Anatomic Localization of Immature and Mature Dendritic Cells in an Ectopic Lymphoid Organ: Correlation with Selective Expression in Rheumatoid This information is current as Synovium of October 1, 2021. Guillaume Page, Serge Lebecque and Pierre Miossec J Immunol 2002; 168:5333-5341; ; doi: 10.4049/jimmunol.168.10.5333

http://www.jimmunol.org/content/168/10/5333 Downloaded from

References This article cites 42 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/168/10/5333.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 by guest on October 1, 2021

*average

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 © 2002 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Anatomic Localization of Immature and Mature Dendritic Cells in an Ectopic Lymphoid Organ: Correlation with Selective Chemokine Expression in Rheumatoid Synovium

Guillaume Page,* Serge Lebecque,† and Pierre Miossec1*

It remains to be clarified whether dendritic cells (DC) reach the rheumatoid arthritis (RA) synovium, considered an ectopic lymphoid organ, as mature cells or undergo local maturation. We characterized by immunohistochemistry the DC subsets and used tonsils as a control. Immature and mature DC were defined by CD1a and DC-lysosome-associated membrane /CD83 expression, respectively. Immature DC were mainly detected in the lining layer in RA synovium. Mature DC were exclusively detected in the lymphocytic infiltrates. The DC-lysosome-associated membrane protein/CD1a ratio was 1.1 in RA synovium and 5.3 in tonsils, suggesting the relative accumulation of immature DC in RA synovium. We then focused on the expression of

CCL20/CCR6 and CCL19/CCR7, CCL21/CCR7 chemokine/receptor complex, which control immature and mature DC migration Downloaded from respectively. A close association was observed between CCL20-producing cells and CD1a؉ cells, suggesting the contribution of CCL20 to CCR6؉ cell homing. Conversely, CCL21 and CCL19 expression was only detected in perivascular infiltrates. The association among CCL19/21-producing cells, CCR7 expression, and mature DC accumulation is in line with the roles of these in mature CCR7؉ DC homing to lymphocytic infiltrates. The role of DC in disease initiation and perpetuation makes chemokines involved in DC migration a potential therapeutic target. The Journal of Immunology, 2002, 168: 5333–5341. http://www.jimmunol.org/ endritic cells (DC)2 are the most potent APCs and play of DC in such tissue. In addition, synovium dissociation with en- a central role in the processing and presentation of Ags zymes may affect phenotypes and functions with a difficult recov- D to T cells during the immune response (1). After Ag ery of such rare cells. Accordingly, we have selected immunohis- exposure, immature DC migrate to draining lymph nodes where tochemical techniques to investigate the different DC subsets they differentiate into mature DC to present Ags to T cells (1). found in RA synovium. We used several newly defined markers of DC are found in the rheumatoid arthritis (RA) synovium and are the immature and mature DC subsets. CD1a, first described as a likely to play an important role in its pathogenesis. This conclusion marker of Langerhans cells (LC) in skin , was later de- is supported by the histological observations demonstrating MHC scribed as a marker of immature DC (13). Conversely, DC-lyso- ϩ by guest on October 1, 2021 class II APC clustered with T cells around blood vessels, some- some-associated membrane protein (LAMP) and CD83 have been times in germinal center (GC) structures, and the association of associated with more mature DC subsets. In vitro studies of DC HLA-DR alleles with disease severity (2–6). However, the func- differentiation have shown that CD83 expression appears before tion of DC according to phenotype and subset in RA synovium that of DC-LAMP, suggesting that CD83ϩ DC are less mature remains to be clarified, in particular regarding their mode of than DC-LAMPϩ DC (14, 15). We selected tonsils, an active lym- migration and accumulation (7–10). Recently, differentiated DC phoid organ, as a control. We analyzed DC frequency and local- expressing nuclear RelB have been detected in RA synovium, ization with the aim of studying the respective immature/mature predominantly located in perivascular mononuclear cell aggre- DC compartmentalization pattern in RA synovium. Such compart- gates (11). mentalization has been observed in breast carcinoma tissue where RA synovium has several, but not all, of the characteristics of immature DC were detected within the tumor and mature DC in lymphoid organs with regard to cellular composition and organi- peritumoral areas (16). zation (12), but it is unknown whether DC reach the synovium as To study the mode of migration of DC, we also looked at the mature cells or undergo local maturation. In particular, an accu- expression of several chemokines and their associated receptors. mulation of immature DC could be secondary to inflammation. Macrophage-inflammatory protein-3␣, a CC chemokine, recently Such studies are difficult to perform because of the low frequency renamed CCL20 (17), and its receptor CCR6 (18) are critical for the recruitment of immature DC (19). CCL20 plays a major role in *Department of Immunology and Rheumatology and Institut National de la Sante et epithelial colonization by LC in response to inflammation (20). de la Recherche Me´dicale Unite´ 403, Hoˆpital Edouard Herriot, Lyon, France; and Macrophage-inflammatory protein-3␤/CCL19 and 6-chemokine/ † Schering-Plough Laboratory for Immunological Research, Dardilly, France CCL21 contribute through their associated receptor CCR7 to the Received for publication October 11, 2001. Accepted for publication February accumulation of Ag-loaded mature DC in T cell-rich areas of lym- 25, 2002. phoid organs (19). These chemokines attract in vitro-derived DC The costs of publication of this article were defrayed in part by the payment of page ␣ charges. This article must therefore be hereby marked advertisement in accordance following CCR7 induction by CD40 ligand, TNF- , or LPS stim- with 18 U.S.C. Section 1734 solely to indicate this fact. ulation (19, 21). In situ detection of CCL19 showed its expression 1 Address correspondence and reprint requests to Dr. Pierre Miossec, Clinical Immu- within the T cell areas of secondary lymphoid organs (22), with nology Unit, Departments of Immunology and Rheumatology, Hoˆpital Edouard Her- CCL21 having a wider distribution. riot, 69437 Lyon Cedex 03, France. E-mail address: [email protected] In this work we report the coexistence of both immature and 2 Abbreviations used in this paper: DC, ; AEC, 3-amino 9-ethylcarba- mature DC subsets in RA synovium. The close association be- zole; GC, germinal center; LAMP, lysosome-associated membrane protein; LC, ϩ Langerhans cell; RA, rheumatoid arthritis. tween immature CD1a DC and CCL20-producing cells suggests

Copyright © 2002 by The American Association of Immunologists 0022-1767/02/$02.00 5334 DC SUBSETS IN RA SYNOVIUM

ϩ the contribution of CCL20 to the homing of immature CCR6 gen (blue color; Vector Laboratories, Burlingame, CA). In negative control cells into RA synovium. Conversely, the colocalization of CCL21/ sections, one of the two primary Abs was omitted. CCL19, CCR7, and mature DC in perivascular infiltrates suggests Detection of chemokines and their associated receptors a role for these chemokines in the homing of mature CCR7ϩ DC For detection of chemokines, goat polyclonal Abs to CCL19, CCL20, or in lymphocytic infiltrates. CCL21 (R&D Systems, Abingdon, U.K.) were used, followed by biotin- ylated rabbit anti-goat IgG (DAKO), and was developed using streptavidin- Materials and Methods peroxidase. CCR6 and CCR7 were detected using anti-CCR6 (R&D Sys- Collection of samples tems) and anti-CCR7 (BD Biosciences) mouse mAbs. In double-step immunohistochemical staining, after an initial blocking Synovial samples were obtained from 12 patients with RA, defined ac- with rabbit serum, avidin-biotin, and BSA, primary Ab to CCL20 was cording to the revised criteria of the American College of Rheumatology added, followed by rabbit biotinylated anti-goat IgG and streptavidin-per- (formerly the American Rheumatism Association) (23). To assess inflam- oxidase. Peroxidase was developed by AEC. Mouse mAb to CD1a (IgG2b) matory synovium, samples were obtained from patients undergoing knee or was followed by rat anti-mouse IgG2a/2b (BD Biosciences) and mouse wrist synovectomy, and not at a late stage, such as for joint replacement. alkaline phosphatase anti-alkaline phosphatase (DAKO). Alkaline phos- Tonsils were obtained from children undergoing tonsillectomy. phatase was revealed using Fast Blue as chromogen (blue color; Vector The study was first initiated with frozen sections. However, immuno- Laboratories). staining using frozen specimens, despite good Ag preservation, was not sufficient to clearly define the morphology of DC and to allow their quan- Quantification of positive cells tification. Accordingly, an effort was made to develop immunostaining in Because of the high heterogeneity of RA synovium and the low frequency paraffin-embedded sections, which allows a better preservation of histo- of DC subsets, quantification was performed with a method first used for logical patterns. Samples were fixed in 4% phosphate-buffered paraformal- the estimation of new blood vessel formation in breast tumor (29). This dehyde and then embedded in paraffin. Four-micrometer sections were cut method allows quantification of cells in hot spots, defined in this study as Downloaded from and mounted on glass slides (adhesive slides, SuperFrost; CML, Nemours, synovium areas containing the highest density of positive cells. Accord- France). To detect Ag expression in paraffin-embedded sections, Ag re- ingly, two hot spots per rheumatoid synovium section were selected. In trieval procedures were performed, including incubation in either citrate each spot positive cells were counted in 10 consecutive high-power fields buffer (10 mM, pH 6) or EDTA buffer (1 mM, pH 8), followed by micro- (ϫ500). One field corresponded to 0.3 mm2, and thus one spot to 3 mm2. wave oven incubation (three times for 3 min each time). The number of positive cells per two hot spots was averaged, and results Following staining with hematoxylin, samples were classified according are expressed as the number of positive cells per square millimeter. to the following criteria (24): diffuse infiltration of T cells and B cells and absence of lymphoid organization (n ϭ 4), T cell- aggregates without Statistical analysis http://www.jimmunol.org/ GC (n ϭ 5), and presence of GC (n ϭ 3). Results were expressed as the mean Ϯ SD. Levels of marker expression Detection of DC subsets were compared using the nonparametric Mann-Whitney U test between RA synovium and tonsils and between subsets of RA synovium. For single staining, an immunoperoxidase technique using 3-amino 9-eth- ylcarbazole (AEC; red color; DAKO, Glostrup, Denmark) or 3,3Ј-diami- Results nobenzidine tetrahydrochloride (DAKO) as chromogen was performed. Expression of HLA-DR, CD11c, and CD40 in RA synovium Briefly, endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Sections were incubated with several mouse mAbs recognizing We investigated DC subsets in sections obtained from 12 RA pa- the following molecules (see also Table I): CD1a (IgG2b; BD Biosciences, tients and used tonsils, an active secondary lymphoid organ as a Pont de Claix, France), CD11c (IgG2b; CamFolio, San Jose, CA), Langerin control. We used several mAbs against specific and nonspecific by guest on October 1, 2021 (IgG1; Schering-Plough, Dardilly, France), CD40 (IgG1; Schering- DC markers (Table I). HLA-DR, CD11c, and CD40, known to be Plough), HLA-DR (IgG2b; BioSource, Camarillo, CA), CD83 (IgG2a; Im- munotech, Marseilles, France), and DC-LAMP (IgG1; Immunotech). expressed by DC, were detected in all RA synovium samples. (Refs. 13–15 and 25–28, see Table I for details). After overnight incubation HLA-DR, CD11c, and CD40 were highly expressed in every sam- at 4°C and washing, the sections were incubated with biotinylated anti- ple (44 Ϯ 11, 32 Ϯ 11, and 35 Ϯ 12 positive cells/mm2, using the mouse IgG, followed by streptavidin-peroxidase (DAKO). The peroxidase hot spot analysis, respectively; Table II). Similar expression was was developed by AEC, and Mayer hematoxylin (DAKO) was used as the counterstain. In negative control sections the primary Abs were omitted, or observed in tonsils (not significantly different vs RA synovium). irrelevant Ab was applied at the same concentration as the primary Ab. These markers were detected in many cell types other than DC, In double-step immunohistochemical staining, after an initial blocking such as T cells, B cells, and synoviocytes, and localized in the with rat serum, avidin-biotin (DAKO), and BSA, primary Ab to DC-LAMP lining and sublining layers as well as in infiltrates. was followed by rat biotinylated anti-mouse IgG1 and streptavidin-perox- idase. Peroxidase was developed by 3,3Ј-diaminobenzidine tetrahydrochlo- Immature DC are present within RA synovium ride. Mouse mAb to CD1a was followed by rat anti-mouse IgG2a/2b (BD Biosciences) and mouse alkaline phosphatase anti-alkaline phosphatase We then focused on more DC-specific markers. Immature DC ϩ (DAKO). Alkaline phosphatase was revealed using Fast Blue as chromo- were defined by CD1a expression. Immature CD1a DC were

Table I. Markers and Abs used for DC subset detection

Expression by DC Abs Catalog Markers Immaturea Mature (origin/isotype) References Source

CD1a ϩϩϩ (13) Ϫ Mouse/IgG2b 347430 BD Biosciences CD11c ϩϩ (25) ϩϩ Mouse/IgG2b 550004 CamFolio CD40 ϩ (26) ϩϩϩ Mouse/IgG1 Schering-Plough CD83 ϩ (14) ϩϩϩ Mouse/IgG2a 2069 Immunotech HLA-DR ϩϩb (27) ϩϩc Mouse/IgG2b AHU0182 BioSource DC-LAMP Ϫ (15) ϩϩϩ Mouse/IgG1 IM3448 Immunotech Langerin ϩϩϩd (28) Ϫ Mouse/IgG1 Schering-Plough

a Corresponding references are listed in parentheses. b Intracellular. c Surface. d LC. The Journal of Immunology 5335

Table II. Quantification of DC subsets in RA synovium and tonsilsa

Immature DC Mature DC RA Patients (histological classification) CD1a Langerin CD83 DC-LAMP CD11c CD40 HLA-DR

1 (lymphoid follicles) 5 0 6 10 21 31 36 2 (lymphoid follicles) 4 0 0 7 18 24 40 3 (diffuse) 0 0 3 3 32 37 51 4 (GC) 26 0 15 18 52 60 50 5 (lymphoid follicles) 8 0 6 8 47 56 52 6 (diffuse) 1 0 1 1 24 23 29 7 (diffuse) 2 0 0 5 42 34 66 8 (lymphoid follicles) 14 0 2 9 31 30 45 9 (diffuse) 0 0 0 0 21 22 29 10 (GC) 9 0 5 11 25 26 33 11 (lymphoid follicles) 3 0 3 7 33 37 42 12 (GC) 10 0 17 12 42 40 54 All (mean Ϯ SD) 7 Ϯ 7* 0** 5 Ϯ 6** 8 Ϯ 5** 32 Ϯ 11** 35 Ϯ 12* 44 Ϯ 11* Control (tonsils)b 15 Ϯ 214Ϯ 235Ϯ 779Ϯ 347Ϯ 465Ϯ 635Ϯ 3

a Results are expressed as mean of DC marker-positive cells per square millimeter based on the hot spot analysis (see Materials and Methods for details). RA synovium vs tonsils: *, NS; **, p Ͻ 0.001. b n ϭ 3. Shown are the means Ϯ SD. Downloaded from detected in 10 of 12 RA synovium samples. The number of CD1aϩ Mature CD83ϩ cells were mainly detected at the periphery of cells per square millimeter in RA synovium ranged from 1 to 26 the perivascular or lymphocytic infiltrates (Fig. 1C). In contrast, cells (mean Ϯ SD, 7 Ϯ 7; Table II). In tonsils, that number was DC-LAMPϩ DC were not observed at the periphery, but inside the ϳ2-fold higher (15 Ϯ 2 positive cells/mm2), but without reaching lymphocytic infiltrates (Fig. 1E), the site of interactions between significance (not significantly different vs RA synovium). mature DC and . In tonsils, DC-LAMPϩ DC were http://www.jimmunol.org/ The LC subset, also expressing CD1a marker in skin epithe- exclusively detected in the T cell zone (Fig. 1F), whereas CD83ϩ lium (13), has been characterized by the specific expression of cells were found in both T cell and GC zones (Fig. 1D). langerin/DCGM4. This represents a marker of epithelial tissue. In synovium, most of the mature DC also had a dendritic-like As expected, no LC were found in RA synovium. Conversely, morphology (inset of Fig. 1, C and E). Some of the DC-LAMPϩ LC in tonsils (14 Ϯ 2 positive cells/mm2; p Ͻ 0.001 vs RA or CD83ϩ cells had a less common aspect, with a plasmacytoid- synovium) were exclusively present in the epithelium. Such like appearance and a remote nucleus in a large cytoplasm (Fig. finding excludes the migration of epithelial-derived DC to the 1E, inset).

synovium. To further clarify the dichotomy between immature and mature by guest on October 1, 2021 In RA synovium immature CD1aϩ DC were preferentially lo- DC subsets in RA synovium, double staining using anti- calized in the lining (inset of Fig. 1A) or sublining layer as well as DC-LAMP and anti-CD1a Abs was performed. The lack of coex- at the periphery of perivascular infiltrates (Fig. 1A). CD1aϩ cells pression of DC-LAMP and CD1a confirms the presence of inde- were detected in the lining layer of all 10 CD1aϩ samples, whereas pendent immature and mature DC subsets (Fig. 1, G and H). The immature DC in perivascular infiltrates were present in only 4 of DC-LAMP/CD1a ratio was 8/7 ϭ 1.1 in RA synovium vs 79/15 ϭ 12 samples (patients 4, 8, 10, and 12). In tonsils, CD1aϩ DC were 5.3 in tonsils ( p Ͻ 0.001 vs RA synovium), suggesting the relative detected in the T cell zone and the epithelium (Fig. 1B). accumulation of immature DC in RA synovium. Most of the immature CD1aϩ DC showed a different dendritic morphology compared with the stellate shape of LC in skin epi- thelium, with large veils expanding in every direction from the cell CCL20 is expressed in RA synovium body with long (10 ␮m) and thin processes. In RA synovium, To study the mode of migration of DC in response to chemokines, ϩ immature CD1a DC had only a few long processes (Fig. 1A). we first analyzed the expression pattern of CCL20 and its receptor CCR6, which is known to control the migration of immature DC. Mature DC are preferentially localized in the perivascular or In RA synovium CCL20 was mainly expressed in the lining layer lymphocytic infiltrates (Fig. 2A), but also in the perivascular infiltrates (Fig. 2C). CCL20 Mature DC were analyzed for the expression of CD83 and DC- expression showed a large degree of heterogeneity between the LAMP. In vitro studies of DC differentiation have shown that different RA synovium samples. In 4 of 12 samples no CCL20 CD83 expression appears before that of DC-LAMP, suggesting expression was detected. In the eight CCL20ϩ RA synovium, the that CD83ϩ DC are less mature than DC-LAMPϩ DC (15). densities of CCL20ϩ cells per square millimeter in the lining layer CD83ϩ cells were detected in 9 of 12 samples, and DC-LAMPϩ and/or perivascular infiltrates were 9 Ϯ 9 and 9 Ϯ 11, respectively, cells were found in 11 of 12 samples. The mean numbers of with a large degree of heterogeneity (Table III). In samples 11 and CD83ϩ cells and DC-LAMPϩ cells per square millimeter in RA 12, CCL20 was only expressed in the lining layer. In tonsils, synovium were 5 Ϯ 6 and 8 Ϯ 5, respectively (Table II). In tonsils CCL20-producing cells were more common (38 Ϯ 5 positive cells/ the expression of CD83 and DC-LAMP was 7-fold higher (35 Ϯ mm2; p Ͻ 0.001 vs RA synovium) and were exclusively detected 7 positive cells/mm2; p Ͻ 0.001 vs RA synovium) and 10-fold in the epithelial crypts (Fig. 2E). higher (79 Ϯ 3 positive cells/mm2; p Ͻ 0.001 vs RA synovium) In RA synovium CCL20-producing cells detected in the lining compared with RA synovium, and the DC-LAMP/CD83 ratio was layer extend the demonstration of secretion of CCL20 by syno- 79/35 ϭ 2.3, indicating a relative accumulation of the most mature viocytes (30). In tonsils CCL20-producing cells in epithelial crypts subset. In RA synovium, that ratio was 8/5 ϭ 1.6 (not significant were identified as epithelial cells characterized by their elongated vs tonsils). shape. 5336 DC SUBSETS IN RA SYNOVIUM

FIGURE 1. Independent subsets of immature CD1aϩ DC and mature DC- LAMPϩ and CD83ϩ DC are detected in RA synovium and tonsils. Immunostain- ing of paraffin-embedded RA synovium and tonsils sections using mouse anti- CD1a mAb shows CD1aϩ cells (red) in the perivascular infiltrates (A) and the

lining layer (inset of A) in RA synovium Downloaded from and in the epithelium and the T cell zone of tonsils (B). Immature CD1aϩ DC pre- sented a dendritic-like morphology char- acterized by a few processes from the cell body (A). CD83ϩ cells (red) in RA synovium are detected at the periphery of the perivascular or lymphocytic infil- http://www.jimmunol.org/ trates in RA synovium (C) and in GC and T cell zone in tonsils (D). Mature CD83ϩ DC have a dendritic-like appearance (in- set of C). DC-LAMPϩ cells (red) are found inside the lymphocytic infiltrates in RA synovium (E) and with a strong expression in the T cell areas of tonsils (F). DC-LAMPϩ DC have a dendritic- like morphology, with very short or long by guest on October 1, 2021 veils (inset of E). Double staining with anti-DC-LAMP (brown) and anti-CD1a (blue) Abs shows a lack of coexpression of DC-LAMP and CD1a in RA syno- vium (G) and tonsils (H). Magnification: A, C, D, E, G, H, and inset of A, ϫ400; B and F, ϫ250; inset of C and F, ϫ1000.

CCR6, the receptor for CCL20, is highly expressed in RA 2F). In both RA synovium and tonsils, CCR6-expressing cells had synovium various morphotypes, including cells with a dendritic morphology. CCR6 expression was detected in all RA synovium samples. Ϯ 2 Close association between CCL20-producing cells and immature CCR6 was highly expressed (39 20 positive cells/mm )inlym- ϩ phocytic infiltrates and at the periphery of blood vessels (Fig. 2D). CD1a DC In some cases CCR6 was also observed in the lining layer where As observed in Table II, the lack of CCL20 expression was asso- CCL20 was highly expressed (Fig. 2B). In tonsils, CCR6 was ciated with the absence (samples 3 and 9) or low number (patients highly detected (66 Ϯ 6 positive cells/mm2; not significant vs RA 6 and 7) of CD1aϩ cells. This suggested a link between the accu- synovium) in the T cell zone, and accumulation of CCR6ϩ cells mulation of immature CD1aϩ DC and the expression of CCL20 in was directly adjacent to CCL20-expressing epithelial cells (Fig. RA synovium. This observation was extended by double staining The Journal of Immunology 5337

FIGURE 2. CCL20 and its associ- ated receptor CCR6 are detected near CD1aϩ cells in the lining layer and the perivascular infiltrates in RA synovium Downloaded from and tonsils. Immunostaining with anti- CCL20 (A and C) and anti-CCR6 (B and D) Abs of serial paraffin-embedded RA synovium sections shows the asso- ciation between CCL20-producing cells and CCR6ϩ cells in the lining layer, as

well as in the perivascular infiltrates. In http://www.jimmunol.org/ tonsils CCL20 expression is found only in epithelial crypts (E), and that of its associated receptor, CCR6, is found in the T cell zone adjacent to the CCL20- producing cells (F). Double staining with anti-CCL20 (red) and anti-CD1a (blue) Abs shows the association of CCL20-producing cells and immature CD1aϩ cells in the lining layer (G) and in perivascular infiltrates (H). Such an by guest on October 1, 2021 association is also observed in tonsils where immature CD1aϩ cells are adja- cent to CCL20-expressing epithelial cells (I). Magnification for all: ϫ400.

with anti-CCL20 and anti-CD1a Abs. In both RA synovium and CCL19 and CCL21 are mainly localized in perivascular tonsils we observed a close association between immature CD1aϩ infiltrates ϩ DC and CCL20-producing cells. In tonsils, CCL20 cells in epi- We then looked at CCL19 and CCL21 expression, the CCR7 li- ϩ thelial crypts localized near, but not in close contact with, CD1a gands known to control the migration of mature DC. In RA syno- cells (Fig. 2I). In RA synovium this association was mainly de- vium the frequencies of CCL21ϩ and CCL19ϩ cells were 13 Ϯ 14 tected in the lining layer (Fig. 2G) and to a lesser degree in perivas- and 7 Ϯ 10/mm2, respectively (Table III). Staining of CCL21ϩ cular infiltrates (Fig. 2H). cells was more intense than that of CCL19ϩ cells. As observed for 5338 DC SUBSETS IN RA SYNOVIUM

Table III. Quantification of chemokines and their associated receptors in RA synovium and tonsilsa

CCL20 RA Patients (histological classification) Lining layer Perivascular infiltrates CCR6 CCL21 CCL19 CCR7

1 (lymphoid follicles) 25 6 52 12 5 7 2 (lymphoid follicles) 9 10 19 13 12 20 3 (diffuse) 0 0 20 0 0 0 4 (GC) 18 33 63 38 24 31 5 (lymphoid follicles) 15 20 45 15 0 9 6 (diffuse) 0 0 28 0 0 0 7 (diffuse) 0 0 23 0 0 0 8 (lymphoid follicles) 26 10 50 8 0 5 9 (diffuse) 0 0 14 0 0 0 10 (GC) 4 26 80 24 19 23 11 (lymphoid follicles) 7 0 37 10 0 6 12 (GC) 8 0 41 37 23 19 All (mean Ϯ SD) 9 Ϯ 9** 9 Ϯ 11** 39 Ϯ 20* 13 Ϯ 14** 7 Ϯ 10** 10 Ϯ 11*** Control (tonsils)b 38 Ϯ 566Ϯ 682Ϯ 543Ϯ 531Ϯ 4

a Results are expressed as mean of positive cells per square millimeter based on the hot spot analysis (see Materials and Methods for details). RA synovium versus tonsils: .p Ͻ 0.005 ,ءءء ;p Ͻ 0.001 ,ءء ;NS ,ء b n ϭ 3. Shown are the means Ϯ SD. Downloaded from

CCL20, the expression of CCL21 and CCL19 revealed a large Discussion degree of heterogeneity between samples. No expression of We demonstrate in this work the presence in RA synovium of CCL21 and CCL19 was detected in samples characterized by a independent immature and mature DC subsets using the CD1a- diffuse infiltrate without organization of the lymphocytic aggre- and DC-LAMP/CD83-specific markers, respectively. We focused http://www.jimmunol.org/ gates (patients 3, 6, 7, and 9). All samples with such features on these three markers as reflecting the different steps of in vitro (patients 1, 2, 5, 8, and 11) showed the expression of CCL21, DC differentiation, with CD1a and DC-LAMP linked to early and whereas CCL19 was only detected in two of those samples (pa- late stages of differentiation, respectively, CD83 expression ap- tients 1 and 2). Both chemokines were detected in samples show- pearing before that of DC-LAMP. ing lymphocytic infiltrates with GC (patients 4, 10, and 12). More Compared with tonsils, accumulation of DC in RA synovium ϩ ϩ Ϯ CCL21 and CCL19 cells were found in samples with GC (33 was lower, with a relative accumulation of immature DC. In ton- 8 and 22 Ϯ 3 positive cells/mm2 , respectively) than in samples sils, the mature/immature ratios (DC-LAMP/CD1a) were 5.3 and without GC (11 Ϯ 3 and 3 Ϯ 5 positive cells/mm2, respectively; Ͻ Ͻ 1.1 in RA synovium. Such a difference may lead to functional

p 0.001 and p 0.001 vs samples with GC). by guest on October 1, 2021 changes between a lymphoid organ in a normal position (in this In RA synovium CCL21 and CCL19 were mainly localized in study, the tonsils) or an ectopic position (in this study, the syno- perivascular or lymphocytic infiltrates (Fig. 3, A and B). We also vium). Immature DC are highly efficient in Ag uptake. In the con- observed the expression of CCL21 and CCL19 in endothelium text of RA, further clarification is difficult because the causative cells (Fig. 3, C and D). In tonsils, CCL21 (82 Ϯ 5 positive cells/ Ag remains unknown. Immature DC could take up Ag directly in mm2; p Ͻ 0.001 vs RA synovium) and CCL19 (43 Ϯ 5 positive 2 synovium, as suggested by the presence of a long process from cells/mm ; p Ͻ 0.001 vs RA synovium) were highly present and ϩ detected in T cell areas, with a wider distribution for CCL21 (Fig. CD1a cells, but immature DC could have already loaded Ag 3, G and H). They were expressed around GC. before their entry into the synovium. Conversely, mature DC are These observations suggested a link between the expression of highly efficient APC and were localized in lymphocytic infiltrates CCL21 or CCL19 and the organization of lymphoid infiltrates of the synovium, where they can interact with T and B cells. Such in GC. interaction could occur at various stages of the RA process (31). Immature CD1aϩ DC were localized in the lining or sublining CCR7, the receptor for CCL21 and CCL19, is expressed in layers, as well as in the perivascular infiltrates. Mature DC were lymphocytic infiltrates mainly detected in perivascular infiltrates. In RA synovium, de- CCR7 expression was detected in 8 of 12 RA synovium samples at tection of both immature and mature DC subsets in perivascular a frequency of 10 Ϯ 11/mm2 and at 31 Ϯ 4/mm2 in tonsils ( p Ͻ infiltrates argues for the lack of compartmentalization, such as is 0.005 vs RA synovium). As observed for CCL19 and CCL21, observed for breast carcinoma tissue, where immature DC were CCR7 was only detected in samples characterized by lymphocytic found within the tumor and mature DC in peritumoral areas (16). aggregates with or without GC. In RA synovium, CCR7ϩ cells However, in RA synovium different localization patterns were ob- ϩ ϩ ϩ were mainly detected in lymphocytic infiltrates and rarely near a served for DC-LAMP and CD83 DC in these infiltrates. CD83 ϩ vascular endothelium without any infiltrate (Fig. 3E). Some of cells, considered less mature than DC-LAMP cells, were exclu- them had a dendritic morphology (Fig. 3E, inset). In tonsils most sively detected at the periphery, whereas the fully mature DC- ϩ of the CCR7ϩ cells localized in GC and the T cell zone (Fig. 3F), LAMP cells were also observed inside the lymphocytic infil- ϩ ϩ some also with a dendritic morphology (Fig. 3F, inset). trates. Differential localization of CD83 and DC-LAMP cells Staining of serial sections with anti-CCR7, anti-CCL21, and anti- could reflect the effect of maturation on the response of DC to the DC-LAMP indicated the close association between CCL21ϩ cells chemokines that control their homing into infiltrates. (Fig. 3K), CCR7ϩ cells (Fig. 3I), and mature DC (Fig. 3J). This Such differences in the mature/immature ratio suggest changes observation is in line with an interaction between the presence of in the chemokine contribution. Indeed, chemokines are nonspecific mature CCR7ϩ DC and that of CCL21- and CCL19-producing soluble factors that control the migration of DC and many other cells. cells into the synovium. We analyzed the expression of CCL20, The Journal of Immunology 5339

FIGURE 3. CCL19 and CCL21 are detected in perivascular infiltrates and vascular endothelium in RA synovium and tonsils. CCR7, their associated re- ceptor, is expressed in the lymphocytic infiltrates. Immunostaining of paraffin- embedded sections was performed with anti-CCL21 (A and C, RA synovium; G, tonsils) and anti-CCL19 (B and D, RA synovium; H, tonsils) Abs. ϩ ϩ CCL21 (A) and CCL19 cells (B) are Downloaded from found in the perivascular infiltrates, with a stronger expression for CCL21. Some vascular endothelial cells in RA synovium are also CCL21ϩ (C)or CCL19ϩ (D), again with a higher in- tensity for CCL21. In tonsils CCL21 (G) is highly expressed only in the T http://www.jimmunol.org/ cell zone, where CC19-producing cells arelesscommon(H).Stainingusinganti- CCR7 Ab (brown) shows CCR7ϩ cells in the lymphocytic infiltrates in RA sy- novium (E). Some CCR7ϩ cells have a dendritic morphology (inset of E). In tonsils, most CCR7ϩ cells are detected in the GC and some in the T cell zone

(F), with a dendritic morphology (inset by guest on October 1, 2021 of F). Immunostaining of serial RA synovium sections with anti-CCR7 (I), anti-DC-LAMP (J), and anti-CCL21 (K) Abs shows that CCL21, a chemoat- tractant for mature DC, and its associ- ated receptor, CCR7, colocalize in the lymphocytic infiltrates with mature DC. Magnification: AÐC, inset of E and F, and IÐK ϫ400; G and H, ϫ100; E and F, ϫ250; D, ϫ600.

the ligand of CCR6, and CCL19 and CCL21, the ligands of CCR7. IL-1, IL-17, and TNF-␣ was associated with an up-regulation of These chemokines have been classified as chemoattractant for in CCL20 production (30). Synoviocytes are predominant in the lin- vitro-derived immature and mature DC, respectively (19, 32). It ing layer and represent the leading CCL20-producing cells, similar was critical to demonstrate whether such findings are in line with to keratinocytes in psoriasis skin and epithelial cells in tonsil the in vivo situation. The close association between CCL20-pro- crypts, which are the sites of CCL20 production (19, 20, 33). The ducing cells and CD1aϩ immature DC observed in this study was expression of CCR6 near the lining layer suggests that it represents indeed an in situ confirmation of these previous results obtained in one site of entry of immature CCR6ϩ DC into the RA synovium. vitro. CCL20 was highly expressed in the lining layer, extending Indeed, such DC were observed inside the lining layer. However, our recent results where in vitro incubation of synoviocytes with high endothelial venules should also be considered a site of 5340 DC SUBSETS IN RA SYNOVIUM immature DC entry, as shown by the high expression of CCR6 with GC and their absence in samples with diffuse infiltrates sug- and the close association between CCL20-producing cells and gest that these chemokines are implicated in GC formation and ϩ CD1a cells in the perivascular infiltrates. migration of mature DC in these lymphocytic aggregates. A sim- Compared with tonsils, where mature DC represent the major ilar conclusion was reached with the B cell-attracting chemokine-1 DC subset, the RA synovium was characterized by the relative (40). However, stronger expression of CCL21 compared with accumulation of immature DC. This suggests a relative defect in CCL19 in RA synovium may indicate a greater contribution of DC maturation in RA synovium. When considering the various CCL21 to the migration of mature DC. In addition, CCL19 ex- ϩ used for in vitro generation of hemopoietic CD34 -de- pression (5 of 12 samples) was less common than that of CCL20 ␣ rived DC (GM-CSF and TNF- ) and monocyte-derived DC (GM- production (9 of 12). When considering the quantification of their CSF and IL-4) (34, 35), the lack of IL-4 in RA synovium contrasts associated receptors, the CCR6/CCR7 ratio was 39/10 ϭ 3.9 in RA ␣ with the high levels of GM-CSF and TNF- (36). The defect in synovium, suggesting a higher contribution of CCL20/CCR6 to DC differentiation compared with tonsils could be linked to the that of CCR7 and its ligands, CCL19 and CCL21, in the homing of lack of IL-4 leading to defective Ag processing. In addition, T DC. Such differences could result in a reduced migration of mature cell-derived cytokines are characterized by high IL-17 and low DC from blood. The combined results observed here are depicted IL-4 production, resulting in a proinflammatory pat- in Fig. 4, which represents a model of DC subsets and migration tern (37). IL-10, another anti-inflammatory cytokine known to patterns in RA synovium. inhibit the capacity of synovial macrophages to function as APC (38), had no immunosuppressive effect on rheumatoid sy- Chemokines acting as chemoattractants of DC are key factors in novial fluid DC compared with peripheral blood DC (39). These stimulating immunity against tumor Ags. Adenovirus-mediated Downloaded from results indicate the contribution of synovium cytokine micro- transfer of CCL20 to murine tumors induced the local accu- environment that results in complex changes in DC recruitment mulation of DC, leading to tumor-specific cellular immunity and and function. suppression of tumor (41). Vaccination of mice with the CCL19- The potent local defect of maturation of mature DC in RA syno- transduced C3L5 resulted in the rejection of CCL19-transduced vium led to evaluation of the expression of CCR7 and its ligands, tumor, showing an antitumor activity for CCL19 (42). The role of CCL19 and CCL21, known chemoattractants for in vitro-derived ma- DC in disease initiation and perpetuation makes DC subsets a ture DC. The detection of both chemokines in vascular endothelium potential therapeutic target in RA. This also suggests the control of http://www.jimmunol.org/ and perivascular infiltrates suggests that migration of mature chemokines involved in DC migration. The link between CCL20 CCR7ϩ DC occurs from blood to RA synovium through vascular production and inflammation makes this chemokine a potential endothelium. The expression of CCL19 and CCL21 in RA samples therapeutic target. by guest on October 1, 2021 FIGURE 4. A model of DC subsets and migration patterns in RA syno- vium. In response to local inflamma- tion and the production of proinflam- matory cytokines (1) following an unknown event, immature DC are at- tracted to the synovium in response to the local production of CCL20 in the perivascular infiltrates and the lining layer (2). CCR6ϩ immature DC mi- grate from blood through vascular en- dothelium and possibly also from sy- novial fluid through the lining layer. It is also possible that CCR6ϩ cells mi- grate to synovial fluid toward CCL20 expressed within the lining layer (3). The presence of mature DC in syno- vium results from the combined effects of cell interactions and cytokine micro- environment (4). A defect in produc- tion of differentiation factors, such as IL-4, may favor a relative accumula- tion of immature DC. However, the de- tection of CCL19 and CCL21 in perivascular infiltrates and vascular en- dothelium argues for a direct migration of CCR7ϩ mature DC from blood into the RA synovium, where they can in- teract with lymphocytes, leading to po- tent local activation. The Journal of Immunology 5341

References 22. Ngo, V. N., H. L. Tang, and J. G. Cyster. 1998. Epstein-Barr virus-induced molecule 1 ligand chemokine is expressed by dendritic cells in lymphoid tissues 1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of and strongly attracts naive T cells and activated B cells. J. Exp. Med. 188:181. immunity. Nature 392:245. 23. Arnett, F. C., S. M. Edworthy, D. A. Bloch, D. J. McShane, J. F. Fries, 2. Itoh, J., K. Kinjoh, A. Ohyama, M. Nose, and M. Kyogoku. 1992. Application of N. S. Cooper, L. A. Healey, S. R. Kaplan, M. H. Liang, H. S. Luthra, et al. 1988. two-color immunofluorescence staining to demonstration of T cells and HLA- The American Rheumatism Association 1987 revised criteria for the classifica- DR-bearing cells in rheumatoid synovitis. J. Histochem. Cytochem. 40:1675. tion of rheumatoid arthritis. Arthritis Rheum. 31:315. 3. Van Dinther-Janssen, A. C., S. T. Pals, R. Scheper, F. Breedveld, and 24. Klimiuk, P. A., J. J. Goronzy, J. Bjo¨rnsson, R. D. Beckenbaugh, and C. J. Meijer. 1990. Dendritic cells and high endothelial venules in the rheumatoid C. M. Weyand. 1997. Tissue cytokine patterns distinguish variants of rheumatoid synovial membrane. J. Rheumatol. 17:11. synovitis. Am. J. Pathol. 151:1311. 4. Cush, J. J., and P. E. Lipsky. 1988. Phenotypic analysis of synovial tissue and 25. Lenz, A., M. Heine, G. Schuler, and N. Romani. 1993. Human and murine dermis peripheral blood lymphocytes isolated from patients with rheumatoid arthritis. contain dendritic cells: isolation by means of a novel method and phenotypical Arthritis Rheum. 31:1230. and functional characterization. J. Clin. Invest. 92:2587. 5. Fox, R. I., S. Fong, N. Sabharwal, S. A. Carstens, P. C. Kung, and J. H. Vaughan. 26. Inaba, K., M. Inaba, M. Witmer-Pack, K. Hatchcock, R. Hodes, and 1982. Synovial fluid lymphocytes differ from peripheral blood lymphocytes in R. M. Steinman.1995. Expression of B7 costimulator molecules on mouse den- patients with rheumatoid arthritis. J. Immunol. 128:351. dritic cells. Adv. Exp. Med. Biol. 378:65. 6. Winchester, R. J., and P. K. Gregersen. 1988. The molecular basis of suscepti- 27. Fearnley, D. B., A. D. McLellan, S. I. Mannering, B. D. Hock, and D. N. Hart. bility to rheumatoid arthritis: the conformational equivalence hypothesis. Spring- 1997. Isolation of human blood dendritic cells using the CMRF-44 monoclonal ers Semin. Immunopathol. 10:119. Ab: implications for studies on antigen-presenting cell function and immunother- 7. Summers, K. L., J. L. O’Donnell, L. A. Williams, and D. N. J. Hart. 1996. apy. Blood 89:3708. Expression and function of CD80 and CD86 costimulator molecules on synovial 28. Valladeau, J., V. Duvert-Frances, J. J. Pin, C. Dezutter-Dambuyant, C. Vincent, dendritic cells in chronic arthritis. Arthritis Rheum. 39:1287. C. Massacrier, J. Vincent, K. Yoneda, J. Banchereau, C. Caux, et al. 1999. The 8. Summers, K. L., P. B. Daniel, J. L. O’Donnell, and D. N. J. Hart. 1995. Dendritic monoclonal Ab DCGM4 recognizes langerin, a protein specific of Langerhans cells in synovial fluid of chronic inflammatory arthritis lack CD80 surface ex- cells, and is rapidly internalized from the cell surface. Eur. J. Immunol. 29:2695. pression. Clin. Exp. Immunol. 100:81. 29. Bertin, N., P. Clezardin, R. Kubiak, and L. Frappart. 1997. Thrombospondin-1 and -2 messenger RNA in normal, benign, and neoplastic breast tissues: corre- 9. Tak, P. P., T. J. M. Smeets, M. R. Daha, P. M. Kluin, K. A. E. Meijers, R. Brand, Downloaded from A. E. Meinders, and F. C. Breedveld. 1997. Analysis of the synovial cell infiltrate lation with prognostic factors, tumor angiogenesis and fibroblastic desmoplasia. in early rheumatoid synovial tissue in relation to local disease activity. Arthritis Cancer Res. 57:396. Rheum. 40:217. 30. Chabaud, M., G. Page, and P. Miossec. 2001. Enhancing effect of IL-1, IL17, and 10. Iglesias, A., H. Deulofeut, D. P. Dubey, M. Salazar, E. Egea, E. J. Yunis, and TNF-␣ on MIP-3␣ production in rheumatoid arthritis: regulation by soluble re- P. A. Fraser. 1992. Functional deficiency of antigen-presenting cells in the sy- ceptors and Th2 cytokines. J. Immunol. 167:6015. novial fluid of rheumatoid arthritis. Hum. Immunol. 35:109. 31. Thomas, R., and P. E. Lipsky. 1996. Presentation of self peptides by dendritic 11. Pettit, A. R., K. P. A. MacDonald, B. O’Sullivan, and R. Thomas. 2000. Differ- cells: possible implications for the pathogenesis of rheumatoid arthritis. Arthritis entiated dendritic cells expressing nuclear RelB are predominantly located in Rheum. 39:183. rheumatoid synovial tissue perivascular mononuclear cell aggregates. Arthritis 32. Kellermann, S. A., S. Hudak, E. R. Oldham, Y. J. Liu, and L. M. McEvoy. 1999. http://www.jimmunol.org/ Rheum. 43:791. The CC -7 ligands CCL21 and macrophage inflammatory ␤ 12. Kratz, A., A. Campos-Neto, M. S. Hanson, and N. H. Ruddle. 1996. Chronic protein-3 are potent chemoattractants for in vitro- and in vivo-derived dendritic inflammation caused by in lymphoid neogenesis. J. Exp. Med. 183: cells. J. Immunol. 162:3859. 1461. 33. Homey, B., M. C. Dieu-Nosjean, A. Wiesenborn, C. Massacrier, J. J. Pin, E. Oldham, D. Catron, M. E. Buchanan, A. Muller, R. deWaal Malefyt, et al. 13. Caux, C., C. Massacrier, B. Vanbervliet, B. Dubois, I. Durand, M. Cella, ϩ 2000. Up-regulation of macrophage inflammatory protein-3␣/CCL20 and CC A. Lanzavecchia, and J. Banchereau. 1997. CD34 hematopoietic progenitors chemokine receptor 6 in psoriasis. J. Immunol. 164:6621. from human cord blood differentiate along two independent dendritic cell path- 34. Caux, C., B. Vanbervliet, C. Massacrier, C. Dezutter-Dambuyant, ways in response to granulocyte-macrophage colony-stimulating factor plus tu- B. de Saint-Vis, C. Jacquet, K. Yoneda, S. Imamura, D. Schmitt, and mor necrosis factor ␣. II. Functional analysis. Blood 90:1458. ϩ ϩ J. Banchereau. 1996. CD34 hematopoietic progenitors from human cord blood 14. Zhou, L. J., and T. F. Tedder. 1996. CD14 blood monocytes can differentiate ϩ differentiate along two independent dendritic cell pathways in response to GM-

into functionally mature CD83 dendritic cells. Proc. Natl. Acad. Sci. USA 93: by guest on October 1, 2021 CSFϩTNF␣. J. Exp. Med. 184:695. 2588. 35. Caux, C., C. Massacrier, B. Dubois, J. Valladeau, C. Dezutter-Dambuyant, 15. De Saint-Vis, B., J. Vincent, S. Vandenabeele, B. Vanbervliet, J. J. Pin, I. Durand, D. Schmitt, and S. Saeland. 1999. Respective involvement of TGF-␤ S. Ait-Yahia, S. Patel, M. G. Mattei, J. Banchereau, S. Zurawski, et al. 1999. A and IL-4 in the development of Langerhans cells and non-Langerhans dendritic novel lysosome-associated membrane glycoprotein, DC-LAMP, induced upon cells from CD34ϩ progenitors. J. Leukocyte Biol. 66:781. DC maturation, is transiently expressed in MHC class II compartment. Immunity 36. Miossec, P., M. Naviliat, A. Dupuy d’Angeac, J. Sany, and J. Banchereau. 1990. 9:325. Low levels of 4 and high levels of transforming ␤ in 16. Bell, D., P. Chomarat, D. Broyles, G. Netto, G. M. Harb, S. Lebecque, rheumatoid synovitis. Arthritis Rheum. 33:1180. J. Valladeau, J. Davoust, K. A. Palucka, and J. Banchereau.1999. In breast car- 37. Chabaud, M., J. M. Durand, N. Buchs, F. Fossiez, G. Page, L. Frappart, and cinoma tissue, immature dendritic cells reside within the tumor, whereas mature P. Miossec. 1999. Human interleukin-17: a T cell-derived proinflammatory cy- dendritic cells are located in peritumoral areas. J. Exp. Med. 190:1417. tokine produced by the rheumatoid synovium. Arthritis Rheum. 42:963. 17. Zlotnik, A., and O. Yoshie. 2000. Chemokines and immunity. Immunity 12:121. 38. Mottonen, M., P. Isomaki, R. Saario, P. Toivanen, J. Punnonen, and O. Lassila. 18. Liao, F., R. L. Rabin, C. S. Smith, G. Sharma, T. B. Nutman, and J. M. Farber. 1998. Interleukin-10 inhibits the capacity of synovial macrophages to function as 1999. CC-chemokine receptor 6 is expressed on diverse memory subsets of T antigen-presenting cells. Br. J. Rheumatol. 37:1207. cells and determines responsiveness to macrophage inflammatory protein 3␣. 39. MacDonald, K. P., A. R. Pettit, C. Quinn, G. J. Thomas, and R. Thomas. 1999. J. Immunol. 162:186. Resistance of rheumatoid synovial dendritic cells to the immunosuppressive ef- 19. Dieu, M. C., B. Vanbervliet, A. Vicari, J. M. Bridon, E. Oldham, S. Aı¨t-Yahia, fect of IL-10. J. Immunol. 163:5599. F. Brie`re, A. Zlotnik, S. Lebecque, and C. Caux. 1998. Selective recruitment of 40. Shi, K., K. Hayashida, M. Kaneko, J. Hashimoto, T. Tomita, P. E. Lipsky, immature and mature dendritic cells by distinct chemokines expressed in different H. Yoshikawa, and T. Ochi. 2001. Lymphoid chemokine B cell-attracting che- anatomic sites. J. Exp. Med. 188:373. mokine-1 (CXCL13) is expressed in germinal center of ectopic lymphoid follicles 20. Dieu-Nosjean, M. C., C. Massacrier, B. Homey, B. Vanbervliet, J. J. Pin, within the synovium of chronic arthritis patients. J. Immunol. 166:650. A. Vicari, S. Lebecque, C. Dezutter-Dambuyant, D. Schmitt, A. Zlotnik, et al. 41. Fushimi, T., A. Kojima, M. A. S. Moore, and R. G. Crystal. 2000. Macrophage 2000. Macrophage inflammatory protein 3␣ is expressed at inflamed epithelial inflammatory protein 3␣ transgene attracts dendritic cells to established murine surfaces and is the most potent chemokine known in attracting Langerhans cell tumors and suppresses tumor growth. J. Clin. Invest. 105:1383. precursors. J. Exp. Med. 192:705. 42. Braun, S. E., K. Chen, R. G. Foster, C. H. Kim, R. Hromas, M. H. Kaplan, 21. Yanagihara, S., E. Komura, E. Nagafune, H. Watarai, and Y. Yamaguchi. 1998. H. E. Broxmeyer, and K. Cornetta. 2000. The CC chemokine CK␤/MIP-3␤/ELC/ EBI1/CCR7 is a new member of dendritic cell chemokine receptor that is up- Exodus 3 mediates tumor rejection of murine breast cancer cells through NK regulated upon maturation. J. Immunol. 161:3096. cells. J. Immunol. 164:4025.