Human Dermal Dendritic Cells Process and Present SolubleProtein Antigens

Frank O. Nestle, Luis Filgueira,* Brian J. Nickoloff,† and Gunter BurgDepartment of Dermatology, University of Zurich Medical School, Zurich, Switzerland; * Institute of Anatomy, Division of Cell Biology, University of Zurich,Zurich, Switzerland; †Department of Pathology, Loyola University Medical Center, Maywood, Illinois, U.S.A.

Recently, a novel type of dendritic antigen-presentingcell has been identified in the dermis of normal humanand mouse skin. These dermal dendritic cells (DDC)occur in higher numbers than epidermal Langerhanscells, represent a distinct differentiation pathway of dend-ritic cells, and are as potent as Langerhans cells in theactivation of superantigen specific T cells. As yet, nothingis known about their capacity to take up, process, andpresent soluble protein antigens. We used the model oftetanus toxoid (TT) driven T cell proliferation to addressthese questions. To test for active internalization of TTprotein, gold labeled TT was incubated with Langerhanscells and DDC and could be traced to multivesicularendo-lysosomal compartments. DDC internalize TTthrough a receptor-mediated, clathrin-independent path-way, whereas Langerhans cells predominantly use macro-pinocytosis. To verify that DDC process TT by theexogenous pathway of antigen presentation, we pulsedDDC with TT protein or TT peptide after preincubation

Dendritic cells (DC) belong to a family of antigen-presenting cells (APC) that are located in the T celldependent area of lymphoid tissues and at the barrierzones where antigen entry into the human bodythrough the skin, lung, and gut may take place. The

best studied DC in the human system is the Langerhans cell, locatedin a suprabasal position in human and murine epidermis (Stingl andBergstresser, 1995). Recently, the dermis has been appreciated as animportant immunoreactive site (Streilein, 1993). The understanding ofcomponents of the dermal immune system may be important not onlyfor most inflammatory and neoplastic skin diseases, but also for theunderstanding of mechanisms involved in anti-tumor vaccination(Gross, 1943). A new member of the DC family has been describedin a perivascular location at the strategic dermal interface between theblood supply and the epidermis (Headington, 1986; Lenz et al, 1993;Meunier et al, 1993; Nestle et al, 1993; Sepulveda-Merrill et al,1994). Recent evidence indicates a role for dermal DC in contacthypersensitivity reactions (Streilein, 1989; Tse and Cooper, 1990)and in diseases such as psoriasis (Nestle et al, 1994), cutaneous

Manuscript received July 17, 1997; revised December 30, 1997; accepted forpublication January 26, 1998.

Reprint requests to: Dr. Frank O. Nestle, Department of Dermatology,University of Zurich Medical School, Gloriastrasse 31, 8091 Zurich, Switzerland.

Abbreviations: DDC, dermal dendritic cells; DC, dendritic cells; TT, tet-anus toxoid.

0022-202X/98/$10.50 · Copyright © 1998 by The Society for Investigative Dermatology, Inc.

762

with chloroquine. Preincubation with chloroquine dimin-ished the capacity of DDC to induce TT protein specificT cell proliferation (70–80%), but was not effective tosuppress TT peptide induced T cell responses. DDC wereas potent as Langerhans cells and 5–10 3 more potentthan plastic adherent monocytes in the presentation ofTT to autologous resting T cells. Furthermore, as few as50 DDC (stimulator:responder ratio of 1:1000) were ableto induce a significant TT specific T cell proliferation.Because a subpopulation of DDC expresses low levels ofCD1a, a phenotypic marker of Langerhans cells, sortingof CD1a positive and negative DDC was performed. Ona per cell basis, CD1a positive and negative DDC wereequally potent at mediating TT specific T cell prolifera-tion. Thus, DDC are able to internalize, process, andpresent soluble protein antigens such as TT and maytherefore play an important role in the regulation ofskin immune responses. Key words: antigen-presenting cells/Langerhans cells/skin immune system/tetanus toxoid. J InvestDermatol 110:762–766, 1998

lymphoproliferative disorders (Nestle and Nickoloff, 1994), and derma-tofibroma (Nestle et al, 1995), as well as Kaposis’ sarcoma (Nickoloffand Griffith, 1989). More recently, two major pathways of DCdifferentiation have been described: one pathway leading to thedevelopment of Langerhans cell-like cells, and a second pathway wheredermal dendritic cells (DDC) seem to be the prototypic cell type(Caux et al, 1996). Therefore, from the comparative study of Langerhanscells and DDC one may learn not only about the skin immune system,but also about the development of DC in general. DDC havebeen extensively studied concerning their phenotype and potency forpresentation of bacterial derived superantigens to T cells (Nestleand Nickoloff, 1995). In contrast, no data are available regardingphagocytosis, processing, and presentation of more widespread solubleprotein antigens. We have investigated these issues using tetanus toxoid(TT) as a model antigen because this system is well defined (Demotzet al, 1989), and circulating TT specific memory T cells are frequentin the European population. In this study we demonstrate that DDCeffectively take up TT through a receptor mediated nonclathrindependent mechanism and transfer it to multilaminar endo-lysosomalcompartments. DDC are as potent as Langerhans cells and more potentthan monocytes in presenting TT protein to autologous CD4 T cells.Furthermore, on a per cell basis, CD1a positive and negative DDCsubsets have similar capacities of antigen presentation.

MATERIALS AND METHODS

Reagents and antibodies RPMI 1640 (Gibco, Basel, Switzerland) wassupplemented with 10% fetal calf serum (Seromed, Basel, Switzerland), 10 U

VOL. 110, NO. 5 MAY 1998 ANTIGEN-PRESENTING FUNCTION OF DERMAL DENDRITIC CELLS 763

penicillin/streptomycin per ml (Gibco), and 50 µg gentamycin per ml (Gibco).TT was obtained from the Swiss Serum Institute (Berne, Switzerland). SyntheticTT derived peptide p30 (TT 947–967; FNNFTVSFWLRVPKVSASHLE) wasobtained from G. Corradin (University of Lausanne, Switzerland). TT-goldcomplexes were prepared as already described (Filgueira et al, 1989). Briefly,colloidal gold particles, 4 nm in size, and TT protein were adjusted at pH 6.1(isoelectric point) and mixed by stirring. TT-gold complex was purified byultracentrifugation (45 min 3 100,000 3 g). Specificity and antigenic propertiesof TT-gold were tested using TT specific FC4-EBV cell clone (kindly providedby A. Lanzavecchia, Basel Institute of Immunology, Switzerland) and TTspecific cell lines (data not shown). TT-gold was found to be equivalent tosoluble TT-protein concerning binding to FC4 cells and activation of TTspecific T cells. Fluoroscein isothiocyanate conjugated isotype controls werepurchased from Becton-Dickinson (Mountain View, CA). CD1a fluorosceinisothiocyanate (OKT6) antibody was purchased from Ortho Diagnostics (Rari-tan, NJ).

Cells Normal human skin was obtained from corrective plastic surgery ofbreast and abdomen. The inclusion criterion was a recent history of tetanusvaccination. Split-thickness skin was prepared using a dermatome (Aesculap,Tuttlingen, Germany) and than incubated in a solution of the bacterial proteasedispase type 2 (Boehringer, Mannheim, Germany) at a final concentration of1.2 U RPMI 1640 per ml at 4°C overnight. After the incubation period, theepidermis and dermis could be easily separated. Epidermal and dermal sheetswere then rinsed several times in phosphate-buffered saline, cut into smallpieces (µ1–10 mm), and placed in RPMI 1640 supplemented with 10% fetalcalf serum, in 10 cm tissue culture plates (Merck, Dietikon, Switzerland). After2–3 d the pieces of tissue were removed and the medium was collected. Thecells that had migrated out of the tissue sections into the medium were spun,resuspended in 1–2 ml fresh medium, and stained with trypan blue. Cellviability and cell number was assessed with a hemocytometer. Further enrichmentof DDC and Langerhans cells was achieved through separation with a metrizam-ide gradient as described (Nestle et al, 1993). Briefly, cells were layered onto3 ml columns of hypertonic 14.5% metrizamide (Nycomed Pharma AS, Oslo,Norway) and sedimented at 650 3 g for 10 min at room temperature. Lowdensity interphase cells were collected and washed in two successive lesshypertonic washes to return cells to isotonicity. DDC expressed high levels ofHLA DR, CD80, CD86, ICAM-1, and CD83 (HB15) in a homogeneousfashion (data not shown).

For some experiments, metrizamide low density fractions were furtherenriched using sterile cell sorting with a FACS Star Plus flow cytometer(Becton-Dickinson, San Jose, CA). The cells were incubated with a fluorosceinisothiocyanate conjugated CD1a monoclonal antibody for 30 min on ice andthen washed three times. CD1a positive and negative DDC were obtained aftersorting with a purity of . 95%.

Highly purified resting CD4 positive T cells were obtained by high gradientmagnetic cell sorting using the MACS technique according to the manufacturer’sprocedure (Miltenyi Biotec, Bergisch Gladbach, Germany). Briefly, humanPBMC were enriched by Ficoll-Hypaque gradients. T lymphocytes weresubsequently obtained by a first round of negative selection of HLA-DRpositive cells, followed by positive selection for CD4 positive T cells. T cellspurified by this procedure were always more than 98% CD21/CD41. NoHLA-DR positive T cells were detected by fluorescence-activated cell sorter(FACS) analysis.

Transmission electron microscopy (TEM) Cells were prefixed with 2.5%glutaraldehyde 1 0.8% paraformaldehyde, and postfixed with an aqueoussolution of 1% OsO4 containing 1.5% K4(FeCN)6. Subsequently, the specimenswere dehydrated in an alcohol serie and embedded into epon. Ultrathin sections(µ50 nm) were contrasted with lead citrate and uranyl acetate and examinedwith a TEM 420 (Phillips, Eindhoven, the Netherlands).

FACS analysis All antibody dilutions and washes were in phosphate-bufferedsaline containing 1% bovine serum albumin (Sigma, Buchs, Switzerland).Negative controls to rule out nonspecific staining or Fc receptor mediatedbinding of antibodies were unrelated monoclonal antibodies of the same isotype.Cells were aliquoted at 1 3 105 per reaction and all staining was performedfor 25 min at 4°C using the indicated primary reagents. Analysis was performedby flow cytometry using a FACScan (Becton Dickinson, San Jose, CA) equippedwith a Lysis 2 software.

Proliferation assays Proliferation assays were performed using standardtechniques. Briefly, for the TT specific T cell proliferation assays, gammairradiated stimulator cells (3000 rad) were cocultured at the indicated concentra-tions with 5 3 104 purified CD4 positive T cells in 96 well round-bottomculture plates (Falcon-Becton-Dickinson, Basel, Switzerland) in 200 µl RPMI1640 1 10% fetal calf serum for 3 d. Cells were pulsed for the final 18 h of

Figure 1. Ultrastructural analysis of TT uptake and processing inepidermal Langerhans cells. (A) Overview of Langerhans cell with typicaldendritic membrane processes. (B) No membrane binding of TT was foundbut occasionally colloidal gold (4 nm) labeled TT (arrowhead) is found in aninvagination of the surface membrane at the beginning of macropinocytosis.(C) TT-gold complex (arrowhead) in an endocytotic vesicle near to a Birbeckgranule (arrow). (D) TT-gold complexes (arrowhead) in a tubular (one arrowhead)and in a multivesicular (two arrowheads) vesicle in the area of the early lysosomalcompartment near to Birbeck granules (arrow).

the incubation period with 1 µCi per well of [3H]TdR (Amersham, LittleChielfont, U.K.). Values are expressed as the mean cpm 6 SD of triplicate wells.

RESULTS

DDC internalize TT proteins Few data are available regardingthe internalization of soluble protein antigens in human Langerhanscells (Schuler et al, 1991) and no data are available for antigen uptakeof DDC. To ensure that TT is taken up by the APC, and to visualizethe endocytic and processing pathway, Langerhans cells and DDCwere studied with TEM. For that purpose, the cells were incubated atculture conditions with gold (4 nm) labeled TT (TT-gold) for differenttime periods before they were processed for standard TEM. Todemonstrate that TT-gold was functional, Langerhans cells or DDCwere pulsed for 2 h at 37°C with TT-gold or TT at a concentrationof 10 µg per ml and incubated with autologous CD4 T cells.Responding T cells proliferated equally well to TT-gold or TT(data not shown). Langerhans cells and DDC demonstrated a similarultramorphology with typical membrane processes or lamellipodia(Fig 1A, 2A). The cytoplasm contained few lysosomes, many mito-chondria, profiles of smooth endoplasmatic reticulum, and manyelectron-lucent vacuoles. Differences concerning uptake of TT-goldwere detected between Langerhans cells and DDC. DDC bound TT-gold at their surface membrane before endocytosis took place (Fig 2B).Also, in the endosomes, TT-gold was found to be bound to thevesicular membrane (Fig 2C). In contrast, no membrane binding ofTT-gold was found in the case of Langerhans cells, either to the surface(Fig 1B) or to the endosomal membrane (Fig 1C). In the endosomeof Langerhans cells, TT-gold was laying without contact to the vesiclemembrane free in the vesicular lumen (Fig 1C). In both cell typesendocytic vesicles were uncoated and were of various different sizes(100–300 nm diameter). Towards the lysosomal compartment, the goldparticle bearing endosomes developed a lamellar or multivesicularsubstructure in Langerhans cells (Fig 1D) and DDC (Fig 2D),which corresponds to the ultramorphology of major histocompatibilitycomplex (MHC) class II loading compartments.

764 NESTLE ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 2. Ultrastructural analysis of TT uptake and processing in DDC.(A) Overview of a DDC with typical dendritic membrane protrusions. (B)Colloidal gold (4 nm) labeled TT (arrowhead) bound to the surface membraneof a DDC. (C) TT-gold complexes (arrowhead) are visualized bound to themembrane of endocytotic vesicles near to the cell surface of a DDC. (D) TT-gold complexes (arrowhead) are visualized bound to membrane structures inlamellar vesicles in the area of the early lysosomal compartment.

DDC present TT to autologous T cells We used TT as a modelantigen because many human beings have a history of vaccination forTT and possess circulating TT specific memory T cells. To test ifDDC present TT to T cells, they were compared with Langerhanscells and plastic adherent monocytes from the same donor, using5 3 104 autologous resting CD4 positive T cells as responders. TheT cell responder population was greater than 98% pure and devoid ofaccessory cells as confirmed by nonresponsiveness to phytohemaggluti-nin. Langerhans cells, DDC, and monocytes were pulsed with TT for2 h at 37°C, washed extensively, and incubated with 5 3 104

responder cells (stimulator:responder ratio of 1:1000, 1:100, and 1:10,respectively). DDC were as effective as Langerhans cells in presentingTT to autologous T cells and 5–10 times more potent than monocytes(Fig 3). Furthermore, as few as 50 DDC were able to induce asignificant TT specific T cell proliferation (. five times backgroundproliferation).

The exogenous pathway of antigen processing is used byDDC To verify if DDC process TT by the exogenous (acidicendosomal) pathway of antigen presentation, we studied the antigen-processing ability in the presence of chloroquine, a selective inhibitorof this pathway (Streicher et al, 1984). Langerhans cells and DDC werepulsed with TT protein (TTprot at 10 µg per ml) or a universallyimmunogenic TT peptide p30 (Panina-Bordignon et al, 1989) (TTpeptat 10 ng per ml) in the presence or absence of 300 µM chloroquine.Five 3 102 APC were then incubated with autologous resting CD4positive T cells at a stimulator:responder ratio of 1:100 and theproliferative response was tested. TTpept induced T cell responseswere used to control for chloroquine induced nonspecific suppressionof T cell proliferation. Chloroquine treatment reduced the stimulatorypotency of TTprot pulsed DDC and Langerhans cells significantly(68% and 64%, respectively) (Fig 4). Using TTpept pulsed APC, norelevant inhibition of T cell stimulation occurred (DDC, 9%;Langerhans cells, 27%) (Fig 4). When chloroquine was added to theAPC after antigen pulsing, there was no significant inhibition of T cellproliferation (data not shown).

Figure 3. Induction of TT specific T cell proliferation by increasingnumbers of stimulator cells. Langerhans cells, DDC, and monocytes (Mono)were pulsed with TT (10 µg per ml) for 2 h and cultured at 50, 500, and 5000cells per well in triplicates with 5 3 104 autologous resting CD4 positive Tcells, representing stimulator:responder ratios of 1:1000, 1:100, and 1:10,respectively. Proliferative responses in these cultures were measured after 3 d.T cells alone or DDC alone or in the presence of TT incorporated less than500 cpm. Note the strong stimulatory capacity of Langerhans cells and DDCrelative to monocytes.

Figure 4. Effects of chloroquine on antigen processing. Langerhans cellsand DDC were used as APC and pulsed with TT (TTprot; 10 µg per ml) ora TT peptide p30 (TTpept at 10 ng per ml) in the presence or absence of300 µM chloroquine (Chloro). Five 3 102 APC were incubated with 5 3 104

resting CD4 positive T cells (stimulator:responder ratio of 1:100) and proliferativeresponses were measured after 3 d. In the presence of chloroquine there was a68% decrease (16744 6 2210 vs 5511 6 466) in the ability of DDC and a 64%decrease (14493 6 1031 vs 5445 6 447) in the ability of Langerhans cells toinduce antigen specific T cell proliferation. T cell proliferation in the absenceof antigen was 2122 6 332 for Langerhans cells and 1880 6 322 for DDC.

Comparison of CD1a positive and negative DDCsubsets Subpopulations of CD1a1 and CD1a– DDC have beendescribed (Meunier et al, 1993; Nestle et al, 1993). To test the possibilitythat these subsets differ in their capacity to present antigen, CD1a1 andCD1a– DDC were differentially sorted on a flow cytometer (purity. 95%). They were pulsed with TT (10 µg per ml, 2 h at 37°C) andincubated at different stimulator:responder ratios with autologousresting CD41 T cells. TT induced T cell proliferation did notsignificantly differ between the CD1a1 and CD1a– DDC and wascomparable with unsorted DDC induced T cell proliferation (Fig 5).This indicates that on a per cell basis, CD1a positive and negativeDDC are equally able to process and present TT protein to autologoushelper T cells.

DISCUSSION

Regional immunity is becoming an increasingly important topicregarding the peripheral activation and de-activation of circulating T

VOL. 110, NO. 5 MAY 1998 ANTIGEN-PRESENTING FUNCTION OF DERMAL DENDRITIC CELLS 765

Figure 5. TT specific T cell proliferation induced by subpopulations ofDDC. Cell populations were differentially sorted using immunofluorescencestaining for CD1a. CD1a negative and positive, as well as unsorted DDC (50,500, or 5000 per well in triplicates), were pulsed with TT (10 µg per ml) for2 h and incubated with 5 3 104 autologous resting CD4 positive T cells,representing stimulator:responder ratios of 1:1000, 1:100, and 1:10, respectively.Proliferative responses in these cultures were measured after 3 d. T cells aloneor DDC alone or in the presence of TT incorporated less than 500 cpm. Notethat CD1a positive and negative, as well as unsorted DDC, have similarcapacities in inducing TT specific T cell proliferation.

cells at various tissue sites (Arnold et al, 1992). In this context, activationof T cells by APC of the dermis that are located at the interfacebetween blood supply and the skin immune system may play a centralrole in vaccination and disease (Streilein, 1993). A new member ofthe DC family of APC has been recently isolated from human andmurine dermis and characterized regarding phenotype and T cellstimulatory capacity (Nestle and Nickoloff, 1995). Interest in thebiology of DDC also stems from the fact that they may be thephysiologic representative of a major DC differentiation pathway (Cauxet al, 1996). No information is currently available concerning uptake,processing, and presentation to T cells of common soluble proteinantigens by these cells. We have investigated these issues using TT asa model antigen because this system is well defined (Demotz et al,1989) and circulating TT specific memory T cells are frequent in theEuropean population.

Before immunogenic peptides are presented on surface MHC classII molecules to T cells, protein antigens have to be phagocytosed andprocessed intracellularily in the endosomal compartment (Germain,1994). The processing pathway can be visualized by use of gold labeledproteins with TEM (Peters et al, 1995). In this study we used goldlabeled TT that had been shown to be functionally active concerningactivation of TT specific T cells by TT pulsed APC. There wasefficient binding of TT-gold to the surface membrane of DDC. Thechemical structure of the molecules binding TT-gold has not beendetermined yet. Despite a receptor mediated uptake of TT-gold thatseems to take place in the case of DDC, the endosomal vesicleswere uncoated, indicating a clathrin-independent pinocytic pathway(Guagliardi et al, 1990; Lamaze and Schmid, 1995). This pathway maysubstitute for clathrin-dependent endocytosis in DDC and is currentlyunder investigation. In contrast, Langerhans cells did not bind TT-gold to the surface membrane, but were still able to take up TT-goldefficiently. In Langerhans cells the endosomal vesicles were alsouncoated and displayed typical characteristics of macropinocytosis(Hewlett et al, 1994). In comparison, Langerhans cells and DDC seemto use different phagocytosis mechanisms for TT uptake as visualizedby TEM, namely macropinocytosis and receptor mediated phagocytosis,respectively. The different uptake mechanisms support the conceptabout two independent DC differentiation pathways (Caux et al, 1996),which is probably also reflected at the level of antigen uptake. Bothuptake mechanisms are known to drive towards a similar Ag processingcompartment (Sallusto et al, 1995). Accordingly, the endo-lysosomalvesicles, where processing and antigen loading takes place, wereidentical in both Langerhans cells and DDC. These vesicles displayedlamellar or multivesicular substructures, characteristic for so-called

MIIC, and have been proposed as the major antigen processing andloading site in DC (Nijman et al, 1995). We did not directlyaddress the question of whether those compartments are identical tomultivesicular or multilaminar MHC class II enriched compartments,although the ultramorphology of the compartments concentrating TTsuggests their presence also in DDC.

To ensure that the immunogenic TT peptides, loaded onto MHCmolecules, are derived from intracellular processing of TT and notfrom processing by exogenous proteases in the culture medium, wetook advantage of the fact that antigen processing requires an acidiclysosomal compartment. TT pulsing of Langerhans cells and DDC inthe presence of chloroquine, which elevates the pH in lysosomalcompartments and blocks processing of TT protein, resulted in signific-ant reduction of TT specific T cell proliferation. In contrast, TTpeptide induced T cell proliferation, which is not dependent onintracellular processing, was not significantly inhibited by chloroquine.

A relationship between the maturational state of DC and their abilityto process complex polypeptides such as TT has been mainly shownfor mouse Langerhans cells (Schuler and Steinman, 1985; Romani et al,1989a) and was later also suggested for human Langerhans cells (Romaniet al, 1989b; Teunissen et al, 1990); however, efficient presentation ofTT by human in vitro generated DC (Sallusto and Lanzavecchia, 1994),as well as human mature cultured Langerhans cells (Cohen and Katz,1992), has been demonstrated. In this study, dependence of thematurational state of DDC on efficiency of antigen processing andpresentation could not be investigated due to a lack of an appropriateisolation method for ‘‘fresh,’’ immature DDC. The maturational statusof the DDC preparation used here has been extensively investigated(Nestle et al, 1993). The high surface expression of CD80, CD86,ICAM-1, CD83 (HB15), and MHC class II molecules, as well as thepotent activation of allogeneic T cells, indicates a mature state of theDDC preparation used in this study. DDC were as efficient asLangerhans cells at presenting TT to autologous, resting CD4 positiveT cells and were 5–10 times more potent than highly enriched plasticadherent monocytes. Furthermore, as few as 50 DDC induced asignificant TT specific T cell proliferation.

DDC express high levels of MHC class II molecules and costimulatorymolecules such as CD80 and C86 in a homogenous manner (Nestleet al, 1993). Strong surface expression of CD1a is considered aphenotypic hallmark of human Langerhans cells. Because a minorsubpopulation (10–15%) of DDC expresses low levels of CD1a (Nestleet al, 1993), we addressed the question of whether the minor CD1a1

subpopulation accounts for the potent APC capacity of DDC. There-fore, DDC were differentially sorted in CD1a1 and CD1a– cells andtested for the induction of TT specific T cell proliferation. On a percell basis, CD1a1 as well as CD1a– subpopulations were equally potentas unsorted DDC in the induction of TT driven T cell proliferation.This is in contrast to a published work investigating fresh dermal cellsuspensions, where, by indirect reasoning, the major antigen-presentingcell function for allogeneic T cell proliferations was found mainly inthe CD1a population (Sepulveda-Merrill et al, 1994). It is thereforeimportant to mention that, for reasons of cell purity and to assess TTspecific T cell proliferation on a per cell basis, our data were obtainedwith highly enriched cultured DDC that could account for thedifferent findings.

In summary, we have provided evidence that DDC are as efficientas Langerhans cells in the internalization, processing, and presentationof soluble protein antigens such as TT to resting T helper cells.Therefore, DDC in their strategic perivascular location in humandermis may play an important role in the activation of skin seeking Tcells in response to soluble protein antigens.

Supported by grants from the Cancer League Zurich and the Swiss Cancer League toFON. We thank the Department of Plastic Surgery (Director Prof Dr. V.E. Meyer),University of Zurich Medical School for providing tissue samples, Mrs. E. Niederer,Flow Cytometry Core Unit, University of Zurich Medical School, for cell sorting,and Mrs. M. Balzer, Mrs. M. Erni, and Mr. H. Sonderegger for their excellenttechnical support.

766 NESTLE ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

REFERENCES

Arnold B, Schonrich G, Hammerling GJ: Extrathymic T-cell selection. Curr Opin Immunol4:166–170, 1992

Caux C, Vanbervliet B, Massacrier C, et al: CD341 hematopoietic progenitors fromhuman cord blood diffeerntiate along two independent dendritic cell pathways inresponse to GM-CSF and TNFa. J Exp Med 184:695–706, 1996

Cohen PJ, Katz SI: Cultured human Langerhans cells process and present intact proteinantigens. J Invest Dermatol 99:331–336, 1992

Demotz S, Matricardi PM, Irle C, Panina P, Lanzavecchia A, Corradin G: Processing oftetanus toxin by human antigen-presenting cells. J Immunol 143:3881–3886, 1989

Filgueira L, Groscurth P, Aguet M: Binding and internalization of gold-labeled IFN-g byhuman RAJI cells. J Immunol 142:3336–3339, 1989

Germain RN: MHC-dependent antigen processing and peptide presentation: providingligands for T lymphocyte activation. Cell 76:287–299, 1994

Gross L: Intradermal immunization of C3H mice against a sarcoma that originated in thesame line. Cancer Res 3:326–333, 1943

Guagliardi LE, Koppelman B, Blum JS, Marks MS, Cresswell P, Brodsky FM: Co-localization of molecules involved in antigen processing and presentation in an earlyendocytic compartment. Nature 343:133–139, 1990

Headington JT: The dermal dendrocyte. In: Callen JP, Dahl HV, Golitz LE, RasmussenJE and Stegman ST (ed.). Advances in Dermatology. Yearbook Medical Publishers,Chicago, 1986, pp. 159–170

Hewlett LJ, Prescott AR, Watts C: The coated pit and macropinocytic pathways servedistinct endosome populations. J Cell Biol 124:689–703, 1994

Lamaze C, Schmid SL: The emergence of clathrin-independent pinocytic pathways. CurrOpin Cell Biol 7:573–580, 1995

Lenz A, Heine M, Schuler G, Romani N: Human and murine dermis contain dendriticcells. J Clin Invest 92:2587–2596, 1993

Meunier L, Gonzales-Ramos A, Cooper KD: Heterogenous populations of class II MHC1cells in human dermal cell suspensions. J Immunol 151:4067–4080, 1993

Nestle FO, Nickoloff BJ: Role of dendritic cells in benign and malignant lymphocyticinfiltrates of the skin. Dermatol Clin 12:271–282, 1994

Nestle FO, Nickoloff BJ: Dermal dendritic cells are important members of the skinimmune system. Adv Exp Med Biol 378:111–116, 1995

Nestle FO, Zheng XG, Thompson CB, Turka LA, Nickoloff BJ: Characterization ofdermal dendritic cells obtained from normal human skin reveals phenotypic andfunctionally distinctive subsets [published erratum appears in J Immunol 152 (1):376,1994]. J Immunol 151:6535–6545, 1993

Nestle FO, Turka LA, Nickoloff BJ: Characterization of dermal dendritic cells in psoriasis.Autostimulation of T lymphocytes and induction of Th1 type cytokines. J ClinInvest 94:202–209, 1994

Nestle FO, Nickoloff BJ, Burg G: Dermatofibroma: an abortive immunoreactive processmediated by dermal dendritic cells? Dermatol 190:265–268, 1995

Nickoloff BJ, Griffith CEM: Factor XIIIa expressing dermal dendrocytes are increased inAIDS-associated Kaposi’s sarcoma. Science 243:1736, 1989

Nijman HW, Kleijmer MJ, Ossevoort MA, et al: Antigen capture and majorhistocompatiblity class II compartments of freshly isolated and cultured human blooddendritic cells. J Immunol 182:163–174, 1995

Panina-Bordignon P, Tan A, Termitjelen A, Demotz S, Corradin G, Lanzavecchia A:Universally immunogenic T cell epitopes: promiscuous binding to human MHCclass II and promiscuous recognition by T cells. Eur J Immunol 19:2237–2242, 1989

Peters PJ, Rapso G, Neefjes JL, Oorschot V, Leijendekker RL, Geuze HJ, Ploegh HL:Major histocompatibility complex class II compartments in human B lymphoblastoidcells are distinct from early endosomes. J Exp Med 182:325–334, 1995

Romani N, Koide S, Crowley M, et al: Presentation of exogenous protein antigens bydendritic cells to T cell cones: intact protein is presented best by immature epidermalLangerhans cells. J Exp Med 169:1169, 1989a

Romani N, Lenz A, Glassel H, et al: Cultured human Langerhans cells resemble lymphoiddendritic cells in phenotype and function. J Invest Dermatol 93:600–609, 1989b

Sallusto F, Lanzavecchia A: Efficient presentation of soluble antigen by cultured humandendritic cells is maintained by granulocyte/macrophage colony-stimulating factorplus interleukin-4 and downregulated by tumor necrosis factor-alpha. J Exp Med179:1109–1118, 1994

Sallusto F, Cella M, Danieli C, Lanzavechia A: Dendritic cells use macropinocytosis andthe mannnose receptor to concentrate macromolecules in the major histocompatibilitycomplex class II compartment: downregulation by cytokines and bacterial products.J Exp Med 182:389–400, 1995

Schuler G, Steinman RM: Murine epidermal langerhans cells mature into potentimmunostimulatory dendritic cells in vitro. J Exp Med 161:526–546, 1985

Schuler G, Romani N, Stossel H, Wolff K: Structural organization and biological propertiesof Langerhans cells. In: Schuler G (ed.). Epidermal Langerhans Cells. CRC Press,Boca Raton, 1991, pp. 87–137

Sepulveda-Merrill C, Mayall S, Hamblin AS, Breathnach SM: Antigen-presenting capacityin normal human dermis is mainly subserved by CD1a 1 cells. Br J Dermatol 131:15–22, 1994

Stingl G, Bergstresser PR: Dendritic cells: a major story unfolds. Immunol Today 16:330–333, 1995

Streicher HZ, Berkower IJ, Busch M, Gurd FRN, Berzofsky JA: Antigen confirmationdetermines processing requirements for T-cell activation. Proc Natl Acad Sci 81:6831–6835, 1984

Streilein JW: Antigen-presenting cells in the induction of contact hypersensitivity in mice:evidence that Langerhans cells are sufficient but not required. J Invest Dermatol93:443–448, 1989

Streilein JW: Dermal cells: Underappreciated components of skin-associated lymphoidtisuue (SALT). In: Nickoloff BJ (ed.). Dermal Immune System. CRC Press, BocaRaton, 1993, pp. 25–38

Teunissen MBM, Wormmeester J, Krieg SR, Peeters PJ, Vogels IMC, Kapsenberg ML,Bos JD: Human and epidermal Langerhans cells undergo profound morphologicand phenotypic changes during in vitro culture. J Invest Dermatol 94:166–173, 1990

Tse H, Cooper KD: Cutaneous dermal Ia1 cells are capable of initiating delayed typehypersensitivity responses. J Invest Dermatol 94:267, 1990