FoxP3+ T Cells Undergo Conventional First Switch to Lymphoid Tissue Homing Receptors in but Accelerated Second Switch to Nonlymphoid Tissue Homing This information is current as Receptors in Secondary Lymphoid Tissues of September 25, 2021. Jee H. Lee, Seung G. Kang and Chang H. Kim J Immunol 2007; 178:301-311; ; doi: 10.4049/jimmunol.178.1.301 http://www.jimmunol.org/content/178/1/301 Downloaded from

References This article cites 58 articles, 37 of which you can access for free at:

http://www.jimmunol.org/content/178/1/301.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 September 25, 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 © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

FoxP3؉ T Cells Undergo Conventional First Switch to Lymphoid Tissue Homing Receptors in Thymus but Accelerated Second Switch to Nonlymphoid Tissue Homing Receptors in Secondary Lymphoid Tissues1

Jee H. Lee, Seung G. Kang, and Chang H. Kim2

Forkhead box P3 (FoxP3)-positive T cells are a specialized subset for immune regulation and tolerance. We investigated the trafficking receptor switches of FoxP3؉ T cells in thymus and secondary lymphoid tissues and the functional consequences of these switches in migration. We found that FoxP3؉ T cells undergo two discrete developmental switches in trafficking receptors to migrate from primary to secondary and then to nonlymphoid tissues in a manner similar to conventional CD4؉ T cells as well as ؉ ؉

unique to the FoxP3 cell lineage. In the thymus, precursors of FoxP3 cells undergo the first trafficking receptor switch (CCR8/ Downloaded from -CCR93CXCR43CCR7), generating mostly homogeneous CD62L؉CCR7؉CXCR4lowFoxP3؉ T cells. CXCR4 expression is re gained in FoxP3؉ thymic emigrants in the periphery. Consistent with this switch, recent FoxP3؉ thymic emigrants migrate exclusively to secondary lymphoid tissues but poorly to nonlymphoid tissues. The FoxP3؉ thymic emigrants undergo the second switch in trafficking receptors for migration to nonlymphoid tissues upon Ag priming. This second switch involves down-regulation of CCR7 and CXCR4 but up-regulation of a number of memory/effector type homing receptors, resulting in generation of ؉ heterogeneous FoxP3 T cell subsets expressing various combinations of trafficking receptors including CCR2, CCR4, CCR6, http://www.jimmunol.org/ CCR8, and CCR9. A notable difference between the FoxP3؉ and FoxP3؊ T cell populations is that FoxP3؉ T cells undergo the second homing receptor switch at a highly accelerated rate compared with FoxP3؊ T cells, generating FoxP3؉ T cells with unconventionally efficient migratory capacity to major nonlymphoid tissues. The Journal of Immunology, 2007, 178: 301–311.

he forkhead box P3 (FoxP3)3-positive T cell plays an im- by Tregs is likely to be mediated through their migration to target portant role in the regulation of immune tolerance (1, 2). tissue sites and suppression of various immune effector cells. Al- FoxP3, the best marker and master transcription factor for though many CD4ϩCD25ϩ Tregs are FoxP3ϩ cells, some FoxP3ϩ Tϩ ϩ Ϫ CD4 CD25 regulatory T cells (Tregs), is a winged-helix family T cells in mice and humans are CD25 cells (11, 16). transcription factor that binds and suppresses NFAT and NF-␬B In a manner similar to conventional T cells, it is thought that by guest on September 25, 2021 (3–6). Defective FoxP3 expression or function leads to Xid syn- CD4ϩCD25ϩ Tregs in mice or humans express various trafficking ϩ ϩ drome IPEX (immune dysregulation, polyendocrinopathy, enter- receptors for migration. It has been reported that CD4 CD25 or ϩ opathy, X-linked syndrome) characterized by an autoimmune or FoxP3 Tregs express CCR4, CCR5, CCR7, CCR8, CCR10, inflammatory disease in many organs (7, 8). CD4ϩ FoxP3ϩ T cells CXCR4, and CXCR5 and L-, E-selectin, and P-selectin ϩ are continuously produced in the thymus as functionally competent ligands (17–30). CD103 Tregs express effector/memory-type Ϫ Tregs (9–11). Transgenic coexpression of Ags and TCR specific trafficking receptors, and CD103 Tregs express lymphoid tissue for the Ags in mice can lead to enhanced generation of Tregs, homing receptors (20). Some of these receptors are associated with underlying the importance of self-Ags in generation of FoxP3ϩ T migration of Tregs to specific tissue sites such as bone marrow cells (12–14). Additionally, FoxP3ϩ Tregs can be induced in the (CXCR4) (27, 28), liver (CCR10) (24), lymphoid tissues (CD62L, periphery upon Ag priming (15). Regulation of immune responses CCR7, and CXCR5) (19, 29), and tumors (CCR4) (25, 26, 31). Although both lymphoid and nonlymphoid tissue homing Tregs have been identified, it has been unclear how FoxP3ϩ Treg subsets acquire heterogeneous migratory capacity to target tissue sites dur- Department of Comparative Pathobiology, Laboratory of Immunology and Hemato- poiesis, Purdue Cancer Center, and Bindley Bioscience Center, Purdue University, ing their development in thymus and secondary lymphoid tissues. West Lafayette, IN 47907 Using FoxP3 as the main cell lineage marker and polyclonal and ϩ Received for publication July 6, 2006. Accepted for publication October 16, 2006. Ag-specific mouse FoxP3 T cell systems, we undertook a series ϩ The costs of publication of this article were defrayed in part by the payment of page of experiments to determine the modes of FoxP3 cell trafficking charges. This article must therefore be hereby marked advertisement in accordance receptor switches in thymus and secondary lymphoid tissues and with 18 U.S.C. Section 1734 solely to indicate this fact. the impact of such trafficking receptor switches on their organ to 1 This study was supported in part by Grant AI063064 from National Institutes of organ migration behavior. Health-National Institute of Allergy and Infectious Diseases, by the American Heart Association, and by the Sidney Kimmel Foundation (to C.H.K). 2 Address correspondence and reprint requests to Dr. Chang Kim, Department of Materials and Methods Comparative Pathobiology, Purdue University, 725 Harrison Street, West Lafayette, Mice IN 47907. E-mail address: [email protected] BALB/c mice were purchased from Harlan Sprague Dawley. Rat insulin 3 Abbreviations used in this paper: FoxP3, forkhead box P3; Treg, regulatory T cell; PLN, peripheral lymph node; MLN, mesenteric lymph node; PP, Peyer’s patch; promoter membrane-bound form OVA (RIP-mOVA; line 296-1B) mice mOVA, membrane-bound OVA; DP, double positive; SP, single positive; DN, double were from Dr. A. Abbas (University of California, San Francisco, CA) with negative. permission from Dr. W. Heath (Walter and Eliza Hall Institute, Melbourne, Australia). DO11.10 mice and RIP-mOVA mice were bred to generate Ϫ Ϫ Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$2.00 mOVA ϫ DO11.10 Treg mice in wild-type or rag2 / background. www.jimmunol.org 302 DEVELOPMENT OF MOUSE FoxP3ϩ T CELL HOMING PROGRAM

FIGURE 1. Tissue distribution of heterogeneous FoxP3ϩ T cell subsets in primary lymphoid, secondary lymphoid (2-LT), and nonlymphoid tissues (Non-LT). A, Definitions of FoxP3ϩ and FoxP3Ϫ CD4 T cell sub- sets. B, Absolute numbers of FoxP3ϩ T cells in the 10 selected organs. C, Percentage shows relative distribu- tion of FoxP3ϩ T cells among the 10 selected organs: thymus, blood, spleen, MLN, six PLN (two inguinal, two auxiliary, and two brachial), PP, small intestine, bone marrow (marrow), and liver. D, Percentages of the relative frequencies of CD62Lϩ and CD62LϪFoxP3ϩ T cell subsets in each organ. Peritoneal cavity (P. cavity) and the lamina propria of small intestine (S. intestine) are also examined. A representative set of data (A) and combined data (B–D) obtained from seven different ex- periments using 6- to 8-wk-old BALB/c mice are Downloaded from shown. http://www.jimmunol.org/

DO11.10 TCR rag2Ϫ/Ϫ transgenic mice were purchased from Taconic Cells and cell culture Farms. Mice were housed at Purdue University (West Lafayette, IN) under a specific pathogen-free condition and used according to approved proto- Single-cell suspensions from lymph nodes (mesenteric, inguinal, auxiliary, cols and institutional guidelines. PCR genotyping was performed to iden- and brachial), Peyer’s patches (PP), spleen, and liver were prepared by tify mOVA ϫ DO11.10 Treg mice that contain both mOVA and DO11.10 grinding tissues through an iron mesh. Peritoneal lavage was prepared in transgenic . Generally, mice used were between age 6 and 8 wk at the ϳ10 ml of PBS. Bone marrow cells were flushed from femurs and tibias start of each experiment. with ϳ10 ml of PBS per tibia or femur. Lamina propria lymphocytes of by guest on September 25, 2021

FIGURE 2. Thymus generates CCR7ϩCXCR4low FoxP3ϩ T cells. A, DN, DP, and SP FoxP3ϩ or FoxP3Ϫ subsets. B, Expression of four major thymo- cyte receptors by DN (FoxP3Ϫ), DP (FoxP3Ϫ and FoxP3ϩ), and SP (FoxP3Ϫ and FoxP3ϩ) subsets and recent FoxP3ϩ thymic emigrants in PLN and MLN was examined. FoxP3ϩ thymic emigrants in PLN and MLN were identified as described for Fig. 3. Combined data from four independent experiments are p Յ 0.05 significant difference between ,ء .shown FoxP3ϩ and FoxP3Ϫ CD4ϩ cells. The Journal of Immunology 303

small intestine were isolated by removing the epithelial cells with 5 mM EDTA (five times) and digesting the remaining tissues for 45 min with 300 U/ml collagenase (type 3; Worthington Biochemical) and 100 ␮g/ml DNase I (Worthington Biochemical) at 37°C. For removal of RBC from spleen, liver, and bone marrow cells, cells were treated with hypotonic RBC lysis buffer. For one experiment (see Fig. 6, B and C), T cells were isolated from mesenteric lymph nodes (MLN) and labeled with CFSE as previously described (32). CFSE-labeled MLN cells were cultured in the ␮ presence of IL-2 and Ag peptide OVA323–339 (final 1 g/ml) for 5 days before examination of cell division and trafficking of receptor expression by FACS. Expression of trafficking receptors Fresh or Ag-primed mouse T cells, harvested from BALB/c, mOVA ϫ DO11.10, or mOVA ϫ DO11.10 (rag2Ϫ/Ϫ) mice were stained with Abs to CCR3 (clone 83103; R&D Systems), CCR5 (clone C34-3448; BD Pharmin- gen), CCR6 (clone 140706; R&D Systems), CCR7 (clone 4B12; BioLegend), CCR9 (clone 242503; R&D Systems), CXCR2 (clone 242216; R&D Sys- tems), CXCR3 (clone 220803; R&D Systems), CXCR4 (clone 2B11; BD Phamingen), CXCR5 (clone 2G8; BD Pharmingen), CXCR6 (clone 221002; R&D Systems), or rat control IgG2b (clone A95-1; BD Pharmingen) for 30 min on ice. Fc fusion of CCL1, CCL17, and CCL27 were used for staining of CCR8, CCR4, and CCR10, respectively. To generate Fc chemo- Downloaded from kine proteins, pcDNA3.1 expression vectors containing Fc proteins of CCL27 (33), CCL1 and CCL17 under the CMV early promoter were used to produce Fc chemokine constructs in HEK-293 cells. The specificity of the staining with FIGURE 3. Tissue destinations for recent FoxP3ϩ thymic emigrants in the constructs was verified by using transfectants. Mouse the periphery. A, FITC was injected into the thymus to label FoxP3ϩ T cells were stained with biotinylated anti-rat IgG2a (clone RG7/1.30; BD cells. Three days later, FITCϩ FoxP3ϩ T cells that had migrated to various Pharmingen), anti-IgG2b (clone RG7/11.1; Caltag Laboratories), anti-rat organs were counted by flow cytometry analysis. FITCϩ FoxP3ϩ cell fre- IgG2c (clone MRG2c-67; BioLegend), or anti-human IgG Ab (Vector Labo- quencies within the FoxP3ϩ cell population in each organ, shown as the http://www.jimmunol.org/ ratories) for 30 min, followed by staining with allophycocyanin-streptavidin percentage after normalization for spleen, are shown. B, Relative distribu- (BD Pharmingen) and Abs to CD62L, DO11.10 TCR, and CD4. The cells tion of thymic emigrants among the selected secondary lymphoid and non- were further stained with FoxP3 (eBioscience) according to the manufacturer’s ϩ ϩ protocol. lymphoid organs. The absolute number of FITC FoxP3 T cells that ϩ migrated from the thymus to each organ are calculated, and the percentage Labeling of thymic FoxP3 T cells and emigration study of distribution among the selected secondary and nonlymphoid organs was BALB/c mice were anesthetized by i.p. injection of 90–120 mg/kg keta- calculated. Combined data from four independent experiments are shown. mine and 10 mg/kg xylazine. A ventral midline incision was made one- third down the sternum to expose the thymus. A Hamilton syringe was used ␮ ϩ to inject 5 l of a 1 mg/ml solution of FITC into each lobe. The skin For assessment of FoxP3 T cell numbers in bone marrow, we used the by guest on September 25, 2021 incision was closed with surgical clips, and mice were allowed to recover number of total bone marrow mononuclear cells (2.8 ϫ 108) determined by ϳ under a heat source. We routinely achieved 70% labeling of others (34, 35). using this method. Three days later, mice were sacrificed and various tis- sues were harvested. Single-cell suspensions from various tissues were Statistical analyses examined by flow cytometry to determine the number of FITCϩFoxP3ϩ T cells. In some experiments, the trafficking receptor expression of Averages with SEM are shown in most of the experiments. Student’s FITCϩFoxP3ϩ T cells in peripheral organs was examined as already paired and unpaired t tests were used to determine the significance of the described. differences between two groups of data. Values of p Յ 0.05 were consid- ered significant. Adoptive transfer, immunization, and homing experiment For short-term (20-h) homing experiments, splenocytes containing ϳ5 ϫ Results 106 KJ1-26ϩFoxP3ϩ T cells from mOVA ϫ DO11.10 Treg mice were FoxP3ϩ T cell subsets and their tissue tropisms in primary, prepared and injected into BALB/c mice via a tail vein. Host mice were secondary, and nonlymphoid tissues immunized with 250 ␮g of OVA with 50 ␮l of complete or IFA per site ϩ (CFA or IFA; Sigma-Aldrich). The mice were immunized s.c. at the flank As the first step, we examined the tissue distribution of FoxP3 T to examine only the T cells in peripheral lymph nodes (PLN), or s.c. and cells in young (6- to 8-wk-old) BALB/c mice using FoxP3 as a cell i.p. at the same time to examine T cells in PLN, MLN, and spleen. Com- lineage marker (Fig. 1, A and B). The majority (on average ϳ65%) parable results were obtained for T cells in PLN using the two immuniza- of the FoxP3ϩ T cells were found in MLN, PLN, and spleen tion methods. For controls, mice were immunized with CFA or IFA only ϩ without OVA. Three or 8 days later, mice were sacrificed to determine the among the 10 organs examined. Other FoxP3 T cells were presence of KJ1-26ϩFoxP3ϩ T cells in various tissue sites. For short-term present in nonlymphoid tissues, thymus, and bone marrow. When homing experiments, Ag-primed KJ1-26ϩFoxP3ϩ T cells were prepared ϩ ϩ ϩ ϩ the frequency of FoxP3 T cells among CD4 T cells was con- by transferring KJ1-26 FoxP3 T cells from mOVA ϫ DO11.10 mice ϩ sidered (Fig. 1C), the most FoxP3 T cell-enriched organ was into BALB/c mice followed by immunization with OVA as described. ϩ Eight days later, 5 ϫ 107 splenocytes (containing ϳ5 ϫ 106 KJ1- marrow (ϳ40% of CD4 T cells) followed by peritoneal cavity ϩ 26ϩFoxP3ϩ T cells) isolated from the immunized BALB/c mice were in- (ϳ16%). On average, ϳ8% of CD4 T cells in the small intestine jected into BALB/c mice. At 20 h later, mice were sacrificed to determine ϩ ϩ ϩ Ϫ were FoxP3 T cells. homing of KJ1-26 FoxP3 and FoxP3 T cell subsets to selected organs. ϩ ϩ ϩ Tregs in the spleen and lymph nodes are divided into CD62L KJ1-26 FoxP3 T cells in various organs were counted by flow cytom- and CD62LϪ cells, which have different phenotype and function etry. Because the absolute number of migrated T cells varies from exper- ϩ ϩ Ϫ iment to experiment due to differences in the number of transferred cells, (19, 36–38). The majority of FoxP3 CD4 CD8 T cells in the we normalized the absolute number of KJ1-26ϩFoxP3ϩ T cells present in thymus were CD62LϩCD44high cells, whereas FoxP3ϪCD4ϩ each organ to the number of spleen (spleen ϭ 1) T cells. This system CD8Ϫ T cells were CD62LϩCD44low cells (data not shown). In generates a parameter termed “homing index,” which indicates the relative ϩ ϩ ϩ MLN, PLN, and spleen, many FoxP3 T cells were CD62L distribution of injected FoxP3 T cells among the organs examined. high Ϫ high Thymus, blood, spleen, MLN, six PLN (two inguinal, two auxiliary, and CD44 , but some were CD62L CD44 cells. In the periph- Ϫ ϩ two brachial), PP, small intestine, bone marrow, and liver were examined. eral blood, the proportion of CD62L FoxP3 cells was increased 304 DEVELOPMENT OF MOUSE FoxP3ϩ T CELL HOMING PROGRAM

FIGURE 4. Ag priming redistributes FoxP3ϩ T cells from secondary lymphoid to nonlymphoid tissues. A, Characteriza- tion of spleen FoxP3ϩ T cells (isolated from mOVA ϫ DO11.10 mice and used as input cells) in expression of FoxP3, CD62L, and/or CD25. CD4ϩ cells from Downloaded from BALB/c and DO11.10 rag2Ϫ/Ϫ mice are shown for comparison. B, Dynamics of FoxP3ϩ T cell subset distribution among the six organs after Ag priming. BALB/c mice were injected i.v. with splenocytes containing mOVA ϫ DO11.10 FoxP3ϩ http://www.jimmunol.org/ CD4ϩ T cells (day 0) followed by immu- nization s.c. at the flank and i.p. with OVA. Relative distribution of FoxP3ϩ T cells in multiple tissues was determined based on absolute numbers of KJ1-26ϩ FoxP3ϩ cells in each organ, which was normalized by that of KJ1-26ϩFoxP3ϩ T cells in spleen (spleen ϭ 1). The average absolute number of KJ1Ϫ26ϩFoxP3ϩ Ϫ and FoxP3 T cells in immunized spleens by guest on September 25, 2021 at day 3 was 123 ϫ 103 (CD62Lϩ FoxP3ϩ), 42.9 ϫ 103 (CD62LϪFoxP3ϩ), 155 ϫ 103 (CD62LϩFoxP3Ϫ), and 106 ϫ 103 (CD62LϪFoxP3Ϫ). The average ab- solute number in immunized spleens at day 8 was 46.8 ϫ 103 (CD62LϩFoxP3ϩ), 3.7 ϫ 103 (CD62LϪFoxP3ϩ), 37.8 ϫ 103 (CD62LϩFoxP3Ϫ), and 11.8 ϫ 103 (CD62LϪFoxP3Ϫ). Combined data from four independent experi- -p Յ 0.05. C, Con ,ء .ments are shown version of OVA-specific FoxP3Ϫ CD4ϩ T cells to FoxP3ϩ T cells is min- imal in the given immunization condition. BALB/c mice were injected i.v. with splenocytes containing DO11.10 rag2Ϫ/Ϫ FoxP3Ϫ CD4ϩ or mOVA ϫ DO11.10 FoxP3ϩCD4ϩ T cells followed by immu- nization with OVA. The Journal of Immunology 305

FIGURE 5. CD62LϩFoxP3ϩ T cells mainly express CCR7 and CXCR4, whereas CD62LϪFoxP3ϩ T cells characteristically express memory/effector type traffick- ing receptors in MLN (A) and PLN (B) of unimmunized BALB/c mice. Expression of chemokine receptors ␣ ␤ (CCR1-CCR10 and CXCR2-CXCR6) and 4 7 by FoxP3ϩ T cell subsets in MLN and PLN were exam- ined. Combined data from four independent experi- Downloaded from ments are shown. Significant differences between CD62Lϩ and CD62LϪ cells. *, Significant differences between CD62Lϩ and CD62LϪ cells (p Ͻ 0.05); **, significant differences between FoxP3ϩ and FoxP3Ϫ cells (p Ͻ 0.05). http://www.jimmunol.org/ by guest on September 25, 2021

to ϳ50% within the FoxP3ϩ T cell population (Fig. 1D). In bone emigration to lymph nodes, CXCR4 was again up-regulated. Con- marrow and nonlymphoid tissue sites, such as liver, peritoneal cav- versely, CCR9 (the CCL25 receptor) was high on DN FoxP3Ϫ ity, and lamina propria of the small intestine, the proportion was cells and DP FoxP3ϩ cells but was sharply down-regulated after- increased even more, up to ϳ95%. In PP (the secondary lymphoid ward. CCR8 (the CCL1 receptor) was high on DN FoxP3Ϫ cells ϩ tissues embedded within the intestine), ϳ70% of FoxP3 cells but was down-regulated afterward. In contrast, CCR7 (the CCL19 Ϫ were CD62L cells. There is a clear trend that the proportion of receptor, a secondary lymphoid tissue homing receptor) was grad- Ϫ ϩ CD62L FoxP3 cells is increased in nonlymphoid tissues com- ually up-regulated at the transition from DP FoxP3ϩ cells to SP pared with secondary lymphoid tissue and thymus, suggesting ϩ ϩ Ϫ ϩ FoxP3 cells. Notable differences between FoxP3 and FoxP3 that the FoxP3 cell subsets have different tissue tropisms from cells include expression of CXCR4 and CCR9: CXCR4 was ex- each other. pressed on many more FoxP3ϩ cells, whereas CCR9 was ex- Ϫ ϩ pressed on many more FoxP3 T cells at the CD4 SP stage. Over- FoxP3 T cells undergo a trafficking receptor switch in the ϩ ϩ thymus to generate CD62LϩCCR7ϩCXCR4lowFoxP3ϩ T cells, all, thymus generates mostly homogeneous CD62L CCR7 low ϩ and these FoxP3ϩ thymic emigrants mainly migrate to CXCR4 FoxP3 T cells as the consequence of the switch. ϩ ϩ ϩ secondary lymphoid tissue Apparently, the CD62L CCR7 CXCR4lowFoxP3 T cells, generated in the thymus, have the homing receptors associated Next, we examined the trafficking receptor switch in thymus, the ϩ with secondary lymphoid tissues. To determine their homing be- primary lymphoid organ for T cells. FoxP3 cells appear at dou- ble-positive (DP) and single-positive (SP) stages in the thymus havior to secondary lymphoid tissues as well as their emigration (11). Consistently, the majority of thymic FoxP3ϩ cells were CD4 behavior to other tissue sites, we used a FITC thymic injection method to label thymic FoxP3ϩ cells (Fig. 3A). Three days after SP, and some were CD4 and CD8 DP (Fig. 2A). We determined ϩ ϩ the trafficking receptors expressed by FoxP3ϩ cells at different the injection with FITC, tissue distribution of FITC FoxP3 cells ϩ ϩ development stages. Thymic , such as CXCL12, in the periphery was determined. Most FITC FoxP3 cells were CCL1, CCL25, and CCL19, are important for T cell localization, found in secondary lymphoid tissues such as spleen and lymph development, and emigration. CXCR4 (the CXCL12 receptor) was nodes among the eight organs examined (Fig. 3B), suggesting that ϩ ϩ ϩ up-regulated at the transition from the double-negative (DN) the thymus-emigrating CD62L CCR7 CXCR4low FoxP3 T FoxP3Ϫ cells to DP FoxP3ϩ cells but was down-regulated as cells preferentially migrate to secondary lymphoid tissues over DP FoxP3ϩ cells become CD4 SP FoxP3ϩ cells (Fig. 2B). After nonlymphoid tissues. 306 DEVELOPMENT OF MOUSE FoxP3ϩ T CELL HOMING PROGRAM Downloaded from http://www.jimmunol.org/

FIGURE 6. Ag priming in secondary lymphoid tissue induces the second chemokine receptor switch in FoxP3ϩ T cells. A, Splenocytes containing by guest on September 25, 2021 OVA-specific KJ1-26ϩFoxP3ϩ T cells, obtained from mOVA ϫ DO11.10 mice, were injected i.v. into BALB/c mice. Chemokine receptor expression by KJ1-26ϩFoxP3ϩ T cells in PLN before and after immunization s.c. at the flank (or s.c. and i.p. at the same time) with OVA plus CFA (or CFA alone) was compared. Multiple control Abs (CTR Ab) with different isotypes were used to determine background staining levels. Combined data from four independent .p Յ 0.05 for significant difference between FoxP3ϩ and FoxP3Ϫ cells. B and C, The second trafficking receptor switch in vitro ,ء .experiments are shown ϩ ϫ MLN KJ1-26 T cells of mOVA DO11.10 mice were labeled with CFSE and cultured with OVA323–339 in the presence of irradiated splenocytes as APCs. On day 5, FoxP3ϩ or FoxP3ϪKJ1-26ϩCD4ϩ T cells at different numbers of cell division were examined for their expression of chemokine receptors. Combined data from three experiments are shown.

Ag priming in secondary lymphoid tissue generates in a manner similar to the recent FoxP3ϩ thymic emigrants. After CD62LϪFoxP3ϩ T cells with a tissue tropism for immunization of the mice with OVA, CD62LϪKJ1-26ϩFoxP3ϩ T nonlymphoid tissue cells were generated and expanded in secondary lymphoid tissues ϩ ϩ Ag priming is a key event that expands the number of T cells and (data not shown). At day 3 after immunization, KJ1-26 FoxP3 (or FoxP3Ϫ) T cells were still found mainly in secondary lymphoid changes their function and tissue tropism. Next, we investigated ϩ ϩ the changes in FoxP3ϩ T cell tissue distribution following Ag tissues (Fig. 4B). By day 8, however, many KJ1-26 FoxP3 T priming in secondary lymphoid tissues. For this experiment, we cells were found in nonlymphoid tissues such as bone marrow and used OVA-specific (KJ1-26ϩ) FoxP3ϩ T cells obtained from lamina propria of the small intestine (Fig. 4B). It was notable that Ϫ ϩ ϩ ϩ mOVA ϫ DO11.10 double transgenic mice. Although the surface it was CD62L FoxP3 T cells, but not CD62L FoxP3 T cells, and functional phenotypes of these Tregs are indistinguishable that were increased in the nonlymphoid tissues (Fig. 4B). The ap- ϩ Ϫ ϩ from those of wild-type mice, the lymphoid tissues of these mice pearance rate for KJ1-26 CD62L FoxP3 T cells in nonlym- Ϫ are highly enriched with OVA-specific FoxP3ϩ cells as the result phoid tissue was significantly higher than the rate for their FoxP3 ϩ of transgenic coexpression of OVA-specific TCR and OVA (39). counterparts. In contrast, the FoxP3 T cells in control mice, in- A 35–50% of KJ1-26ϩ CD4ϩ T cells from secondary lymphoid jected with adjuvant only without the Ag, were gradually de- tissues (spleen) of these mice were FoxP3ϩ T cells, and ϳ85% of creased in numbers in secondary lymphoid tissue at days 3 and 8 these FoxP3ϩ T cells were CD25ϩ (Fig. 4A). Most (Ͼ95%) of the and did not appear in the nonlymphoid tissue at significant num- FoxP3ϩ T cells in spleen and lymph nodes were CD62Lϩ. When bers, suggesting that the redistribution (potentially the result of ϩ splenocytes of mOVA ϫ DO11.10 mice were injected i.v. into migration and/or expansion) of FoxP3 T cells to nonlymphoid BALB/c mice, most KJ1-26ϩ FoxP3ϩ T cells migrated back to tissue is an Ag-specific response. secondary lymphoid tissues such as spleen, PLN, and MLN, but It has been reported that FoxP3Ϫ T cells can convert to FoxP3ϩ not to nonlymphoid tissues among the organs examined (Fig. 4B) T cells following Ag priming. Therefore, it is possible that the The Journal of Immunology 307

FIGURE 7. The FoxP3ϩ T cells that populate non- lymphoid tissues following immunization show the traf- ficking receptor phenotype after the second switch. Splenocytes containing OVA-specific KJ1-26ϩFoxP3ϩ T cells, obtained from mOVA ϫ DO11.10 mice, were adoptively transferred into BALB/c mice. The traffick- ing receptor phenotype of KJ1-26ϩFoxP3ϩ T cells in secondary lymphoid tissue (2° LT) and nonlymphoid tissue (Non-LT) at day 8 postimmunization with OVA plus CFA is shown. PLN, MLN, peritoneal cavity (PC), and small intestine (SI) are represented. Representative data from three independent experiments are shown.

FoxP3ϩ cells in the periphery of the immunized mice in Fig. 4B FoxP3ϩ cells, obtained from mOVA ϫ DO11.10 double trans- Ϫ are from FoxP3 T cells. To examine this possibility, we trans- genic mice, into BALB/c mice and examined their up-regulation of Downloaded from ferred DO11.10 rag2Ϫ/Ϫ T cells, which are free of KJ1- chemokine receptors upon antigenic challenge. Before cell trans- 26ϩFoxP3ϩ T cells as shown in Fig. 4A, into BALB/c mice and fer, the injected KJ1-26ϩFoxP3ϩ T cells largely expressed only immunized the host mice with OVA (Fig. 4C). The average con- two receptors (CCR7 and CXCR4) among the 14 chemokine re- version rate of KJ1-26ϩFoxP3Ϫ T cells to FoxP3ϩCD4ϩ T cells ceptors examined (only seven receptors are shown in Fig. 6A). On was ϳ0.5% at day 3 and ϳ1.0% at day 8 in the spleen and lymph day 3 after immunization with OVA, however, many FoxP3ϩ T

nodes of BALB/c mice after immunization. These conversion rates cells found in PLN acquired CCR4, CCR6, CCR8, and CXCR3 at http://www.jimmunol.org/ are almost negligible, compared with the high FoxP3ϩ T cell fre- high frequencies (Fig. 6A). Importantly, FoxP3ϩ T cells gained the quency (40–60% of KJ1-26ϩCD4ϩ T cells) in the mice injected expression of the memory/effector/inflammation-associated che- with mOVA ϫ DO11.10 CD4ϩ cells. These data support that most mokine receptors at much higher frequencies than FoxP3Ϫ T cells. Ag-specific FoxP3ϩ T cells, present in host mice after immuniza- During the 3-day period, 50–60% of FoxP3ϩ and FoxP3Ϫ T cells tion, were probably from injected FoxP3ϩ but not FoxP3Ϫ T cells. lost the expression of CCR7 and CXCR4. The decrease in CCR7

ϩ and increase in memory/effector/inflammation-associated receptor Differential chemokine receptor expression by FoxP3 T cell expression suggest the presence of a chemokine receptor switch in subsets in secondary lymphoid tissues FoxP3ϩ T cells following antigenic stimulation. Compared with

Because trafficking receptors are important for organ to organ mi- by guest on September 25, 2021 gration of T cells, we next examined the trafficking receptor phe- notype of the CD62Lϩ and CD62LϪFoxP3ϩ T cell subsets in MLN and PLN of unimmunized BALB/c mice (Fig. 5). CD62LϩFoxP3ϩ T cells in MLN mainly expressed the two lym- phoid tissue-associated chemokine receptors CCR7 and CXCR4, along with CCR6, CCR8, and CCR9 at low frequencies (10–20%) (Fig. 5A). This pattern was a very different from that of the che- mokine receptors expressed by CD62LϪFoxP3ϩ T cells, many of which expressed CCR2, CCR4, CCR5, CCR6, CCR8, CXCR3, and CXCR6 in addition to CCR7 and CXCR4. In a manner similar to the FoxP3ϩ T cell subsets in MLN, the two FoxP3ϩ T cell subsets in PLNs differentially expressed che- mokine receptors (Fig. 5B). CD62LϩFoxP3ϩ T cells in PLN mainly expressed CCR7 and CXCR4, whereas CD62LϪFoxP3ϩ T cells had the two receptors plus a number of memory/effector/ inflammation-associated chemokine receptors. Reduced expres- sion of CCR7 by some CD62LϪFoxP3ϩ T cells and up-regulation of memory/effector/inflammation-associated chemokine receptors by FoxP3ϩ T cells compared with FoxP3Ϫ cells were again evident.

Ag priming in secondary lymphoid tissue induces the second ϩ switch in trafficking receptors in FoxP3 T cells in an FIGURE 8. The accelerated second switch in FoxP3ϩ T cells, compared Ϫ accelerated manner compared with FoxP3 T cells with FoxP3Ϫ T cells, is not due to relative suppression of FoxP3Ϫ T cells by coinjected FoxP3ϩ T cells. Splenocytes containing KJ1-26ϩ T cells The differences in trafficking receptor expression between Ϫ/Ϫ ϩ Ϫ ϩ were transferred from DO11.10 (rag2 ) mice or mOVA ϫ DO11.10 CD62L and CD62L FoxP3 T cells, shown in Fig. 5, and the Ϫ/Ϫ Ϫ ϩ ϩ (rag2 ) mice into BALB/c mice. Twenty hours later, the host mice were generation of CD62L FoxP3 T cells from CD62L cells upon immunized with OVA. Three days later, FoxP3ϩ and FoxP3ϪKJ1- Ag priming, shown in Fig. 4, suggest the presence of another 26ϩCD4ϩ T cells were examined for their expression of chemokine re- switch in trafficking receptors. To demonstrate the switch, we ceptors. Combined data from four independent experiments are shown. .p Յ 0.05 is significant difference compared with FoxP3Ϫ T cells ,ء adoptively transferred splenocytes containing OVA-specific 308 DEVELOPMENT OF MOUSE FoxP3ϩ T CELL HOMING PROGRAM

FIGURE 9. FoxP3ϩ T cells after the accelerated sec- ond switch are more efficient, than FoxP3Ϫ T cells, in mi- gration to nonlymphoid tissue. Primed KJ1-26ϩFoxP3ϩ T cells were prepared by transferring KJ1-26ϩFoxP3ϩ T cells from mOVA ϫ DO11.10 mice into BALB/c mice followed by immunization with OVA. Eight days later, splenocytes containing KJ1-26ϩFoxP3ϩ and FoxP3Ϫ T cells were harvested from the immunized mice and in- jected i.v. into BALB/c mice for short-term (20-h) homing experiments. A, Homing index indicates the relative dis- tribution of FoxP3ϩ T cells among the six organs exam- ined (normalized for the spleen) as shown in Fig. 4B. Ab- solute number (ϫ103) of the FoxP3ϩ T cells from representative data to show actual number of FoxP3ϩ T cells present in each organ. The y-axis (left) is for FoxP3ϩ p Յ 0.05 ,ء .cells and the y-axis (right) is for FoxP3Ϫ cells represents significant differences. B, Representative dot ϩ/Ϫ ϩ Ϫ plots of CD62L FoxP3 T cells and FoxP3 T cells Downloaded from that migrated to each organ. C, Percentage of frequency of CD62LϪFoxP3ϩ T cells within the CD4ϩKJ1-26ϩ T cell population that migrated to each organ. Representative dots are shown to reflect degree of cell migration. Data from three different experiments were combined to show averages (ϮSEM). http://www.jimmunol.org/

day 3, the numbers of FoxP3ϩ T cells expressing the nonlymphoid Ag-primed and nonlymphoid tissue-residing FoxP3ϩ T cells tissues homing chemokine receptors were decreased at day 8 in have the postsecond switch trafficking receptor phenotype ϩ PLN, MLN, and spleen, whereas the proportions of CXCR4 and ϩ ϩ ϩ To determine the trafficking receptor switch status of the FoxP3 CCR7 FoxP3 cells were increased (data not shown). Control by guest on September 25, 2021 ϩ T cells that appear in nonlymphoid vs secondary lymphoid tissues mice were injected with the adjuvant without OVA, and FoxP3 T after Ag priming, KJ1-26ϩ (OVA-specific), FoxP3ϩ T cells were cells in these mice did not undergo the chemokine receptor switch isolated from selected nonlymphoid tissues (small intestine and that occurred in mice immunized with OVA, suggesting that the ϩ peritoneal cavity) of BALB/c mice, previously injected i.v. with trafficking receptor switch of FoxP3 T cells in secondary lym- mOVA ϫ DO11.10 cells and immunized with OVA, and exam- phoid tissues is an Ag-specific response. Similar switches in traf- ϩ ined for their trafficking receptor profiles (Fig. 7). The KJ1- ficking receptors of the FoxP3 T cells were also observed in 26ϩFoxP3ϩ T cells in the nonlymphoid tissues of mice had very spleen, PP, and MLN (data not shown). It has been reported that different trafficking receptor phenotypes than that of secondary TLR ligands (e.g., mycobacterial products in CFA) may affect the ϩ lymphoid tissue at day 8 postimmunization. Although the majority biology of FoxP3 cells (40, 41). To rule out the potential mod- of KJ1-26ϩFoxP3ϩ T cells in secondary lymphoid tissue ulation effects of CFA on expression of FoxP3 or chemokine re- expressed CCR7, the majority of KJ1-26ϩFoxP3ϩ T cells in ceptors, we also used IFA for immunization. The results with IFA nonlymphoid tissues did not express it. Instead, many of KJ1- (data not shown) were almost identical with those found in mice 26ϩFoxP3ϩ T cells in nonlymphoid tissues expressed tissue- immunized with CFA, and thus, the differences in trafficking re- ϩ Ϫ specific homing receptor such as CCR9 (small intestine) and ceptor switch between FoxP3 and FoxP3 T cells are not attrib- memory/effector type homing receptors such as CCR4, CCR6, utable to the type of adjuvant used. and CXCR3 (peritoneal cavity). We, next, examined whether mouse FoxP3ϩ T cells undergo the same accelerated switch in vitro. For this study, CFSE-labeled ϩ Ϫ The hyperexpression of nonlymphoid tissue homing receptors by MLN cells containing FoxP3 and FoxP3 T cells obtained from ϩ Ϫ mOVA ϫ DO11.10 double transgenic mice were cultured for 5 Ag-primed FoxP3 T cells compared with FoxP3 T cells after the second switch is not due to relative suppression of FoxP3Ϫ days with the Ag peptide OVA323–339. Consistent with the in vivo ϩ data (Fig. 6A), the hyperexpression of nonlymphoid tissue homing T cells by FoxP3 T cells during Ag priming receptors by primed FoxP3ϩ T cells compared with FoxP3Ϫ T It is possible that the apparent hyperexpression of nonlymphoid cells was also detected in vitro (Fig. 6, B and C). The difference tissue homing receptors by secondary lymphoid tissue FoxP3ϩ T between FoxP3ϩ and FoxP3Ϫ T cells in expression of major re- cells (and the relative hypoexpression by FoxP3Ϫ T cells) would ceptors was evident for the cells after various numbers of cell be due to suppression of the up-regulation of the trafficking recep- divisions. CCR7 expression was similarly down-regulated on both tors on FoxP3Ϫ T cells by the FoxP3ϩ T cells cotransferred to- FoxP3ϩ and FoxP3Ϫ cells as they divided. In contrast, CCR4 and gether. To rule out this possibility, BALB/c mice were adoptively ␣ ␤ Ϫ/Ϫ 4 7 were gradually up-regulated with cell division. Other che- injected i.v. with splenocytes of DO11.10 (rag2 ) mice, which mokine receptors such as CCR8, CCR9, and CXCR6 were grad- do not contain KJ1-26ϩ FoxP3ϩ cells) (Fig. 4A) and immunized ually down-regulated. with OVA. In parallel, a separate group of BALB/c mice were The Journal of Immunology 309 injected i.v. with mOVA ϫ DO11.10 (rag2Ϫ/Ϫ) splenocytes con- Discussion ϩ ϩ Ϫ taining both KJ1-26 FoxP3 and FoxP3 subsets and immunized We investigated the development of FoxP3ϩ T cell homing pro- with OVA for comparison. Chemokine receptor switch on the gram in primary and secondary lymphoid tissues, a process that is ϩ Ϫ KJ1-26 FoxP3 T cells, Ag-primed in the presence or absence of important for migration of FoxP3ϩ T cells to target tissue sites, but ϩ ϩ KJ1-26 FoxP3 T cells, was compared (Fig. 8). In the situation in has been incompletely understood so far. Our results show the ϩ Ϫ ϩ which splenocytes containing KJ1-26 FoxP3 CD4 T cells orderly switches of FoxP3ϩ T cell trafficking receptors and mi- ϩ ϩ without KJ1-26 FoxP3 cells were transferred, the KJ1- gration behavior during their development in the thymus and sec- ϩ Ϫ 26 FoxP3 cells up-regulated memory/effector-associated recep- ondary lymphoid tissues. FoxP3ϩ T cells, generated in the thymus, tors (e.g., CCR2, CCR4, CCR6, CCR8, CXCR3, and CXCR5) at undergo the first trafficking receptor switch for emigration to sec- levels similar to (or even lower for some receptors than) the KJ1- ondary lymphoid tissues. FoxP3ϩ thymic emigrants migrate to ϩ Ϫ ϩ ϩ 26 FoxP3 T cells, primed in the presence of KJ1-26 FoxP3 T secondary lymphoid tissues and undergo the second round of traf- cells. Therefore, the accelerated chemokine receptor switch by ficking receptor switch to migrate to nonlymphoid tissues. Inter- ϩ FoxP3 T cells appears not to be due to potential suppression of estingly, the second switch of FoxP3ϩ T cells is highly robust, Ϫ ϩ chemokine receptor expression in FoxP3 T cells by FoxP3 T compared with that of conventional CD4ϩ T cells. Only the cells with the same Ag specificity. When compared with KJ1- FoxP3ϩ T cells that underwent the second switch in trafficking ϩ Ϫ ϩ ϩ 26 FoxP3 CD4 cells, FoxP3 T cells, primed in both MLN receptor expression were able to migrate efficiently to nonlym- and PLN, more efficiently up-regulated CCR2, CCR4, CCR6, and phoid tissues. CCR8, again providing evidence for the presence of the acceler- Ϫ Ϫ ϩ Prothymocytes sequentially undergo CD4 CD8 (DN), CD4 Downloaded from ated second switch in chemokine receptors. It was notable that CD8ϩ (DP), and CD4ϩCD8Ϫ or CD4ϪCD8ϩ (SP) stages of ϩ only the FoxP3 T cells, primed in MLN but not PLN, expressed development. These thymocyte subsets are differentially located in ␣ ␤ gut homing-related 4 7 and CCR9 at significant levels. Skin the thymus: DN cells are found in the subcapsular region, whereas ϩ homing-related CCR8 was expressed on many more FoxP3 T DP cells are located in the cortex. CD4ϩCD8Ϫ or CD4ϪCD8ϩ SP cells primed in PLN than in MLN. Again, these results suggest that T cells are located in the medulla of thymus. Previously, others and FoxP3ϩ T cells are programmed to efficiently express the memory/ we found that thymocytes undergo developmental changes in che- http://www.jimmunol.org/ effector-type trafficking receptors upon Ag priming, and, impor- mokine receptor expression (42–45), and chemokine receptors ϩ tantly, this occurs at highly accelerated rates for FoxP3 T cells play important roles in T cell maturation and/or localization within Ϫ ϩ compared with FoxP3 CD4 T cells. the thymus (46–49). Our data show that FoxP3ϩ thymocytes un- dergo a developmental switch in trafficking receptor expression medium high high low Ϫ ϩ from CXCR4 CCR9 CCR8 CCR7 (DN FoxP3 Ag-primed FoxP3 T cells, after the accelerated second switch, cells) to CXCR4highCCR9highCCR8lowCCR7low (DP FoxP3ϩ can migrate more efficiently to nonlymphoid tissues than low low low high Ϫ cells), and finally to CXCR4 CCR9 CCR8 CCR7 (ma- FoxP3 T cells ture SP FoxP3ϩ cells). The CCR9 ligand CCL25 is expressed by To directly determine the functional consequence of the acceler- thymic epithelial cells and dendritic cells (50, 51), and provides the by guest on September 25, 2021 ated chemokine receptor switch, we performed a short-term (20 h) signal important for their localization in the subcapsular region of in vivo homing study of Ag-primed FoxP3ϩ T cells (Fig. 9). The the thymus (44, 52). The major CCR8 ligand CCL1 is expressed 20-h time period for homing was chosen because it was long by thymic macrophages and fibrous septa epithelial cells (53). A enough for the primed cells to migrate to various tissue sites but CCR7 ligand CCL19 is predominantly expressed in the medulla not enough for their expansion or induction at these sites. Ag- including endothelial venules (54). Considering their expression primed FoxP3ϩ T cells were prepared in vivo by transfer of OVA- sites, these chemokine receptors may regulate localization of DP specific FoxP3ϩ T cells into BALB/c mice followed by immuni- and SP FoxP3ϩ T cells in different microenvironments between zation. Eight days later, splenocytes containing Ag-primed subcapsular area and medulla. Consistent with the trafficking re- FoxP3ϩ and FoxP3Ϫ T cells were harvested from the immunized ceptor phenotype (CXCR4lowCCR9lowCCR8lowCCR7high cells), mice and then reinjected into BALB/c mice to determine their our data show that recent FoxP3ϩ thymic emigrants almost exclu- migration capacity to various tissues within 20 h. Absolute num- sively migrate to secondary lymphoid tissues. Once within the sec- bers of the OVA-specific FoxP3ϩ and FoxP3Ϫ T cells that mi- ondary lymphoid tissues, they regain CXCR4 to become CXCR4high grated to various organs were determined. As shown in Fig. 9A, CCR7high cells, which are mainly localized in the T cell zones. KJ1-26ϩCD62LϩFoxP3ϩ and FoxP3Ϫ T cells preferentially mi- As shown in Fig. 1, certain FoxP3ϩ T cells are highly enriched grated to secondary lymphoid tissues (spleen, MLN, and PLN) at in selected nonlymphoid tissues, data that support and expand pre- levels similar to each other. Some CD62LϩFoxP3ϩ cells migrated vious reports by others on unconventional enrichment of Tregs in to nonlymphoid tissues at low levels (Fig. 9A). KJ1-26ϩCD62LϪ certain tissue sites (22, 24, 27). The fact that FoxP3ϩ T cells, FoxP3ϩ also migrated to secondary lymphoid tissue at somewhat freshly made in the thymus, almost exclusively migrate to second- reduced rates but more efficiently migrated to nonlymphoid tissues ary lymphoid tissue points out the need for additional trafficking (i.e., bone marrow and lamina propria of small intestine) than their receptor switches to migrate from secondary to nonlymphoid tis- FoxP3Ϫ counterparts. A notable difference between FoxP3ϩ and sues. Alternatively, it is possible that the enrichment may occur FoxP3Ϫ cells was that CD62LϪFoxP3ϩ T cells significantly better because FoxP3ϩ T cells can be converted from FoxP3Ϫ naive T migrated than their FoxP3Ϫ counterparts into the nonlymphoid cells and/or expanded from existing FoxP3ϩ cells in nonlymphoid tissues (Fig. 9A). Only FoxP3ϩ, but not FoxP3Ϫ, memory T cells tissues. In the experimental models used in this study, the conver- migrated into bone marrow at significant levels. Although the ab- sion rate of FoxP3Ϫ T cells to FoxP3ϩ cells within the given time solute numbers of the CD62LϪFoxP3ϩ cells that migrated to the period was negligible (Fig. 4C). Furthermore, the results of our peritoneal cavity were small due to the small organ size (Fig. 9A), 20-h short-term homing study revealed that migration of Ag- cells were better enriched in the peritoneal cavity than in the sec- primed FoxP3ϩ cells does occur in a highly efficient manner. ondary lymphoid tissue (Fig. 9, B and C). Therefore, the results support that FoxP3ϩ T cells found in the 310 DEVELOPMENT OF MOUSE FoxP3ϩ T CELL HOMING PROGRAM nonlymphoid tissues were probably the FoxP3ϩ T cells that mi- The FoxP3ϩ cells that efficiently migrate to nonlymphoid tissue grated from secondary lymphoid tissues after Ag priming and/or sites are CD62LϪ, but not CD62Lϩ, cells. In contrast, the originated from them if the FoxP3ϩ T cells underwent expansion CD62Lϩ T cell subset mainly migrates to secondary lymphoid in the nonlymphoid tissue. tissues, which is consistent with another report by tracking CD25ϩ We hypothesize that FoxP3ϩ T cells acquire the ability to mi- or CD103ϩ T cells (20). This difference in trafficking behavior is grate to nonlymphoid tissues through the trafficking receptor in line with the different yet complementary expression patterns of switch in secondary lymphoid tissues. Our data indeed show that chemokine receptors by the two FoxP3ϩ T cell subsets. FoxP3ϩ T cells in secondary lymphoid tissues undergo an Ag- CD62LϩFoxP3ϩ cells mainly express CCR7 and CXCR4. The driven switch in trafficking receptors: down-regulation of lym- ligands of these chemokine receptors, such as CXCL12 (CXCR4), phoid tissue homing receptors (L-selectin, CCR7, and CXCR4) but CCL19, and CCL21 (CCR7), are expressed in secondary lymphoid up-regulation of a number of memory/effector chemokine recep- tissues and play important roles in lymphocyte recruitment to sec- tors (CCR2, CCR4, CCR5, CCR6, CCR8, CCR9, CXCR3, ondary lymphoid tissue (56–58). In contrast, CD62LϪFoxP3ϩ CXCR5, and CXCR6). These mouse FoxP3ϩ T cells are particu- cells express reduced levels of CCR7 and CXCR4 but high levels larly efficient at up-regulation of memory/inflammation/nonlym- of a panel of memory/effector type chemokine receptors such as phoid tissue-related trafficking receptors, a universal feature shared CCR2, CCR4, CCR5, CCR6, CCR8, CXCR3, and CXCR6. These with human FoxP3ϩ T cells (30). We found in this study that chemokine receptors are implicated in memory/effector T cell mi- FoxP3ϩ T cells and FoxP3Ϫ T cells are intrinsically different in gration to nonlymphoid tissues or sites of inflammation. Impor- up-regulation of many trafficking receptors in secondary lymphoid tantly, many more CD62LϪFoxP3ϩ cells than CD62LϪFoxP3Ϫ cells tissue. Furthermore, we were able to demonstrate the differences in express such chemokine receptors, consistent with the relative abun- Downloaded from vitro. It remains to be determined whether the transcription factor dance of FoxP3ϩ T cells in the selected nonlymphoid tissues. FoxP3, together with the TCR signal, would directly or indirectly Taken together, our results show that FoxP3ϩ T cells undergo control expression of the chemokine receptors at the molecular level. programmed switches in migratory behavior and trafficking recep- Recently, a report raised a possibility that regulatory T cells may tors in the thymus and secondary lymphoid tissues. The first switch suppress activation of non-FoxP3ϩ T cells and up-regulation of a occurring in the thymus is consistent with the specific emigration ϩ chemokine receptor CXCR3 (55). It is possible that activation of of FoxP3 T cells into secondary lymphoid tissues. The second http://www.jimmunol.org/ FoxP3Ϫ T cells may be suppressed by the FoxP3ϩ T cells sharing the switch in secondary lymphoid tissues upon Ag priming is also same Ag specificity to make FoxP3Ϫ T cells less efficient in the che- consistent with their migration to nonlymphoid tissues or sites of mokine receptor switch. However, our data that FoxP3ϩ and FoxP3Ϫ inflammation. Moreover, our study revealed the high expression of T cells lost CCR7 and CXCR4 at similar rates following antigenic nonlymphoid tissue homing receptors by FoxP3ϩ T cells com- stimulation (Fig. 6A) suggest that both FoxP3ϩ and FoxP3Ϫ T cells pared with FoxP3Ϫ T cells after the second switch, which is con- were comparably activated. To more rigorously demonstrate the dif- sistent with their efficient migration and enrichment in major non- ferences in chemokine receptor switch, we performed experiments lymphoid tissues. Also at the second switch, FoxP3ϩ T cells can also with FoxP3Ϫ T cells from DO11.10 rag2Ϫ/Ϫ mice (Fig. 8). These diverge into heterogeneous groups in terms of trafficking receptors. mice have OVA-specific FoxP3Ϫ CD4 T cells but not FoxP3ϩ cells, Although the first switch is a programmed one leading to gener- by guest on September 25, 2021 while mOVA ϫ DO11.10 rag2Ϫ/Ϫ mice are highly enriched with ation of FoxP3ϩ T cells with a largely homogeneous trafficking FoxP3ϩ cells. We did not use wild-type FoxP3Ϫ T cells because it is potential to secondary lymphoid tissues, the robust second switch not feasible to observe synchronized chemokine receptor switches upon Ag priming in secondary lymphoid tissues appears to gen- using polyclonal wild-type FoxP3Ϫ T cells in an Ag-dependent man- erate FoxP3ϩ cell subsets with heterogeneous homing potentials to ner in vivo. These experiments demonstrated that up-regulation of different secondary lymphoid and nonlymphoid tissues. memory/effector type chemokine receptors in KJ1-26ϩFoxP3Ϫ T cells was not suppressed by coinjected KJ1-26ϩFoxP3ϩ T cells. Acknowledgments Therefore, the available evidence, in vivo and in vitro, presented in We thank C. W. Wang and H. W. Lim for helpful input (Kim laboratory). ϩ this report supports the hypothesis that FoxP3 T cells are intrinsi- We also thank Dr. A. Abbas (University of California, San Francisco, CA) cally more efficient in up-regulation of nonlymphoid tissue trafficking and Dr. W. Heath (Walter and Eliza Hall Institute) for RIP-mOVA mice receptors than FoxP3ϪCD4ϩ T cells. Additionally, we found that and Dr. K. Hieshima (Kinki University) for the CCL27-Fc fusion vector. FoxP3Ϫ T cells, transferred from two different mouse strains (DO11.10 rag2Ϫ/Ϫ mice and mOVA ϫ DO11.10 rag2Ϫ/Ϫ mice) be- Disclosures have similarly in expression of many chemokine receptors. However, The authors have no financial conflict of interest. ␣ ␤ up-regulation of a few receptors such as CCR4, CCR7, and 4 7 was weaker on FoxP3Ϫ T cells transferred from DO11.10 rag2Ϫ/Ϫ mice References than those from mOVA ϫ DO11.10 rag2Ϫ/Ϫ mice. This result is 1. Sakaguchi, S. 2004. Naturally arising CD4ϩ regulatory T cells for immunologic unexpected and warrants further investigation in the future. self-tolerance and negative control of immune responses. Annu. Rev. Immunol. ϩ 22: 531–562. It has been undetermined where and when Foxp3 T cells ac- 2. Baecher-Allan, C., V. Viglietta, and D. A. Hafler. 2004. Human CD4ϩCD25ϩ quire tissue-specific trafficking receptors such as gut homing-re- regulatory T cells. Semin. Immunol. 16: 89–98. ␣ ␤ 3. Bettelli, E., M. Dastrange, and M. Oukka. 2005. Foxp3 interacts with nuclear lated CCR9 and 4 7 and skin homing-related CCR4 and CCR8. factor of activated T cells and NF-␬B to repress cytokine expression and ϩ ␣ ␤ We found that FoxP3 T cell expression of CCR9 and 4 7 is effector functions of T helper cells. Proc. Natl. Acad. Sci. USA 102: 5138–5143. induced in MLN, but not in PLN (Figs. 7 and 8). In contrast, CCR8 4. Hori, S., T. Nomura, and S. Sakaguchi. 2003. Control of regulatory T cell de- ϩ velopment by the transcription factor Foxp3. Science 299: 1057–1061. is expressed by many more FoxP3 T cells primed in PLN than 5. Fontenot, J. D., M. A. Gavin, and A. Y. Rudensky. 2003. Foxp3 programs the MLN. CCR4 is highly expressed by the FoxP3ϩ cells in both or- development and function of CD4ϩCD25ϩ regulatory T cells. Nat. Immunol. 4: ϩ 330–336. gans, and thus, cannot distinguish FoxP3 T cells primed in skin 6. Khattri, R., T. Cox, S. A. Yasayko, and F. Ramsdell. 2003. An essential role for vs gut. These results suggest that it is the second switch after which Scurfin in CD4ϩCD25ϩ T regulatory cells. Nat. Immunol. 4: 337–342. the naive-like FoxP3ϩ T cells with homogeneous trafficking po- 7. Chatila, T. A., F. Blaeser, N. Ho, H. M. Lederman, C. Voulgaropoulos, C. Helms, ϩ and A. M. Bowcock. 2000. JM2, encoding a fork head-related , is mutated tential to secondary nonlymphoid tissue diverge into FoxP3 T in X-linked autoimmunity-allergic disregulation syndrome. J. Clin. Invest. 106: cells with heterogeneous trafficking potentials to different tissue sites. R75–R81. The Journal of Immunology 311

8. Brunkow, M. E., E. W. Jeffery, K. A. Hjerrild, B. Paeper, L. B. Clark, roles in mucosal immunity as a chemokine with broad-spectrum antimicrobial S. A. Yasayko, J. E. Wilkinson, D. Galas, S. F. Ziegler, and F. Ramsdell. 2001. activity. J. Immunol. 170: 1452–1461. Disruption of a new forkhead/winged-helix protein, Scurfin, results in the fatal 34. Abkowitz, J. L., S. N. Catlin, M. T. McCallie, and P. Guttorp. 2002. Evidence lymphoproliferative disorder of the scurfy mouse. Nat. Genet. 27: 68–73. that the number of hematopoietic stem cells per animal is conserved in mammals. 9. Itoh, M., T. Takahashi, N. Sakaguchi, Y. Kuniyasu, J. Shimizu, F. Otsuka, and Blood 100: 2665–2667. S. Sakaguchi. 1999. Thymus and autoimmunity: production of CD25ϩCD4ϩ 35. Boggs, D. R. 1984. The total marrow mass of the mouse: a simplified method of naturally anergic and suppressive T cells as a key function of the thymus in measurement. Am. J. Hematol. 16: 277–286. maintaining immunologic self-tolerance. J. Immunol. 162: 5317–5326. 36. Fu, S., A. C. Yopp, X. Mao, D. Chen, N. Zhang, M. Mao, Y. Ding, and 10. Papiernik, M., M. L. de Moraes, C. Pontoux, F. Vasseur, and C. Penit. 1998. J. S. Bromberg. 2004. CD4ϩ CD25ϩ CD62ϩ T-regulatory cell subset has optimal Regulatory CD4 T cells: expression of IL-2R␣ chain, resistance to clonal deletion suppressive and proliferative potential. Am. J. Transplant. 4: 65–78. and IL-2 dependency. Int. Immunol. 10: 371–378. 37. Taylor, P. A., A. Panoskaltsis-Mortari, J. M. Swedin, P. J. Lucas, R. E. Gress, 11. Fontenot, J. D., J. L. Dooley, A. G. Farr, and A. Y. Rudensky. 2005. Develop- B. L. Levine, C. H. June, J. S. Serody, and B. R. Blazar. 2004. L-Selectinhigh but mental regulation of Foxp3 expression during ontogeny. J. Exp. Med. 202: not the L-selectinlow CD4ϩ25ϩ T-regulatory cells are potent inhibitors of GVHD 901–906. and BM graft rejection. Blood 104: 3804–3812. 12. Kawahata, K., Y. Misaki, M. Yamauchi, S. Tsunekawa, K. Setoguchi, 38. You, S., G. Slehoffer, S. Barriot, J. F. Bach, and L. Chatenoud. 2004. Unique role J. Miyazaki, and K. Yamamoto. 2002. Generation of CD4ϩCD25ϩ regulatory T of CD4ϩCD62Lϩ regulatory T cells in the control of autoimmune diabetes in T cells from autoreactive T cells simultaneously with their negative selection in the cell receptor transgenic mice. Proc. Natl. Acad. Sci. USA 101(Suppl. 2): thymus and from nonautoreactive T cells by endogenous TCR expression. J. Im- 14580–14585. munol. 168: 4399–4405. 39. Walker, L. S., A. Chodos, M. Eggena, H. Dooms, and A. K. Abbas. 2003. An- 13. Jordan, M. S., A. Boesteanu, A. J. Reed, A. L. Petrone, A. E. Holenbeck, tigen-dependent proliferation of CD4ϩ CD25ϩ regulatory T cells in vivo. J. Exp. M. A. Lerman, A. Naji, and A. J. Caton. 2001. Thymic selection of CD4ϩCD25ϩ Med. 198: 249–258. regulatory T cells induced by an agonist self-peptide. Nat. Immunol. 2: 301–306. 40. Crellin, N. K., R. V. Garcia, O. Hadisfar, S. E. Allan, T. S. Steiner, and 14. Apostolou, I., A. Sarukhan, L. Klein, and H. von Boehmer. 2002. Origin of M. K. Levings. 2005. Human CD4ϩ T cells express TLR5 and its ligand flagellin regulatory T cells with known specificity for antigen. Nat. Immunol. 3: 756–763. enhances the suppressive capacity and expression of FoxP3 in CD4ϩCD25ϩ T 15. Kretschmer, K., I. Apostolou, D. Hawiger, K. Khazaie, M. C. Nussenzweig, and regulatory cells. J. Immunol. 175: 8051–8059. H. von Boehmer. 2005. Inducing and expanding regulatory T cell populations by 41. Liu, H., M. Komai-Koma, D. Xu, and F. Y. Liew. 2006. Toll-like receptor 2 Downloaded from foreign antigen. Nat. Immunol. 6: 1219–1227. signaling modulates the functions of CD4ϩ CD25ϩ regulatory T cells. Proc. Natl. 16. Lim, H. W., P. Hillsamer, A. H. Banham, and C. H. Kim. 2005. Cutting edge: Acad. Sci. USA 103: 7048–7053. direct suppression of B cells by CD4ϩCD25ϩ regulatory T cells. J. Immunol. 42. Kim, C. H., L. M. Pelus, J. R. White, and H. E. Broxmeyer. 1998. Differential 175: 4180–4183. chemotactic behavior of developing T cells in response to thymic chemokines. 17. Iellem, A., M. Mariani, R. Lang, H. Recalde, P. Panina-Bordignon, F. Sinigaglia, Blood 91: 4434–4443. and D. D’Ambrosio. 2001. Unique chemotactic response profile and specific 43. Campbell, J. J., J. Pan, and E. C. Butcher. 1999. Cutting edge: developmental expression of chemokine receptors CCR4 and CCR8 by CD4ϩCD25ϩ regulatory switches in chemokine responses during T cell maturation. J. Immunol. 163:

T cells. J. Exp. Med. 194: 847–853. 2353–2357. http://www.jimmunol.org/ 18. Colantonio, L., A. Iellem, F. Sinigaglia, and D. D’Ambrosio. 2002. Skin-homing 44. Uehara, S., K. Song, J. M. Farber, and P. E. Love. 2002. Characterization of CLAϩ T cells and regulatory CD25ϩ T cells represent major subsets of human CCR9 expression and CCL25/thymus-expressed chemokine responsiveness dur- peripheral blood memory T cells migrating in response to CCL1/I-309. Eur. ing T cell development: CD3highCD69ϩ thymocytes and ␥␦TCRϩ thymocytes J. Immunol. 32: 3506–3514. preferentially respond to CCL25. J. Immunol. 168: 134–142. 19. Szanya, V., J. Ermann, C. Taylor, C. Holness, and C. G. Fathman. 2002. The 45. Youn, B. S., C. H. Kim, F. O. Smith, and H. E. Broxmeyer. 1999. TECK, an subpopulation of CD4ϩCD25ϩ splenocytes that delays adoptive transfer of dia- efficacious chemoattractant for human thymocytes, uses GPR-9-6/CCR9 as a spe- betes expresses L-selectin and high levels of CCR7. J. Immunol. 169: 2461–2465. cific receptor. Blood 94: 2533–2536. 20. Huehn, J., K. Siegmund, J. C. Lehmann, C. Siewert, U. Haubold, M. Feuerer, 46. Herna´ndez-Lo´pez, C., A. Varas, R. Sacedo´n, E. Jime´nez, J. J. Mun˜oz, A. G. Zapata, G. F. Debes, J. Lauber, O. Frey, G. K. Przybylski, et al. 2004. Developmental and A. Vicente. 2002. Stromal cell-derived factor 1/CXCR4 signaling is critical for stage, phenotype, and migration distinguish naive- and effector/memory-like early human T-cell development. Blood 99: 546–554. CD4ϩ regulatory T cells. J. Exp. Med. 199: 303–313. 47. Wurbel, M. A., M. Malissen, D. Guy-Grand, E. Meffre, M. C. Nussenzweig,

21. Wysocki, C. A., S. B. Burkett, A. Panoskaltsis-Mortari, S. L. Kirby, A. D. Luster, M. Richelme, A. Carrier, and B. Malissen. 2001. Mice lacking the CCR9 CC- by guest on September 25, 2021 K. McKinnon, B. R. Blazar, and J. S. Serody. 2004. Differential roles for CCR5 chemokine receptor show a mild impairment of early T- and B-cell development expression on donor T cells during graft-versus-host disease based on pretrans- and a reduction in T-cell receptor ␥␦ϩ gut intraepithelial lymphocytes. Blood 98: plant conditioning. J. Immunol. 173: 845–854. 2626–2632. 22. Kleinewietfeld, M., F. Puentes, G. Borsellino, L. Battistini, O. Rotzschke, and 48. Kwan, J., and N. Killeen. 2004. CCR7 directs the migration of thymocytes into K. Falk. 2005. CCR6 expression defines regulatory effector/memory-like cells the thymic medulla. J. Immunol. 172: 3999–4007. within the CD25ϩCD4ϩ T-cell subset. Blood 105: 2877–2886. 49. Ueno, T., F. Saito, D. H. Gray, S. Kuse, K. Hieshima, H. Nakano, T. Kakiuchi, 23. Lee, I., L. Wang, A. D. Wells, M. E. Dorf, E. Ozkaynak, and W. W. Hancock. M. Lipp, R. L. Boyd, and Y. Takahama. 2004. CCR7 signals are essential for 2005. Recruitment of Foxp3ϩ T regulatory cells mediating allograft tolerance cortex-medulla migration of developing thymocytes. J. Exp. Med. 200: 493–505. depends on the CCR4 chemokine receptor. J. Exp. Med. 201: 1037–1044. 50. Wurbel, M. A., J. M. Philippe, C. Nguyen, G. Victorero, T. Freeman, 24. Eksteen, B., A. Miles, S. M. Curbishley, C. Tselepis, A. J. Grant, L. S. Walker, P. Wooding, A. Miazek, M. G. Mattei, M. Malissen, B. R. Jordan, et al. 2000. The and D. H. Adams. 2006. Epithelial inflammation is associated with CCL28 pro- chemokine TECK is expressed by thymic and intestinal epithelial cells and at- duction and the recruitment of regulatory T cells expressing CCR10. J. Immunol. tracts double- and single-positive thymocytes expressing the TECK receptor 177: 593–603. CCR9. Eur. J. Immunol. 30: 262–271. 25. Ishida, T., T. Ishii, A. Inagaki, H. Yano, H. Komatsu, S. Iida, H. Inagaki, and 51. Vicari, A. P., D. J. Figueroa, J. A. Hedrick, J. S. Foster, K. P. Singh, S. Menon, R. Ueda. 2006. Specific recruitment of CC chemokine receptor 4-positive regu- N. G. Copeland, D. J. Gilbert, N. A. Jenkins, K. B. Bacon, and A. Zlotnik. 1997. latory T cells in Hodgkin lymphoma fosters immune privilege. Cancer Res. 66: TECK: a novel CC chemokine specifically expressed by thymic dendritic cells 5716–5722. and potentially involved in T cell development. Immunity 7: 291–301. 26. Yang, Z. Z., A. J. Novak, M. J. Stenson, T. E. Witzig, and S. M. Ansell. 2006. 52. Benz, C., K. Heinzel, and C. C. Bleul. 2004. Homing of immature thymocytes to Intratumoral CD4ϩCD25ϩ regulatory T-cell-mediated suppression of infiltrating the subcapsular microenvironment within the thymus is not an absolute require- CD4ϩ T cells in B-cell non-Hodgkin lymphoma. Blood 107: 3639–3646. ment for T cell development. Eur. J. Immunol. 34: 3652–3663. 27. Zou, L., B. Barnett, H. Safah, V. F. Larussa, M. Evdemon-Hogan, P. Mottram, 53. Annunziato, F., L. Cosmi, F. Liotta, E. Lazzeri, R. Manetti, V. Vanini, S. Wei, O. David, T. J. Curiel, and W. Zou. 2004. Bone marrow is a reservoir for P. Romagnani, E. Maggi, and S. Romagnani. 2002. Phenotype, localization, and CD4ϩCD25ϩ regulatory T cells that traffic through CXCL12/CXCR4 signals. mechanism of suppression of CD4ϩCD25ϩ human thymocytes. J. Exp. Med. Cancer Res. 64: 8451–8455. 196: 379–387. 28. Wei, S., I. Kryczek, and W. Zou. 2006. Regulatory T cell compartmentalization 54. Ueno, T., K. Hara, M. S. Willis, M. A. Malin, U. E. Hopken, D. H. Gray, and trafficking. Blood 108: 426–431. K. Matsushima, M. Lipp, T. A. Springer, R. L. Boyd, et al. 2002. Role for CCR7 29. Lim, H. W., P. Hillsamer, and C. H. Kim. 2004. Regulatory T cells can migrate ligands in the emigration of newly generated T lymphocytes from the neonatal to follicles upon T cell activation and suppress GC-Th cells and GC-Th cell- thymus. Immunity 16: 205–218. driven B cell responses. J. Clin. Invest. 114: 1640–1649. 55. Sarween, N., A. Chodos, C. Raykundalia, M. Khan, A. K. Abbas, and 30. Lim, H. W., H. E. Broxmeyer, and C. H. Kim. 2006. Regulation of trafficking L. S. Walker. 2004. CD4ϩCD25ϩ cells controlling a pathogenic CD4 response receptor expression in human forkhead box P3ϩ regulatory T cells. J. Immunol. inhibit cytokine differentiation, CXCR-3 expression, and tissue invasion. J. Im- 177: 840–851. munol. 173: 2942–2951. 31. Curiel, T. J., G. Coukos, L. Zou, X. Alvarez, P. Cheng, P. Mottram, 56. Scimone, M. L., T. W. Felbinger, I. B. Mazo, J. V. Stein, U. H. Von Andrian, and M. Evdemon-Hogan, J. R. Conejo-Garcia, L. Zhang, M. Burow, et al. 2004. W. Weninger. 2004. CXCL12 mediates CCR7-independent homing of central Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune memory cells, but not naive T cells, in peripheral lymph nodes. J. Exp. Med. 199: privilege and predicts reduced survival. Nat. Med. 10: 942–949. 1113–1120. 32. Kim, C. H., K. Nagata, and E. C. Butcher. 2003. Dendritic cells support sequen- 57. Fo¨rster, R., A. Schubel, D. Breitfeld, E. Kremmer, I. Renner-Mu¨ller, E. Wolf, and tial reprogramming of chemoattractant receptor profiles during naive to effector M. Lipp. 1999. CCR7 coordinates the primary immune response by establishing T cell differentiation. J. Immunol. 171: 152–158. functional microenvironments in secondary lymphoid organs. Cell 99: 23–33. 33. Hieshima, K., H. Ohtani, M. Shibano, D. Izawa, T. Nakayama, Y. Kawasaki, 58. Kim, C. H. 2005. The greater chemotactic network for lymphocyte trafficking: F. Shiba, M. Shiota, F. Katou, T. Saito, and O. Yoshie. 2003. CCL28 has dual chemokines and beyond. Curr. Opin. Hematol. 12: 298–304.