Mast Cell Migration from the Skin to the Draining Lymph Nodes upon Ultraviolet Irradiation Represents a Key Step in the Induction of Immune Suppression¹

Due to their abundant expression of Fcε receptors and their ability to secrete histamine following IgE binding, mast cells have been traditionally associated with allergic-type immune reactions. However, newer findings indicate that mast cells influence a wide variety of nonallergic immune responses (1) and participate in inducing immune tolerance (2). Immunosuppression and tolerance are necessary counterbalances for hyperactive inflammatory-mediated immune responses, in that they inhibit the severity of allergy and prevent the onset of autoimmune disease. In contrast, unwarranted or ill-timed immune suppression can have significant consequences on the ability of the immune system to combat infections and destroy tumors. The UV wavelengths in sunlight are a prime example of an environmentally acquired immunosuppressant, and suberythemal UV doses are known to cause significant systemic immune suppression and induce tolerance (3). Although the DNA damaging properties of sunlight are well known, the mechanisms of how UV suppresses Th1-immune responses and induces tolerance are not as well understood.

Following UV exposure, a cytokine cascade that biases the immune response toward a Th2 reaction is initiated, which ultimately leads to the formation of CD4⁺CTLA-4⁺ regulatory T cells (4, 5). However, the cells and inflammatory mediators involved in the initial steps toward suppression and tolerance (i.e., those within the first hours following UV exposure) are still unknown. Hart et al. (6) and later Alard et al. (7) demonstrated that mast cells are required for both systemic and local UV-induced immune suppression, respectively. In these studies, mast cell-deficient mice were resistant to the immunosuppressive effects of UV radiation, and suppression was restored in knockout mice reconstituted with wild-type bone marrow-derived mast cells (BMMC).³ In addition, mast cell density in human skin correlates with susceptibility to both melanoma (8) and nonmelanoma skin cancers (9), suggesting that the immunomodulatory function of mast cells is likely to be important for the development of skin tumors. This is perhaps not surprising when one considers the wide range of inflammatory mediators and cytokines that mast cells have been shown to produce (10). Indeed, many of the inflammatory mediators released by mast cells including, histamine (10, 11), PGE₂ (12), serotonin (13), platelet-activating factor (PAF) (14, 15) TNF, IL-4, and IL-10 (16) are critical mediators of UV-induced immunosuppression. Grimbaldeston et al. (17) recently demonstrated that mast cell-derived IL-10 limits the skin pathology associated with contact dermatitis and chronic inflammation induced by UV exposure, again reinforcing the growing appreciation for the ability of mast cells to regulate inflammation and the immune response.

One hallmark of UV immunosuppression is the generation of suppressor lymphocyte populations. During the early phase (i.e., within hours of UV exposure) an IL-10-producing suppressor B cell is activated (18, 19), followed a few days to weeks later by CD4⁺ regulatory T cells (4, 20, 21). Mast cells are not only potent B cell activators (22, 23), they are capable of producing Th2-polarizing cytokines (24) that preferentially activate CD4⁺ Th2 cells (25). Most of these earlier studies used in vitro-cultured BMMC and therefore it is still not clear exactly how a mast cell in the periphery influences lymphocyte activation, although their ability to reach draining lymph nodes (DLN) where lymphocyte activation occurs would seem to be a necessary prerequisite. In this study, we show that UV exposure triggers mast cell migration to the DLN through CXCR4 expressed on mast cells and CXCL12 expressed on lymph node cells. Blocking mast cell migration into the DLN by a CXCR4 antagonist abrogates UV-induced immune suppression.

C57BL/6 wild-type mice, mast cell-deficient mice on a C57BL/6 background (Kit^W-sh/W-sh), and GFP⁺ mice (C57BL/6-Tg(UBC-GFP) 30Scha/J) were obtained from The Jackson Laboratory. The mice were housed in facilities approved by the Association for Assessment and Accreditation of Laboratory Animal Care International. The University of Texas M.D. Anderson Cancer Center Institutional Animal Care and Use Committee approved all of the animal procedures described here.

On day 0, the mice were exposed to an immunosuppressive dose of UV radiation (80 kJ/m² of solar- simulated radiation; 290–400 nm; containing ∼8 kJ/m² of UVB; 290–320 nm) supplied by a 1000 W xenon arc solar simulator (Oriel), as described previously (26). Four days later, the mice were sensitized by applying 50 μl of 0.3% 2,4-dinitro-1-fluorobenzene (DNFB; Sigma-Aldrich diluted in 4:1 acetone:olive oil) to the unirradiated, shaved abdominal skin. Six days later, the ears of each mouse were measured with a micrometer and the animals were challenged by applying 5 μl of 0.2% DNFB in the same diluent to the ventral and dorsal surface of each ear. Twenty-four hours later, the change in ear thickness (after challenge − before challenge ear thickness) was determined (18, 19).

In some experiments, the effect of UV exposure on mast cell-deficient mice reconstituted with wild-type BMMC was examined. Bone marrow stem cells were isolated from the femurs and tibiae of 6-wk-old C57BL/6 mice and then cultured at a concentration of 10⁶ cells/ml in complete RPMI 1640 supplemented with murine rIL-3 (10 ng/ml; PeproTech) and stem cell factor (10 ng/ml; PeproTech). Nonadherent cells were transferred to fresh culture medium twice a week for 4–5 wk, at which point >98% of viable cells were mast cells as verified by flow cytometry (CD45⁺CD117⁺FcεR1α⁺CD3⁻B220⁻) and positive staining for toluidine blue. A total of 1 × 10⁶ BMMC was injected into multiple sites underlying the dorsal skin of mast cell-deficient mice (6). Six weeks later, the mice were exposed to UV radiation as described above.

To activate BMMC in vitro, 10⁶ cells were incubated for 6 h with 5 μg/ml purified mouse IgE (Sigma-Aldrich) to cross-link Fcε receptors. BMMC were then washed and incubated in RPMI 1640 containing 0.5% BSA and 100 ng/ml DNP-keyhole limpet hemocyanin (Sigma-Aldrich) for 18 h. Seventy-two hours after activation, BMMC were labeled with Abs against CD117, FcεR1α, CXCR4 (BD Pharmingen), and CCR7 (eBioscience).

Twenty-four hours after UV exposure, the inguinal lymph nodes were removed, snap frozen in liquid nitrogen, and pulverized with a mortal and pestle. Control groups were shaved but unirradiated. Total RNA was extracted with TRIzol (Invitrogen Life Technologies) and further purified by treating with RNeasy RNA cleanup protocol (Qiagen). The concentration of isolated RNA was measured and 0.5 μg was converted to cDNA using the Retroscript RT Kit (Ambion). Twenty-five nanograms of cDNA was subjected to real-time RT-PCR using a sequence detector (model Applied Biosystems Prism 7500) and target mixes for CXCL12 and GAPDH (TaqMan Gene Expression Assay; Applied Biosystems). Cycle threshold (C_T) values for CXCL12 were normalized to GAPDH using the following equation: (1.8^{(GAPDH −CXCL12)} × 1000), where GAPDH is the C_T of each GAPDH control, CXCL12 is the C_T of CXCL12, and 1000 is an arbitrary factor to bring all values above 1. There were four mice in each group; RNA was isolated from each individual mouse.

Skin samples from control or UV-irradiated mice were embedded in paraffin and 7-μm serial sections were cut. One section was labeled for CXCR4 using a rat anti-mouse mAb (clone 2B11; BD Pharmingen) while the other section was stained for mast cells using toluidine blue. DLN from control or UV-irradiated mice were frozen in liquid nitrogen, 7-μm sections were cut, fixed, and then stained with toluidine blue. Care was taken to ensure that sectioning occurred in the same area of each individual lymph node. Lymph node mast cell density was determined by counting the total numbers of mast cells per lymph node section and dividing this count by the area of the lymph node section calculated using NIH Image J software (http://rsb.info.nih.gov/nih-image/).

Mice were exposed to different doses of UV radiation (0–80 kJ/m²) and, 24 h later, the inguinal lymph nodes were removed. Single lymph node cell suspensions were enzymatically digested with collagenase (400 U/ml) and DNase (300 U/ml) (Sigma-Aldrich) before labeling for mast cells by flow cytometry. The following Abs were used: CD45, CD3, CD4, CD8, CD11c, CD49b, NK1.1, Gr-1, CD117, and FcεR1α.

Mice were supplied with AMD3100 (Sigma-Aldrich) in their drinking water (60 μg/ml) beginning 2 days before UV exposure. For analysis of mast cell densities, AMD3100 was provided 2 days before UV and maintained throughout the experiment. In experiments where contact hypersensitivity (CHS) was measured, AMD3100-supplemented water was provided 2 days before UV radiation. Four days after UV irradiation and 1 day before hapten sensitization (5 days after UV), the mice were put on normal drinking water.

In the CHS experiments, the mean change in ear thickness (left ear + right ear ÷ 2) was calculated for each animal in each group. There were at least five mice per group. The change in thickness ± the SEM was then calculated for the group. Statistical differences between the control and experimental groups were determined using an unpaired two-tailed Student’s t test (GraphPad Prism Software version 4). In experiments measuring changes in lymph node mast cell numbers, there were at least three mice per group. The number of mast cells per mm² for each individual animal was calculated. The mean ± the SEM was then calculated for the group. Similarly, when RT-PCR was used to determine fold increases in chemokine mRNA levels, values were calculated from back skin samples isolated from four individual mice. The means and the SEM for each treatment group were calculated and statistical differences between the experimental groups were determined using an unpaired two-tailed Student’s t test. Representative experiments are shown; each experiment was repeated at least three times.

Exposing mice to 80 kJ/m² of solar-simulated UV radiation (290–400 nm) significantly (p = 0.006) suppressed contact hypersensitivity (CHS) (Fig. 1,a). This was accompanied by an increase in the size and cellularity of skin DLN 24 h after UV irradiation of back skin (Fig. 1,b). We verified that mast cells are critical mediators of UV immunosuppression by exposing various groups of mice to UV radiation. In contrast to their wild-type littermates, UV-irradiated mast cell-deficient mice (Kit^W-sh/W-sh) were resistant to the immunosuppressive effects of UV (Fig. 1 c). Injecting 10⁶ wild-type BMMC into the backs of Kit^W-sh/W-sh mice restored immune suppression (p = 0.0003 vs no UV control), thus confirming that mast cells were required for UV-induced immunosuppression.

Next, we examined the effect of UV radiation on lymph node mast cell density. Twenty-four hours after UV exposure, a substantial (75%) and significant (p = 0.0001 vs no UV control) increase in mast cell density was observed in the DLN of irradiated mice (Fig. 1,d). This mast cell increase was confirmed by flow cytometry. Mast cells were identified by double staining with anti-CD117 and FcεR1 (Fig. 1,e). There was a doubling in mast cell numbers in the skin DLN 24 h after exposure to 80 kJ/m² of UV radiation (p = 0.02 vs no UV control; Fig. 1 f). We did not observe a significant increase in mast cell densities or numbers in non-DLN (data not shown).

Next, we examined changes in dermal mast cell numbers following UV exposure. We observed a significant increase in mast cell density 6 h after UV radiation (Fig. 1 g, p = 0.004 vs 0 time), which returned to normal at the 24-h time point. This rise and fall in skin mast cell densities was not due to UV-induced changes to dermal thickness because there was no significant difference in skin area between unirradiated (0.4 ± 0.03 mm²) and UV-exposed groups (6 h = 0.3 ± 0.02 mm²; 24 h = 0.3 ± 0.01 mm²). Exposure to UV radiation, therefore, results in an initial increase in dermal mast cell density, with a return to baseline levels at 24 h. At the same time, we noted a concordant increase in the density and number of lymph node mast cells.

A limitation of the experiment described above is the difficulty in distinguishing between skin-derived and blood-derived lymph node mast cells. To differentiate between these two populations, we grafted skin from GFP⁺ mice onto the backs of congenic mast cell-deficient (Kit^W-sh/W-sh) mice. After allowing 5 wk for the skin grafts to take, the mice were exposed to UV radiation and, 24 h later, the draining (inguinal, brachial, axillary) as well as nondraining (popliteal) lymph nodes were excised and analyzed by flow cytometry. GFP⁺ cells were only found in lymph nodes draining the back skin (Fig. 2,a, R1). The DLN from mast cell^−/− mice grafted with GFP⁺ skin, but not exposed to UV radiation, were infiltrated by a small number of GFP⁺ cells (0.63%), which was not much greater than background (0.45%; data not shown). We attribute this small increase of GFP⁺ cells in unirradiated animals to the migration of dendritic cells from the graft to recipient DLN (27), which is supported by the fact that the majority of these cells (>97%) were positive for CD11c, CD4, CD8, or CD19 (Fig. 2,b). Gating on CD11c⁻CD3⁻CD4⁻CD8⁻CD19⁻ cells (Fig. 2,b, R2) revealed a population of GFP⁺ cells found exclusively in the DLN of UV-irradiated mice (Fig. 2,b, R2). This subset of GFP⁺ cells also had high forward and side scatter profiles and consistent with a mast cell phenotype were CD117⁺FcεR1α⁺ (Fig. 2 c). Since the transplanted skin was the only source of GFP⁺ cells, we conclude that UV exposure triggers the migration of mast cells from the skin to DLN.

The chemokine receptor CXCR4 is abundantly expressed on cultured mast cells and these cells migrate toward the CXCR4-specific ligand CXCL12 (28). Although many cells in skin express CXCR4⁺ (Fig. 3, b and d), toluidine blue-positive mast cells were CXCR4⁺ in serial skin sections (Fig. 3, c–f). Similarly, in vitro- activated BMMC express CXCR4⁺ (Fig. 3 g).

UV exposure significantly increased the expression of the CXCR4-specific ligand CXCL12 (stromal-derived factor 1α) mRNA in DLN (Fig. 4,a). In contrast, UV exposure had no effect on the expression of CXCL12 in the skin (Fig. 4 b). Thus, UV radiation establishes a CXCL12 chemokine gradient potentially directing CXCR4⁺ cells toward CXCL12⁺ DLN.

In the experiment described above (Fig. 4,a), whole lymph nodes were used to isolate mRNA for CXCL12 analysis. It was unclear which cells within the DLN up-regulates CXCL12. To address this question, we sorted CD19⁺, CD4⁺, and CD8⁺ lymphocytes by FACS (>98% purity) and isolated mRNA from the purified cells. As can be seen in Fig. 4,c, CD19⁺ B cells were the major source of CXCL12 in lymph nodes and UV exposure up-regulates the expression of CXCL12 on B cells. Lymph node high endothelial venules also express CXCL12 (29): therefore, nonhematopoietic-derived populations (i.e., CD45⁻ cells) were also analyzed. As expected, these cells also expressed CXCL12, although no difference in expression between control and UV-irradiated groups was observed (Fig. 4 d). These results demonstrate that the increase in lymph node CXCL12 expression observed after UV radiation was predominantly B cell derived.

To determine the significance of the UV-induced CXCL12 production by B cells, we analyzed the localization of the infiltrating mast cells. Using a two-step immunohistochemical staining procedure, we were able to visualize both CD19⁺ B cells and toluidine blue-stained mast cells (Fig. 5). Most of the resident mast cells in DLN from unirradiated animals were found in the subcapsular sinus and medulla regions of the node (Fig. 5,a). In contrast, mast cells in the DLN of UV-irradiated animals were often found in close association with CD19⁺ B cells (Fig. 5, b and c). We quantified this association by counting the number of B cells that were in physical contact with mast cells (Fig. 5,d). In resting lymph nodes isolated from unirradiated mice, the majority of mast cells (>0%) were not associated with B cells. UV irradiation significantly enhanced mast cell-B cell interactions so that almost half of all mast cells in the DLN were in direct physical contact with B cells. Indeed, mast cells were sometimes found deep within the B cell follicles (Fig. 5, b and c). Almost one in four mast cells (24.3 ± 4.7%) in the DLN of UV-exposed mice were in direct contact with five or more B cells and a significant number (almost 5%) were in contact with more than nine B cells (Fig. 5, c and d). In comparison, only 1 in 50 mast cells in resting DLN (2.3 ± 0.7%) were in contact with 5 or more B cells and no lymph node mast cells were observed to be in contact with >9 B cells.

An UV-induced CXCL12 gradient toward the B cell areas of DLN, combined with mast cell CXCR4 expression, suggested that this chemokine pathway maybe driving the UV-induced mast cell migration. To investigate this possibility, we treated mice with AMD3100, a CXCR4 antagonist (30). Normal mice treated with AMD3100 (Fig. 6,a, filled bars, No UV) generate a CHS reaction that is not statistically different (>0.05) from that found in mice maintained on normal drinking water (Fig. 6 a, open bars, No UV). As expected, UV suppressed CHS in mice maintained on normal drinking water (p = 0.015 vs No UV control). In contrast, when mice maintained on AMD3100-supplemented drinking water were exposed to an immunosuppressive dose of UV radiation, no immune suppression was noted (p > 0.05 UV vs No UV).

The effect of AMD3100 on UV-induced mast cell migration is found in Fig. 6,b. AMD3100 was supplied in the drinking water for the entire experiment. Twenty-four hours following UV exposure (80 kJ/m²), the mice were killed and their lymph nodes were removed and stained with toluidine blue. As before, UV exposure causes a significant increase (p = 0.02) in DLN mast cell density (Fig. 6 b, open bars, UV vs No UV). However, when the AMD3100-treated mice were exposed to UV radiation, no increase in lymph node mast cell density was observed (p > 0.05). These findings indicate that AMD3100, a drug known to interfere with the binding of CXCL12 to its receptor CXCR4 blocks UV-induced immune suppression and interferes with the ability of mast cells to migrate into the DLN, supporting our hypothesis that mast cell migration to the DLN is a critical step in the pathway leading to immune suppression.

UV exposure also depletes epidermal Langerhans cell (LC) numbers in the skin, which is thought to be responsible for the ability of UV to inhibit local immune responses (31). Because LC express CXCR4 (32) it was possible that AMD3100 might prevent UV-induced immune suppression (Fig. 6,a) by interfering with LC migration. To rule out this possibility, we prepared epidermal sheets from the backs of AMD3100-treated, UV-irradiated mice. Counting the density of IA^b+ LC revealed that AMD3100 had no effect on the ability of UV radiation to alter LC morphology or deplete LC from the skin (Fig. 7). This indicates that LC migration is not involved in AMD3100-induced abrogation of UV-induced immune suppression.

The UV radiation present in sunlight damages DNA, induces inflammation, and suppresses the immune response, including the rejection of highly antigenic sunlight-induced skin cancers (21, 33). Exposure to suberythemal doses of UV radiation is all that is required to damage DNA and induce immune suppression, and humans are frequently exposed to these doses on a regular basis (34). This makes sunlight one of the most significant and potent human environmental carcinogens and human immunosuppressant. The precise mechanism by which UV suppresses antitumor immunity is still unknown, even though understanding this process is crucial to our ability to design new treatment regimens aimed at reducing the incidence of skin cancer. A number of different cell types are known to be involved including, dendritic cells (31), immunoregulatory T cells (4, 21), suppressor B cells (18, 19), NKT cells (20), macrophages (35), and mast cells (6). What remains to be shown is how inflammatory events in the skin (i.e., UV exposure) affect the induction of regulatory cells in distant lymphoid tissues, leading to Ag-specific immune suppression and tolerance.

Exposure to UV radiation induces systemic immune suppression. This is illustrated by the fact that UV exposure at one site will suppress the immune response to hapten or Ags introduced at a distant nonirradiated site (16). Because the skin effectively absorbs UV radiation and none of the UV wavelengths penetrate to the DLN, it is still not entirely clear how the suppressive signal is transmitted from the skin to the immune system. In this study, we present data supporting a novel mechanism by which dermal UV exposure induces immune suppression, mast cell migration from the skin to the DLN. Early after UV exposure, we noted a modulation of mast cell density in the skin, and, 24 h after UV exposure, we observed a doubling of lymph node mast cell density. When skin from GFP⁺ mice was grafted onto mast cell-deficient animals and the donor grafts were exposed to UV radiation, we observed the appearance of GFP⁺ mast cells in the lymph nodes of the recipient mast cell-deficient mice, confirming the hypothesis that UV irradiation is triggering the migration of mast cells from the skin to the DLN. The significance of UV-induced mast cell migration was highlighted by the fact that mast cell migration was required for the UV-induced immune suppression. When we used the CXCR4 inhibitor AMD3100 (36), also known as Mozobil in phase III clinical trials, to block CXCR4 binding to CXCL12, we blocked UV-induced mast cell migration and prevented UV-induced immune suppression.

The use of AMD3100 as an inhibitor of mast cell migration and immunosuppression is novel and, to our knowledge, has never been reported before. Unfortunately, verifying these results by UV-irradiating CXCR4^−/− mice is not possible due to embryonic lethality. Similarly, reconstituting mast cell knockout mice with CXCR4^−/− embryonic liver-derived mast cells is not possible because as we found, this results in the repopulation of both the skin and the DLN (data not shown). Finally, although the use of neutralizing anti-CXCR4 Abs might confirm these results, there are questions surrounding Ag specificity, the potential for agonistic effects, and problems associated with CXCR4 heterogeneity that could result in the Ab inhibiting one cell population over another (37), thus confusing the interpretation of such an experiment.

The creation of a chemokine gradient is necessary for directing cellular traffic. In vitro studies have established that mast cells express CXCR4 and migrate toward CXCL12 (28), although demonstrating the existence and importance of a CXCL12 gradient in vivo has not been shown. Mast cells in the skin were found to express CXCR4 and UV radiation increased the expression of the CXCR4-specific chemokine CXCL12 in the DLN. It is not clear how UV exposure sets up a CXCL12 chemokine gradient in the DLN, although we suggest that the cytokines and biological response modifiers released by keratinocytes after UV irradiation may play a role. For example, Silva et al. (38) showed that oxidized lipids, including PAF, induced the expression of a wide range of chemokines. Because of the critical role PAF plays in UV-induced immune suppression (14, 15), it is tempting to speculate that PAF is driving UV-induced CXCL12 expression. Another possible CXCL12 trigger might be the multitude of cytokines released following UV exposure including IL-4, IL-10, and TNF (39). TNF for example has been shown to increase CXCL12 production in osteoblasts (40).

It has recently become clear that mast cells not only mediate allergic type immune responses but also have the capacity to influence adaptive immune responses (1) and even induce tolerance (2). The physical separation of these cell populations (lymphocytes being activated in lymphoid tissues and mast cells residing predominantly in the periphery) may be the reason that the immunomodulating function of mast cells has received little attention. However, mast cells can migrate to sites other than the periphery. Ag sensitization induces dermal mast cell migration to DLN (41). Using experimental allergic encephalomyelitis as a model of multiple sclerosis, Tanzola et al. (42) observed that mast cells only migrated into the lymph nodes after the induction of the disease state. More recently, using a model of glomerulonephritis, it was shown that mast cell accumulation in the lymph nodes (but not the kidneys) was an essential feature of the ability of mast cells to inhibit disease progression (43). We extend these observations by indicating that a ubiquitous environmental carcinogen, UV radiation, activates mast cell migration to lymph nodes. Perhaps, then it would be more accurate to conclude that mast cells migrate toward sites of inflammation and that the DLN can be considered one such site. In our model, UV radiation not only induces inflammation locally in the skin, but also in the DLN. At early time points after exposure (i.e., 6 h), mast cells migrate into the skin, but at 24 h after UV radiation, CXCL12 is increased in DLN redirecting mast cells from the skin to “inflamed” hypertrophic nodes. Further supporting this hypothesis was the fact that mast cells did not migrate into uninflamed, non-DLN, or to unexposed skin (data not shown).

In summary, our findings indicate that UV-induced mast cell migration from the skin into the DLN represents a critical step in the induction of immune suppression. Blocking mast cells migration, by interfering with CXCR4/CXCL12 interactions blocks both mast cell migration and the induction of immune suppression. We note increased migration of dermal mast cells into B cell regions of the lymph node, suggesting this may be the mechanism by which tolerance-inducing, IL-10-secreting immunoregulatory B cells are activated (18, 19). These findings support the growing appreciation for the ability of mast cells to regulate adaptive immune reactions. We suggest that mast cell migration represents a critical mechanism for transmitting immunoregulatory signals from the periphery to the immune system after exposure to dermal immune modulating environmental toxins.

We thank Nasser Kazimi for help with the skin grafting experiments and Prof. Yong-Jun Liu for comments and critical review of this manuscript.

The authors have no financial conflict of interest.