Induction of C3 and CCL2 by C3a in Keratinocytes: A Novel Autocrine Amplification Loop of Inflammatory Skin Reactions¹ Free

The anaphylatoxin complement fragment-3a (C3a)³ is a pleiotropic molecule that plays an important role in innate and adaptive immune responses (1). The effects of C3a are mediated upon binding to the C3aR, which is a G protein-coupled, pertussis toxin-sensitive, and seven-transmembrane-spanning receptor (2). The expression of C3aR has been described on a variety of cell types. C3aR-positive bone marrow-derived cells that participate in host immune responses include mast cells, eosinophils, basophils, neutrophils, dendritic cells (DCs), macrophages, T cells, and B cells (3, 4, 5, 6). The C3aR is also expressed in nonmyeloid tissues, including kidney, brain, lung, heart, and liver (7, 8, 9, 10).

Unlike C5a, the role of C3a in inflammatory diseases is not well defined. Recent studies suggest that C3a-C3aR is involved in allergic diseases. C3a has been shown as an activating and chemotactic factor for human eosinophils, mast cells, and leukocytes (11, 12, 13). C3a induces a variety of inflammatory mediators such as IL-1, IL-6, IL-8, and TNF-α (14, 15). C3a provides costimulatory signals to B cells and enhances the humoral response against specific Ags, thus providing an interface between the innate and adaptive immune systems (16).

C3a can exhibit anti-inflammatory properties by suppressing LPS-induced TNF-α, IL-1β, and IL-6 production from freshly isolated PBMCs (17). C3aR^−/− mice showed an increased mortality, compared with control mice in a LPS model of septic shock (18), demonstrating an important protective role of C3aR in endotoxin shock by attenuating LPS-induced proinflammatory cytokine production. In a murine model of allergic airway disease, genetic deletion of the C3aR protected the development of Ag-induced airway hyperresponsiveness (19), suggesting a role of C3a in the pathogenesis of allergic asthma. These studies indicate that C3aR may have a dual role in modulating the immune responses.

The liver is a major source of C3 present in the serum (20); however, locally synthesized complement components play an important role in inflammation at the target site (21). Epidermal keratinocytes (KCs) constitutively secrete C3, factor B, and factor H and have been shown to release high amounts of C3 after stimulation with proinflammatory cytokines (22, 23). Increased levels of C3a in the circulation have been found in diseases such as adult respiratory distress syndrome (24), asthma (25), psoriasis, and atopic dermatitis (AD) (26). C3 fragments are also found intralesionally in inflammatory diseases such as psoriasis (27), eczema (28), and asthma (19, 29). Recently, C3b deposition was detected in the stratum corneum of the lesions of psoriasis vulgaris, and a cytoplasmic distribution of C3b as well as C3a was observed in the subcorneal to spinous KCs of allergic contact dermatitis, AD, and psoriatic lesional KCs (30). In the skin, C3 can be activated by mechanical injury and UV radiation injury (31). These data suggest that complement activation is pathogenic and takes place in the epidermal compartment of the skin.

These findings prompted us to characterize the expression and function of C3aR on epidermal KCs. Our study demonstrates the expression of the functional C3aR on epidermal KCs. We observed the induction of C3 and CCL2 by C3a in KCs and the activation of C3 by skin mast cells tryptase independent of activation of the entire complement cascade. In conclusion, our data suggest a novel mechanism by which the complement system may influence the course of skin inflammation.

Human C3a, C3a-desArg, C5a, and SB 290157 (C3aR antagonist) were purchased from Calbiochem. Recombinant human IFN-γ and TNF-α were obtained from ImmunoTools; IL-1β and IL-4 from R&D Systems; human IFN-α-2a from Roche; and IL-13, IL-1β, and IFN-β from PromoCell. Human C3 was provided by Dr. M. Oppermann (Georg-August-University, Göttingen, Germany).

Primary cultures of normal human KCs were prepared as described previously (32). In brief, the single-cell suspension of KCs was cultured in serum-free KCs growth medium (Keratinocyte Growth Medium 2 kit; PromoCell). All cell cultures were incubated in a humidified atmosphere containing 5% CO₂ at 37°C and were used at the passage two to six. Hydrocortisone-free medium was used for all experiments. The purity of KCs was verified by the expression of the epithelial marker cytokeratin (anti-human cytokeratin Ab, clone MNF-116; DakoCytomation). All cells (>95%) were found to be uniformly positive for cytokeratin. However, no cells were stained with CD1a (DC), CD3 (T cells), or CD90 (fibroblasts) mAb.

The expression of C3aR on the surface of KCs was assessed by indirect staining as described previously (5). The following unlabeled mouse anti-human mAbs against C3aR detecting different epitopes of the C3aR (33) were used: clones 17/hC3aRZ8 and hC3aRZ1 (provided by Dr. J. Zwirner, Georg-August-University, Göttingen, Germany), and clone 8H1 (BD Pharmingen). The expression of C5L2 was detected by rabbit anti-human C5L2 polyclonal Ab (HyCult). In a second step, cells were incubated for another 30 min on ice with an R-PE-labeled goat anti-mouse Ig (Dianova). Stained cells were measured by flow cytometry (FACSCalibur; BD Biosciences) and analyzed using CellQuestPro software (BD Biosciences). Cytokeratin (DakoCytomation) was detected by intracellular staining using the Cytofix/Cytoperm kit (BD Biosciences).

Skin punch biopsies from multiple sclerosis patients undergoing type 1 IFN (IFN-β) therapy were taken as described previously (5). The study was approved by the ethics committee of the Hannover Medical School and was conducted according to the Declaration of Helsinki Principles. Immunohistology was performed as described elsewhere in detail (34). Briefly, tissue specimens obtained by punch biopsy were shock-frozen in liquid nitrogen. Cryostat tissue sections (5-μm thick) were dried and then fixed for 10 min in acetone. The endogenous peroxidase of the cells was inhibited by an incubation for 15 min with 150 ml of PBS (to which 3 ml of 1 M sodium azide was added) and 0.5 ml of peroxide (30%). After three washes in PBS, the fixed sections were incubated in 5% normal goat serum (Invitrogen Life Technologies) in PBS to block FcR. The fixed sections were overlaid with a predetermined optimal concentration of anti-C3aR (hC3aRZ1) mAb containing 2% goat serum or with corresponding concentrations of isotype control mAbs in the same buffer. After a 1-h incubation in a moist chamber and three washing steps in PBS, the sections were overlaid with biotin-conjugated sheep anti-mouse Ig (Amersham Biosciences) and diluted 1/400 for 40 min at room temperature (RT), followed by three washing steps in PBS. The sections were incubated for 30 min at RT with a 1/1000 dilution of streptavidin-peroxidase (Dianova), washed, and then stained by immersion in 150 ml of chromogenic solution of 3-amino-9-ethylcarbazole (Sigma-Aldrich) containing N,N-dimethylformamide (Merck) and 0.1 ml of hydrogen peroxide (30%) for 8 min. The sections were counterstained with hemalum and then mounted in Faramount mounting medium (DakoCytomation).

KCs were cultured in KC growth medium without supplements for 2 h before stimulation with C3a (1000 ng/ml) and C3a-desArg (1000 ng/ml) for 30 min at 37°C. Nuclear extraction was performed using NE-PER nuclear and cytoplasmic extraction reagents according to the manufacturer’s instruction (Pierce). EMSA has been conducted as described previously to detect AP-1 (35). In brief, nuclear protein-DNA binding reactions were performed for 20 min at RT using LightShift Chemiluminescent EMSA kit (Pierce) according to the manufacturer’s instructions.

KCs were either prestimulated with IFN-γ + IFN-α or left unstimulated for 24 h. After 24 h, cells were washed two times carefully and stimulated with C3a for 6 h. Total RNA was isolated by using the RNeasy kit (Qiagen). Microarray experiments were performed using the human inflammation microarray (MWG Biotech), which contains 136 gene probes for inflammatory genes and 19 gene probes for housekeeping genes. The inflammation array is the second version of an inflammation microarray as described previously (36).

For quantification of gene expression, KCs were exposed to appropriate stimuli (as indicated) for 6 h and total RNA was isolated by using the High Pure mRNA isolation kit (Roche Molecular Biochemicals). The cDNA was synthesized with integrated removal of genomic DNA in the samples by using a QuantiTect reverse transcription kit (Qiagen). Quantitative real-time RT-PCR (qRT-PCR) was performed on a LightCycler PCR (Roche Molecular Biochemicals) using QuantiTect SYBR Green PCR kit (Qiagen) and Quantitect Primer (Qiagen) for C3aR, C3, CCL2, C5L2, and GAPDH. Specific targets were amplified using the following program: (PCR initial activation step) 15 min, 95°C and ramp 20°C per min; (denaturation) 15 s, 94°C, ramp 2°C per min; (annealing) 20 s, 55°C, ramp 2°C per min; (extension) 20 s, 72°C, ramp 2°C per min. For quantitative analysis, standard curves for the C3aR, C3, and CCL2 were created. These standard curves describing the PCR efficiencies of the target and the reference gene (GAPDH) allowed an efficiency-corrected quantification using the Relative Quantification software (Roche Molecular Biochemicals).

To assay C3a generation, either purified C3 or cell-free supernatants derived from KCs stimulated with IFN-γ (10 ng/ml) and TNF-α (10 ng/ml) for 24 h were incubated with 0.2 μg of skin mast cell-derived tryptase (Promega) for 1 h at 37°C. C3a was measured by a C3a-specific ELISA according to manufacturer’s instructions (Quidel).

In all experiments, KCs were grown in a 24-well plate (Nunc) in a 60–90% cell density in 0.5 ml of KC medium. Concentrations of C3 and chemokines (CCL2, CCL5) were measured in cell-free supernatants collected after stimulation with appropriate stimuli. Chemokines (CCL2 and CCL5) were quantified at protein levels by DuoSet ELISA (R&D Systems) according to manufacturer’s instructions. LTB₄ was measured using an LTB₄ assay (R&D Systems) according to the manufacturer’s instructions.

C3 was measured by sandwich ELISA using matched Ab pairs (a gift from Dr. M. Oppermann). ELISA plates (Nunc) were coated with 10 μg/ml capture Ab (clone B3/6) in coating buffer (NaHCO₃/Na₂CO₃, pH 9.5) overnight at RT. ELISA plates were washed with washing buffer (0.05% Tween 20 in PBS) and blocked with blocking buffer (PBS containing 1% BSA, 0.05% NaN₃) for 1 h. After washing, undiluted samples (100 μl) and standards were incubated for 2 h at RT, washed, and incubated with biotinylated C3 Ab (clone K13/16) at RT for 2 h. After washing, streptavidin-HRP (1/200 dilution, R&D Systems) was added for 20 min at RT, washed, and developed with substrate solution (R&D Systems) for 15 min. The reaction was stopped by adding 50 μl of 2 N H₂SO₄. The plate was read at 450 nm in a FluoStar ELISA plate reader (BMG Lab Technologies).

Although the C3aR is expressed on a variety of cells, the role of C3a–C3aR axis in the epidermal compartment of the skin has not been previously studied. To investigate whether human KCs respond to C3a, we first evaluated whether KCs express C3aR. Using three different mAbs against C3aR, we could show that KCs express C3aR (Table I). All anti-C3aR-specific mAbs showed positive staining (Table I). As depicted in Fig. 1, using C3aR mAb (clone 17/hC3aRZ8), we detected a low-level C3aR expression in ∼15–30% of KCs. Type 1 IFNs and IFN-γ are present at the site of skin inflammation in various skin disorders (37, 38). Stimulation of KCs with IFN-γ or IFN-α increased the expression of C3aR (Fig. 1, A and B). However, IFN-γ and IFN-α in combination showed significant up-regulation of C3aR expression on KCs (Fig. 1, A and B). In contrast, we could not observe significant effects of other cytokines such as TNF-α, IL-1β, IL-13, and IL-4 on C3aR expression (data not shown). After acquisition of samples on flow cytometer, surface staining of C3aR on KCs was further confirmed by immunofluorescence microscopy (data not shown).

In addition, we investigated the C3aR expression in type 1 IFN (IFN-β)-treated human skin by immunohistology. At the site of IFN-β injection, expression of C3aR was observed in the epidermal layer (Fig. 1 C). The C3aR expression was particularly prominent near the basement membrane. However, we did not observe C3aR expression in normal skin (data not shown), which might be due to the relatively low sensitivity of immunohistology.

To demonstrate the specificity of anti-C3aR mAb binding to KCs, we incubated KCs with C3a (2.5 μg/ml) for 30 min before C3aR staining. Preincubation of KCs with C3a led to the down-modulation of the binding sites for anti-C3aR mAb (Fig. 2,A). Furthermore, we investigated C3aR expression at the mRNA level. Using qRT-PCR, we demonstrated an up-regulation of the C3aR by IFN-γ and/or IFN-α at the mRNA level in KCs (Fig. 2 B).

To investigate the C3a induced signaling in KCs, we used EMSA to detect binding of the transcription factor AP-1 to its specific DNA consensus sequence. As depicted in Fig. 2 C, C3a induced AP-1 in KCs; however, no binding was observed with C3a-desArg, suggesting the specificity of the C3a-C3aR in KCs. Similar results were obtained with KCs pretreated with IFN-γ and IFN-α (data not shown). To identify potential molecular targets of C3a in KCs, we first used microarray gene expression analysis. We investigated the expression of a selected set of 136 genes with known relevance to inflammation (36), such as cytokines, chemokines, and matrix metalloproteinases. KCs, prestimulated with IFN-γ and IFN-α (up-regulation of C3aR expression) and unstimulated KCs were exposed to C3a (1000 ng/ml) for 6 h. Microarray analysis revealed up-regulation of C3 in prestimulated KCs (1.6-fold) after C3a stimulation. Additionally, some chemokine genes were induced by C3a stimulation in KCs after up-regulation of C3aR, namely CCL5 (2.3-fold induction), CCL2 (1.7-fold induction), CXCL5 (1.7-fold induction), CXCL10 (1.6-fold induction), and CXCL8 (1.5-fold induction). Without prestimulation, no induction of the above-mentioned genes by C3a was detected in KCs.

In the next series of experiments, we focused on the induction of C3 by its own cleavage product C3a. We confirmed our microarray data of C3 induction by C3a at the protein level by using a sensitive C3-ELISA. We observed that after C3a (1000 ng/ml) stimulation, KCs secreted 228 ± 28.52 ng/ml C3 after 24 h incubation at 37°C (Fig. 3,A). Induction of C3 in KCs by C3a stimulation was dose dependent, and 1000 ng/ml C3a was the optimal concentration to induce C3 production (Fig. 3 A).

We next studied the induction of C3 by C3a in KCs after up-regulation of C3aR (prestimulation with IFN-γ + IFN-α) at the RNA and protein level. We conducted qRT-PCR to demonstrate the induction of C3 by C3a stimulation in KCs. As shown in Fig. 3,B, C3a induced the expression of C3, suggesting a direct effect of C3a in generation of its own precursor molecule C3. Furthermore, the exposure of KCs prestimulated with IFN-γ + IFN-α to C3a (1000 ng/ml) resulted in a strong induction of C3 at protein level in a time-dependent manner (Fig. 3, C and D).

In skin inflammation, mast cells have been observed mostly in upper dermis close to basal lamina and in the epidermis of AD (39, 40). In a previous study, C3 cleavage by tryptase derived from pulmonary mast cells in lung inflammation was shown (41). In this study, we tested the effects of skin mast cell tryptase on C3. Tryptase derived from skin mast cells in different concentration was incubated with C3 for 1 h at 37°C. As shown in Fig. 4 A, skin mast cell-derived tryptase generated C3a from C3. KCs stimulated with IFN-γ and TNF-α are rich sources of C3 (23). Incubation of supernatant from IFN-γ and TNF-α stimulated KCs with 0.2 μg of skin mast cell tryptase for 1 h at 37°C resulted in generation of C3a (Fig. 4 B).

Recently, C3a has been shown to stimulate chemokine production in epithelial cells (15). Since our microarray analysis revealed the up-regulation of CCL2 expression after C3a stimulation in KCs prestimulated with IFN-γ + IFN-α, we investigated whether C3a could influence the expression of CCL2 at the protein level. There was no effect of C3a on CCL2 expression in non-prestimulated KCs (Fig. 5,A). Basal expression of CCL2 in KCs prestimulated with IFN-γ + IFN-α was high; however, C3a stimulation further increased the secretion of CCL2 significantly (Fig. 5,A). Low doses of C3a (100 ng/ml) up-regulated the expression of CCL2 in KCs prestimulated with IFN-γ + IFN-α. Maximal response in CCL2 induction was observed with the optimal dose of 1000 ng/ml (Fig. 5 B). The stimulation of C3a clearly up-regulated the production of CCL2 in KCs prestimulated with IFN-γ + IFN-α, suggesting that KCs respond to C3a in terms of chemokine production. Thus, the induction of C3 and CCL2 after C3a stimulation in KCs suggests that KCs express functional C3a receptor. Furthermore, we also investigated the secretion of CXCL8 and CXCL10 (which have been observed to be up-regulated at the mRNA level in microarray analysis), and LTB₄. We did not observe a significant induction of these chemokines at the protein level and of LTB₄ by using specific ELISAs (data not shown). Using CFSE labeling-based proliferation assay, we did not observe an effect of C3a on KCs proliferation (data not shown).

To study whether the induction of CCL2 and C3 by C3a in KCs is due to C3a–C3aR interactions, we used a potent small molecule described as an antagonist of the C3aR, SB 290157 (42). Unexpectedly, we observed that SB 290157 acted as an agonist on KCs. After incubation of KCs (pretreated with IFNs) with SB 290157 for a period of 24 h, CCL2 and C3 secretion were increased by 1.44 ± 0.14-fold (n = 3) and 1.29 ± 0.23-fold (n = 3).

To describe the specificity of C3a–C3aR interaction, which is responsible for C3a response in KCs, we stimulated KCs with C3a-desArg, a split product that does not signal via C3aR (43). C3a-desArg stimulation did not induce C3 (Fig. 6,A) or CCL2 (Fig. 6,B) significantly in KCs. However, we observed only a trend of C3 and CCL2 induction in our experiments, which was not as prominent as after C3a stimulation (Fig. 6). Recently, an alternative C3a and C3a-desArg receptor named C5L2 has been described (44). We did not detect the expression of C5L2 on KCs. C5L2 expression was analyzed in unstimulated and KCs stimulated with IFN-γ, TNF-α, IFN-α, or in combinations of these cytokines for 6 h (mRNA) and 24 h (protein level) using qRT-PCR and flow cytometry (data not shown). As a further control, we used C5a, which has similar electrochemical properties like C3a but does not bind to a specific receptor on KCs (34). As shown in Fig. 6 A, we did not observe the induction of C3 after stimulation with C5a. Therefore, we conclude that C3a specifically interacts with C3aR on KCs.

In this study, we demonstrate for the first time the expression and regulation of functional C3aR on human epidermal KCs. In the skin, monocytes, mast cells, neutrophils and activated T cells express C3aR (4, 5, 45, 46). The C3aR expression induced under inflammatory conditions as shown by the up-regulation of C3aR by type 1 (IFN-α and IFN-β) and type 2 (IFN-γ) IFNs in vitro. This was in line with other recent findings of the regulation of C3aR. IFN-γ, IFN-α, and IFN-β have been shown to up-regulate the C3aR on T cells, monocyte-derived DCs, U937 and other myeloblastic cells lines (5, 45, 46). However, we did not observe modulation of C3aR expression in KCs by IL-1 or TNF-α in KCs as described in monocyte-derived DCs and mononuclear cells (45, 47). Type 1 IFNs (IFN-α and IFN-β) signal via the same receptor (48) and have similar effects. To demonstrate the C3aR expression in vivo, we used skin biopsies from multiple sclerosis patients undergoing IFN-β therapy. The expression of C3aR was observed in the epidermal layer of skin biopsies obtained from sites of IFN-β injection in multiple sclerosis patients.

To further show the specificity of the C3aR expression, reduction of mAb binding could be demonstrated following preincubation of cultured KCs with C3a (C3aR ligand). Down-regulation of surface receptor expression upon exposure to its ligand is a well-documented phenomenon for members of the family of G protein-coupled receptors with seven transmembrane domains and has also been demonstrated for the C3aR (49).

We have demonstrated the effects of C3a and C3a-desArg on the induction of AP-1 binding. C3a but not C3a-desArg induced AP-1 binding activity. Previously, C3a has been described to induce MAPK, in particular ERK, and AP-1 in a mast cell line (50). In another study, C3a as well as C3a-desArg have been described to enhance LPS-induced AP-1 binding activity in PBMCs (51). In this study, we have used purified KCs, which do not express the alternative C3a and C3a-desArg receptor C5L2. This may explain why we did not observe C3a-desArg mediated effects in our system.

Several studies envisaged the role of C3a in skin inflammatory disorders particularly in psoriasis and allergic contact dermatitis (52). Little is known about the mechanism by which C3a may influence the pathogenesis of skin inflammatory disorders. To define the effects of C3a in KCs, microarray analysis was used to identify potential targets. In microarray analysis, we did not observe significant effects of C3a on KCs, which had not been prestimulated with IFN-γ and IFN-α. However, after up-regulation of C3aR by IFNs, expression of five genes was increased, and two of those were confirmed at RNA and protein level. The inability of KCs to respond toward C3a before up-regulation of C3aR may be explained by the low surface expression of C3aR on KCs, the short period of stimulation used (6 h), or by the stringency of the statistical condition of microarray analysis.

An interesting and novel finding of the present work was the induction of C3 by its own cleavage product C3a in KCs and C3 activation by tryptase derived from skin mast cells. Recently, KCs have been described as a rich source of C3 (23). In KCs, induction of C3 by C3a clearly suggests the amplification of C3a mediated effects in an autocrine fashion. Local expression of C3, pivotal to complement activation pathways, was described in a range of inflammatory skin disorders such as psoriasis and allergic contact dermatitis (30). Inhibition of the C3 synthesis in KCs by psoralen and UVA light therapy and UVA has been suggested as one of the mechanism to suppress the effector phase of the immune response for the many skin diseases such as psoriasis (53). Recently, in a C3^−/− mouse, reduction in priming of Th cells and cytotoxic T cells and impairment of the migration of virus-specific T cells after influenza virus infection have been described (54). In a murine model, absence of C3aR resulted in exaggerated Th2 responses against epicutaneously introduced Ag by inhibiting IL-12 derived from APCs (52). Thus, C3 synthesis by C3a locally in the epidermis and C3a signaling through C3aR may contribute as one of the key factors in the pathogenesis of T cell mediated skin diseases.

Previously, tryptase from pulmonary mast cells has been shown to cleave C3 into C3a (41). In this study, we confirm that skin mast cell tryptase can generate C3a from C3 as well. Tryptase is collectively believed to be involved in the immuno-pathogenesis of psoriasis; however no operative mechanism has been proposed for its role in the genesis of skin inflammation (55). In the skin inflammation, mast cells have been observed mostly in upper dermis close to basal lamina (39) and recently in stratum granulosum of the epidermis in AD (40). Thus, it is conceivable that, during inflammation, high C3 levels in skin and tryptase released from activated mast cells could lead to C3a generation. Additional mechanisms for the activation of complement in inflammatory skin diseases have been suggested. First, activation of an Ab-independent alternative pathway by stratum corneum was proposed (56). Second, an activation of mannose-binding lectin pathway by surface Ag of carbohydrate nature, is also possible. Since C3a is considered as a chemotactic factor for mast cells, monocytes, and granulocytes, our results point to the fact that C3a generated in skin compartment can recruit leukocytes. The synthesis of C3 by C3a and its further activation into C3a by skin mast cell tryptase reveal a novel mechanism by which complement system may influence the course of skin inflammation.

We confirmed the induction of CCL2 by C3a stimulation in KCs after up-regulation of C3aR at the protein level. The induction of CCL2 by C3a may account for biologically important effects such as transepithelial migration of monocytes, DCs, and CD4⁺ T cells during skin inflammation, which are the major cause of damage to the epidermis seen in inflammatory skin diseases (37, 57). Basal KCs strongly express CCL2 in patients with AD (58). However, other cytokines, such as IFN-γ, TNF-α, and IL-4/IL-13 in combination with IFN-γ, have also been shown to induce CCL2 in KCs (57). Thus, it is tempting to speculate that signaling through up-regulated C3aR in KCs may contribute to the pathogenesis of skin inflammation by recruiting leukocytes at the site of inflammation. In a murine model of experimental lupus nephritis, signaling via C3aR has been illustrated as one of the key factors for the pathogenesis of lupus nephritis (59).

To investigate whether the induction of CCL2 and C3 by C3a was due to C3a–C3aR interactions, we used the small nonpeptide SB 290157 considered as a C3aR antagonist (42). Surprisingly, we observed that SB 290157 had an agonistic property in our experiments as observed by induction of C3 and CCL2 in KCs. Recently, SB 290157 has been described to induce calcium mobilization in a variety of primary cells and cell lines, including transfected RBL cells and U-937 cells (60). Thus, SB 290157 may also act as an agonist of C3aR on human KCs.

The microarray analysis has indicated the increase in CXCL8, CXCL10, and CCL5 expression. However, using ELISAs, we could not confirm the induction of CXCL10 and CCL5 at protein level in KCs. We observed inconsistent induction of CXCL8 in KCs after C3a stimulation (data not shown). Previously, C3a has been shown as a regulator of CXCL8 production in renal epithelial cells (15). In this study, we used skin KCs, and therefore discrepancy in CXCL8 induction may be explained by the differences in anatomical position or by the growth conditions used.

C3a-desArg stimulation did not induce C3 and CCL2 in KCs significantly. However, we observed a trend of CCL2 and C3 induction after C3a-desArg stimulation. Recently, another functional receptor of C3a and C3a-desArg, named C5L2, has been described (44). However, we could not detect C5L2 expression on unstimulated KCs and KCs stimulated with variety of cytokines. Roles of C3a-desArg have been described in regulation of TNF-α, IL-1β, and IL-6 synthesis in human PBMCs (51, 17). To further describe the specificity of C3a-C3aR system, we used C5a, which has similar electrochemical properties as C3a. We did not observe the induction of C3 after stimulation with C5a. Thus, these results suggest that C3a specifically interacts with C3aR on KCs. Nevertheless, further studies are required to decipher the effects of C3a-desArg on KCs.

In conclusion, our data demonstrate a physiological role of C3a–C3aR axis in regulating skin inflammation. Finally, the induction of C3 and CCL2 in KCs and C3 activation in the epidermal compartment might represent one of the key factors in the pathogenesis of inflammatory skin diseases.

We thank Gabriele Begemann and Susanne Mommert for helping in immunohistology and qRT-PCR.

The authors have no financial conflict of interest.