Involvement of Dectin-2 in Ultraviolet Radiation-Induced Tolerance ¹

Ultraviolet radiation is one of the, if not the most, significant environmental factors affecting humans. Besides its well-known beneficial and indispensable effects for human life, UV, in particular the middle wave length range (290–320 nm, UVB), can be a hazard to human health by inducing skin cancer, premature skin aging, inflammation, and cell death (1, 2, 3, 4). In addition, UV exhibits immunosuppressive properties which may be relevant for photocarcinogenesis and the exacerbation of infectious diseases following UV exposure (5, 6). The immunosuppressive properties of UV are best illustrated by the inhibition of cellular immune reactions, such as contact hypersensitivity (CHS)³ (7, 8). Epicutaneous sensitization in mice is impaired when the hapten is applied onto skin which has been exposed to UV immediately before (7). In addition to the failure to generate hapten sensitization, tolerance develops because mice treated in this way cannot be resensitized against the same hapten at a later time point (7). UV-induced tolerance appears to be hapten-specific because the sensitization against other nonrelated haptens is not affected (7). Hapten-specific unresponsiveness can be adoptively transferred; injection of lymph node cells and splenocytes from UV-tolerized mice into syngeneic naive mice inhibits sensitization against the relevant hapten in the recipients (9). Consequently, it has been suggested that UV-induced tolerance is mediated via the induction of hapten-specific T suppressor cells, now renamed regulatory T cells (reviewed in Ref. 10). Although the cellular transfer of suppression and the hapten specificity were convincingly demonstrated almost two decades ago, the cells transferring hapten-specific unresponsiveness are still poorly characterized and their mode of action is unclear in major parts.

Using a subtractive cDNA cloning strategy, we previously isolated genes encoding for C-type lectin receptors (termed dectin-1 and dectin-2); dectin-1 is expressed by dendritic cells and macrophages (11). Preliminary in vitro data showed that a soluble receptor of dectin-1 binds to the cell surface of T cells, thereby delivering T cell costimulatory signals (11). Recently, dectin-1 was reported to bind to β-glucan, a polysaccharide synthesized in yeasts in an opsonin-independent fashion, indicating that dectin-1 is a pattern recognition receptor (12, 13). This ligand does not block binding of dectin-1 to T cells (12). Thus, dectin-1 has been well-characterized.

Dectin-2 is also a C-type lectin receptor with a type II configuration, consisting of, from the N terminus, a short cytoplasmic domain, a transmembrane domain, and an extracellular domain that contains a neck domain and a single carbohydrate recognition domain in the C terminus (14). Unlike dectin-1, dectin-2 expression appears more restricted to dendritic cells including epidermal dendritic cells (Langerhans cells); using a transgenic mouse bearing a dectin-2 promoter-driven luciferase gene, Bonkobara et al. (15) found the highest luciferase activity in the skin, lower levels in spleen, lymph node, and thymus and only background levels in nonlymphoid organs. Despite having a perfect C-type lectin motif, responsible for recognition of carbohydrates, the ligand for dectin-2 has not yet been identified and even its carbohydrate recognition has not been proven. Accordingly, the function of dectin-2 still remains to be determined.

To study the function of dectin-2, we have prepared a soluble dectin-2 receptor in the format of histidine-fusion proteins, consisting of the extracellular domains (aa 43–209) of dectin-2. This soluble dectin-2 (sDec2) should bind to the putative ligand and thus presumably inhibit dectin-2-triggered stimulation. We used this soluble receptor protein to examine whether dectin-2 is involved in the mediation of CHS. Surprisingly, injection of sDec2 affected neither the induction nor the elicitation of CHS, but blocked UV-induced suppression of CHS. Even more importantly, injection of sDec2 was able to break established tolerance. T cells transferring suppression were found to bind sDec2. Depletion of the sDec2-binding fraction of T cells was associated with a loss of transfer of suppression. Taken together, these data indicate that dectin-2 and its yet unidentified ligand may play a crucial role in the mediation of UV-induced immunosuppression. In addition, sDec2 may be a useful tool to further characterize the regulatory T cells mediating UV-induced tolerance.

Recombinant soluble protein of dectin-2 (sDec2) was generated from an expression vector encoding for 6xHis-tagged extracellular domain of dectin-2 (aa 43–209) as described previously (14). Recombinant soluble protein of dihydrofolate reductatse (DHFR) was derived from 6xHis-tagged DHFR and used as a control protein. Both proteins were expressed in Escherichia coli M15 and purified by using a Ni-NTA resin (Life Technologies, Grand Island, NY) followed by refolding.

C3H/HeN mice (8- to 10-wk-old) were obtained from Japan Clea (Hamamatsu, Japan). Mice were sensitized on day 0 by painting 25 μl of 0.5% 2,4-dinitrofluorobenzene (DNFB; solved in acetone-olive oil, 4:1) onto the shaved back. After 5 days, mice were challenged by painting 20 μl of 0.3% DNFB onto the left ear. The right ear was painted with the acetone-olive oil solution without DNFB. Ear swelling was quantified 24 h later using a spring-loaded micrometer. CHS was determined as the amount of swelling of the hapten-challenged ear compared with the thickness of the vehicle-treated ear in sensitized animals and expressed in millimeters × 10⁻² (mean ± SD). After 2 wk, mice were resensitized through nonirradiated abdominal skin and challenged on the right ear after 5 days. Each group consisted of at least seven mice. Data were analyzed by Student’s t test, and differences were considered as significant at p < 0.05.

Mice were exposed to 1000 J/m² UV on the shaved back daily for 4 consecutive days. FS-20 fluorescent lamps (Westinghouse Electric, Pittsburgh, PA) which emit most of their energy within the UVB range (290–320 nm) with an emission peak at 311 nm were used. DNFB was applied onto the surface of the irradiated skin area 24 h after the last UV exposure.

Donor mice were exposed to UV and sensitized with DNFB through UV-exposed skin as described above. Five days after sensitization, spleens and regional lymph nodes were removed, and single cell suspensions were prepared. T cells were prepared by passing the cell suspension through a nylon wool column twice, which yields ∼95% pure T cells as measured by FACS analysis using Abs directed against T cell markers (Thy 1.2, CD3). The cell number was adjusted to 1 × 10⁸ cells/ml, and 200 μl were injected i.v. into naive recipient mice. Recipients were sensitized 24 h later by epicutaneous application of DNFB on the shaved abdomen. After 5 days, mice were challenged on the left ear and ear swelling was evaluated 24 h later. For control purposes, identical numbers of cells obtained from untreated or DNFB-sensitized mice were injected.

Lymphocytes obtained from regional lymph nodes and spleens were incubated in nylon wool columns. T cells were first reacted with biotinylated-sDec2, -DHFR, or -sDec2 which was boiled. Biotinylation was conducted by use of the ECL Protein Biotinylation Module (Amersham Biosciences, Buckinghamshire, U.K.). One hour after the incubation of the cells with the biotinylated proteins, cells were incubated with streptavidin-coupled magnetic beads for 30 min. Magnetoseparation was performed by placing the tubes into a magnetic field (Dynal, Oslo, Norway) for 4 min. The supernatants containing the cells not binding to sDec2 were removed. Cells were washed and used for i.v. injection. Cells bound to the magnetobeads were detached by incubating cells overnight in cell culture medium at 37°C in a humidified atmosphere containing 5% CO₂. Magnetobeads which had spontaneously detached from the cells after overnight incubation were removed with a magnet. The remaining (dectin-2 binding) cells were harvested, washed, and adjusted to 2.5 × 10⁷ cells/ml for i.v. injection (200 μl/mouse). The efficacy of separation was determined by flow cytometry (FACSCalibur; BD Biosciences, Mountain View, CA).

T cells prepared from superficial regional lymph nodes and spleens were first incubated in PBS supplemented with 1% BSA on ice for 30 min. After extensive washing in PBS/BSA, T cells were incubated with biotinylated sDec2 on ice for 1 h. To perform competition of the binding, boiled sDec2, DHFR, and sDec2 were added to the reaction as a nonspecific and a specific competitor, respectively. After extensive washing in PBS/BSA, the samples were reacted with streptavidin-coupled with FITC on ice for 15 min. After washing, the samples were analyzed by flow cytometry (FACSCalibur). In some experiments, three-color staining was conducted with an allophycocyanin-conjugated Ab directed against CD4 (clone RM4-5; BD Biosciences, San Diego, CA), with a PE-conjugated Ab directed against CD25 (clone 7D4; BD Biosciences), and with biotinylated sDec2, which was finally reacted with FITC-coupled streptavidin (BD Biosciences).

T lymphocytes were prepared from mice which were tolerized by applying DNFB onto UV-irradiated skin and separated into sDec2⁺ and sDec2⁻ fractions by magnetobead separation. T cells (5 × 10⁶/ml) were incubated with dendritic cells (1 × 10⁶/ml) which were isolated from bone marrow of syngeneic naive mice as described previously (16). Coincubations were performed in the absence or presence of the water soluble DNFB analog 2,4-dinitrobenzene-sulfonic sodium salt (DNBS, 0.1 mM). Supernatants were collected 48 h after stimulation and tested for the amounts of IL-4, IL-10, IFN-γ, and TGF-β using ELISAs (Bender Medsystems, Vienna, Austria).

To elucidate whether interaction of dectin-2 with its putative ligand is of relevance in the generation of CHS, sDec2 was used. The soluble form of dectin-2 should bind to the putative though yet unknown ligand, inhibit the interaction of dectin-2 and therefore act as an antagonist of dectin-2. To test whether dectin-2 is involved in the mediation of CHS, mice received an i.v. injection of 50 μg of sDec2 or of DHFR. Three hours later, the animals were sensitized against DNFB. The ear swelling response caused by the ear challenge performed 5 days later was not affected by the injection of sDec-2 (data not shown). Likewise, ear challenge was not altered when sDec2 was injected i.v. into sensitized mice 3 h before challenge (data not shown). These data suggest that dectin-2 does not play an important role either in the induction or in the elicitation of CHS.

UV radiation is well known to inhibit the induction of CHS and to induce hapten-specific tolerance, although the detailed underlying mechanisms still remain to be determined. Therefore, we were interested in whether interaction of dectin-2 is involved in UV-induced immunosuppression. For that purpose, mice were sensitized against DNFB by epicutaneous application of the hapten on UV-exposed or unirradiated back skin. Ear challenge was performed 5 days later. In accordance with previous observations, ear swelling response was significantly suppressed in mice which were sensitized through UV-exposed skin (Fig. 1). To test whether dectin-2 is involved in UV-induced immunosuppression, mice were injected i.v. with various amounts of sDec2 3 h before application of DNFB onto UV-exposed skin. Mice which received 50 μg of sDec2 mounted an ear swelling response upon challenge comparable to that of positive control mice, despite the fact that sensitization was performed in UV-exposed skin (Fig. 1). sDec2 prevented UV-induced immunosuppression in a dose-dependent manner. Injection of 50 μg of the control protein DHFR did not affect inhibition of the induction of CHS by UV. Together, these data demonstrate that i.v. injection of sDec2 prevents UV-induced suppression of CHS, suggesting that interaction of dectin-2 with its putative ligand may play an important role in UV-induced immunosuppression.

Application of haptens onto UV-exposed skin does not only result in the failure to induce a CHS response, but also induces long-term immunosuppression because the respective mice cannot be sensitized at a later time against the specific hapten. This tolerance is hapten-specific because the animals can be sensitized at later time points against other unrelated haptens (7). Because both impairment of CHS and induction of hapten-specific tolerance are caused by the identical procedure, i.e., sensitization through UV-exposed skin, it was postulated for quite a long time that the same mechanisms are involved and that induction of tolerance is the direct consequence of the inhibition of CHS induction. However, there is accumulating evidence that the molecular basis of UV-induced tolerance is different from the mechanism responsible for UV-impaired induction of CHS (17). Whereas a key event in the suppression of CHS by UV appears to be the depletion of Langerhans cells from the epidermis (7, 18), hapten-specific tolerance seems to be due to the generation of T suppressor cells (reviewed in Refs. 10 and 19).

Therefore, we next checked whether injection of sDec2 also prevents UV-mediated tolerance. For this purpose, the mice which were sensitized through UV-exposed skin after injection of sDec2 were resensitized with DNFB on the abdomen 2 wk after the first sensitization. Five days after resensitization, the second challenge was performed on the right ear. There was no difference in the ear swelling response between sensitized mice and UV-exposed mice which had initially received sDec2 i.v. (Fig. 2). In contrast, mice which were sensitized through UV-exposed skin but had not received sDec2 were unresponsive to resensitization, indicating that tolerance had developed. Taken together, these data demonstrate that injection of sDec2 prevents the UV-induced inhibition of sensitization and the development of UV-induced tolerance.

We next were interested in whether injection of sDec2 is not only able to prevent the development of UV-induced tolerance but can also break tolerance once developed. To address this issue, mice were tolerized against DNFB by painting the hapten onto UV-exposed back skin. As demonstrated above, this procedure induces tolerance because these animals cannot be resensitized against DNFB 2 wk later. In one group of mice, sDec2 was injected i.v. before resensitization, while the other group received DHFR. Mice which were injected with DHFR did not show an ear swelling response after resensitization, indicating development of tolerance against DNFB (Fig. 3). In contrast, mice which received an i.v. injection of 50 μg of sDec2 revealed a specific ear swelling response after resensitization. Because tolerance had already developed at the time of injection of sDec2, one can conclude that injection of sDec2 is able to break established tolerance.

UV-induced tolerance is mediated via T cells with suppressor activity because adoptive transfer of T cells obtained from lymph nodes and spleens of mice which had been UV-tolerized against DNFB renders the recipient mice unresponsive against DNFB (9). The mechanism and the phenotype of these T cells with regulatory/suppressor function still remain to be determined. To address whether injection of sDec2 prevents the development of these regulatory T cells, donor mice were tolerized against DNFB by painting the hapten onto UV-exposed skin. Five days later, lymph nodes and spleens were removed and single cell suspensions were prepared. T cells (2 × 10⁷) were injected i.v. into naive syngeneic mice. Recipients were sensitized against DNFB 24 h later. Mice which received cells from UV-tolerized donors could not be sensitized against DNFB (Fig. 4). Mice which received cells from untreated donors were not impaired in their CHS response (data not shown). Transfer of cells from mice which received an injection of sDec2 before tolerization against DNFB did not suppress the CHS response against DNFB in the recipients. In contrast, transfer of suppression was observed upon injection of cells from DHFR-treated and UV-tolerized mice (Fig. 4). These data indicate that injection of sDec2 before tolerization either inhibits the development of UV-induced regulatory T cells or alternatively inhibits their suppressive activity.

The data so far suggested that interaction of dectin-2 with its putative ligand is crucial in the development and mediation of UV-induced tolerance. Because dectin-2 is expressed on dendritic cells, we postulated that the putative ligand may be expressed on T cells. To test this hypothesis, T cells were incubated with biotinylated sDec2 followed by incubation with streptavidin-coupled FITC. After washing, FACS analysis was performed. Although in naive and DNFB-sensitized animals ∼0.3 and 2.48% T cells, respectively, bound sDec2, the number of sDec2-positive T cells was increased (5.95%) in UV-tolerized animals (Fig. 5). The specificity of the binding was confirmed by competition with excess amounts of unlabeled sDec2, while the control protein DHFR did not compete (Fig. 5).

Both T cells (10) and NKT cells (20) have been recognized to mediate UV-induced immunosuppression, the latter being involved in the suppression of delayed-type hypersensitivity and tumor immunity (20). T cells with immunosuppressive properties have been renamed regulatory T cells (21). Different subsets of regulatory T cells have been described in mouse and man (reviewed in Ref. 22). An important subtype of regulatory T cells expresses CD4 and CD25 and comprises 8–10% of the peripheral CD4⁺ T cell subset in the mouse (23). In addition, these cells constitutively express the negative regulatory molecule CTLA-4 (24, 25). How these cells exert their suppressive activity is not yet clear. It is still under debate whether some regulatory T cells mediate suppression via the release of the immunosuppressive cytokines IL-10 and/or TGF-β, while others require cell-to-cell contact. There are recent indications that CD4⁺CD25⁺ T cells may also be involved in the mediation of UV-induced tolerance (26).

Due to the increased binding of sDec2 on T cells from UV-tolerized mice, we surmised that sDec2 binds to the regulatory T cells which are responsible for transferring suppression. Therefore, T cells obtained from UV-tolerized mice were incubated with biotinylated sDec2. Subsequently, cells were incubated with streptavidin-coupled beads. Finally, UV-tolerized T cells were depleted of the sDec2-bound fraction. The sDec2-depleted fraction of cells was injected into naive recipients, which were sensitized against DNFB 24 h later. Although transfer of bulk T cells from UV-tolerized donors inhibited sensitization in the recipients, animals which received the depleted fraction could be successfully sensitized against DNFB (Fig. 6 A). Depletion of sDec2-bound cells from T cells obtained from DNFB-sensitized donors did not affect induction of CHS in the recipients.

In turn, sDec2-bound T cells from UV-tolerized mice were obtained by magnetobead separation. Magnetobeads which had spontaneously detached from the cells after overnight incubation were removed with a magnet and the remaining cells were harvested, 5 × 10⁶ cells were injected i.v. into naive recipients, which were then DNFB-sensitized and challenged (Fig. 6 B). When compared with the transfer of equal numbers of bulk T cells obtained from tolerized mice, inhibition of sensitization was more pronounced upon injection of sDec2-binding cells. This indicates that sDec2-bound T cells are predominantly responsible for transfer of suppression. However, we cannot exclude that only a subfraction of sDec2-binding T cells transfers suppression, hence it is difficult to determine the minimum number of cells required for effective suppression.

To further characterize the phenotype of sDec2-bound T cells, FACS analysis was performed. T cells obtained from lymph nodes and spleens of either naive, DNFB-sensitized, or UV-tolerized mice were triple-stained with the Abs directed against CD4 and CD25, respectively, and with biotinylated sDec2. FACS analysis (Fig. 7) revealed an increase in the number of CD4⁺CD25⁺ T cells in UV-tolerized animals as compared with naive and sensitized mice. sDec2-binding cells were detected by incubating the cells with FITC-coupled streptavidin. sDec2-bound CD25⁺ T cells were almost exclusively observed in UV-tolerized animals, the same was observed for CD4. Together, these findings indicate that sDec2-bound T cells also express CD4 and CD25 and thus may belong to the group of CD4⁺CD25⁺ regulatory T cells. We are currently investigating whether these cells also express CTLA-4 which has been suggested to be involved in mediating UV-induced tolerance (25).

To elucidate the cytokine secretion of sDec2-bound T cells, T cells were obtained from mice which were tolerized against DNFB by UV and separated into sDec2⁺ and sDec2⁻ fractions. These subtypes were put into culture with bone marrow-derived DC. Cells were cultured for 48 h in the absence or presence of the water soluble DNFB analog DNBS. Supernatants were harvested and tested for the amounts of IL-4, IL-10, IFN-γ, and TGF-β (Table I). Stimulation of sDec2⁺ cells with hapten-coupled DC resulted in the induction of the release of IL-4, IL-10 and TGF-β, while IFN-γ was not affected. Cytokine induction was hapten-dependent because incubation of DC in the absence of DNBS did not induce the release of cytokines. In contrast, no induction was observed when sDec2⁻ cells were stimulated with DC either in the absence or presence of DNBS.

Finally, we tested whether injection of sDec2 into recipients of UV-tolerized T cells prevents transfer of suppression. Therefore, T cells isolated from UV-tolerized animals were transferred into naive recipients, which had been injected with sDec2 3 h before transfer. Mice which received T cells from tolerized mice could not be sensitized, while recipients of the same fraction of cells which in addition were treated with sDec2 revealed a normal CHS responses against DNFB (Fig. 8). This indicates that injection of sDec2 into recipients enables sensitization in the presence of regulatory T cells. When T cells from DNFB-sensitized animals were transferred, injection of sDec2 did not affect the onset of CHS in the recipients (data not shown).

In summary, the present data indicate that dectin-2 and its yet unidentified ligand may play a crucial role in the mediation of UV-induced immunosuppression. Interaction between these two molecules appears to be involved both in the inhibition of the induction of CHS by UV and in the mediation of UV-induced tolerance. Because sDec2 binds to the fraction of T cells which are crucial for transferring suppression, the putative ligand may serve as a future marker for UV-induced regulatory T cells. Hence, identification of the ligand for dectin-2 on the T cells is urgently required and ongoing in our laboratory.

However, the binding of sDec2 must not be regarded as a specific feature of the cells transferring UV-induced suppression because it cannot be excluded that only a minority of cells transferred are responsible for mediating hapten-specific unresponsiveness in the recipients. In contrast, interaction of dectin-2 with its ligand appears to be functionally relevant for mediating UV-induced immunosuppression because injection of sDec2 in vivo prevented the suppression of CHS by UV and even broke UV-mediated tolerance. It has been previously reported that established tolerance can be broken by the injection of IL-12 (27, 28). We do not know whether these two phenomena are related or independent, but we are currently investigating whether the binding sites for sDec2 on T cells are regulated by IL-12. In addition, injection of neutralizing Abs against CTLA-4 is also able to break UV-induced tolerance (29). In this context, it was proposed that triggering of CTLA-4 by B7 molecules expressed on the APCs induces release of IL-10 which subsequently inhibits resensitization, thereby rendering the respective animal tolerant to the hapten. Accordingly, T cells obtained from UV-tolerized mice were found to release high amounts of IL-10 upon in vitro stimulation with hapten-coupled APCs (29). It remains to be determined whether the release of IL-10 by T cells is suppressed in the presence of sDec2.

UV-induced regulatory T cells are still poorly characterized. A major obstacle in this context may be the poor proliferative capacity which makes cloning extremely difficult probably even impossible. Therefore, the detection of any further surface marker expressed on UV-induced regulatory T cells will help to better characterize these cells and to better understand their mode of action. The detection of the binding sites for dectin-2 as demonstrated in this study may add to this.