gene therapy in genodermatoses, because plucked hair follicles are easily available and keratinocytes close to the bulge area should be used selectively.

Merkel cells are difficult to identify in tissue sections. Previous studies have used cytokeratins (CK) 8, 18, and 19 as histologic markers of Merkel cells. However, these CKs are also expressed in some outer root sheath keratinocytes and some early fetal epidermal cells and thus are not truly specific of Merkel cells in general. Using selective antibodies against a newly described CK, number 20--originally found in intestinal epithelium and Merkel cell carcinomas--in comparison to a key protein of neuroendocrine cells, chromogranin A, we established CK 20 as a specific Merkel cell marker in skin of humans, pigs, and mice. CK 20 seems to be an even more general and sensitive Merkel cell marker as compared to CgA. In double-labeling experiments with stratified-squamous epithelial CK (numbers 5 and 13-17) and simple epithelial CK (numbers 8, 18, and 20) antibodies evaluated by confocal laser scanning microscopy, no cell expressing CKs of both types (i.e., no cell of so-called "transitional" character between Merkel cells and keratinocytes) was identified in human skin. In addition, various neuronal markers present in Merkel cell carcinomas including neurofilaments, peripherin, nerve growth factor receptor, and neuronal cell adhesion molecule appear to be absent in normal Merkel cells. Thus, Merkel cells exhibit a distinct and unique marker profile, with CK 20 being of particularly high value in various species.

Two Merkel cell tumor cultures (MC-MA1, MC-MA2) have been established from metastases of typical Merkel cell tumors. The mestastases in vivo were characterized by co-expression of cytokeratins 8, 18, 19, 20 and neurofilaments, presence of intermediate filament whirls, expression of synaptophysin, neuron-specific enolase, and chromogranin A, rare and weak immunostaining for plakoglobin but absence of cadherins and desmoplakins. Both cultures grow, using supplemented RPMI medium on human irradiated fibroblast feeder layers, as loosely arranged floating small aggregates. Their karyotypes are mostly hyperdiploid. The mean doubling times were about 84 h in the first 8 months and later increased. Ultrastructural and immunoelectron microscopic studies of the Merkel cell tumor cells in vitro (MC-MA1, MC-MA2) revealed sparse membrane-bound neuroendocrine granules and typical IFs that were partly arranged in paranuclear whirls and were labeled by antibodies against cytokeratins and neurofilaments. In immunocytochemical studies using antibodies to cytokeratins 8, 18, 19, and 20 and neurofilament protein NF-L, Merkel cell tumor cells in vitro showed a uniform staining appearing as paranuclear whirls and cytoplasmic fibrils as well. Double-labeling experiments showed a co-localization of both intermediate filament types in most cells. Biochemically we found cytokeratins 8, 18, 19, and 20, and NF-L in tumor cells in vitro. Immunocytochemical staining was negative for desmoplakins, various cadherins, and cell adhesion molecules, whereas plakoglobin was only rarely detectable in some Merkel cell tumor cells in vitro. By immunoluminometric assay chromogranin A was detected in cell homogenates and culture supernatants as well. Immunocytochemically, synaptophysin and neuron-specific enolase were detectable additionally in some of the cells. These established cell cultures will allow further studies devoted to the biology, differentiation, and hormone secretion of Merkel cell tumors that may also increase our knowledge about normal Merkel cells.

Differentiation of human plantar and palmar epidermis is characterized by the suprabasal synthesis of a major special intermediate-sized filament (IF) protein, the type I (acidic) cytokeratin 9 (CK 9). Using partial amino acid (aa) sequence information obtained by direct Edman sequencing of peptides resulting from proteolytic digestion of purified CK 9, we synthesized several redundant primers by 'back-translation'. Amplification by polymerase chain reaction (PCR) of cDNAs obtained by reverse transcription of mRNAs from human foot sole epidermis, including 5'-primer extension, resulted in multiple overlapping cDNA clones, from which the complete cDNA (2353 bp) could be constructed. This cDNA encoded the CK 9 polypeptide with a calculated molecular weight of 61,987 and an isoelectric point at about pH 5.0. The aa sequence deduced from cDNA was verified in several parts by comparison with the peptide sequences and showed the typical structure of type I CKs, with a head (153 aa), and alpha-helical coiled-coil-forming rod (306 aa), and a tail (163 aa) domain. The protein displayed the highest homology to human CK 10, not only in the highly conserved rod domain but also in large parts of the head and the tail domains. On the other hand, the aa sequence revealed some remarkable differences from CK 10 and other CKs, even in the most conserved segments of the rod domain. The nuclease digestion pattern seen on Southern blot analysis of human genomic DNA indicated the existence of a unique CK 9 gene. Using CK 9-specific riboprobes for hybridization on Northern blots of RNAs from various epithelia, a mRNA of about 2.4 kb in length could be identified only in foot sole epidermis, and a weaker cross-hybridization signal was seen in RNA from bovine heel pad epidermis at about 2.0 kb. A large number of tissues and cell cultures were examined by PCR of mRNA-derived cDNAs, using CK 9-specific primers. But even with this very sensitive signal amplification, only palmar/plantar epidermis was found positive. By in situ hybridization and immunolocalization we further showed that CK 9 is only expressed in the suprabasal cell layers of this special epidermal tissue. We discuss the molecular properties of CK 9 and its cell type- and body site-specific expression in relation to the special differentiation of palmar/plantar epidermis and to diseases specific for this body site.

Human haarscheiben, epidermal Merkel cell-rich sensory organs of hairy skin, were studied for the expression of various cytokeratin (CK) polypeptides and other epithelial and neuronal differentiation markers by applying immunoperoxidase and immunofluorescence microscopy to frozen sections and by two-dimensional gel electrophoresis. The basal clusters of Merkel cells were specifically detected by antibodies against CK 20. Haarscheiben keratinocytes were unique mainly by the prominent expression of CK 17 in the lower and middle layers. Further differences as compared to keratinocytes of usual epidermis included the enlargement of the basal compartment, characterized by the expression of CK 5 and the absence of the maturation-associated CKs 1/10/11, and the reduction of CK 15, which is a constituent of normal basal cells. Using CK 17 as a highly sensitive Haarscheibe marker in skin tissue sections, variabilities in the spatial relationship of the haarscheibe and the corresponding hair follicle were recorded. The results show that haarscheibe keratinocytes express a special program of differentiation that may be important for optimal stimulus perception. Immunohistochemical stainings for CK 17 will facilitate further studies on the distribution and biology of haarscheibe.

The intermediate filament expression in fetal and adult human eccrine sweat glands was studied by immunoperoxidase microscopy performed on cryostat sections using monoclonal antibodies against various cytokeratins (CK), vimentin, and actin. In palmar skin of 14-week-old fetuses, the early dermal cords showed a primitive CK pattern similar to that of epidermal basal cells. From week 15 on (distal finger skin), inner cells of the proximal (ductal) portion of the glandular anlagen expressed CK 1/10/11 and 19 (markers of adult eccrine ductal luminal cells). In addition, CK 4 was expressed in ductal luminal cells mainly in the fetal period. In the distal portion of the sweat gland anlagen the increased or new expression of the simple-epithelium-type CK 7, 8, 18, and 19 was detected at week 15, indicating the onset of the secretory differentiation pathway. Two subsegments of the prospective secretory portion could be distinguished (elongated part and end bud). Interestingly, in fetuses, most secretory portion cells co-expressed vimentin in addition to CK. From week 22 on, peripheral cells of the secretory portion were stained for CK 17 and smooth-muscle-type actin, suggesting myoepithelial differentiation. In newborn and adult eccrine glands, secretory cells expressed mainly CK 7, 8, 18, and 19, whereas myoepithelial cells were conspicuous by their co-expression of certain CK (including CK 5 and 17), vimentin, and smooth-muscle-type actin and sometimes even glial filament protein (GFP), similar to myoepithelial cells of other glands. These results throw further light onto the complex processes of fetal development of eccrine sweat glands and their cellular diversification. The possible biologic significance of the differential CK expression in the various glandular cell types is discussed.

An experimental transplantation model using human fetal skin was applied to approach the question of the embryologic origin of human Merkel cells. Palmar and plantar skin from five fetuses, between 8 and 11 weeks of estimated gestational age (EGA), was xenografted to subcutaneous beds of nude mice. After 4 or 8 weeks of growth, biopsies were taken from these xenografts and examined for the presence of Merkel cells, using immunocytochemistry with antibodies specific for simple epithelial-type cytokeratins and neuron-specific enolase (NSE) as well as using electron microscopy. Skin from the same fetuses at the time of transplantation was screened in the same way. In all fetuses, no (or very scarce) epidermal Merkel cells were detected at the transplantation time, but in all cases abundant epidermal Merkel cells of apparent human origin were found after 4 or 8 weeks in xenograft culture. Dermal nerve fibers, as recognized by neurofilament antibodies, were scarce or essentially absent in the xenografts. These results indicate that Merkel cells regularly develop in epidermis dissected and xenografted in an early fetal stage, although the dissection implies the interruption of the dermal nerves. The results strongly support the notion of the origin of Merkel cells from epidermal precursor cells. The apparent absence of dermal Merkel cells and dermal nerve fibers in the xenografts suggests that the presence of dermal sensory nerve fibers may be required for the dropping off of epidermal Merkel cells into the upper dermis, which occurs in normal fetal development.

Biochemical analyses have previously shown that palmar and plantar epidermis, unlike the epidermis of other body sites, contain cytokeratin 9 (Mr 64,000), an unusually large acidic (type I) cytokeratin. Guinea-pig antibodies that specifically and selectively react with bovine and human cytokeratin 9 were used for the immunocytochemical identification of cytokeratin 9 in adult and fetal human epidermis from various body sites. In the epidermis of palms and soles, antibodies against cytokeratin 9 stained a high proportion of the keratinocytes in suprabasal locations. These suprabasal cytokeratin-9-positive keratinocytes were often arranged in vertical columns and concentrated around intraepidermal sweat-gland ducts, but they sometimes also formed extended continuous sheets. In contrast, another type-I component, cytokeratin(s) 10/11, was uniformly distributed among suprabasal keratinocytes. By double-labeling immunofluorescence microscopy using a monoclonal antibody against cytokeratin(s) 10/11, we found that cytokeratin 9 usually appears in cells located one or two layers above the cells in which cytokeratin(s) 10/11 is detected, indicating that most keratinocytes expressing cytokeratin 9 also express cytokeratin(s) 10/11. At other body sites, cytokeratin 9 was only detected in sparsely distributed keratinocytes localized in upper epidermal layers; these cells were scattered or formed small clusters, and often exhibited a conspicuous association with the epidermal portion of eccrine sweat-gland ducts. During human fetal development, cytokeratin 9 was first detected at week 15 of gestation in some suprabasal cells of the foot-sole epidermis and, occasionally, in basal cells. At later fetal stages, most of the cytokeratin-9-positive cells appeared in clusters that were mainly concentrated in glandular ridges and interridges. Our results show that two major types of terminally differentiating keratinocytes can be distinguished in human epidermis, i.e. those that do and those that do not express cytokeratin 9. This special program of keratinocyte differentiation identified by the presence of cytokeratin 9 appears to be related to the morphogenesis of palm and sole epidermis, where this protein is expressed early in fetal life. Possible biological functions of this subset of cytokeratin-9-positive keratinocytes are discussed.

The origin of Merkel cells is still a matter of debate, specifically the question of whether they are derived from epithelial cells of the epidermis or from immigrated neural crest cells. As an argument for the latter hypothesis the occurrence of dermal, nerve-associated Merkel cells in human fetal skin has often been mentioned. Therefore, we analyzed the distribution of Merkel cells in epidermis and dermis of plantar skin of human embryos and fetuses, ranging in gestational age between 7 and 17 weeks. Merkel cells were identified by immunocytochemistry on frozen sections using antibodies against simple epithelium-type cytokeratins and by electron microscopy. In the 17-week-old fetus, 17% of the total cutaneous (epidermal and dermal) Merkel cells were located in the upper dermal compartment, whereas in the 14-week-old fetus only 3.9% of the Merkel cells were dermal, including some cells that seemed to be in the process of traversing the dermal-epidermal junction. Thirteen-week-old fetuses showed even fewer dermal Merkel cells. Twelve-week-old fetuses exhibited 660 epidermal Merkel cells per 100 mm total section length, but none in the upper or deep dermis. In 7- to 9-week embryos, no Merkel cells were recognized. However, at this stage, but not in later stages, the basal cells of the plantar epidermis expressed certain simple epithelium-type cytokeratin polypeptides. These results speak against an invasion of Merkel cells or putative neural crest-derived precursor cells into the epidermis via a dermal passage. They suggest that in plantar skin Merkel cells arise, between weeks 8-12, from precursor stages of epithelial cells of the early fetal epidermis which still express simple epithelium-type cytokeratins. The results further suggest that in subsequent stages of skin development some epidermal Merkel cells detach from the epithelium and migrate into the upper dermis where some of them may associate with small nerves.

The cytokeratin family of intermediate filament (IF) proteins can be grouped into the epithelial polypeptides ("soft alpha-keratins"), of which at least 19 exist in the various human epithelia, and the hair-type cytokeratins ("hard alpha-keratins"), which are typical of trichocytes, i.e., the living hair-forming cells. We have recently shown [34] that the hair follicles from diverse mammalian species contain a set of eight major cytokeratin polypeptides, four each of the acidic (type I) and the basic (type II) subfamily, which are different from all known epithelial cytokeratins. In addition, we have identified two new minor trichocytic cytokeratin polypeptides, designated Hax (type I) and Hbx (type II). Antibodies against trichocytic cytokeratins that do not crossreact with any of the epithelial cytokeratins have enabled us to study the expression of both kinds of cytokeratin in the various cell types of human and bovine hair follicles. Using immunofluorescence microscopy, we have observed intense reactions of trichocytic cytokeratins only in cells contributing to the forming hairs, i.e., hair shaft, medulla and cuticle, whereas immunostaining of the peribulbar matrix cells was weaker, if at all detectable. In contrast, epithelial cytokeratins were localized in both the inner and outer root sheath epithelia but, surprisingly, also in certain portions of the trichocyte column, notably cells of the cuticle, certain medullary cells, and trichocytes of the basalmost peripapillary cell layers. Cells coexpressing trichocytic and epithelial cytokeratins have been identified by double-label immunofluorescence microscopy. Epithelial cytokeratins of the inner and outer root sheath epithelia include, most remarkably, "simple-epithelium-type" cytokeratins 8, 18, and 19; these occur in certain peribulbar regions, in distinct patterns, but with variable frequencies. The occurrence of simple epithelial cytokeratins in hair follicles has also been confirmed by high-sensitivity immunoblotting of follicular polypeptides separated by gel electrophoresis. Vimentin-positive cells were abundantly interspersed (in some follicles, but not in all) between the trichocytes of the peripapillary cone, most of them probably being melanocytes. The cell-type complexity of the hair follicle and the different patterns of cytoskeletal protein expression in the various hair follicle cells are discussed in relation to the development and growth of this organ.

The various structural components of pilomatrixoma (calcifying epithelioma of Malherbe) were studied for the expression of hair-specific (trichocytic) cytokeratins as well as epithelial cytokeratins, using immunoperoxidase and immunofluorescence microscopy of frozen sections as well as two-dimensional gel electrophoresis and immunoblotting. Trichocyte-type cytokeratins were detected in only a minor subpopulation of basophilic cells but more prominently in most "transitional" cells as well as in "shadow" cells. In contrast, antibodies against certain epithelial cytokeratins (including antibody KA1 against cytokeratins of stratified squamous epithelia and antibodies against cytokeratin 19) revealed an extensive but heterogeneous staining of basophilic cells. In regions of squamoid cells, epithelial cytokeratins in an unusual pattern were found. An antibody against vimentin was negative on pilomatrixoma cells. An antibody against desmoplakins decorated true desmosomes in basophilic and transitional cells. Biochemically, trichocytic cytokeratin polypeptides as well as epithelial cytokeratins 5, 6, 14, 16, 17, and 19 were positively detected. These results provide evidence in support of the notion of a trichocytic differentiation and probably derivation of pilomatrixoma. According to the cytokeratin expression patterns the majority of pilomatrixoma cells resemble the normal hair cortex lineage but some enter a pathway of squamous cell differentiation.

The hair-forming cells (trichocytes) and the mature hair contain four major trichocytic cytokeratins from each of the subfamilies, basic (Hb1-4) and acidic (Ha1-4); these are related - but not identical - to the epithelial cytokeratins. Here we show, by biochemical methods and immunofluorescence microscopy using antibodies specific for either epithelial or trichocyte cytokeratins, that the same set of hair-type cytokeratins, including two newly identified minor components, designated Hax (type I) and Hbx (type II), are also expressed in cells forming nails, in the filiform papillae of the dorsal surface of human and bovine tongue, and, most surprisingly, in some cells of the epithelial reticulum of bovine and human thymus. By double-label immunofluorescence microscopy, we also show that the expression of the two subsets of cytokeratins, i.e., the epithelial and the trichocytic ones, is not necessarily mutually exclusive, but that certain cells of hair follicles, nail matrix and bed, lingual papillae, and the nonlymphoid cell system of the thymus contain both trichocytic and certain epithelial cytokeratins. This indicates that these cells coexpress representatives of both kinds of cytokeratin. Implications of these findings with respect to problems of regulatory control of cytokeratin synthesis in tissue development and differentiation, and the possible functional meaning of the occurrence of trichocytic cytokeratins in such histologically diverse tissues, are discussed.

Cells forming hair and nail material are characterized by the synthesis of members of a particular group of alpha-keratin polypeptides (trichocytic cytokeratins. "T cytokeratins") different from epithelial cytokeratins ("E cytokeratins"). As the precursor cells to trichocytes are derived from fetal epidermal keratinocytes expressing only E cytokeratins, we have studied the patterns of expression of both T and E cytokeratins in developing human hair-and nail-forming tissues of different fetal stages, by immunocytochemistry using antibodies specific for certain T or E cytokeratins and by two-dimensional gel electrophoresis and immunoblotting. In developing hair follicles up to the early bulbous-peg stage (weeks 12-15 of gestational age), only certain E but no T cytokeratins were identified. T cytokeratins were first detected in the late bulbous-peg stage (in week-14 scalp skin) in certain cells of the central part of the hair cone. In hair-producing follicles (weeks 18-25), the lower hair matrix cells were positive for certain E cytokeratins, whereas T cytokeratins appeared in the uppermost portion of the matrix and, most prominently, in the maturing trichocytes. From the late bulbous-peg stage on. E cytokeratin antibody Ks13.1 selectively decorated the inner root sheath. In finger nail "anlagen", T cytokeratins were detected first in week 12 and 13 fetuses, specifically in cells of the lunula region. In more-advanced stages of nail formation, expression of T cytokeratins extended not only to the upper layers of the ventral nail matrix but was also found, albeit more sparsely, in cells of the whole nail-bed epithelium. Throughout these developmental stages, coexpression of T and E cytokeratins was noted in certain cells, including E cytokeratin 19. While in earlier stages E cytokeratins 10/11, characteristic of epidermal-type cornification, were noted in different regions, including the superficial stratum of the nail bed epithelium, they were later restricted to the epithelium of the proximal nail fold. The results show that terminal trichocytic differentiation starts, both in ontogeny and during the steady growth of hairs and nails, in cells expressing E cytokeratins and that coexpression of E and T polypeptides occurs in both kinds of appendages. While in the hair follicle, the change to the exclusive synthesis of T cytokeratins appears to take place relatively abruptly and simply, the development of nail structures from the ventral nail matrix seems to be more gradual and is characterized by more-complex patterns of coexpression of both kinds of cytokeratins.

Merkel cells are special neurosecretory cells which, in adult human skin, are usually very scarce. By immunofluorescence microscopy using antibodies to human cytokeratin polypeptide no. 18, we localized distinct non-keratinocyte cells in the glandular ridges of human fetal and adult plantar epidermis. Using electron and immunofluorescence microscopy, these cells were identified as Merkel cells containing typical neurosecretory granules as well as bundles of intermediate-sized filaments and desmosomes. Two-dimensional gel electrophoresis of the cytoskeletal fractions of microdissected epidermal preparations highly enriched in Merkel cells indicated the presence of cytokeratin polypeptides nos. 8, 18 and 19 which are typical of diverse simple epithelia of the human body. Double immunofluorescence microscopy showed that these human Merkel cells contain neither neurofilaments nor vimentin filaments. In human fetuses of 18-24 weeks of age, conspicuously high concentrations of Merkel cells, reaching a density of approximately 1,700 Merkel cells/mm2 skin, were found in the glandular ridges of plantar skin. The concentration decreased considerably at newborn and adult stages. Thin cell processes (up to 20 microns long) were observed in many fetal epidermal Merkel cells. In addition, we detected isolated Merkel cells deeper in the dermis (i.e. at distances of, at most, 100 microns from the epidermis) in fetal and newborn plantar skin. Our results show that Merkel cells are true epithelial cells which, however, differ profoundly from epidermal keratinocytes in their cytokeratin expression. The findings are discussed in relation to the much disputed question of the origin of Merkel cells. The present data speak against the immigration of Merkel cells from the neural crest, but rather suggest that they originate from epithelial cells of the skin, although most probably not from differentiated keratinocytes.

The patterns of expression of cytokeratin polypeptides which are closely correlated to routes of differentiation of epithelial cells were studied in extramammary Paget's disease. Cytokeratins of uninvolved and involved epidermis were analyzed by two-dimensional gel electrophoresis of microdissected tissue preparations as well as by immunofluorescence microscopy using cytokeratin antibodies with different specificities. In uninvolved epidermis, cytokeratins Nos. 1, 5, 6, 10, 11, 14, and 16, characteristic of keratinocytes, were found. Epidermis infiltrated by Paget's cells contained the same components and, in addition, cytokeratins Nos. 7, 8, 18, and 19, the latter being characteristic of simple and glandular epithelia, including apocrine and eccrine skin glands. By immunohistochemistry, broad-spectrum antibodies to cytokeratins decorated both keratinocytes and Paget's cells. Antibodies selective for cytokeratins Nos. 1, 10, and 11 stained suprabasal keratinocytes but not Paget's cells. In contrast, antibodies to cytokeratin No. 18 were negative on keratinocytes but the Paget's cells were selectively stained, as were the secretory cells but not the ductal cells in apocrine and eccrine glands. The results show that the cytoskeleton of Paget's cells is different from that of keratinocytes and ductal cells of skin glands and suggest that these tumor cells express the glandular type cytokeratins Nos. 7, 8, 18, and 19. This provides cell biologic support for a relationship of cells of Paget's disease to secretory cells of apocrine and eccrine glands. The histogenesis of extramammary Paget's cells is discussed in relation to these findings.