The Met tyrosine kinase has a pivotal role in embryonic development and tissue regeneration, and deregulated Met signaling contributes to tumorigenesis. After binding of its cognate ligand hepatocyte growth factor, Met signaling confers mitogenic, morphogenic, and motogenic activity to various cells. Met expression in the hematopoietic compartment is limited to progenitor cells and their Ag-presenting progeny, including dendritic cells (DCs). In this study, we demonstrate that Met signaling in skin-resident DCs is essential for their emigration toward draining lymph nodes upon inflammation-induced activation. By using a conditional Met-deficient mouse model (Metflox/flox), we show that Met acts on the initial step of DC release from skin tissue. Met-deficient DCs fail to reach skin-draining lymph nodes upon activation while exhibiting an activated phenotype. Contact hypersensitivity reactions in response to various contact allergens is strongly impaired in Met-deficient mice. Inhibition of Met signaling by single-dose epicutaneous administration of the Met kinase-specific inhibitor SU11274 also suppressed contact hypersensitivity in wild-type mice. Additionally, we found that Met signaling regulates matrix metalloproteinase MMP2 and MMP9 activity, which is important for DC migration through extracellular matrix. These data unveil Met signaling in DCs as a critical determinant for the maintenance of normal immune function and suggest Met as a potential target for treatment of autoimmune skin diseases.
Antigen-presenting dendritic cells (DCs) play a crucial role in the induction of primary immune responses (1). DCs are scattered in lymphoid and nonlymphoid organs throughout the body and act as sentinels of the immune surveillance system. In skin, dermal DCs (dDCs) and Langerhans cells (LCs), the epidermal contingent of DCs, constitute the first immune barrier for invading pathogens and contact allergens. Additionally, dDCs and LCs have been implicated in the induction and maintenance of tolerance (2, 3). DC precursors populate and functionally integrate as immature DCs into peripheral tissues, such as skin. Following Ag capture, DCs are activated and migrate via lymphatic vessels to the T cell areas of lymphoid organs where they encounter T cells to present processed Ags to induce specific immune responses. Thus, during the DC life cycle, migration and homing are tightly regulated (e.g., by specific expression of chemokines and chemokine receptors, adhesion molecules, and homing receptors) to ensure the selective positioning of DCs at specific sites of function (4, 5). Before entering afferent lymphatics to reach draining lymph nodes (LNs), migration of activated tissue-resident DCs requires an initial sequence of steps, including mobilization and detachment from surrounding tissues and transmigration through interstitial space. In skin, proinflammatory cytokines, such as TNF-α and IL-1β, have been recognized as important mediators of DC mobilization after activation, leading to the emigration from peripheral tissues (4). Similarly, hepatocyte growth factor (HGF; also known as scatter factor) induces emigration of LCs from mouse epidermis in skin biopsy assays (6). Dermal fibroblasts were identified as a major source of HGF in skin upon infection or stimulation with proinflammatory cytokines, such as IL-1α, IL-1β, TNF-α (7), or IL-6 (8). More recently, keratinocytes were also reported to produce HGF upon skin injury and inflammation (9, 10).
The proto-oncogene Met encodes the high-affinity transmembrane tyrosine kinase receptor for HGF. Met signaling after ligand binding confers mitogenic, morphogenic, and motogenic activity to various cells and is essential for embryogenesis and tissue regeneration. Aberrant Met signaling contributes to invasive growth and metastasis of tumors (11, 12). Met is predominantly expressed on epithelial cells but has also been implicated in hematopoiesis (13). Expression of Met in the hematopoietic compartment was found in progenitor cells, macrophages, B cells, and DCs. Met signaling has been proposed in monocyte–macrophage development (14) and B cell homing to LNs (15). Met expression was also found on splenic DCs (16, 17), bone marrow (BM)-derived DCs, and LCs (6). Like Met, its ligand HGF is indispensable during embryonic development, as shown by the lethal phenotype in utero of conventional Met- or HGF-null mutations (18–20). HGF is a pleiotropic factor that promotes cell survival, tissue protection, and regeneration and, hence, exerts anti-inflammatory activities (21). Moreover, HGF was found to regulate various functions of immune cells, including cytokine production, migration, and adhesion (6, 16, 17, 22–26). However, the mechanisms by which HGF impacts on immunological responses have remained elusive.
In this article, we describe the impact of Met signaling for DC/LC mobilization from skin and homing into peripheral tissues. By using a conditional Met-deficient mouse model (Metflox/flox), we demonstrate that Met signaling in skin-resident DCs is essential for their detachment and emigration from surrounding tissue upon activation. As a consequence, Met signaling-deficient skin DCs/LCs fail to reach draining LNs. Met kinase inactivation, either genetically or pharmacologically, resulted in strongly impaired contact hypersensitivity (CHS) reactions in response to contact allergens. We found that Met signaling in DCs modulates their matrix metalloproteinase (MMP)2 and MMP9 activity required for transmigration through extracellular matrix (ECM). Our findings unveil Met signaling in DCs as a critical determinant for the maintenance of the immune surveillance system and suggest Met as a potential target for the treatment of autoimmune skin diseases, such as psoriasis.
Materials and Methods
C57BL/6, CD45.1 (B6.SJL-Ptprca Pepcb/BoyJ), OT-II (C57BL/6-g(TcraTcrb)425Cbn/J), Metflox/flox (C57BL/6), and mice heterozygous for the Met knockout allele (Met+/−) and crossed to transgenic mice expressing Cre under the control of the IFN-inducible Mx promoter (Mx-cre × Met+/−) were maintained under specific pathogen-free conditions in the central animal facility of the Rheinisch-Westfälische Technische Hochschule University Hospital Aachen. All animal experiments were approved by local authorities in compliance with the German animal protection law. Metflox/flox mice were interbred with Mx-cre × Met+/− mice, and offspring (6–8 wk) received five injections of 300 μg polyinosinic/polycytidylic acid (pI:pC) i.p. at 2-d intervals for excision of floxed exon 15 of Met to generate Met-deficient MetΔ/− mice (Met knockout [KO]) or control (wild-type [WT]) MetΔ/+ mice (Met WT). Further functional studies in Met KO and Met WT mice were performed ≥4 wk after the last pI:pC injection.
To generate LC chimeric mice, 2 × 106 BM cells from CD45.2+ Met WT or Met KO mice were transplanted into CD45.1+ recipient mice, as previously described (27). Donor chimerism in BM was evaluated by flow cytometry and was ≥95% for donor cells 7 wk after transplantation. Then mice were exposed to UV light (254 nm) for 30 min (total dose 16.8 kJ) to induce ear skin inflammation and LC chimerism in epidermis (28). The presence of donor- and host-derived LCs was investigated 9 wk after UV irradiation by immunofluorescence microscopy and flow cytometry. Ears of BM chimeric mice were subjected to ex vivo migration assays.
Local skin inflammation was induced by applying 50 μl acetone/dibutyl phthalate solution (1:1, v/v) on ear skin for 24 h. A systemic inflammatory response was induced by a single-dose injection of 50 ng LPS i.v. into the tail vein, and Met expression on DCs was analyzed 24 h later.
Skin DC migration toward LNs was analyzed by flow cytometry 24–72 h after painting with 50 μl 0.5% FITC in acetone/dibutyl phthalate solution (1:1, v/v) on the shaved abdomen.
To induce CHS, mice were sensitized by applying 400 μl acetone/dibutyl phthalate/0.5% FITC solution or 20 μl 0.5% 2,4-dinitrofluorobenzene (DNFB; (Sigma-Aldrich) in acetone/olive oil (4:1, v/v) on the shaved abdomen. Immunization with DNFB was repeated after 24 h. In some experiments, the Met-specific kinase inhibitor SU11274 (Merck) dissolved in DMSO (200 μM, final concentration) or DMSO alone was applied 30 min before immunization with DNFB. Basal ear thickness was measured 5 d after immunization with a digital caliper (Hogetex) before applying 20 μl 0.5% FITC or 0.2% DNFB solutions onto the right ears. Left ears received solvents only as control. CHS responses were determined by gauging the ear swelling 24 h later.
Flow cytometry analysis was performed on FACSCalibur and FACSCanto (BD Biosciences) devices, as previously described (6), using fluorochrome-conjugated Abs to stain for Met (clone eBioclone 7), langerin (929F3), CD8α (53-6.7), CD11b (M1/70), CD11c (N418), CD40 (1C10), CD83 (Michel-17), CD86 (GL1), MHC class II (I-A/I-E; clone 2G9), panCD45 (30-F11), CD45.1 (A20), CD45.2 (104), CD103 (M290), and CD184 (2B11), as well as biotin-conjugated Abs for CD80 (1G10), CD209 (5H10), and CD205 (NLDC-145). For detection of langerin, cells were fixed with PBS/4% paraformaldehyde and permeabilized with PBS/1% saponin (Sigma-Aldrich). Detection of CCR7/CD197 was done with purified chimeric CCL19-human Fc fusion protein (29). Data were analyzed using FlowJo software (Tree Star).
LC/dDC analyses in skin
Ears from mice were split into dorsal and ventral halves and incubated in PBS/0.02 M EDTA (Sigma-Aldrich) for 90 min at 37°C to separate epidermis from dermis. Dermal and epidermal cell suspensions were obtained by digestion with type V collagenase (1 mg/ml; Sigma-Aldrich) at 37°C for 30 min, passed through a 21-gauge syringe needle and a cell strainer, and assessed by flow cytometry.
Ex vivo migration assay
Dorsal and ventral ear halves were cultivated split-side down for 3 d at 37°C in RPMI 1640 medium containing 10% FCS and GM-CSF (200 U/ml), plus HGF (40 ng/ml; eBiosciences), TNF-α (100 ng/ml; PeproTech), LPS (1 μg/ml); or no factor. Epidermis was isolated, fixed in acetone, and subjected to immunofluorescence microscopy. Cells that had migrated into culture medium were stained for CD11c, langerin, and Met and analyzed by flow cytometry. Numbers of migrated cells were normalized using Dynabeads (15 μm diameter; Dynal Polymers), as previously described (6). Ear halves of mice that received UV irradiation were cultivated with or without LPS (1 μg/ml) for 24 h. Epidermal cell suspensions and cells that had migrated into culture medium were stained for panCD45, CD45.1, CD45.2, CD11c, CD11b, MHC class II, and langerin and analyzed by flow cytometry.
Cryosections of LNs (5 μm thickness) and epidermal sheets were fixed in 100% ice-cold acetone for 20 min, rinsed in PBS, and blocked with PBS/3% BSA (Sigma-Aldrich) for 30 min at room temperature before staining. Samples were stained with fluorochrome-conjugated Abs for detection of Met, langerin, and MHC class II and counterstained with DAPI (Vector Laboratories). Images were acquired using an Axiovert 200 fluorescence microscope (Zeiss) and a digital CCD camera (Roper Scientific) operated with IPlab software. Image processing was done with Adobe Photoshop software.
RNA isolation and RT-PCR
To isolate RNA from mouse ear skin, mouse ear tissues were dissociated into single-cell suspensions using the gentleMACS Dissociator (Miltenyi Biotec). Total RNA was isolated using the NucleoSpin RNA II kit (MACHEREY-NAGEL) with on-column DNAse digestion, according to the manufacturer’s instructions, and mRNA was reverse-transcribed with a High Capacity cDNA RT Kit (Applied Biosystems). A total of 50 ng cDNA was used as a template for PCR reactions using gene-specific primers and conditions. Five-microliter aliquots of PCR reactions were subjected to 1% agarose gel containing ethidium bromide, and signals were recorded with GelDocX gel documentation system (Bio-Rad). Expression of Gapdh mRNA was used as control. Primers used were as follows: Gapdh, forward 5′-ggggTgAggCCggTgCTgAgTAT-3′, reverse 5′-CATTgggggTaggAACACggAAgg-3′; Met, forward 5′-CATTCTggTgCTgcTgTTgTCC-3′, reverse 5′-TgTCTTTCCAAACCCCTCCTg-3′; Hgf, forward 5′-CATTCAAggCCAAggAgAAg-3′, reverse 5′-AACTCggATgTTTgggTCAg-3′; Tnfa, forward 5′-ggCAggTCTACTTTggAgTCATTgC-3′, reverse 5′-ACATTCgAggCTCCAgTgAATTCgg-3′; Mmp2, forward 5′-CCCCTATCTACACCTACACCAAGAAC-3′, reverse 5′-CATTCCAggAgTCTgCgATgAgC-3′; and Mmp9, forward 5′-AGTTTGGTGTCGCGGAGCAC-3′, and reverse 5′-TACATgAgCgCTTCCggCAC-3′.
Functional analyses of Met in DCs
Immature DCs were generated from mouse BM using a two-step amplification and differentiation protocol, as previously described (27). In some experiments, 0.1 μg/ml LPS (Sigma-Aldrich) was added 24 h prior to treatment with HGF (40 ng/ml) to induce DC maturation.
Ag-specific T cell activation was performed with CD4+ T cells from OT-II mice expressing a transgenic TCR that recognizes OVA323–339 peptide (ISQAVHAAHAEINEAGR). Splenic CD4+ T cells were prepared by immunomagnetic bead purification using MACS anti-CD4 MicroBeads (Miltenyi Biotec), according to the manufacturer’s instructions. Immature and LPS-matured DCs from Met WT and Met KO mice, stimulated with 40 ng/ml HGF for 16 h or left untreated, were pulsed with 0.1 μM OVA323–339 peptide for 30 min and then cocultured at various ratios with isolated OT-II CD4+ T cells in 96-well microtiter plates. At day 4 of coculture, 0.75 μCi/well [3H]thymidine (GE Healthcare) was added per well. Cells were harvested 5 h later, and [3H]thymidine incorporation was measured with a scintillation counter (MicroBeta TriLux; Wallac). Experiments were done with triplicate samples.
Immunoblotting was carried out essentially as described before (30). Total and phosphorylated p44/42 MAPK were identified using rabbit anti-p44/42 MAPK and anti–p-p44/42 MAPK Ab (Erk1/2, clone 137F5, and pErk1/2, clone D13.14.4E; Cell Signaling Technology), respectively. MMP-2 and MMP-9 were detected using polyclonal rabbit Abs from Cell Signaling Technology. Detection of β-actin (AC-74; Sigma-Aldrich) was used as loading control.
Chemotaxis of DCs in response to CCL19 or CXCL12 chemokines (100 ng/ml each; PeproTech) was performed essentially as described previously (27), with minor modifications. Transwell inserts (5 mm pore size; Costar) were coated with Matrigel (BD Biosciences) at 37°C for 16 h or left uncoated. LPS-activated DCs were stimulated with HGF 30 min prior to the assay start. Transmigration of DCs through uncoated Transwell inserts and through Matrigel matrix was analyzed by flow cytometry after 90 min and 5 h, respectively. The numbers of transmigrated cells were normalized using Dynabeads, as done in ex vivo migration assays.
Gel zymography was used to analyze membrane-bound MMP-2 and -9 activities. Native protein lysates of DCs were mixed with 2× sample buffer (125 mM Tris-HCl, 20% glycerol, 4% SDS, 0.005% bromophenol blue) for 10 min before gel electrophoresis. After electrophoresis, gels were renatured with 2.5% Triton X-100 for 30 min and incubated with 50 mM Tris base/0.2 M NaCl/5 mM CaCl2 overnight at 37°C. Degradation of gelatin was visualized by Coomassie Blue R-250 staining.
Statistical analyses were performed using the Student t test. A p value < 0.05 was considered statistically significant.
Activation-dependent Met expression on DCs
DC homeostasis is differentially regulated in steady state and inflammation (31, 32); thus, we analyzed various DC subtypes in mice for Met expression in lymphoid tissues and in skin before and after DC activation. We found Met expression on all DC subtypes analyzed, albeit at different levels according to subtype, location, and activation status (Fig. 1A, 1C–H, Supplemental Figs. 1, 2D). DC subsets in the skin, identified as langerin+ LCs in the epidermis and langerin−CD103− and langerin+ CD103+ in the dermis (33–36), expressed Met, with the majority of cells showing low expression in the steady state (Fig. 1A, 1C, 1D, 1H). By applying a skin-irritating acetone/dibutyl phthalate solution on ear skin, we induced local inflammation, as evidenced by the increased expression of the proinflammatory cytokine TNF-α (Fig. 1B). Under such conditions, increased expression of Met and its ligand HGF was also observed (Fig. 1B). Met expression on skin DC subsets in skin and in skin-draining LNs was analyzed under such inflammatory conditions and after LPS injection 24 h later. An increased expression of Met on skin DC subsets was found under both local and systemic inflammatory conditions (Fig. 1C, 1D, 1F, 1H). The upregulated expression of Met was more pronounced in epidermal LCs after skin irritation, whereas it was greater in dDC subsets after LPS injection (Fig. 1C, 1D, 1H). In LNs, the DC compartment comprises LN-resident DC subsets and tissue-derived DCs immigrating from peripheral tissues. In the steady state, migratory DCs constitute only a minor population of all DCs in LNs (37), but skin irritation results in an increased influx of skin DCs into draining LNs. Under such inflammatory conditions, the greatest Met expression was found on CD8α−CD11b+ MHC class II-high migratory DCs compared with LN-resident CD8α+CD11b− MHC class II-intermediate DCs (Fig. 1G). The predominant migratory DC subsets in skin-draining LNs, which have transported Ags from peripheral tissues, are langerin+ LCs and langerin− dDCs (37, 38); both subsets exhibited elevated Met expression compared with steady-state expression in skin (Fig. 1C, 1D, 1F, 1H).
LCs and dDCs were shown to migrate with different kinetics to draining LNs upon activation, with dDCs preceding LCs (38, 39). To study migration kinetics and Met expression of LCs and dDCs, DCs in skin were labeled in vivo by applying 0.5% FITC on the shaved abdomen of mice (38). Met expression was analyzed on FITC+ LCs and FITC+ dDCs that had migrated to draining inguinal LNs (Fig. 2). Most of the faster migrating DCs (24 h) showed the greatest Met expression, whereas after 48 h more DCs with intermediate levels of Met were found; numbers of FITC+ DCs started to decline at later time points (Fig. 2A). LCs and dDCs differentially express endocytic receptors DEC205/CD205 and DC-SIGN/CD209 (38, 40). LCs are CD209− and upregulate the expression of CD205 upon activation, whereas dDCs express CD209 and intermediate levels of CD205. This expression pattern was used in this study to discriminate FITC+ LCs and FITC+ dDCs upon migration to draining LNs (Fig. 2B) (38, 40). Met expression levels were lower at all time points in FITC+ LCs than in FITC+ dDCs, coinciding with the different migration kinetics of these DC subtypes (Fig. 2). Taken together, these data show that Met expression on peripheral skin DCs is upregulated upon activation, suggesting a functional role of Met for DC migration toward draining LNs.
Met is essential for emigration of skin DCs from residential tissue
Met-signaling deficient mouse models are embryonically lethal; therefore, we used a floxed Met transgenic mouse model (Metflox/flox) (41) to investigate Met function in skin DCs. Metflox/flox mice were crossed with mice heterozygous for the Met knockout allele that express Cre under the control of the IFN-inducible Mx promoter (Mx-cre/Met+/−, Fig. 3A). Repeated injections of pI:pC (1:1; w/w) into animals with an Mx-cre/Metflox/− genotype induced efficient excision of the single floxed Met allele, resulting in a Met-deficient (Met KO) phenotype. Mx-cre/Metflox/+ littermates retain the WT phenotype (Met WT) after pI:pC injections and were used as controls. Met KO and Met WT mice show the same number of LCs in epidermis (Supplemental Fig. 2A).
DC emigration from skin upon activation was quantified in skin explant assays (Fig. 3B, 3C, Supplemental Fig. 2) (6). Ear sheets from Met KO and Met WT mice were cultured in the presence or absence of LPS, TNF-α, or HGF (Supplemental Fig. 2B). All factors significantly enhanced emigration of DCs from skin tissue of Met WT mice into culture medium (Fig. 3B). In contrast, none of the stimuli was able to enhance the release of DCs from skin of Met KO mice (Fig. 3B). Immunofluorescence microscopy and flow cytometry analyses revealed that LCs and dDCs remained in the skin of Met KO mice (Fig. 3C, 3D, Supplemental Fig. 2C). The stationary Met KO DCs showed an activated cell surface marker profile after LPS stimulation of the ear sheets identical to that of emigrated DCs from Met WT ear skin (Fig. 3D).
To investigate whether the role of Met signaling for emigration of skin DCs is cell intrinsic, we performed BM transplantation of CD45.2 Met KO and Met WT BM mice into CD45.1 Met WT recipients. We first analyzed LC repopulation in the epidermis of CD45.2 Met KO → CD45.1 Met WT BM chimeric mice (Fig. 4A). LCs develop during embryogenesis from a skin-seeding LC precursor (42, 43) and are resistant to radiation-induced ablation. However, UV irradiation causes skin inflammation and results in loss of LCs that are then replaced by circulating BM-derived precursors (28). LC chimerism in skin epidermis of recipient mice was analyzed by fluorescence microscopy 9 wk after UV irradiation. We found no difference in the number of donor-derived LCs in the epidermis of Met KO versus Met WT BM chimeric mice and a comparable high degree of donor LC chimerism (Fig. 4B, 4C). These results show that Met signaling is dispensable for LC precursor immigration into skin.
In a second set of experiments using CD45.2 Met KO → CD45.1 Met WT BM chimeric mice, emigration of Met KO LCs from skin was analyzed in skin explant assays 9 wk after UV irradiation using flow cytometry (Fig. 4A, 4D–F). Again, no differences in LC chimerism between Met KO and Met WT BM chimeric mice were observed (Fig. 4D). Ear sheets from BM chimeric mice were cultured in the presence or absence of LPS for 24 h, and emigration of donor-derived LCs into culture supernatant was analyzed. Emigration of Met WT LCs from BM chimeric ear skin explants was significantly induced by LPS stimulation after 24 h (Fig. 4E). Concomitantly, the number of Met WT LCs in the epidermis diminished (Fig. 4F). In contrast, Met KO LCs from BM chimeric mice were severely impaired in their ability to emigrate from skin into culture supernatant after LPS stimulation and were retained in epidermis (Fig. 4E, 4F). These data confirm and further extend our results from skin explant assays in conditional Met KO and Met WT mice and demonstrate the cell-intrinsic function of Met in LCs.
Next, migration kinetics of skin DCs from Met KO and Met WT mice toward draining LNs were analyzed in vivo after labeling with FITC. Brachial LNs are distant from the site of FITC painting; therefore, direct labeling of LN-resident DCs by passively transported FITC through lymphatics is minimized (5, 38). Brachial LNs and nondraining mesenteric LNs were analyzed for FITC+ DCs. In Met WT mice, a rapid and sustained migration of FITC+ MHC class IIhigh DCs to the draining LNs was observed (Fig. 5A, 5B, Supplemental Fig. 3). In contrast, the presence of FITC+ DCs was severely diminished in Met KO mice.
Met-deficient mice were then examined for their ability to establish CHS. Met KO and Met WT mice were sensitized using the chemical haptens FITC (Fig. 5C) or DNFB (Fig. 5D, 5E). In Met WT mice, both DNFB and FITC elicited a strong CHS response, whereas in Met KO mice the CHS response was substantially attenuated (Fig. 5C, 5D). Next, we addressed whether inhibition of Met signaling by pharmacological means could repress CHS. The shaved abdominal skin of WT mice was treated with a single dose of the Met inhibitor SU11274 for 30 min before applying DNFB. CHS responses were elicited and analyzed as before; they revealed that pharmacological inhibition of Met signaling significantly diminished ear swelling (Fig. 5E).
Met regulates transmigration through ECM barriers and MMP activities in DCs
To investigate molecular mechanisms of Met signaling in DCs, homogenous DC populations from BM of Met KO and Met WT mice were generated that exhibit a phenotype similar to skin DCs (Fig. 6A) (27). We found no differences in the surface marker profile of immature DCs from both Met KO and Met WT mice and of mature DCs after stimulation with LPS (Fig. 6A) or TNF-α (Supplemental Fig. 4A). The marker profile of mature Met KO and Met WT DCs was also virtually identical to that of emigrated and retained DCs of Met KO and Met WT mice in skin explant assays (Figs. 3D, 6A). Activation of Met KO and Met WT DCs caused efficient upregulation of MHC class II and costimulatory molecules; accordingly, no differences in their capacity to activate Ag-specific naive T cells were observed (Fig. 6B, Supplemental Fig. 4B). It is noteworthy that Met KO DCs still express Met on their cell surface, because Cre-mediated excision of floxed exon 15 affects only the intracellular ATP-binding site, similar to the classical Met KO (18, 41). Met-signaling deficiency in Met KO DCs was confirmed by analyzing MAPK activation in response to HGF stimulation, which showed a rapid and sustained ERK1/2 activation, particularly of ERK1/p44 MAPK (Fig. 6C). In contrast, Met KO DCs were Met-signaling incompetent, because no ERK1/2 activation was observed in response to HGF (Fig. 6C).
Next, the migratory functions of HGF-stimulated DCs from Met WT and Met KO mice were analyzed. DCs express CCR7 and CXCR4 chemokine receptors and respond to their cognate chemokines CCL19 (also known as MIP3-β/ELC) and CXCL12 (also known as SDF-1), respectively (4). Chemotaxis responses to CCL19 and CXCL12, using uncoated transmembrane inserts in Transwell-chamber assays, showed no differences in the migratory activity of both Met KO and Met WT DCs (Fig. 6D). This is in line with the observed expression profiles of CCR7 in BM and skin-derived DCs from Met KO and Met WT mice (Figs. 3D, 6A).
Then, transmembrane inserts were coated with an ECM compound that requires partial degradation by DCs for transmigration toward a chemokine gradient (44). Transmigration of DCs from Met KO and Met WT mice through ECM in response to CCL19 and CXCL12 revealed a deficiency of Met KO DCs to pass the ECM barrier (Fig. 6E). Mature DCs express proteolytic enzymes to overcome ECM and basement membrane barriers (e.g., in skin) (4). Particularly, maturing skin DCs upregulate the expression of MMP-2 and MMP-9 (44, 45), and both MMPs were expressed in Met KO and Met WT DCs (Fig. 6F, 6G). However, we found no Mmp2 and Mmp9 gene expression regulation after HGF stimulation within the time period of DC migration through the ECM barrier in transmigration assays (Fig. 6F). Therefore, we next addressed regulation of proteolytic activities in DCs by LPS and HGF stimulation using gel zymography. Met WT DCs showed high MMP-2 and low MMP-9 activities of membrane-bound forms after LPS activation, which correlated with protein expression levels (Fig. 6G). HGF stimulation of LPS-matured Met WT DCs enhanced MMP-9 activity, whereas MMP-2 activity was abrogated (Fig. 6G). In contrast, Met KO DCs showed complete absence of membrane-bound MMP-2 and MMP-9 activities, whereas protein expression levels were similar to those in Met WT DCs (Fig. 6G). Taken together, these findings show that Met signaling regulates MMP activities in DCs and suggests a potential role of the Met–MMP pathway in the emigration of LCs and dDCs from peripheral skin tissue.
The important role of the receptor tyrosine kinase Met and its cognate ligand HGF in embryonic development has been well recognized. In adults, Met signaling is involved in tissue repair and invasive tumor growth; thus, it has received special attention as a potential target in cancer therapy (46). Expression of Met in B cells, monocytes, and DCs, but not T cells, has been described, yet Met function on the immune system is not understood. Moreover, results from recent studies that focused on a potential role for Met in modulating immune responses are highly controversial (17, 47). We demonstrate in this study the impact of Met signaling on DC emigration from peripheral skin tissue toward lymphoid organs, which is in line with the motogenic activity of Met signaling in other cells, including monocytes (22, 23).
The Met ligand HGF is produced by various mesenchymal cell types and acts as a paracrine factor on Met-expressing cells located in close proximity (48). Stromal fibroblasts, vascular endothelial and smooth muscle cells, neutrophils, and mast cells are major sources of HGF after activation by tissue injury or inflammatory cytokines (49). Thus, we propose a model where, upon skin injury and/or inflammation, HGF is produced (e.g., by fibroblasts), which, in a paracrine fashion, acts on and mobilizes skin DCs. In contrast, anti-inflammatory factors, such as glucocorticoids (50), 1,25-dihydroxyvitamin D3 (51), and TGF-β (52), inhibit HGF production.
Fibroblast-like stromal cells from human lymphoid tissues were found to constitutively produce HGF that is modulated by activated T cells (53). Thus, HGF might play additional roles within lymphoid organs on Met-expressing cells, such as regulation of cell survival (54) and cytokine production (16, 17, 24–26), thereby influencing immune responses.
Met interaction with other cell surface receptors important for cellular migration was described, including integrin α6β4, plexin-B1, CD44, and Mif receptor (54, 55). Plexin-A1 was recently found to regulate DC entry into lymphatics (56), which opens the possibility that this occurs through interaction with Met. Additionally, a similar role in DC migration was described for plexin-C1 (57), and the expression of plexin-B2 on DCs was recently described (58). Plexins and Met share a highly homologous Sema domain that allows physical interactions between Met and plexins (59). Consequently, semaphorin binding to plexins can lead to HGF-independent Met transactivation that might account for some Met-dependent activities (60). Such versatile interactions are expected to allow integration of various extracellular stimuli supplied at different tissue environments to the signaling and locomotive machinery of Met-expressing cells, including DCs.
In many biological systems, Met signaling leads to the transient conversion of cells from an epithelial to a mesenchymal phenotype (epithelial-to-mesenchymal transition [EMT]), followed by well-controlled cell migration (frequently referred to as invasive growth) (11, 12). In epithelial cells, physical contact with neighboring cells is mediated by adherens junctions, and various mechanisms have been described that result in adherens junction destabilization by Met signaling, including activation of MMPs (55, 61). Met signaling in tumor cells can stimulate the proteolytic activity of MMPs, thus facilitating their dissociation and invasive scattering (55, 61). Interstitial DCs express E-cadherin and tight junction proteins (4, 32, 62) and, thus, exhibit features common to epithelial cells, which might be required for their integration into the skin epithelial cell layer. The impact of E-cadherin on DC function has been well recognized (63). Thus, Met signaling in DCs appears to recapitulate properties of EMT, including detachment from the tissue of residence, adherence to and migration through ECM, and crossing tissue boundaries. The important roles for MMPs, particularly MMP-2 and MMP-9, in interstitial DC migration have been well recognized (4, 44, 45). Our data provide evidence that regulation of MMP activities by Met in DCs, as shown in this study, is involved in this process.
In tumor cells, Met regulates MMP activities by both enhancing the transcriptional levels of various MMPs and by stimulating conversion of the inactive precursor forms into active enzymes (61). We found no evidence for immediate regulation of MMP-2 and MMP-9 activities on the transcriptional level by HGF stimulation in DCs, suggesting a posttranscriptional control mechanism. Met-deficient LCs and dDCs remained in their residential tissue, indicating that Met signaling in DCs acts at the initial step of the EMT program (i.e., disruption of the intercellular network junctions). Regulation of MMP activities in DCs by Met is likely not sufficient to account for all Met-induced activities. Thus, additional Met-dependent mechanisms that control this program in DCs remain to be identified.
The findings reported in this article also suggest a role for Met signaling in trafficking of interstitial DCs in other peripheral tissues, such as lung and gut; further studies are needed to scrutinize this notion. Hence, the question arises whether treatment of cancer with Met inhibitors interferes with immune functions, leading to not-yet-recognized side effects. In addition, the proposed model emphasizes the importance of the tissue environment as a source of cytokines and mediators that modulate DC functions. Finally, the discovery of Met as a fundamental regulator of DC emigration from peripheral tissues, such as skin, makes it an attractive target for the treatment of autoimmune skin diseases or other autoimmune disorders and in transplantation.
We thank S. Rose-John (University of Kiel, Kiel, Germany) for rIL-6/IL-6R fusion protein, S. Krautwald (University of Kiel) for purified chimeric CCL19/ELC-human Fc fusion protein, Z. Guhe (Rheinisch-Westfälische Technische Hochschule) for technical assistance, A. Teubner (Rheinisch-Westfälische Technische Hochschule) for supporting animal experiments, and S. Diebold (King's College London, London, U.K.) for critical reading of the manuscript.
This work was supported in part by grants from the Deutsche Forschungsgemeinschaft (to M.Z.) and the Interdisciplinary Center for Clinical Research within the Faculty of Medicine at Rheinisch-Westfälische Technische Hochschule Aachen (VV B112-a to T.H. and TV M7 to M.Z.). T.H. was funded by the START program of the Faculty of Medicine at Rheinisch-Westfälische Technische Hochschule Aachen.
The online version of this article contains supplemental material.
Abbreviations used in this article:
dermal dendritic cell
hepatocyte growth factor
The authors have no financial conflicts of interest.