Human Formyl Peptide Receptor 2 (FPR2/ALX) Senses Highly Pathogenic Staphylococcus aureus

Formylated and Nonformylated PSMs Induce Ca²⁺ Influx in Human Neutrophils, which Cannot be Blocked by the FPR1 Inhibitor CHIPS

S. aureus PSMs have been shown to attract neutrophils but structure-function relationships of PSMs with regard to their chemotactic potential are poorly defined (Wang et al., 2007). We found PSMα3, the most active of the PSMs to attract human neutrophils at nanomolar concentrations (Figure 1A,B). The decrease of chemotaxis at PSMα3 concentrations above 100 nM reflects the activation of neutrophils by high amounts of GPCR ligands leading to halt of migration (Christophe et al., 2001) and causing the typical bell-shaped dose-response curve (Figure 1A). PSMα3 also elicited intracellular calcium ion flux in neutrophils, which is known to accompany chemotactic migration (Hartt et al., 1999; Kalmar and Van Dyke, 1994) and can be measured very sensitively. In contrast, a PSM variant with scrambled amino acid sequence had no such activity (Figure S1). Since PSMs usually retain their N-terminal formyl-groups during secretion (Wang et al., 2007) we assumed them to be recognized by the human FPR1. Unexpectedly even the non-formylated PSMs, which are produced under certain environmental conditions depending on the varying activity of S. aureus peptide deformylase (Somerville et al., 2003), elicited chemotaxis and Ca²⁺ ion flux in neutrophils albeit at lower concentrations as the formylated peptides (Figure 1B). Moreover, the FPR1-inhibitory protein CHIPS blocked neutrophil response to the synthetic control ligand fMLF but had no impact on the activity of formylated or non-formylated PSMs (Figure 1C), which was another indication that a receptor different from FPR1 is responsible for the proinflammatory neutrophil responses to PSMs.

PSMs Are Specifically Detected by FPR2/ALX

We confirmed that PSMs are recognized by a G protein-coupled receptor as pertussis toxin (PTX), a general antagonist of this type of receptor blocked the calcium flux-inducing activities of representative formylated or non-formylated PSMs (Figure 1C). Notably, we found that FLIPr, a specific inhibitor of the FPR1 paralog FPR2/ALX leads to almost complete inhibition of neutrophil response to both, the formylated and non-formylated PSMs suggesting that these peptide toxins represent a new class of FPR2/ALX ligands. Control experiments with the synthetic FPR2/ALX ligand MMK1 confirmed the specificity of FLIPr (Figure 1C). Similar results were also obtained when the same set of inhibitors was tested for blocking neutrophil chemotaxis (data not shown).

In order to further evaluate the role of FPR2/ALX in recognizing PSM peptides, we analyzed the response to PSMs of HL60 cells stably transfected with the human FPR1, FPR2/ALX, or FPR3 receptors. While untransfected HL60 cells were unresponsive (data not shown), the FPR1 and FPR2/ALX-transfected cells responded to the corresponding control ligands fMLF and MMK1, respectively, as expected (Figure 2A). All eight formylated PSMs activated FPR2/ALX-transfected cells very efficiently and, depending on the individual peptide, led only to residual activation of FPR1 or FPR3-transfected cells. Similar results were obtained with the corresponding non-formylated PSMs. Dose-response curves of representative PSMs revealed effective concentrations in the low nanomolar range for formylated PSMα3 and δ-toxin (Figure 2B) and the other alpha PSMs (data not shown). pEC₅₀ (negative logarithm of half-maximal effective concentration) values of formylated PSMs reached 8.67 (PSMα3), 7.37 (δ-toxin) or 6.47 (PSMβ2) (Table 1). These values were considerably lower for non-formylated peptides thereby confirming that the formyl groups are not essential but have a considerable impact on PSM activities. Neutrophils revealed similar or somewhat lower potencies for PSMs as FPR2/ALX-transfected HL60. (Figure 2B, Table 1). Thus, formylated PSM alpha peptides are among the most potent naturally occurring FPR2/ALX peptide ligands, which have reported pEC₅₀ values between 4.60 and 9.00 (Ye et al., 2009).

Direct binding of representative PSMs to FPR2/ALX was confirmed by three different approaches. (i) PSM peptides C-terminally labeled with the fluorescent dye 5-carboxytetramethylrhodamin revealed only residual binding to untransfected or FPR1-transfected HL60 but bound strongly to FPR2/ALX-transfected cells and neutrophils (Figure 2C, Figure S2). (ii) PSM peptides prevented binding of the FPR2-specific monoclonal antibody #304405 to FPR2-transfected cells (Figure 2D). (iii) PSMα3 displaced the ¹²⁵I-labelled synthetic FPR2/ALX ligand WKYMVm (Elagoz et al., 2004; Christophe et al., 2001) from FPR2/ALX-transfected cells (Figure 2E). From the latter experiment we calculated pIC₅₀ and K_i values for PSMα3 of 7.3 and 13.3 nM, respectively. Taken together, these data confirm that FPR2/ALX is the major receptor on human neutrophils binding and recognizing S. aureus PSM peptides while FPR1 and FPR3 play only minor roles in recognizing PSM peptides.

FPR2/ALX Does not Affect the Ability of PSMs to Lyse Leukocytes

PSM peptides destroy leukocytes at micromolar concentrations, most probably by membrane disruption (Wang et al., 2007). However, it has remained unknown whether PSMs use a receptor on target cells to exert these cytotoxic activities. When release of cytoplasmic lactate dehydrogenase was monitored as a measure for PSM-mediated cytolysis, there were no notable differences in susceptibility of untransfected or FPR1, FPR2/ALX, or FPR3-transfected HL60 (Figure S3). Moreover, FLIPr had no considerable impact on PSM-mediated disruption of human neutrophils, indicating that PSMs lyse leukocytes independently of FPR2/ALX.

PSMs Induce Ca²⁺ Influx in Mouse Leukocytes, which Can be Specifically Blocked by an Inhibitor of Human FPR2/ALX or its Mouse Counterpart(s)

MMK1 and PSMs also induced calcium fluxes in mouse neutrophils (Figure 3A), which is in agreement with our previous finding that expression of PSM peptides contributes significantly to neutrophil recruitment by CA-MRSA in a mouse local infection model (Wang et al., 2007). Thus, mice probably also use FPR-like receptor(s) to sense PSM-like peptide toxins. However, mice have a different set of FPR paralogs with seven to eight members and it has remained unclear how many and which of these represent(s) the mouse ortholog(s) of human FPR2/ALX (Migeotte et al., 2006; Ye et al., 2009). FLIPr did not interfere with PSM-mediated activation of mouse neutrophils (data not shown) and is probably highly human-specific as previously described for CHIPS (Prat et al., 2006; de Haas et al., 2004). However, we noted that the synthetic peptide WRW4, an inhibitor of human FPR2/ALX and of at least one mouse FPR paralog (Fpr-rs2) (Karlsson et al., 2006; Onnheim et al., 2008), had a potent antagonistic effect on mouse neutrophil stimulation by MMK1 and PSMs while fMLF-mediated stimulation was not affected (Figure 3A).

In order to evaluate the impact of FPR-like receptors on PSM-mediated leukocyte influx in vivo, we used two different mouse models. (i) Injection of PSMs-producing S. aureus USA300 into the mouse peritoneum caused a considerable increase in peritoneal neutrophil numbers (Figure 3B). When WRW4 was injected prior to infection the influx was inhibited indicating that the majority of the leukocyte-recruiting activity was dependent on the mouse counterpart(s) of FPR2/ALX. (ii) The capacity of synthetic PSMs to recruit leukocytes in a mouse air pouch model, which allows to monitor a 1.5 to 3-fold increased leukocyte influx upon local injection of chemoattractants into the air pouch (Perretti et al., 2002; Devosse et al., 2009) was evaluated. Injection of PSMα2, which was the most potent PSM peptide toward mouse neutrophils (data not shown), led to strong influx of leukocytes (Figure 3C). However, when WRW4 was injected into the air pouch prior to PSMα2, the influx of leukocytes was blocked. In contrast, a scrambled variant of WRW4 did not inhibit leukocyte influx into air pouches in response to PSMs (Figure 3C). These data confirm that PSMs are potent neutrophil attractants and that FPR2/ALX or its functional mouse ortholog(s) are crucial for neutrophils to migrate toward PSMs in vitro and in vivo.

FPR2/ALX Mediates very Strong Neutrophil Responses to CA-MRSA but not to HA-MRSA

We have recently described that CHIPS can block most of the chemotactic activity of neutrophils toward culture supernatants from the S. aureus laboratory strain RN4220 demonstrating that formylated peptides represent major neutrophil chemoattractants of this hardly pathogenic strain (Dürr et al., 2006). In contrast, we noted that CHIPS had no significant impact on neutrophil calcium fluxes induced by culture supernatants from USA300 or USA400 (Figure 4A), indicating that a receptor other than FPR1 is responsible for neutrophil recognition of CA-MRSA. Notably, CA-MRSA supernatants had about hundred-fold stronger proinflammatory activities compared to the HA-MRSA strains COL, Mu50, and N315, which corresponds to the extraordinary virulence of CA-MRSA (Wang et al., 2007) (Figure 4C). The potent neutrophil activation by CA-MRSA was clearly mediated by PSMs as deletion of the psmα gene cluster encoding the four most potent PSMs that together constitute the vast majority of the proinflammatory PSM activity produced by CA-MRSA (Wang et al., 2007) abrogated most of the neutrophil response (Figure 5A,B). Production of the remaining PSMs in this mutant has been shown to be unaltered (Wang et al., 2007; Queck et al., 2008) but their comparatively low expression and activity levels appears to be the reason for their marginal contribution to the FPR2/ALX-dependent stimulatory capacity of CA-MRSA. Moreover, the strong calcium flux-inducing capacity of CA-MRSA was almost entirely mediated by FPR2/ALX as FLIPr inhibited this activity completely and FPR2/ALX-transfected HL60 responded magnitudes higher than FPR1-transfected HL60 to CA-MRSA (Figure 5A–C). In contrast, the two cell lines responded with similar efficiency to supernatants from the HA-MRSA strains COL, Mu50, and N315, albeit at moderate levels. USA300-induced chemotaxis and IL-8 secretion by neutophils were also largely FPR2/ALX-dependent (Figure 5D). Collectively, these date indicate that FPR2/ALX has a critical, unrecognized role in responding to PSM-like peptide toxins and that leukocytes use PSMs as potent proinflammatory signals from CA-MRSA.

Synthetic Peptides

PSM peptides with the recently published sequences (Wang et al., 2007; Queck et al., 2009), the FPR2/ALX-specific control ligand MMK1 (LESIFRSLLFRVM-NH2) (Fu et al., 2006), and the FPR2/ALX inhibitor WRW4 (WRWWWW-NH2) (Karlsson et al., 2006; Onnheim et al., 2008) were synthesized (EMC Microcollections). PSM-mec was synthesized by American Peptide Company, Inc. (Sunnyvale, CA). PSMα1 (MGIIAGIIKVIKSLIEQFTGK), PSMα2 (MGIIAGIIKFIKGLIEKFTGK), PSMα3 (MEFVAKLFKFFKDLLGKFLGNN), PSMα4 (MAIVGTIIKIIKAIIDIFAK), PSMβ1 (MEGLFNAIKDTVTAAINNDGAKLGTSIVSIVENGVGLLGKLFGF), PSMβ2 (MTGLAEAIANTVQAAQQHDSVKLGTSIVDIANGVGLLGKLFGF), δ-toxin (MAQDIISTISDLVKWIIDTVNKFTKK), and PSM-mec (MDFTGVITSIIDLIKTCIQAFG) were synthesized in their formylated and non-formylated form. Some peptides were C-terminally modified with 5-carboxytetramethylrhodamin. fMLF was purchased from Sigma. Scrambled variant of WRW4 with all-D amino acids (wwrwww) and of alpha PSM with retrograde amino acid sequence (KAFIDIIAKIIKIITGVIAM) were obtained from EMC Microcollections.

Bacteria and Cell Lines

USA300 and USA400 or COL, Mu50, and N315 are prevalent CA-MRSA or HA-MRSA strains, respectively (Diep and Otto, 2008). Defined PSM mutants have recently been described in detail (Wang et al., 2007). Bacterial culture supernatants were obtained by centrifugation of overnight cultures grown in tryptic soy broth (TSB) and filtered through 0.2-µm pore size filters. No chemotactic or stimulatory activity was detected in non-inoculated medium at relevant concentrations.

HL60 cells stably transfected with human FPR1, FPR2/ALX, and FPR3 have been described recently (Christophe et al., 2001; Dahlgren et al., 2000). These cells were grown in RPMI medium (Biochrom) supplemented with 10% FCS (Sigma-Aldrich), 20 mM Hepes (Biochrom), penicillin (100 units/ml), streptomycin (100 µg/ml) (Gibco), and 1 × Glutamax (Gibco). Transfected cells were grown in the presence of G418 (Biochrom) at a final concentration of 1 mg/ml.

Neutrophil Chemotaxis

Human neutrophils were isolated by standard Ficoll/Histopaque gradient centrifugation (Dürr et al., 2006). Chemotaxis of neutrophils towards staphylococcal supernatants or synthetic peptides was determined by using fluorescence-labeled neutrophils that migrated through a 3-µm-pore size polycarbonate trans-well filter as described recently (Dürr et al., 2006). Synthetic chemoattractants were used at concentrations in the linear range of the dose/response curves. Initial control experiments were performed to verify that PSM peptides mobilize neutrophils by chemotactic rather than by undirected chemokinetic stimulation. Neutrophil migration from the upper to the lower trans-well chamber was only observed when the peptides were added to the lower chamber. When an even PSM concentration was present in the two chambers no such migration was observed therby confirming that PSMs induce in fact chemotactic migration in neutrophils. USA300 culture supernatants were used at 1% dilution. The relative fluorescence measured was corrected for the buffer control (only buffer added to lower compartment) and divided through thousand.

Measurement of Calcium Ion Fluxes in Human Neutrophils and HL60 Cells

Calcium fluxes were monitored in many experiments as a surrogate marker for chemotaxis since it can be measured more robustly and accurately than chemotaxis. Calcium fluxes were analyzed by stimulating cells loaded with Fluo-3-AM (Molecular Probes) and monitoring fluorescence with a FACSCalibur flow cytometer (Becton Dickinson) as described recently (de Haas et al., 2004). For EC₅₀ determination, Ca²⁺ flux measurements were carried out at 37°C. In order to study the influence of pertussis toxin (PTX), cells were preincubated with 1 µg/ml PTX (List Biological Laboratories) for three hours at 37°C under 5% CO ₂. For measuring the influence of CHIPS or FLIPr, 1 × 10⁶ cells/ml were preincubated with CHIPS or FLIPr at final concentrations of 1.4 µg/ml or 0.5 µg/ml, respectively, for 20 min at room temperature under agitation. The CHIPS and FLIPr proteins were prepared as described recently (de Haas et al., 2004; Prat et al., 2006). Synthetic chemoattractants were used at concentrations in the linear range of the dose-response curves. In order to stimulate neutrophils, peptides were used at final concentrations of 250 nM, 2 µM for formylated PSMsα3, β1, or 1 µM for δ–toxin and PSM-mec, respectively; 10 µM for non-formylated PSMα3; and 10 nM or 50 nM for fMLF or MMK1, respectively. HL60 cells were stimulated with final peptide concentrations of 62.5 nM (formylated PSMα3), 125 nM (formylated PSMα2), 250 nM (formylated δ–toxin, PSM-mec), 500 nM (formylated PSMsα1, α4, β1, β2), or 2 µM (all non-formylated PSMs). fMLF and MMK1 were used for HL60 cells at final concentrations of 20 nM and 10 nM, respectively. USA300 culture supernatants were used at 0.047% dilution unless otherwise noted. Measurements of 2,000 events were performed and calcium flux was expressed as relative fluorescence corrected for buffer controls.

Peptide Binding and Displacement Assays

10⁵ HL60 cells or 5 × 10⁵ PMN were washed with PBS containing 0.1% HSA and incubated with increasing concentrations of 5-carboxytetramethylrhodamin-labeled peptides for 30 min (neutrophils) or 2 h (HL60) at 37°C under agitation. Cells were washed twice with ice-cold PBS containing 0.1% HSA and cell-associated fluorescence was measured using a FACSCalibur flow cytometer.

For displacement experiments with phycoerythrin-labeled monoclonal mouse anti-human FPR2/ALX antibody Clone #304405 (R and D Systems) FPR2/ALX-transfected HL60 cells were first blocked 15 min at room temperature with 1% HSA in PBS. Then, 1 × 10⁵ cells were washed with PBS, sedimented by centrifugation, resuspended in PBS containing increasing concentrations of PSMα2, PSMα3 or δ-toxin, and incubated for 15 min at 300 rpm and 37°C. Then either anti-FPR2/ALX antibody or an appropriate isotype control (R and D Systems) was added and cells were incubated for further 45 min under agitation on ice. After washing with 200 µl PBS fluorescence was measured using a FACSCalibur flow cytometer. Control experiments confirmed that this antibody does not bind to untransfected or FPR1-transfected HL60 (data not shown).

The radioactively-labeld ligand competition assay was performed by MDS Pharma Services (Taipei, Taiwan) as described recently (Elagoz et al., 2004; Christophe et al., 2001). Briefly, purified membranes of CHO cells transiently transfected with FPR2/ALX were incubated at room temperature for 90 min with 0.025 nM of the synthetic ¹²⁵I-labeled FPR2/ALX ligand WKYMVm in the absence or presence of increasing concentrations of either formylated PSMα3 or of the established FPR2/ALX control ligand peptide CKβ8-1 (Elagoz et al., 2004). Unbound tracer was washed and bound label was counted using a TopCount microplate scintillation and luminescence counter (PerkinElmer Life and Analytical Sciences, Waltham, MA).

Mouse Experiments

Mouse leukocytes were obtained from the peritoneum of 12-week old female CD1 mice (Charles River) upon inducing an intraperitoneal inflammatory reaction by repeated casein injection as described previously (Dürr et al., 2006). Calcium fluxes in mouse neutrophils were determined as described above for human neutrophils with final concentrations of 250, 12.5 and 50 nM of formylated PSMα3, α2, and δ–toxin, respectively, or 500 nM of fMLF and MMK1.

The mouse air pouch model was conducted essentially as described elsewhere (Perretti et al., 2002; Devosse et al., 2009). Briefly, air pouches were raised on the dorsum of 6 to 8-week old female BALB/c mice (Harlan Winkelmann) by injection of 5 ml (day one) and 3 ml (day 4) sterile air. The test peptides were dissolved in DMSO at 100 mM (WRW4 and the scrambled variant wwrw3), and 10 mM (PSMα2) and then diluted with PBS. On day seven 250 µl PBS with 0.27 mM WRW4 or wwrw3 (corresponding to ca. 7.5 mg/kg body weight) or the corresponding amount of DMSO were injected into the air pouch. After 15 min the mice received a second injection into the air pouch of 750 µl PBS with 27 µM PSMα2 or the corresponding amount of DMSO. Four hours later mice were sacrificed, air pouches were rinsed with 2 ml of ice-cold PBS, and leukocytes were counted.

In the mouse peritonitis model WRW4 at 7.5 mg/kg body weight or the vehicle control (2% DMSO) in 300 µl PBS was injected into the peritoneum of 6–8 weeks-old female BALB/c mice. 20 min later 300 µl PBS or PBS containing 2 × 10⁷ live S. aureus USA300 were intraperitoneally injected. Two hours after infection the mice were euthanized with CO₂. Subsequently, peritoneal exudates were collected and leukocytes were stained and counted as described previously (Wang et al., 2007). FITC-conjugated antibody to mouse Gr1⁺ and a corresponding isotype control was purchased from BD Bioscience. Samples were analyzed on a FACSCalibur flow cytometer. Initial control experiments showed that application of WRW4 prior to infection had no impact on numbers of life bacteria after the two-hours infection period.

Animal experiments were performed according to German law with permission of the responsible authorities (Regierungspräsidium Tübingen)

IL-8 Production and Cell Lysis

Human IL-8 was measured using an ELISA kit (R&D Systems) according to the manufacturer’s instructions. Lysis of neutrophils and HL60 cells by PSM peptides was determined by measuring the release of cytoplasmic lactate dehydrogenase (Cytotoxicity Detection Kit, Roche) as described recently (Wang et al., 2007).

Statistical Methods

Statistical analyses were performed with the Prism 4.0 package (GraphPad Software) and the between-group differences were analyzed for significance with the two-tailed Student’s t-test unless otherwise noted.