Signal Inhibitory Receptor on Leukocytes-1 Limits the Formation of Neutrophil Extracellular Traps, but Preserves Intracellular Bacterial Killing

Neutrophils are key effector cells in infection, inflammation, and tissue damage, making up the vast majority of circulating blood cells (1–5). Neutrophils form neutrophil extracellular traps (NETs) in response to microbial invasion, and these are thought to prevent overwhelming infection (6–10). NETs are extracellular structures composed of extruded DNA and decorated with histones and antimicrobial factors (6).

Initially, NETs were described as an antimicrobial strategy, but the tissue damaging potential of NETs has gained salient attention (11), with aberrant NET formation now being suggested to contribute extensively to the pathogenesis of sepsis (12), autoimmunity (13–16), vascular inflammation (17), and thrombosis (18–21). The release of NETs within the circulation, as well as their interaction with platelets and RBCs, has devastating procoagulant and prothrombotic consequences (18, 19, 22–24). Histones directly cause epithelial and endothelial cell death (25, 26), whereas release of NETs exposes self-antigens (27–29), possibly leading to induction and perpetuation of autoimmunity (16, 30). This is most evident in systemic lupus erythematosus (SLE), as immune complexes detected in SLE were found to trigger NET release (25, 31, 32).

In view of a suggested role for NETs in early innate immune defense, it seems dangerous to inhibit NET formation owing to the risk of increased microbial burden in settings of acute or chronic infection (33, 34). Therefore, the ideal NET inhibitor should be a neutrophil-specific, NET-inhibitory agent that preserves neutrophil respiratory burst, phagocytosis, intracellular bacterial killing, and other antimicrobial functions (35).

Signal inhibitory receptor on leukocytes-1 (SIRL-1) is a member of the transmembrane receptor Ig superfamily of immune inhibitory receptors, and it is exclusively expressed on myeloid cells, including neutrophils, eosinophils, and monocytes (36). SIRL-1 contains two canonical ITIMs that are essential for its inhibitory function. Cross-linking of SIRL-1 limits the production of reactive oxygen species (ROS) by neutrophils following isolated cross-linking of FcRs, whereas neutrophil phagocytosis is not affected (37). Because the generation of oxidants has been reported to drive the release of NETs, targeting SIRL-1 represents a promising strategy to arrest NET formation and improve outcomes in settings where NETs cause harm.

Because we previously demonstrated that cross-linking of SIRL-1 suppresses the release of NETs in response to autoantibodies and plasma from SLE patients (38), we now investigated whether vital antimicrobial functions, other than NET formation, remain intact when SIRL-1 is cross-linked on the surface of neutrophils. Our present findings set the stage for SIRL-1 as an ideal therapeutic target to inhibit NET release in NET-mediated diseases.

Histopaque 1119, LPS (Salmonella typhosa), PMA, dichlorofluorescein diacetate (DCF), HRP, gentamicin, and poly-l-lysine were purchased from Sigma-Aldrich. Additional reagents included Ficoll (GE Healthcare), micrococcal nuclease (Worthington Biochemical), DNase I and PicoGreen (Invitrogen), Sytox Green, and Hoechst 33342 (Molecular Probes), diphenylene iodonium (DPI; Sigma-Aldrich), cytochalasin D (cytoD; Sigma-Aldrich), methyl-β-cyclodextrin (Sigma-Aldrich), PP2 (Sigma-Aldrich), wortmannin (Sigma-Aldrich), Ly294002 (Cell Signaling Technology), U0126 (Cell Signaling Technology), piceatannol (Sigma), Bay11-7082 (InvivoGen), celastrol (InvivoGen), and Amplex Red (Molecular Probes). Triclinic monosodium urate (MSU) crystals were synthesized and characterized as previously described by Naccache et al. (39).

Human neutrophils were isolated from sodium-heparin anticoagulated venous blood of healthy adults under protocols approved by the Medical Ethical Committee of the University Medical Center Utrecht. All donors gave informed consent. Neutrophil suspensions were prepared as previously described (38). Unless stated otherwise, freshly purified cells were resuspended in RPMI 1640 (Life Technologies) supplemented with 2% (v/v) heat-inactivated (HI) FCS.

Staphylococcus aureus Wood 46, Staphylococcus epidermidis, Klebsiella pneumoniae, and Salmonella typhimurium (all provided by J. A. van Strijp, Medical Microbiology, University Medical Center Utrecht, Utrecht, the Netherlands) were grown to exponential phase in Todd–Hewitt broth medium with aeration. Bacteria were quantified by measuring A_600nm. Bacteria were washed twice with PBS and opsonized for 30 min with 10% HI human pooled serum. In some experiments, bacteria were heat-killed for 60 min at 70°C before opsonization.

Human neutrophils were incubated with 100 μg/ml MSU crystals at 37°C with gentle agitation for the indicated times. Uptake of crystals was determined by measuring the increase in side scatter by flow cytometry (FACSCalibur; BD Biosciences) and analyzed with FlowJo software (Tree Star, version 10.0.7r2). A threshold value for side scatter was determined in crystal-free samples, and MSU-challenged samples were evaluated for percentage of cells with higher side scatter than the threshold value.

The following stimuli were used at indicated concentrations: PMA at 25 ng/ml, MSU at 100 μg/ml, LPS at 1 μg/ml, opsonized and nonopsonized S. aureus at a multiplicity of infection of 10, and anti-LL37 Abs (Hycult Biotech) at 10 μg/ml. For inhibitor studies, neutrophils were preincubated with 10 μM DPI (an NADPH oxidase [Nox]-2 inhibitor), 10 μg/ml human IgG Fc fragments (Bethyl Laboratories), 20 μM cytoD (inhibitor of actin polymerization), 10 μM PP2 (Src kinase inhibitor), 20 μM wortmannin (PI3K inhibitor), 5 μM U0126 (ERK inhibitor), 20 μM piceatannol (Syk inhibitor), 5 mM methyl-β-cyclodextrin (lipid raft inhibitor), 5 μM Bay11-7082 (NLRP3 inflammasome inhibitor, IκB phosphorylation inhibitor), 5 μM celastrol (a triterpenoid compound), or DMSO (vehicle control) for 30 min before stimulation. Additionally, neutrophils were preincubated with 10 μg/ml isotype-matched control IgG or anti–SIRL-1 mAb 1A5, followed with 20 μg/ml goat anti-mouse F(ab′)₂ fragments.

NET formation by human neutrophils was analyzed by fluorescence microscopy as described previously (38). In short, a total of 0.5 × 10⁶ neutrophils were seeded on coated glass coverslips (0.001% poly-l-lysine) and challenged with the indicated stimuli for 30 and 180 min at 37°C. Cells were stained with Sytox Green (0.5 μM), gently washed, fixed with 4% paraformaldehyde, and stained with Hoechst 33342 (1 μM). Fixed cells were imaged with an Olympus IX71 wide-field inverted microscope with a UPlanSApo ×20/0.75 air objective in Fluoromount-G (SouthernBiotech). Fluorescence was detected with a Photometrics EMCCD 1024 × 1024 pixel camera and softWoRx acquisition software. For the quantification of NET formation, images were processed with ImageJ software (National Institutes of Health) as previously described (38, 40). Briefly, at least four fields of view (each 659 × 659 μm) per condition were captured. Contrast was adjusted to minimize background autofluorescence and a fluorescent threshold was set to result in positive staining only. The same contrast and threshold were applied to all images from all conditions within the experiment. To minimize differences in fluorescence, the same exposure times for excitation filters were applied between experiments. Typical exposure time for Sytox Green fluorescence (490/20, green channel) was 100 ms. Sytox-positive pixel counts were divided by the total number of pixels of thresholded 8-bit images using ImageJ software and expressed as the percentage of image area covered by positive fluorescence staining in each field of view.

For live cell imaging, neutrophils were allowed to settle for 30 min, and 1 × 10⁵ cells were seeded in each well of a black 96-well clear-bottom plate (Costar). Neutrophils were incubated in phenol red-free RPMI 1640 supplemented with 2% HI-FCS and 10 mM HEPES or challenged with indicated stimuli in RPMI 1640 containing 2% HI-FCS and 10 mM HEPES and recorded at 37°C in 5% CO₂/95% air on the BD Pathway 855 bioimaging system with a ×20 objective during a period of 180 min. NETs were detected with a mixture of cell-permeable (Hoechst 33342; 1 μM) and impermeable (Sytox Green; 2 nM) DNA fluorescent dyes. Every 2 min, a set of three images (phase contrast, blue and green fluorescence) was taken with a Hamamatsu Orca high-resolution CCD camera. The system was controlled by the BD AttoVision software (version 1.7/855). Individual frame overlays were prepared with ImageJ software.

To assess the kinetics of NET release, Sytox Green fluorescence was monitored in real time as described previously (41). A total of 1 × 10⁵ neutrophils were resuspended in phenol red-free RPMI 1640 supplemented with 10 mM HEPES, 2% HI-FCS, and 1 μM Sytox Green and seeded into each well of a white 96-well plate. Sytox Green fluorescence (reflecting extracellular DNA) was measured every 5 min for the indicated times in a preheated fluorescence plate reader (Fluoroscan; Thermo Scientific) at 37°C with a filter setting of 480 (excitation)/520 (emission).

NET-DNA in neutrophil supernatants was quantified with a PicoGreen dsDNA detection kit as previously described (42). After stimulation of neutrophils (2 × 10⁵; in phenol red-free RPMI 1640, without FCS), the cells were incubated with micrococcal nuclease (500 mU/ml) for 15 min at room temperature to release NETs formed in response to stimulation. Nuclease activity was stopped with 5 mM EDTA. The supernatant was gently removed after centrifugation at 1200 rpm for 5 min. NET-DNA in cell-free supernatants was quantified with a PicoGreen dsDNA detection kit according to the manufacturer’s instructions. Extracellular DNA was measured in a fluorescence plate reader (Fluoroscan; Thermo Scientific) with a filter setting of 480 (excitation)/520 (emission).

Neutrophils were seeded on glass coverslips coated with 0.001% poly-l-lysine, allowed to settle, and challenged with opsonized S. aureus (multiplicity of infection of 10) or left untreated for 10 min. Neutrophil elastase (NE) was immunostained as described elsewhere (38). Briefly, cells were fixed with 4% PFA, permeabilized with 0.25% Triton X-100 in PBS, blocked (1% BSA and 0.1% Tween 20 in PBS), and incubated overnight with anti-NE Abs (sc-9518, Santa Cruz Biotechnology), which were detected with F(ab′)₂ fragments of DyLight 594–coupled secondary Abs (Jackson ImmunoResearch Laboratories). For detection of DNA, Hoechst 33342 was used. Specimens were mounted in Fluoromount-G and analyzed with a UPlanSApo ×20/0.75 air objective on a wide-field inverted microscope (IX71; Olympus).

Extracellular ROS production was measured in real time by chemifluorescence as previously described (37). Alternatively, real-time intracellular generation of ROS was monitored in a DCF-based assay. Isolated neutrophils were allowed to settle (60 min, 37°C) and 1 × 10⁵ cells were preloaded for 20 min at 37°C with the fluorescent probe DCF (10 μM). After incubation, cells were washed and carefully resuspended in RPMI 1640 supplemented with 2% HI-FCS in white 96-well plates. Fluorescence was measured every 5 min for the indicated times in a preheated fluorimeter at 37°C (Fluoroscan; Thermo Scientific) at 480 (excitation)/520 (emission).

Total killing of S. aureus was performed in the presence of 100 U/ml DNase I to prevent NET-mediated extracellular killing as previously described (37). To inhibit Nox-2–dependent intracellular killing, we incubated selected wells with DPI (10 μM) for 30 min before adding the bacteria.

Phagocytic killing of S. aureus, S. epidermidis, K. pneumoniae, and S. typhimurium by neutrophils was measured in a gentamicin protection assay. Opsonized bacteria were added to the neutrophils at a multiplicity of infection of 10. After 15 min, gentamicin was added to the medium at 100 μg/ml, followed by continued incubation for 20 min. Wells were then washed with PBS, the neutrophils were permeabilized with 0.1% Triton X-100 for 10 min, and bacterial counts were determined as described above.

Statistical analysis was performed with GraphPad Prism software (version 6.0). Data are presented as mean ± SD of independent experiments. A Student t test was used to compare two groups. For comparing more than two groups, nonparametric or parametric one-way ANOVA or Kruskal–Wallis test with Dunn post hoc testing was used where appropriate. A p value of ≤0.05 was considered to be statistically significant.

We have previously shown that ligation of SIRL-1 inhibits NET formation induced by autoantibodies (38). We now addressed whether cross-linking of SIRL-1 altered NET formation in response to other stimuli by direct microscopic observation. Tissue deposition of MSU crystals causes a prevalent sterile inflammatory condition, called gout. Several studies have demonstrated that MSU crystals activate neutrophils to form NETs (43, 44). As previously reported (38), SIRL-1 inhibited NET formation when neutrophils were exposed to anti-LL37 autoantibodies (Fig. 1A), and likewise crystal-induced NET release was suppressed by cross-linking of SIRL-1 (Fig. 1B). As expected, no inhibition by SIRL-1 cross-linking was observed when NETs were induced with the potent protein kinase C activator PMA (Fig. 1B). Also, cross-linking of SIRL-1 did not inhibit the release of NETs when neutrophils were exposed to nonopsonized S. aureus or to the TLR4 agonist LPS (Fig. 1C). In contrast, in response to opsonized S. aureus, cross-linking of SIRL-1 on the surface of neutrophils reduced NET formation (Fig. 1C).

We previously published that SIRL-1 ligation does not affect phagocytic uptake of opsonized bacteria (37). Similarly, cross-linking of SIRL-1 did not affect uptake of inflammatory MSU crystals by neutrophils as evidenced by flow cytometry in the side light–scattering properties (Fig. 1D). These findings suggest that the release of NETs in response to opsonized S. aureus can occur independently of phagocytosis. Indeed, pretreatment with an inhibitor of actin polymerization, cytoD, failed to cause significant inhibition of NET release (Fig. 1E), whereas control experiments confirmed that the uptake of opsonized bacteria is reduced in the presence of cytoD (data not shown). Inhibition of lipid raft formation by methyl-β-cyclodextrin also had no effect on the formation of NETs in response to opsonized S. aureus, further suggesting that NET formation does not require FcR-mediated phagocytosis of opsonized bacteria.

The selectivity for suppression of NET release through SIRL-1 may indicate that alternative cellular processes trigger NET formation, depending on the stimulus.

NETs are extracellular lattices of DNA that can be visually determined (Fig. 1). The presence of extracellular DNA associated with NETs in the culture supernatant of activated neutrophils can be used as an alternative measure for NET formation. In a complementary strategy, we used PicoGreen to measure DNA release and found that treatment of neutrophils with LPS, S. aureus, MSU crystals, and PMA induced significant DNA release after 180 min in a concentration-dependent manner (Fig. 2A).

We assessed the kinetics of NET release by real-time quantification of DNA released in the supernatant with the cell nonpermeable DNA-binding dye Sytox Green (Fig. 2B) as previously described (41). In contrast to PMA-induced NET release, NETs in response to opsonized S. aureus and MSU crystals occurred earlier. Neutrophils form NETs in response to nonopsonized S. aureus, and opsonization of bacteria significantly accelerated S. aureus–induced NET release. Fluorescence microscopy images confirm that neutrophils challenged with S. aureus and MSU release NETs, and that opsonization of S. aureus enhanced NET formation (Fig. 2C). NE is a granular protein that associates with NETs. The presence of extracellular DNA that stains positively for NE is consistent with the process of NET formation. No NE was released when cells where left unstimulated (Fig. 2D). In contrast, after challenge with opsonized S. aureus for 10 min, neutrophils released extracellular DNA where NE colocalizes.

We activated neutrophils and monitored their release of NETs over time by live cell imaging (Fig. 2E). Upon exposure of neutrophils to nonopsonized S. aureus, the cells did not release DNA during the early phase (Fig. 2F). At later time points following stimulation, progressively more cells lost their condensed nuclear material and released NET-DNA, assessed as Sytox Green fluorescence. In contrast, neutrophils exposed to opsonized S. aureus rapidly released a high amount of NETs that increased with time (Fig. 2G). Similarly, inflammatory activation of neutrophils with MSU crystals caused robust NET formation that was detectable early after challenge (Fig. 2H). With similar kinetics as nonopsonized bacteria, stimulation with PMA resulted in an increase of Sytox Green fluorescence during the late phase after challenge. Ultimately, most neutrophils challenged for 180 min with PMA showed decondensed chromatin and formed NETs (Fig. 2I). Taken together, the time course analysis of NET release shows that the formation of NETs induced by S. aureus and PMA follows distinct kinetics, suggesting that distinct forms of NET release are at play.

Opsonized S. aureus and MSU crystals can activate neutrophils and significantly increase their intracellular ROS concentration (39, 45). To detect intracellular ROS produced by Nox-2, we used DCF, a fluorescent indicator of intracellular ROS. Pretreatment of neutrophils with DPI, a flavoprotein inhibitor of Nox-2, before challenge with opsonized S. aureus completely abolished the generation of ROS (Fig. 3A). In contrast, DPI failed to inhibit the release of NETs in response to opsonized S. aureus or MSU crystals, but it completely abrogated PMA-stimulated NET formation as previously reported (42) (Fig. 3B). Additionally, DPI had no effect on NET release induced by nonopsonized bacteria. This is in line with the relative absence of intracellular ROS generated after exposure to nonopsonized S. aureus (data not shown). Therefore, S. aureus triggers the release of NETs in a manner that does not depend on ROS production, distinct from PMA-induced Nox-2–dependent NET formation.

Activation of neutrophils by MSU crystals occurs in part through FcγRIIIB (46). Blocking of FcγRs partially inhibited NET formation when neutrophils were challenged with opsonized bacteria or MSU crystals, but not in response to nonopsonized S. aureus (Fig. 3B), suggesting that ROS-independent NET release to these stimuli involves FcγR-mediated contact. The interaction of MSU crystals with FcγRIIIB is likely to be opportunistic in nature, because opsonization with IgGs is not a prerequisite (46). In fact, it is highly unlikely that FcγRs provide a complete repertoire of the surface molecules with which MSU crystals and opsonized S. aureus interact. Indeed, we obtained only partial inhibition of NET formation in response to MSU crystals and opsonized bacteria with FcγR block.

We studied the contribution of FcR signaling to NET formation by using inhibitors of signaling molecules engaged by the ITAM-coupled FcγRs. Neutrophils were pretreated with inhibitors of Syk (piceatannol), Src (PP2), PI3K (wortmannin), or ERK1/2 (U0126) before exposure to opsonized S. aureus. Treatment of neutrophils with Syk inhibitors only partially inhibited NET formation in response to opsonized bacteria, whereas inhibition of Src, PI3K, and ERK1/2 had no effect (Fig. 4A). In contrast, blocking of Syk and ERK1/2 completely suppressed bacteria-induced ROS production (Fig. 4B). Treatment with inhibitors of Src and PI3K also resulted in diminished generation of ROS. These results suggest that S. aureus–induced receptor/Syk activation contributes to, but is not essential for, NET formation, whereas it is required for the generation of ROS.

The fact that Syk kinase–mediated signaling pathways play a minor role in the release of NETs in response to opsonized bacteria suggests that other intracellular signaling pathways are involved in NET release. Treatment with either Bay11-7082 or celastrol, inhibitors of IκBα phosphorylation and NF-κB, respectively, completely abolished rapid S. aureus–induced NET formation (Fig. 4C), whereas it had no effect on ROS production (Fig. 4D). Taken together, these results clearly highlight that differences exist in the requirement for signaling events involved in ROS production and NET formation after challenge with opsonized S. aureus.

In line with our previous findings (37), we observed an inhibitory effect of SIRL-1 on extracellular ROS when FcγRIIA (CD32) was triggered on neutrophils (Fig. 5A). Because neutrophil ROS production is an essential effector function involved in intracellular bacterial killing, we next aimed to evaluate whether ligation of SIRL-1 affects the generation of ROS in response to other stimuli. Activation of neutrophils through opsonized bacteria and MSU crystals is mediated by FcγRs (46). Pretreatment with human IgG Fc fragments nearly completely abolished the intracellular generation of ROS (Fig. 3A). However, cross-linking of SIRL-1 on the surface of neutrophils had no effect on intracellular levels of ROS in response to opsonized S. aureus (Fig. 5B). Using F(ab′)₂ fragments against SIRL-1, we excluded effects of the Fc part of the cross-linking Ab. Furthermore, ligation of SIRL-1 had no effect on extracellular ROS production by neutrophils exposed to MSU crystals (Fig. 5A). MSU crystals increase neutrophil intracellular ROS concentration in a dose-dependent manner (Fig. 5C). At none of the MSU concentrations tested did SIRL-1 ligation interfere with intracellular generation of ROS (Fig. 5D).

Defects in Nox-2–mediated ROS production enhance intracellular survival of S. aureus (47). Also, S. aureus has been shown to be killed by NETs in vitro (6). Therefore, we determined the effect of cross-linking SIRL-1 on total bacterial killing.

Killing activity of neutrophils was determined on the basis of changes in the number of viable bacteria over time (Supplemental Fig. 1). In the absence of both DNase and DPI, neutrophils can kill S. aureus by phagocytosis and through NET formation. Initially, efficient phagocytic killing was observed, and the presence of DNase had no effect, indicating that no NET-mediated killing occurred at 10 min. In contrast, 30 min after challenge with opsonized S. aureus, when neutrophils start to release NETs (Fig. 1C), phagocytic killing was diminished and most of the antimicrobial activity was mediated through NET formation. By 90 min, most neutrophils form NETs (Fig. 2G) and no changes in the number of viable bacteria were observed in the presence or absence of DPI (Supplemental Fig. 1).

We chose to use a 30-min incubation time in subsequent experiments, because at this time point neutrophils kill both by phagocytosis and through NET formation. Exposure of human neutrophils to opsonized S. aureus for 30 min in the presence of DNase I to dismantle NETs results in significantly increased total bacterial survival, which was not further enhanced by cross-linking of SIRL-1 (Fig. 6A). When human neutrophils were pretreated with DPI and then exposed to S. aureus for 30 min, total bacterial survival was also significantly increased compared with control cells. In this case, ligation of SIRL-1 further enhanced the inhibitory effect of DPI (Fig. 6A). This result is consistent with inhibition of NET formation by SIRL-1 and suggests that SIRL-1 specifically regulates cellular pathways required for extracellular NET-mediated, but not phagocytic, microbial killing. We incubated human neutrophils with various opsonized bacterial strains, including S. aureus, in an in vitro gentamicin protection assay. Treatment with DPI significantly increased intracellular survival of S. aureus (Fig. 6B), indicating that phagocytic bacterial killing depends on activity of Nox-2. In contrast, no difference in intracellular survival was detected with or without ligation of SIRL-1 (Fig. 6C). Thus, engagement of SIRL-1 on neutrophils inhibits extracellular bacterial killing while phagocytic killing is preserved.

Accumulating evidence supports that dysregulated NET formation can cause harm and perpetuate tissue damage in autoimmune disorders and other inflammatory conditions. Strategies that aim to limit the release of NETs are only now beginning to emerge (12, 24, 48). We have previously proposed that targeting immune inhibitory receptors to arrest NET formation could be beneficial in the context of autoimmunity (38). We showed that cross-linking of SIRL-1 suppresses the release of NETs in response to autoantibodies from SLE patients. Given the essential activity of neutrophils in innate immunity and their importance in preventing infections, therapeutic inhibitors of NET release should ideally preserve other neutrophil antimicrobial functions, such as ROS production and intracellular killing (35). In this study, we show that SIRL-1 specifically controls NET formation, without compromising other important neutrophil antimicrobial functions, and advocate for the inhibition of NET release by cross-linking of SIRL-1.

Recently, concern about experimental challenges in studying NET formation has been raised by others (49–51). Detection of NETs by fluorescence microscopy still remains the most informative experimental approach. In the present study, we visualized NETs as extracellular structures released by neutrophils that stain positive for DNA. Furthermore, we confirmed the true nature of observed NETs by costaining the extracellular DNA with the granule protein NE, a specific marker of NET formation. Analysis of data obtained by fluorescence microscopy, however, remains challenging and difficult owing to different ways of expressing the extent of observed NET formation. In this study, we translated the microscopic observations into comparable semiquantitative data with a standardized methodology previously described by others in the field (40, 41). We complemented our findings with additional experiments to assess NET formation, in which we quantify extracellular DNA by staining extracellular DNA with PicoGreen or Sytox Green. Although more quantitative and high throughput in nature, this parameter is less sensitive.

Consistent with our previous report (37), we show that cross-linking of SIRL-1 suppresses FcγRIIa-mediated ROS production in neutrophils (Fig. 5A). In contrast, in the present study, we show that SIRL-1 does not affect intracellular ROS production in neutrophils following challenge with opsonized S. aureus. It has been shown that TLR signaling cooperates with FcRs in the killing of intracellular bacteria by promoting assembly and thus activity of the Nox complex (52). Most likely, signaling through SIRL-1 is not able to suppress synchronized FcR and TLR engagement and the resulting synergistic activation of Nox-2 in neutrophils.

Release of NETs after several hours has often been reported. This cellular process requires the generation of ROS (38, 42, 53). However, ROS-mediated signaling is not the only way that NET release is triggered, as rapid NET formation (within minutes) was described, which is independent of oxidants (41, 54). Our present study shows that S. aureus triggers the formation of NETs through a mechanism that does not depend on ROS. Also, although NETs are formed in response to nonopsonized S. aureus, there is little, if any, generation of intracellular ROS (data not shown), suggesting little interplay between Nox-2 activity and NET formation. Following exposure to opsonized S. aureus, neutrophils produce large amounts of intracellular ROS. Activation of the kinase ERK has been implicated to be involved in ROS-dependent NET release (38, 55). Our data, however, show that inhibition of ERK does not suppress ROS-independent NET formation in response to opsonized S. aureus, whereas the ERK inhibitor U0126 completely abolished ROS production. Similarly, Src kinase and PI3K inhibitors did not block the release of NETs, but they inhibited the generation of ROS after exposure to opsonized S. aureus. Alternatively, NF-κB inhibitors Bay11-7082 and celastrol abrogated S. aureus–induced NET release, whereas ROS production was not affected in the presence of these inhibitors. Thus, distinct signaling events are responsible for the rapid release of NETs in response to S. aureus.

Celastrol and Bay 11-7082 are widely known for their potential to inhibit the transcription factor NF-κB (56, 57). Both compounds, however, have been shown to directly or indirectly modulate numerous cellular targets, including JAK kinase, ERK, and JNK (58, 59). Interestingly, celastrol was recently shown to act on activation of the kinase SYK, a very early signaling event during NET formation in response to serum IgG from SLE and RA patients (60). Therefore, we suggest that celastrol and/or Bay 11-7082 could act on molecular targets upstream of NF-κB, rather than directly modulating NF-κB.

Possible therapeutic approaches to prevent NET formation and its damage to the host are needed. DNase has been used in animal models to remove NETs, and it is given as a therapeutic agent, for instance, in patients with cystic fibrosis. However, DNase might not be effective in removal of the cytotoxic mix of NET components, such as histones and NE. Indeed, others suggest that prevention of NET release might provide more protection against the pathogenicity of NETs than removal of NETs. Anti–Mac-1 blocking Abs and protein arginine deiminase 4 (PAD4) inhibition have been proposed as strategies to arrest NET formation (12, 24). However, blocking Mac-1 is expected to resemble leukocyte adhesion deficiency in situations with a component of infection (such as sepsis). Although neutrophils that lack PAD4 remain capable of killing bacteria by means other than NET formation (12), PAD4 is expressed by many cell types, and the systemic consequences of PAD4 inhibition are not known. Also, the requirement for PAD4 is likely not common to all mechanisms of NET release (61, 62), limiting the spectrum of NET-mediated disorders that could be targeted by PAD4 inhibition.

SIRL-1 is also expressed on the surface of monocytes and eosinophils. Thus, systemic effects of cross-linking SIRL-1 cannot be excluded, and it remains to be determined whether inhibiting NET formation by cross-linking SIRL-1 may improve outcomes in preclinical in vivo model systems. Nonetheless, this study highlights SIRL-1 as a target that is capable of suppressing the formation of NETs in response to autoantibodies, MSU crystals, and bacteria. Importantly, neutrophils retain their intracellular antibacterial activity when SIRL-1 is cross-linked on the surface of the cells. These findings warrant further exploration of SIRL-1 as a therapeutic target in settings where NETs harm.

We thank Prof. Dirk Roos (Sanquin Blood Supply, Amsterdam, the Netherlands) for helpful comments and critical revision of the manuscript, and Prof. Jos A. van Strijp (University Medical Center Utrecht, Utrecht, the Netherlands) for providing valuable reagents.

The authors have no financial conflicts of interest.