Langerhans Cells Serve as Immunoregulatory Cells by Activating NKT Cells

Abstract

UV exposure alters the morphology and function of epidermal Langerhans cells, which plays a role in UV-induced immune suppression. It is generally believed that UV exposure triggers the migration of immature Langerhans cells (LC) from the skin to the draining lymph nodes, where they induce tolerance. However, because most of the previous studies employed in vitro UV-irradiated LC, the data generated may not adequately reflect what is happening in vivo. In this study we isolated migrating Langerhans cells from the lymph nodes of UV-irradiated mice and studied their function. We found prolonged LC survival in the lymph nodes of UV-irradiated mice. LC were necessary for UV-induced immune suppression because no immune suppression was observed in Langerhans cells-deficient mice. Transferring LC from UV-irradiated mice into normal recipient animals transferred immune suppression and induced tolerance. We found that LC co-localized with lymph node Natural Killer T (NKT) cells. No immune suppression was observed when LC were transferred from UV-irradiated mice into NKT cell-deficient mice. NKT cells isolated from the lymph nodes of UV-irradiated mice secreted significantly more IL-4 than NKT cells isolated from non-irradiated controls. Injecting the wild type mice with anti-IL-4 blocked the induction of immune suppression. Our findings indicate that UV exposure activates the migration of mature LC to the skin draining lymph nodes where they induce immune regulation in vivo by activating NKT cells.

Introduction

Epidermal Langerhans cells (LC) ⁴ are immature dendritic cells residing in the skin that are distinguished from other dendritic cells by the presence of cytoplasmic organelles, known as Birbeck granules (1), and by strong expression of the transmembrane type-II Ca²⁺-dependent lectin langerin/CD207 (2). Langerhans cells capture antigens in the skin, undergo maturation and migrate to LN (3). Langerhans cells are traditionally thought to play a crucial role in activating adaptive cutaneous immune responses, such as contact hypersensitivity (CHS). Recent studies using LC deficient mice, however, have caused a re-evaluation of the exact role of LC in CHS because three distinct results were obtained when these mice were sensitized with hapten: a diminished CHS response (4), an enhanced CHS response (5), or a CHS response that was no different from that found in the wild type controls (6). The concept that LC may not play a major role in cutaneous immunology is further supported by data demonstrated that dermal dendritic cells (dDC), and not LC act as the principal antigen-presenting cell in leishmaniasis (7) and CHS (8).

Langerhans cells also play a role in regulating the immune response. The enhanced CHS response in LC-deficient mice was an early hint that LC can function as immune regulatory cells (5). LC are required for graft acceptance across an HY minor histocompatibility barrier, indicating they regulate tolerance induction in vivo (9). Similarly, LC regulate the induction of graft versus host disease in vivo (10). Furthermore, Waithman and colleagues found that presentation of self-antigen (OVA constitutively expressed under the control of the K5 promoter in OVA transgenic mice) by LC resulted in the deletion of antigen-specific T cells, resulting in immune tolerance (11).

Exposing skin to UV radiation, prior to hapten sensitization, suppresses the induction of CHS in both humans (12) and mice (13). Because the morphology and function of epidermal LC is profoundly altered by UV-irradiation (14), their role in UV-induced immune suppression has been intensively studied (15). Hapten-bearing cells, isolated from the draining LN of UV-irradiated mice fail to induce CHS when transferred to normal recipient mice; rather they induce immunological tolerance (16). In vitro UV exposure of LC renders them incapable of presenting antigen to Th1 cells (17). Similarly, in vitro UV exposure, down regulates the expression of CD80 and CD86 on LC (18), suggesting that one mechanism by which UV exposure induces immune suppression and tolerance is by promoting the migration of immature dendritic cells to the LN. Studies by Kripke and colleagues indicate that UV-induced DNA damage, specifically pyrimidine dimer formation, depresses LC APC function (19, 20).

Although much is published about the role of LC in UV-induced immune suppression, two major caveats must be kept in mind when reviewing the data in this area. First, studies using in vitro irradiated LC may not adequately reflect what happens when LC are exposed to UV radiation in vivo. Second, in experiments where hapten-bearing DC are isolated from the LN of UV-irradiated mice, FITC is often used. Because FITC is a weak contact allergen, a relatively large volume is required to induce CHS, raising the concern that free hapten gets to the LN and is taken up and presented to T cells by resident DC.

Because the UV radiation in sunlight is the primary cause of skin cancer (21), and because UV-induced immune suppression is a major risk factor for skin cancer induction (22), a major focus of our research is to delineate the mechanisms involved. In view of the developing controversy regarding the role of LC in cutaneous immunology we decided to examine the mechanism(s) by which UV-irradiated LC induce immune suppression and tolerance. We were particularly interested in studying the immune function of LC that migrated from the skin to the LN. Rather than using fluorescent hapten to track migrating LC, we employed cell sorting to isolate LC from the LN of UV-irradiated mice. As expected, transferring UV-irradiated LC to normal mice suppressed CHS and induced immunologic tolerance. Moreover, we confirmed that LC are essential for inducing UV-induced immune suppression using LC-depleted mice. We noted that LC co-localized with Natural Killer T (NKT) cells in the LN and NKT cells isolated from the LN secreted IL-4. Moreover, when LC from UV-irradiated wild type mice were transferred into CD1d−/− or Jα-18−/− NKT deficient mice, a vigorous CHS reaction was generated. Our findings indicate that LC play an immune regulatory role and induce immune regulation by activating NKT cells.

Materials and Methods

Animals

Specific pathogen-free female C57BL/6J (CD45.2⁺) and congenic female CD45.1⁺ (B6.SJL-Ptprc^a Pepc^b/BoyJ) mice were purchased from the Jackson Laboratory. Dr. Dapeng Zhou (University of Texas, M. D. Anderson Cancer Center) supplied us with the Jα-18 deficient mice; CD1d deficient mice were obtained from Dr. Luc Van Kaer (Vanderbilt University Medical Center). Transgenic mice with the Langerin promoter driving expression of the diphtheria toxin receptor (Lang-EGFP-DTR) (6) were obtained from Dr. Bernard Malissen (Institut National de la Santé et de la Recherche Médicale). All procedures were reviewed and approved by the University of Texas, M.D. Anderson Cancer Center Animal Care and Use Committee and the Animal Experimentation Committee at Kobe University Graduate School of Medicine.

UV Source

UV radiation was supplied by a bank of 6 FS-40 sunlamps (National Biologic). These lamps emit a continuous spectrum from 270 to 390 nm (≈ 65% UVB, peak emission at 313 nm), as measured with IL-1700 research radiometer (International Light). The back of mice was shaved and then irradiated with UV. During irradiation, the mice were anesthetized and the ears were covered with black adhesive tape.

Antibodies and reagents

Monoclonal antibodies recognizing CD8α, CD11c, CD3, NK1.1, CD1d, IL-4, I-A/I-E, corresponding isotype controls, and secondary reagents were purchased from BD Bioscience. Monoclonal antibodies specific for I-A/I-E, CD80, CD86 and CD103 were purchased from eBioscience. Monoclonal antibodies specific for Langerin (CD207), were purchased from eBioscience (clone RMUL.2), Dendritics (clone 929F3) and the monoclonal recognizing the extracellular domain of Langerin (clone 205C1) (23) was purchased from AbCys. The monoclonal specific for epithelial cell adhesion molecule (Ep-CAM) was purchased from Biolegend. Alexa Flour 594 goat anti-rat IgG (H+L) and stereptavidin-conjugated Alexa Flour 350 were purchased from Invitrogen. Dinitrofluorobenzene (DNFB) was purchased from Sigma-Aldrich. The PE-conjugated CD1d tetramer loaded with PBS-57, an alpha-galactosylceramide analogue, as well as the empty tetramer control was obtained from the NIH Tetramer Facility.

Preparation of murine epidermal sheets and immunofluorescence analysis

Murine epidermal sheets were prepared as described previously (24). After fixation, the sheets were incubated at 23° overnight with rat anti-mouse CD207 (cloneRMUL.2). They were then washed with PBS, were incubated at 23° for 1h with Alexa Flour 594 goat anti-rat IgG (H+L). After washing with PBS, they were mounted using VECTASHIELD HardSet Mounting Medium with DAPI (Vector Laboratories). The samples were analyzed using a fluorescence microscope (Olympus). The number of LC found in the epidermis was determined by counting at least 10 fields/sample.

Generation of bone marrow chimeric mice

Bone marrow chimeras were prepared according to the method of Ginhoux et al (25). Six to 8 week old recipient C57BL/6 mice (CD45.2) were lethally irradiated with 1,200 rads (660 rads/per exposure; two exposures 3–4 h apart). The mice then received 10⁶ CD45.1 bone marrow cells via the tail vein. Six to eight weeks after reconstitution, chimerism was confirmed by measuring the number of CD45.1 cells in the peripheral blood, as described previously (8).

Suppression of CHS and tolerance induction by UV radiation

A modification of the procedure described by Miyauchi et al (26) was used to induce immunosuppression and a modification of the procedure described by Matsumura et al (27) was used to induce tolerance. The dorsal hair of the mice was removed and the mice were exposed to 1 kJ/m² of UVB. One to 35 days later, the mice were sensitized by applying 50 µl of 0.5% DNFB solution diluted in acetone/olive oil (4:1) to the UV-irradiated skin. Six days later, the thickness of the right ear was measured with a micrometer (Mitutoyo) and 5 µl of 0.2% DNFB solution was applied to the dorsal and ventral aspects of the ear. One day later, the thickness of right ear was re-measured. Three weeks later, the mice were re-sensitized by applying 50 µl of 0.5% DNFB solution to the un-irradiated shaved abdominal skin. Six days later, the left ear was measured and 10 µl of 0.2% DNFB solution was applied. One day later, ear swelling was measured. Langerhans cells were depleted by injecting Lang-EGFP-DTR mice with 1µg diphtheria toxin (DT, i.p. injection, Calbiochem).

In some experiments immune suppression was transferred by LC that had migrated to the LN. Inguinal LN from control or UV-irradiated mice were removed 7 days after UV and stained for CD11c, CD8α, and CD207. The CD11c⁺ CD8α⁻ CD207⁺ cells and CD11c⁺ CD8α⁻ CD207⁻ cells were sorted using a FACSAria (BD Biosciences). The sorted cells (1×10² to 2×10⁵ DC) were injected subcutaneously into the backs of wild type, CD1d-deficient, or Jα-18-deficient mice. Three days later the mice were sensitized with DNFB, and CHS was measured. In some experiments, 100 µg of anti-IL-4 Abs (clone 11B11) and isotype matched control Ab (eBiosciences) were injected i.p. 1 day after transferring the LC.

IL-4 production by NKT cells

CD4⁺ NK1.1⁺ lymph node cells from UV-irradiated or control mice were isolated by FACSAria. The cells (2 × 10⁵/culture) were stimulated for 16h with plate-bounded anti-CD3 (10 µg/ml) and soluble anti-CD28 (5 µg/ml). The supernatant was collected and IL-4 production measured by ELISA (R&D Systems).

Statistics

In the CHS experiments the mean change in ear thickness (left ear + right ear ÷ 2) was calculated for each animal in the group (N = 5). The change in thickness ± the SEM was then calculated for the group. In experiments measuring changes in lymph node dendritic cell numbers, the number of cells per lymph node for each individual animal (N= 3) was calculated. The mean ± SD was then calculated for the group. Statistical differences between the control and experimental groups were determined using an un-paired two-tailed Student’s t-test. (GraphPad Prism Software V4, GraphPad Inc).

Results

UV-induced migration and long-term survival of epidermal LC in draining lymph nodes

After UV-irradiation of the skin, epidermal LC density decreases. Our initial focus was to determine the duration of UV-induced LC depletion (Fig. 1). After treatment with relatively low doses of UV radiation (1000 to 2000 J/m²), epidermal LC numbers started to decrease 1 day post irradiation, and 7 days later almost no epidermal LC were observed. LC gradually re-populated the epidermis and normal numbers were noted 63 days post-irradiation (Fig. 1A & B). Lower UV doses (400 J/m²) had a similar effect, although the depression was not as severe and re-population occurred earlier (Fig. 1B).

Figure 1.

LC migration from the skin after UV radiation. A, Epidermal sheets were obtained from mice 1, 3, 7, 21 or 63 days after treatment with 1 kJ/m² of UVB radiation. They were stained with anti-CD207. Bar = 100 µm. B, Mice were exposed to 400 J/m² (-□-), 1 kJ/m² (-■-) or 2 kJ/m² (-▲-) of UVB radiation and the density of LC was determined. The data is expressed as % of control vs. normal LC density. Results are expressed as means ± SD. Each experiment was repeated twice; representative data are shown.

Conventional wisdom suggests that UV-irradiated LC migrate from the skin to the LN where they activate immune suppression. It is also clear that other dendritic cells, particularly dDC migrate to the LN following trauma to the skin. Moreover, others have shown that dDC express CD207 on their surface adding to the complexity of differentiating LC from dDC (28). With this in mind, we took advantage of the fact that Merad and colleagues demonstrated that LC precursors are radioresistant (25), whereas Hogquist and colleagues have shown that dDC precursors are radiosensitive (28). We reconstituted lethally x-irradiated CD45.2⁺ mice with CD45.1⁺ bone marrow cells, which results in a mouse whose LC are CD45.2+ and whose dDC are CD45.1⁺ (25). We then exposed the chimeric mice to 1 kJ/m² of UVB radiation. Seven days post irradiation, the mice were killed and the LN draining the dorsal skin were removed, and the phenotype of the CD45.1 and CD45.2 cells was analyzed (Fig. 2). We gated on the CD45.1⁺ or CD45.2⁺ cells, and then further gated on the CD8a⁻ CD11c⁺ cells to discriminate the resident DC from the infiltrating cells. The CD45.2⁺ cells (LC) were positive for CD207, CD24a, CD80, CD86 and MHC Class II, and a small number of cells expressed CD103 (Fig. 2A). The CD207+, CD103+ cells may reflect a small population of radio-resistant dDC, in the face of complete chimerism, recently described by Heni and colleagues (29). The CD45.1⁺ cells (dDC), did not express CD207 or CD24a, stained positively for CD103, CD86 and MHC Class II (Fig. 2B). Because CD24a and CD103 expression discriminate between LC and dDC (28, 30) , and in light of the fact that in this chimera, the LC are derived from radioresistant CD45.2⁺ precursors, we conclude that UV exposure induces the migration of both LC and dDC to the LN; the LC are CD207+ CD24a+ and the dDC are CD207−, CD24a−.

Figure 2.

UV-induced DC migration. Bone marrow chimeric (CD45.1 bone marrow x-irradiated CD45.2 recipients) mice were exposed to UV and 7 days later LN were removed. A, CD45.2⁺, CD11c⁺, CD8α⁻ cells were stained with anti-207 (clone 205C1), CD24a, CD103, CD80, CD86 and MHC Class II (IgG controls in red). B, CD45.1⁺, CD11c⁺, CD8α⁻ cells were stained with anti-CD207 (clone 205C1), CD24a, CD103, CD80, CD86 and MHC Class II (IgG controls in red). C. Wild type mice were exposed to UV and 7 days later LN were removed. The CD11c+, CD8α−, cells were stained with anti-CD207 (clone 929F3), CD24a, CD103 or Ep-CAM.

Because the protocol employed (UV-irradiation of chimeric mice) differs from most of the previous work described in the literature, we were concerned that the differences in the protocols may introduce unwanted artifacts. Therefore, we repeated the experiment using wild type C57BL/6 mice. The animals were shaved and exposed to 1 kJ/m² of UV radiation. Seven days later the mice were killed, and their lymph nodes were removed. As before we gated on the CD11c+, CD8α− population and stained these cells with CD207, CD24a, Ep-Cam, which is expressed on LC but not dDC (31) and CD103. Because the expression of the epitope recognized by anti-CD207 clone 929F3 does not change with maturation (28), this antibody was used in this experiment. We observed (Fig. 2C) that the CD 207+ cells co-expressed CD24a and Ep-CAM, but not CD103. From these data, using both chimeric and normal mice, we conclude that UV exposure activates the migration of CD207+, CD24a+, Ep-CAM+, CD80+, CD86+, CD103− LC into the draining LN.

Next we assessed the time course of LC migration into the LN. Mice were exposed to 1 kJ/m² of UV radiation. One to 35 days later, LN were collected. As expected we noted a significant increase in the cellularity of the LN 3, 7, and 14 days post irradiation (* P < 0.05 vs. time 0). This increase in cell number is undoubtedly due to the migration of a variety of myeloid cells into LNs, including dDC and LC (6, 8, 23), as well as other cells such as mast cells (32). However, because we are primarily focused on the role of LC in UV-induced-immune suppression, we concentrated on LC migration (Fig. 3B). We noted a significant increase in CD207 positive cells in the LN at days 7, 14 and 21 days post irradiation (Fig. 3C, *P < 0.05 vs. 0 time control). To test the relation between LC migration and immunosuppression, 2,4-dinitro-1-fluorobenzene (DNFB) was applied directly to the UV-irradiated skin (1 kJ/m²) at various times post irradiation. Immune suppression was most prominent when DNFB was applied 7 days after UVB radiation (Fig. 3D). When DNFB was applied 21 or 35 days after UV radiation, no suppression was generated, as the ear swelling response was not significantly different from the positive control (Day 0; no UV). To measure UV-induced immune tolerance, the mice were rested and then re-sensitized with DNFB 21 days after the first sensitization. Tolerance induction was most striking in mice where DNFB was initially applied 7 or 14 days after UV radiation (Fig. 3E). These results indicate that UV-induced immunosuppression and tolerance is correlated with a decrease in epidermal LC numbers from the skin (Fig. 1) and correlates with the migration of dendritic cells to the LN (Fig. 3).

Figure 3.

UV-induced LC migration and immune suppression. A, total number of cells in the draining LN was measured at various times after UVB exposure (1kJ/m²). *P < 0.05 Student’s T-test vs. time 0. B. Langerin (anti-CD207; clone 205C1) expression was measured on gated CD11c+, CD8α− cells. White histograms represent CD207 expression: grey histograms represent isotype controls. C. The number of LC cells in the LN is shown at different times post UVB exposure. *P < 0.05 vs. time 0. D, Suppression of the primary CHS reaction in UV-irradiated mice. Mice were irradiated with 1kJ/m² of UVB radiation, and at the time indicated, were sensitized with hapten. All mice were challenged with hapten (on the right ear) 6 days after sensitization; mean ear swelling ± SEM (N = 5) is shown. * P = 0.011; ** P = 0.15; † P = 0.003; ‡ P = 0.0028 vs. No UV). E, tolerance induction. The mice used in panel C were rested, and re-sensitized with DNFB 21 days after the first sensitization. They were challenged 6 days (left ear) after re-sensitization. Data represents mean ear swelling ± SEM (N = 5). * P = 0.0082; ** P = 0.0001; † P = 0.0001; ‡ P = 0.0001 vs. No UV. Each experiment was repeated twice; representative data are shown.

LC are essential for UV-induced immune suppression

To further confirm a role for LC in UV-induced immune suppression, we used Lang-EGFP–DTR mice, in which LC are depleted by administration of DT. Both epidermal LC and CD207⁺ dDCs are depleted following DT injection. Dermal DC repopulate the skin 3–7 days post treatment, whereas LC remain absent 14–28 days post injection (28). LC-deficient mice were exposed to UV 7 days post DT-treatment, and DNFB was applied to the skin 7 days after UV exposure. We saw no UV-induced immune suppression in these mice (Fig. 4A). In contrast, CHS was not affected by injecting non-irradiated mice with DT 14 days before sensitization. These results indicate that epidermal LC are essential for activating UV-induced immune suppression.

Figure 4.

Migrating LC induce immunosuppression by activating NKT cells. A, Lang-EGFP-DTR mice were injected with DT 7days before irradiation with 1 kJ/m² of UVB. The mice were sensitized with hapten 7 days post UV (DT + UVB + DNFB). Un-irradiated Lang-EGFP-DTR mice (DT+DNFB) received DT 14 days before sensitization. Wild type mice (UVB+DNFB) were irradiated with 1 kJ/m² of UVB, and sensitized 7 days after irradiation. The positive control mice (DNFB) were sensitized and were challenged. Negative control mice (negative) were not sensitized but were challenged (DNFB vs. UVB+DNFB; *P = 0.0012 vs. DNFB only control; N =5). B, Epidermal cell suspensions were prepared from normal mice or mice exposed to 1 kJ/m² of UVB. Langerhans cells were identified by gating on the CD11c⁺, I-A^high+ cells (blue box on dot blot). Blue lines indicate CD1d expression; red lines indicate isotype control. C, LN were collected from normal mice and mice exposed to UVB 7 days previously. The LC in the LN were identified by positive CD207 staining (blue scale), and the CD1d expression by these cells was measured. Blue lines indicate CD1d expression; red lines indicate isotype control. D, LN were sectioned and stained with CD3 (blue), NK1.1 (green) and Langerin (red) 7 days post UVB exposure. LC (red) and NKT cells (green + blue) expressing both NK1.1 and CD3 were found in close approximation in the T cell area of the node (arrows). Insert shows a close up of a LC-NKT interaction. Bar = 50 µm. E, Transferring LC isolated from the LN of UVB-irradiated mice suppresses CHS. 7 days post UV; migrating LC (UVB-LC) and dDC (UVB-dDC) were collected. LC isolated from untreated mice (NT-LC) were used as controls. The cells were injected into wild-type recipient mice and 3 days later the recipient mice were sensitized with hapten and CHS was measured. One group of recipient mice received no cells but were exposed to UV (UVB 1kJ/m²). Positive indicates mice that were sensitized and challenged. Negative indicates mice that were not sensitized but were challenged. *, P = 0.02 compared NT-LC control. **, P = 0.001 compared NT-LC control. F, LC from UV-irradiated wild type mice (UVB-LC) were transferred into CD1d−/−mice; 3 days later the recipients were sensitized with hapten and CHS was measured. UVB 1 kJ/m² shows the results in a group of control CD1d −/− mice exposed to UV radiation. F, LC from UV-irradiated wild type mice (UVB-LC) were transferred into Jα-18−/− mice; 3 days later the recipients were sensitized with hapten and CHS was measured. UVB 1 kJ/m² shows the results in a group of control Jα-18−/− mice exposed to UV radiation. Results are expressed as mean ear swelling ± SEM, N = 5. Each experiment was repeated twice; representative data are shown.

Migrating epidermal LC induce immune suppression by activating NKT cells

During our examination of the cell surface markers expressed on migrating LC, we noted that UV-induced the up-regulation of CD1d on the migrating LC (Fig. 4B & C). Because of the prominent role that CD1d plays in activating NKT cells (33), and in light of previous findings showing that the transfer of NKT cells from UV-irradiated mice to non-irradiated controls transfers UV-induced immune suppression (34), we examined the ability of migrating LC to activate NKT cells. Immunohistochemical analysis of LN from UV-irradiated mice demonstrated that CD207⁺ LC migrated to the T cell area of the node, and were found in the vicinity of NK1.1⁺, CD3⁺ T cells (Fig. 4D, insert). To determine if LC activate NKT cells we performed the following experiments. Langerhans cells were isolated from the LN of UV-irradiated mice. The LC were then injected into the subcutaneous space underlying the back skin of normal syngeneic mice. Three days later these mice were painted with hapten. Six days later the mice were challenged and CHS was measured (Fig. 4E). Injecting normal LC (NT-LC) had no effect on the CHS reaction (P = 0.16 vs. the positive control). Injecting LC isolated from the LN of UV-irradiated mice (UVB-LC) caused a dose-dependent suppression of CHS (*P = 0.02, ** P = 0.001 vs. normal LC control). Injecting dDC from UV-irradiated mice (UVB-dDC) did not suppress CHS, indicating that UV does not activate immune regulatory dDC. As expected, sensitizing mice through UVB-irradiated skin (UVB 1 kJ/m²) induced immune suppression. To test the hypothesis that immune suppression results from CD1d⁺ LC activating NKT cells, we transferred LC from UV-irradiated wild-type mice into NKT-deficient animals. We noted no immune suppression when LC from UV-irradiated mice were injected into CD1d−/− recipients, lacking all NKT cells (Fig. 4F), or Jα-18−/−mice, which lack invariant NKT cells (Fig. 4G). Nor did direct irradiation of NKT cell-deficient mice induce immune suppression, as reported previously (34). These data indicate that transferring LC isolated from the LN of UV-irradiated mice, activate immune suppression, and do so in an NKT cell dependent manner.

To measure immune tolerance, the mice were rested for three weeks, and then re-sensitized on the un-irradiated abdominal skin (Fig. 5). Langerhans cells isolated from the LN of UV-irradiated mice LC activated immune tolerance in wild-type mice (Fig. 5A) but not in the CD1d-deficeint or Jα-18-deficeint mice (Fig. 5B). These data indicate UV-irradiated LC induce immune suppression and tolerance, and do so by activating NKT cells.

Figure 5.

Migrating CD1d⁺ LC induce tolerance by activating NKT cells. 7 days post UV migrating LC (UVB-LC) and dDC (UVB-dDC) were collected. LC isolated from untreated mice (NT-LC) were used as a control. A, The cells were transferred to wild-type mice; B, CD1d−/− mice; C, Jα18−/− mice. Three days later they were sensitized with hapten. The animals were rested for three weeks and then re-sensitized with hapten and CHS was re-measured. UVB 1 kJ/m² shows the results in a group of recipient mice that received no cells but were exposed to UVB, as a positive suppressor control. *, P = 0.02 compared to NT-LC. Results are expressed as mean ear swelling ± SEM, N = 5. Each experiment was repeated twice; representative data are shown.

NKT cells isolated from the LN of UV-irradiated mice secrete IL-4

Natural killer T cells modulate immunity by secreting cytokines such as IL-4 (35). Antibodies to IL-4 block UV-induced immune suppression, and IL-4-deficient mice are resistant to the immunosuppressive effects of UV radiation (36, 37). Therefore, we measured IL-4 secretion by NKT cells (CD4⁺ NK1.1⁺) sorted from the LN of UV-irradiated mice (Fig. 6). The CD4⁺ NK1.1⁺ cells stained positively with a CD1d-tetramer confirming they are NKT cells (Fig. 6A). Ultraviolet exposure significantly enhanced IL-4 secretion by NKT cells (Fig. 6B; *P = 0.0006 UVB CD4+, NK1.1 + vs. NT CD4⁺NK1.1⁺). Transferring LC isolated from the LN of UV-irradiated mice (UVB-LC) into wild-type recipients suppressed CHS (Fig. 6C; P = 0.0005 UVB-LC + control Ab vs. positive + control Ab). When anti-IL-4 was injected into mice that were injected with UVB-LC, the suppression was reversed (**, P = 0.002 UVB-LC + control Ab vs. UVB-LC + anti-IL-4). Injecting anti-IL-4 had no effect on the induction of CHS in the positive control mice. These data indicate that migrating LC induce NKT to secrete IL-4, a cytokine critical for UV-induced immunosuppression.

Figure 6.

IL-4 production by NKT cells. A, LN from normal mice (no treatment) or UVB-irradiated mice (1 kJ/m²) were stained with NK1.1 and CD4. Cells gated on CD4⁺ and NK1.1⁺ (red rectangle) were stained with CD1d tetramer. Blue line indicate CD1d tetramer positive; and red line indicates cells stained with control tetramer. B, CD4⁺ NK1.1⁺ or CD4⁺ NK1.1⁻ lymph node cells from UVB-irradiated mice (UVB) or normal mice (NT) were isolated and stimulated in culture with anti-CD3 and anti-CD28. IL-4 secretion was measured by ELISA. *, P = 0.0006 compared to NT CD4⁺NK1.1⁺ cells. C, Anti-IL-4 blocks the transfer of immune suppression. Transferring LC isolated from the LN of UV-irradiated mice (UVB-LC) induced immune suppression (*, P = 0.0005 UVB-LC + control Ab vs. positive + control Ab). When UVB-LC were transferred into mice that received anti-IL-4, no immune suppression was noted (**, P = 0.002 UVB-LC + control Ab vs. UVB-LC + anti-IL-4). Results are expressed as mean ear swelling ± SEM, N = 5. Each experiment was repeated twice; representative data are shown.

Discussion

Conventional wisdom suggests that LC are the major antigen presenting cell of the skin, whose function is to take up antigen, migrate to the LN, and initiate an immune response. However, the observation that dermal DC and not LC are the primary antigen presenting cells in leishmaniasis and CHS (7, 8), and the fact that a vigorous CHS reaction is found in LC knockout mice (4, 5), suggests that our understanding of LC function is incomplete. It has been recognized for years that UV irradiation of the skin modulates the morphology and function of LC, transforming them from immune stimulatory cells into immune regulatory cells (38, 39). One suggested mechanism is the migration of immature LC to the LN, whose APC function is impaired, thereby resulting in tolerance (17, 18). In view of recent reports questioning the exact role of LC in the generation of, and/or the regulation of the immune response, we re-examined the role of LC in UV-induced immune regulation. For some time Langerin (CD207) expression has been used to identify LC in the LN. However, it is now clear that dDC also express Langerin (28). For this reason we took advantage of the observation that LC precursors are radioresistant, and dDC precursors are radiosensitive to generate bone marrow chimeric mice (CD45.1 bone marrow transplanted into CD45.2 X-irradiated recipient) to follow DC migration in response to UV exposure. The CD45.2⁺ cells were positive for CD207, CD24a, CD80, CD86 and MHC Class II, and expressed low levels of CD103. The CD45.1⁺ cells, on the other hand, did not express CD207 or CD24a, and stained intensely positive for CD103, CD86 and MHC Class II. Because CD24a and CD103 expression discriminate between LC and dDC (28, 30), and in light of the fact that in this chimera, the LC are derived from radioresistant CD45.2⁺ precursors, these data indicate that UV exposure induces the migration of both LC and dDC to the LN, but the CD207+ population found in the lymph nodes of UV-irradiated mice are LC. We confirmed the same migration pattern when wild type mice were UV-irradiated (Fig. 2C). The CD207+ cells simultaneously expressed CD24a, Ep-CAM and did not express CD103, indicating they are LC. Moreover, the findings reported here are similar to those recently reported by Yoshiki et al (31) and support their conclusion that UV exposure activates the migration of CD207+ LC, but not CD207+ dDC from the skin to the draining lymph node.

We found that LC rapidly migrate to the LN following UV exposure, and persist in the lymph nodes for up to 63 days. We found that the suppression of CHS and tolerance induction strongly correlated with LC migration to the LN. The LC isolated from the LN of UV-irradiated mice expressed I-A, CD80 and CD86 and CD1d, indicating that they are mature cells, suggesting that the induction of immune suppression and tolerance cannot be attributed solely to the migration of immature LC to the LN. Finally we found that injecting LC into normal recipient mice suppresses the induction of CHS in the recipient animals, and the induction of immune suppression is associated with the activation of IL-4 secreting NKT cells.

No immune suppression was noted in LC-deficient mice. At least 3 other groups have used Lang-EGFP-DTR mice to study the role of LC in UV-induced immune suppression; one reports LC are dispensable, whereas two other report they are essential (31, 40, 41). What is puzzling about the results reported by Wang and colleagues (40) is that although they report immune suppression, they see no epidermal LC depletion after exposing their mice to 1350 J/m² of UVB radiation. This is in contrast to the results reported here, which use a similar dose of UV, and findings frequently reported in the literature (13, 14). It appears that the major difference between Wang’s study, and the data demonstrating that LC are essential for UV-induced immune suppression reflects differences in UV dosimetry.

Another hallmark of UV-immunosuppression is the generation of regulatory cells in the LN (34, 42, 43). Because all the energy contained within UV radiation is absorbed in the upper layers of the skin, it is still unclear how the immunosuppressive signal is transmitted from the epidermis to the lymph nodes. Loser and colleagues have shown that Langerhans cell migration plays a role in the generation of classic T regulatory cells (44). In addition, Moodycliffe et al previously demonstrated that IL-4 secreting NKT cells can transfer UV-induced immune suppression (34) and El-Ghorr and Norval failed to note immune suppression in UV-irradiated IL-4-deficeint mice (37). Our present study expands on these previous observations by demonstrating that LC migration to the T cell area of the LN activates IL-4 secretion by NKT cells, and activates immune suppression. Whether the NKT cells are the only source of IL-4 is open to question. Although we note a significant increase in IL-4 secretion by NKT cells isolated from UV-irradiated mice, and we can completely block immune suppression by neutralizing IL-4, the increase in IL-4 production by NKT cells is only two-fold. Whether the NKT cells (or their secretions) are inducing other cells in the lymph node to release IL-4 remains to be seen.

Natural Killer T cell function and development is restricted by CD1d, which preferentially presents glycolipids to NKT cells (45). Although the identity of the natural ligand for NKT cells is still open to question, it is interesting to note that glucosylceramide is found in high concentrations in the skin, and UV exposure suppressed β-glucocerebrosidase activity, an enzyme which converts glucosylceramide to ceramide, thereby up-regulating the concentration of glucosylceramides in the epidermis (46). We suggest that CD1d⁺ LC carry the epidermal derived glycolipids to the LN, and activate NKT cells to secrete IL-4. This potential mechanism links epidermal damage with the induction of immune regulatory cells in the LN. This observation may have clinical relevance. Phototherapy with UVB has been used since the 1920’s as an effective treatment for psoriasis, an inflammatory Th-1 mediated skin disease. IL-4 therapy is also effective in treating psoriasis by inducing a Th2 response (47). Our findings provide a potential mechanism for the therapeutic effects of UV radiation in psoriasis.

Exposure to UV radiation in vivo, depresses Th1 immunity in part through the production of IL-4 and IL-10 (48). The data presented here provide a mechanism linking events in the skin to immune regulation. These findings may have implications beyond UV-induced immune suppression. In parasite infections, such as schistosomiasis, the induction of a Th2 reaction is protective. The most common route of human Schistosoma infection is through the skin. Recent findings have indicated the generation of an immune response to S. mansoni involves antigen presentation by CD1d⁺ dendritic cells (49), which activate NKT cells to secrete IFN-γ, IL-4, and IL-5 (50). Perhaps the migration of CD1d positive LC, as described here, plays a role in activating the immune response to parasites that enter via the skin.

In summary, our findings demonstrate that LC are essential for UV-induced immune suppression and immune tolerance, supporting the concept that these cells are important immunoregulatory elements. They demonstrate that CD1d⁺ LC migrate to the LN and activate NKT cells to secrete IL-4. These findings provide a mechanistic link between events that occur in the skin and the activation of immune regulation in lymph nodes.