The challenging human pathogen Staphylococcus aureus has highly efficient immune evasion strategies for causing a wide range of diseases, from skin and soft tissue to life-threatening infections. Phenol-soluble modulin (PSM) peptides are major pathogenicity factors of community-associated methicillin-resistant S. aureus strains. In previous work, we demonstrated that PSMs in combination with TLR2 ligand from S. aureus induce tolerogenic dendritic cells (DCs) characterized by the production of high amounts of IL-10, but no proinflammatory cytokines. This in turn promotes the activation of regulatory T cells while impairing Th1 response; however, the signaling pathways modulated by PSMs remain elusive. In this study, we analyzed the effects of PSMs on signaling pathway modulation downstream of TLR2. TLR2 stimulation in combination with PSMα3 led to increased and prolonged phosphorylation of NF-κB, ERK, p38, and CREB in mouse bone marrow–derived DCs compared with single TLR2 activation. Furthermore, inhibition of p38 and downstream MSK1 prevented IL-10 production, which in turn reduced the capacity of DCs to activate regulatory T cells. Interestingly, the modulation of the signaling pathways by PSMs was independent of the known receptor for PSMs, as shown by experiments with DCs lacking the formyl peptide receptor 2. Instead, PSMs penetrate the cell membrane most likely by transient pore formation. Moreover, colocalization of PSMs and p38 was observed near the plasma membrane in the cytosol, indicating a direct interaction. Thus, PSMs from S. aureus directly modulate the signaling pathway p38–CREB in DCs, thereby impairing cytokine production and in consequence T cell priming to increase the tolerance toward the pathogen.
The Gram-positive bacterium Staphylococcus aureus is an opportunistic pathogen that causes soft-tissue and systemic infections. Twenty percent of the population is permanently colonized with S. aureus, and antibiotic treatment is often ineffective because the strains develop resistance. Methicillin-resistant S. aureus (MRSA) occurs frequently in hospital-associated (HA) infections and community-associated (CA) diseases (1, 2). CA-MRSA strains cause mainly skin and soft tissue infections in healthy individuals, with USA 300 as the most prominent strain worldwide. CA-MRSA strains express a wide range of virulence factors consisting of Panton-Valentine Leukocidin, α-toxin and phenol-soluble modulin (PSM) peptide toxins (1, 2). PSM peptides are secreted by CA-MRSA strains in much higher concentrations than by HA-MRSA strains, and they are essential virulence factors in mouse models of sepsis and soft tissue infection (3).
PSM peptides comprise seven different members, all arranged into an amphipathic α-helix. These members include five α-peptides (PSMα1-4 and δ-toxin) with 20–25 aa length and two β-peptides (PSMβ1-2) with 44 aa (3, 4). PSM peptides can affect the generation of bacterial biofilms because of their physical and chemical characteristics and their detergent activities (4). They can attract and activate human neutrophils at nanomolar concentrations, whereas at micromolar concentrations, they induce neutrophil lysis with their ability to form transient pores (3, 5). Nanomolar concentrations of PSM peptides are recognized by the human formyl peptide receptor 2 (FPR2) on neutrophils. Furthermore, this receptor recognizes the pathogenicity status of bacteria and adapts the immune reaction (6). There is a substantial body of research on the influence of PSM peptides on innate immune cells; however, little is known about cells of the adaptive immune system.
Dendritic cells (DCs) are the most important APCs that trigger immune responses. They link the innate and adaptive immune system by their ability to recognize pathogens and to activate B and T cells (7). Moreover, DCs are mediators of anti-inflammatory immune responses inducing tolerance (8). Mouse DCs express mFPR2 and are attracted by PSM peptides such as neutrophils, although higher concentrations are needed (9). In contrast to neutrophils, DCs are not lysed by micromolar PSM concentrations (9). Their phenotype changes by PSM peptide treatment showing decreased endocytosis and increased TLR2 ligand-induced secretion of IL-10, whereas TNF, IL-12, and IL-6 secretion is abrogated. Consequently, DCs treated with PSM peptides demonstrate a decreased priming ability for T helper 1 cells, but enhanced induction of Foxp3+ regulatory T cells (Tregs). However, the tolerogenic phenotype of DCs caused by PSM peptides is mFPR2 independent (9). A possible explanation is the ability of PSM peptides to generate transient pores into the cell membrane (5), thereby enabling access to the cytosol; however, the signaling pathway involved in the induction of tolerogenic DCs remains elusive.
DCs detect pathogens via microbial products by pattern recognition receptors (10), such as TLRs and NOD-like receptors (11, 12). TLRs are important regulators of the immune response because they initiate the production of different cytokines and chemokines (13). In myeloid DCs, TLR2 agonists induce the expression of anti-inflammatory IL-10 via activation of NF-κB, p38, and ERKs (10).
In this study, we show that TLR2 ligand stimulation of DCs in combination with PSMα3 induces the phosphorylation of p38-cAMP response element binding-protein (CREB) pathway, independently of FPR2. Inhibition of the p38–CREB pathway reduced IL-10 secretion and induction of Tregs by DCs. Furthermore, our results point toward a specific interaction of cytosolic PSMs with p38, thereby potentially increasing the tolerance toward the pathogen.
Materials and Methods
Female C57BL/6JolaHsd mice were purchased from Janvier (St. Berthevin, France). FPR2−/− mice (14) with a genetic C57BL/6 background were bred in the animal facilities of the University Clinic of Tübingen. All mice were held under specific pathogen-free conditions, provided food and water ad libitum, and used for experiments between 6 and 12 wk of age. Animal experiments were performed in strict accordance with the German regulations of the Society for Laboratory Animal Science and the European Health Law of the Federation of Laboratory Animal Science Associations. The protocol was approved by the Regierungspräsidium Tübingen (Anzeige 09.01.2014).
Generation of bone marrow–derived DCs
RPMI 1640 medium (Merck) supplemented with 10% FCS (FBS; Sigma-Aldrich), 2 mM glutamine (Life Technologies), 100 U/ml penicillin–streptomycin (Life Technologies), 50 μM 2-ME (Roth), 1 mM sodium pyruvate (Merck) and 1× nonessential amino acids (Merck) was used in all cell culture experiments. Bone marrow–derived DCs (BMDCs) were prepared using GM-CSF as described previously (9, 15, 16). Briefly, 2 × 106 bone marrow cells, flushed from the femurs and tibias of C57BL/6 and FPR2−/− mice, were seeded in 100-mm dishes in 10 ml medium containing 200 U/ml GM-CSF. After 3 d, an additional 10 ml of fresh medium containing 200 U/ml GM-CSF was added to the cultures. On day 6, half of the culture supernatant was replaced with fresh medium containing GM-CSF. At day 7–8, the slightly attached cells were used for the experiments described in this report.
Formylated PSMα3 and δ-toxin peptides with the recently published sequence (3) and FITC labeled PSMα2 and PSMα3 and OVA323–339 were synthesized in house. BMDCs were treated with S. aureus cell lysates specifically activating TLR2 (9). Treatment of BMDCs with 3 μg/ml S. aureus cell lysates was done simultaneously in combination with PSMα3 peptide (10 μM). Where indicated, BMDCs were pretreated with the following inhibitors in different concentrations for 1 h: BAY11-7082 (NF-κB inhibitor; Selleckchem), PD98059 (p-ERK inhibitor; Merck), SB203580 (p-p38 MAPK inhibitor; Merck) and Gö6976 (MSK1 inhibitor; Cell Signaling).
Cytokine production by BMDCs
BMDCs (2.5 × 105) were seeded in 96-well plates, pretreated with inhibitors, and incubated with S. aureus cell lysate and PSM peptides. Supernatants were collected after 24 h for IL-10 (BD Biosciences), and ELISAs were performed according to the manufacturer’s instructions.
BMDCs (5 × 105) were seeded in 48-well plates and treated as described above. Cells were removed from the plate using Accutase (Sigma-Aldrich) and stained with 7-aminoactinomycin D (Biomol) or Aqua Life/Dead (Invitrogen) according to the manufacturer’s instructions to exclude dead cells. Cells were stained for 20 min at 4°C with extracellular Abs against CD11c-PE (N418; eBioscience) and MHC class II-FITC (M5/114.15.2; Miltenyi). For p-CREB staining, cells were fixed and permeabilized with Foxp3 Staining Buffer Set (eBioscience) and stained with primary Ab phospho-CREB mAb (Ser133; clone 87G3; Cell Signaling) for 30 min in the dark at room temperature followed by secondary goat anti-rabbit IgG-DyLight649 (Jackson ImmunoResearch) for 15 min at 4°C. To detect intracellular p-ERK, p-p38 and p-NF-κB BMDCs were fixed with 2% paraformaldehyde (VWR) in PBS, permeabilized with 90% freezing methanol (Applichem) and stained with the primary Abs to phospho-p44/42 MAPK (Erk1/2; Thr202/Tyr204; clone 197G2), phospho-p38 MAPK (Thr180/Tyr182; clone 12F8) and phospho-NF-κB p65 (93H1) (all from Cell Signaling) for 60 min in the dark at room temperature followed by goat anti-rabbit IgG-PE-Cy7 (Santa Cruz Biotechnology) for 15 min at 4°C. PBS with 0.5% BSA (Biomol) was used for all incubations and washing steps. At least 50,000 cells were acquired using a Canto-II or LSRFortessa flow cytometer (BD Biosciences) with DIVA software (BD Biosciences) and were further analyzed using FlowJo 10.0.7r2 software (Tree Star).
Multispectral imaging flow cytometry
BMDCs (1 × 106) were seeded in 1.5 ml Eppendorf tubes and stimulated for different times with FITC-labeled PSMα2 and PSMα3 (0.5 μM) alone or in combination with OVA-Alexa647 (0.5 μM). Cells were washed three times and stained with Zombie NIR (BioLegend) according to the manufacturer’s protocol to exclude dead cells. Cells were stained with CD11c-PE (N418, eBioscience) and MHC class II-eFluor 450 (M5/114.15.2, eBioscience) for 15 min at 4°C. Next, cells were fixed and permeabilized using the Foxp3 Staining Buffer Set (eBioscience). For localization analysis, cells were incubated with a primary Ab against p38 (clone 27, BD Biosciences) or phospho-p38 (T180/Y182 12F8 Rabbit mAb Lot #9; Cell Signaling) for 30 min at room temperature; 5% NGS in PBS was used to block unspecific binding sites. Goat anti-mouse IgG-DyLight594 (Abcam) or Goat anti-rabbit IgG-PE/Cy7 (Santa Cruz Biotechnology) were used for 15 min at 4°C as secondary Abs to stain p38 and phospho-p38, respectively. Images of up to 100,000 BMDCs were then acquired with multispectral imaging flow cytometry (MIFC) using the ImageStreamx mkII with the INSPIRE instrument controller software. The data were analyzed using the IDEAS analysis software (run on the Amnis flow cytometer manufactured by Merck Millipore), which allows an objective and unbiased analysis of thousands of images per sample on the single cell level. The same range of pixel intensity was set for all samples within an experiment, and all samples were gated on CD11c+MHC II+ cells as shown in Supplemental Fig. 4B.
T cell assay
BMDCs (5 × 104) were seeded in 96-well U-bottom plates and treated as described above. Splenic CD4+ T cells from C57BL/6 mice were purified using CD4+ T Cell Isolation Kit II (Miltenyi Biotec) according to the manufacturer’s instructions. CD4+ T cells (2 × 105) were added to the BMDCs and cultured in RPMI 1640 medium (Merck) supplemented with 20% FCS (FBS; Sigma-Adrich), 2 mM glutamine (Life Technologies), 100 U/ml penicillin–streptomycin (Life Technologies), 50 μM 2-ME (Roth), 1 mM sodium pyruvate (Merck), 10 mM HEPES-Buffer (Biochrom AG) and 1× nonessential amino acids (Merck). Ninety-six hours later, T cells were stained first with Zombie NIR (BioLegend) according to the manufacturer’s protocol to exclude dead cells, followed by CD4-BrilliantViolet510 (V4), CD3e-PerCP/Cy5.5 (BM10-37), CD25-PE-Cy7 (B6.1), and Foxp3-APC (FJK-16s [BioLegend], Foxp3 Staining Buffer Set [eBioscience]). One hundred thousand cells were acquired using LSRFortessa flow cytometer (BD Biosciences) with DIVA software (BD Biosciences) and were analyzed using FlowJo 10.0.7r2 software (Tree Star).
LDH release assay
BMDCs (2 × 105) were seeded in 96-well U-bottom plates and treated with 10 μM PSMα2, PSMα3, δ-toxin, or OVA for 10 min. Lactate dehydrogenase (LDH) was analyzed in the supernatant using the Cytotoxicity Detection Kit (Roche) according to the manufacturer’s protocol. Absorbance was measured at 490 nm using an ELISA reader.
Statistical analysis was performed with the GraphPad Prism 6 software (GraphPad, San Diego, CA) using one-way ANOVA with Bonferroni posttest. The differences were considered as statistically significant if p < 0.05 (*), p < 0.005 (**), p < 0.001 (***), or p < 0.0001 (****).
PSMs induce a sustained NF-κB p65 phosphorylation in TLR2-stimulated DCs
It has been shown that TLR2-stimulated DCs coincubated with PSMα peptides increase the production of the anti-inflammatory cytokine IL-10 (9). However, the intracellular signaling pathways involved in the PSM-induced cytokine modulation have not been elucidated. It was shown previously that the TLR-dependent activation of the NF-κB subunit p65 leads to an extended and enhanced IL-10 transcription in DCs (17). To investigate whether NF-κB signaling plays a role in the cytokine modulation by PSMs, BMDCs were treated with synthetic PSMα3, the TLR2 ligand S. aureus cell lysate, or a combination of both. We have previously shown that S. aureus cell lysate specifically activates TLR2 and no other PRRs (9). Phosphorylation of NF-κB p65 (p-NF-κB) was analyzed by flow cytometry.
The level of p-NF-κB in DCs was slightly increased after 60 min of treatment with PSMα3 and increased 1.5-fold with S. aureus cell lysate compared with untreated DCs (Fig. 1). DCs treated with S. aureus cell lysate and PSMα3 revealed a significant 2-fold increase of p-NF-κB (Fig. 1). Moreover, the increased NF-κB phosphorylation was prolonged from 30 to 240 min (Supplemental Fig. 1A). Similar results were observed for mFPR2 deficient DCs in comparison with DCs from WT mice (Fig. 1B), demonstrating an mFPR2-indepentent effect. These data show a cooperating effect of PSMα3 and the TLR2 ligand S. aureus leading to enhanced and prolonged NF-κB activation in DCs.
Enhanced MAPK phosphorylation in DCs induced by PSMs and TLR2
The virulence factor β hemolysin/cytolysin of group B streptococcus was shown to induce IL-10 secretion in macrophages by activating p38 MAPK (18). Furthermore, TLR2 stimulation of DCs leads to phosphorylation of ERK1/2, which induces IL-10 production (19). To investigate whether the MAPKs play a role in the cytokine modulation by PSMs, we stimulated DCs as described above and analyzed phopho-p44/42 MAPK (p-ERK1/2) or phospho-p38 MAPK (p-p38) by flow cytometry. Treatment of DCs with PSMα3 did not affect p-ERK, whereas a strong increase for S. aureus cell lysate was observed 15 min after treatment compared with untreated DCs (Fig. 2). DCs incubated with S. aureus cell lysate combined with PSMα3 revealed a 2.2-fold increase of ERK phosphorylation (Fig. 2). Similar results were observed for mFPR2-deficient DCs (Fig. 2B). These results were visible only shortly after treatment, as no increased p-ERK could be detected 30 min after stimulation (Supplemental Fig. 1B), indicating a strong but short activation of the ERK pathway.
No significant change in the phosphorylation of p38 was observed in DCs treated with PSMα3 compared with untreated cells over time, whereas in DCs treated with S. aureus, cell lysate phosphorylation of p38 was increased by 1.5-fold starting 30 min after treatment (Fig. 3, Supplemental Fig. 1C). DCs incubated with S. aureus cell lysate and PSMα3 for 30 min revealed a significant 2.5-fold increase of p38 phosphorylation, which was independent of mFPR2 (Fig. 3). The data likewise show a cooperating effect of PSMα3 and the TLR2 ligand S. aureus cell lysate for the activation of MAPKs in DCs independently of mFPR2.
Enhanced CREB phosphorylation in DCs induced by PSMs and TLR2
It was shown that, in macrophages and myeloid DCs p38, and ERK activate MSK1/2 that directly phosphorylate CREB, which eventually binds to the IL-10 promoter (8, 20). To investigate whether CREB signaling plays a role in the cytokine modulation by PSMs, DCs were stimulated with PSMα3, S. aureus cell lysate, and in combination, and CREB phosphorylation (p-CREB) was analyzed by flow cytometry (Fig. 4). Similar results as for p-p38, p-ERK and p-NF-κB were observed for p-CREB, with the highest phosphorylation levels in DCs treated with the combination of PSMα3 and S. aureus cell lysate for 30 and 60 min (Fig. 4, Supplemental Fig. 1D). PSMα3 in collaboration with TLR2 stimulation enhances the activation of various signaling pathways downstream of TLR2 involved in the production of the anti-inflammatory molecule IL-10.
The p38–CREB axis mediates IL-10 secretion in TLR2 and PSM-treated DCs
The production of IL-10 was analyzed to address whether the enhanced activation of the signaling pathways NF-κB, p38, and ERK has an influence on the modulation of cytokine secretion by PSMα3. IL-10 is exclusively produced by DCs stimulated with the combination of PSMα peptides and S. aureus cell lysate (Fig. 5) (9).
Pretreatment of DCs with various concentrations of the NF-κB inhibitor BAY 11-7082 had no effect on IL-10 production, except for the 10-μM concentration (Fig. 5A); however, this concentration was cytotoxic for the cells, as determined with a cell viability assay using 7-aminoactinomycin D (Supplemental Fig. 2A). Thus, enhanced NF-κB activation by PSMα3 is not involved in IL-10 production by DCs.
Preincubation of DCs with various concentrations of the p-ERK inhibitor PD 0325901 did not affect the IL-10 production after stimulation with S. aureus cell lysate and PSMα3, indicating no influence of enhanced ERK phosphorylation on IL-10 production (Fig. 5B). In contrast, pretreatment with the p-p38 MAPK inhibitors SB 203580 and BIRB 0796 lead to a concentration-dependent inhibitory effect of IL-10 production (Fig. 5C, 5D). Furthermore, inhibition of MSK1 by Gö6976, which acts downstream of p38 and ERK and upstream of CREB, revealed a concentration-dependent decrease of IL-10 secretion by DCs (Fig. 5E). No toxic effect was observed for the used inhibitor concentrations or for the vehicle (Supplemental Fig. 2B–F). The specificity of the inhibitors was assessed with flow cytometry for p38 or CREB phosphorylation. SB 203580 and BIRB 0796 prevented the phosphorylation of p38 and Gö6976 of CREB in a dose-dependent manner (Supplemental Fig. 3), respectively, whereas these inhibitors had no effect on activation of other signaling pathways (data not shown). Together, our data demonstrate the involvement of the p38–CREB axis in IL-10 production induced by PSMα3.
Enhanced activation of p38–CREB–IL-10 axis by PSMs in TLR2-stimulated DCs primes Tregs
Previously, we showed that increased IL-10 production by DCs upon TLR2 and PSM stimulation primes Tregs (9). To address whether the p38–CREB–IL-10 axis is involved in the priming of Tregs by PSMs, DCs were treated with p38 inhibitors prior to incubation with S. aureus cell lysate and PSMα3. Twenty-four hours later DCs were incubated with naive CD4+ T cells and T cell priming was assessed 4 d later by flow cytometry. The frequency of CD4+CD25+Foxp3+ T cells was significantly increased when DCs were treated with S. aureus cell lysate and PSMα3 compared with S. aureus cell lysate alone (Fig. 6, Supplemental Fig. 4A: gating strategy). Inhibition of p38 signaling by the inhibitors SB 203580 and BIRB 0796 prevented this increase in CD4+CD25+Foxp3+ T cells mediated by PSMα peptides in a dose-dependent manner (Fig. 6). Thus, PSMs induce Tregs by modulating the p38–CREB–IL-10 axis in DCs.
PSMs penetrate DCs by transient pore formation and directly interact with p38 MAPK in DCs
It has been reported that PSM peptides can induce an effective inflammatory immune response by binding to the FPR2 receptor, whereas their cytolytic activity is FPR2 independent (6, 9). It is assumed that PSMα peptides like δ-toxin are able to form transient pores (5). To address whether PSMs are internalized by DCs via mechanisms of Ag uptake (e.g., receptor mediated endocytosis WT and FPR2−/−) mice were incubated with fluorescently labeled PSMα peptides and analyzed with multispectral imaging flow cytometry. PSMα2 was located in the cytosol in WT and FPR2−/− DCs after 10, 30, and 60 min of incubation (Fig. 7A, Supplemental Fig. 4B [gating strategy], and data not shown). The frequency of PSMα2+ DCs was comparable in WT and FPR2−/− DCs (Fig. 7A), showing that PSMs penetrate DCs by an FPR2-independent mechanism. To address whether PSMs are actively internalized by DCs via macropinocytosis or receptor-mediated endocytosis, DCs from WT mice were incubated with FITC-labeled PSMα2 on ice, preventing actin rearrangement and thereby endocytosis, and were compared with incubation at 37°C enabling endocytosis. Incubation of FITC-labeled PSMα2 with DCs on ice did not prevent PSM penetration into DCs (Fig. 7B). Furthermore, simultaneous incubation of DCs with OVA-Alexa647, which is taken up by macropinocytosis and receptor-mediated endocytosis, and PSMα2-FITC on ice revealed intracellular PSMα2-FITC, whereas no OVA-Alexa647 was taken up by DCs (Fig. 7B). These data show that PSMs penetrate DCs independently of endocytosis most likely by pore formation.
To address the hypothesis that pore formation by PSMs is responsible for cell penetration, LDH release by DCs upon PSM-treatment was analyzed. LDH is a soluble cytoplasmic enzyme that is present in almost all cells and is released into the supernatant when the plasma membrane is damaged (21, 22). Indeed, 10 min after treatment with PSMα3, a significant amount of LDH was released from DCs that was comparable with the LDH release of δ-toxin, which are known to induce transient pore formation (23) (Fig. 7C). In contrast, almost no LDH release was observed when DCs were treated with OVA peptide, further supporting that PSMs induce transient pore formation in DCs.
PSMs were predominantly localized close to the plasma membrane in spots devoid of endosomal and lysosomal markers, independently of whether they were activated via TLR2 stimulation (Fig. 8 and data not shown). Instead, PSMs colocalized with p38 (Fig. 8A) and p-p38 (Fig. 8B) molecules, as shown by imaging flow cytometry. DCs treated with PSMs and S. aureus cell lysate showed an increased max pixel intensity of p-p38 gated on PSMα-FITC+p-p38+ DCs compared with DCs treated with PSMs alone, indicating an increased phosphorylation of p38 upon TLR2 activation (Fig. 8B and 8C). These data point toward a direct interaction of PSMα peptides with the p38 MAPK signaling pathway.
PSMs play a key role in the pathogenicity of CA-MRSA strains (3). They bind to human and mouse FPR2, thereby initiating chemotaxis of neutrophils and DCs (6, 9). We have previously shown that PSMα peptides induce a tolerogenic phenotype in DCs upon TLR2 stimulation, characterized by the production of IL-10 and impaired secretion of proinflammatory cytokines (9). Consequently, these tolerogenic DCs favored the priming of regulatory T cells (9). How PSMs induce IL-10 secretion and which signaling pathways are involved remained elusive. In this study, we show that PSMs penetrate DCs via transient pore formation, directly interact with p38, and upon TLR2 activation enhance its phosphorylation and downstream CREB activation. This consequently increased IL-10 production and induction of Tregs.
Pore-forming toxins (PFTs) constitute ∼25% of all bacterial toxins and represent the largest class of bacterial virulence factors (24, 25). The amphipathic PSMα peptide δ-toxin was shown to form receptor-independent transient pores in solution (5, 23). We hypothesize on the basis of these experiments, that δ-toxin and potentially other PSMs form dimers and bind to the cytoplasmic membrane at low peptide density, and oligomers span the membrane and induce pore formation at high peptide density (5, 23). However, the PSM side of action in DCs remains unclear. Do PSMs act as PFTs in DCs? Can PSMs reach the cytosol via pore formation or are they internalized by DCs via macropinocytosis or receptor-mediated endocytosis? We show that PSMs are located in the cytosol independently of FPR2 (Fig. 7). Although the localization of PSMs appears point-shaped, they are not in endosomes or lysosomes (data not shown) arguing against their active internalization by DCs via macropinocytosis or receptor-mediated endocytosis. In agreement, PSMs are similarly found in the cytosol when endocytosis (shown with OVA) is prevented during ice incubation. Furthermore, the direct interaction of PSMs with p38 supports this conclusion, as p38 is located in the cytoplasm and not in subcellular fractions (26, 27). Moreover, FPR2 is not involved in the cytotoxic activity of PSMs on neutrophils (6) or the production of IL-10 by DCs (9), which is strongly supporting the hypothesis that receptor independent processes are responsible for cytosolic localization of PSMs. Recently, Grosz et al. (28) demonstrated that PSMα peptides are required for phagosomal escape of various cytolytic S. aureus strains in professional and nonprofessional phagocytes enabling cytoplasmic replication of these strains. In conclusion, our data obtained by imaging flow cytometry and LDH release provide evidence that transient pore formation is mediating the transport of PSMs to the cytoplasm.
TLR2 agonists are able to induce the expression of anti-inflammatory IL-10 via activation of NF-κΒ, p38, and ERK in DCs (10). Furthermore, pathogens triggering the C-type lectin DC-SIGN can modify TLR signaling in DCs. Upon TLR-dependent signaling, DC-SIGN activates the kinase Raf-1, which acetylates the p65 subunit of NF-κB leading to an extended and enhanced IL-10 transcription (17). In TLR2-stimulated DCs, the phosphorylation of ERK is increased, which induces the IL-10 production and represses IL-12(p70) (19). Although NF-κB and ERK signaling were increased by PSMs, blocking of these signaling pathways with chemical inhibitors had no effect on IL-10 production by PSM-treated DCs. Our data indicate a direct interaction of PSMα peptides with p38, thereby enhancing its phosphorylation and via CREB activation (20) eventually leading to high IL-10 production in DCs upon TLR2 ligand stimulation. This conclusion is supported by Bebien et al. (18) showing that the virulence factor β hemolysin/cytolysin of group B streptococcus induces IL-10 secretion via p38 MAPK activation. PFTs activate MAPK signaling pathways in different eukaryotic cells; whether this is beneficial or detrimental for the pathogen is species dependent (29–36).
Besides PFTs, other pathogenicity factors of Gram-negative bacteria or viruses modulate p38 signaling in host cells (29, 37). YopJ from Yersinia pseudotuberculosis inhibits p38 and JNK phosphorylation thereby preventing the production of TNF-α in macrophages (15, 37). In addition, YopP from Yersinia enterocolitica (also termed YopJ) prevents IL-10 production by DCs (15, 29). Accordingly, many effector proteins of Gram-negative bacteria were shown to impair p38 MAPK activation, thereby preventing proinflammatory cytokine secretion (29, 38). Moreover, Leghmari et al. (38–41) suggest a new immune escape mechanism for HIV-1 infection, by which the Tat protein induces IL-10 production in monocytes in a p38 MAPK-dependent manner. Thus, we demonstrate a new function of the PFT PSMα peptides acting as further pathogenicity factors by modifying p38 MAPK signaling pathway.
The capacity to induce Tregs via production of anti-inflammatory molecules that may be secreted, membrane bound, or both define tolerogenic DCs. A variety of Treg differentiation models demonstrated the necessity of IL-10 secretion by tolerogenic DCs for tolerance induction (39–42) and for the maintenance of suppressive Tregs upon strong inflammatory signals (9, 42–44). Like PSMα peptides, Candida albicans, Cryptococcus neoformans, and Fasciola hepatica subvert the immune system by promoting DC tolerogenicity and Treg differentiation (9, 43–45). How these pathogens impair recognition and signaling remains unknown. In this study, we describe the p38–CREB–IL-10 axis as molecular mechanism for DC tolerogenicity and Treg differentiation induced by PSMα peptides in vitro (Fig. 6). Whether this holds true in vivo has to be shown.
For the treatment of increasingly antibiotic-resistant bacteria, the development of new, narrower-spectrum or virulence-targeted antimicrobial therapeutics is necessary (45, 46). PFTs were used as live vaccines in various disease models, but successful immunization against a PFT does not always prevent disease (46). Other examples such as inhibiting PFTs and using competitive inhibitors were shown to prevent or cure the infection (46). Furthermore, boosting host defense (e.g., by using drugs modulating MAPK signaling pathways) is discussed as immunotherapy against infections. Our data point toward the use of a p38 inhibitor in the case of CA-MRSA infection, possibly preventing the induction of tolerogenic DCs and thereby immune escape.
We thank the Core Facility Flow Cytometry and Core Facility ImageStreamX mk II for technical assistance; Stefan Stevanović, Patricia Hristić, and Nicole Zuschke for providing PSM peptides; Andreas Peschel for discussions; and Kristin Bieber and Ingo Autenrieth for critical reading of the manuscript.
This work was supported by the German Research Foundation Grant SFB685 and the European Social Fund of Baden-Württemberg (Margarete von Wrangell Program; to S.E.A.) and by a grant from the Ministry of Science, Research and the Arts of Baden-Württemberg (SI-BW 01222-91) and the German Research Foundation (INST 2388/33-1).
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.