Prostaglandin E2-EP3 Axis in Fine-Tuning Excessive Skin Inflammation by Restricting Dendritic Cell Functions

Prostaglandin E2 (PGE2) is produced in the skin and is suggested to play a role in the regulation of cutaneous immune homeostasis and responses. However, the multifaceted functions of PGE2 continue to elude our understanding, especially because of the multiplicity of PGE2 receptors—EP1, EP2, EP3, and EP4. While cAMP-elevating EP4 is known to activate the functions of cutaneous dendritic cells (DCs), including Langerhans cells (LCs) and dermal DCs, the role of cAMP-suppressing EP3 in this process remains unknown. Here we demonstrated that an EP3 receptor selective agonist, ONO-AE-248, inhibited chemotaxis and co-stimulatory molecule expressions of DCs in vitro. A suboptimal dose of antigen was sufficient to induce contact hypersensitivity in EP3-deficient mice. Intriguingly, EP3 deficiency did not impair skin inflammation at all when the antigen dose was sufficiently high. EP3 limited the functions of cutaneous DCs only when the antigen dose was low. In contrast to EP4, the observed unappreciated function of EP3 may stabilize the cutaneous DCs to halt the impetuous response to a suboptimal dose of antigen. Taken together, PGE2-EP3 signaling is essential for fine-tuning excessive skin inflammation by restricting DC functions.


Introduction
The skin is continually exposed to multiple foreign antigens throughout life and it exerts antigen-specific immune responses in which skin dendritic cells (DCs) play an important role as antigen-presenting cells (APCs) [1]. DCs in the skin are roughly divided into three subsets: Langerhans cells (LCs) in the epidermis and Langerin + or Langerin -DCs in the dermis [2]. These cutaneous DCs play a central role in the immunity and tolerance of the skin. Responding to foreign antigen, cutaneous DCs migrate to draining lymph nodes undergoing terminal differentiation and maturation. Matured cutaneous DCs then activate naive T cells to induce antigen-specific effector/ memory T cells in the lymph nodes [3]. The migration and maturation of cutaneous DCs are, therefore, crucial for the initiation of specific immune responses in the skin.
Lines of evidence suggest that prostanoids, including prostaglandins (PGs), engage in this DC alteration step [4,5]. On exposure to physiological or pathological stimuli, arachidonic acid is liberated from cell membrane phospholipids and is converted to prostanoids, including PGD 2 , PGE 2 , PGF 2α , PGI 2 , and thromboxane A 2 , through cyclooxygenases-mediated oxygenation followed by respective synthases. Prostanoids are produced in large amounts during inflammation and they exert complicated actions, including swelling, pain sensation, and fever generation. Among the prostanoids, PGD 2 and PGE 2 are abundantly produced in the skin during the elicitation phase of contact hypersensitivity (CHS)-a murine model for allergic contact dermatitis [3,6,7]. Therefore, it is of interest to evaluate the roles of PGD 2 and PGE 2 on DC functions.
It has been reported that PGD 2 suppresses cutaneous DC functions via DP1 receptor [8], while it enhances these functions via CRTH2 [9]. PGE 2 is produced abundantly in the skin on exposure to antigen [10], and is supposed to play a key role in determining the direction of immune response. Indeed, PGE 2 affects an immune response differently in a contextdependent fashion, showing some inconsistency at first glance. This contradictory effect is partially explained by the complexity of the four subtypes for the EP-the type E prostanoid receptors for PGE 2 , i.e., EP1, EP2, EP3, and EP4, each of which couples a different type of G protein. EP1 mediates the elevation of intracellular Ca 2+ concentration to promote Th1 differentiation [11]. On the other hand, EP2 and EP4 couple Gs protein that activates the cyclic adenosine monophosphate (cAMP)-dependent pathway by activating adenylate cyclase. EP2 is a potent suppressor of T cell proliferation in vitro [12,13]. EP4 suppresses T cell proliferation in vitro [12][13][14] and reinforces immunosuppression by expanding the number of Treg cells in vivo [15]. However, in a contradictory manner, EP4 also initiates the CHS response by inducing the migration and maturation of cutaneous DCs [10]. EP3 couples the Gi protein that inhibits cAMP-dependent pathways. We previously demonstrated that EP3 inhibited CHS by restraining keratinocytes from producing CXCL1, a neutrophil-attracting chemokine ligand CXCL1 [16]. EP3 is highly expressed in cutaneous DCs; however, the role of EP3 in APCs has not been studied in detail.
In this study, we demonstrated that EP3 downregulated the functions of DCs and that CHS was induced in mPger3 (EP3)deficient (EP3KO) mice upon exposure to suboptimal doses of antigens. Our results suggest that EP3 signaling inhibits undesired skin inflammation by limiting the maturation and migration of cutaneous DCs.

Expression of EP3 in bone marrow-derived DCs
EP subtypes are differentially expressed in the organs depending on the cell types. While the role of cAMP-elevating EP4 is known to enhance the functions of cutaneous DCs, the role of cAMP-decreasing EP3 remains unclear. It has been reported that EP3 is widely expressed in immune cells in mice [17], such as DCs [17], macrophages [18], and B cells [19], but not T cells [20].
We first confirmed the expression levels of EP subtypes in bone marrow-derived DCs (BMDCs) by reverse transcription-PCR. The expression level of EP3 mRNA was the second to third among the four EP subtypes. It was about one hundredths of the level of EP4, which was the most abundant in BMDCs ( Figure 1a). However, we speculated that EP3 potentially modulates the functions of cutaneous DCs because the binding affinity of EP3 for PGE 2 is much higher than that of EP4 [21]. PGE 2 is known to promote the migration of cutaneous DCs to draining lymph nodes in an EP4-dependent manner during cutaneous inflammatory conditions when the local concentration of PGE 2 is high (~10 μM) [10]. Since the binding affinity of EP3 is much higher than that of EP4 [21], we hypothesized that PGE 2 -EP3 signaling may function when the local concentration of PGE 2 is low (pM scale), for example, under the steady states or upon innocuous stimuli. Therefore, to validate the physiological significance, we examined the effect of low-dose PGE 2 on the chemotactic activity of BMDCs to CCL21, one of the chemokines that induce the homing of DCs to draining lymph nodes. We found that as low as 10 pM of PGE 2 reduced the chemotaxis of BMDCs to CCL21 when the concentration of CCL21 is 30 and 100 ng/mL, but not 300 ng/mL (Figure 1b and Figure S1). These findings suggest that PGE 2 may have opposite effects on DC functions, possibly through EP3 signaling, which is consistent with the cAMPlowering effect of EP3, differing from EP4. Intriguingly, when the dose of CCL21 was high, the inhibitory effect of PGE 2 was not observed. Therefore, although the inhibitory effect of PGE 2 on BMDC chemotaxis was about 50%, we assume that this inhibitory mechanism is important for fine-tuning of skin homeostasis.

EP3 agonist inhibited migration and maturation of BMDCs in vitro
Thus, we investigated the effect of EP3 signaling on DC migration responding to CCL21. We used BMDCs instead of cutaneous DCs to exclude the possible influences of the contaminating EP3-sensitive epidermal keratinocytes. We placed BMDCs in the upper chamber of Transwell ® and counted the number of major histocompatibility complex (MHC) class II + CD11c + BMDCs that migrated to the lower chamber filled with CCL21. The addition of the EP3 agonist to the upper chamber inhibited the migration of BMDCs (Figure 1c). We measured the amount of intracellular cAMP in the presence or absence of the EP3 agonist after the treatment of 0.5 μM 3isobutyl-1-methylanthine (IBMX), which inhibits endogenous phosphodiesterase to fix the amount of the intracellular cAMP. We found that the EP3 agonist reduced the amount of cAMP to 50% of the IBMX treatment group (Figure 1d).
We also examined the effect of EP3 signaling on BMDC maturation. We cultured BMDCs with or without the EP3 agonist for 2 days and counted the MHC class II high and CD86 high populations, which correspond to the mature BMDCs. The number of total and mature BMDCs was significantly reduced in the presence of the EP3 agonist ( Figure 1e).

EP3 agonist inhibited migration and maturation of LCs in vitro
To evaluate the effect of EP3 signaling on cutaneous DCs, we prepared epidermal LCs as a representative of cutaneous DCs that play an important role in the development of epicutaneous sensitization [22]. Consistent with the finding in BMDCs, EP3 mRNA was observed in LCs under the steady and repeated hapten applied conditions by reverse transcription-PCR (Figure 2a, left). EP3 protein was also detected in LCs in the steady state of C57BL/6 (B6) mice by western blot analysis (Figure 2a, right). We then prepared epidermal cell suspensions from ears of B6 and EP3KO mice, and incubated them for 24 hours with or without the EP3 agonist. These cells were applied to the upper chambers of Transwell ® with lower chambers containing CCL21. We then counted the number of migrated LCs as CD11c + MHC class II + cells in the lower chamber after 3 hours. The migration of B6derived LCs was inhibited by the EP3 agonist, whereas that of EP3KO-derived LCs was not affected (Figure 2b), indicating that EP3 signaling specifically inhibits the migration of LCs. To examine the effect of the EP3 agonist on the maturation of LCs, we cultured epidermal cell suspensions from B6 mouse ears for 24 hours with or without the EP3 agonist. The expression levels of co-stimulatory molecules, such as CD80 EP3 Signaling Regulates the Cutaneous DC Functions PLOS ONE | www.plosone.org and CD86, and adhesion molecule CD54 were decreased by the addition of the EP3 agonist ( Figure 2c). Taken together, EP3 signaling suppressed the migration and maturation of LCs in vitro.

EP3 restricted migration of cutaneous DCs under suboptimal antigen stimulation in vitro
To further test our hypothesis that PGE 2 -EP3 signaling functions in vivo, we applied different doses of fluorescein-5isothiocyanate (FITC) onto the abdominal skin of mice under the steady state or upon innocuous/mild stimuli. Then we counted the number of FITC + MHC class II + cutaneous DCs BMDCs were treated with 0, 1, 10 and 100 pM PGE 2 and applied to a transwell. 100 ng/mL CCL21 was administrated to the lower chamber. Migrated BMDCs were identified as MHC class II + CD11c + subset in the lower chamber. The % input was calculated as follows: (the number of BMDCs migrated into the lower chamber)/(the number of BMDCs applied into the upper chamber) *100. (c) EP3 agonist attenuated the CCL21-induced migration of BMDCs. BMDCs were treated with 10 μM AE248 for 2 days and applied to a transwell. The migration assay was carried out the same way as previously descried. (d) Intracellular cAMP concentration. BMDCs were treated with 0.5 mM IBMX for15min prior to EP3 agonist treatment (30min). (e) The number of total (CD86 high MHC class II middle to high ) and mature (CD86 high MHC class II high ) BMDCs were counted after EP3 agonist incubation for 2 days (n=6). Each data represents the mean + SD. *p<0.05. doi: 10.1371/journal.pone.0069599.g001 that migrated from the skin to the draining lymph nodes after 96 hours of immunization (Figure 3a and b). At the dose of 2% that is commonly used in this FITC assay, there was no significant difference in the degree of migration between EP3KO and B6 mice in accordance with our previous report [10]. Intriguingly, however, the number of migrated FITC + cells was apparently increased in EP3KO mice when the concentration of FITC was 0.5% (Figure 3a and b).
We measured the PGE 2 levels in FITC-applied skin. We collected the skin samples using punch biopsy (8 mm in diameter) 24 hours after 0.5% FITC application. Skin sections were homogenized in PBS and PGE 2 levels in the supernatant were measured with a PGE 2 EIA kit (Cayman Chemical). PGE 2 levels with or without FITC applications were comparable between B6 and EP3KO mice (Table S1). These data suggest that the difference of migrated DCs between B6 and EP3KO mice depends not on the expression level of PGE 2 in the skin, but on the EP3 signaling after binding of PGE 2 to EP3 receptor. This indicates that EP3 attenuates the migration of cutaneous DCs only when the antigen is applied at a suboptimal dose. We further examined the number of migrated FITC + cutaneous DCs by dividing into Langerin + DCs (including LCs and Langerin + dermal DCs) and Langerindermal DCs, and found that both subsets were increased in EP3KO mice ( Figure 3c). Therefore, EP3 signaling restrains the migration of cutaneous DCs when the external stimuli are subtle.

Suboptimally primed CHS response was inhibited by EP3
We examined whether the restriction on cutaneous DC functions by EP3 is also involved in the CHS response that is a murine allergic contact dermatitis model. CHS is a cascade of sequential events that starts with the sensitization phase and is terminated by the priming of hapten-specific T cells. Reexposure of the hapten initiates the second phase of CHS, which is called the elicitation phase. It is well known that cutaneous DCs play a pivotal role in the sensitization phase by migrating into the draining lymph nodes to activate T cells.
By sensitizing abdomens with 50 µl of 0.5% 2,4-dinitro-1fluorobenzene (DNFB) five days before, we could elicit ear swelling in both B6 and EP3KO mice after 24 hours of challenge with 20 µl of 0.3% DNFB. Importantly, sensitization with 0.05% DNFB was suboptimal for normal B6 mice and did not elicit ear swelling after the challenge. However, the same sensitization did elicit ear swelling in EP3KO mice at a comparable level to that of optimally sensitized mice ( Figure  4a). Histological analysis confirmed significant inflammation in the ears of 0.05% DNFB-sensitized EP3KO mice (Figure 4b and Table S2). Accordingly, there was approximately twice the amount of IFN-γ transcripts in the draining lymph nodes of 0.05% DNFB-sensitized and challenged EP3KO mice compared to that of B6 mice (Figure 4c). Moreover, careful observation revealed that EP3KO mice exhibited mildly enhanced ear swelling and a mild increase in the amount of  (Figure 4a and c). This observation strongly indicates that EP3 engagement constitutively operates against undesirable CHS to a suboptimal dose of antigen exposure.
In addition, to assess the contribution of PGE 2 -EP3 signaling on DCs/LCs circulation, we measured the quantifications of the frequencies of DCs/LCs in skin and lymph nodes using B6 and EP3KO mice in the steady state. The frequency of LC in epidermal cell suspension from the earlobes, and resident DCs (MHC class II middle CD11c high ) and migrated DCs (MHC class II high CD11c middle ) in lymph nodes were comparable between B6 and EP3KO mice in the steady state (Table S3). Therefore, PGE 2 -EP3 signaling may not contribute DCs/LCs circulation between skin and lymph node in the steady state.
To further characterize the effect of EP3 signaling during the sensitization phase of CHS, we analyzed T cell subsets in the lymph nodes. After the challenge, the number of CD44 -CD62L + naïve T cells, CD44 + CD62L + central memory T cells, and CD44 + CD62L -effector memory T cells were markedly increased in the lymph nodes of EP3KO mice that were suboptimally sensitized 5 days before (Figure 5a and b). However, the proportion of each subset did not show significant difference between EP3KO mice and B6 mice (Figure 5a). Therefore, together with the previous report that EP3 is not expressed in T cells [20], we speculated that the effect of EP3 is mainly seen in the migration and maturation of cutaneous DCs but not in the development of T cell subsets.  IL-1, and IL-18 [3], however, it remains unclear how the homeostasis of cutaneous DCs is maintained in relation to CHS. Our results have demonstrated that PGE 2 -EP3 signaling plays a unique role in the regulation of the CHS response. PGE 2 -EP4 signaling promotes the migration and maturation of cutaneous DCs, thereby initiating the CHS response on exposure to antigens. On the contrary, the PGE 2 -EP3 signaling functions in the opposite direction, balancing the cutaneous immune homeostasis in a subtle way. When the amount of antigens does not reach the threshold for appropriate immune responses, the PGE 2 -EP3 signaling actively limits the migration and maturation of cutaneous DCs through Gi protein to avoid unwanted inflammation. Once the antigen dose crosses the threshold, the PGE 2 -EP4 signaling axis overcomes the restriction by PGE 2 -EP3 signaling in order to switch the mode of cutaneous DCs toward activation.
We do not yet know the molecular mechanism that determines which EP subtype will dominate in response to PGE 2 . As DCs express both EP3 and EP4, the cell surface expression level of EP3 and EP4 may be differentially regulated depending on the antigen dose. However, at the transcript level, the mRNA level of EP4 was at least tenfold higher than that of EP3 (see Figures 1a and 2b). Therefore, it is not easy to explain the mechanism simply by the expression level of these subtypes. Alternatively, there may be cross talk between EP3 and EP4. For example, EP3-coupling Gi may somehow over-rule EP4-coupling Gs under non-inflammatory conditions when the production of PGE 2 is low. On the other hand, once DCs are exposed to a large dose of antigens, EP4coupling Gs now dominates the relation (see Figure 6). In line with the above hypothesis, the binding affinity of EP3 for PGE 2 is much higher than that of EP4 [21].
EP3 is the only prostanoid receptor that couples Gi and functions in a cAMP-inhibitory manner. Other prostanoid receptors work in an either Ca 2+ stimulatory (EP1, FP, TP) or cAMP-stimulatory fashion (EP2, EP4, DP, IP). This multiplicity of EP subtypes makes PGE 2 the most versatile prostaglandin in vivo. Here we revealed another unexpected dual role of PGE 2 on the CHS response (see Figure 6). In the steady state, low-dose PGE 2 limits migration and maturation of cutaneous DCs through EP3 to halt impetuous response to suboptimal stimuli. Thus, PGE 2 -EP3 axis seems to exhibit fine-tuning excessive skin inflammation by restricting DC functions. This limitation is easily cancelled under inflammatory state by highdose PGE 2 , which now acts on EP4 to switch the state of cutaneous DCs to an activation mode.
The mechanism to initiate skin immune responses have been vigorously studied, but the mechanism how to keep skin homeostasis has not been revealed well. In this study, we focused on the role of DCs. On the hand, other possible candidates to maintain skin homeostasis include regulatory T cells (Tregs). In the absence of Tregs, mice lead to spontaneous skin inflammation [23] and enhanced CHS to hapten exposure [24][25][26]. It remains unclear whether PGE 2 -EP3 signaling on DCs modulates the induction of Tregs, which will be addressed in the future.
It has also been reported that PGE 2 -EP3 signaling suppressed conjunctivitis and airway inflammation by inhibition EP3 Signaling Regulates the Cutaneous DC Functions PLOS ONE | www.plosone.org of epithelial cells and mast cells [27,28]. Since DCs are widely distributed to the tissues, PGE 2 -EP3 signaling may exhibit antiinflammatory effects in other tissues, including the conjunctiva and respiratory tract. Therefore, to exploit this versatile PGE 2 -EP3/EP4 pathway can lead to a novel approach for treating skin inflammation, including allergic contact dermatitis.

Animals and reagents
BALB/c and B6 mice were purchased from Japan SLC (Hamamatsu, Japan). EP3KO mice were backcrossed more than ten times onto B6. Mice were maintained in specific pathogen-free facilities at the Kyoto University, and eight-to thirteen-week-old mice were used for experiments. All experimental procedures were approved by the institutional animal care and use committee of Kyoto University Graduate School of Medicine (Permit Number: 080150).
Epidermal cell suspensions were obtained from the earlobes of mice with trypsin treatment [30]. LCs were purified from epidermal cell suspensions as CD11c + cells (purity >70%) by means of auto MACS (Miltenyi Biotec, Auburn, CA, USA).
BMDCs were prepared as described [31]. In brief, 5 × 10 6 bone marrow cells were cultured in 10 cm tissue culture dishes Figure 6. Hypothesis of the dual roles of PGE 2 on cutaneous DCs. In the steady state when the concentration of PGE 2 is low, endogenous PGE 2 binds to EP3 preferentially (binding affinity of PGE 2 to EP3 is higher than EP4), resulting in the prevention of impetuous immune responses to innocuous stimuli. On the other hand, in the inflammatory state, abundant PGE 2 is produced by keratinocytes. High expression of PGE 2 binds to EP4, which overcomes the inhibitory effect of EP3 signaling and promotes the maturation and migration of cutaneous DCs to initiate skin immune responses.

Flow cytometry
Cell suspensions were prepared from lymph nodes by mechanical disruption on 70 μm nylon cell strainers (BD Falcon, San Jose, CA, USA). For flow cytometry, cells were prepared and stained with antibodies (Abs) as described previously [32].

Quantitative reverse transcription-PCR
Total mRNA was extracted with the SVTotal RNA Isolation system (Promega, Madison, WI). Target gene expression was quantified in a two-step reverse transcription-PCR. cDNA was reverse transcribed from total RNA samples using the TaqMan RT reagents (Applied Biosystems, Foster City, CA, USA). Murine EP1, 2, 3, and 4 and IFN-γ (Assay ID: Mm00443097_m1, Mm00436051_m1, Mm00441045_m1, Mm00436053_m1, Mm01168134_m1) expressions were quantified using TaqMan Gene Expression Assay in the ABI PRISM 7000 sequence detection system (Applied Biosystems). As an endogenous reference for these PCR quantification studies, β-actin gene expression was measured using the TaqMan β-actin control reagents. The relative expression was calculated using the 2 -ΔΔCT method [33]. The expression of the target gene normalized to an endogenous reference and relative to a calibrator is given by the formula 2 -ΔΔCT . Gene expression in untreated mice was used as a calibrator expression to calculate ΔΔCT.

cAMP assay
We treated 1 × 10 6 cells of BMDC with 0.5 μM IBMX (Wako, Osaka, Japan) for 15 minutes. These cells were further incubated with or without the EP3 agonist (10 μM) for 30 minutes, and cAMP levels in the culture supernatant were measured with ELISA (R&D Systems, MN, USA).

Chemotaxis assay and FITC-induced cutaneous DC migration
BMDCs and epidermal cell suspensions were tested for transmigration across uncoated 5-µm Transwell ® filters (Corning Costar Corp., Corning, NY, USA) for 3 hours to 100 ng/mL CCL21 (R&D systems) in the lower chamber [32]. The number of MHC class II + CD11c + cells in the lower chamber was counted as migrating cells by flow cytometry. The percent input was calculated as follows: (the number of cells migrated into the lower chamber)/(the number of cells applied to the upper chamber) × 100.
For FITC-induced cutaneous DC migration, mice were painted on their shaved abdomen with 100 μL of 0.5% FITC or 200 μL of 2% FITC dissolved in a 1:1 (v/v) acetone/dibutyl phthalate (Sigma-Aldrich) mixture, and the number of migrated cutaneous DCs into draining inguinal and axillary lymph nodes was enumerated by flow cytometry.

DNFB-induced CHS model
For the CHS model, mice were immunized by application of 50 μL of 0.5 or 0.05% (wt/v) DNFB in 4:1 (v/v) acetone/olive oil to their shaved abdomens on day 0. They were challenged on the right ear on day 5 with 20 μL of 0.3% (wt/v) DNFB 21 . Ear thickness was measured before and 24 hours after the challenge to assess inflammation.
To examine mRNA expression and histological examination, ears were collected after ear thickness measurement. For histological examination, tissues were fixed with 10% formalin in phosphate buffered saline and then embedded in paraffin. Sections with a thickness of 4 μm were prepared and subjected to staining with hematoxylin-eosin. The histological findings were evaluated for the severity and character of the inflammatory response using a subjective grading scale as reported previously [34].

Statistical analysis
Data were analyzed using an unpaired Student's t-test or Dunnett's multiple comparison tests. A p value of less than 0.05 was considered to be significant. Unless otherwise indicated, data are presented as means + standard deviations (SDs). Figure S1. Effects of PGE 2 on migration of BMDCs. Effects of PGE 2 on migration of BMDCs to CCL21 BMDCs were treated with 0, 1, 10 and 100 pM PGE 2 and applied to a transwell. 30 ng/mL CCL21 (left panel) and 300 ng/mL CCL21 (right panel) were administrated to the lower chamber. Migrated BMDCs were identified as MHC class II + CD11c + subset in the lower chamber. The % input was calculated as follows: (the number of BMDCs migrated into the lower chamber)/(the number of BMDCs applied into the upper chamber) *100 (n=4-5). Each data represents the mean + SD. *p<0.05. (TIF) Table S1. PGE 2 concentration in the skin. B6 and EP3KO mice were painted with 0.5% of FITC. The skin samples were collected 24 hours after 0.5% FITC application using skin punch biopsies (8 mm in diameter). The skin samples were homogenized in 1 mL of PBS, and PGE 2 levels in the supernatant were measured with a PGE 2 EIA kit (Cayman Chemical). Data indicate the mean ± SD of 3 mice.

Supporting Information
(DOC) Table S2. Histological evaluation of. CHS.Mice were sensitized with 0.05% of DNFB and challenged. Histology scores are calculated as the sum of four elements (inflammation, neutrophils, edema, and epithelial hyperplasia). Data indicate the mean ± SD of 5 mice. (DOC) Table S3. The frequency of DCs and LCs in the skin draining lymph nodes and skin. Earlobes and inguinal lymph nodes were collected from B6 and EP3KO mice. Epidermal cell suspensions and lymph node cell suspensions were prepared and subjected to FACS analysis for quantifications of the frequencies of resident and migratory DCs and LCs. Data indicate the mean ± SD of 3 mice. (DOC)