Cannabinoid Receptor 2 Modulates Maturation of Dendritic Cells and Their Capacity to Induce Hapten-Induced Contact Hypersensitivity

Contact hypersensitivity (CHS) is an established animal model for allergic contact dermatitis. Dendritic cells (DCs) play an important role in the sensitization phase of CHS by initiating T cell responses to topically applied haptens. The cannabinoid receptors 1 (CB1) and 2 (CB2) modulate DC functions and inflammatory skin responses, but their influence on the capacity of haptenized DCs to induce CHS is still unknown. We found lower CHS responses to 2,4-dinitro-1-fluorobenzene (DNFB) in wild type (WT) mice after adoptive transfer of haptenized Cnr2−/− and Cnr1−/−/Cnr2−/− bone marrow (BM) DCs as compared to transfer of WT DCs. In contrast, induction of CHS was not affected in WT recipients after transfer of Cnr1−/− DCs. In vitro stimulated Cnr2−/− DCs showed lower CCR7 and CXCR4 expression when compared to WT cells, while in vitro migration towards the chemokine ligands was not affected by CB2. Upregulation of MHC class II and co-stimulatory molecules was also reduced in Cnr2−/− DCs. This study demonstrates that CB2 modulates the maturation phenotype of DCs but not their chemotactic capacities in vitro. These findings and the fact that CHS responses mediated by Cnr2−/− DCs are reduced suggest that CB2 is a promising target for the treatment of inflammatory skin conditions.


Introduction
Allergic contact dermatitis (ACD) is an inflammatory skin reaction to subthreshold exposures to low-molecular-weight compounds called haptens [1]. Skin manifestations such as pruritic skin lesions

Reduced Contact Hypersensitivity upon Adoptive Transfer of Haptenized Cnr2 −/− Dendritic Cells
We utilized an in vivo adoptive transfer model to study the effect of CB1 and/or CB2 deficiency on the capacity of haptenized DCs to sensitize naïve mice for contact hypersensitivity (CHS). For this, we injected DNFB-haptenized DCs generated from the bone marrow (BM) of Cnr1 −/− /Cnr2 −/− , Cnr1 −/− , Cnr2 −/− or wild type (WT) mice into naïve WT recipients. Animals were subsequently challenged twice with DNFB and ear swelling was measured after 48 h (Figure 1a). DNFB-challenged WT mice showed allergic ear swelling after sensitization with haptenized WT DCs (Figure 1b). However, the CHS response was significantly reduced in mice sensitized with haptenized Cnr1 −/− /Cnr2 −/− DCs as compared to recipients injected with WT DCs. A reduced ear swelling reaction to DNFB was also observed in WT animals after adoptive transfer of Cnr2 −/− DCs, whereas induction of CHS was not affected in Cnr1 −/− DC-treated WT recipients. cells into Cnr1 −/− /Cnr2 −/− or Cnr1 −/− mice, but not Cnr2 −/− mice, induced exacerbated allergic responses when compared to WT recipients (Figure 1b). These findings indicate that CB1 affects the host immune response during CHS, while CB2 expression on DCs in the sensitization phase alters their ability to initiate CHS responses.
Thus, CB1 signaling in DCs is dispensable during naive T cell priming but is critical in host cells. We therefore focused our subsequent analyses on the specific role of CB2 in DC maturation and migration. Experiments were repeated two times with similar results. * p < 0.05, ** p < 0.01.

Reduced Expression of Chemokine Receptors on Cnr2 -/-and WT BM-DCs
It is well established that the chemokine receptors CCR7 and CXCR4 and their cognate ligands CCL19/CCL21 and CXCL12, respectively, are regulators of skin DC migration under inflammatory conditions [29,30]. We therefore assessed expression of CCR7 and CXCR4 on BM-DCs of both genotypes by flow cytometry. Levels of both chemokine receptors did not differ between WT and Cnr2 −/− DCs ( Figure 2). However, upon Toll-like receptor 4 (TLR4) stimulation with LPS, upregulation of CCR7 expression was significantly lower on Cnr2 −/− DCs when compared to WT cells. Furthermore, Cnr2 −/− DCs exhibited lower CXCR4 expression levels as compared to WT cells after TLR4 or TLR9 ligation with LPS or CpG oligonucleotide 1668 (CpG), respectively ( Figure 2). To investigate whether CB1 and CB2 also modulated the response of the host, haptenized DCs were injected into Cnr1 −/− /Cnr2 −/− , Cnr1 −/− , and Cnr2 −/− recipient mice as well as into WT controls. Again, recipient mice were subsequently challenged twice with DNFB ( Figure 1a). Adoptive transfer of WT cells into Cnr1 −/− /Cnr2 −/− or Cnr1 −/− mice, but not Cnr2 −/− mice, induced exacerbated allergic responses when compared to WT recipients ( Figure 1b). These findings indicate that CB1 affects the host immune response during CHS, while CB2 expression on DCs in the sensitization phase alters their ability to initiate CHS responses.
Thus, CB1 signaling in DCs is dispensable during naive T cell priming but is critical in host cells. We therefore focused our subsequent analyses on the specific role of CB2 in DC maturation and migration.

Reduced Expression of Chemokine Receptors on Cnr2 −/− and WT BM-DCs
It is well established that the chemokine receptors CCR7 and CXCR4 and their cognate ligands CCL19/CCL21 and CXCL12, respectively, are regulators of skin DC migration under inflammatory conditions [29,30]. We therefore assessed expression of CCR7 and CXCR4 on BM-DCs of both genotypes by flow cytometry. Levels of both chemokine receptors did not differ between WT and Cnr2 −/− DCs ( Figure 2). However, upon Toll-like receptor 4 (TLR4) stimulation with LPS, upregulation of CCR7 expression was significantly lower on Cnr2 −/− DCs when compared to WT cells. Furthermore, Cnr2 −/− DCs exhibited lower CXCR4 expression levels as compared to WT cells after TLR4 or TLR9 ligation with LPS or CpG oligonucleotide 1668 (CpG), respectively ( Figure 2). To determine whether CB2 affects migration of DCs, we next examined the chemotactic capacity of WT and Cnr2 −/− DCs using in vitro transwell assays. For this, LPS or CpG stimulated BM-DCs were loaded into the upper well with or without addition of CCL19 in the bottom well. While migration of DCs was greatly increased in the presence of CCL19, WT and Cnr2 −/− BM-DCs migrated toward CCL19 in equivalent numbers ( Figure 3). These data demonstrate that CB2 deficiency does not alter the chemotactic behavior of DCs in response to CCL19.

Reduced Expression of MHC Class II (MHC II) and Co-Stimulatory Molecules by Cnr2 −/− BM-DCs upon TLR Stimulation
A prerequisite for CHS is the maturation of DCs during their migration to the draining LNs following exposure to antigen. We therefore examined the influence of CB2 on the maturation of BM-DCs from Cnr2 −/− and WT mice by analyzing surface expression of MHC II and the co-stimulatory molecules CD40 and CD86 upon stimulation with LPS or CpG. We found reduced expression of MHC II on BM-DCs isolated from Cnr2 −/− mice in comparison to WT BM-DCs in the naïve state and upon stimulation. Notably, higher expression levels of CD40 were observed in naïve Cnr2 −/− BM-DCs when compared to WT BM-DCs. However, stimulation with LPS induced lower upregulation of CD40 on Cnr2 −/− BM-DCs when compared to WT cells ( Figure 4). In addition, CD86 expression was markedly enhanced in WT BM-DCs upon stimulation with CpG, while it was not affected in Cnr2 −/− BM-DCs. Thus, CB2 affects upregulation of MHC II and co-stimulatory molecules upon maturation of DCs, a process required for naïve T cell stimulation. To determine whether CB2 affects migration of DCs, we next examined the chemotactic capacity of WT and Cnr2 −/− DCs using in vitro transwell assays. For this, LPS or CpG stimulated BM-DCs were loaded into the upper well with or without addition of CCL19 in the bottom well. While migration of DCs was greatly increased in the presence of CCL19, WT and Cnr2 −/− BM-DCs migrated toward CCL19 in equivalent numbers ( Figure 3). These data demonstrate that CB2 deficiency does not alter the chemotactic behavior of DCs in response to CCL19. To determine whether CB2 affects migration of DCs, we next examined the chemotactic capacity of WT and Cnr2 −/− DCs using in vitro transwell assays. For this, LPS or CpG stimulated BM-DCs were loaded into the upper well with or without addition of CCL19 in the bottom well. While migration of DCs was greatly increased in the presence of CCL19, WT and Cnr2 −/− BM-DCs migrated toward CCL19 in equivalent numbers ( Figure 3). These data demonstrate that CB2 deficiency does not alter the chemotactic behavior of DCs in response to CCL19.

Reduced Expression of MHC Class II (MHC II) and Co-Stimulatory Molecules by Cnr2 −/− BM-DCs upon TLR Stimulation
A prerequisite for CHS is the maturation of DCs during their migration to the draining LNs following exposure to antigen. We therefore examined the influence of CB2 on the maturation of BM-DCs from Cnr2 −/− and WT mice by analyzing surface expression of MHC II and the co-stimulatory molecules CD40 and CD86 upon stimulation with LPS or CpG. We found reduced expression of MHC II on BM-DCs isolated from Cnr2 −/− mice in comparison to WT BM-DCs in the naïve state and upon stimulation. Notably, higher expression levels of CD40 were observed in naïve Cnr2 −/− BM-DCs when compared to WT BM-DCs. However, stimulation with LPS induced lower upregulation of CD40 on Cnr2 −/− BM-DCs when compared to WT cells ( Figure 4). In addition, CD86 expression was markedly enhanced in WT BM-DCs upon stimulation with CpG, while it was not affected in Cnr2 −/− BM-DCs. Thus, CB2 affects upregulation of MHC II and co-stimulatory molecules upon maturation of DCs, a process required for naïve T cell stimulation.

Reduced Expression of MHC Class II (MHC II) and Co-Stimulatory Molecules by Cnr2 −/− BM-DCs upon TLR Stimulation
A prerequisite for CHS is the maturation of DCs during their migration to the draining LNs following exposure to antigen. We therefore examined the influence of CB2 on the maturation of BM-DCs from Cnr2 −/− and WT mice by analyzing surface expression of MHC II and the co-stimulatory molecules CD40 and CD86 upon stimulation with LPS or CpG. We found reduced expression of MHC II on BM-DCs isolated from Finally, we investigated possible CB2-mediated effects on immunoregulatory mechanisms in DCs. For this purpose, we analyzed expression of programmed death-ligand 1 (PD-L1) and PD-L2 on BM-DCs since both PD-1 ligands play a regulatory role in immune responses to contact allergens. Blocking PD-L1/PD-1 with monoclonal antibodies has been shown to enhance CHS reactions [25,31], while interfering RNA-mediated silencing of PD-L2 on DCs inhibited the elicitation of CHS [32]. Here, we found comparable expression of PD-L1 and PD-L2 on unstimulated and LPS-stimulated Cnr2 −/− and WT BM-DCs (Figure 5a). Finally, we investigated possible CB2-mediated effects on immunoregulatory mechanisms in DCs. For this purpose, we analyzed expression of programmed death-ligand 1 (PD-L1) and PD-L2 on BM-DCs since both PD-1 ligands play a regulatory role in immune responses to contact allergens. Blocking PD-L1/PD-1 with monoclonal antibodies has been shown to enhance CHS reactions [25,31], while interfering RNA-mediated silencing of PD-L2 on DCs inhibited the elicitation of CHS [32]. Here, we found comparable expression of PD-L1 and PD-L2 on unstimulated and LPS-stimulated Cnr2 −/− and WT BM-DCs (Figure 5a).
Finally, we studied the production of IL-10 by activated Cnr2 −/− and WT BM-DCs. This anti-inflammatory cytokine has previously been shown to efficiently inhibit ear swelling responses in hapten-induced CHS when overexpressed in virally transduced DCs upon in vivo transfer [33]. In our study, we found equivalent levels of IL-10 in the supernatants of stimulated Cnr2 −/− and WT BM-DCs (Figure 5b), underscoring that neither expression of PD-1 ligands nor IL-10 production in BM-DCs in vitro is affected by CB2.
DCs. For this purpose, we analyzed expression of programmed death-ligand 1 (PD-L1) and PD-L2 on BM-DCs since both PD-1 ligands play a regulatory role in immune responses to contact allergens. Blocking PD-L1/PD-1 with monoclonal antibodies has been shown to enhance CHS reactions [25,31], while interfering RNA-mediated silencing of PD-L2 on DCs inhibited the elicitation of CHS [32]. Here, we found comparable expression of PD-L1 and PD-L2 on unstimulated and LPS-stimulated Cnr2 −/− and WT BM-DCs (Figure 5a).

Discussion
In this study, we investigated the role of CB1 and CB2 in CHS induction by DNFB-haptenized DCs in vivo and the specific influence of CB2 in maturation and migration of DCs in vitro. Our data demonstrate that the absence of CB2 reduced the potential of haptenized DCs to induce CHS responses in mice. Adoptive transfers of DNFB-haptenized Cnr2 −/− or Cnr1 −/− /Cnr2 −/− deficient DCs resulted in attenuated ear swelling while transfer of haptenized Cnr1 −/− DCs did not affect CHS in WT animals. These findings suggest that specifically CB2 affects DCs in a cell-intrinsic manner. Notably, host Cnr2 deficiency did not alter ear swelling responses, whereas Cnr1-deficient recipients showed enhanced allergic responses upon transfer of WT DCs. Thus, CB1 expression by host cells, but not adoptively transferred donor cells, mediates protective effects. These results are in line with our previous findings showing that lack of CB1 expression in host keratinocytes leads to exacerbated CHS responses associated with increased expression of chemokines including CCL8 and the alarmin thymic stromal lymphopoietin TSLP [21,27,28]. With these opposing roles of CB1 (host protection) and CB2 (maturation signal for haptenized DCs) in mind, we investigated changes in key functions of DCs that may be operative in CHS induction by haptenized DCs.
Various chemokines and their receptors have been functionally implicated in migration of DCs into the lymph node during CHS [34][35][36]. Accordingly, it has been demonstrated that CCR7 mediates entry of both dermal and epidermal DCs into the lymphatic vessels and thus acts as a master regulator of DC migration under steady-state and inflammatory conditions [29,37]. Additional studies have highlighted the functional involvement of the CXCL12-CXCR4 axis in CHS [30,38,39]. CHS responses were impaired by CXCR4 antagonist administration during the sensitization phase in mice indicating that CXCL12-CXCR4 engagement was required for migration of cutaneous DCs [30]. Here, we demonstrate that CB2 deficiency did not alter CCR7 and CXCR4 expression levels on unstimulated BM-DCs, but reduced TLR-mediated upregulation of both receptors on activated BM-DCs. This is surprising in the context of an earlier study demonstrating that activation of BM-DCs with a cocktail of TNF, IL-1β, IL-6, PGE2, and the CB2 agonist GP1a did not affect CCR7 upregulation. Here, Adhikary and coauthors harvested BM-DCs on day 7 of culture and further matured them for up to 2 days in the presence of the cytokine cocktail and PGE2 and the CB2 agonist [25]. In our study, however, BM-DCs were utilized at day 6 and stimulated with the TLR agonist LPS according to established maturation protocols. Both maturation protocols induce distinct gene expression profiles and immune related functions in BM-DCs [40]. Thus, the diverse findings from both studies can be explained by differences in the maturation states of BM-DCs.
Furthermore, Adhikary et al. showed that CB2 deficient BM-DCs matured with various cytokines migrated equally well towards CCL19 as WT BM-DCs [25]. In our transwell migration assays, both, unstimulated as well as LPS or CpG-stimulated Cnr2 −/− DCs, showed an equivalent migratory behavior towards CCL19 when compared to WT DCs. However, it is yet unresolved whether CCR7 expression levels potentially altered under in vivo conditions in activated Cnr2 −/− DCs may affect their migratory behavior in CHS.
Additionally, we found that Cnr2 −/− BM-DCs were impaired in TLR ligand-induced upregulation of MHC class II and the co-stimulatory molecules CD40 and CD86. It is possible that these phenotypic changes in Cnr2 −/− DCs affect their capacity to activate T cells in CHS. In this context, blockade of the CD40-CD40L pathway during sensitization has been shown to inhibit T cell mediated responses and thus induce tolerance in murine CHS [41]. CD40-CD40 ligand interactions have further been demonstrated to regulate migration of antigen-bearing DCs from skin to draining lymph nodes in vivo, ameliorating CHS responses in CD40L knockout mice after hapten sensitization [42]. Furthermore, CD86 expression on DCs has been reported to be required for induction of CHS. Accordingly, injection of anti-CD86 antibodies before DNFB sensitization inhibited CHS development associated with reduced upregulation of CD80 and CD86 on DCs in the lymph nodes [43,44]. This involvement of CD86 in CHS was further highlighted by a prior study demonstrating that topical application of cream-emulsified CD86 siRNA in mice after sensitization reduced CD86 expression of DCs in skin-draining lymph nodes and ameliorated the clinical manifestations of CHS [45]. Hence, it is possible that Cnr2 deficiency in haptenized DCs dampens induction of CHS responses due to reduced upregulation of co-stimulatory molecules, resulting ultimately in impaired induction of effector T cell responses [46].
In conclusion, our findings confirm the importance of CB1 on host cells for protection from CHS and identified CB2 expression on DCs as a factor contributing to the development of the disease. CB2 plays a DC-intrinsic role that affects CHS induction in vivo. Loss of CB2 signaling resulted in reduced upregulation of CCR7 and CXCR4 on activated BM-DCs but did not affect in vitro migration behavior of unstimulated BM-DCs in response to a CCR7 ligand. CB2 deficiency further impaired upregulation of MHC class II and co-stimulatory molecules in DCs under inflammatory conditions in vitro and may thus be operative in DC-dependent mechanisms involved in T cell activation in CHS. Thus, targeting CB2 signaling specifically in DCs has therapeutic potential for the treatment of atopic dermatitis, which represents a considerable burden on patients and healthcare systems.

Animals
Mice with a genetic deletion of the Cnr1/2 (Cnr1/2 −/− ), Cnr1 (Cnr1 −/− ), and Cnr2 (Cnr2 −/− ) gene on the C57BL/6J background [46] and their littermate controls were bred and housed in the Specific Pathogen Free (SPF) animal facility of the House for Experimental Therapy (University of Bonn, Germany) and ZTE Münster. All experiments were conducted according to the institutional and national guidelines for the care and use of laboratory animals and were approved by the local government authorities (Landesamt für Natur, Umwelt und Verbraucherschutz NRW, Germany, date of document: 17/04/2008).

Contact Hypersensitivity
Mature DCs were haptenized with 2,5 mM dinitrobenzene sulfonic acid (DNBS, MP Biomedicals, Solon, OH, USA), the water-soluble analogue of the obligate contact sensitizer 1-fluoro-2,4-dinitrobenzene (DNFB). For sensitization, naïve mice received two inguinal s.c. injections of 5 × 10 5 haptenized DCs. For elicitation of CHS, ears of mice were painted with 10 µL of 0.3% DNFB on day 5, a second challenge was performed on day 12. Ear thickness was measured 48 h hours after the second challenge using a spring-loaded caliper (Kroeplin, Schlüchtern, Germany). Ear swelling was calculated in each mouse as the difference in ear thickness between the unchallenged and challenged ear.

Transwell Migration Assays
GM-CSF generated BM-DCs were stimulated at day 5 of culturing with LPS (Sigma-Aldrich, Saint Louis, MO, USA) or CpG1668 (TIB MOLBIOL, Berlin, Germany) and 4 × 10 5 cells transferred at day 6 to the upper chamber compartment of a 5 µm pore size transwell plate (Corning, Kennebunk, ME, USA). The lower well compartments were filled either with media only or 200 ng/mL CCL19 (R&D Systems, Minneapolis, MN, USA). The concentration of 200 ng/mL recombinant CCL19 was selected in accordance with earlier studies demonstrating optimal chemotactic responses at these concentrations [41,48]. Cells were allowed to migrate for 4 h at 37 • C in 5% CO 2 . Cell counts of the migrated DCs harvested from the lower chambers were determined by FACS.

Elisa
IL-10 in cell culture supernatants was measured using ELISAs according to the manufacturer's instructions (Biolegend, San Diego, CA, USA). Detection limits for IL-10 were 2000 pg/mL.