Where does the antigen of cutaneous sarcoidosis come from?

Background: The antigen pathway of cutaneous sarcoidosis remains obscure. We have investigated topographic involvement of inflammatory cells and lymphatic vessels. Methods: Eleven cutaneous biopsies from eight patients were studied, along with controls from other granulomatous disorders and various skin lesions. Markers for lymphocytes, dendritic cells (DCs), and lymphatic vessel endothelial cells were detected using immunohistochemistry. Results: S100⁺ and CD1a⁺ immature DCs (Langerhans cells) occurred more frequently within the epidermis, whereas S100⁺, fascin⁺, or CD83⁺ maturing DCs occurred more frequently beneath the epithelium in cutaneous sarcoidosis cases than in controls (e.g. CD83, cutaneous sarcoidosis vs. other granulomatous disorders: r = 0.557, p = 0.011). Fascin⁺ and CD83⁺ mature DCs were often closely attached to CD3⁺ T‐lymphocytes around dermal granulomas. D2‐40⁺ lymphatic vessels were often found surrounding dermal granulomas, especially those located in the deeper dermis, in contrast to fascin⁺ blood vessels. Conclusions: Antigen‐capturing by immature DCs seems to take place initially in the epidermis, followed by maturation of DCs. These mature DCs may present the processed antigen to T‐lymphocytes that cause dermal granulomas either in the interstitium of the upper dermis, or in or around lymphatic vessels of the lower dermis. Environmental antigen could be verified by skin test. Kurata A, Terado Y, Izumi M, Fujioka Y, and Franke FE. Where does the antigen of cutaneous sarcoidosis come from?     

LICHEN PLANUS: EVALUATION OF CELLS IN SKIN LESIONS AND OF T-LYMPHOCYTE SUBSETS IN BLOOD

The purpose of the present study was to examine the phenotype of cutaneous immunocompetent cells and to quantify Langerhans cells in Lichen planus by the use of monoclonal antibodies directed against T-cell populations. Helper cells (OKT4+) and Suppressor/cytotoxic cells (OKT8+) were observed in all cutaneous infiltrates, and numerous Langerhans cells were identified by OKT6, BL6, and BL2 (HLA-DR) in the epidermis and dermis. The quantification of Langerhans cells demonstrated that the number of dendritic cells in epidermis is greater in involved skin than in non-involved skin. In recent lesions, Langerhans cells are more abundant than in older lesions. The results in peripheral blood indicated a T Helper/Suppressor imbalance with a decreased T-Suppressor/cytotoxic subpopulation in patients with lichen planus diseases. In lichen planus, our results suggest an immunological reaction involving all the immunocompetent cell subpopulations with a first stage of information by Langerhans cells (OKT6+, BL6+, BL2+ (HLA-DR)) and Helper cells (OKT4+), and second stage mediated by Suppressor/cytotoxic cells (OKT8+).     

Immunopathologic Demonstration of T Lymphocyte Subpopulations and Interleukin 2 in Graneloma Annulare

Immunopathologic aspects of granuloma annulare were studied in frozen sections of nine skin biopsy specimens with monoclonal antibodies directed against T lymphocytes, Langerhans' cells, interleukin 2, and interleukin 2 receptors in conjunction with immunoperoxidase techniques. The predominant lymphocyte was an activated T lymphocyte (Leu 1+, HLA-DR+) with an excess of helper/inducer phenotype (Leu 3a+) as compared with suppressor/cytotoxic phenotype (Leu 2a+). Langerhans' cells were increased in the epidermis and numerous OKT6+ cells were observed in the perivascular and granulomatous infiltrate. Both interleukin 2-positive cells and interleukin 2 receptor-positive cells were identified in the dermal lesions according to observed reactivity with the corresponding monoclonal antibodies. These findings suggest that a cell-mediated immune response producing cytokines may be important in the pathogenesis of granuloma annulare. Comparison of these results with skin specimens from patients with sarcoidosis and from a patient with granuloma annulare having some of the histologic features of sarcoidosis, suggests that the cutaneous infiltrate in granuloma annulare represents a response distinct from that of sarcoidosis.     

Expression of OKM5 antigen on epidermal cells in mycosis fungoides plaque stage

To characterize and quantitate potential antigen-presenting cell subsets in the epidermis of patients with cutaneous T-cell lymphoma, epidermal cells in suspension were obtained from involved and uninvolved skin. Involved epidermis contained increased numbers of OKT6+HLA-DR+ Langerhans cells and a variable number of OKM5+ epidermal cells (ECs) in all mycosis fungoides (MF) patients tested (N = 14). The OKM5+ EC population from involved epidermis of MF patients were heterogeneous and comprised both OKM5+HLe1- keratinocytes and OKM5+HLe1+ leukocytes. Uninvolved epidermis, in 6 of 14 patients with MF, contained a small number of OKM5+ leukocytes; however, no OKM5+ keratinocytes were detected. Neither OKM5+ leukocytes nor OKM5+ keratinocytes were detected in the epidermis obtained from healthy controls. The increased number of potential antigen-presenting cells, that is, OKT6+HLA-DR+ Langerhans cells and OKM5+HLA-DR+ monocytic leukocytes, in the epidermis of patients with MF may be important for the activation of abnormal T cells contained within the epidermis of these patients. Such activated T cells may release gamma-interferon and induce expression of both HLA-DR and OKM5 antigens on keratinocytes. OKM5+ keratinocytes are present in the epidermis of patients with MF, but not in normal skin, and may thus play a role in the pathogenetic mechanisms of mycosis fungoides by recruitment of immunocompetent cells to the epidermis.     

Inflammatory cell types in normal human epidermis--an immunohistochemical and morphometric study

Inflammatory cells in 15 specimens of normal human epidermis were selectively stained by a monoclonal antibody immunoperoxidase technique. The quantitative assessment using an interactive image analysis system revealed OKT 11 positive cells (T lymphocytes) and Leu 2a positive cells (suppressor/cytotoxic cells). As there was no significant difference in the distribution of these markers, helper/inducer cells obviously are not present in considerable amounts. OKM 5 positive cells outnumbered OKM 1 positive cells, indicating the presence of a OKM 5+, OKM 1-macrophage subset. The epidermal dendritic cells clearly showed a striking heterogeneity regarding the expression of HLA-DR (62% of OKT 6-positive cells) and Leu 3a (47%), suggesting the existence of immunologically distinct subsets of human epidermal dendritic cells.     

Cutaneous T-cell lymphoma lesional epidermal cells activate autologous CD4+ T lymphocytes: involvement of both CD1+OKM5+ and CD1+OKM5- antigen-presenting cells

Cutaneous T-cell lymphoma is characterized by infiltration of the skin by activated CD4+ T lymphocytes. The mechanism by which these T lymphocytes achieve and maintain their activated state is unknown. Antigen-specific activation of T lymphocytes is dependent upon antigen-presenting cells which express HLA-DR class II major histocompatibility complex molecules, such as epidermal Langerhans cells. In addition to CD1+DR+ Langerhans cells, cutaneous T-cell lymphoma lesional epidermis contains major histocompatibility complex class II positive non-Langerhans cell populations, including CD1+OKM5+ bone-marrow-derived cells and DR+ keratinocytes. We asked whether any of these epidermal cell populations demonstrate capacity to activate T lymphocytes. Various numbers of epidermal cells from uninvolved and involved cutaneous T-cell lymphoma plaques were therefore used to stimulate autologous CD4+ and CD8+ T lymphocytes in the absence of exogenous antigen. Involved epidermal cells potently induced proliferation of CD4+ T lymphocytes (S.I. +/- SEM = 466 +/- 45). In contrast, uninvolved epidermal cells only induced background levels of proliferation (S.I. +/- SEM = 2 +/- 0.5, N = 8, p less than 0.01). Neither involved nor uninvolved epidermal cells were able directly to activate CD8+ lymphocytes. The capability of involved epidermal cells to activate CD4+ T lymphocytes was dependent upon CD1+DR+ leukocytes and not DR+ keratinocytes, because depletion of either HLA-DR+, CD1+ or HLe1+ epidermal cells totally abrogated the T-lymphocyte proliferation. Interestingly, on a cell per cell basis CD1+DR+ cells obtained from involved skin, demonstrated relative to CD1+DR+ cells from uninvolved skin, enhanced capacity to activate CD4+ T lymphocytes. Furthermore, CD1+OKM5+ cells from involved epidermis stimulated autologous CD4+ T lymphocytes. This indicates that a unique hitherto undescribed CD1+OKM5+ epidermal antigen-presenting cell population may participate in T-lymphocyte activation. These findings provide support for the concept that the epidermal cells in cutaneous T-cell lymphoma patients, particularly the antigen-presenting cells, may contribute significantly to the activation of CD4+ malignant and/or non-malignant inflammatory T lymphocytes within the skin.     

Use of the fluorescence-activated cell sorter to quantitate and enrich for subpopulations of human skin cells

A variety of immunologic staining techniques were compared in a quantitative study of antigen expression by human epidermal cells. Virtually all nucleated epidermal cells express beta 2-microglobulin, which is associated with HLA-A, -B, and -C antigens, whereas only about 4% expressed T6, an antigen expressed by Langerhans cells but not other cells in the skin. With the fluorescence-activated cell sorter (FACS), epidermal cell suspensions were selectively enriched 10- to 15-fold for T6-positive Langerhans cells. An average of 6.5% of cells were specifically stained by anti-HLA-DR antibody. When dispersed cells stained with anti-DR plus peroxidase were examined with the technique of immunoelectron microscopy, only mononuclear leukocytes (probably Langerhans cells) were stained. After separating HLA-DR positive skin cells with the FACS, the DR-positive population but not the DR-negative population stimulated proliferation of allogeneic responder lymphocytes, indicating that sorted cells are metabolically active. We conclude that HLA-DR antigen is not expressed by keratinocytes in normal human skin cell suspensions and that the FACS can be used to selectively enrich or deplete skin cell suspensions of antigenically distinct subpopulations such as Langerhans cells.     

Expression of HLA-DR and OKT6 antigens on keratinocytes and dendritic cells in pemphigus

Summary The expression of HLA-DR and OKT6 antigens by epidermal cells in pemphigus was examined by immunohistochemical methods using skin biopsy specimens from human patients with pemphigus vulgaris. Changes in the distribution and subpopulations of Langerhans cells were observed in lesional (involved) skin. HLA-DR-positive, OKT6-negative dendritic cells were numerous and generally confined to the margins of epidermal clefts. OKT6-positive dendritic cells were rare or absent in lesional skin. The HLA-DR-positive, OKT6-negative dendritic cells associated with intraepidermal clefts are most likely functionally active antigen-presenting cells, which may play a role in lesion development. Aberrant expression of HLA-DR molecules by keratinocytes was observed in both involved and uninvolved skin, and may function to facilitate the recognition of surface-bound pemphigus antigen(s) by immunocompetent cells.     

Langerhans cells in skin from patients with psoriasis: quantitative and qualitative study of T6 and HLA-DR antigen-expressing cells and changes with aromatic retinoid administration

Using a monoclonal antibody against human HLA-DR antigens and OKT6, we investigated by indirect immunofluorescence the distribution of Langerhans cells in normal human skin and involved and uninvolved skin from patients with psoriasis before, during, and after systemic aromatic retinoid administration. In parallel, enumeration of HLA-DR and of OKT6+ cells was also performed. In involved psoriatic epidermis the distribution of positive cells was disturbed; OKT6+ cells were reduced in number, as were HLA-DR+ cells which were seen in clusters. In control skin sections, a regular pattern of fluorescent dendritic epidermal cells was noted. In normal-appearing human skin, in nonlesional psoriatic skin, but not in diseased psoriatic skin, the number of OKT6+ cells per epidermal section surface unit was higher than that of HLA-DR expressing cells. Changes in the number and distribution of OKT6 and HLA-DR+ cells in psoriatic involved epidermis were corrected by oral retinoid treatment.     

The ELAM-1 ligand sialosyl-Lex is present on Langerhans cells isolated from stratified epithelium

In this study we show the expression of the newly identified carbohydrate ligand, sialosyl-Le(X) on Langerhans cells. The receptor for sialosyl-Le(X) is the endothelial leukocyte adhesion molecule-1 (ELAM-1) present on activated endothelial cells. Using flow cytometry, Langerhans cells were selected due to positivity for an antibody against CD1a and low orthogonal light scatter. The CD1a antigen stained by the OKT6 antibody is considered a maturational marker of Langerhans cells in agreement with the specific labeling of dendritic cells in the epithelium only. Double immunostaining (OKT6/anti-sialosyl-Le(X)) using flow cytometry and immunohistochemistry demonstrated that almost all OKT6-positive cells in normal stratified epithelium expressed sialosyl-Le(X). Conversely, by immunohistochemistry of oral epithelium with acute inflammation, additional dendritic cells negative for OKT6 were found to express sialosyl-Le(X). In addition, sialosyl-Le(X)-positive but not OKT6-positive dendritic cells were found in the submucosa. These findings indicate that the carbohydrate antigen sialosyl-Le(X) is expressed earlier than the CD1a antigen in the maturation of the Langerhans cell lineage. Future studies should aim at investigating the importance of adhesion between sialosyl Le(X) and ELAM-1 in epithelial recruitment of Langerhans cells.     

T-cell subsets in cutaneous sarcoidosis

Skin lesions from four patients with systemic and cutaneous sarcoidosis were studied, by the use of monoclonal antibodies, for the presence of T cells and T-cell subsets. Large numbers of lymphoid cells reacting with anti-pan T-cell (LEU-1) and anti-helper and inducer subset (LEU-3) monoclonal antibodies were observed around and within the sarcoid granulomas in three of the four patients. Only rare LEU-2-reactive suppressor cells were observed in all four patients. Activated T lymphocytes with focal acid phosphatase activity, together with epithelioid cells and multinucleated giant cells with strong diffuse activity of acid phosphatase and nonspecific esterase, were identified within the granulomas. The two patients with active disease demonstrated substantially more T cells in the sarcoid granulomas than did the two patients with chronic disease. Our study results suggest the importance of helper T cells in the formation of the sarcoid granuloma by mononuclear phagocytes and imply that the activity and duration of disease may be related to the T-cell populations.     

Detection of HTLV-I proviral DNA in sarcoidosis

'Sarcoidosis-lymphoma syndrome' is known as an association of sarcoidosis with malignant lymphoma. We report a 56-year-old woman with systemic sarcoidosis who was seropositive for antibody against human T cell lymphoma/leukemia virus type I (HTLV-I). This patient showed integration of HTLV-I proviral DNA within cutaneous sarcoid nodules, but not in peripheral blood mononuclear cells. Neither atypical lymphocytes nor a T cell receptor beta1 gene rearrangement were observed in peripheral blood mononuclear cells or in cutaneous nodules, indicating that the patient did not have a smouldering type of adult T cell lymphoma/leukemia. Detection of integration of HTLV-I proviral DNA in cutaneous sarcoid nodules could suggest that the sarcoid nodules might have been generated as a protective response to chronic stimuli of HTLV-I.     

The semi-quantitative distribution of T4 and T6 surface antigens on human Langerhans cells

We have used indirect immunogold electron microscopy to compare the respective density of cell membrane determinants revealed by OKT6 and OKT4 monoclonal antibodies on normal human Langerhans cells (LC): 12.9 +/- 3.5 gold granules were noted per cell section on OKT4-positive LC whereas 236.8 +/- 23.5 granules were counted per cell section on OKT6-reactive cells. These results confirm that human LC react with OKT4 antibody and they demonstrate a marked quantitative difference on LC surface between the antigenic determinants recognized by OKT6 and OKT4 antibodies.     

Lymphocytes and Langerhans cells in patch tests

The distribution of immunocompetent cells was analysed in allergic (nickel) and irritant (dithranol) patch tests using conventional transmission electron microscopy and labelling with monoclonal antibodies in an avidin-biotin immunoperoxidase study. The biopsies were taken 24 or 48 h after the allergen/irritant application. In allergic and irritant reactions, most inflammatory cells were OKT11 positive (pan T lymphocytes). The majority of these cells were also OKT4 positive (helper/inducer T lymphocytes), while the minority were OKT8 positive (suppressor/cytotoxic T lymphocytes). NK9 positive cells (natural killer cells) were observed in small numbers. The number of dendritic OKT6 and OKIal positive cells (Langerhans cells) in the epidermis was unaffected in allergic reactions. In irritant reactions, a normal number of OKT6 positive Langerhans cells was observed, while the number of OKIal positive cells had increased in the epidermis. Dithranol caused prominent fine structural changes in the mitochondria of the Langerhans cells, while the keratinocytes appeared largely unaffected. The present study indicates that allergic and irritant patch tests cannot be differentiated reliably using current immunohistopathological or electron microscopic techniques, in spite of the small differences observed.     

Intermediate or chronic cutaneous leishmaniasis: leukocyte immunophenotypes and cytokine characterisation of the lesion

The American cutaneous forms of leishmaniasis include immune-responder individuals with localised cutaneous leishmaniasis (LCL) and non-responder individuals with diffuse cutaneous leishmaniasis (DCL). Patients with intermediate or chronic cutaneous leishmaniasis (ICL) have increased morbidity due to the length of their illness, atypical forms and areas of compromise. In the present study, we evaluated the expression of the leukocyte antigens (CD4, CD8, CLA: cutaneous lymphocyte antigen, CD69, CD83 and CD1a) and cytokines (IFN-gamma, IL-4, IL-10 and TGF-beta 1) in the lesions of patients with ICL (n = 18) using an immunocytochemical procedure. ICL results were compared with the information for LCL (n = 19) and DCL (n = 4). The numbers of CD4+ and CD8+ T cells in ICL were similar to those of LCL lesions, but significantly different (P < or = 0.05) from DCL lesions. LCL lesions have about half the numbers of early activated CD69+ cells as ICL, but most are CLA+ skin homing memory T cells, whereas ICL lesions have the highest number of CD69+ T cells, but about one-third of these cells expressed CLA. This suggests that the granuloma of ICL patients contains many activated T cells that are unprimed to cutaneous-launched antigens, thus contributing to an aberrant immune response. In contrast, DCL granulomas presented the lowest numbers of activated CD69+ and CLA+ cells, associated with the characteristic tolerogenic state of these patients. The immunolocalisation of cytokines showed a mixed cytokine pattern in ICL lesions with many positive cells for IL-10, TGF-beta 1, IL-4 and IFN-gamma, with a preponderance of the first two, and different from the prevalent Th1 and Th2 responses associated with LCL and DCL lesions, respectively. CD1a+ Langerhans cells were decreased (P < or = 0.05) in both ICL (271 +/- 15 cells/mm2) and DCL (245 +/- 19 cells/mm2) as compared to LCL (527 +/- 54 cells/mm2) epidermis. The percentage of IL-10+ epidermal Langerhans cells in ICL (33.69), from the total CD1a+ population, was higher than in LCL (17.45). In addition, fewer CD83+ primed Langerhans cells were present in ICL epidermis. The diminished participation of epidermal Langerhans cells, causing a defective signalling by the epidermis, in ICL lesions may account for the tissue-damaging state observed in these patients.     

Quantitation of cutaneous Langerhans cells of sarcoidosis patients

Langerhans cells play a role in cell-mediated immune reactions which are often depressed in sarcoidosis. We examined the epidermis of 17 anergic patients with sarcoidosis (Kveim-reactive and/or biopsy-proved) for the number of Langerhans cells in noninvolved skin and in any cutaneous sarcoidal lesions. Skin biopsies of 10 healthy volunteers served as controls. In comparison to controls, the epidermis overlying noninvolved (p less than 0.05), sarcoidal (p less than 0.0005), and Kveim-reactive (p less than 0.005) skin contained significantly fewer detectable Ia and T6 antigen-bearing Langerhans cells. The reductions within noninvolved skin were most pronounced in patients with multisystem disease. Lower epidermal Langerhans cell densities, in comparison to controls, were detected in both prednisone-treated and untreated patients. Epidermis overlying sarcoidal skin of untreated patients contained significantly fewer Ia and T6 antigen-bearing Langerhans cells (p less than 0.05, p less than 0.0025, respectively) than epidermis from noninvolved skin. Whether reduced numbers of cutaneous Langerhans cells are due to either a local and/or systemic effect of sarcoidosis, or reflect the anergic state of these patients is unknown.     

T6 is superior to Ia (HLA-DR) as a marker for Langerhans cells and indeterminate cells in normal epidermis: a monoclonal antibody study

Previous studies in our laboratory using immunoelectron microscopy have shown that anti-T6 monoclonal antibody reacts with all epidermal Langerhans cells in normal skin. Comparison of the number of T6-positive (+) epidermal cells with Ia (HLA-DR) (+) cells, as defined by the monoclonal antibodies, anti-I1 and anti-I2, disclosed that these latter markers significantly underestimated Langerhans cell and indeterminate cell numbers (p less than 0.01 and p less than 0.001, respectively) when employed in a sensitive 4-step immunoperoxidase procedure. Thus, it appears that all epidermal Langerhans cells and indeterminate cells are not Ia-positive as defined in this system and that Ia(+)/T6(+) and Ia(-)/T6(+) subsets exist. These subsets may be analogous to the Ia(+) and IA(-) subsets of macrophages, in which the former are responsible for antigen interaction with T cells.     

In situ characterization of cellular infiltrates in lupus vulgaris indicates lesional T-cell activation.

Skin biopsy specimens from nine patients with lupus vulgaris were examined in situ by means of monoclonal antibodies directed against phenotypes of lymphocyte subsets, Langerhans cells, HLA-DR antigens, and interleukin 2 receptor. The epidermis showed prominent changes, including intense expression of HLA-DR on keratinocytes, increase in epidermal cell layers, moderate to high Langerhans cell hyperplasia, and infiltration by CD3+ pan-T cells as well as CD8+ (cytotoxic/suppressor) and CD4+ (helper/inducer) T cells. The predominant lymphocyte in the dermal granulomas was the activated CD3+ T cell, expressing major histocompatibility complex class II antigens and interleukin 2 receptor. CD4+ and CD8+ cells were randomly distributed among the epithelioid cells, which showed intense staining for major histocompatibility complex class II antigens. In all except two patients, the CD4+ population was greater than that of the CD8+ cells. CD1+ Langerhans cells were scattered in moderate numbers in the dermal granulomas. Acid-fast bacilli were conspicuously absent in the biopsy specimens. These features suggest that T-cell activation and Langerhans cell hyperplasia are prominent features of dermal tuberculosis.     

Downregulation of class II molecules on epidermal Langerhans cells in Lyme borreliosis

BACKGROUND: Borrelia burgdorferi can be isolated from the skin of patients with acrodermatitis chronica atrophicans (ACA), a late-stage manifestation of Lyme borreliosis; despite a marked T-cell infiltrate in lesional skin and high antibody titres in patients' sera. OBJECTIVES: To determine whether antigen-presenting Langerhans cells (LCs), which reportedly show signs of injury in erythema chronicum migrans (ECM), the early stage of disease, are altered in ACA. PATIENTS/METHODS: We studied the immunophenotype of cutaneous leucocytes on cryostat sections of lesional skin from both ECM and ACA patients. RESULTS: The total number of CD1a+ cells evaluated by semiautomatic image analysis was lower in ECM (594 +/- 263 cells mm(-2) epidermis) than in ACA (835 +/- 317 cells mm(-2) epidermis). HLA-DR expression was remarkably downregulated on CD1a+ LCs to 29% in ECM and 18% in ACA, whereas in normal skin, most of the epidermal CD1a+ dendritic cells were HLA-DR+. The inflammatory infiltrate was mainly composed of CD68+ macrophages and CD45RO+ memory T cells, with a predominance of CD4+ helper T cells. CONCLUSIONS: It is conceivable that the downregulation of major histocompatibility complex class II molecules on LC in both the early and late skin manifestations of Lyme borreliosis is indicative of a poorly effective anti-B. burgdorferi immune response and thus at least partly responsible for the insufficient elimination of this micro-organism from ACA skin.     

HLA-DR antigen expression on the keratinocyte surface in dermatoses characterized by lymphocytic exocytosis (e.g. pityriasis rosea)

We have investigated the immunoperoxidase staining pattern of the epidermis in several dermatoses characterized by exocytosis of mononuclear cells into the epidermis. We found that HLA-DR antigens showed an intercellular distribution in localized areas of the epidermis in nine of ten cases of pityriasis rosea, and in all four cases of spontaneously regressing flat warts, two cases of pityriasis lichenoides chronica, two of Schamberg's disease, and one case of lichen striatus. Lichen planus and mycosis fungoides cases were used as positive controls. OKT6 antigen was recognized only on the dendritic cells of the epidermis in all these cases. Judging from the distribution of Langerhans cells, the epidermal intercellular HLA-DR antigen seems to be expressed on the keratinocytes in such diseases, and this feature was confirmed by immunoelectron microscopy. These findings support the hypothesis that the expression of HLA-DR antigen on keratinocytes in these dermatoses is linked to cellular immune reactions involving the epidermis.