Ultraviolet B–Induced Maturation of CD11b-Type Langerin⁻ Dendritic Cells Controls the Expansion of Foxp3⁺ Regulatory T Cells in the Skin

Ultraviolet B irradiation is known to induce immune tolerance and has been proven to be clinically effective for immunological skin diseases such as psoriasis (1–3). After UVB exposure, thymus-derived Foxp3⁺ regulatory T cells (tTreg) with Treg-specific CpG hypomethylation, which expanded to ∼60% of the CD4⁺ T cells in the skin, formed clusters with dendritic cells (DCs) (4). This indicates involvement of DCs in the expansion of Treg cells. It remains obscure, however, whether skin DCs are involved in UVB-induced immune tolerance.

DCs, which play critical roles in linking innate and adaptive immunity, are heterogeneous and divided into several subsets with distinctive functions (5). CD8⁺DEC205⁺ DCs or Langerin⁺ DCs are able to induce Treg cells from naive CD4⁺ T cells in the steady-state (6–8). In the skin, both Langerhans cells (LCs) and Langerin⁺ dermal DCs have been shown to bear this capacity (9–11). LCs also contribute to Treg induction after ionizing radiation or UVB exposure (12–14). However, despite these studies, which mainly investigated the role of DC subsets for Treg induction in hapten-applied draining lymph nodes (DLN), it remains unclear how DC subsets in the skin elicit Treg expansion after UVB exposure without exogenous Ag, contributing to immunological homeostasis and tolerance in the skin.

In this study, we investigated the roles of DC subsets for UVB-mediated immune tolerance and found that it was not LCs or Langerin⁺ dermal DCs, but Langerin⁻ DCs that are required for Treg expansion in UVB-exposed skin. Upon UVB exposure, Langerin⁻ DCs, which were CD11b⁺, upregulated the expression of CD86 along with several tolerogenic genes related to Treg function and proliferation. Furthermore, the Langerin⁻ DCs from UVB-exposed skin induced Treg proliferation in vitro without Ag, suggesting that the Langerin⁻ DCs would present self-antigens released from UVB-damaged skin for Treg expansion.

C57BL/6 mice were obtained from Japan SLC (Shizuoka, Japan). We received gifts of Langerin–diphtheria toxin receptor (DTR) mice (15) from Dr. B. Malissen (Centre National de la Recherche Scientifique UMR, Marseille, France) and CD11c-Cre mice (16) from Dr. B. Reizis (Columbia University, New York, NY). To obtain CD11c–DT A subunit (DTA) mice, CD11c-Cre mice were crossed with R26:lacZbpA^flox DTA mice carrying the DTA gene, which was designed to be expressed specifically in the cells or tissues expressing Cre recombinase under the control of the ubiquitously active Rosa26 promoter (16–18). Mice were maintained at the Nagoya City University Animal Facility and the Osaka University Animal Facility. All mice were kept under specific pathogen-free conditions. The Institutional Animal Care and Use Committees of Nagoya City University and Osaka University approved this study. A single 1-μg dose of DT can deplete Langerin⁺ dermal DCs for ∼2 wk and LCs for ∼4 wk (15, 19–21). Using flow cytometry, we confirmed that both epidermal LCs and Langerin⁺ dermal DCs were successfully decreased at day 1 and day 8 after one shot of DT injection (data not shown).

The anti-mouse Foxp3 (FJK-16s) staining kit and anti-CD86 Ab were purchased from eBioscience (San Diego, CA). Anti–MHC class II (MHC II)-FITC Ab was from BD Biosciences (San Diego, CA). Anti-Langerin Ab was from Dendritics (Lyon, France). Other Abs were purchased from BioLegend (San Diego, CA). CFSE, Live/Dead fixable aqua, and TOPRO-3 were from Molecular Probes (Eugene, OR). Mitomycin C (MMC) and DT were from Sigma-Aldrich (St. Louis, MO). Anti-CD4 and streptavidin microbeads were from Miltenyi Biotec (Bergisch Gladbach, Germany). Human IL-2 was from Chiron (Emeryville, CA).

UVB was produced by TL20/W12 lamps (Phillips, Eindhoven, the Netherlands), which emit most of their energy within the UVB range (290–320 nm; emission peak 313 nm). Mice were shaved and irradiated with systemic exposure of 500 mJ/cm² as reported previously (4).

Cells were isolated as previously reported (4, 19, 22–25). For skin DCs, epidermis and dermis from the ear were detached by trypsin or dispase treatment (Invitrogen). The epidermal cell suspension was collected using a cell strainer, whereas the dermis was further incubated with collagenase D (Roche, Basel, Switzerland) and digested with a gentleMACS dissociator (Miltenyi Biotec). Cell suspensions were stained with Abs specific for CD45.2 (104), CD11c (N418), MHC II (M5/114.15.2), CD11b (M1/70), CD86 (GL1), and Live/Dead fixable aqua. To assess DC cell numbers, all cells were acquired from each sample using a FACSVerse (BD Biosciences) flow cytometer and analyzed using FlowJo software (FlowJo, Ashland, OR).

The suppression assay was performed in vitro as previously described (4, 7, 26). Briefly, responder CD25⁻CD4⁺ T cells were negatively purified from CD45.1 mice by microbeads (Miltenyi Biotec), CFSE labeled, and stimulated with anti-CD3 (2C11) Ab and mitomycin C–treated spleen (Sp) APCs. CD25^high+CD4⁺ Treg cells were purified by a FACSAria II (BD Biosciences). After 3 d, the cells were stained with Abs specific for CD45.1 (A20), CD4 (RM4-5), and Live/Dead fixable aqua or TOPRO-3. Dead cells were gated out by Live/Dead fixable aqua or TOPRO-3.

The Treg and DC coculture was performed as previously described (27). CD25^high+CD4⁺ Treg cells (1 × 10⁴ cells per well) were sorted from the DLN and Sp of naive CD45.1 mice as in Fig. 2A, and then CFSE labeled and cultured with sorted skin DC subsets (2000–3500 cells per well) from Langerin-DTR mice in the presence of 10 U/ml IL-2. After 3 d, the cells were analyzed for CFSE dilution. Cells were stained with Abs specific for CD45.1 (A20), CD4 (RM4-5), Foxp3 (FJK-16s), and Live/Dead fixable aqua. Dead cells were gated out by Live/Dead fixable aqua.

Epidermal sheets were obtained from ear, back, or ventral skin and prepared by ammonium thiocyanate treatment as previously described (28, 29). Back skin sections were stained as previously described (4). Anti–Foxp3 (FJK-16s)-Alexa Fluor 647, anti–CD4 (RM4-5)-Pacific Blue, anti–MHC II (2G9)-FITC, and anti–Langerin (929F3.01)-Alexa Fluor 647 Abs were used. We used an FV1000 confocal microscope (Olympus, Tokyo, Japan) with ×40 or ×60 objective lenses and FV10-ASW4.2a data acquisition software (Olympus). The line between epidermis and dermis was added using Adobe Photoshop CS6. The brightness of the Alexa Fluor 647 signal was increased equally in all images using FV10-ASW4.2a data acquisition software.

We used human CTLA-4–Ig fusion protein (CTLA4-Ig) because it has been reported that human CTLA4-Ig is effective in mice as well as humans (30–32). Mice were i.p. injected with 25 mg/kg CTLA4-Ig (Abatacept; Bristol-Myers Squibb, New York, NY) on day 0 and day 2 after UVB exposure.

We used the protocol reported by Naik et al. (23). Mice were i.p. injected with 2 mg of anti–CSF1R Ab (clone AFS98) from Bio X Cell (Lebanon, NH) on days −5 and −3. Then, 0.5 mg of anti-CSF1R Ab was injected on days −2, −1, 0, 2, 4, and 6. Control mice were injected with same dose of rat IgG (Sigma-Aldrich) or PBS on the same days. Mice were irradiated with UVB on day 0 and analyzed on day 7.

Skin DC subsets from UVB-exposed skin were purified as in Fig. 4A and lysed with RLT buffer (Qiagen, Hilden, Germany). Reverse transcription was performed using SMART-Seq v4 Ultra low input RNA kit for sequencing (Clontech, Mountain View, CA). cDNAs were then fragmented by a S22 Focused-ultrasonicator (Covaris, Wobrun, MA), subjected to library preparation with a KAPA Library Preparation Kit for Ion Torrent (KAPA Biosystems, Wilmington, MA) according to the manufacturer’s instructions, and sequenced using Ion Proton (Thermo Fisher Sceintific, Waltham, MA). Sequences were mapped to mm9 using hisat2 and bowtie2 (two-step mapping). Normalized fragments per kilobase per million reads were generated with Cuffnorm. All RNA sequencing data have been deposited in the DNA Data Bank of Japan database under accession number SSUB006796 (http://www.ddbj.nig.ac.jp/intro-e.html).

All statistical analyses were performed using Prism (GraphPad Software, CA). The p values were obtained from a paired or unpaired Student t test. A p value <0.05 was considered significant. In the graphs, horizontal bars indicate mean values and vertical bars signify SD.

To test whether classical DCs were required for UVB-induced Treg expansion, we first used CD11c-DTA mice (16–18), which constitutively lack classical CD11c⁺ DCs. Treg frequency was not reduced in the skin without UVB exposure (Fig. 1A), presumably because approximately half of skin DCs were retained in the CD11c-DTA mice (Supplemental Fig. 1). By irradiating the mice with UVB on day 0, and analyzing Treg cells on day 7 when UVB-mediated Treg expansion peaked in the skin (4), we observed a significant reduction in Treg expansion in the skin (Fig. 1A). Although the Treg expansion was not completely blocked, as skin DCs were not completely depleted in the CD11c-DTA mice (Supplemental Fig. 1), the results indicate that classical CD11c⁺ DCs contribute to Treg expansion in UVB-exposed skin.

Next, given that LCs and Langerin⁺ DCs are able to induce Treg cells under several conditions (8–14), we attempted to determine possible roles of Langerin⁺ DCs in the expansion of Treg cells following UVB exposure. There are two kinds of Langerin-DTR mouse models to deplete Langerin⁺ DCs in the literature, and both of them have normal Foxp3⁺ Treg development and do not develop autoimmune diseases in the steady-state (15, 33). By use of Langerin-DTR mice, which are knocked in with a human DTR fused to enhanced GFP at the Langerin gene locus (15), we depleted Langerin⁺ DCs by DT treatment on day −1, prior to UVB irradiation on day 0. Epidermal sheet staining revealed depletion of LCs in the DT-injected mice (DT+) but not in PBS-injected control mice (DT−) (Fig. 1B, top). Seven days after UVB exposure, we found that Treg cells were expanded up to 50 to 60% of CD4⁺ T cells in the ear skin even in Langerin⁺ DC-depleted mice (Fig. 1C, DT+). As in the literature (15, 33), without UVB exposure, Langerin⁺ DC-depleted mice had similar numbers of Treg cells (data not shown). Foxp3⁺CD4⁺ T cells were detected by confocal microscopy in the dermis of both DT− and DT+ mice (Fig. 1D, Supplemental Fig. 2). Because the condition to acquire the images in Supplemental Fig. 2 was adjusted to the brightness of MHC II⁺ cells in the dermis, there were almost no MHC II⁺ signals in the epidermis in contrast to the epidermal sheet staining (Fig. 1B). However, when the brightness signal for MHC II was increased equally throughout the image, MHC II⁺ LCs were observed in the epidermis (data not shown). We confirmed that MHC II expression on LCs was lower than MHC II⁺CD11c⁺ DCs by flow cytometry (Supplemental Fig. 3). The frequency of Treg cells in the DLN and Sp was similar between DT− and DT+ mice (Fig. 1C). Epidermal sheet staining and flow cytometry confirmed continued reduction of LCs and Langerin⁺ dermal DCs in the ear skin 7 d after UVB exposure in the DT+ mice (Fig. 1B, Supplemental Fig. 4A). To deplete Langerin⁺ cells completely, we additionally injected Langerin-DTR mice with DT on days 2 and 5 after UVB exposure, but failed to see a significant reduction of Treg expansion (Supplemental Fig. 4B).

To determine then whether Treg cells from the UVB-exposed Langerin⁺ DC-depleted mice were functional, we performed an in vitro suppression assay using CD25^high+CD4⁺ Treg cells from the DLN in DT- or PBS-treated Langerin-DTR mice. When sorted CD25^high+CD4⁺ Treg cells (>90% Foxp3⁺ as in Fig. 2A) were added to CFSE-labeled responder CD25⁻CD4⁺ T cells, the proliferation of responder cells was blocked similarly by CD25^high+CD4⁺ Treg cells from DT− or DT+ mice (Fig. 2B). CD25^high+CD4⁺ Treg cells from the mice DT treated on days −1, 2, and 5 were also similarly suppressive compared with those from PBS-treated mice (data not shown).

Taken together, these results demonstrated that depletion of LCs or Langerin⁺ DCs scarcely affected the induction and function of UVB-expanded Treg cells, implicating the possible contribution of other types of CD11c⁺ DCs to the Treg induction.

To identify the critical CD11c⁺ DC subsets responsible for UVB-induced Treg expansion, we next investigated the composition of dermal DC subsets on day 6 after UVB exposure because the expanded Treg cells were located in the dermis on day 7 (Fig. 1C) (4). The frequency of Langerin⁺CD11c⁺ cells was reduced, whereas that of Langerin⁻CD11c⁺ cells was not (Fig. 3A).

In the dermis, CD11c⁺MHC II⁺ DCs are mainly divided into three subsets (23, 34): CD11b-type DCs are Langerin⁻CD11b^high+, LCs are Langerin⁺CD11b^inter+, and CD103-type DCs are Langerin⁺CD11b^low+ as gated in Fig. 3B. In UVB-irradiated skin, we found that the population of CD11b-type dermal DCs, which represent Langerin⁻CD11c⁺ dermal DCs, was increased in frequency, whereas LCs and CD103-type DCs were not (Fig. 3B), with an expansion of Langerin⁻CD11c⁺ or Langerin⁻MHC II⁺ cells in the epidermis (Fig. 3C, data not shown). Thus, by day 6 after UVB exposure, the frequency of Langerin⁻ DCs was significantly increased in both the epidermis and dermis, although total cell numbers were not significantly different. These results suggest that migratory Langerin⁺ DCs, such as LCs and CD103-type Langerin⁺ dermal DCs, migrate to the DLN after UVB exposure, as reported previously (35–38), whereas CD11b-type Langerin⁻ DCs are retained, or may be recruited to the skin.

It is known that CD11b-type dermal DCs, which represent Langerin⁻CD11c⁺ dermal DCs, are heterogeneous (39). Especially upon UVB exposure, Gr1 (Ly6C/Ly6G)^high+ monocyte-derived inflammatory DCs may be contained within this population (39, 40). To discriminate between populations, we further investigated Langerin⁻CD11c⁺ DCs in the UVB-exposed dermis and found that only a small fraction of them are monocyte-derived inflammatory DCs (2.802 ± 1.56%, n = 7) (Fig. 3D). Thus, even after UVB exposure, the major population in dermal Langerin⁻CD11c⁺ DCs was CD11b⁺Gr1⁻, which represented CD11b-type dermal DCs (Fig. 3D).

We previously reported that the expansion of Treg cells was controlled by CD86 on mature DCs (27, 41). This suggested the possibility that the CD86 was upregulated in dermal DCs following UVB exposure. Indeed, on day 6 after UVB exposure, among MHC II⁺CD11c⁺ DCs, MHC II⁺CD11c⁻ cells, and MHC II⁻ cells in the dermis, significant CD86 upregulation was only observed with MHC II⁺CD11c⁺ DCs (Fig. 4A). Within the MHC II⁺CD11c⁺ DCs, CD86 expression in CD11b-type dermal DCs, which were Langerin⁻MHC II⁺, was significantly increased after UVB exposure (Fig. 4B). Within epidermal DCs, we also found that CD86 expression in Langerin⁻MHC II⁺ DCs was highly upregulated after UVB exposure (Fig. 4C).

To determine the contribution of this CD86 upregulation to Treg expansion, we examined CD80/CD86 blockade by the use of CTLA4-Ig. Blocking CD80/CD86 completely reduced the UVB-induced Treg expansion in the skin (Fig. 4D). UVB-expanded Treg cells in the Sp and DLN were also significantly reduced by CTLA4-Ig treatment (Fig. 4D), suggesting the key roles CD80/CD86 play in maintaining Treg homeostasis (42, 43).

We previously reported that Treg cells proliferated in vitro in the presence of IL-2 and CD86^high+ mature DCs without exogenous Ags (27, 41). To investigate the actual capacity of Langerin⁻ DCs to expand Treg cells, CD25^high+CD4⁺ Treg cells sorted from naive mice were CFSE labeled and cocultured with DC subsets sorted from UVB-exposed skin. Skin DC subsets were sorted from UVB-exposed skin as in Fig. 5A. In the presence of a low concentration IL-2 without anti-CD3 stimulation, we found that Treg cells with high expression of Foxp3 more actively divided with Langerin⁻ DC subsets from UVB-exposed skin when compared with LCs (Fig. 5B). Taken together with our previous finding that UVB exposure expanded tTreg cells in the skin in situ (4), these results suggest that the Langerin⁻ DCs may present self-antigens in UVB-damaged skin to stimulate the expansion of tTreg cells, which are better at recognizing self-antigens.

Next, we analyzed gene expression of skin DC subsets sorted from UVB-exposed skin by RNA sequencing analysis. We found the induction of several genes associated with immunological tolerance and Treg proliferation (44, 45) in the UVB-exposed dermal Langerin⁻ DCs (Fig. 5C). It was noted that the UVB-exposed dermal Langerin⁻ DCs highly expressed amphiregulin (Areg) (46, 47) and 4-1BBL (Tnfsf9) (48), which are important for Treg function and proliferation. Moreover, the RNA sequencing data also showed that dermal Langerin⁻ DCs from UVB-exposed skin had unique gene expression profiles compared with Langerin⁺ DCs from UVB-exposed dermis and epidermis (Fig. 5D). They highly expressed Gzmc, Xcl1, Iftm1, Cd274 (PDL1), and chemokine receptors such as Ccr2 and Ccr7 (Fig. 5D). There were some genes related to inflammatory DCs such as Cd209a (49); however, the contamination of inflammatory DCs was only a small fraction as in Fig. 3D. The flow cytometry analysis confirmed that only Gr1⁻CD11b⁺Langerin⁻CD11c⁺ DCs in the UVB-exposed dermis, not Gr1⁺CD11b⁺ inflammatory DCs, highly expressed PDL1, 4-1BBL, CCR7, and CCR2 on their surface (Fig. 5E). Therefore, although they may be heterogeneous and a mixture of resident and migratory DCs, CD11b⁺Langerin⁻CD11c⁺ DCs in the UVB-exposed dermis have a unique gene expression profile and have the capacity to expand Treg cells.

Next, we tried to deplete dermal CD11b⁺ Langerin⁻CD11c⁺ DCs using anti-CSF1R Ab based on the protocol of Naik et al. (23). Anti-CSF1R Ab treatment induced a significant reduction of LCs and CD11b⁺F4/80⁺ macrophages in the UVB-exposed skin (Fig. 6A, 6B). However, we did not observe a significant reduction of CD11b⁺ Langerin⁻CD11c⁺ DCs in the dermis (Fig. 6C), and the Treg expansion was not significantly reduced (Fig. 6D). In contrast to steady-state, upon UVB exposure, anti-CSF1R Ab treatment might not be enough to reduce CD11b⁺Langerin⁻CD11c⁺ DCs in the dermis. Therefore, using anti-CSF1R Ab, we reconfirmed that LC depletion did not influence Treg expansion in the UVB-exposed skin.

Collectively, these results indicate that UVB-matured CD11b-type Langerin⁻ DCs contribute to the expansion of Treg cells in the skin. In the UVB-exposed skin, the dermal Langerin⁻ DCs expressed a tolerogenic gene signature compared with LCs. Recently, Price et al. (12) reported that radio-resistant LCs induced Treg cells upon exposure to ionizing radiation. In contrast to the response to ionizing radiation, in this study we show that Treg expansion was independent of LCs in the case of UVB exposure. Migratory skin DCs, including LCs, constantly migrate to DLN to present Ags and maintain tolerance (8). Although UVB induces migration of LCs from the skin to DLN (37, 38), LC depletion prior to UVB exposure did not result in any significant change in the frequency or activity of Treg cells from the DLN (Figs. 1, 2). It is reported that LCs are important for UV-induced tolerance (14, 50). However, Schwarz et al. (50) used a contact hypersensitivity model with adaptive transfer of whole splenocytes and lymph nodes from hapten-treated mice with or without UVB exposure. In contrast, in this study, we used UVB exposure without any hapten application or exogenous Ags and focused on expansion of Treg cells responding to self-antigens released from UVB-damaged skin. Therefore, the experimental systems and aims are different from those from Schwarz et al.

We showed that Treg expansion in the UVB-exposed skin was decreased in CD11c-DTA mice (Fig. 1A). In our CD11c-DTA mice, without UVB exposure, we observed a significant reduction of Treg cells in DLN and Sp, but not in the skin (Fig. 1A). Ohnmacht et al. (51) reported that their CD11c-DTA mice had constitutive loss of all classical DCs and LCs. Birnberg et al. (52) reported that their CD11c-DTA mice constitutively lacked classical DCs, but contained spared amount of LCs. Both CD11c-DTA mice contained normal Treg cells in the thymus and periphery (51, 52). Even though there were retained skin DCs in our CD11c-DTA mice, we observed significant reduction of Treg expansion in the skin after UVB exposure (Fig. 1A, top right). It contrasted to no reduction of Treg cells in Langerin⁺ DC-depleted mice after UVB exposure (Fig. 1C). Thus, these findings indicate that classical CD11c⁺ DCs, rather than Langerin⁺ DCs, play an important role in expanding Treg cells in the UVB-exposed skin.

We showed that the Langerin⁻ DCs, rather than LCs, from UVB-exposed skin were able to stimulate Foxp3^high+ Treg proliferation in the presence of low-concentration IL-2 without extra TCR stimulation (Fig. 5B). This indicates that the dermal CD11b-type Langerin⁻ DCs are specialized to present the self-antigens released by UVB-damaged skin and thereby expand Treg cells to dampen autoimmune response. Dermal CD11b-type Langerin⁻ DCs in the steady-state have been reported to induce Treg cells by producing retinoic acid (53). They are also the key APC for Treg cells and effector T cells when Ag is s.c. injected with IFA (54). Although dermal CD11b-type Langerin⁻ DCs are heterogeneous (55, 56), CD11b-type Langerin⁻ DCs appear to be the main APCs for self-antigen presentation in UVB-exposed skin.

The RNA sequencing analysis revealed that Langerin⁻ DCs from UVB-exposed dermis had unique gene expression profile such as chemokines, chemokine receptors, costimulatory molecules, Gzmc, Xcl1, and Ifitm1 (Fig. 5D). In protein level, PDL1, 4-1BBL, CCR2, and CCR7 were also highly expressed on CD11b⁺Langerin⁻CD11c⁺ DCs in the UVB-exposed dermis (Fig. 5E). It is possible that high PDL1 expression on CD11b⁺Langerin⁻CD11c⁺ DCs reduces the expansion of effector T cells and high 4-1BBL expression helps Treg expansion (48). CCR2 plays a role in development of monocyte-derived inflammatory DCs, whereas it is also expressed on conventional CD11b⁺ dermal DCs (39, 40). As CCR2 ligand is reported to be expressed in the skin upon UVB exposure (57), some CCR2⁺CD11b⁺Langerin⁻CD11c⁺ DCs may be recruited into UVB-exposed skin by CCR2. CCR7 is required for DCs to egress from the skin via lymphatic vessels and migrate into the DLN (58, 59). Thus, CD11b⁺Langerin⁻CD11c⁺ DCs that also express CCR7 can migrate to DLN at a later time point. Thus, CD11b⁺Langerin⁻CD11c⁺ DCs in the UVB-exposed dermis may be a mixture of resident and migratory DCs, but only a small fraction of them were Gr1 (Ly6C/Ly6G)^high+ inflammatory DCs as in Fig. 3D. Collectively, our finding in the present study is that CD11b⁺Langerin⁻CD11c⁺ DCs in the UVB-exposed dermis have a unique gene expression profile and play a role in expanding Treg cells in the skin.

We also found that CD11b⁺Langerin⁻CD11c⁺ DCs matured after UVB exposure with upregulation of CD86 (Fig. 4), which is important for Treg expansion. Damage-associated molecular patterns released from UVB-damaged skin may bind on pattern recognition receptors on DCs to induce maturation. For example, double-stranded self-RNA released from UVB-damaged skin could stimulate TLR-3 (60). Further investigation is required for the mechanisms of how CD11b⁺Langerin⁻CD11c⁺ DCs are matured upon UVB exposure.

Our work highlights the key role of dermal CD11b-type Langerin⁻ DCs in promoting Treg expansion in the skin upon UVB exposure. Further understanding of the functions of Langerin⁻ DCs in the UVB-exposed skin would enable expanding Ag-specific Treg cells for the treatment of autoimmunity, allergy, and graft rejection in the skin.

We thank Shigeru Ohshima and Yamami Nakamura for technical assistance, Bernard Malissen for Langerin DTR mice, Boris Reizis for CD11c-Cre mice, Kenji Kabashima for transferring Langerin DTR mice, and Guido Ferlazzo, Anthony Bonito, Kyoko Ikumi, Takuma Matoba, and Maiko Watanabe for critical reading of the manuscript.

The authors have no financial conflicts of interest.