Blood-derived dermal langerin⁺ dendritic cells survey the skin in the steady state

Injection of DT into Langerin-DTR/EGFP mice leads to elimination of langerin⁺ DCs in the absence of overt skin or systemic toxicity, and therefore can be used to examine the repopulation of langerin⁺ DCs. As previously described, DT eliminated the entire LC pool in the epidermis, in addition to langerin^loCD8⁺ DCs and langerin^hiCD8⁻ DCs in the skin DLNs (Fig. 1, A and B). LCs were totally eliminated 2 d after DT injection and remained absent >14 d after treatment in the epidermis (Fig. 1, A and B). Langerin⁺ DCs in the skin DLNs were also efficiently depleted after DT, but in contrast to LCs, their repopulation occurred much earlier (Fig. 1, A and B). Langerin⁺ DCs in the skin DLNs include CD8⁺ DCs and CD8⁻ DCs. Langerin^loCD8⁺ DCs recovered to normal levels in <5 d after DT administration. Their rapid repopulation after DT administration is consistent with the fact that langerin^loCD8⁺ DCs are thought to derive from circulating precursors with a turnover time of a few days (25, 26). In contrast, langerin⁺CD8⁻ DCs, a population that is thought to derive exclusively from epidermal LCs, started to repopulate the skin DLNs long before LCs reappeared in the epidermis (Fig. 1, A and B), accounting for 25–30% of total langerin⁺CD8⁻ DCs in normal untreated skin DLNs. Langerin⁺CD8⁻ DCs were not detected in nonskin DLNs, such as mesenteric LNs, after DT treatment (unpublished data). Based on these results, we conclude that a population of LC-independent langerin⁺ DCs, distinct from CD8⁺ DCs, exists in the skin DLNs. The level of langerin expression on repopulating DCs in the skin DLNs was measured by EGFP fluorescence and by langerin intracellular staining. LC-independent langerin⁺ DCs in the skin DLNs expressed higher levels of langerin compared with CD8⁺ DCs and similar levels compared with skin DLN langerin⁺ DCs in untreated mice (Fig. 1 C). Using immunostaining and confocal analysis of DT-treated skin DLNs, we were able to visualize in situ the presence of LC-independent langerin⁺ DCs in the T cell areas of the skin DLNs 2 wk after DT, a time when no LCs are present in the skin (Fig. 1 D).

Langerin⁺CD8⁻ DCs in the skin DLNs can derive either from the skin or from blood circulating precursors that enter directly into the LN from the blood. Because LCs are absent when langerin⁺CD8⁻ DCs reappear in the skin DLNs, we looked for langerin⁺ DCs in the dermis. We detected langerin⁺ DCs in the dermis long before LCs reappeared in the skin (Fig. 2, A and B). Dermal langerin⁺ DCs appeared around day 5 and increased up to day 10 after DT injection, after which they remained stable in the dermis, reaching 20% of total dermal langerin⁺ cells, 10% of total dermal DCs, and 2% of total dermal leukocytes.

The aforementioned results suggest that a population of langerin⁺ DCs, independent of LCs, resides in the dermis and in the skin DLNs. To directly examine the origin of skin DLN langerin⁺ DCs, we used parabiotic mice in which C57BL/6 CD45.2⁺ mice were paired with congenic CD45.1⁺ mice so that they shared the same blood circulation, but separate organs, for prolonged periods of time. As previously described, epidermal LCs remained of recipient origin >6 mo after parabiosis, which is consistent with their origin from skin-resident precursors (Fig. 3, A and B). Therefore, the presence of CD45.1⁺langerin⁺ DCs in the dermis of the CD45.2⁺ parabiont should establish their independence from LCs and their recruitment from the blood to the dermis in the steady state. As shown in Fig. 3, we were able to detect blood-derived langerin⁺CD8⁻ DCs in the skin (Fig. 3, A and B) and the skin DLNs (Fig. 3, C and D) of each parabiont. Total dermal DC chimerism was low, but significant, compared with LC chimerism as previously described (27), reflecting either prolonged or partial self-replenishment (27) or, as recently suggested, an uneven distribution of DC precursors (26). Nevertheless, the presence of donor-derived dermal langerin⁺ DCs suggests that, in contrast to LC-derived dermal DCs, dermal langerin⁺ DCs can originate from blood-derived precursors in the steady state. Interestingly, the chimerism of langerin⁺ DCs was lower than langerin⁻ DCs. Because host langerin⁺ DCs consist of a mixture of dermal langerin⁺ DCs and migrating LCs, it is likely that the lower chimerism of langerin⁺ compared with langerin⁻ DCs reflects the contribution of migrating LCs. Therefore, to better reveal the recruitment of blood-derived langerin⁺ cells to the skin and to the skin DLNs, we parabiosed Langerin-DTR/EGFP CD45.2⁺ C57BL/6 animals with CD45.1⁺ C57BL/6 and followed the repopulation of langerin⁺ DCs in the CD45.2⁺ parabionts at different times after DT treatment. 2 d after DT injection, langerin⁺ DCs were eliminated from the skin and the skin DLNs of the CD45.2⁺ Langerin-DTR/EGFP parabionts (unpublished data). 18 d after DT administration, a time point where LCs are absent (Fig. 3 E, top), the percentage of donor CD45.1⁺langerin⁺ DCs increased to the level of langerin⁻ DCs in the dermis (Fig. 3, E and F, bottom) and in the skin DLNs (Fig. 3, G and H) of the CD45.2⁺ Langerin-DTR/EGFP parabionts. These results establish that in the steady state, circulating langerin⁺ DCs are recruited to the dermis and to the skin DLNs independently of LCs.

Congenic BM chimeric mice provide another model in which the CD45 congenic allele can be used to distinguish LCs from blood-derived DCs in the skin DLNs. The advantage of this model is that the majority of circulating leukocytes express the congenic allele, whereas in parabiotic mice, no more than 50% of leukocytes express the congenic allele (26). To generate congenic BM chimeras, C57BL/6 CD45.2⁺ mice were lethally irradiated and reconstituted with BM cells isolated from C57BL/6 CD45.1⁺ mice. As previously described, a few weeks after transplant, most blood leukocytes originate from donor BM–derived precursors and express CD45.1, whereas the majority of epidermal LCs are maintained by radioresistant precursors, and remain of host (CD45.2⁺) origin throughout life, in the absence of skin injury (24) (Fig. 4 B). Using flow cytometry analysis of skin samples isolated at different times after transplant, we found that a substantial population of langerin⁺ DCs in the dermis of BM chimeras expressed donor CD45.1 antigens, suggesting that they derived from donor BM-derived circulating precursors and not from the host remaining CD45.2⁺ LCs (Fig. 4, A and B). Consistent with our results (Fig. 2), this population corresponded to ∼10% of dermal MHC class II⁺ cells. Blood-derived CD45.1⁺langerin⁺ DCs in the dermis were also MHC class II⁺, CD11c^int, CD11b⁺, CX₃CR1⁻, CD8⁻, a phenotype similar to that expressed by classical LCs with the exception of the langerin levels, but different from that of dermal langerin⁻ DCs (Fig. 4 C). We also discovered that, in the skin DLNs, an increased proportion of langerin⁺CD8⁻ DCs express CD45.1 (60–70%), and therefore are derived from blood-precursors and not from epidermal LCs (Fig. 4, D and E). Blood- and epidermal-derived langerin⁺CD8⁻ DCs were indistinguishable phenotypically and expressed higher levels of MHC class II and langerin and lower levels of CD11c compared with langerin^loCD8⁺ DCs (Fig. 4 F). Interestingly, langerin expression formed a smear on dermal blood–derived langerin⁺ DCs, which may suggest that the langerin receptor was being acquired upon entry in the dermis and up-regulated during the migration to the skin DLNs. These results are consistent with the data from a companion paper showing that blood-derived langerin⁺ DCs that emigrate from skin explants have similar langerin levels compared with LCs (28). Donor-derived langerin⁺ DCs were visualized in situ by confocal microscopy analysis of skin sections isolated from (CD45.2⁺Langerin-EGFP BM→CD45.1⁺C57BL/6) or the reverse (CD45.1+C57BL/6 BM→CD45.2+Langerin-EGFP) mice chimeras (Fig. 4 G). Donor-derived langerin⁺ DCs appear scattered in the dermis, without any conspicuous localization near the hair follicle. Donor-derived CD45.2⁺ MHC class II⁺ Langerin-EGFP⁺ DCs were also purified by FACS and stained for langerin on cytospins (Fig. 4 H). Altogether, our results suggest that, although undistinguishable phenotypically, blood- and epidermal-derived langerin⁺CD8⁻ DCs coexist in the skin DLNs of BM chimeric mice.

The presence of blood-derived langerin⁺CD8⁻ DCs in the skin DLNs always paralleled their presence in the skin. However, the percentage of blood (CD45.1⁺)-derived langerin⁺ DCs among total langerin^hi DCs was always higher in the skin DLNs compared with the dermis (compare Fig. 4 B with Fig. 4 E), suggesting that these cells either transit more rapidly than migratory LCs, but do not accumulate in the skin, or that they can also be recruited to the skin DLNs directly from the blood. To address these questions, we reconstituted lethally irradiated CD45.1⁺C57BL/6 recipient mice with BM cells isolated from congenic CD45.2⁺ mice that lack L-selectin (CD62L KO) or the chemokine receptor CCR7 (CCR7 KO). L-selectin is constitutively expressed on most leukocytes and plays an important role in leukocyte homing into lymphoid tissue from the blood (29). CCR7 is a chemokine receptor that plays a critical role in the migration of DCs from the skin to the DLNs (30); therefore, absence of CCR7 should interfere with the migration of DCs from the skin to the skin DLNs. 2 mo after reconstitution, the percentage of CCR7 KO and CD62L KO circulating langerin⁺ DCs in the dermis was similar to their wild-type counterpart (Fig. 5, A and B). In contrast, the percentage of CCR7 KO langerin⁺CD8⁻ DCs was markedly reduced compared with CCR7^+/+ langerin⁺CD8⁻ DCs in the skin DLNs (Fig. 5, C and D). No differences were observed between the percentages of CD62LKO versus CD62L^+/+ langerin⁺CD8⁻ DCs. These results suggest that circulating langerin⁺CD8⁻ DCs reach the skin DLNs through the skin and not directly from the blood.

Previous results suggest that circulating DC precursors roll continuously along noninflamed mouse dermal endothelium in vivo, an interaction shown to be strictly dependent on endothelial selectins (31). To examine if endothelial selectins play a role in the homeostasis of dermal langerin⁺ DCs, we reconstituted lethally irradiated C57BL/6 mice that lack both P- and E-selectins (P/E-selectin KO), or wild-type C57BL/6 controls with BM cells isolated from congenic CD45.1⁺C57BL/6 mice. 2 mo after reconstitution, we found that P/E-selectin KO have similar numbers of CD45.1⁺ circulating B cells, monocytes, and neutrophils compared with wild-type controls, suggesting that absence of P- or E-selectins did not affect BM engraftment (unpublished data). In contrast, the percentage of dermal langerin⁺ DCs in P/E-selectin KO mice was 30–50% lower than the percentage of dermal langerin⁺ DCs in wild-type mice (Fig. 5 E). Similar results were observed for dermal langerin⁻ DCs (unpublished data).

We previously established that CCR2 chemokine ligands are required for the recruitment of circulating dermal DC precursors to the skin (27). We also established that CCR2 and CCR6 chemokine ligands are required for the recruitment of circulating LC precursors and the repopulation of LCs in inflamed skin (32, 33). To examine the role of these molecules in the recruitment of circulating donor langerin⁺ DC precursors to the dermis, we reconstituted lethally irradiated CD45.1⁺C57BL/6 mice with a 1:1 mixture of hematopoietic progenitors cells that consisted of CD45.2⁺ BM cells that either lacked CCR2 (CCR2 KO) or CCR6 (CCR6 KO) and wild-type CD45.1⁺ BM. 2 mo after reconstitution, we found that CCR2^−/−, CCR6^−/−, and wild-type CD45.1⁺ BM gave rise to similar numbers of circulating B cells and neutrophils, suggesting that the absence of CCR2 and CCR6 did not affect BM engraftment (unpublished data). In contrast, CCR2^−/−CD45.2⁺CD115⁺ circulating monocytes represented 30–40% of circulating CD45.2⁺CCR6^−/− or CD45.1⁺ wild-type monocytes (unpublished data), confirming that absence of CCR2 expression on hematopoietic progenitors affects monocyte repopulation in the blood (27, 34). We found that CCR2^−/− dermal langerin⁺ DCs were reduced in the skin compared with their wild-type counterpart (Fig. 5 F), whereas CCR6 KO langerin⁺ dermal DCs were not affected (Fig. 5 F). Similar results were observed for dermal langerin⁻ DCs (Fig. 5 F). These differences underline the essential role of CCR2 in the recruitment of dermal langerin⁺ DCs to the skin, whereas the role of CCR6 seems to be restricted to the repopulation of epidermal LCs.

Our results reveal for the first time that circulating langerin⁺ DC precursors are recruited to the dermis in the steady state or after minor injuries (i.e., DT or radiation injuries). Our results also suggest that circulating langerin⁺ precursors that are recruited to the dermis do not establish in the epidermis, but emigrate to the skin DLNs. To examine if blood-derived dermal langerin⁺ DCs play a role in skin immunity, we sought to analyze whether they can capture, process, and present skin-derived antigen in the skin DLNs under steady-state conditions. To assess antigen presentation, we took advantage of the YAe monoclonal antibody that recognizes an MHC class II I-E–derived peptide presented in the context of MHC class II I-A^b molecules. We reconstituted lethally irradiated F1 (C57BL/6 I-A^b–/– x BALB/c I-E^d+) recipient mice with BM cells isolated from CD45.1⁺ C57BL/6 I-A^b+/+ Langerin-EGFP BM. In transplanted mice, only I-A^b+/+ donor cells, and not host residual LC I-A^b−/−, can present I-E^d–derived peptide recognized by YAe. In the skin of F1 (C57BL/6xBALB/c) animals, I-E^d is expressed by host radioresistant LCs (24) and some activated keratinocytes (35) (unpublished data). Therefore, donor I-A^b+/+ DCs should be able to present I-E^d–derived peptides and stain positive for YAe only if they can capture, process, and present skin-derived antigens. We found that, in the dermis, donor-derived circulating dermal langerin⁺ DCs stained positive for YAe, as did langerin⁻ dermal DCs, establishing their capacity to capture and present exogenous skin-derived antigens (Fig. 6, A and B, top). Our results also show that blood-derived dermal langerin⁺ DCs that migrate to the skin DLNs also stain positive for YAe (Fig. 6, A and B, bottom). A higher percentage of blood-derived dermal langerin⁺ DCs were positive for YAe in the skin DLNs compared with the dermis (Fig. 6, A and B, bottom), suggesting that dermal-derived langerin⁺ DCs process and present skin-derived antigens en route to the skin DLNs. Similar results were obtained when corresponding purified dermal DC populations were examined on cytospins for langerin and YAe expression (Fig. 6 C).

Using three separate models to trace the turnover of circulating leukocytes, our study reveals the presence of a population of langerin⁺ DCs, independent of LCs, in the dermis and skin DLNs of mice in the steady state.

We used parabiotic mice to establish the constitutive recruitment of blood-derived langerin⁺ DCs to the dermis and their subsequent emigration to the skin DLNs. Parabiotic mice provide a useful model to study the physiological turnover of leukocytes (36). Although recent data suggest that circulating DC precursors do not reach complete equilibrium in parabionts (26), this model continues to be very useful to compare trafficking patterns among DC subsets. Parabiotic mice are particularly useful to distinguish LC-derived DCs from blood-derived dermal langerin⁺ DCs in the steady state because LCs fail to mix in the epidermis >18 mo after the parabiosis is established (24). Using the congenic CD45.1 marker to trace circulating leukocytes in the CD45.2 parabiont, our results reveal that blood-borne dermal langerin⁺ DCs are constitutively recruited to the dermis and the skin DLNs in the absence of overt injuries. Similar results are reported in a companion paper (28). In this study, the authors show that in congenic BM chimeric mice, langerin⁺ DCs express higher levels of CD45 compared with LCs and used the levels of CD45 expression to identify langerin⁺ DCs, independent of LCs, in the dermis of naive nontransplanted animals. Altogether, these results strongly support the recruitment of circulating langerin⁺ DCs, independent of LCs, to the dermis in the steady state.

Congenic BM chimeric mice provide another tool to distinguish LCs from blood-derived DCs, as LCs remain of host origin months after transplant despite complete donor-derived chimerism in the blood. In this model, blood-derived langerin⁺ DCs were also present in the dermis and accounted for 30% of total langerin⁺ dermal DCs. Surprisingly, the percentage of blood-derived langerin⁺ DCs is increased in the skin DLNs, reaching up to 60% of langerin⁺ DCs. The higher percentage of donor-derived langerin⁺ DCs in the skin DLNs, compared with dermal langerin⁺ DCs in congenic BM chimeras, is intriguing, and could be partially caused by a minor contribution of donor-derived LCs, which represent 5–10% of total LCs in areas of skin exposed to fight- or grooming-related injuries, such as the neck and the back skin. However, the constant percentage of donor-derived langerin⁺ in skin DLNs draining different parts of the skin, including those containing host LCs exclusively, argues against this hypothesis (unpublished data). Alternatively, the radiation-induced injuries that might occur in transplanted BM chimera mice may also affect the homeostasis of dermal langerin⁺ DCs and increase their recruitment to the skin DLNs. Another possibility is that blood-derived langerin⁺ DCs are recruited directly from the blood to the skin DLNs. Several lines of evidence argue against this possibility. First, the presence of langerin⁺CD8⁻ DCs in the skin DLNs always paralleled their presence in the skin, with the latter reconstituting the dermis before the skin DLNs in depletion experiments. Second, their recruitment to the skin DLNs was independent of the L-selectin (CD62L), but required the chemokine receptor CCR7. CD62L is constitutively expressed on most leukocytes, including DC precursors, and plays a critical role in the first steps of leukocyte extravasation from the blood to the LN, allowing the rolling of blood leukocytes on high endothelial venules (HEVs) (29). CCR7 is the receptor for two structurally related chemokines, CCL19 and CCL21 (37). CCL19 and CCL21 are expressed by stromal cells within the lymphoid T cell zones, and CCL21 is expressed by HEVs and at lower levels by the lymphatic endothelium (38, 39). In mice lacking CCR7 (30) or CCR7 ligands (40), DCs fail to migrate from the skin to the T cell areas of the skin DLNs, providing the most compelling evidence for an essential role of CCR7 in this process. Altogether, these studies led to the current concept that migrating DCs express CCR7 and become chemotactically attracted to CCL21-expressing lymphatic vessels. The migrating DCs reach the subcapsular sinus of the DLNs and follow a gradient of CCL21 and CCL19 to move into the lymphoid T zone (38). Unlike all other chemokine receptors, CCR7 is resistant to ligand-induced down-regulation (41), which may explain how DCs can use CCR7 to perform all the critical steps of migration. The failure of blood-derived langerin⁺ DCs to enter the skin DLNs in the absence of CCR7, but not in the absence of CD62L, strongly suggests that these cells reach the skin DLNs through the skin lymphatics and not directly from the HEVs. However, CD8⁺ DCs, a population thought to be recruited to the LN through the HEVs, were also able to reach the skin DLNs in the absence of CD62L; however, in contrast to CD8⁻ langerin⁺ DCs, the lack of CCR7 did not affect their recruitment to the skin DLNs (unpublished data). These results were surprising, as we expected that, similar to naive T cells (39), expression of CD62L and CCR7 would both be required for efficient homing to the HEVs. However, these data support previous findings, showing that CD62L is not required for the migration of spleen DCs on resting or activated endothelial cells in vitro (42), and suggest that other molecules might orchestrate the recruitment of blood DCs through the HEVs. A potential candidate might be “chemerin,” a proteolytically regulated chemoattractant protein expressed at high levels by HEVs and shown to chemoattract in vitro ChemR23⁺ DCs, which include human blood DCs, but not LCs (43, 44).

We also examined the mechanisms that regulate the recruitment of blood-derived langerin⁺ DCs to the dermis. Our data suggest that blood-derived langerin⁺ DCs use E- and P-selectin to enter the skin in the absence of skin injuries. These data are consistent with a previous study showing that human CD34⁺-derived DCs injected into immunocompromised mice roll continuously along noninflamed mouse dermal endothelium in an E- and P-selectin–dependent manner (31). Our data also show that CCR2 controls the recruitment of blood-derived langerin⁺ DCs to the quiescent dermis. Expression of CCR2 (45) on circulating leukocytes was shown to be required for the repopulation of LCs (32, 46) and dermal langerin⁻ DCs in inflamed skin (27). Our results represent the first indication for a role of CCR2 in cutaneous DC homeostasis in the absence of overt skin inflammation. However, it remains possible that the transplant model used to reveal the role of CCR2 might lead to x ray–induced skin injuries, triggering the release of CCR2 ligands and the recruitment of circulating DCs that will not be encountered in the steady state. CCR2 also plays a critical role in the release of BM monocytes into the bloodstream (34), and mice reconstituted with CCR2^−/− BM have a decreased number of circulating monocytes (27). Therefore, it is possible that the decrease of CCR2^−/− blood-derived langerin⁺ DCs in the dermis reflects a reduced capacity for CCR2^−/− monocytes to reach the peripheral blood and differentiate into dermal DCs in the quiescent skin. These results support the recent findings that monocytes participate in DC homeostasis in peripheral tissues, but not in lymphoid organs (47), and the role of CCR2 in the development of nonlymphoid DC populations other than skin is currently being analyzed in the laboratory.

Consistent with previous studies, our results show that after systemic depletion of langerin⁺ cells in vivo, the repopulation of dermal and skin DLN langerin⁺ DCs precedes the repopulation of epidermal LCs (18, 23). We also demonstrate that dermal langerin⁺ DCs, independent of LCs, participate in the repopulation of the langerin⁺ DC pool in the skin DLNs, whereas the epidermis remains empty of LCs. It is intriguing that langerin⁺ dermal DCs migrate to the skin DLNs before seeding an epidermis depleted of LCs, suggesting that blood-derived langerin⁺ DCs cannot differentiate into LCs in vivo. Another possibility is that the toxin used to deplete langerin⁺ cells leads to inflammatory signals that trigger the migration of blood-derived dermal langerin⁺ DCs to the skin DLNs before they can reach the epidermis. This is unlikely because we have shown that circulating LC precursors can seed an inflamed epidermis and differentiate into LCs (33). In addition, we and others have not detected signs of skin inflammation or injuries after DT injection ([18] and unpublished data). A more likely explanation is that in vivo conditional LC ablation with DT frees available LC niches, but fails to induce epidermal injury signals, precluding the recruitment of blood-derived precursors to the epidermis. It will be interesting to examine if induction of inflammatory cytokines and chemokines in the epidermis after DT treatment accelerates the recruitment of dermal langerin⁺ DCs to the epidermis. If this is true, these data, together with our previous data showing that monocytes are the precursors of LCs in inflamed skin, will support the hypothesis that monocytes are constitutively recruited to the dermis where they differentiate into langerin⁺ DCs dedicated to survey the epidermis or the dermal–epidermal junction.

The acquisition of the langerin receptor on skin-infiltrating monocytes or a committed DC precursor might represent the signature of a “LC-like” differentiation program induced by environmental factors such as TGFβ, for example (48). Supporting this hypothesis is our finding that dermal langerin⁺ DCs seem to progressively acquire the LC marker langerin while losing the monocyte marker CX3CR1, a marker that is also absent on LCs, but partially expressed by dermal langerin⁻ DCs (Fig. 4 C). Whether dermal langerin⁺ and dermal langerin⁻ DCs represent the progeny of a precursor that follows a separate differentiation pathway, with the dermal langerin⁺ DCs being an intermediate precursor to LCs, is plausible, but difficult to test because of the limitation of an experimental system dealing with a low number of cells. More information will be needed to compare the nature of the antigen cargo displayed by these different sources of antigen-presenting cells in the skin DLNs to assess the immune relevance of each of these DC populations.

Our study also reports a striking difference between the homeostasis and trafficking pattern of LCs and dermal langerin⁺ DCs. After conditional ablation of langerin⁺ DCs in vivo, we were surprised to discover that dermal langerin⁺ DCs reappear in 5 d and start to emigrate to the skin DLNs and repopulate the langerin⁺ DC pool immediately after, whereas LCs remained absent from the epidermis for >2 wk after their elimination. The constitutive recruitment of blood-derived langerin⁺ DCs to the dermis, together with their rapid transit through the skin before reaching the skin DLNs, strongly support their role in skin immunosurveillance. This role is further supported by our findings showing that blood-derived dermal langerin⁺ DCs capture and process skin-derived antigens before emigrating to the T cell areas of the skin DLNs, where they present skin-derived peptides in the context of MHC class II molecules. These findings establish that, in addition to LCs and dermal langerin⁻ DCs, an additional population of antigen-presenting cells may play a role in skin immunity. The expression of the langerin receptor on the surface of dermal DCs should allow the sampling of antigen patterns by LCs, but not by dermal langerin⁻ DCs.

Our results reveal a novel pathway of cutaneous DC differentiation and homeostasis that add to the complexity of the DC network system developed to support skin immune responses (Fig. 7). We identified a population of circulating langerin⁺ DCs, independent of LCs, that constitutively patrols the dermis to capture and process skin antigens before emigrating to the skin DLNs, where they present skin-derived peptides–MHC complexes (Fig. 7). This study uncovers a previously unappreciated element of skin immunosurveillance that is likely to impact the design of vaccine strategies.

BALB/c, C57BL/6 (CD45.2⁺), and congenic C57BL/6 CD45.1⁺ mice 5–8 wk of age were purchased from The Jackson Laboratory. Langerin-EGFP, Langerin-DTR/EGFP mice were generated as previously described (18). CCR2-deficient C57BL/6 (45) and CCR6-deficient C57BL/6 (49), CD62L^−/− (50) and CCR7^−/− (30) (a gift from J. Lowe, Cleveland Clinic, Cleveland, OH), and P/E-selectin–deficient (P/E^−/−) mice (51) (a gift from P. Frenette, Mount Sinai School of Medicine, New York, NY) were bred and maintained at our animal facility (Mount Sinai School of Medicine, New York, NY). CD45.1⁺ Langerin-EGFP mice were generated by crossing of congenic C57BL/6 CD45.1⁺ with CD45.2⁺ Langerin-EGFP. C57BL/6 BALB/c F1 and I-A^b−/−C57BL/6 BALB/c F1 were generated by breeding BALB/c with C57BL/6 or I-A^b−/−C57BL/6 (The Jackson Laboratory), respectively. All mice used for experiments were between 8 and 12 wk of age. Parabiotic mice were generated as previously described (36). Pairs of parabiotic mice consisted of congenic C57BL/6 CD45.1⁺ mice linked to either C57BL/6 CD45.2⁺ or Langerin-DTR/EGFP C57BL/6 CD45.2⁺ mice. Parabiotic mice were killed ∼2 mo after initiation of parabiosis and subjected to tissue analysis. To confirm efficient blood mixing in parabiotic mice, the percentage of CD45.1⁺ and CD45.2⁺ cells among blood leukocytes was analyzed in each animal. All animal protocols were approved by the Institutional Committee on Animal Welfare of the Mount Sinai Medical School.

Multiparameter analyses of stained cell suspensions were made on an LSR II flow cytometer (Becton Dickinson) and were analyzed with FlowJo software (Tree Star, Inc.). Fluorochrome- or biotin-conjugated monoclonal antibodies specific to mouse B220 (RA3-6B2), CD8a (53–6.7), I-A^b (AF6-120.1), IA/IE (M5/114.15.2), CD11b (M1/70), CD11c (N418), CD45 (30F11), CD45.1 (A20), CD45.2 (104), CD115 (AFS98), Gr-1Ly6C (1A8), Gr-1Ly6C/G (RB6-8C5), CD3 (17A2), CD4 (clone L3T4), anti–mouse Eα52-68 peptide bound to the I-A^b/MHC Class II (YAe), the corresponding isotype controls, and the secondary reagents (allophycocyanin, peridinin chlorophyll protein, and phycoerythrin–indotricarbocyanine–conjugated streptavidin) were purchased either from BD Biosciences or eBioscience. For YAe staining, naive untreated cells from C57BL/6 mice that lack I-E molecules were used as negative controls, and cells from C57BL/6 BALB/c F1 mice were used as positive controls. Anti–mouse CD16-32 (BD Biosciences) and biotinylated mouse IgG2b isotype control (BD Biosciences) were used to block Fc III/II receptors and as a YAe isotype control antibody, respectively. F4/80 (A3-1) was purchased from Serotec. Polyclonal antibody to langerin was purchased from Santa Cruz Biotechnology. Intracellular staining against langerin was performed with the BD Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer's protocol. For cytospin immunostaining, stained cell suspensions were sorted using an inFlux cell sorter (Cytopeia, Inc.) and achieve 97–99% purity.

8-wk-old recipient CD45.2⁺ or CD45.1⁺ C57BL/6 mice were lethally irradiated with 1,200 rads, delivered in 2 doses of 600 rads each, 3 h apart, and were injected i.v. with 10⁶ BM cells obtained from congenic CD45.1⁺ or CD45.2⁺ C57BL/6 adult mice, respectively. To address the role of CCR2 and CCR6 in dermal DC langerin^+/− homeostasis, lethally irradiated CD45.1⁺ C57BL/6 mice were reconstituted with a 1:1 mixture of WT CD45.1⁺ BM cells and CD45.2⁺ BM isolated from WT, CCR2 KO, or CCR6 KO mutant mice. In all transplanted mice, levels of blood donor chimerism were analyzed by measuring the percentage of donor cells among total B220⁺ B cells, Ly6C/G⁺CD115⁻ granulocytes, and CD115⁺ monocytes in the blood 3 wk after transplantation.

Skin cell suspensions were isolated as previously described (24) and were analyzed by flow cytometry. In brief, mouse ears were split in two parts (dorsal and ventral) and incubated for 60 min in PBS containing 0.5% trypsin with 5 mM EDTA (Invitrogen) to allow for separation of dermal and epidermal sheets. Epidermal and dermal sheets were then cut into small pieces and incubated for 2.5 h in collagenase D (1.6 mg/ml, working activity of 226 U/mg; Worthington) to obtain homogeneous cell suspension. The analysis of hematopoietic cell populations present in skin cell suspensions was assessed by flow cytometry by gating on DAPI⁻CD45⁺ cells.

For staining of frozen sections, skin samples or skin DLNs were first fixed for 2 h in PBS 4% paraformaldehyde containing 20% sucrose to preserve EGFP fluorescence and were rinsed in PBS before freezing in optimum cutting temperature compound. 8-μm-thick skin sections were prepared and stained with anti-langerin (polyclonal goat IgG; Santa Cruz Biotechnology) for 1 h, washed in PBS, and incubated with secondary reagent donkey anti–goat Cy3 (Jackson Immunoresearch Laboratories). Skin DLN sections were first stained with anti-langerin (polyclonal goat IgG), followed by donkey anti–goat Cy3, and were stained after with biotinylated anti-B220 (RA3-6B2; BD Biosciences) revealed with streptavidin-Cy5 (Vector Laboratories). Slides were mounted with Vectashield containing DAPI (4,6-diamidino-2-phenylindole; Vector Laboratories). Images were acquired with a laser scanning confocal microscope (TCS-SP; Leica) and analyzed with Image J software (National Institutes of Health).

Cytospins were prepared from 5–10,000 sorted cells, air-dried, and fixed in 3% paraformaldehyde. Slides were incubated with 0.8 μg/ml anti-langerin goat polyclonal antibody or corresponding control (goat IgG; Sigma-Aldrich). For YAe staining, after blocking with 5 μg/ml mouse IgG2b (eBioscience) and 2.5 μg/ml anti-CD16/32 antibodies (BD Biosciences), slides were incubated with 2.5 μg/ml anti-YAe biotinylated antibody (eBioscience) and 0.8 μg/ml anti-langerin goat polyclonal antibody or isotype controls. Secondary detection of YAe was achieved using biotinyl-tyramide amplification (kit NEL700A; PerkinElmer) combined with 1 μg/ml Cy5-conjugated streptavidin (Jackson Immunoresearch Laboratories) and 1 μg/ml Cy3-conjugated donkey anti–goat (Jackson Immunoresearch Laboratories). GFP fluorescence was usually not detectable in these preparations. Images were acquired using an Axiophot microscope at 40× magnification (Carl Zeiss, Inc.).

Fig. S1 depicts the gating strategy used in the dermis and in the skin DLNs to analyze all DC subsets by flow cytometry. The online version of this article is available.

The authors would like to thank Drs. J. Donovan and Y. Kirkorian for their review of the manuscript.

This work was supported by grants from the National Institutes of Health (RO1-CA112100), the Leukemia and Lymphoma Society (FG, 3220-08), the Leukaemia Research Fund Bennett Senior Fellowship (M.P. Collin), and the Amy Strelzer Manasevit Research program (M. Bogunovic).

The authors have no conflicting financial interests.