Epidermal Langerhans cells (LC) are potent dendritic cells in the induction of primary T cell-mediated immune responses in the skin. They capture foreign Ags and migrate to regional lymph nodes to carry and present these Ags to naive T cells. We investigated the role of matrix metalloproteinase-9 (MMP-9) in LC migration using an anti-MMP-9 mAb. Intradermal injection of anti-MMP-9 mAb before rhodamine B or oxazolone painting markedly inhibited these hapten-induced decreases in LC number in the epidermis and the accumulation of dendritic cells in the regional lymph nodes, indicating that MMP-9 plays some important roles in LC migration in the induction phase of contact sensitization. Treatment with anti-MMP-9 mAb also blocked the increase in cell size, dendrite development, and the enhanced expression of MHC class II Ags in LC induced by hapten painting. In addition, intradermal injection of purified MMP-9 induced marked increases in cell size, dendrite extension, and enhanced expression of MHC class II Ags in LC. These results strongly suggested that MMP-9 is involved not only in LC migration, but also in their morphological and phenotypic maturation in the skin.
Epidermal Langerhans cells (LC)2 belong to the family of dendritic APC (1). They are bone marrow-derived cells and reside in the basal and suprabasal layers in the epidermis. When LC encounter foreign Ags, they capture these Ags and move to the T cell area in the draining lymphoid organs, where they present these Ags to naive T cells to generate Ag-specific T cells (2, 3, 4). Therefore, LC migration from the epidermis to the regional lymph nodes is a crucial step in the induction of primary T cell-mediated immune responses in the skin.
Although the properties of LC as migrating cells are not fully understood, it has been reported that treatment with contact sensitizers induces several phenotypic changes in LC, including decreased expression of E-cadherin that mediates adhesion between LC and surrounding keratinocytes in the epidermis (5). It has also been shown that LC motility is stimulated by contact with several haptens (4, 6, 7). These studies indicated that various events allowing LC to migrate out of the epidermis are induced by contact of these cells with sensitizers.
We recently found that painting of mice with strong sensitizers such as 2,4,6-trinitrochlorobenzene and 2,4-dinitrofluorobenzene markedly stimulated the expression of matrix metalloproteinase-9 (MMP-9) in LC (8). Similarly to MMP-2, MMP-9 is capable of degrading type IV collagen in the basement membrane, and thus is regarded as one of the key enzymes in migration and invasion of several cell types through the basement membrane (9, 10). The present study was performed to study the role of MMP-9 in LC emigration in vivo using an anti-MMP-9 mAb that is capable of inhibiting type IV collagenase activity.
Materials and Methods
Female BALB/c mice were obtained from SLC Japan (Hamamatsu, Japan), and were used at the age of 8–11 wk.
Rhodamine B isothiocyanate and oxazolone (4-ethoxymethylene-2-phenyloxazol-5-one) were obtained from Sigma (St. Louis, MO). A specific mAb against MMP-9 (clone 6-6B, IgG1) (11) was obtained from Oncogene Research Products (Cambridge, MA). Isotype-matched control mouse IgG1 (clone MOPC21) was purchased from Sigma. The sodium azide contained in this reagent was removed by ultrafiltration using Microcon-10 (Grace Japan, Tokyo, Japan). The hybridoma M5/114.15.2 producing rat anti-I-Ab,d,q and I-Ed,k was obtained from American Type Culture Collection (Manassas, VA), and its culture supernatant was used in this study. FITC-conjugated mouse mAb against mouse I-Ad (clone AMS-32-1) was purchased from PharMingen (San Diego, CA). Purified MMP-9 and MMP-2 derived from murine macrophages were obtained from Elastin Products (Owensville, MO), the gelatin-degrading activities of both of which were 50–70 U/μg protein (1 U degrades 1 μg of gelatin/h). Antibodies and purified MMPs used in this study were endotoxin free, as confirmed by the Limulus test.
Antibody, enzyme, and chemical treatment
Mice were injected intradermally on the dorsal side of both ear pinnae with 30 μl of anti-MMP-9 mAb (1, 5, or 10 μg/ear) or isotype-matched IgG (10 μg/ear) dissolved in PBS using 50-μl microsyringes and 26½-gauge needles. Thirty minutes after mAb injection, 25 μl of 2% rhodamine B dissolved in acetone-dibutylphthalate (1:1) or 0.25% oxazolone dissolved in acetone-olive oil (4:1) was painted onto the dorsal side of both pinnae. In some experiments, mice were injected intradermally on the dorsal side of both pinnae with 30 μl of purified MMP-9 (1.0 μg/ear) or MMP-2 (1.0 μg/ear) dissolved in PBS.
Preparation of epidermal sheets
Eighteen hours after sensitizer painting, the ears of mice were obtained, and epidermal sheets were prepared by incubating dorsal ear halves with 0.02 M EDTA in PBS for 2 h at 37°C. Epidermal sheets were fixed in acetone for 20 min at −20°C, washed with PBS, and incubated with rat anti-I-A mAb (M5/114.15.2) overnight at 4°C. They were then incubated with peroxidase-conjugated anti-rat IgG (Zymed, South San Francisco, CA) for 3 h at room temperature, followed by incubation with the substrate 3-amino-9-ethylcarbazole. The sheets were then mounted on glass slides and photographed under phase-contrast microscopy (Olympus, Tokyo, Japan), and the numbers of MHC class II-positive cells in five random high power fields (0.33 mm2) were counted. In some experiments FITC-conjugated rabbit polyclonal Ab against rat IgG (Dako Japan, Kyoto, Japan) was used to distinguish I-A-positive cells under fluorescence microscopy (Olympus).
Preparation of lymph node dendritic cells
Dendritic cells (DC) in lymph nodes were obtained by the method of Cumberbatch (12) with slight modifications. Briefly, 18 h after rhodamine B or oxazolone application, auricular lymph nodes were obtained and pooled from groups of mice (n = 3). Lymph node cells were prepared by mechanical disaggregation using glass homogenizers and were washed with RPMI 1640 medium containing 10% heat-inactivated FCS for 10 min at 300 × g. Aliquots of 2–5 ml of these cells (2 × 106 cells/ml prepared with RPMI 1640-FCS medium) were layered onto 2 ml of 14.5% metrizamide (Nycomed, Oslo, Norway) and centrifuged for 15 min at 600 × g, and interface cells were collected. They were then washed once with RPMI 1640-FCS medium, and treated with FITC-conjugated anti-I-Ad mAb for 30 min at 4°C. After incubation, the numbers of DC (I-A-positive and dendritic-shaped cells) were counted under fluorescence and phase-contrast microscopy. The frequency of DC was expressed as DC number per lymph node.
Preparation of epidermal cell suspensions and immunocytological analysis
Epidermal cells were obtained from mouse ear halves by trypsinization (0.25%) for 90 min at 37°C. They were then stained with FITC-conjugated mouse anti-I-Ad mAb for 30 min at 4°C. The epidermal cells were also prepared from unsensitized normal mice, and labeled with FITC-conjugated isotype-matched control mAb (FITC-anti-I-Ak: clone 11-5.2, PharMingen) as a negative control. Cells were washed once with PBS, and at least 60,000 cells were examined using FACSCalibur (Becton Dickinson Immunocytometry Systems, San Jose, CA). For quantitative evaluation, the I-A-positive populations were gated out manually, and the mean fluorescence intensity (MFI) of each population was measured using CellQuest software (Becton Dickinson Immunocytometry Systems). The MFI of the negative control was also measured, and the relative fluorescence intensity (RFI) was calculated as follows: RFI = (MFI of I-A positive cells)/(MFI of negative control). Propidium iodide-permeable cells were gated out and were excluded from analysis.
Effects of anti-MMP-9 mAb on hapten-induced LC migration
We first examined the effects of anti-MMP-9 mAb on hapten-induced LC emigration. Some preliminary experiments using a variety of haptens with different sensitizing capacities revealed that the intradermal injection of anti-MMP-9 mAb 30 min before hapten painting effectively inhibited hapten-induced LC emigration, although larger amounts of mAb appeared to be needed when we used strong sensitizers in conventional doses, such as 1% 2,4,6-trinitrochlorobenzene or 1% 2,4-dinitrofluorobenzene. We therefore chose 2% rhodamine B and 0.25% oxazolone as contact sensitizers, since they induced significant, but relatively moderate, LC emigration.
As shown in Fig. 1,A, marked LC emigration was observed following rhodamine B painting, and the number of LC 18 h after rhodamine B treatment was ∼74% that in untreated controls. Treatment with anti-MMP-9 mAb resulted in dose-dependent inhibition of the rhodamine B-induced LC emigration. In contrast, injection of the same amount of isotype-matched control IgG did not prevent LC emigration. When we used 0.25% oxazolone as a sensitizer, marked inhibition of LC emigration was also observed following treatment with anti-MMP-9 mAb at 10 μg/ear (Fig. 1 B), although little effect was observed following treatment with isotype-matched control IgG.
We next examined the effects of the same mAb on DC accumulation in regional lymph nodes under the same experimental conditions. Following painting of 25 μl of both 2% rhodamine B and 0.25% oxazolone on the dorsal ear skin, significant DC accumulation was observed in the auricular lymph nodes (Fig. 2, A and B). Eighteen hours after hapten painting, the numbers of DC in both rhodamine B- and oxazolone-treated mice were increased by ∼3-fold compared with those in untreated controls. Intradermal injection of anti-MMP-9 mAb 30 min before rhodamine B (Fig. 2,A) or oxazolone (Fig. 2 B) painting markedly inhibited DC accumulation in lymph nodes. In contrast, little effect was observed on DC accumulation induced by either rhodamine B or oxazolone following injection of the same amount of isotype-matched control IgG.
Effects of anti-MMP-9 mAb on morphological and phenotypic changes in LC induced by hapten painting
Fig. 3 shows the representative changes in LC morphology in the epidermal sheets prepared from mAb-treated or untreated mice. Eighteen hours after rhodamine B or oxazolone painting, I-A-positive cells in the epidermis appeared to be larger and extended longer dendrites among surrounding keratinocytes. Also, the expression of MHC class II Ags on LC appeared to be more intense. When anti-MMP-9 mAb was injected intradermally 30 min before rhodamine B or oxazolone painting, dendrite extension was markedly reduced, and the shape of I-A-positive cells remained unchanged (Fig. 3, C and G). In addition, the intensity of the reaction product of peroxidase for MHC class II Ags on LC prepared from anti-MMP-9 mAb-treated mice appeared to be unaffected. In contrast, little inhibitory effect was observed when the same amount of isotype-matched control IgG was injected instead of anti-MMP-9 mAb.
The effects of anti-MMP-9 mAb on morphological and phenotypic changes in LC were further examined by FACS using epidermal cell suspensions prepared from mAb-treated or untreated mice. Dot-plot analysis (log FITC-I-A fluorescence intensity vs forward scatter) revealed that a subset of I-A-positive cells was increased in size, as assessed by forward scattering, and showed an enhanced level of Ia Ag in the epidermal cells prepared from oxazolone-treated mice. Anti-MMP-9 mAb treatment partially, but clearly, blocked these changes. However, isotype-matched control IgG had no effect (Fig. 4,A). The RFI of I-A-positive LC prepared from 0.25% oxazolone-treated mice exhibited a 2- to 3-fold increase in I-A expression compared with those from untreated mice. Injection of anti-MMP-9 mAb before oxazolone painting inhibited this enhanced I-A expression by 50.0% at 10 μg mAb (n = 5). In contrast, injection of the same amount of isotype-matched control IgG showed little effect (Fig. 4 B).
Effects of purified MMP-9 injection on LC morphology and phenotype
Because the inhibition of MMP-9 activity using anti-MMP-9 mAb prevented dendrite development and the up-regulation of I-A expression in LC, it is possible that MMP-9 was directly involved in LC activation. To confirm this possibility, we next injected purified MMP-9 into ear pinnae and examined its effects on cell size, dendrite development, and intensity of I-A expression on LC. Also, purified MMP-2, another type IV collagenase with substrate specificity similar to that of MMP-9, was examined for comparison. As shown in Fig. 5, injection of 1.0 μg of purified MMP-9 induced marked extension of dendrite in I-A-positive cells in the epidermis, and their level of I-A appeared to be enhanced. In contrast, injection of the same amount of MMP-2 with the equivalent enzymatic activity did not induce LC dendrite extension and did not affect their level of I-A expression. FACS analysis revealed that injection of purified MMP-9, but not purified MMP-2, induced a marked increase in cell size (Fig. 6,A) and enhanced I-A expression in LC (Fig. 6 B).
It has been reported that MMPs play some important roles in degradation of extracellular matrixes in many biological events, such as tumor invasion, angiogenesis, wound healing, etc. (13). The gelatinase-type MMP, MMP-9 and MMP-2, are capable of degrading type IV collagen in the basement membrane and thus they are regarded as key enzymes in migration and invasion by several cell types, including neutrophils, macrophages, T cells, and some invasive tumor cells (14, 15, 16, 17). The present study provided evidence that MMP-9 also participates in LC emigration in vivo, because the intradermal injection of a blocking mAb effectively prevented the decrease in LC number in the epidermis and DC accumulation in regional lymph nodes induced by hapten painting. We previously found that MMP-9 production and secretion were induced by hapten painting in LC (8). Immunohistochemical analysis using human skin explants also revealed that LC were actually capable of producing MMP-9 in the epidermis (18). These and the present results suggested that MMP-9 produced by LC in response to contact with haptens plays some important role(s) in LC emigration, probably by degrading type IV collagen in the basement membrane.
The induction of MMP-9 is thought to be regulated by various factors, such as cytokines, growth factors, etc. (19). Among these stimuli, it is of particular interest that both TNF-α and IL-1β stimulate MMP-9 production in several cell types. In the epidermis, TNF-α is produced by keratinocytes in response to various stimuli, including hapten application, and act on LC in a paracrine fashion (20). Interleukin-1β is constitutively produced by LC, and its production is increased rapidly by treatment with contact sensitizers (20, 21). Recent studies have shown that both cytokines induce LC migration in vivo (22). It is thus conceivable that in the induction phase of contact sensitization these cytokines provide a signal to induce MMP-9 production by LC. Interestingly, both TNF-α and IL-1β also reduce keratinocyte-LC adhesion mediated by E-cadherin in vitro and in vivo (5, 23). The precise roles of these cytokines should be examined in future studies. However, they appear to participate in various events that allow LC to migrate out of the epidermis in the early phase of initiation of contact sensitization.
The present study also revealed that the intradermal injection of anti-MMP-9 mAb markedly inhibited hapten-induced morphological changes in LC, such as increases in cell size and dendrite development. Also, intradermal injection of purified MMP-9 was capable of inducing these changes, suggesting that MMP-9 plays some important roles in LC activation. In addition, it is also interesting to note that MMP-9 secretion appeared to be closely related to the up-regulation of MHC class II molecule expression on LC. The results of experiments using an anti-MMP-9 mAb to block and those using purified MMP-9 to mimic phenotypic maturation of LC indicated that the production and secretion of MMP-9 from LC are induced before the enhanced expression of MHC class II molecules on LC.
Although the present study provided phenomenological evidence for the need for MMP-9, at present it is difficult to estimate why and how this particular protease is required for maturation processes. However, our results suggested several possible roles of MMP-9 in LC maturation. First, there may be a specific substrate(s) for MMP-9 in or on the cell surface of surrounding keratinocytes, and they could prevent or regulate the extension of LC dendrites. Also, the degradation products derived from these specific substrate(s) may act as a trigger or mediator to up-regulate the expression of MHC class II molecules on LC. However, these specific substrates for MMP-9 have not yet been identified. MMP-9 degrades gelatin; type III, IV, and V collagens; and elastin (10, 24, 25), although these extracellular matrix components have not been detected in situ among keratinocytes in the epidermis, at least at the immunohistochemical level. Thus, to clarify the precise mechanism by which MMP-9 participates in LC activation, it is necessary to first identify the specific substrate(s) for MMP-9.
It is also possible that MMP-9 could act on LC themselves in an autocrine manner and is directly involved in their maturation process. At present, little is known about the precise mechanism of LC maturation. However, recent findings support the possibility that MMPs could directly modulate cellular functions in several cell types by cleaving and shedding certain membrane molecules, such as FcγRIIIB on neutrophils (26), chemokine receptors on neutrophils (27), and CD43, CD44, and CD16 on granulocytes (28). To examine this possibility in the case of LC maturation, in vitro experiments are necessary in future studies. It will be important to identify first the target molecule(s) of MMP-9 to determine whether their effects on LC maturation are direct or indirect.
In conclusion, MMP-9 is likely to be involved in various functions of LC, i.e., not only LC emigration but also their accommodation in the epidermis and their phenotypic maturation. The function of MMP-9 appears to be crucial for the initiation phase and probably the elicitation phase of contact sensitization.
Abbreviations used in this paper: LC, Langerhans cells; DC, dendritic cells; MFI, mean fluorescence intensity; MMP, matrix metalloproteinase; RFI, relative fluorescence intensity.