Immunoelectron microscopic visualization of the transcytosis of low density lipoproteins in perfused rat arteries.

A postembedding labeling technique was employed to visualize human native low density lipoproteins (LDL) during transcytosis in rat arterial endothelium. For this purpose human LDL was perfused through rat vasculature before fixation and processing for immunoelectron microscopy. The LDL particles were located on sections by anti-human apolipoprotein B-100 (LDL) antibodies and secondary antibodies or protein-A conjugated to 10-nm colloidal gold. LDL molecules were seen in plasmalemmal vesicles as well as in the subendothelial space. No colloidal gold was found in the intercellular junctions. Perfusion with reductively methylated LDL, which cannot bind to the LDL receptor, gave a similar labeling pattern, indicating that transcytosis of LDL via plasmalemmal vesicles is most likely receptor independent. Furthermore, the passage of LDL through intact vascular endothelium is a vesicular transport rather than an intercellular diffusion process.

Functional human epidermal Langerhans cells that lack Birbeck granules.

Birbeck granules (BG) are cytoplasmic organelles that are only found in Langerhans cells (LC). The function of BG is still unclear, although it has been claimed that they are actively involved in receptor-mediated endocytosis and participate in the antigen-processing/presenting function of LC. We have identified a healthy white 29-year-old man whose LC completely lack the presence of BG as determined by electronmicroscopic studies. This was observed repeatedly using skin biopsy specimens taken from several places on the body during a period of 2.5 years. The absence of BG in these LG was documented further by the lack of staining with a BG-specific monoclonal antibody. Despite the complete lack of BG, LC were present in normal numbers, had all the usual morphologic characteristics, and were CD1a and human leukocyte antigen (HLA) class II positive. Two observations indicate that these BG-negative LC display normal antigen-presenting capacity. First, the individual could be sensitized by the hapten diphenylcyclopropenone. This was accompanied by a strong increase in the cell surface expression of HLA class II antigens on his LC, suggesting LC activation. Second, his epidermal cells elicited a normal positive response in an allogeneic mixed epidermal cell lymphocyte reaction. Together these observations strongly suggest that BG are not a prerequisite for normal LC function in vivo and in vitro.

Complement-mediated endothelial cell damage in immune complex vasculitis of the skin: ultrastructural localization of the membrane attack complex.

Activation of the complement system is an important element in our concept of the pathomechanism of immune complex (IC) vasculitis. Both deposition of IC and attraction of polymorphonuclear leukocytes (PMN) are effected by products of complement activation. Actual tissue damage, however, is believed to be caused by PMN penetrating the vessel wall. Our former finding that deposits of membrane attack complex of complement (MAC) are found predominantly in skin lesions of patients with IC vasculitis and not in perilesional skin, has raised the question whether the complement system itself (by way of the MAC) contributes to tissue damage. Our present study shows the ultrastructural localization of MAC in lesional and clinically uninvolved skin in two patients with a cutaneous IC vasculitis. Lesional skin deposits of MAC were found on endothelial cells (EC) of upper dermal vessels and on infiltrating PMN. Uninvolved skin deposits of MAC were found on some EC, but clearly to a lesser extent than on EC of the lesional skin. In the skin of two healthy controls MAC was only found sporadically on EC. Deposits of MAC on EC in the lesional skin were often associated with a typical form of local cell swelling. This local form of endothelial cell swelling was incidentally seen in vessels of clinically uninvolved skin, but not in the skin of the two controls. The association of the endothelial cell swelling with deposits of MAC suggests that the complement system can have a direct damaging effect on EC in IC vasculitis by the assembly of MAC on the endothelial cell membrane.

Macrophages, but not Langerhans cell-like cells of dendritic lineage, express the CD36 molecule in normal human dermis: relevance to downregulatory cutaneous immune responses?

The CD36 molecule has been shown to be associated with subsets of peripheral blood monocyte/macrophages and, in cells isolated from either ultraviolet (UV)-irradiated or diseased skin, to induce downregulatory immune responses. Although macrophages are certainly present within normal human dermis, whether they normally express CD36 is still a matter of debate. In this study, we investigated dermal CD36-expressing macrophages in situ using the gold immunoelectron microscopic technique on tissue ultracryosections. This is a very sensitive and specific method, and its results clearly reflect the in vivo immunophenotypic constitutive situation. Macrophages in normal human dermis were variously shaped from round to dendritic and were localized either immediately beneath the epidermis, in perivascular areas, or in intervascular zones. Macrophages showed consistent gold-positive staining on their cell surface. In contrast, other dermal cells, including fibroblasts, lymphocytes, and mast cells, as well as dermal fibers, were not decorated with gold; dermal Langerhans cell-like dendritic cells (LC/DC), though they did show gold labeling in some intracytoplasmic organelles, did not show any gold particles along their plasma membranes. Therefore, although macrophages in normal human dermis exhibit variability with regard to their localization and shape, they regularly and constitutively expressed CD36. CD36 molecules may be considered a useful marker for macrophages in normal human dermis and may furthermore confer on macrophages, or a subpopulation thereof, intriguing functional properties (e.g., downregulatory capacity versus upregulatory capacity subserved by LC/DC) within the cutaneous immune system.

Freeze-fracture electron microscopy of in vitro reconstructed human epidermis.

Epidermis has been reconstructed in vitro by seeding human keratinocytes on a human dermal substrate in an air-exposed culture. The end product has been examined by freeze-fracture electron microscopy, transmission electron microscopy (TEM) of thin sections, light microscopy, and lipid analysis using thin-layer chromatography. Light microscopic observation of hematoxylin-eosin stained, paraffin embedded cross-sections of the cell culture revealed a strong resemblance to its intact human counterpart, especially with respect to the morphologic organization in basal, spinous, granular, and horny layers. Freeze-fracture electron microscopy and TEM of thin sections generally confirmed the observed resemblances and additionally suggested the presence of lamellar bodies in the stratum granulosum, and of lamellar (lipid) structures between the corneocytes. However, some imperfections were also observed, including some anomalous lipid structures in the intercellular space. Lipid analyses in conjunction with essential fatty acid enrichment studies suggested that the structural anomalies observed in the cultured system may be caused by a lack of linoleyl-ceramides resulting from "immobilization" of linoleyl moieties in the form of triglycerides and phospholipids. In its present form, the air-exposed cell culture already looks very promising as a model for studies of, e.g., skin differentiation disorders such as psoriasis or ichthyosis, studies of the percutaneous penetration and intra(epi)dermal biotransformation of drugs, and skin toxicity screenings. It is furthermore expected that the aforementioned imperfections in the air-exposed cell culture should be avoidable by changing culture conditions such as the relative humidity and the pH, the composition of the medium, or both.

Distribution of low-density lipoprotein receptors and apolipoprotein B on normal and on reconstructed human epidermis.

Immunoelectronmicroscopy was employed to investigate the ultrastructural localization of the low density lipoprotein (LDL) receptor and the protein moiety of LDL (apo B) on normal human epidermis in situ and reconstructed human epidermis in vitro. For this purpose ultrathin sections of Lowicryl K4M embedded material were incubated with monoclonal antibodies against the LDL receptor and apo B followed by a second antibody conjugated to 15 nm colloidal gold. Examination of the sections revealed a similar distribution of the LDL receptor and apo B in normal human epidermis and in reconstructed epidermis. In both systems the amount of LDL receptors decreases during keratinocyte differentiation. In contrast, apo B molecules are more abundantly expressed in the upper layers of the stratum spinosum and the stratum granulosum. The great similarities in the distribution of the LDL receptor and apo B between the in vitro reconstructed epidermis and its in vivo counterpart provide additional proof that reconstructed epidermis is an excellent tool to investigate proliferation and differentiation processes of keratinocytes in vitro.

Differences in low density lipoprotein receptor expression in the suprabasal layer of normal and psoriatic epidermis.

Previous morphological experiments on the distribution of binding sites for low density lipoprotein (LDL) on normal and psoriatic epidermis in situ, done with the LDL-gold technique [Mommaas-Kienhuis AM, et al. J Invest Dermatol 89: 513-517, 1987.] showed an unequivocal correlation between the ability to bind LDL-gold complexes and the state of keratinocyte differentiation. To determine the involvement of the LDL receptor in this phenomenon, we applied immunoelectronmicroscopical methods in conjunction with a monoclonal anti-LDL receptor antibody. Biopsy specimens of normal and psoriasis skin were fixed before being embedded in Lowicryl K4M. Ultrathin sections were incubated first with the anti-LDL receptor antibody, and then with a second antibody conjugated to colloidal gold. On basal cells of both normal and psoriatic epidermis the LDL receptor was distributed evenly between the cell surface and the cytoplasm. No obvious differences in the density of LDL receptors were observed. However, cells from the suprabasal layer showed two striking differences in the localization of the LDL receptor: 1) normal epidermis showed fewer LDL receptor molecules, whereas in psoriasis epidermis the number increased relative to those on basal cells; and 2) in normal suprabasal cells most of the LDL receptors were located inside the cell, but in psoriasis the majority was found on the cell surface. Both phenomena are discussed and we postulate that the higher expression of LDL receptors in psoriasis suprabasal cells and the high expression of the receptor on the cell surface is connected with the hyperproliferative state of the disorder.

Lipid and ultrastructural characterization of reconstructed skin models.

The study aimed at evaluating tissue architecture and quality of the permeability barrier in commercially available reconstructed human skin models; EpiDerm, SkinEthic and Episkin in comparison to native tissue. For this purpose, tissue architecture was examined by electron microscopy and epidermal lipid composition was analyzed by HPTLC. Stratum corneum lipid organization was investigated by electron microscopy in combination with RuO(4) post-fixation and by SAXD. Ultrastructurally, the overall tissue architecture showed high similarities with native epidermis. In the stratum corneum extracellular space, lipid lamellae consisting of multiple alternating electron-dense and electron-lucent bands were present. This regular pattern was not seen throughout the whole stratum corneum probably due to the observed irregular lamellar body extrusion in some areas. Lipid analyses revealed the presence of all major epidermal lipid classes. Compared with native epidermis the content of polar ceramides 5 and 6 was lower, ceramide 7 was absent, and the content of free fatty acids was very low. These differences in lipid composition may account for differences observed in SAXD pattern of Episkin and EpiDerm penetration models. In the latter only the long-distance periodicity unit of about 12 nm was observed and the short periodicity unit was missing. In conclusion, all three skin models provide a promising means for studying the effects of topically applied chemicals, although the observed deviations in tissue homeostasis and barrier properties need to be optimized.

Barrier function in reconstructed epidermis and its resemblance to native human skin.

One of the prerequisites for the use of human skin equivalents for scientific and screening purposes is that their barrier function is similar to that of native skin. Using human epidermis reconstructed on de-epidermized dermis we demonstrated that the formation of the stratum corneum (SC) barrier in vitro proceeds similarly as in vivo as judged from the extensive production of lamellar bodies, their complete extrusion at the stratum granulosum/SC interface, and the formation of multiple broad lamellar structures in the intercorneocyte space. The presence of well-ordered lipid lamellar phases was confirmed by small-angle X-ray diffraction. Although the long periodicity lamellar phase was present in both the native and the reconstructed epidermis, the short periodicity lamellar phase was present only in native tissue. In addition, the SC lipids predominantly formed the hexagonal sublattice. Analysis of lipid composition revealed that all SC lipids are synthesized in vitro. Differences in SC lipid organization in reconstructed epidermis may be ascribed to the differences in fatty acid content and profile indicating that further improvement in culture conditions is required for generation of in vitro reconstructed epidermis with stratum barrier properties of the native tissue.

Keratinocytes resident in normal human skin constitutively express, at low levels, the intercellular adhesion molecule-1. An in situ immunoelectronmicroscopy study.

The expression of intercellular adhesion molecule-1 (ICAM-1) on keratinocytes (KC) was previously demonstrated in biopsies from various inflammatory skin lesions. KC were, however, found virtually ICAM-1 negative, in normal skin, when the same immunocytochemical techniques were employed. By contrast, epithelial cells resident in different organs constitutively express ICAM-1, albeit weakly. The aim of the present study was to use an immunostaining system more sensitive than the conventional immunocytochemistry, namely the in situ immunogold labelling of ultracryosections, to investigate the constitutive ICAM-1 expression by resting KC in normal skin, in vivo. The semiquantitative analysis performed on 500 resident KC, visualized within tissue ultracryosections of normal human skin, revealed that gold granules were present long the cell membrane in a small percentage (14.6%) of resident KC. The density of gold particles (10 nm sized) observed on the cell surface per KC section was as scarce as 13.72 +/- 4.6 (mean +/- standard deviation), although highly significant when compared with controls (P < 0.005). This indicates the presumably low expression of ICAM-1 moieties on the plasma membrane of this KC subset. This ICAM-1 expression could be important in modulating the trafficking to and from normal epidermis of migrating Langerhans cells and occasional leucocytes. The fact that the ICAM-1 expression on KC in normal skin is limited can be considered favourable, because it can account for the prevention of inappropriate KC/leucocyte interactions in the resting cutaneous environment.

The tolerogenic molecule CD95-L is expressed on the plasma membrane of human activated, but not resting, Langerhans' cells.

Although dendritic cells (DC) are well known for their immunogenic capacities, they may even induce peripheral T-cell tolerance, and such a tolerogenic potential can be exerted in mouse through the expression on the DC plasma membrane of the CD95-ligand (CD95-L) molecule, which is able to trigger apoptosis of CD95-expressing antigen-specific T cells. We therefore asked whether epidermal DC, namely Langerhans' cells (LC), either resting (i.e. within the epidermis, 'in situ') or activated (i.e. suspended from the epidermis) or both, could express the CD95-L molecule on the plasma membrane. For such a purpose, two colloidal gold-immunoelectron microscopy (IEM) double-step procedures were carried out: an 'in situ' method, able to investigate resting LC, was performed on ultrathin frozen sections obtained by ultracryomicrotomy (UCMT) of normal skin biopsies; a pre-embedding (P-E) method, able to investigate suspended LC, was performed on epidermal cells (EC) suspended from normal skin specimens. In UCMT/IEM sections, resting LC showed gold particles within the cytoplasm but very rarely within organelles and never along the plasma membrane: resting LC are therefore capable of synthesizing CD95-L but not of expressing it in a functional location, thus autoreactive phenomena against CD95-expressing EC being avoided in normal epidermis. On the other hand, in P-E/IEM preparations, suspended LC showed several gold particles along the plasma membrane: activated LC are therefore capable of expressing CD95-L in a functional location, thus bearing the potential to exert tolerogenic capabilities against CD95-expressing T cells, e.g. to prevent inflammatory/autoimmune cutaneous disorders and/or favor the resolution thereof.