Human thymus contains major dendritic cell (DC) subsets, myeloid DCs (mDCs), and plasmacytoid DCs (pDCs). We previously showed that mDCs, educated by thymic stromal lymphopoietin (TSLP) produced by the epithelial cells of the Hassall’s corpuscles, induced differentiation of CD4+CD25 thymocytes into Forkhead Box P3+ (FOXP3+) regulatory T cells (TR) within the medulla of human thymus. In this study, we show that pDCs expressed the TSLP receptor and IL-7 receptor α complexes upon activation and became responsive to TSLP. TSLP-activated human pDCs secrete macrophage-derived chemokine CCL-22 and thymus- and activation-regulated chemokine CCL-17 but not Th1- or Th2-polarizing cytokines. TSLP-activated pDCs induced the generation of FOXP3+ TR from CD4+CD8CD25 thymocytes, which could be strongly inhibited by Th1-polarizing cytokine IL-12 or Th2-polarizing cytokine IL-4. Interestingly, the FOXP3+ TR induced by the TSLP-pDCs expressed more IL-10 but less TGF-β than that induced by the TSLP-mDCs. These data suggest that TSLP expressed by thymic epithelial cells can activate mDCs and pDCs to positively select the FOXP3+ TR with different cytokine production potential in human thymus. The inability of TSLP to induce DC maturation without producing Th1- or Th2-polarizing cytokines may provide a thymic niche for TR development.

Nondeletional tolerance leading to the generation of CD4+CD25+FOXP3+ regulatory T cells (TR) within the thymus represents a key mechanism to establish immunologic self-tolerance (16). How some of the medium- to high-affinity self-reactive TR precursors escape negative selection and what type of APCs positively selects them to become TR within the thymus are unsolved questions.

Pioneering studies using three different models suggested that thymic epithelial cells are critical for the induction of nondeletional tolerance or the generation of TR. The first model was based on xeno- or allotransplantation of embryotic thymic rudiments of chick, quail, or mice into a developing embryo before the colonization of hematopoietic progenitors (79). These embryonic thymic rudiments at this stage were called precolonized embryotic thymic rudiments and were believed to give rise to thymic epithelial cells but not hematopoietic cells. Therefore, it has been generally accepted that the resulting chimeric thymus contains epithelial cells of donor origin and hematopoeitic cells of host origin. The results of these experiments suggested that thymic epithelial cells, but not hematopoietic cells, are responsible for the induction of central tolerance, either by deletion or by a nondeletional mechanism involving the induction of TR (1, 10, 11). A major question is whether the precolonized thymic rudiments contain hematopoietic progenitors that will give rise to thymic dendritic cells (DCs) in the medulla. Because the window during which precolonized thymic rudiments can be isolated without colonization of the hematopoietic stem cells is very narrow and variable in all three species, it is likely that the hematopoietic progenitors within these thymic rudiments were contaminated in these early studies. In particular, donor DCs might have been mistakenly classified as reticular cells or stromal cells in these earlier studies.

The second model involved the transfer of thymic epithelial cells or thymic stroma cells derived from thymic grafts treated with irradiation or with deoxyguanosine that kills the cycling hematopoietic cells within the thymus (12). However, it is unclear whether such treatment would have efficiently eliminated all the noncycling thymic DCs in the medulla.

The third model involved targeted expression of MHC class II (MHC-II) molecules on thymic cortical epithelial cells in class II-deficient mice and showed that the CD4+CD25+ TR are selected by thymic cortical epithelial cells (5, 13, 14). The questions posed were: how did the CD4+CD25+ TR that were selected by cortical epithelial cells escape the negative selection mediated by medullary DCs, and if cortical epithelial cells selected the CD4+CD25+ TR at the CD4+CD8+ double-positive stage, why were so few T cells expressing Forkhead Box P3+ (FOXP3+) TR found in the cortex (15, 16)?

We have recently shown that a subset of myeloid DCs (mDCs) educated by thymic stromal lymphopoietin (TSLP) produced by epithelial cells of the Hassall’s corpuscles induced differentiation of CD4+CD25 thymocytes into FOXP3+ TR within the medulla of human thymus (16). The ability of TSLP-activated mDCs to induce TR appeared to be linked with their unique features, including: 1) expression of high MHC-II and costimulatory molecules (CD80, CD86); 2) sparse production of proinflammatory cytokines IL-1, IL-6, and IL-12, which inhibit TR development; and 3) prolonged formation of stable conjugate with T cells that provide sustained TCR signaling and costimulation (17). We believe that DCs have several advantages over epithelial cells as the APCs for selecting TR in thymus: 1) DCs can transport peripheral Ags into the thymus; 2) DCs have the ability to cross-present self-Ags, including epithelial cell-derived Ags; and 3) TSLP-matured DCs can provide better quality of costimulation to prevent the developing T cells’ propensity for apoptosis when they are signaled through medium- to high-affinity self-reactive TCR.

Human thymus contains at least two major subsets of DCs, including conventional mDCs and plasmacytoid DCs (pDCs) (18, 19). We have previously shown that TSLP selectively activates ex vivo-derived peripheral blood mDCs but not other blood cell types (20). Recently, we found that although resting pDCs do not express the TSLP receptor (TSLPR), they do rapidly express mRNA encoding for TSLPR and IL-7 receptor α (IL-7Rα) following activation, suggesting that pDCs may acquire the ability to respond to TSLP following activation.

In our current study, we showed that pDCs rapidly expressed surface TSLPR and IL-7Rα complexes and became responsive to TSLP activation. TSLP-activated human pDCs expressed higher levels of CD80 and CD86 and secreted thymus and activation-regulated chemokine CCL-17 and macrophage-derived chemokine CCL-22 but not Th1- or Th2-polarizing cytokines. TSLP-activated pDCs could efficiently induce the generation of FOXP3+ TR from the CD4+CD8CD25 thymocytes, which could be strongly inhibited by Th1-polarizing cytokine IL-12 or Th2-polarizing cytokine IL-4. Interestingly, the FOXP3+ TR induced by the TSLP-activated pDCs appeared to produce more IL-10 and to express less TGF-β than that induced by the TSLP-activated mDCs. Our study suggests that there are multiple types of APCs such as TSLP-activated mDCs and TSLP-activated pDCs in thymus that are responsible for the selection of functional distinct subsets of FOXP3+ TR. The inability of TSLP to induce DC maturation without producing Th1- or Th2-polarizing cytokines may provide a suitable thymic niche for TR development.

This study was approved by the Institutional Review Board for Human Research at The University of Texas M. D. Anderson Cancer Center, Houston, TX. Thymuses from human newborns and children 0 d to 2 y old were obtained from the Texas Children’s Hospital, Houston, Texas. Adult blood buffy coats from healthy donors were obtained from the Gulf Coast Regional Blood Center, Houston, TX. For isolation of thymic pDCs, thymuses were cut into small pieces and digested as previously described (18). After separation of the mononuclear cells by Ficoll centrifugation, blood DC Ag 4+ (BDCA-4+) cells were obtained from thymus or PBMCs using a Microbeads-BDCA-4 isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany). The enriched cells were stained with FITC-conjugated anti-CD3, CD14, CD16, CD56, CD19, CD20, PE-conjugated anti-CD11c, allophycocyanin-conjugated anti-CD123, and allophycocyanin-Cy7–conjugated anti-CD4 Abs (BD Biosciences, San Jose, CA), and then linageCD11cCD4+CD123+ cells were purified using FACS on an FACS Aria (BD Biosciences) to reach >99% purity, as described (21). Isolated pDCs were cultured for 48 h with various stimuli with or without TSLP (50 ng/ml; recombinant protein was prepared in house using an adenovirus vector system) as described (17). After 48 h of culture, we determined the cell-surface marker characteristics of activated pDCs also as described (17).

For isolation of T cell linage thymocytes, thymuses were digested, and mononuclear cells were separated by Ficoll centrifugation as described (17). Briefly, T cell linage thymocytes were obtained by negative deletion using a mixture of mouse mAbs against the linage markers CD11c, CD14, CD15, CD20, CD56, and CD235a (produced in-house). This was followed by incubation of these thymocytes with goat anti-mouse IgG-coated magnetic beads (M-450; Dynal, Oslo, Norway) and microbeads goat anti-mouse IgG (Miltenyi Biotec). Enriched cells were stained with FITC-conjugated anti-CD11c, CD14, CD16, CD19, CD20, CD56, and BDCA-2, PE-conjugated anti-CD8, PE-Cy7–conjugated anti-CD25, and allophycocyanin-Cy7 conjugated anti-CD4 Abs (BD Biosciences), and then LinageCD4+CD8CD25 cells were purified using FACS on an FACSAria (BD Biosciences) to reach >99% purity.

After 2 d of culture, pDCs were collected and washed three times to remove any cytokines. Viable DCs were counted by trypan blue exclusion of dead cells. The remaining cells were cocultured with 2 × 104 freshly isolated allogeneic thymocytes or peripheral blood-naive CD4+ T cells in round-bottomed 96-well culture plates in IMDM containing 2% human AB serum and 10 ng/ml of IL-7. Cultures were done in triplicate at a 1:2 or 1:1 ratio of DCs:T cells. Neutralizing anti–HLA-DR (BD Biosciences), anti-CD80, anti-CD86, and anti–IL-2 (R&D Systems, Minneapolis, MN) mAbs were used in culture at a concentration of 30 μg/ml. The labeling of T cells with CFSE (Molecular Probes, Carlsbad, CA) or PKH26 (Sigma-Aldrich, St. Louis, MO) was performed as described (17). In some experiments, isolated T cells were cultured with plate-bound anti-CD3 Ab (OKT-3; 2 μg/ml), soluble anti-CD28 (28.2; 1 μg/ml), 20 U/ml IL-2, and 10 ng/ml IL-7 with or without 10 μg/ml TGF-β (R&D Systems).

After 7 d of culture, cells were collected and resuspended in an EDTA-containing medium to dissociate the clusters. Viable cells were counted by trypan blue exclusion of dead cells. The remaining cells were stained with PE-conjugate anti-CD25, and Alexa Fluor 647-conjugated anti-FOXP3 Abs (259D or 236A/E7) (Biolegend, San Diego, CA, and eBioscience, San Diego, CA) and analyzed with a LSR-II (BD Biosciences). Dead cells were excluded on the basis of side- and forward-scatter characteristics, and viable CD4+CD25+ and FOXP3+ T cell numbers were calculated as: the percentage of cells in the cell type × the number of viable cells.

After 7 d of culture with TSLP-DCs, CD4+ thymocytes were stained with PE-conjugated anti-CD25 (Miltenyi Biotec), and then CD25high cells and CD25dim cells, which divided >5 times, were isolated by cell sorting. The mixture of these sorted cells and peripheral CD4+CD25 T cells at a 1:1 ratio or multiple ratios were stimulated with immobilized anti-CD3 (OKT-3, 2 μg/ml) and anti-CD28 Abs (CD28.2, 1 μg/ml) for 4 d. Cellular proliferation was assessed by [3H]thymidine incorporation, as described in Ref. 17. Peripheral CD4+CD25 T cells were also labeled with CFSE and cultured with sorted CD25high cells and CD25dim cells at a 1:1 ratio in the presence of immobilized anti-CD3 (OKT-3, 2 μg/ml) and soluble anti-CD28 Abs (CD28.2, 0.5 μg/ml). After 4 d of cultures, cell division of CFSE-labeled CD4+CD25 T cells was evaluated by flow cytometry.

DC culture supernatants were collected at 24 h, and chemokine and cytokine production was assessed as described Ref. 17. CD4+CD25 T cells from adult peripheral blood, fetal, or neonatal thymuses were cultured for 7 d with TSLP-DCs and IL-7 or with immobilized anti-CD3, soluble anti-CD28 Abs, IL-2, and IL-7. T cells were then washed and restimulated with 50 ng/ml PMA and 2 μg/ml ionomycin for 6 h (Sigma-Aldrich). A total of 10 μg/ml of brefeldin A (Sigma-Aldrich) was added during the last 2 h. The T cells were stained with PE-conjugated anti–IL-10 (BD Biosciences) and Alexa Fluor 647-conjugated anti-FOXP3 (259D or 236A/E7) Abs using CALTAG FIX and PERM kit (Invitrogen, Carlsbad, CA) and the FOXP3 staining kit (eBioscience). The T cells were also stimulated with plate-bound 2 μg/ml of anti-CD3 and 1 μg/ml of soluble anti-CD28 Abs at a concentration of 106 cells/ml for 24 h. The level of IL-10 in the supernatants was measured by ELISA (R&D Systems). For detection of membrane-bound latency associated peptide (LAP) (TGF-β1) on the surface, cells were stained with LAP (TGF-β1) Ab (27232, R&D Systems), followed by staining with PE-conjugated anti-mouse IgG1 Abs (BD Biosciences).

After 7 d of culture with DCs, T cells were stained with PE-conjugated anti-CD25 and then fixed with Fixation/Permeabilization working solution (eBioscience) for 1 h. After washing, the cells were treated with Permeabilization Buffer (eBioscience) in the presence of Alexa Fluor 647-conjugated anti-FOXP3 Ab (259D or 236A/E7) for 30 min.

For immunofluorescence staining, the slides were incubated with biotinylated mouse anti-CD123 mAb (BD Biosciences) at room temperature for 1 h. After washing, anti-CD123 mAb-dependent tissue deposition of biotin was increased using a tyramide signal biotin amplification system and then was visualized by Alexa Fluor 549-conjugated streptavidin (Molecular Probes). The slides were then stained with FITC-conjugated anti-FOXP3 mAb (eBioscience) for 1 h. Images were acquired by using an inverted microscope, BX41 (Olympus, Tokyo, Japan). Final image processing was performed by using Photoshop (Adobe, San Jose, CA).

We have previously shown that TSLP could strongly activate mDCs to express costimulatory molecules CD80 and CD86 and to produce Th2 chemokines CCL-17 and CCL-22 without producing Th1-polarizing cytokine IL-12 (20). Through gene expression analyses, we recently found that pDCs rapidly express mRNA encoding for TSLPR and IL-7Rα following activation through TLR-7 or TLR-9 (data not shown). This stimulated us to further investigate whether pDCs have the ability to express surface TSLPR at the protein level and to become functionally responsive to TSLP. We found that whereas pDCs constitutively expressed surface IL-7Rα, pDCs expressed surface TSLPR following activation by TLR7-ligand R848 and TLR9-ligand CpG-B (Fig. 1A). IL-3 plus CD40L, CpG-C, HSV, and influenza A moderately upregulated the expression of TSLPR by pDCs. We found that TSLP strongly upregulated surface expression of CD80 and CD86 on pDCs activated by CpG-B or R848 (Fig. 1B) and promoted pDCs to produce CCL-17 and CCL-22 without inducing or effecting IFN-α or IL-12 production in culture (Fig. 1C). These data suggest that, unlike CD40L or ligands for different TLRs, TSLP represents a very unique stimulus that induces activation of both mDCs and pDCs to upregulate costimulatory molecules without producing either Th1-polarizing cytokines or proinflammatory cytokines.

In the human system, we have recently shown that TSLP is expressed by epithelial cells of the Hassall’s corpuscles in the thymic medulla, and TSLP-conditioned mDCs could induce the differentiation of CD4+CD25 thymocytes into CD4+CD25+FOXP3+ TR cells (16). Because human thymus contains both mDC and pDC, we next investigated whether TSLP-activated human pDCs could induce the differentiation of CD4+CD25 thymocytes into CD4+CD25+FOXP3+ TR cells. CD4+CD25FOXP3 thymocytes isolated from human thymus were labeled with CFSE and cultured for 7 d with allogeneic pDCs precultured for 2 d with CpG-B in the presence or absence of TSLP. The sorted CD4+CD25 thymocytes contained <0.2% of FOXP3+ cells as assessed by staining with anti-human FOXP3 mAbs 259D (Fig. 2A) and 236A/E7 (data not shown). We found that TSLP-conditioned CpG-B preactivated pDCs (TSLP-pDCs) induced a vigorous expansion of CD4+ thymocytes (Fig. 2B, 2D, 2F) and the generation of CD4+CD25+FOXP3+ cells (Fig. 2B, 2E, 2G), whereas pDCs activated by CpG-B alone induced less expansion and few generation of FOXP3+ cells. The number of the FOXP3+ cells induced by TSLP-pDCs followed the number of TSLP-pDCs added in the cultures (Fig. 2G). In contrast, whereas CD4+CD25 thymocytes differentiated into FOXP3+ cells by stimulating with anti-CD3, anti-CD28, IL-2, and TGF-β–like peripheral naive CD4+ T cells (22), stimulation of CD4+CD25 thymocytes with anti-CD3 and IL-2 or anti-CD3, anti-CD28, and IL-2 induced vigorous expansion but not the differentiation of CD4+CD25 cells into FOXP3+ cells (Fig. 2C). These data suggest that the stimulation by anti-CD3, anti-CD28, and IL-2 is not sufficient to induce FOXP3+ cells from human CD4+CD25 thymocytes.

We then examined whether TSLP-pDCs induced the differentiation of CD4+CD25FOXP3 cells into CD4+CD25+FOXP3+ cells or the expansion of a small number of CD4+CD25+FOXP3+ TR contaminated in sorted CD4+CD25 thymocytes. To determine this, we performed cell-mixed culture experiments with CFSE-labeled CD4+CD25+ TR cells and PKH26-labeled CD4+CD25 cells (16). Sorted CD4+CD25+ thymocytes were labeled with CFSE and mixed with PKH26-labeled CD4+CD25 thymocytes at a 1:9 ratio. This ratio was chosen because the original CD4+ thymocyte population contains CD25+ and CD25 cells at 1:9 ratios (Fig. 2A) (16). These mixed cells were then cultured together with TSLP-pDCs, and the proliferative response of the two populations was compared (Fig. 3A, 3B). We found that thymic CD4+CD25+ TR in mixed culture expanded ∼4-fold, which is a similar extent as the expansion of CD4+CD25 cells. Thus, it seems unlikely that the contaminated CD4+CD25+FOXP3+ TR, which is <0.2% in CD4+CD25 cells, can reach to 4.8%. Based on the cell number, contaminated CD4+CD25+FOXP3+ cells have to expand >30–200-fold to reach 4.8% of cultured cells. Therefore, these data indicate that TSLP-pDCs preferentially induce CD4+CD25 thymocytes to differentiate into CD4+CD25+FOXP3+ T cells.

The generation of CD4+CD25+FOXP3+ TR induced by TSLP-pDCs was completely blocked by the addition of Abs to MHC-II (HLA-DR), CD80/CD86, or IL-2 (Fig. 4A, 4B), suggesting the critical role of the interaction of TCR and Ag/MHC-II complex, CD28 signaling through CD80/CD86, and IL-2 in the generation and expansion of CD4+CD25+FOXP3+ TR in thymus, as previously observed in both humans and mice (2326).

We next examined the phenotype, cytokine profile, and suppressive activity of FOXP3+ cells generated from CD4+CD25FOXP3 thymocytes by TSLP-pDCs. CD4+CD25FOXP3 thymocytes were labeled with CFSE and cultured with TSLP-pDCs. After 7 d of culture, we evaluated the surface phenotype of FOXP3+ or FOXP3 cells in the same culture, which went through similar rounds of cell division after culture with pDCs. As shown in Fig. 5A, FOXP3+ cells expressed much higher levels of CD25, CTLA-4, and ICOS than FOXP3 cells and a similar level of glucocorticoid-induced TNF receptor to that expressed by FOXP3 cells. Furthermore, FOXP3+ cells induced by TSLP-pDCs did not produce IL-2 and IFN-γ, whereas FOXP3 cells in the same culture produced significant levels of IL-2 and IFN-γ (Fig. 5B).

We next examined whether CD4+CD25+FOXP3+ T cells generated by TSLP-pDCs exhibited suppressive function. After 7 d of culture, we sorted both CD25high and CD25dim cells, which went through >5 times of division. More than 60% of the sorted CD25high cells and <5% of CD25dim cells expressed FOXP3 (Fig. 6A). These cells were cultured with CD4+CD25 T cells isolated from peripheral blood in the presence of immobilized anti-CD3 and soluble anti-CD28. We found that CD4+CD25highFOXP3+ cells induced by TSLP-pDCs were anergic and displayed the ability to strongly inhibit the proliferation of CD4+CD25 T cells (Fig. 6B, 6D). In contrast, CD4+CD25dimFOXP3 cells were neither anergic nor suppressive (Fig. 6C, 6D). They even promoted the expansion of responder CD4+CD25 T cells rather than inhibited their proliferation. We confirmed these results by CFSE-labeling experiments. The proliferation of CFSE-labeled responder CD4+CD25 T cells was significantly inhibited by CD25+CD25highFOXP3+ cells but not CD4+CD25dimFOXP3 cells (Fig. 6E).

These findings show that newly generated CD4+CD25+FOXP3+ T cells by TSLP-pDCs possess some of the key features of naturally occurring TR developed in mice and human thymus.

We have recently found that human thymus contains two distinct FOXP3+ TR subsets defined by surface expression of the costimulatory molecule ICOS (27). Although the ICOS+FOXP3+ TR produced high levels of IL-10 that suppressed DC function and expressed low levels of LAP (TGF-β1) that directly suppressed T cell proliferation, the ICOSFOXP3+ TR expressed high levels of LAP (TGF-β1) but produced low levels of IL-10, suggesting that there may be different populations of APCs selecting the FOXP3+ TR. Thus, we next questioned whether there are phenotypic differences between TSLP-pDC–induced and TSLP-mDC–induced TR. Interestingly, the FOXP3+ TR generated by TSLP-pDCs expressed lower LAP (TGF-β) (Fig. 7A) and produced higher IL-10 levels (Fig. 7B, 7C) than did the FOXP3+ TR induced by TSLP-mDCs, suggesting that TSLP-pDCs and TSLP-mDCs may play different roles in TR development.

Unlike CD40L or ligands for different TLRs that all induce mDCs to produce Th1-polarizing cytokine IL-12, TSLP does not induce mDCs or pDCs to produce IL-12. We hypothesized that the absence of Th1-polarizing cytokines provided by TSLP-activated DCs may provide a permissive environment that allows the development of TR. To test this hypothesis, CFSE-labeled CD4+CD25FOXP3 thymocytes were cultured for 7 d with TSLP-pDCs or TSLP-mDCs in the presence or absence of IL-4, IFN-γ, IL-12, TNF-α, or TGF-β. IL-12 and IL-4 strongly inhibited the differentiation of CD4+CD25 thymocytes into CD4+CD25+FOXP3+ cells (Fig. 8A, 8B). However, IFN-γ and TNF-α did not inhibit the generation of FOXP3+ cells. Interestingly, TGF-β, which is known to induce FOXP3 in peripheral CD4+ T cells activated by DCs (Fig. 8C, 8D) (28) or by anti-CD3 and anti-CD28 (22, 29, 30) and seems to be important to convert peripheral T cells to FOXP3+ T cells and to maintain the expression of FOXP3 in peripheral TR, significantly inhibited the generation of FOXP3+ thymocytes induced by TSLP-pDCs. These data suggest that both Th1-polarizing cytokine IL-12 and Th2-polarizing cytokine IL-4 can inhibit the ability of TSLP-pDCs to induce the generation of FOXP3+ TR from thymocytes in culture and that the expression of FOXP3 in thymus and in periphery may be regulated by different molecular mechanisms.

We next investigated whether human thymic pDCs expressed TSLPR and searched for evidence of in vivo interaction between pDCs and FOXP3+ TR in situ. pDCs were enriched from total thymocytes by anti-BDCA4 magnetic bead sorting. The pDCs were further defined by a phenotype of linageCD4+CD123+ using flow cytometry. These cells showed the typical phenotype of peripheral pDCs, including expression of BDCA-2, BDCA-4, CD45RA, and IL-7Rα (Fig. 9A). Freshly isolated thymic pDCs also had the ability to produce IFN-α but not IL-12 and increased the expression of HLA-DR, CD80, and CD86 in response to HSV or CpGs (data not shown). As shown in Fig. 9B, significant numbers of thymic pDCs (5–20%) expressed TSLPR without any stimulation. Importantly, these TSLPR+ pDCs expressed higher levels of CD80, CD86, CD40, and HLA-DR than TSLP pDCs. These results clearly indicate that TSLPR is expressed on pDCs residing in human thymus.

To provide further anatomical evidence that pDCs are involved in FOXP3+ TR development in thymus, we investigated the localization of thymic pDCs and FOXP3+ TR in human thymus by microscopic analysis using specific Abs to pDCs (anti-CD123 mAb) and to FOXP3+ TR (anti-FOXP3 mAb). Both CD123+ pDCs and FOXP3+ TR were found in thymic medulla, and some pDCs were in close contact with FOXP3+ cells (Fig. 9C). These findings suggest that thymic pDCs as well as mDCs instructed by TSLP produced by Hassall’s corpuscles may have a critical role in inducing the generation and expansion of CD4+CD25+FOXP3+ TR through direct interaction in thymus.

Human thymus contains two major DC subsets, mDCs and pDCs (18, 19). We have previously shown that mDCs from human peripheral blood and thymus expressed the TSLPR/IL-7Rα complex and could be activated by TSLP (20). We presented experimental evidence suggesting that thymic mDCs educated by TSLP produced by the epithelial cells of the Hassall’s corpuscles may induce the differentiation of CD4+CD25 thymocytes into FOXP3+ TR within the medulla in human thymus (16).

In this study, we found that the subpopulation of pDCs isolated from human thymus expressed TSLPR at steady state. Furthermore, this pDC subpopulation is located close to and associated with FOXP3+ cells in the thymic medulla. We also found that pDCs in peripheral blood express both TSLPR and IL-7Rα after activation through TLRs. TSLP activated pDCs to express higher levels of costimulatory molecule and to produce chemokines CCL-17 and CCL-22, which are important in guiding the traffic of developing immature thymic T cells into the medulla (3133). We showed that TSLP-activated pDCs induced the expansion and differentiation of CD4+CD8CD25FOXP3 thymocytes into CD4+CD25+FOXP3+ TR. The TR generated by TSLP-pDCs possess a typical feature of naturally occurring TR as previously reported in mice and humans (3, 34).

Interestingly, the FOXP3+ TR induced by TSLP-pDCs showed distinct cytokine expression potential compared with the FOXP3+ TR induced by TSLP-mDCs. The TSLP-pDC–induced TR were IL-10high/TGF-βlow and the TSLP-mDC–induced TR were IL-10low/TGF-βhigh after activation. These data are consistent with our recent observation that human thymus, peripheral blood, and secondary lymphoid tissues contain two subsets of CD25highFOXP3+ TR according to ICOS expression (27). Whereas the ICOS+ TR subset had the potential to express high IL-10 and less TGF-β, the ICOS TR subset expressed high TGF-β but lower IL-10. These results suggest that pDCs and mDCs in thymus not only play a critical role in the selection of TR but may also imprint the two cell subsets of TR to produce different suppressive cytokines, IL-10 and TGF-β, in response to self-Ag in the periphery.

Previous studies have shown that circulating immature DCs may migrate into the thymus and induce nondeletional tolerance (3537). A more recent study by Proietto et al. (38) has demonstrated that the circulating peripheral Sirpα+CD8low DC subset migrates into the thymus and selects the FOXP3+ TR. Another interesting study provided direct evidence that pDCs are able to acquire and process allogeneic cell-derived Ags and to cross-present the alloantigen to induce FOXP3+ TR cells in periphery (39). In our study, only ∼4–20% of pDCs in the thymus expressed TSLPR. This together with the observation that the peripheral blood pDCs expressed TSLPR only following activation through TLR-7/TLR-9 suggests that peripheral pDCs that carry the peripheral Ags may migrate into the thymus upon activation and participate in the selection of TR together with TSLP-mDCs (16) and medullar epithelial cells (40).

Differentiation of T cells into distinct cell linages is controlled by several key transcription factors that have pivotal roles in cell fate choice during the early stage of lymphoid cell development. These linage specification factors not only promote a particular T cell fate but also are responsible for repressing alternative differentiation pathways. In our experiment, IL-12 and IL-4 actively repressed the adoption of CD4+ thymocytes’ fate to be TR. It seems that transcription factors induced by these cytokines, which are critical to promote Th1 or Th2 differentiation, override FOXP3-dependent development of TR. Our current study suggests that primary CD4+ thymocytes have less ability to produce cytokines, including IL-4 and IFN-γ, by stimulation with DCs or anti-CD3 plus anti-CD28 than do peripheral naive CD4+ T cells (S. Hanabuchi and Y.J. Liu, unpublished observation). Furthermore, both TSLP-pDCs and TSLP-mDCs produce little Th1-polarizing (IL-12) or Th2-polarizing (IL-4) cytokines that inhibit TR development. These features of DCs and thymocytes are likely to be essential for the differentiation of FOXP3+ TR in the thymic microenvironment. Although the existence of a unique niche in thymus has been proposed, its exact nature has remained unclear (40, 41). Decoding the molecular signature of DCs, including surface molecules and soluble factors such as cytokines that may positively or negatively regulate thymic selection, should be very important to accomplish the differentiation of TR in thymus. Thus, we speculate that TR differentiation in thymus is regulated in a very subtle way by DCs, providing a suitable niche that allows TR to develop.

In conclusion, this study provides experimental evidence showing that activated pDCs rapidly express TSLPR and IL-7Rα and that TSLP could strongly activate human pDCs to enhance costimulatory molecules and to initiate production of CCL-17 and CCL-22. Furthermore, TSLP-activated pDCs could efficiently induce the generation and expansion of FOXP3+ TR. Interestingly, the FOXP3+ TR induced by TSLP-pDCs produced more IL-10 and less TGF-β than that of the FOXP3+ TR induced by TSLP-mDCs. These data suggest that two subsets of TR with different cytokine production potential can be selected by TSLP-activated mDC or pDCs, respectively, in thymus. The inability of TSLP to induce DC maturation without producing Th1- or Th2-polarizing cytokines may provide a thymic niche for TR development.

We thank Karen Ramirez, Zhiwei He, and Eric Wieder for cell sorting and support and M.J. Fingold and K. Sternberg for tissue materials.

Disclosures The authors have no financial conflicts of interest.

This work was supported by National Institutes of Health Grant AI062888-01 (to Y.-J.L.) and the Keck Foundation (to Y.-J. L.).

Abbreviations used in this paper:

BDCA

blood dendritic cell Ag

DC

dendritic cell

FOXP3+

Forkhead Box P3+

IL-7Rα

IL-7 receptor α

LAP

latency associated peptide

mDC

myeloid dendritic cell

MFI

mean fluorescence intensity

MHC-II

MHC class II

pDC

plasmacytoid dendritic cell

TR

regulatory T cell

TSLP

thymic stromal lymphopoietin

TSLPR

thymic stromal lymphopoietin receptor.

1
Modigliani
Y.
,
Thomas-Vaslin
V.
,
Bandeira
A.
,
Coltey
M.
,
Le Douarin
N. M.
,
Coutinho
A.
,
Salaün
J.
.
1995
.
Lymphocytes selected in allogeneic thymic epithelium mediate dominant tolerance toward tissue grafts of the thymic epithelium haplotype.
Proc. Natl. Acad. Sci. USA
92
:
7555
7559
.
2
Modigliani
Y.
,
Coutinho
A.
,
Pereira
P.
,
Le Douarin
N.
,
Thomas-Vaslin
V.
,
Burlen-Defranoux
O.
,
Salaün
J.
,
Bandeira
A.
.
1996
.
Establishment of tissue-specific tolerance is driven by regulatory T cells selected by thymic epithelium.
Eur. J. Immunol.
26
:
1807
1815
.
3
Sakaguchi
S.
2004
.
Naturally arising CD4+ regulatory t cells for immunologic self-tolerance and negative control of immune responses.
Annu. Rev. Immunol.
22
:
531
562
.
4
Shevach
E. M.
,
McHugh
R. S.
,
Piccirillo
C. A.
,
Thornton
A. M.
.
2001
.
Control of T-cell activation by CD4+ CD25+ suppressor T cells.
Immunol. Rev.
182
:
58
67
.
5
Bensinger
S. J.
,
Bandeira
A.
,
Jordan
M. S.
,
Caton
A. J.
,
Laufer
T. M.
.
2001
.
Major histocompatibility complex class II-positive cortical epithelium mediates the selection of CD4(+)25(+) immunoregulatory T cells.
J. Exp. Med.
194
:
427
438
.
6
Fontenot
J. D.
,
Rudensky
A. Y.
.
2005
.
A well adapted regulatory contrivance: regulatory T cell development and the Forkhead family transcription factor FOXP3.
Nat. Immunol.
6
:
331
337
.
7
Le Douarin
N. M.
,
Jotereau
F. V.
.
1975
.
Tracing of cells of the avian thymus through embryonic life in interspecific chimeras.
J. Exp. Med.
142
:
17
40
.
8
Guillemot
F. P.
,
Oliver
P. D.
,
Peault
B. M.
,
Le Douarin
N. M.
.
1984
.
Cells expressing Ia antigens in the avian thymus.
J. Exp. Med.
160
:
1803
1819
.
9
Fontaine-Perus
J. C.
,
Calman
F. M.
,
Kaplan
C.
,
Le Douarin
N. M.
.
1981
.
Seeding of the 10-day mouse embryo thymic rudiment by lymphocyte precursors in vitro.
J. Immunol.
126
:
2310
2316
.
10
Houssaint
E.
,
Torano
A.
,
Ivanyi
J.
.
1986
.
Split tolerance induced by chick embryo thymic epithelium allografted to embryonic recipients.
J. Immunol.
136
:
3155
3159
.
11
Salaün
J.
,
Bandeira
A.
,
Khazaal
I.
,
Calman
F.
,
Coltey
M.
,
Coutinho
A.
,
Le Douarin
N. M.
.
1990
.
Thymic epithelium tolerizes for histocompatibility antigens.
Science
247
:
1471
1474
.
12
Ramsdell
F.
,
Lantz
T.
,
Fowlkes
B. J.
.
1989
.
A nondeletional mechanism of thymic self tolerance.
Science
246
:
1038
1041
.
13
Laufer
T. M.
,
DeKoning
J.
,
Markowitz
J. S.
,
Lo
D.
,
Glimcher
L. H.
.
1996
.
Unopposed positive selection and autoreactivity in mice expressing class II MHC only on thymic cortex.
Nature
383
:
81
85
.
14
Laufer
T. M.
,
Fan
L.
,
Glimcher
L. H.
.
1999
.
Self-reactive T cells selected on thymic cortical epithelium are polyclonal and are pathogenic in vivo.
J. Immunol.
162
:
5078
5084
.
15
Fontenot
J. D.
,
Rasmussen
J. P.
,
Williams
L. M.
,
Dooley
J. L.
,
Farr
A. G.
,
Rudensky
A. Y.
.
2005
.
Regulatory T cell lineage specification by the Forkhead transcription factor fOXP3.
Immunity
22
:
329
341
.
16
Watanabe
N.
,
Wang
Y. H.
,
Lee
H. K.
,
Ito
T.
,
Wang
Y. H.
,
Cao
W.
,
Liu
Y. J.
.
2005
.
Hassall’s corpuscles instruct dendritic cells to induce CD4+CD25+ regulatory T cells in human thymus.
Nature
436
:
1181
1185
.
17
Watanabe
N.
,
Hanabuchi
S.
,
Soumelis
V.
,
Yuan
W.
,
Ho
S.
,
de Waal Malefyt
R.
,
Liu
Y. J.
.
2004
.
Human thymic stromal lymphopoietin promotes dendritic cell-mediated CD4+ T cell homeostatic expansion.
Nat. Immunol.
5
:
426
434
.
18
Bendriss-Vermare
N.
,
Barthélémy
C.
,
Durand
I.
,
Bruand
C.
,
Dezutter-Dambuyant
C.
,
Moulian
N.
,
Berrih-Aknin
S.
,
Caux
C.
,
Trinchieri
G.
,
Brière
F.
.
2001
.
Human thymus contains IFN-alpha-producing CD11c(-), myeloid CD11c(+), and mature interdigitating dendritic cells.
J. Clin. Invest.
107
:
835
844
.
19
Wu
L.
,
Shortman
K.
.
2005
.
Heterogeneity of thymic dendritic cells.
Semin. Immunol.
17
:
304
312
.
20
Soumelis
V.
,
Reche
P. A.
,
Kanzler
H.
,
Yuan
W.
,
Edward
G.
,
Homey
B.
,
Gilliet
M.
,
Ho
S.
,
Antonenko
S.
,
Lauerma
A.
, et al
.
2002
.
Human epithelial cells trigger dendritic cell mediated allergic inflammation by producing TSLP.
Nat. Immunol.
3
:
673
680
.
21
Ito
T.
,
Kanzler
H.
,
Duramad
O.
,
Cao
W.
,
Liu
Y. J.
.
2006
.
Specialization, kinetics, and repertoire of type 1 interferon responses by human plasmacytoid predendritic cells.
Blood
107
:
2423
2431
.
22
Chen
W.
,
Jin
W.
,
Hardegen
N.
,
Lei
K. J.
,
Li
L.
,
Marinos
N.
,
McGrady
G.
,
Wahl
S. M.
.
2003
.
Conversion of peripheral CD4+CD25- naive T cells to CD4+CD25+ regulatory T cells by TGF-beta induction of transcription factor FOXP3.
J. Exp. Med.
198
:
1875
1886
.
23
Salomon
B.
,
Lenschow
D. J.
,
Rhee
L.
,
Ashourian
N.
,
Singh
B.
,
Sharpe
A.
,
Bluestone
J. A.
.
2000
.
B7/CD28 costimulation is essential for the homeostasis of the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes.
Immunity
12
:
431
440
.
24
Jordan
M. S.
,
Boesteanu
A.
,
Reed
A. J.
,
Petrone
A. L.
,
Holenbeck
A. E.
,
Lerman
M. A.
,
Naji
A.
,
Caton
A. J.
.
2001
.
Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide.
Nat. Immunol.
2
:
301
306
.
25
Malek
T. R.
,
Yu
A.
,
Vincek
V.
,
Scibelli
P.
,
Kong
L.
.
2002
.
CD4 regulatory T cells prevent lethal autoimmunity in IL-2Rbeta-deficient mice. Implications for the nonredundant function of IL-2.
Immunity
17
:
167
178
.
26
Tai
X.
,
Cowan
M.
,
Feigenbaum
L.
,
Singer
A.
.
2005
.
CD28 costimulation of developing thymocytes induces FOXP3 expression and regulatory T cell differentiation independently of interleukin 2.
Nat. Immunol.
6
:
152
162
.
27
Ito
T.
,
Hanabuchi
S.
,
Wang
Y. H.
,
Park
W. R.
,
Arima
K.
,
Bover
L.
,
Qin
F. X.
,
Gilliet
M.
,
Liu
Y. J.
.
2008
.
Two functional subsets of FOXP3+ regulatory T cells in human thymus and periphery.
Immunity
28
:
870
880
.
28
Luo
X.
,
Tarbell
K. V.
,
Yang
H.
,
Pothoven
K.
,
Bailey
S. L.
,
Ding
R.
,
Steinman
R. M.
,
Suthanthiran
M.
.
2007
.
Dendritic cells with TGF-beta1 differentiate naive CD4+CD25- T cells into islet-protective FOXP3+ regulatory T cells.
Proc. Natl. Acad. Sci. USA
104
:
2821
2826
.
29
Rao
P. E.
,
Petrone
A. L.
,
Ponath
P. D.
.
2005
.
Differentiation and expansion of T cells with regulatory function from human peripheral lymphocytes by stimulation in the presence of TGF-beta.
J. Immunol.
174
:
1446
1455
.
30
Rubtsov
Y. P.
,
Rudensky
A. Y.
.
2007
.
TGFbeta signalling in control of T-cell-mediated self-reactivity.
Nat. Rev. Immunol.
7
:
443
453
.
31
Campbell
J. J.
,
Pan
J.
,
Butcher
E. C.
.
1999
.
Cutting edge: developmental switches in chemokine responses during T cell maturation.
J. Immunol.
163
:
2353
2357
.
32
Chantry
D.
,
Romagnani
P.
,
Raport
C. J.
,
Wood
C. L.
,
Epp
A.
,
Romagnani
S.
,
Gray
P. W.
.
1999
.
Macrophage-derived chemokine is localized to thymic medullary epithelial cells and is a chemoattractant for CD3(+), CD4(+), CD8(low) thymocytes.
Blood
94
:
1890
1898
.
33
Annunziato
F.
,
Romagnani
P.
,
Cosmi
L.
,
Beltrame
C.
,
Steiner
B. H.
,
Lazzeri
E.
,
Raport
C. J.
,
Galli
G.
,
Manetti
R.
,
Mavilia
C.
, et al
.
2000
.
Macrophage-derived chemokine and EBI1-ligand chemokine attract human thymocytes in different stage of development and are produced by distinct subsets of medullary epithelial cells: possible implications for negative selection.
J. Immunol.
165
:
238
246
.
34
Sakaguchi
S.
2005
.
Naturally arising FOXP3-expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self.
Nat. Immunol.
6
:
345
352
.
35
Donskoy
E.
,
Goldschneider
I.
.
2003
.
Two developmentally distinct populations of dendritic cells inhabit the adult mouse thymus: demonstration by differential importation of hematogenous precursors under steady state conditions.
J. Immunol.
170
:
3514
3521
.
36
Beschorner
W. E.
,
Yao
X.
,
Divic
J.
.
1995
.
Recruitment of semiallogeneic dendritic cells to the thymus during post-cyclosporine thymic regeneration.
Transplantation
60
:
1326
1330
.
37
Goldschneider
I.
,
Cone
R. E.
.
2003
.
A central role for peripheral dendritic cells in the induction of acquired thymic tolerance.
Trends Immunol.
24
:
77
81
.
38
Proietto
A. I.
,
van Dommelen
S.
,
Zhou
P.
,
Rizzitelli
A.
,
D’Amico
A.
,
Steptoe
R. J.
,
Naik
S. H.
,
Lahoud
M. H.
,
Liu
Y.
,
Zheng
P.
, et al
.
2008
.
Dendritic cells in the thymus contribute to T-regulatory cell induction.
Proc. Natl. Acad. Sci. USA
105
:
19869
19874
.
39
Ochando
J. C.
,
Homma
C.
,
Yang
Y.
,
Hidalgo
A.
,
Garin
A.
,
Tacke
F.
,
Angeli
V.
,
Li
Y.
,
Boros
P.
,
Ding
Y.
, et al
.
2006
.
Alloantigen-presenting plasmacytoid dendritic cells mediate tolerance to vascularized grafts.
Nat. Immunol.
7
:
652
662
.
40
Aschenbrenner
K.
,
D’Cruz
L. M.
,
Vollmann
E. H.
,
Hinterberger
M.
,
Emmerich
J.
,
Swee
L. K.
,
Rolink
A.
,
Klein
L.
.
2007
.
Selection of FOXP3+ regulatory T cells specific for self antigen expressed and presented by Aire+ medullary thymic epithelial cells.
Nat. Immunol.
8
:
351
358
.
41
Fontenot
J. D.
,
Dooley
J. L.
,
Farr
A. G.
,
Rudensky
A. Y.
.
2005
.
Developmental regulation of FOXP3 expression during ontogeny.
J. Exp. Med.
202
:
901
906
.