Langerhans cells are negative regulators of the anti-Leishmania response

Langerin-DTR mice were injected i.p. with DT to deplete all Langerin⁺ DCs, including epidermal LCs, Langerin⁺ dDCs, and LN-resident Langerin⁺ DCs (Fig. S1, A and B; Bennett et al., 2005, 2007). 2 d after DT treatment, mice were infected intradermally with 1,000 metacyclic L. major promastigotes and were subsequently treated weekly with DT to maintain depletion of all Langerin⁺ cells. DT treatment was well tolerated without any side effects as reported previously (Bennett et al., 2005; Bennett and Clausen, 2007). After infection, mice depleted of Langerin⁺ cells developed significantly smaller ear lesions as compared with PBS-treated controls (Fig. 1 A). Lesion sizes in the PBS control group were similar to those in C57BL/6 wild-type mice (Fig. S1 C). Decreased lesion sizes in Langerin-DTR + DT mice correlated with significantly lower parasite burdens in infected ears at week 6 after infection (5.3 ± 1.9 × 10⁴ in DT-treated mice compared with 3.9 ± 1.10⁵ in control mice, n = 8, P = 0.008; Fig. 1 B). Spleen parasite burdens were very low in both groups without significant differences.

We next assessed the cytokine profile 6 wk after infection. IFN-γ, IL-12, and IL-10 levels were determined after 48 h of restimulation of skin-draining LN cells with soluble Leishmania antigen (SLA). In Leishmania infections, Th1-inducing IL-12 is primarily released by infected DCs (von Stebut et al., 1998; Ahuja et al., 1999; Sacks and Noben-Trauth, 2002) and is critical to mediate protection (Berberich et al., 2003). In this study, Langerin⁺ DCs did not contribute to this IL-12 release (Fig. 1 C). Importantly, IFN-γ production was significantly increased in DT-treated Langerin-DTR mice as compared with controls (57 ± 6 vs. 31 ± 8 ng/ml, P = 0.021). No difference was observed in IL-10 production in week 6 after infection (Fig. 1 C). IL-4 levels in LN supernatants were low as expected (C57BL/6 Leishmania-resistant background) without differences between treatment groups.

Our findings demonstrate a novel and unexpected suppressive role of Langerin⁺ DCs, which is only visible in the physiologically relevant low-dose infection model. When we used high-dose inocula (intradermal infection with 2 × 10⁵ metacyclic promastigotes) in Langerin⁺ DC-depleted mice, we did not observe comparable changes regarding lesion development, parasite burdens, and cytokine profile (Fig. S2, A–C). This is consistent with data from Brewig et al. (2009), who used high doses of parasites (3 × 10⁶) and subcutaneous inoculation, which masques any regulatory role of LCs in experimental leishmaniasis. Thus, other differences between that prior study and ours (other mouse strain, different route of parasite inoculation, and distinct parasite strains) are not responsible for the incongruent results.

As both types of skin DC, i.e., epidermal LCs and Langerin⁺ dDCs, are depleted by continuous DT treatment in our experiments, it was not clear which cell type was responsible for the suppressive effects shown in Fig. 1. To analyze the role of LCs in more detail, we performed timed toxin treatments. After conditional ablation of Langerin⁺ DCs in vivo using a single injection of DT, Langerin⁺ dDCs started reappearing already at days 3–5 and were largely restored by day 14 (Fig. S3 A; Noordegraaf et al., 2010), whereas this one-time DT administration depleted LCs for at least 2–4 wk with, occasionally, very few cells detectable at 2 wk and only partial reconstitution as late as 10 wk after DT (Fig. S3 B; Bennett et al., 2005). We next took advantage of this differential repopulation kinetics of LCs and Langerin⁺ dDCs to discriminate between the two DC subsets. Specifically, Langerin-DTR mice were injected once with DT 2 wk before Leishmania inoculation such that essentially all LCs were still depleted (Bennett et al., 2005), whereas the majority of Langerin⁺ dDCs had already reappeared (Fig. S3 A; Noordegraaf et al., 2010). Upon selective depletion of LCs alone, lesion sizes were significantly reduced as compared with controls, with the largest difference in week 7 (9.6 ± 1 and 5 ± 0.6 mm³ in PBS and DT 1× groups, respectively, P = 0.0001; Fig. 2 A). No obvious difference between the two different DT treatment protocols was seen.

To confirm these results, we generated hematopoietic BM chimeras. C57BL/6 donor BM was used to reconstitute lethally irradiated (9 Gray) Langerin-DTR recipient mice. Epidermal LCs are radioresistant and remain of host (Langerin-DTR) origin after irradiation and BM transplantation. In contrast, all other DC subtypes are radiosensitive and replaced by donor-derived cells (C57BL/6). The resulting chimeric animals were treated once with DT 48 h before L. major infection, causing selective depletion of LCs (Fig. S3 C). Subsequently, mice were infected with low-dose inocula of L. major and lesion development was assessed weekly. As shown in Fig. 2 B, selective depletion of LCs resulted in significantly reduced lesion sizes. In contrast, in irradiated C57BL/6 mice reconstituted with Langerin-DTR BM, treatment with DT led to depletion of all Langerin⁺ dDCs, except LCs (Fig. S3 C). Interestingly, in these mice, lesion development was somewhat worsened as compared with controls, but the differences were not significant (Fig. 2 C). Moreover, these experiments ruled out an unspecific DT-induced phenomenon, as lesion sizes were only reduced after selective depletion of LCs but not after DT-mediated depletion of Langerin⁺ dDCs. Collectively, using two different experimental approaches, our results demonstrate that depletion of LCs alone is sufficient to significantly reduce lesion sizes, indicating that LCs are the relevant DC subset suppressing efficient effector T cell (T eff cell) responses in this skin infection model.

To elucidate the mechanism underlying the observed phenotype, we analyzed the CD4⁺ and CD8⁺ T cell populations in skin draining LN and infected ears in the presence or absence of Langerin⁺ cells biweekly after infection. No difference was observed in the total numbers of CD4⁺ and CD8⁺ T cells during the course of disease in DT- and PBS-treated groups, neither in the skin-draining LN nor in the infected ears (unpublished data).

Next, we asked if improved disease outcome in Langerin⁺ DC–ablated mice was associated with altered numbers of regulatory T cells (T reg cells) infiltrating the infected ears. It was shown previously that T reg cells are essential for the development and maintenance of chronic cutaneous infection in L. major–resistant C57BL/6 mice (Belkaid et al., 2002; Mendez et al., 2004; Yurchenko et al., 2006). T reg cells accumulate at sites of infection, where they suppress anti-pathogen CD4⁺ T cell responses and consequently control concomitant immunity (Mendez et al., 2004). Using IL-10^−/− mice or administration of anti–IL-10 antibody led to complete eradication of the parasite in these mice, whereas an increase in the number of T reg cells promoted development of a new systemic infection (Belkaid et al., 2001, 2002). Interestingly, the important contribution of T reg cells in L. major infections was revealed only by using the physiological low-dose infection model (Belkaid et al., 2002; Yurchenko et al., 2006).

In this study, infiltrating ear cells of DT- and PBS-treated mice were isolated at the indicated time points after infection and analyzed for their CD4 and Foxp3 expression (Fig. 3 A). In the absence of Langerin⁺ DCs, the percentage of CD4⁺ Foxp3⁺ T reg cells in infected ears was significantly reduced in weeks 5 and 7 after Leishmania infection as compared with control mice (8.2 ± 1.7% vs. 14.1 ± 1.7%, n = 8 mice/group in week 5, P = 0.027; and 6.6 ± 1.4% vs. 11.3 ± 1.1%, n ≥ 10 mice/group in week 7, P = 0.015, in DT- and PBS-treated mice, respectively). Similar results were obtained when total numbers of CD4⁺ Foxp3⁺ T reg cells infiltrating infected ears of DT- and PBS-treated mice were compared. In this case, the T reg cell number was significantly reduced from, for example, 11.0 ± 1.9 × 10³ in the PBS group to 3.3 ± 1.0 × 10³ in DT-treated mice at week 7 (n ≥ 10, P = 0.0029; unpublished data).

We also assessed antigen-specific IFN-γ production over time after infection of PBS- and DT-treated mice. In parallel with the time points of maximal differences in T reg cell numbers, IFN-γ production of LN cells was significantly increased in DT-treated Langerin-DTR mice as compared with controls in weeks 5 and 7 (60 ± 15 vs. 21 ± 6 ng/ml in week 5, P = 0.031; and 83 ± 8 vs. 53 ± 10 ng/ml in week 7, P = 0.036, respectively; Fig. 3 B). Studies from Belkaid et al. (2000) revealed two distinct phases in the pathogenesis of cutaneous leishmaniasis (low dose infection). The initial silent phase, lasting 4–5 wk, favored dermal parasite amplification without lesion formation (Belkaid et al., 2000). The second phase corresponded to the development of lesions associated with an acute infiltration of inflammatory cells into the dermis and an expansion of T cells capable of producing IFN-γ in the draining LN, resulting in parasite killing (Belkaid et al., 2000). Accordingly, in our study the strongest difference in lesion sizes was detected between weeks 7 and 12 after infection after depletion of Langerin⁺ DCs, together with significant differences in IFN-γ production between weeks 5 and 7 (and not in the silent phase of infection).

In a separate set of experiments, we assessed if selective depletion of LCs leads to similar results. As indicated in Fig. 3 C, significantly reduced numbers of CD4⁺ Foxp3⁺ T reg cells were observed in the ears of both DT-treated groups (e.g., week 7: PBS, 12.4 ± 1.9 × 10³; DT 1×, 6.9 ± 1.5 × 10³; and DT weekly, 6.5 ± 1.3 × 10³, P = 0.049 and P = 0.016, respectively, n ≥ 14), indicating that LCs alone are sufficient for T reg cell expansion in vivo. In addition, similar to continuous DT treatment, the IFN-γ production and IFN-γ/IL-4 ratio were significantly increased in the DT 1× treatment group in comparison with PBS controls in week 7 (Fig. 3 D).

Because T reg cells exert their suppressive function largely by secretion of IL-10 (Belkaid et al., 2002), we asked whether there was a shift in the balance between CD4⁺ T eff cells producing IFN-γ and T reg cells producing IL-10 when LCs were absent. As shown in Fig. 3 D, IL-10 production was significantly reduced in LC-depleted mice as compared with PBS controls and mice depleted of both LCs and Langerin⁺ dDCs (0.7 ± 0.1, 1.4 ± 0.2, and 1.4 ± 0.2 ng/ml in DT 1×, PBS, and DT weekly treatment groups, respectively, P = 0.007 and P = 0.02). Consequently, we detected an increase in the IFN-γ/IL-10 ratio, indicating that there were more CD4⁺ T eff cells producing IFN-γ and fewer CD4⁺ Foxp3⁺ T reg cells producing IL-10 present in the skin draining LN of LC-depleted as compared with LC-competent mice (94 ± 22 vs. 23 ± 8 in DT 1× as compared with PBS groups, P = 0.0068). In conclusion, conditional ablation of LCs resulted in enhanced protective immunity against physiological low-dose infection with L. major associated with an increased Th1 and an attenuated T reg cell immune response.

Our in vivo data strongly suggest that LCs contribute to the expansion of T reg cell at the site of infection. To confirm these results, we analyzed the effect of SLA-loaded Langerin⁺ DCs on regulatory CD4⁺ T cells ex vivo. Epidermal cells (ECs; including LCs) were isolated from murine ear skin and pulsed with SLA. Subsequently, these ECs were co-cultured with CD4⁺ T cells isolated from Leishmania-infected and healed C57BL/6 mice. After 24 h, T cells were analyzed for their CD25 and Foxp3 expression using flow cytometry (Fig. 4 A). A significant increase of the percentage of CD25⁺ T cells was found when antigen-loaded ECs were used for antigen presentation as compared with negative controls without ECs (11.5 ± 1.0 vs. 6.2 ± 0.9%, respectively, P = 0.017; Fig. 4 B). In addition, this was also true for the frequency of Foxp3⁺ CD4⁺ cells (9.5 ± 1.4 vs. 4.2 ± 0.5%, P = 0.021; Fig. 4 B). These results strongly suggest that epidermal LCs can expand the T reg cell population.

In a subsequent experiment, DCs were isolated using anti-CD11c–coated MACS beads from LN of Langerin-DTR mice treated with either DT (Langerin⁺ DC-depleted) or PBS. As described in the previous paragraph, these DCs were loaded with SLA and co-cultured with CD4⁺ T cells (Fig. 4 C). After 24 h, expansion of CD25⁺ T cells (18.2 ± 1.4 vs. 12.7 ± 0.6%, P = 0.012) and CD25⁺ Foxp3⁺ T cells (5.4 ± 0.4 vs. 3.4 ± 0.5%, P = 0.011) was significantly increased when the antigen-presenting DCs comprised Langerin⁺ DCs as compared with those lacking Langerin⁺ DCs (Fig. 4 D). Thus, Langerin⁺ DCs, and not Langerin^neg DCs, drive the expansion of antigen-specific T reg cell and, corroborating the in vivo data (Fig. 3, A and C), expansion of T reg cell was significantly reduced after Langerin⁺ DCs had been depleted.

Finally, to address the question of whether Langerin⁺ DCs promote T reg cell survival or whether they are able to induce proliferation of T reg cell, SLA-pulsed LN DCs (including/excluding Langerin⁺ DCs) were co-cultured with CD4⁺ CD25^neg T cells, and induction of T reg cell was assessed after 3 d. Absolute numbers of CD4⁺ CD25⁺ Foxp3⁺ and, additionally, the ratio of T eff/T reg cell was strongly increased when Langerin⁺ DCs had been present in the co-culture (Fig. 4 E). DCs without antigen showed only a weak capacity to induce T reg cell.

Collectively, our results reveal a novel inhibitory role of LCs in orchestrating the anti-Leishmania response by suppressing efficient development of protective Th1 immunity. In addition, LCs favor the recruitment of T reg cell to the site of infection, resulting in aggravated disease. Mice on a Leishmania-resistant C57BL/6 background exhibit attenuated disease in the absence of epidermal LCs, suggesting that the parasites may possess well adapted strategies to exploit the regulatory capacities of LCs to escape efficient immune surveillance. Nevertheless, LC-competent C57BL/6 mice still display a self-healing phenotype, arguing that the presence of LCs is insufficient to mediate disease susceptibility in this strain. Our data are also relevant for DC-based vaccination strategies, as infected LCs may induce Leishmania-specific immunosuppression rather than parasite control. Indeed, our findings suggest that efforts to selectively deplete LCs from the skin or avoid targeting LCs during vaccination may benefit patients suffering from (cutaneous) leishmaniasis.

Langerin-DTR mice (C57BL/6) were bred as heterozygotes (Bennett et al., 2005) and housed under specific pathogen-free conditions in the animal care facility in Mainz. Animal experimentation was conducted in accordance with current federal guidelines and approved by the national authorities (Landesuntersuchungsamt, LUA 23 170–07/ G07-1-022).

Metacyclic promastigotes of L. major clone VI (MHOM/IL/80/Friedlin) were prepared as described previously (von Stebut et al., 2003). Groups of three to five mice were infected with low-dose (10³) inocula by intradermal injection into ear skin. In some experiments, DT (Sigma-Aldrich) was diluted in sterile PBS, and ∼26 ng/g body weight was injected i.p. as indicated. Lesion volumes were measured weekly in three dimensions and are reported as ellipsoids: [(a/2 × b/2 × c/2) × 4/3π]. Parasites present in lesional tissue were enumerated using a limiting dilution assay as previously described (von Stebut et al., 2003).

The dorsal and ventral halves of mouse ears were split mechanically and placed dermis down into Petri dishes containing 0.5 M NH₄SCN dissolved in 0.1 M Na-KHPO₄ buffer at 37°C for 20 min. Epidermis was removed, fixed with ice-cold acetone, and rehydrated in PBS. Epidermal sheets were stained with FITC anti–mouse I-A/I-E (2G9; BD) for 1.5 h at room temperature. After washing with PBS, the epidermal sheets were mounted on slides with Fluorescent Mounting Medium (Dako). Images were analyzed at room temperature using a fluorescent microscope (Axiovert S100; Carl Zeiss).

For measurement of antigen-specific cytokine production, retroauricular LNs were recovered and single cell suspensions were prepared. One million LN cells/200 µl complete RPMI 1640 (BioWhittaker) were added to 96-well plates in the presence of 25 µg/ml SLA. Supernatants were harvested 48 h after stimulation and assayed using ELISAs specific for IL-12p40 and IFN-γ (R&D Systems), as well as IL-4 and IL-10 (BD).

Ears infected with L. major were excised, soaked in 70% ethanol, and washed with PBS. The ears were split into halves and placed in 0.5 mg/ml liberase (Sigma-Aldrich) diluted in RPMI 1640 with 5% penicillin/streptomycin for 1.5 h at 37°C. Liberase was inactivated by adding complete RPMI 1640 containing 5% FCS. The ears were put into 50 µM Medicon homogenizers (BD) with 1 ml of complete RPMI 1640 and homogenized in a Medimachine (BD) for 7 min. The homogenized ears were then passed through a 70-µm-pore size filter and centrifuged at 200 g for 8 min. Cells were resuspended in PBS and labeled with antibodies for flow cytometry.

For DC isolation, LNs from DT- or PBS-treated Langerin-DTR mice were harvested and single cell suspensions were prepared. DC isolation was performed using MACS CD11c MicroBeads (Miltenyi Biotec) according to the manufacturer’s instructions. For isolation of CD4⁺ T cells, LN cells from L. major–infected and healed C57BL/6 were collected and enriched using MACS CD4⁺ MicroBeads (Miltenyi Biotec). For isolation of CD4⁺ CD25⁻ T cells, LN cells from L. major–infected and healed C57BL/6 were collected and enriched using MACS CD4⁺ CD25⁺ regulatory T cell isolation kit (Miltenyi Biotec), finally using the CD4⁺ CD25⁻ cell fraction.

ECs were prepared from ear skin by trypsinization (0.5% trypsin [GE Healthcare] in HBSS [BioWhittaker] for 30 min at 37°C) and dissociation of epidermal sheets by gentle pipetting in HBSS/0.05% DNase/30% FCS. Cell clumps were removed by passing the cell suspension through a 70-µm nylon mesh.

Dermal cell suspensions were prepared from ear skin after removal of the epidermis. Dermal tissue was minced into small pieces and digested for 2 h at 37°C with a solution of HBSS, 0.5% Collagenase (Sigma-Aldrich), 0.05% DNase, and 2% FCS. Dermal Langerin⁺ cells were identified by expression of CD11c, Langerin, and CD103.

The following antibodies were purchased from BD: PerCP-Cy5.5 rat anti–mouse CD8α (53–6.7), PE rat anti–mouse CD4 (L3T4), and FITC rat anti–mouse CD25 (7D4). Intracellular APC anti–mouse Foxp3 (FJK-16s; eBioscience) staining was performed using eBioscience reagents according to the manufacturer’s protocol. Single cell suspensions were stained first for surface markers, followed by a permeabilization step and staining for Foxp3. For analysis of skin and LN DCs the following antibodies were used: Pacific blue anti–mouse MHCII (M5/114.15.2; BioLegend), PerCPCy5.5 anti–mouse CD103 (2E7; BioLegend), APC anti–mouse CD11c (HL3; BD), PE anti–mouse CD11c (N418; eBioscience), and Alexa Fluor 488 Langerin (929F3.01; Dendritics) for intracellular staining. Cells were analyzed using a FACSCalibur or LSRII cytometer (BD) and FlowJo software (Tree Star).

Langerin DTR and C57BL/6 donor BM was used to reconstitute lethally irradiated (9 Gray) recipient mice (C57BL/6 and Langerin-DTR, respectively). The resulting chimeric mice were allowed to reconstitute their hematopoietic compartment for at least 6 wk before analysis. Subsequently, mice were infected with physiologically relevant low dose inocula of L. major and lesion development was assessed weekly.

Statistical analysis was performed using StatView software and Student’s t test.

Fig. S1 shows the depletion of Langerin⁺ DCs in the epidermis, dermis, and LN of Langerin-DTR mice 48 h after administration of DT. Fig. S2 illustrates the course of disease in high-dose–infected Langerin-DTR mice with and without DT treatment. Fig. S3 depicts the selective depletion/slow reconstitution of LCs 2 wk after DT administration and confirms specific depletion of LCs and Langerin⁺ dDCs in different BM chimeras.

We thank Susanna Lopez Kostka, Drs. Sven Brosch, Klaus Griewank, Helmut Jonuleit, Kerstin Steinbrink, and Mark C. Udey for helpful discussions, Mark C. Udey and Jon D. Laman for critical reading of the manuscript, and Kurt Reifenberg and staff for excellent assistance in animal experimentation.

This work was supported by grants from the MAIFOR program of the University of Mainz (K. Kautz-Neu), from the Deutsche Forschungsgemeinschaft (DFG; SFB548 to E. von Stebut), and from the Bilateral Cooperation Program DFG-NWO (The Netherlands Organization for Scientific Research; DN 93-525 to B.E. Clausen and Ste 833/6-1 to E. von Stebut). B.E. Clausen is a fellow of the Landsteiner Foundation for Blood Transfusion Research (LSBR) and a VIDI fellow of the NWO.

The authors have no financial conflict of interests.