CD11cCRE mice express the Cre-recombinase under the CD11c promoter (19). In RelB^fl/fl mice, the exon 4 of the relB gene is flanked by loxP sites (20). RelB^fl/fl mice were mated with CD11cCRE mice to investigate the effects of a conditional deletion of RelB in DCs. Resulting CD11cCRE × RelB^fl/fl (subsequently referred to as RelB^DCko) mice and corresponding control mice (RelB^fl/fl or CD11cCRE) were used at 5–12 weeks of age, if not otherwise indicated. C57BL/6 mice were purchased from Charles River (Sulzfeld, Germany) and bred in our respective facilities. OT-II mice were kindly provided by Francis Carbone, Melbourne, Australia and were crossed for some experiments with RAG1^−/− mice (gift from Thomas Winkler, University of Erlangen, Germany). Animal care, housing, and all experiments were performed according to institutional guidelines by the Animal Ethics Committee of the local authorities in Würzburg, Jena and New York, with age- and sex-matched animals.

Skin-draining lymph nodes (inguinal, axillary, brachial, and popliteal), spleens, and thymi were cut into small pieces by using forceps and digested for 20 min at RT with 1 mg/ml DNase I (Roche) and 1 mg/ml collagenase III (Worthington) in RPMI 1640 medium supplemented with 10% FCS, 2 mM l-glutamine (PAA), 100 U/ml penicillin (PAA), and 100 µg/ml streptomycin (PAA). To disrupt multicellular complexes that include DC and T cells, 0.01 mM EDTA (Sigma) was added and incubation at RT was extended for 5 min. Cells were resuspended in ice cold PBS containing 5% FCS and 1 mM EDTA, transferred through a 70 µm cell strainer (BD) and counted. Splenic cells were additionally subjected to erythrocyte lysis before enzymatic digestion. BM-DC were prepared as described in detail before (21).

For cell surface staining the following monoclonal antibodies (mAbs) and conjugates from Biolegend were used: B220 (clone RA3-6B2), CD4 (clone RM4-5 and GK1.5), CD8α (clone 53-6.7), CD11c (clone N418), CD25 (clone PC61 and 7D4), CD40 (clone 3/23), CD80 (clone 16-10A1), CD86 (clone GL-1), CD103 (clone M290), GITR (clone DTA-1), MHC II (clone 2G9 and M5/114.15.2), PD-1 (clone J43), and PDCA-1 (clone 927). Antibodies for CD44 (clone IM7) and Vβ5.1/5.2 (clone MR9-4) were purchased from BD. If primary antibodies were biotinylated, the following secondary antibodies were used: streptavidin–FITC, –PE, –PECy7 (Biolegend), or streptavidin–Pacific Blue (Invitrogen).

Intracellular staining of Foxp3 (clone FJK-16s, eBiosciences), Helios (clone 22F4, Biolegend), and Ki67 (clone B56, BD) was performed in accordance with the manufacturer’s protocol of the Anti-Mouse Foxp3 Staining Set (eBiosciences). Flow cytometric staining of RelB was done by using first a polyclonal rabbit anti-mouse RelB antibody (C-19, Santa Cruz) followed by a goat anti-rabbit-DyLight488 antibody (Jackson) as described previously (15). For detection of intracellular Langerin, cells were fixed and permeabilized in Cytofix/Cytoperm (BD) for 20 min at RT, washed in Perm/Wash Buffer (BD), and incubated with a rat anti-mouse Langerin-AlexaFluor488 (clone 929F3.01, Dendritics) for 30 min at 4°C.

All flow cytometric analyses were performed on a FACSCalibur, FACSCanto II, or LSRII (BD) and data were analyzed by FlowJo Software (Tree Star).

Single cell suspensions from skin-draining lymph nodes were prepared as described above and resuspended at a density of 4 × 10⁶ cells per ml in complete RPMI 1640 medium. Cells were then stimulated with 10 ng/ml PMA (Sigma) and 1 µg/ml ionomycin (Sigma) in the presence of 5 µg/ml brefeldin A (Sigma) for 5 h at 37°C. After stimulation, cells were stained for surface markers (CD4, CD11c, CD25, and CD44), fixed in 2% formaldehyde for 20 min at RT, and stained intracellularly with an IL-2 specific antibody (clone JES6-5H4, Biolegend) in Perm Buffer (PBS supplemented with 0.5% saponin, 0.1% BSA, and 0.05% sodium azide) for 30 min at 4°C. Control straining was performed with a PE-conjugated rat IgG2b isotype.

Micro-osmotic pumps (ALZET, model 1002, Charles River) were filled with PBS or OVA_327-339 peptide, respectively, to deliver 10 µg antigen per day as described previously (12). To decrease the start-up time for continuous and reliable pumping, prefilled micro-osmotic pumps were placed in PBS for at least 6 h at 37°C before implantation. For subcutaneous implantation, a small incision was made on the back next to the hips, a subcutaneous pocket was formed, and the pump was inserted. The wound was closed by saturation and using AutoClip Wound Closure System (Charles River). Adoptive transfer of T cell was carried out 1 day after implantation.

Single cell suspensions of lymph nodes (pooled skin-draining and mesenteric lymph nodes) and spleens from WT, OT-II, and OT-II.Rag1^−/− mice were prepared as described above. CD4⁺ T cells were purified by negative magnetic separation using either the Mouse CD4 T Cell Enrichment Kit (StemCell) or the Mouse CD4⁺ T Cell Isolation Kit II (Miltenyi Biotec) according to the manufacturer’s instructions. For further isolation of CD25⁺ CD4⁺ T cells from WT and OT-II mice a positive magnetic separation was performed using Mouse CD25 Micro Beads (Miltenyi Biotec) in accordance to the recommendations of the manufacturer. The purity of the isolated T cell populations was assessed by flow cytometry (CD4⁺ > 90%, CD4⁺ CD25⁺ > 60%). For labeling with carboxyfluorescein diacetate succinimidyl ester (CFSE) cells were resuspended at a density of 1 × 10⁷ cells per ml in PBS containing 2.5 µM CFSE (Molecular Probes, Invitrogen) for 15 min at RT. Subsequently, cells were washed with complete RPMI 1640 medium, resuspended in an appropriate volume of PBS, and injected into the lateral tail vein of recipient mice. After the indicated time points recipient mice were sacrificed and cell suspensions from skin-draining lymph nodes and spleens were prepared. Cells were stained with antibodies for CD4 and Foxp3 and analyzed by flow cytometry.

To neutralize IL-2, mice received intraperitoneally 2 mg anti-IL-2 mAb (clone S4B6, BioXCell). Before injection and at the indicated time points after injection blood was taken from submandibular vein, stained for CD4, Foxp3, and Ki67, and analyzed by flow cytometry.

Total RNA was isolated from sorted T cells and DC using Micro Rneasy Kit (Qiagen) according to manufacturer’s instructions. mRNA was amplified (Nano or Pico AmpTec Express Art Kit), digested with DNAse I (MBI Fermentas), and reverse transcribed using iScript (Biorad) in accordance to the recommendations of the manufacturer. Quantitative expression analysis of IL-2, IL-7, IL-15, and RelB was determined with a iCycler iQ (Biorad) using the following primers: mIL-2-FOR TTTGAGTGCCAATTCGATGA, mIL-2-REV AGGGCTTGTTGAGATGATGC, mIL-7-FOR2 TCAGCATCGATGAATTGGAC, mIL-7-REV2 CCAGTGTTTGTGTGCCTTGT, mIL-15-FOR CATTTTGGGCTGTGTCAGTG, mIL-15-REV TGCAACTGGGATGAAAGTCA, RT-RelB-F CCGAGCTAGGGGCCTTGGGTTCC, and RT-RelB-R AGCTCGATGGCGGGCAGGGTCTTG. qRT-PCR was performed in 25 µl SYBR Green Master mix (MBI Fermentas) containing 1 µl cDNA (undiluted or 1:10 diluted) and 0.2 µM of forward and reverse primers. The PCR program was: 95°C for 14 min and then 40 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s followed by a melting curve analysis. Relative quantification of the samples was performed by the comparative ΔΔ cycle threshold method. For normalization of the samples the housekeeping gene β-actin (β-Actin-F QRT CATTGCTGACAGGATGCAGA, β-Actin-R QRT TTGCTGATCCACATCTGCTG) was used.

To detect and calculate DC in skin, immunofluorescence stainings of MHC II⁺ cells were performed using either epidermal sheets or frozen cryostat sections of ear skin. Epidermal sheets were prepared as described previously (22). Briefly, ears were washed in 70% ethanol and dried at RT. Ventral and dorsal sides of each ear were separated with fine forceps and the sides with dermis down were floated on 0.5 M ammonium thiocyanate for 25 min at 37°C. Sides were washed with PBS, and the epidermis was peeled off the dermis, cut into small pieces, and fixed for 20 min with acetone at RT. After several washing steps (2× PBS, 2× 1% BSA/PBS), epidermal sheets were incubated with 10% BSA/PBS for 30 min to block unspecific binding of antibodies, stained with a pure rat anti-mouse MHC II (clone 2G9, BD) or a rat IgG2a isotype (BD) overnight at 4°C followed by a donkey anti-rat IgG F(ab′)2-AlexaFluor488 (Jackson) for 30 min at RT. Immunolabeled epidermal sheets were mounted in Vectashield Mounting Medium (Vector Laboratories) and viewed on a fluorescence microscope (DM IRE220, Leica). Skin cryostat sections (9 µm) were fixed with ice cold acetone for 15 min, dried for 30 min, and rehydrated in PBS followed by blocking in 10% BSA/PBS for 30 min at RT. Immunofluorescence staining and microscopically imaging of MHC II⁺ cells were performed as described for epidermal sheets.

Statistical analyses were performed using Prism 5.0 software (GraphPad Prism). The unpaired, two-tailed Student’s t-test was used, if data sets of two independent groups were normally distributed. The Mann–Whitney test was performed when data sets of two groups were not normally distributed.

Previously, we could show in RelB^+/− mice that ssmDCs of skin-draining PLN expressed the highest levels of RelB under steady-state conditions (15). Here, we tested whether DC-intrinsic deletion of RelB in RelB^DCko mice affected the ssmDC populations in PLNs. Intracellular FACS analysis of DCs from thymus, peripheral lymph nodes, and spleen confirmed highest RelB expression in PLNs and this could not be detected in RelB^DCko mice (Figures S1A,B in Supplementary Material), while similar levels of surface MHC II and CD80/CD86 costimulatory molecules were detected (Figures S2A,B in Supplementary Material). Further analyses of PLN DC indicated that the CD11c⁺ CD40^high ssmDCs but not the CD11c⁺ CD40^low resident DCs were increased in RelB^DCko mice (Figures 1A,B), while the total lymph node and spleen cell numbers remained unchanged (Figure S3 in Supplementary Material). In addition, also the skin distribution and frequency of DCs remained unaltered (Figure S4 in Supplementary Material). The ssmDCs in PLN homogenously express RelB (Figures S1C,D in Supplementary Material) and consist of three major subsets composed of Langerin⁺ CD103⁻ epidermal Langerhans cells (LCs), Langerin⁺ CD103⁺ dDCs, and Langerin⁻ CD103⁻ dDCs (Figure 1C) (23). When these ssmDCs were analyzed, we found increased proportions of LCs and Langerin⁻ dDCs, while Langerin⁺ dDCs remained unchanged in RelB^DCko mice (Figure 1D). Increased frequencies of DC subsets may result from enhanced proliferation or reduced apoptosis. However, ssmDCs did not show altered proliferation rates as detected by Ki67 staining (Figure S5A in Supplementary Material) or changes in Annexin V binding as a sign of apoptosis (Figure S5B in Supplementary Material). Thus, the reason for the selective increase in some but not all ssmDC subsets remains open.

Since we showed before that Langerin⁺ dDCs were the decisive subset of ssmDCs in PLNs that converted naive CD4⁺ T cells into Foxp3⁺ Tregs (15), we analyzed the Treg population of RelB^DCko mice. Frequencies and absolute numbers of CD4⁺ Foxp3⁺ Tregs were increased in PLNs of RelB^DCko mice, while CD4⁺ Foxp3⁻ conventional T cells (Tconv) were decreased (Figure 2A). This was not observed for total CD4⁺ or CD8⁺ T cells or B cells in PLNs or spleen (Figure S6 in Supplementary Material). The CD4⁺ Foxp3⁺ Tregs of RelB^DCko mice showed an activated phenotype by expressing significantly higher levels of PD-1 and Helios, while no or only trends of higher CD25, CD69, CD44, and GITR surface expression levels could be observed (Figure 2B). In addition, these activated Tregs also showed higher proliferation rates as indicated by elevated Ki67 expression (Figures 2C,D).

In order to test whether increased Treg frequencies could be a result of enhanced thymic output, the thymocyte frequencies of these mice were analyzed. The data indicate that already in the thymus CD4⁺ CD8⁻ Foxp3⁺ T cells appeared at a higher frequency and were proliferating whereas other thymocyte populations remained normal (Figures S7A–E in Supplementary Material). These data indicate that the frequency of proliferating nTregs is increased in thymus and PLNs of RelB^DCko mice.

Although the results may point to the thymic output as the responsible mechanism for a higher Treg frequency observed in spleen and lymph nodes, these data cannot explain the continuation of proliferation of Tregs in the PLN and spleen. Furthermore, it remained unclear whether the Treg increase was due to enhanced iTreg generation or expansion of nTregs and whether the effects depend on self-antigen. To clarify these points adoptive transfer experiments were performed using T cells from naive CD4⁺ OT-II × RAG1^−/− mice. These mice lack both the endogenous repertoire of T cells and CD4⁺ Foxp3⁺ transgenic OT-II Tregs. Thus, the transferred cells constituted exclusively of naive CD4⁺ Foxp3⁻ transgenic OT-II T cells and the appearance of CD4⁺ Foxp3⁺ transgenic OT-II Tregs cells would indicate de novo conversion of naive CD4⁺ OT-II T cells into iTregs. After transfer of naive CD4+ OT-II × RAG1^−/− T cells into control or RelBDCko mice OVA-loaded osmotic minipumps were implanted under the skin as a source of soluble peripheral self-antigen (15). The results showed a lower antigen-specific iTreg conversion into CD4⁺ Foxp3⁺ Tregs (Figures 3A,B) and Tconv a trend for an increased proliferation rate in RelB^DCko mice (Figure 3C). Importantly, de novo conversion of naive T cells into iTregs is not increased in RelB^DCko mice.

Conversely, adoptively transferred thymus-derived nTregs can be tested for their antigen-dependent expansion in peripheral lymph nodes. To distinguish whether only self-antigen-specific nTreg can be stimulated in lymph nodes or also nTreg with foreign antigen specificity we used two different settings of nTreg transfer. In naive Wt mice, the CD4⁺ CD25⁺ nTregs are predominantly specific for self-antigens (24). After transfer of self-antigen-specific nTregs we detected their enhanced proliferation in PLN of RelB^DCko but not control mice or the spleen (Figure 4A). Since this increased nTreg proliferation correlated with increased frequencies of self-antigen transporting LCs and CD103⁻ Langerin⁻ dDCs we hypothesized that antigen presentation of these two DC subsets may play a role.

In the second setting, we transferred nTregs from OT-II mice that are specific for foreign OVA antigen and express a high affinity TCR, since OVA is not expressed in the thymus of OT-II mice. Then PBS- or OVA-loaded minipumps as a source of antigen were provided. Here, an increased expansion of CD4⁺ CD25⁺ nTregs from OT-II mice could be observed in mice carrying OVA-pumps, indicating a supporting role of antigen for nTreg proliferation (Figure 4B). However, a higher expansion rate of nTregs in RelB^DCko mice than in control mice could not be observed after the implantation of OVA-loaded minipumps (Figure 4B). The results above (Figure 3) and our previous findings showed that the Langerin⁺ dDCs were responsible for OVA transport and presentation as well as de novo conversion of iTregs (15). These results are consistent with the fact that in PLNs of RelB^DCko mice with the frequency of CD103⁺ Langerin⁺ dDC was not increased also the iTreg conversion rate was not enhanced. In contrast, increased frequencies of CD103⁻ Langerin⁻ dDCs and LCs in PLNs of RelB^DCko mice correlated with an enhanced proliferation rate of nTreg. These data may indicate a division of labor among ssmDC subsets from the skin for iTreg conversion by CD103⁺ Langerin⁺ dDCs and nTreg stimulation by CD103⁻ Langerin⁻ dDCs and/or LCs. Although the role of LCs cannot be clearly deducted from our experiments, others have shown that resting LCs in the epidermis stimulate skin resident Treg activation and proliferation (25) and that glucocorticoids further promote TGF-β production by LCs to enhance Treg numbers in contact hypersensitivity patients (26).

Together, these data obtained from RelB^DCko mice indicate that beyond the thymic selection also self-antigen presentation by Langerin⁻ dDCs and/or LCs in PLNs regulates the pool of self-antigen-specific nTregs.

During the steady state, the niche size for nTreg in the thymus is controlled by the availability of IL-2 provided by DCs (27). Also, peripheral maintenance of nTregs is dependent on IL-2 (28, 29), which is secreted from presumably self-reactive CD4⁺ CD25^low non-Tregs (30). Here, we sought to identify the source of increased IL-2-competent cells in PLNs by real-time PCR and FACS analyses in RelB^DCko mice. Total PLN cells were stimulated with PMA + ionomycin and the only IL-2 producing cells were identified as CD11c⁺ DC (Figures 5A,B) and CD4⁺ CD25^low CD44^high Tml (Figures 5C,D). An increase in the frequency of IL-2 producing cells in RelB^DCko mice was observed only for the Tml but not the DCs (Figures 5A–D) together with the total frequency of CD4⁺ CD44^high T cells in PLNs of RelB^DCko mice (Figure 5E). These data suggest that IL-2⁺ Tml and not DCs may contribute to the elevated Treg proliferation in RelB^DCko mice. To substantiate the results, we sorted the ssmDC populations from PLN of RelB^DCko and control mice and tested their IL-2, IL-7, and IL-15 production at real-time PCR level. These data indicate that IL-2 is not produced by DCs in PLNs and no differences were found for the other cytokines between RelB^DCko and control mice (Figure 5F). To demonstrate that IL-2 is functionally required for Treg proliferation in the RelB^DCko mice we injected a blocking anti-IL-2 antibody (clone S4B6) and found that proliferating Ki67⁺ CD4⁺ Foxp3⁺ Treg but not Ki67⁺ CD4⁺ Foxp3⁻ Tconv frequencies in blood dropped with a maximum between day 4 and 7 after injection and recovered thereafter (Figure 5G). In contrast to Treg and Tconv, Tml frequencies increased at days 4 and 7. Thus, the CD4⁺ Tml population in RelB^DCko mice is not only increased in their cellular frequency but also the proportion of IL-2 producers among them is elevated as compared to control mice, thereby enabling higher nTreg proliferation and pool sizes in PLNs of RelB^DCko mice.