Dermal Fibroblasts Induce Maturation of Dendritic Cells

Dermal Fibroblasts Induce Maturation of Dendritic Cells Anja Saalbach, Claudia Klein, Jonathan Sleeman, Ulrich Sack, Friederike Kauer, Carl Gebhardt, Marco Averbeck, Ulf This information is current as Anderegg and Jan C. Simon of October 2, 2021. J Immunol 2007; 178:4966-4974; ; doi: 10.4049/jimmunol.178.8.4966 http://www.jimmunol.org/content/178/8/4966 Downloaded from

References This article cites 40 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/178/8/4966.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Dermal Fibroblasts Induce Maturation of Dendritic Cells1

Anja Saalbach,2* Claudia Klein,* Jonathan Sleeman,‡ Ulrich Sack,† Friederike Kauer,* Carl Gebhardt,* Marco Averbeck,* Ulf Anderegg,* and Jan C. Simon*

To trigger an effective T cell-mediated immune response in the skin, cutaneous dendritic cells (DC) migrate into locally draining lymph nodes, where they present Ag to naive T cells. Little is known about the interaction of DC with the various cellular microenvironments they encounter during their migration from the skin to lymphoid tissues. In this study, we show that human ␤ DC generated from peripheral blood monocytes specifically interact with human dermal fibroblasts via the interaction of 2 integrins on DC with Thy-1 (CD90) and ICAM-1 on fibroblasts. This induced the phenotypic maturation of DC reflected by expression of CD83, CD86, CD80, and HLA-DR in a TNF-␣- and ICAM-1-dependent manner. Moreover, fibroblast-matured DC potently induced T cell activation reflected by CD25 expression and enhanced T cell proliferation. Together these data demonstrate that dermal fibroblasts that DC can encounter during their trafficking from skin to lymph node can act as potent regulators

of DC differentiation and function, and thus may actively participate in the regulation and outcome of DC-driven cutaneous Downloaded from immune responses. The Journal of Immunology, 2007, 178: 4966–4974.

endritic cells (DC)3 are specialized APCs that are pivotal types. There is accumulating evidence that the stromal microen- in initiating immune responses. As sentinel cells of pe- vironment plays an important role in the regulation of DC func- D ripheral tissues, DC reside in and traffic through non- tions and in the priming of DC. Dudda et al. (5, 6) demonstrated lymphoid peripheral tissues, continuously surveying the environ- that intracutaneously injected bone marrow-derived DC can induce http://www.jimmunol.org/ ment for invading microorganisms. Under steady-state conditions skin-homing CD8ϩ T cells that express the cutaneous vascular and in the absence of microbial stimulation or inflammation, ap- entry code (cutaneous leukocyte Agϩ). T cells induced after i.p. parently immature DC capture Ags from apoptotic cells, transport injection of the same bone marrow-derived DC, however, them to secondary lymphoid organs, and are involved in the main- ␣ ␤ expressed the intestinal entry code (integrin 4 7). These experi- tenance of tolerance (1, 2). In contrast, the contact of DC with ments prove that DC possess enough plasticity to acquire tissue- danger signals such as microbial agents or inflammatory mediators specific signals from their local microenvironment and subse- induces the maturation of DC, resulting in an increase in their quently to instruct T cells for tissue-selective homing. Stromal expression of MHC molecules, costimulatory molecules, and cells such as fibroblasts, endothelial cells, and macrophages dis- cytokines, and an enhancement of their migratory capacity (3). play different functional properties, including migratory capacity, by guest on October 2, 2021 DC subsequently migrate from the periphery to the T cell zone contractility, and extracellular matrix production and degradation of secondary lymphoid organs (4). Thus, activated DC are spe- that are dependent on the anatomical site or disease status (7). cialized cells that obtain information from the periphery and Thus, it seems plausible that DC may receive instructions from transfer it to lymphoid organs, where they display Ags to naive stromal cells and/or their secreted products. Furthermore, recent Ag-specific T cells and thereby initiate primary T cell immune evidence suggests that the stromal microenvironment directly responses. affects the behavior and differentiation of DC. For example, When traveling from peripheral to lymphoid tissues, DC interact endothelial cells are able to influence the differentiation and with the stromal microenvironment that is composed of the extra- activation of DC (8, 9). cellular matrix, soluble mediators, and a variety of different cell Fibroblasts are an extremely heterogeneous multifunctional cell population, and their role in wound healing, developmental processes, and tumor development is well established (10). *Department of Dermatology, Venerology, and Allergology, Medical Faculty of Leipzig University, Leipzig, Germany; †Institute of Clinical Immunology and Trans- Nevertheless, immunologists have regarded fibroblast activa- fusion Medicine, Medical Faculty of Leipzig University, Leipzig, Germany; and ‡Re- tion as relatively unimportant in regulating immune responses search Centre Karlsruhe, Institute of Toxicology and Genetics, Karlsruhe, Germany and have focused on cellular interactions between lymphocytes, Received for publication October 4, 2006. Accepted for publication January 22, 2007. macrophages, and dendritic cells, all of which generate Ag- The costs of publication of this article were defrayed in part by the payment of page specific responses (11). However, fibroblasts are capable of pro- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ducing various paracrine immune modulators, such as peptide growth factors, cytokines, chemokines, and inflammatory me- 1 This work was supported by the German Research Foundation (Deutsche For- schungsgemeinschaft SA863/1-3, Te284/2-2) and the Interdisciplinary Center of Clin- diators (12). Additionally, they express cell surface adhesion ical Research Leipzig at the Faculty of Medicine of the Universita¨t Leipzig (Project molecules and costimulatory molecules such as CD40 (13). Fur- Z14(INT05)). thermore, there is abundant evidence that fibroblasts taken from 2 Address correspondence and reprint requests to Dr. Anja Saalbach, Department of Dermatology, Venerology, and Allergology, Medical Faculty of Leipzig University, diseased tissues display a fundamentally different phenotype Johannisallee 30, 04103 Leipzig, Germany. E-mail address: Anja.Saalbach@medizin. compared with fibroblasts taken from normal tissues of the uni-leipzig.de same anatomical site (14, 15). Fibroblasts might therefore not 3 Abbreviations used in this paper: DC, dendritic cell; CFDA, carboxyfluorescein only be purely structural elements, but potentially could also diacetate succinimidyl ester; CHO, Chinese hamster ovary. actively participate in the regulation of inflammation and im- Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 mune responses. www.jimmunol.org The Journal of Immunology 4967

In this study, we found human DC in close proximity to dermal incubated with 1 ␮g of anti-TNF-␣ Ab (R&D Systems) or an isotype fibroblasts in vivo. We also report that monocyte-derived DC have control Ab for 60 min at 4°C. After extensive washing with PBS/10% the capacity to adhere specifically to human dermal fibroblasts in Gelafusal, cells were incubated with a Cy2-labeled goat anti-mouse Ab (Dianova) for 60 min at 4°C. For the identification of the cell source of vitro. This interaction of immature DC with fibroblasts induced the TNF-␣, DC were identified by an anti-CD11c PE Ab. maturation of DC in a manner that was dependent on cell-cell contact as well as on soluble mediators. Fibroblast-induced DC Analysis of phagocytic ability maturation was accompanied by up-regulation of HLA, costimu- Immature DC, LPS-stimulated DC, or fibroblast-stimulated DC (5 ϫ 105) latory molecules and cytokines, and the acquisition of an allo- were incubated for 45 min at 37°C with 1 ϫ 107 FITC-conjugated Esch- stimulatory capacity to activate T cells. In summary, we demon- erichia coli (Invitrogen Life Technologies) in RPMI 1640 medium con- strate for the first time that fibroblasts actively participate in the taining 10% FCS. They were then washed twice with ice-cold PBS, resuspended in ice-cold PBS, and immediately analyzed using flow cytometry. regulation of DC function. This results in the maturation of DC and Cells incubated at 4°C were used as negative controls. To quench the subsequent effective T cell activation, demonstrating that fibro- fluorescence signal from attached, but not internalized E. coli, cells were blasts can be involved in the control of DC-driven immune incubated with 0.4% trypan blue in 10 mM sodium citrate and 0.9% responses. sodium chloride. Immunohistochemistry Materials and Methods Antibodies Thy-1 and ICAM-1 protein expression was detected in frozen sections of skin biopsies from patients with atopic dermatitis (n ϭ 5) with an anti- CD1a PE, CD83 FITC, CD80 FITC, HLA-DR PE, CD86 FITC, and CD154 Thy-1 Ab (20) or anti-ICAM-1 Abs, respectively. Bound Abs were FITC Abs were purchased from BD Biosciences. CD11c PE and CD3 PE Abs detected using the avidin-biotin complex technique, according to the Downloaded from were purchased from Caltag Laboratories. CD40 FITC Ab was provided by manufacturer’s protocol (supersensitive multilink alkaline phosphatase Dr. S. Hauschildt (Institute of Biology, University Leipzig, Leipzig, ready-to-use detection system; Biogenix) and visualized colorimetri- Germany). Anti-ICAM-1 Ab (clone R6.5D6) was purchased from cally using the New Fuchsin substrate system (DakoCytomation). DC American Type Culture Collection. were subsequently stained using a biotin-labeled anti-CD1a Ab in the same tissue sections, using streptavidin-peroxidase (Vectorstain) and Cell culture diaminobenzidine for detection. Tissue sections were extensively washed with TBS/0.3% Tween 20 after every Ab incubation step. In Primary fibroblast cultures were obtained by outgrowth from skin biopsies.

control sections, primary Abs were replaced with an isotype control Ab. http://www.jimmunol.org/ The fibroblasts were cultured at 37°C, 5% CO2 in DMEM medium (Bio- Tissue sections were photographed using a DP70 charge-coupled device chrom) containing 10% FCS (Biochrom) and 1% penicillin/streptomycin camera mounted on a light microscope (Olympus). (Biochrom). After reaching confluence, cells were passaged using 0.05% trypsin and 0.02% EDTA (Biochrom). Thy-1-, ICAM-1-, and vector-trans- Adhesion assay fected Chinese hamster ovary (CHO) cells were cultured similarly. Thy-1- and vector-transfected CHO cells were generated, as described before (16– Thy-1-, ICAM-1-, control vector-transfected CHO cells or primary fibro- 18), and cultured in the presence of 7.5 ␮g/ml blasticidin (Invitrogen Life blasts in the third to fifth passage were seeded onto 8-well chamber slides Technologies) to select for Thy-1-expressing cells. ICAM-1-transfected (Nunc). Confluent cell monolayers were subsequently fixed with ice-cold CHO cells were provided by Dr. A. E. May (German Heart Center, methanol and stored at Ϫ20°C. DC were biotinylated with D-biotinoyl-␧- Munich, Germany). aminocaproic-acid-hydroxysuccinimide-ester (Roche Biochemicals), ac- ϫ 5 cording to the manufacturer’s protocol. DC (2 10 ) in RPMI 1640 with- by guest on October 2, 2021 Preparation of DC out serum were added to Thy-1-, ICAM-1-, control vector-transfected CHO cells or to primary fibroblasts, and then incubated for 30 min at 37°C. PBMC were isolated using an endotoxin-free Ficoll-Paque gradient (Am- Unbound DC were removed by several washes with RPMI 1640. Adherent ersham Biosciences). CD14-positive monocytes were purified using anti- cells were fixed with ice-cold methanol for 10 min at Ϫ20°C, air dried, and CD14 magnetic beads (Miltenyi Biotec), according to manufacturer’s pro- blocked by incubation with PBS/1% BSA. Adherent biotinylated cells were ϫ 6 tocol. DC were generated by culture of 1 10 /ml CD14-positive cells for subsequently visualized by incubation with CY3-conjugated streptavidin 5 days in DC-RPMI 1640 (RPMI 1640 (Invitrogen Life Technologies) (Beckman Coulter; 1:500 in PBS/1% BSA) for 60 min at room tempera- supplemented with 2% FCS (Promo Cell), 1% penicillin/streptomycin (In- ture, and finally washed twice with PBS. The number of DC that adhered vitrogen Life Technologies), 1% L-glutamine (Invitrogen Life Technolo- to each confluent monolayer was assessed microscopically in 10 indepen- gies), 0.1 mM nonessential amino acids (Invitrogen Life Technologies), 10 dent fields (magnification ϫ20) by two independent observers. mM HEPES (Invitrogen Life Technologies), 100 U/ml IL-4 (PeproTech), and 1000 U/ml GM-CSF (Leukine; Berlex). The quality of DC prepara- Ag-specific T cell activation tions with an immature phenotype was controlled by morphologic analysis and by verification of high expression of CD1a and CD11c. Furthermore, Immature DC were cultured alone or in the presence of LPS or fibroblasts. immature DC did not express CD83 and displayed only low expression of The DC were challenged 24 h later with 0.5 ␮g/ml tetanus toxoid (Cal- CD86, CD80, and HLA-DR. biochem) for 4 h. Subsequently, 2 ϫ 104 DC in 100 ␮l of proliferation medium (RPMI 1640 (Invitrogen Life Technologies) supplemented with Coculture of DC and fibroblasts 10% FCS (Promo Cell), 1% penicillin/streptomycin (Invitrogen Life Tech- nologies), 1% L-glutamine (Invitrogen Life Technologies), nonessential Fibroblasts were seeded onto 24-well plates and cultured until they became amino acids (Invitrogen Life Technologies), and 10 mM HEPES (Invitro- confluent. Immature day 5 DC (5 ϫ 105) were directly added to fibroblast gen Life Technologies)) were seeded into 96 round-bottom wells. To ex- monolayers and were cocultured for 24 h in DC-RPMI 1640 supplemented clude the possibility that contaminating fibroblasts remaining from the DC with 2% FCS. In other experiments, fibroblasts and DC were separated by fibroblast coculture might affect T cell function, all fibroblasts were re- culture inserts (BD Biosciences; 3-␮m pore size) to prevent direct cell-cell moved using an anti-Thy-1 Ab coupled to goat anti-mouse magnetic beads contact. As controls, immature DC were cultured for 24 h either in DC- (Dynal Biotech), as previously described (17, 19, 20). Autologous T cells RPMI 1640 supplemented with 2% FCS or in the presence of 1 ␮g/ml LPS were isolated from peripheral mononuclear cells using the negative T cell to induce full DC maturation. isolation kit II from Miltenyi Biotec, according to the manufacturer’s pro- 5 Flow cytometry tocol. T cells (2 ϫ 10 ) were resuspended in 100 ␮l of proliferation medium. In preliminary experiments, an optimal DC:TC ratio of 1:10 was Cells were harvested and washed twice in PBS, and 2 ϫ 105 cells were established. Thus, DC were cocultured with purified T cells at a ratio of incubated with the requisite fluorescently labeled mAbs. Cells were ana- 1:10 in round-bottom wells for up to 5 days. lyzed by flow cytometry (Epics XL; Beckman Coulter) after washing twice For detection of proliferation autologous, purified T cells (1 ϫ 107 with PBS/10% Gelafusal (Serumwerke Bernburg). For staining of intra- cells/ml PBS) were labeled with 1 ␮M carboxyfluorescein diacetate suc- cellular TNF-␣, cells were stimulated for2hinthepresence of 100 ng/ml cinimidyl ester (CFDA) (Molecular Probes) for 5 min at 37°C. Cells were brefeldin A (Calbiochem). Cells were then harvested, fixed with 4% para- washed, resuspended (1 ϫ 107/ml proliferation medium), and incubated for formaldehyde in PBS for 5 min at room temperature, and permeabilized 30 min at 37°C with 5% CO2. Proliferation was detected after 5 days of using 0.1% saponin (Sigma-Aldrich) for 10 min. Cells were subsequently coculture by measuring the decrease in the fluorescence staining (17, 21). 4968 FIBROBLASTS INDUCE MATURATION OF DC

FIGURE 1. Fibroblasts and DC colocalize in human skin. Sections of atopic skin lesions were stained with anti-Thy-1 Abs to detect fibroblasts and with anti-CD1a Abs to detect DC. a, Isotype control Ab. b and c, Double staining with anti-Thy-1 (red) and anti-CD1a (brown). Colocalization of CD1aϩ cells (DC) and Thy-1ϩ fibroblasts is indicated (arrows). Similar expression patterns were detected in five different specimens. Downloaded from http://www.jimmunol.org/

T cells were identified by staining with an anti-CD3 PE Ab. Dead cells adhesion of immature DC to fibroblasts (Fig. 2b). Incubation of were excluded from the analyses using propidium iodide (Calbiochem) DC with an isotype control Ab or with an anti-MHC I Ab did not staining. The number of proliferating cells per 5000 nonproliferating cells significantly affect the adhesion. However, blocking the counter- was calculated from flow cytometry data. ␤ receptors for Thy-1 and ICAM-1 on DC, the 2 integrins CD11a/ Mixed leukocyte reaction CD18 and CD11b/CD18, respectively, also inhibited the adhesion Immature DC were cultured alone or in the presence of LPS or fibroblasts. of DC to fibroblasts (Fig. 2b). Allogeneic T cells were isolated and labeled, as described above. For MLR, To prove the involvement of Thy-1 and ICAM-1 in the adhesion by guest on October 2, 2021 1 ϫ 105 allogeneic T cells were cultured with DC in different ratios for 6 of immature DC to fibroblasts, adhesion of immature DC to days. Proliferation was detected, as described above. ICAM-1-, Thy-1-, and vector-transfected CHO cells was studied in Cytokine determination static adhesion assays. As seen in Fig. 2c, DC adhered strongly to Thy-1- and ICAM-1-transfected cells, whereas very few cells TNF-␣ was detected in conditioned medium from DC cocultures using an bound to vector-transfected control cells. These results together ELISA (BD Biosciences). Experiments were performed in duplicate. suggest that the adhesion of monocyte-derived DC to fibroblasts is Statistical analysis mediated by ICAM-1 and Thy-1. Data were analyzed using Student’s t test. Values of p Ͻ 0.05 were con- sidered significant. The interaction of DC with fibroblasts induces maturation of DC in vitro Results Next, we determined whether the interaction of DC with fibro- ␤ DC interact with fibroblasts via ICAM-1 and Thy-1 and their 2 blasts affects the maturation of DC. Immature monocyte-derived integrin counterreceptors DC were cultured for 24 h either alone or on a confluent monolayer To study whether DC interact with stromal cells, we analyzed the of fibroblasts. As a control, DC were stimulated with 1 ␮g/ml LPS relative localization of CD1a-positive DC and fibroblasts in skin. to induce full DC maturation. DC stimulated with LPS or by con- Frozen tissue sections of human atopic eczema lesions were tact to fibroblasts exhibited typical veils, as described for LPS- stained with an anti-CD1a biotin Ab to identify DC. Fibroblasts stimulated DC (data not shown). Subsequently, DC were har- were identified using the anti-Thy-1 Ab AS02 (17, 19, 20, 22). A vested, and the expression of DC maturation markers was analyzed close colocalization of CD1a-positive DC (brown) and Thy-1-pos- using flow cytometry. Stimulation of DC with LPS induced the itive fibroblasts (red) was found in the dermis of atopic eczema maturation of DC, as demonstrated by the enhanced expression of lesions (Fig. 1). CD80, CD86, CD83, and HLA-DR (Fig. 3a). Interestingly, the Next, the interaction of human DC and fibroblasts was investi- coculture of DC with allogeneic (data not shown) as well as gated in vitro. DC were generated from CD14ϩ PBMC using GM- syngeneic (Fig. 3a) fibroblasts also induced the expression of CSF- and IL-4-containing medium. Immature (5-day) DC adhered these maturation markers on DC (Fig. 3a). However, activation to fibroblasts in static adhesion assays (Fig. 2a). To identify the of DC by fibroblasts resulted in less CD83-positive DC com- molecules mediating this interaction, Thy-1- and ICAM-1-medi- pared with LPS-stimulated DC (Fig. 3a). To investigate ated adhesion was inhibited using blocking Abs to Thy-1 (clone whether soluble factors or direct cell-cell contact were respon- BC9-G2) (16, 17) and ICAM-1 (clone R6.5D6) that bind to fibro- sible for this effect, we separated DC and fibroblasts using blasts. Blocking either Thy-1 or ICAM-1 significantly reduced the Transwell inserts. As shown in Fig. 3a, expression of CD80, The Journal of Immunology 4969

FIGURE 2. DC adhere to fibroblasts via Thy-1 and ICAM-1. Adhesion of immature DC to Thy-1-, ICAM-1-, vector-transfected CHO cells or primary fibroblasts was examined using static adhesion assays. a, Representative picture of immature DC adhering to confluent monolayers of fibroblasts. Examples of DC are indicated by arrows. b, Adhesion of immature DC to human dermal fibroblasts was studied after blocking of Thy-1 and ICAM-1 on fibroblasts, and CD11a, CD11b, and CD18 on DC using function-blocking Abs. Isotype control Abs (isotype ctr) and Abs that bind to the cell surface, but do not influence adhesion (anti-HLA-ABC), were used as negative controls to exclude unspecific effects due to Ab binding. Adherent cells were counted in 10 Downloaded from independent fields using a fluorescence microscope (magnification, ϫ20). The adhesion of untreated cells was set to 100%. Data represent the mean Ϯ SD p Ͻ 0.05 compared with the adhesion without Ab. c, The attachment of immature DC to Thy-1-, ICAM-1-, and ,ء .of four independent experiments vector-transfected CHO cells (CHO.Thy-1, CHO.ICAM-1, and CHO.vec, respectively) was studied in static adhesion assays. The number of adherent cells p Ͻ 0.05 compared with ,ء .was counted in 10 independent fields (magnification, ϫ20). Data represent the mean Ϯ SD of four independent experiments vector-transfected CHO cells. http://www.jimmunol.org/ CD86, and CD83 was also induced on DC when DC and fibro- Another hallmark of DC maturation is the decrease in their blasts were cocultured in the absence of direct cell-cell contact. phagocytic activity (23). Immature DC are highly active phago- However, this expression was significantly lower in comparison cytic cells, as detected by the ingestion of FITC-labeled E. coli at with DC that were directly cocultured with fibroblasts, indicat- 37°C (Fig. 3b). However, maturation induced by LPS or by co- ing that both direct cell-cell contact and soluble factors are culture of DC with fibroblasts both resulted in a distinct loss of involved in the fibroblast-induced DC maturation. Stimulation phagocytic activity (Fig. 3b). DC incubated at 4°C did not inter- of DC maturation by coculturing with fibroblasts from the same nalize FITC E. coli (Fig. 3b). Together these data demonstrate that donor excluded the possibility of an unspecific allogenic stim- the interaction of DC with fibroblasts results in phenotypic DC by guest on October 2, 2021 ulation of DC by fibroblasts. maturation.

FIGURE 3. Adhesion of immature DC to dermal fibroblasts induces maturation of DC. Immature DC were cultured either alone, in the presence of 1 ␮g/ml LPS, or on a confluent monolayer of dermal fibroblasts (FB) for 24 h. Fibroblasts were obtained from the same donor as DC (FB syngeneic). In some experiments, fibroblasts and DC were separated by semipermeable culture inserts to prevent direct cell-cell contact (DC-[FB syngeneic-TW]). a, Expression of CD80, CD83, CD86, and HLA-DR was analyzed by flow cytometry. One representative experiment of five is shown. b, Phagocytotic capability of immature DC, LPS-stimulated DC, and fibroblast-stimulated DC (DCFB) was assessed by the uptake of FITC-labeled E. coli (black line). Cells without E. coli incubation (gray line) or with E. coli incubation at 4°C (black dotted line) were used as negative control. Data represent one of three experiments. 4970 FIBROBLASTS INDUCE MATURATION OF DC Downloaded from http://www.jimmunol.org/ FIGURE 4. TNF-␣ is involved in the fibroblast-induced maturation of DC. Immature DC were cultured either alone, in the presence of 1 ␮g/ml LPS, or on a monolayer of dermal fibroblasts (DCϩFB) for 24 h. In some experiments, DC and fibroblasts were separated by a semipermeable membrane to prevent direct cell-cell contact (DC ϩ [fibroblasts]). a, TNF-␣ was detected in conditioned medium from the different cultures using ELISA. Stimulation p Ͻ 0.05 compared with ,ء .of DC with LPS or with fibroblasts induced TNF-␣ secretion. Data represent the mean Ϯ SD of four independent experiments DC. b, To determine the cellular source of TNF-␣ secretion, fibroblasts, DC, and cocultures of fibroblasts and DC were cultured in the presence of brefeldin A. Intracellular TNF-␣ was detected after2hbyflowcytometry in DC (CD11cϩ cells) and fibroblasts (CD11cϪ cells). CD11cϩ DC produced TNF-␣. Data represent one of three independent experiments. c, DC and fibroblasts were cocultured in the presence of a function-blocking anti-TNF-␣ Ab (black dotted line) or an isotype control (black line). After 24 h, DC were stained with Abs against CD80, CD83, CD86, HLA-DR, and an isotype control Ab, respectively (gray line). Blocking of TNF-␣ resulted in a significantly reduced expression of the costimulatory molecules. Data represent one representative experiment of four. by guest on October 2, 2021

TNF-␣ plays a central role in the fibroblast-induced CD86, and HLA-DR on DC (Fig. 4c), the control Ab having no maturation of DC effect. Because TNF-␣ plays a central role in the regulation of DC mat- The importance of the CD40/CD40L interaction for DC regu- uration (23–25), we analyzed the involvement of TNF-␣ in the lation is well established (26). We therefore studied its role in the fibroblast-induced maturation of DC. TNF-␣ levels in the medium fibroblast-induced DC maturation. CD40 is expressed on DC and from DC fibroblast cocultures were increased dramatically, similar fibroblasts (data not shown). However, CD40L could not be de- to the levels in the medium from LPS-stimulated DC (Fig. 4a). In tected either on DC or on fibroblasts (data not shown). Addition- contrast, untreated DC or fibroblasts alone did not secrete detect- ally, we could not detect soluble CD40L in conditioned medium able amounts of TNF-␣. To determine whether induction of from immature DC or stimulated DC (data not shown). Finally, TNF-␣ secretion is dependent on cell-cell contact, DC and fibro- addition of a CD40L-blocking Ab to the coculture of DC and fi- blasts were cocultivated, but separated by a semipermeable mem- broblasts did not affect fibroblast-induced DC maturation (data not brane to prevent direct cell-cell contact. Interestingly, virtually no shown). Due to the absence of CD40L expression in the coculture TNF-␣ secretion could be detected, indicating that direct cell-cell of DC and fibroblasts, we consider it highly unlikely that CD40/ contact is essential to induce TNF-␣ secretion in DC/fibroblast CD40L interactions are involved in the fibroblast-induced matu- cocultures (Fig. 4a). ration of DC. To identify the cells responsible for TNF-␣ secretion in DC Together, these data suggest that the interaction of DC with ␣ fibroblast cocultures, DC were cocultured with fibroblasts in the fibroblasts induces the secretion of TNF- by DC, which in turn presence of brefeldin A to block TNF-␣ secretion and thereby to stimulates the maturation of DC in an autocrine manner, as re- increase intracellular TNF-␣ protein levels. Intracellular TNF-␣ in flected by the up-regulated expression of costimulatory molecules. both cell populations was subsequently analyzed using flow cytometry. As shown in Fig. 4b, CD11cϩ DC expressed TNF-␣, ICAM-1 is involved in the fibroblast-induced maturation of DC whereas CD11cϪ fibroblasts did not. The results above indicate that the adhesion of DC and fibroblasts Next, we investigated the role of TNF-␣ secretion on fibroblast- is at least partially mediated by the interaction of Thy-1 and ␤ induced DC maturation. DC and fibroblasts were cocultured for ICAM-1 with 2 integrins (Fig. 2). To investigate whether this 24 h in the presence of either a blocking anti-TNF-␣ Ab or an interaction is able to directly induce TNF-␣ secretion, DC were isotype control Ab. Blocking of TNF-␣ resulted in a substantially cultured with ICAM-1-, Thy-1-, or vector-transfected CHO cells. reduced expression of the costimulatory molecules CD80, CD83, As shown in Fig. 5a, DC produced high amounts of TNF-␣ when The Journal of Immunology 4971

FIGURE 5. Interaction of DC with ICAM-1 induces maturation of DC. Immature DC were cultured on monolayers of ICAM-1 (CHO.ICAM1)-, Thy-1 (CHO. Thy1)-, or vector-transfected CHO cells (CHO.vec), or on dermal fibroblasts (FB). As control, DC were cultured alone or in the presence of 1 ␮g/ml LPS. a, TNF-␣ was detected in conditioned medium after 24-h cocul- .p Ͻ 0.05 compared with CHO.vec ,ء .ture using ELISA b, Expression of CD80 and CD86 on CD11cϩ DC was detected by flow cytometry after 24 h of coculture. -p Ͻ 0.05 compared with CHO.vec; #, p Ͻ 0.05 com ,ء pared with DC without stimulation. c, Immature DC were cultured on fibroblasts (DCFB) that had been pre- incubated with a function-blocking anti-ICAM-1 or an anti-Thy-1 Ab. An anti-MHC-I Ab was used as a negative control. TNF-␣ was detected in conditioned medium using ELISA. Secretion of TNF-␣ in the presence Downloaded from of the control Ab (anti-MHC I) was set to 100%. Data represent the mean Ϯ SD of five independent experi- .p Ͻ 0.05 ,ء .ments http://www.jimmunol.org/ cocultured with ICAM-1-transfected cells, whereas no TNF-␣ se- costimulatory molecules such as CD80 and CD86 on DC, similar cretion was induced in cocultures of DC with Thy-1- or vector- to the situation with LPS treatment or fibroblast coculture of DC transfected CHO cells. In addition, coculture of DC with ICAM- (Fig. 5b). In contrast, Thy-1- and vector-transfected cells did not 1-transfected cells also up-regulated the expression of the affect the expression of the costimulatory molecules. by guest on October 2, 2021 FIGURE 6. Fibroblast-stimulated DC (DCFB) activate T cells (TC). a, DC, LPS-stimulated DC, and FB- stimulated DC (DCFB) were pulsed with tetanus toxoid and were cocultured with autologous, purified T cells in a ratio of 1:10. Expression of CD25 on T cells was analyzed after 3 days by flow cytometry. Data represent 1 of 10 independent experiments. b, Proliferation of CD3ϩ T cells cocultured with tetanus toxoid-pulsed DC was detected by flow cytometry as the decrease of CFDA labeling after 5 days. The number of proliferated CD3ϩ cells per 5000 nonproliferated cells is indicated in brackets. The data show the FACS plots of one representative of five independent experiments. c, The data represent the p Ͻ ,ء .mean Ϯ SD of triplicate wells 0.05 compared with basal T cell proliferation. d, In MLR DC, LPS-stimulated DC and FB-stimulated DC (DCFB) were cocultured with allogeneic T cells in the indicated ratios. Proliferation of CD3ϩ T cells was detected by flow cytometry as the decrease of CFDA labeling after 6 days. Data show one representative experiment of three. 4972 FIBROBLASTS INDUCE MATURATION OF DC

Next, we investigated the importance of ICAM-1 in the fibro- fibroblasts to induce DC to express TNF-␣. Nevertheless, partial blast-induced DC maturation. Cultured fibroblasts expressed maturation of DC as demonstrated by their expression of costimu- ICAM-1, as detected by flow cytometry (data not shown). Treat- latory molecules was observed in these physically separated co- ment of fibroblast DC cocultures with a blocking anti-ICAM-1 Ab cultures (Fig. 3a), suggesting that TNF-␣ is not the only signal for (clone R6.5D6) resulted in only a 25% reduction in the TNF-␣ fibroblast-induced DC maturation. These observations also suggest production. Treatment of similar cocultures with function-blocking that in addition to the TNF-␣ produced by DC, fibroblast-derived anti-Thy-1 Abs (BC9-G2) (16, 17) or anti-MHC I Abs (HLA- soluble mediators play a role in fibroblast-mediated DC matura- ABC) did not affect the TNF-␣ production in cocultures of DC and tion, a notion supported by the observation that blocking anti- fibroblasts. Additionally, blocking of ICAM-1 did not affect the TNF-␣ Abs only partially inhibited the expression of costimula- up-regulation of costimulatory molecules in DC that were cocul- tory molecules on DC in response to direct coculture with tured with fibroblasts (data not shown). In conclusion, ICAM-1 is fibroblasts (Fig. 4c). involved in the fibroblast-induced TNF-␣ secretion, but other mol- The data we present in this study show that the direct cell-cell ecules on DC and fibroblasts must also be engaged for full DC contact required for fibroblast-induced maturation of DC has both maturation. adhesive and inductive components. The importance of the inter- ␤ action of 2 integrins (LFA-1 and Mac-1) on leukocytes for the Fibroblast-matured DC activate T cells adhesion to the endothelium via ICAM-1 and Thy-1 is well estab- To characterize the fibroblast-matured DC functionally, we inves- lished (17), and our data suggest that these interactions also me- tigated their capacity to activate T cells. Immature DC, LPS-stim- diate the adhesion of DC to fibroblasts, at least in part. In the ulated DC, and fibroblast-matured DC were pulsed with tetanus leukocyte-endothelium interaction, ICAM-1 acts not only as an toxoid and cocultured with purified autologous CD3ϩ T cells. T adhesion molecule, but also as a signaling receptor that induces the Downloaded from cells were harvested after 3 days, and the expression of CD25 on reorganization of F-actin filaments and thereby regulates the af- the CD3ϩ T cells was detected by flow cytometry. T cells alone finity of LFA-1 for its receptors (27). Consistently, ICAM-1 on did not express significant amounts of CD25 (Ͻ2%) (Fig. 6a). fibroblasts represents a signal for the initiation of DC maturation, Untreated DC only stimulated CD25 expression on Ͻ10% of the T because ICAM-1 on the cell surface induced the maturation of DC, cells (Fig. 6a). As expected, LPS-stimulated DC robustly induced as shown by the up-regulation of costimulatory molecules on DC CD25 expression on 30% of T cells. Importantly, fibroblast-ma- and the expression of TNF-␣ (Fig. 5). Furthermore, blocking of http://www.jimmunol.org/ tured DC also strongly activated T cells, as demonstrated by CD25 ICAM-1 on fibroblasts reduced TNF-␣ production by DC. The role expression on 36% of the T cells. The activation of T cells by of ICAM-1 expression by fibroblasts in the maturation of DC in fibroblast-matured DC was further reflected in the enhanced T cell vivo is likely to be physiologically and pathologically relevant, as proliferation that was detected as a decrease in CFDA labeling increased expression of ICAM-1 on fibroblasts during inflamma- after 5 days of coculture. Activation of DC by LPS or fibroblasts tory and immune reactions has been reported in several diseases, enhanced the T cell proliferation two times more efficiently com- such as renal allograft rejection, rheumatoid arthritis, and hyper- pared with T cells cocultured with immature DC (Fig. 6, b and c). trophic scars (28–30). We also found ICAM-1 expressed on en- In accordance, in MLR, fibroblast-matured DC also significantly dothelial cells and fibroblasts in inflamed skin, whereas in healthy by guest on October 2, 2021 stimulated proliferation of allogeneic T cells. LPS- and fibroblast- skin only very few ICAM-1-positive cells were detectable (data matured DC significantly increased the proliferation up to a ratio not shown). of DC/TC of 1:40 (Fig. 6d). Although our data implicate an involvement of ICAM-1 in ␤ the fibroblast-induced maturation of DC, the ICAM-1/ 2 inte- Discussion grin interaction is not the only mechanism involved. Ab-medi- DC in peripheral tissues mature in response to antigenic stimula- ated blocking of ICAM-1 on fibroblasts only partially reduced tion and migrate to secondary lymphoid organs, where they acti- TNF-␣ production by DC (Fig. 5). Moreover, this blocking had vate T cells. During their migration to secondary lymphoid organs, no influence on the expression of costimulatory molecules on DC have to travel through the stromal microenvironment com- DC. As we show that direct cell-cell contact is partially required prised of the extracellular matrix and stromal cells, such as for the fibroblast-induced maturation of DC (Fig. 3a), these data fibroblasts, macrophages, and endothelial cells. In this study, suggest that cell surface molecules in addition to ICAM-1 must we show that DC are located in close proximity to stromal also be involved in fibroblast-induced maturation of DC. How- fibroblasts in inflamed skin. In accordance, a Thy-1- and ICAM- ever, anti-Thy-1-blocking Abs had no effect on fibroblast-in- 1-dependent adhesion of DC to fibroblast could be detected in duced DC maturation. Despite the importance of the CD40/ vitro. Coculture experiments demonstrate that fibroblasts are CD40L interaction for DC maturation, we could find no effective in inducing both phenotypic and functional maturation evidence for a role of the CD40/CD40L interaction in the fi- of DC in a manner that is dependent on both direct cell-cell broblast-induced maturation of DC (26). CD40L could not be contact as well as soluble mediators. The resulting fibroblast- detected either on DC, on fibroblasts, or in the coculture of both matured DC are able to support T cell-driven immune re- (data not shown); thus, we consider it unlikely that CD40/ sponses. Together, these data suggest that dermal fibroblasts CD40L play a role in the fibroblast-induced DC maturation. can be active regulators of DC functions and of DC-driven cu- We found that fibroblast-matured DC were functionally active, taneous immune responses. as they were capable of Ag-dependent activation of autologous T The fibroblast-dependent expression of TNF-␣ in DC plays a cells as well as of stimulation of T cell proliferation in allogeneic central role in the fibroblast-induced maturation of DC, because MLR in vitro. TNF-␣ production in DC was strongly induced as a result of the Why should fibroblasts induce DC maturation and subsequent T direct coculture of DC with fibroblasts, and the blocking of TNF-␣ cell activation? We suggest that fibroblasts are not activated in activity inhibited this DC maturation. Virtually no TNF-␣ expres- resting skin (e.g., no ICAM-1) and thus are not able to stimulate sion was induced in DC when they were cocultured with fibro- DC maturation. In contrast, in inflammatory conditions, in the blasts, but physically separated from them with a semipermeable tumor environment, or in granulation tissue, resident cells are ac- membrane, indicating that direct cell-cell contact is required for tivated (7, 11, 31–33). We suppose that under such conditions The Journal of Immunology 4973

fibroblasts are activated and may acquire the capability to stimu- 4. Sallusto, F., B. Palermo, D. Lenig, M. Miettinen, S. Matikainen, I. Julkunen, late DC. In vitro, culture of fibroblasts (culture on plastic, mono- R. Forster, R. Burgstahler, M. Lipp, and A. Lanzavecchia. 1999. Distinct patterns and kinetics of chemokine production regulate dendritic cell function. Eur. J. Im- layer conditions, presence of serum) might mimic this activation of munol. 29: 1617–1625. fibroblasts, resulting in a phenotype that is capable of stimulating 5. Dudda, J. C., J. C. Simon, and S. Martin. 2004. Dendritic cell immunization route ϩ DC. The high expression of ICAM-1 and secretion of IL-6, both determines CD8 T cell trafficking to inflamed skin: role for tissue microenvironment and dendritic cells in establishment of T cell-homing subsets. J. Immu- markers of cell activation by cultured fibroblasts, might reflect this nol. 172: 857–863. activation of cultured fibroblasts. Thus, we suggest that fibroblasts 6. Lappin, M. B., J. M. Weiss, V. Delattre, B. Mai, H. Dittmar, C. Maier, K. Manke, may act as mediators of inflammation and adaptive immunity. In S. Grabbe, S. Martin, and J. C. Simon. 1999. Analysis of mouse dendritic cell migration in vivo upon subcutaneous and intravenous injection. Immunology 98: accordance, the activation of DC by activated neutrophils and ac- 181–188. tivated NK cells has been shown. The authors suggested that these 7. Parsonage, G., A. D. Filer, O. Haworth, G. B. Nash, G. E. Rainger, M. Salmon, activated neutrophils or NK cells might act as mediators of innate and C. D. Buckley. 2005. A stromal address code defined by fibroblasts. Trends Immunol. 26: 150–156. and adaptive immunity (34–36). 8. Manna, P. P., B. Duffy, B. Olack, J. Lowell, and T. Mohanakumar. 2001. Acti- Our data underline the impact of the fibroblast from the dermal vation of human dendritic cells by porcine aortic endothelial cells: transactivation stromal microenvironment for the differentiation and regulation of of naive T cells through costimulation and cytokine generation. Transplantation 72: 1563–1571. DC function. Several studies have reported the importance of the 9. Randolph, G. J., S. Beaulieu, S. Lebecque, R. M. Steinman, and W. A. Muller. stromal microenvironment for immune responses in other contexts. 1998. Differentiation of monocytes into dendritic cells in a model of transendo- For example, Dudda et al. (5) and Johansson-Lindbom et al. (37) thelial trafficking. Science 282: 480–483. 10. Silzle, T., G. J. Randolph, M. Kreutz, and L. A. Kunz-Schughart. 2004. The have demonstrated the role of the stromal microenvironment for fibroblast: sentinel cell and local immune modulator in tumor tissue. Int. J. Can- the imprinting of a target organ-specific homing of T cells. Fur- cer 108: 173–180. thermore, CD34ϩ progenitor cells differentiate into defined subsets 11. Buckley, C. D., D. Pilling, J. M. Lord, A. N. Akbar, D. Scheel-Toellner, and Downloaded from of DC when cocultured with splenic fibroblasts in a manner that is M. Salmon. 2001. Fibroblasts regulate the switch from acute resolving to chronic ϩ persistent inflammation. Trends Immunol. 22: 199–204. dependent on M-CSF and CD40 (38). Consistently, CD34 stem 12. Smith, R. S., T. J. Smith, T. M. Blieden, and R. P. Phipps. 1997. Fibroblasts as cells differentiated into DC in response to cell-cell contact with sentinel cells: synthesis of chemokines and regulation of inflammation. cutaneous fibroblasts and M-CSF (39). However, little is known Am. J. Pathol. 151: 317–322. 13. Smith, T. J. 2005. Insights into the role of fibroblasts in human autoimmune about the effect of stromal cells on immature DC. Splenic endo- diseases. Clin. Exp. Immunol. 141: 388–397.

ϩ http://www.jimmunol.org/ thelial-like CD106 stromal cells have been reported to induce DC 14. Hogaboam, C. M., M. L. Steinhauser, S. W. Chensue, and S. L. Kunkel. 1998. maturation (40). These stromal cells induced a defined regulatory Novel roles for chemokines and fibroblasts in interstitial fibrosis. Kidney Int. 54: 2152–2159. DC phenotype that was able to induce T cell activation without the 15. Pap, T., U. Muller-Ladner, R. E. Gay, and S. Gay. 2000. Fibroblast biology: role promotion of proliferation. In contrast, we show in this study that of synovial fibroblasts in the pathogenesis of rheumatoid arthritis. Arthritis Res. dermal fibroblasts stimulate DC that are capable of supporting an 2: 361–367. 16. Saalbach, A., A. Wetzel, U. F. Haustein, M. Sticherling, J. C. Simon, and immune response and concomitant T cell proliferation. These data ␣ ␤ U. Anderegg. 2005. Interaction of human Thy-1 (CD 90) with the integrin v 3 therefore suggest that different compartments of the stromal mi- (CD51/CD61): an important mechanism mediating melanoma cell adhesion to croenvironment may have distinct regulatory effects on the im- activated endothelium. Oncogene 24: 4710–4720. mune response. Thus, in the future, defined stromal components 17. Wetzel, A., T. Chavakis, K. T. Preissner, M. Sticherling, U. F. Haustein, U. Anderegg, and A. Saalbach. 2004. Human Thy-1 (CD90) on activated endo- by guest on October 2, 2021 may be used for directed regulation of DC activation and sub- thelial cells is a counterreceptor for the leukocyte integrin Mac-1 (CD11b/CD18). sequent promotion/inhibition of an immune response. More- J. Immunol. 172: 3850–3859. 18. Wetzel, A., T. Wetzig, U. F. Haustein, M. Sticherling, U. Anderegg, J. C. Simon, over, given that the stromal microenvironment can determine and A. Saalbach. 2006. Increased neutrophil adherence in psoriasis: role of the organ-specific T cell homing, it will be interesting in future human endothelial cell receptor Thy-1 (CD90). J. Invest. Dermatol. 126: studies to determine whether stromal fibroblasts play any role in 441–452. this process. 19. Saalbach, A., G. Aust, U. F. Haustein, K. Herrmann, and U. Anderegg. 1997. The fibroblast-specific MAb AS02: a novel tool for detection and elimination of hu- In summary, we demonstrate in this study for the first time that man fibroblasts. Cell Tissue Res. 290: 593–599. dermal fibroblasts actively participate in the regulation of DC 20. Saalbach, A., R. Kraft, K. Herrmann, U. F. Haustein, and U. Anderegg. 1998. The function, resulting in the maturation of DC and the subsequent monoclonal antibody AS02 recognizes a protein on human fibroblasts being highly homologous to Thy-1. Arch. Dermatol. Res. 290: 360–366. T cell activation. Our data thus identify a new pathway of Ag- 21. Mannering, S. I., J. S. Morris, K. P. Jensen, A. W. Purcell, M. C. Honeyman, independent DC maturation induced by the local stromal P. M. van Endert, and L. C. Harrison. 2003. A sensitive method for detecting microenvironment. proliferation of rare autoantigen-specific human T cells. J. Immunol. Methods 283: 173–183. 22. Saalbach, A., U. Aneregg, M. Bruns, E. Schnabel, K. Herrmann, and U. F. Haustein. 1996. Novel fibroblast-specific monoclonal antibodies: properties and Acknowledgments specificities. J. Invest. Dermatol. 106: 1314–1319. We are grateful to H. Gedicke and A. Rohmer for excellent technical 23. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of assistance. immunity. Nature 392: 245–252. 24. Lutz, M. B., and G. Schuler. 2002. Immature, semi-mature and fully mature dendritic cells: which signals induce tolerance or immunity? Trends Immunol. 23: Disclosures 445–449. 25. Voigtlander, C., S. Rossner, E. Cierpka, G. Theiner, C. Wiethe, M. Menges, The authors have no financial conflict of interest. G. Schuler, and M. B. Lutz. 2006. Dendritic cells matured with TNF can be further activated in vitro and after subcutaneous injection in vivo which converts their tolerogenicity into immunogenicity. J. Immunother. 29: 407–415. References 26. Mackey, M. F., J. R. Gunn, C. Maliszewsky, H. Kikutani, R. J. Noelle, and 1. Balazs, M., F. Martin, T. Zhou, and J. Kearney. 2002. Blood dendritic cells R. J. Barth, Jr. 1998. Dendritic cells require maturation via CD40 to generate interact with splenic marginal zone B cells to initiate T-independent immune protective antitumor immunity. J. Immunol. 161: 2094–2098. responses. Immunity 17: 341–352. 27. Anderson, M. E., and T. J. Siahaan. 2003. Targeting ICAM-1/LFA-1 interaction 2. Huang, F. P., N. Platt, M. Wykes, J. R. Major, T. J. Powell, C. D. Jenkins, and for controlling autoimmune diseases: designing peptide and small molecule in- G. G. MacPherson. 2000. A discrete subpopulation of dendritic cells transports hibitors. Peptides 24: 487–501. apoptotic intestinal epithelial cells to T cell areas of mesenteric lymph nodes. 28. Castagnoli, C., M. Stella, G. Magliacani, S. Ferrone, and P. M. Richiardi. 1994. J. Exp. Med. 191: 435–444. Similar ectopic expression of ICAM-1 and HLA class II molecules in hypertro- 3. Tan, J. K., and H. C. O’Neill. 2005. Maturation requirements for dendritic cells phic scars following thermal injury. Burns 20: 430–433. in T cell stimulation leading to tolerance versus immunity. J. Leukocyte Biol. 78: 29. Hill, P. A., I. W. Main, and R. C. Atkins. 1995. ICAM-1 and VCAM-1 in human 319–324. renal allograft rejection. Kidney Int. 47: 1383–1391. 4974 FIBROBLASTS INDUCE MATURATION OF DC

30. Ishikawa, H., S. Hirata, Y. Andoh, H. Kubo, N. Nakagawa, Y. Nishibayashi, and 35. Walzer, T., M. Dalod, E. Vivier, and L. Zitvogel. 2005. Natural killer cell-den- K. Mizuno. 1996. An immunohistochemical and immunoelectron microscopic dritic cell crosstalk in the initiation of immune responses. Exp. Opin. Biol. Ther. study of adhesion molecules in synovial pannus formation in rheumatoid arthritis. 5(Suppl. 1): S49–S59. Rheumatol. Int. 16: 53–60. 36. Walzer, T., M. Dalod, S. H. Robbins, L. Zitvogel, and E. Vivier. 2005. Natural- 31. Wandel, E., A. Grasshoff, M. Mittag, U. F. Haustein, and A. Saalbach. 2000. killer cells and dendritic cells: “l’union fait la force.” Blood 106: 2252–2258. Fibroblasts surrounding melanoma express elevated levels of matrix metallopro- 37. Johansson-Lindbom, B., M. Svensson, O. Pabst, C. Palmqvist, G. Marquez, teinase-1 (MMP-1) and intercellular adhesion molecule-1 (ICAM-1) in vitro. R. Forster, and W. W. Agace. 2005. Functional specialization of gut CD103ϩ Exp. Dermatol. 9: 34–41. dendritic cells in the regulation of tissue-selective T cell homing. J. Exp. Med. 32. McClain, S. A., M. Simon, E. Jones, A. Nandi, J. O. Gailit, M. G. Tonnesen, 202: 1063–1073. D. Newman, and R. A. Clark. 1996. Mesenchymal cell activation is the rate- 38. Briard, D., B. Azzarone, and D. Brouty-Boye. 2005. Importance of stromal de- limiting step of granulation tissue induction. Am. J. Pathol. 149: 1257–1270. terminants in the generation of dendritic and natural killer cells in the human 33. Buckley, C. D., D. Pilling, J. M. Lord, A. N. Akbar, D. Scheel-Toellner, and spleen. Clin. Exp. Immunol. 140: 265–273. M. Salmon. 2001. Fibroblasts regulate the switch from acute resolving to chronic 39. Mollah, Z. U., S. Aiba, H. Manome, Y. Yoshino, and H. Tagami. 2002. Cord persistent inﬂammation. Trends Immunol. 22: 199–204. blood CD34ϩ cells differentiate into dermal dendritic cells in co-culture with 34. van Gisbergen, K. P., M. Sanchez-Hernandez, T. B. Geijtenbeek, and Y. cutaneous ﬁbroblasts or stromal cells. J. Invest. Dermatol. 118: 450–460. van Kooyk. 2005. Neutrophils mediate immune modulation of dendritic cells 40. Zhang, M., H. Tang, Z. Guo, H. An, X. Zhu, W. Song, J. Guo, X. Huang, T. Chen, through glycosylation-dependent interactions between Mac-1 and DC-SIGN. J. Wang, and X. Cao. 2004. Splenic stroma drives mature dendritic cells to differ- J. Exp. Med. 201: 1281–1292. entiate into regulatory dendritic cells. Nat. Immunol. 5: 1124–1133. Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021