Staphylococcus aureus promotes receptor-interacting protein kinase 3- and protease-dependent production of IL-1β in human neutrophils

Abstract

Microbial infection elicits robust immune responses that initially depend on polymorphonuclear neutrophils (PMN), which ingest and kill invading bacteria. However, community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) remain viable within PMN and prompt their lysis with concomitant release of damage-associated molecular patterns and proinflammatory cytokines that promote additional inflammation. Here, we show that ultrapure human PMN (>99.8% pure) that have ingested CA-MRSA released interleukin (IL)-1β but not IL-18. The ingested CA-MRSA needed to be viable, and phagocytosis alone was insufficient to stimulate IL-1β secretion from PMN fed CA-MRSA. In contrast to PMN response to the canonical NLRP3 inflammasome agonist nigericin, IL-1β secretion by PMN fed CA-MRSA occurred independently of NLRP3 inflammasome or caspase-1 activation and required instead active receptor-interacting protein kinase 3 (RIPK3) but not RIPK1. Furthermore, inhibition of neutrophil serine proteases blocked pro-IL-1β cleavage in PMN fed CA-MRSA. Taken together, our data suggest that with respect to secretion of IL-1β and IL-18, PMN differ from human macrophages and exhibit agonist-specific responses. After phagocytosis of CA-MRSA, human PMN secreted IL-1β through a previously unrecognized mechanism dependent on RIPK3 and serine proteases but independent of canonical NLRP3 inflammasome and caspase-1 activation.

1 |. INTRODUCTION

Optimal host defense against infection includes both elimination of invading microbes and restitution of homeostasis.¹ At the onset of most bacterial infections, polymorphonuclear neutrophils (PMN) act as initial cellular responders, recruited early to sites of microbial invasion to ingest and kill bacteria.² Having executed their antimicrobial activities, spent PMN typically become apoptotic and then a target for ingestion by macrophages, which clear dead and dying cells from the site and promote resolution of inflammation (reviewed in Karaji³). In circumstances in which the virulence or number of invading microbes is great or host defense is compromised, homeostasis is not restored and disease occurs (reviewed in Kobayashi⁴).

Infection with community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) exemplifies the clinical consequences of the failure to resolve inflammation. CA-MRSA infections frequently elicit a robust inflammatory response and produce extensive tissue damage and necrosis, findings that suggest dysregulation of host responses that otherwise promote resolution of inflammation.^5,6 Phagocytosed CA-MRSA remain viable within PMN, and PMN that harbor viable CA-MRSA undergo a novel form of programmed cell death and lyse,⁷ thereby releasing damage-associated molecular patterns (DAMPs) that can promote inflammation and tissue necrosis.

Innate immune cells respond to DAMPs and other agonists to release proinflammatory cytokines, including interleukin (IL)-1β and IL-18.⁸ In general, the proteolytic maturation of pro-IL-1β and pro-IL-18 and secretion of the mature cytokines depend on active caspase-1 and an inflammasome, a multicomponent protein complex that cleaves procaspase-1.⁹ Staphylococcus aureus (SA) stimulates IL-1β secretion from macrophages by activation of caspase-1 and the NLRP3 inflammasome, the most extensively studied inflammasome in myeloid cells.^10,11 In contrast to the extensive studies of macrophage responses, the presence and importance of NLRP3 and IL-1β production by human PMN have infrequently been examined.

In this study, we demonstrate that highly purified human PMN fed viable CA-MRSA (PMN-SA) released IL-1β but not IL-18 and that this IL-1β secretion required active receptor-interacting protein kinase 3 (RIPK3) and serine proteases but occurred independently of the NLRP3 inflammasome or caspase-1 as well as PMN lysis. The behavior of PMN-SA contrasts both with IL-1β secretion by human monocyte-derived macrophages (HMDM), which utilize NLRP3 and caspase-1, and with that by nigericin-stimulated human PMN, which rely on NLRP3 and serine proteases. Together, these findings suggest that a previously unrecognized signaling pathway, dependent on RIPK3 and serine proteases but independent of NLRP3 or caspase-1, supports IL-1β secretion from human PMN that have phagocytosed CA-MRSA.

2 |. MATERIAL AND METHODS

2.1 |. Reagents

All reagents were purchased from Fisher Scientific (Pittsburgh, PA) unless otherwise indicated. Heparin was purchased from Fresenius Kabi USA LLC (Lake Zurich, IL), clinical grade dextran T500 from Pharmacosmos (Holbaek, Denmark), Ficoll-Hypaque PLUS from GE Healthcare (Piscataway, NJ). Sterile endotoxin-free water and 0.9% sterile endotoxin-free sodium chloride were obtained from Baxter (Deerfield, IL). Human serum albumin (25%) was purchased from USP (Grifols, Canada), and EasySep human neutrophil enrichment kit from StemCell Technologies (Vancouver, Canada). Tryptic Soy Broth was purchased from Becton Dickson (Sparks, MD). Cytotoxicity Detection Kit PLUS and cDNA Synthesis Kit were obtained from Roche (Indianapolis, IN), GSK’872 from BioVision (Milpitas, CA), and GSK’843 from Glixx Laboratories (Hopkinton, MA). GSK’963 was kindly provided by Glasko-Smith and Kline (Collegeville, PA). IL-1β and IL-18 ELISA kits were purchased from R&D Systems (Minneapolis, MN), N-acetyl-L-tyrosyl-L-valyl-N-[(1S)-1-(carboxymethyl)-3-chloro-2-oxo-propyl]-L-alaninamide (Ac-YVAD-cmk), MCC950, nigericin, and diisopropylfluorophosphate (DFP) from Sigma-Aldrich (St. Louis, MO). Silica was obtained from U.S. Silica (Berkeley Springs, WV). PerfeCTa SYBR Green FastMix from Quantabio (Beverly, MA), RNeasy Mini Kit from Qiagen (Germantown, MD). All primers were synthesized by Integrated DNA Technologies (Coralville, IA). NLRP3 antibody was obtained from AdipoGen (San Diego, CA), anti-human caspase-1 antibody was obtained from Cell Signaling (Danvers, MA), actin antibody from Aviva Systems Biology (San Diego, CA), and anti-human IL-1β from R&D Systems (Minneapolis, MN). Caspase-Glo 1 Inflammasome assay kit and Caspase-8-Glo assay kit were obtained from Promega (Madison, WI). Anti-Fas IgM (human-activating clone CH11) were obtained from MilliporeSigma (Darmstadt, Germany). RPMI 1640 was purchased from Lonza (Hopkinton, MN), and ultrapure lipopolysaccharide (LPS) from Escherichia coli O111:B4 was obtained from InvivoGen (San Diego, CA).

2.2 |. S. aureus culture

Experiments were performed with the USA300 LAC wild-type strain. S. aureus USA300 LAC (referred to throughout as SA) was cultured overnight in tryptic soy broth (TSB) at 37°C with agitation at 180 rpm. Bacteria were diluted to an OD₅₅₀ of 0.05 in TSB and incubated at 37°C with agitation at 180 rpm for 2.5 h until SA reached midlogarithmic growth (OD₅₅₀ between 0.6 and 0.9). Unless otherwise indicated, SA were opsonized in HBSS containing 10% pooled human serum, 20 mM HEPES, and 1% human serum albumin (HSA) for 20 min while tumbling at 37°C.

2.3 |. Human PMN and PBMC isolation

PMN and peripheral blood mononuclear cells (PBMC) were isolated from venous blood collected from healthy volunteers, as previously described by Nauseef.¹² Written consent was obtained from each volunteer in accordance with a protocol approved by the Institutional Review Board for Human Subjects at the University of Iowa. Briefly, heparinized blood was collected, PMN and PBMC were isolated using dextran sedimentation followed by density gradient separation on Ficoll-Paque PLUS. After hypotonic lysis of erythrocytes, ultrapure PMN were isolated by using the EasySep Human neutrophil enrichment kit based on negative selection. The purity of ultrapure PMN was routinely > 99.8% as defined by flow cytometry detects CD15⁺/CD16⁺ cells. PMN were resuspended in HBSS with divalent cations containing 20 mM HEPES and 1% HSA and adjusted to 20 × 10⁶ cells/ml.

For differentiation of monocytes into macrophages, PBMC (2 × 10⁶ cells/ml) in RPMI 1640 with 20% autologous serum, were incubated in Teflon jars for 6 days at 37°C, 5% CO₂. HMDM (0.5 × 10⁶ cells/ml) adhered to plastic wells overnight in RPMI 1640 containing 10% pooled human serum, and nonadherent cells were washed away with warm RPMI 1640 the following day. Replicate experiments utilized PMN and PBMC from different donors.

2.4 |. Cytokine analysis

Ultrapure PMN (10 × 10⁶ cells) or HMDM (0.5 × 10⁶ cells) were tumbled in buffer alone or primed with 10 ng/ml LPS for 4 h, and pretreated with buffer or inhibitors (50 μM GSK’872; 50 μM GSK’843; 100 nM GSK’963; 10 μM Ac-YVAD-cmk;1 μM MCC950) 30 min before stimulation with nigericin or SA). In experiments assessing the role of neutrophil serine proteases, PMN were incubated with 1 mM DFP for 20 min on ice, washed and resuspended in HBSS with divalent cations containing 20 mM HEPES and 1% HSA followed with LPS priming. Subsequently, PMN or HMDM were stimulated with 5 μM nigericin for 2 h, with 500 μg/ml silica for 15 h, or fed SA (5:1 MOI) for 10 min. After 10 min of phagocytosis, uningested SA were removed by centrifugation at 300 × g for 5 min and PMN that harbored viable SA (PMN-SA) were resuspended in HBSS containing 20 mM HEPES and 1% HSA and incubated for 2 h, after which the PMN-SA suspensions were centrifuged (300 × g, 5 min). Supernatants were collected, pellets were lysed in RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.1% sodium deoxycholate, 0.1% SDS, 1 mM PMSF), and IL-1β and IL-18 were measured by ELISA.

2.5 |. Cytotoxicity assay

Ultrapure PMN (10 × 10⁶ cells) were in buffer alone or primed with 10 ng/ml LPS for 4 h, and stimulated with 5 μM nigericin for 2 h or fed SA (5:1 MOI) for 10 min. After 10 min of phagocytosis, uningested SA were removed by centrifugation at 300 × g for 5 min and PMN-SA were resuspended in HBSS containing 20 mM HEPES and 1% HSA and incubated for 2 h, after which the PMN-SA suspensions were centrifuged (300 × g, 5 min). Supernatants were collected and assayed for lactate dehydrogenase (LDH). LDH was measured using the Cytotoxicity Detection Kit PLUS according the manufacturer’s instructions.

2.6 |. qPCR analysis

Ultrapure PMN (10 × 10⁶ cells) were in buffer alone or primed with 10 ng/ml LPS for 4 h with tumbling. Subsequently, PMN were stimulated with 5 μM nigericin for 2 h or fed SA or heat-killed SA (HK SA) (5:1 MOI). After 10 min of phagocytosis, uningested SA were removed by centrifugation at 300 × g for 5 min and PMN-SA were resuspended in HBSS containing 20 mM HEPES and 1% HSA and incubated for 2 h. PMN-SA were spun at 300 × g for 5 min and pellets were lysed. RNA was extracted using RNeasy Mini Kit. Total RNA (250 ng) was used for cDNA synthesis with Transcriptor First Strand cDNA Synthesis Kit according to the manufacturer’s protocol. Real-time quantitative polymerase chain reactions (RT-qPCR) were performed using PerfeCTa SYBR Green FastMix and Mastercycler ep realplex (Eppendorf, Germany). The cycling conditions used were: an initial denaturation step at 95°C for 30 s, followed by 40 cycles of denaturation at 95°C for 10 s, and annealing/extension at 55°C for 30 s. The melting curve analysis was set as a final step of the qPCR program by gradually heating the qPCR product from 55 to 95°C.

The following primers were used: IL-1β forward, 5′-CAGCCAATCTTCATTGCTCAA-3′ and reverse, 5′-GAACAAGTCAT CCTCATTGCC-3′; IL-18 forward, 5′-GTCGCAGGAATAAAGATGG CTG-3′ and reverse, 5′-CCAAAGTAATCTGATTCCAGGTTT-3′; NLRP3 forward, 5′-CAGACTTCTGTGTGTGGGACTGA-3′ and reverse, 5′-TCCTGACAACATGCTGATGTGA-3′. Data were normalized to GAPDH and the gene expression levels (fold difference) were presented as 2^−ΔΔct. Control PMN (t = 0 h) mRNA level was set to 1.0, and mRNA from unstimulated or stimulated PMN are shown in relation.

2.7 |. Immunoblotting

PMN or HMDM were in buffer alone or treated with 10 ng/ml LPS for 4 h. The caspase-1 inhibitor Ac-YVAD-fmk (10 μM) was added 30 min before treatment with 5 μM nigericin. PMN were fed SA (5:1 MOI) for 10 min, after which uningested SA were washed away by centrifugation at 300 × g for 5 min. PMN-SA were resuspended in HBSS containing 20 mM HEPES and 1% HSA. PMN-SA, PMN with nigericin or HMDM with nigericin were incubated for 2 h at 37°C. PMN were collected by centrifugation and suspended in lysis buffer (50 mM Tris-HCl, pH 8, 5 mM EDTA, 150 mM NaCl, 1% Triton X-100, proteases inhibitor cocktail) on ice for 10 min. Lysis buffer was added to each well of HMDM, cells were scratched, transferred into micro tubes, and incubated for 10 min on ice. Lysates were spun at 20,000 × g for 5 min to remove insoluble material. Total protein (20 μg) was separated by poly-acrylamide gel electrophoresis and transferred to PVDF membrane, which were incubated with specific antibodies. Anti NLRP3 and anti-human caspase-1 were diluted 1:1000, anti-human IL-1β was used at a final concentration 1 μg/ml. Horseradish peroxidase-conjugated secondary anti-mouse or anti-rabbit antibodies were diluted 1:15,000.

2.8 |. Caspase activity assay

PMN tumbled in buffer alone or primed with 10 ng/ml LPS for 4 h and then fed SA (5:1 MOI) for 10 min. After phagocytosis, uningested SA were washed away. PMN-SA or PMN with 5 μM nigericin were incubated with tumbling for 2 h at 37°C. Activities of caspase-1 and caspase-8 were measured using Caspase-Glo 1 Inflammasome Assay Kit or Caspase-8 Assay Kit, respectively, according to the manufacturer’s instructions. Enzyme activity was assessed by quantifying chemiluminescence generated upon cleavage of luminogenic caspase-1-or caspase-8-specific substrate. Briefly, 1 × 10⁶ cells (for caspase-1) or 5 × 10⁴ cells (for caspase-8) were transferred into a white-walled 96-well microplate in triplicate. Equal volumes of the Caspase-Glo reagent were added to the wells, mixed thoroughly, and incubated with shaking at room temperature for 1 h. Luminescence was detected using a LUMIstar Galaxy luminometer (BMG Labtech, Cary, NC). Caspase activity was expressed as relative light units (RLU).

In experiments assessing pharmacologic inhibition of caspase-1 activity, PMN or HMDM were in buffer alone or primed with 10 ng/ml LPS for 4 h and then incubated in vehicle or 10 μM Ac-YVAD-fmk for 30 min before being fed SA (5:1 MOI) for 10 min or 5μM nigericin for 2 h. Subsequent processing was as described above.

2.9 |. Statistical analysis

Results are expressed as the mean ± SEM of at least three independent experiments. Paired Student’s t-test was used for comparison between two groups, and one-way ANOVA followed by Dunnett’s or Tukey’s post hoc tests were used for multiple group comparison. All analyses were performed using GraphPad Prism software. P < 0.05 was regarded as statistically significant.

3 |. RESULT S

3.1 |. PMN-SA secrete IL-1β

As previously reported by Bakele et al.,¹³ the prototypical NLRP3 agonist nigericin stimulated secretion of the proinflammatory cytokine IL-1β by LPS-primed ultrapure human PMN (Fig. 1A). Because we were interested in the behavior of PMN that harbored viable SA within their phagosomes (referred to as PMN-SA), we fed LPS-primed ultrapure PMN SA for 10 min, after which extracellular SA were removed and PMN-SA incubated for 2 h. PMN-SA released IL-1β (Fig. 1A), whereas LPS-primed PMN fed opsonized heat-killed SA or opsonized zymosan (OpZ) did not release IL-1β in excess of that by unstimulated PMN (Fig. 1B). The release of extracellular IL-1β was not a result of a loss of PMN integrity, since what relatively little lysis occurred was to the same extent without or with stimulation (Fig. 1C).

FIGURE 1. PMN-SA secrete IL-1β.

(A) Ultrapure PMN were left in buffer or treated with 10 ng/ml LPS for 4 h prior to stimulation with 5 μM nigericin for 2 h or fed SA (MOI 5:1) for 10 min. After phagocytosis, uningested SA were removed, and PMN harboring viable SA (PMN-SA) were incubated for 2 h. Unstimulated PMN, nigericin-stimulated PMN, or PMN-SA were spun down, supernatants were collected and ELISA for IL-1β was performed. Bars represent the average from three independent experiments ± SEM. P-values were determined using a one-way ANOVA and Dunnett posttest (***P < 0.001 vs. untreated PMN). (B) Ultrapure PMN were primed with 10 ng/ml LPS and incubated for 4 h prior to stimulation with opsonized SA (MOI 5:1), opsonized heat-killed SA (HK SA) (MOI 5:1) or opsonized zymosan (OpZ) for 10 min. After phagocytosis, uningested SA or OpZ were removed and PMN-SA, PMN-HK SA or PMN-OpZ incubated for 2 h. Supernatants were collected and ELISA for IL-1β was performed. Bars represent the average from three independent experiments ± SEM. P-values were determined using a one-way ANOVA and Dunnett posttest (*P < 0.05 vs. PMN + LPS + SA). (C) Ultrapure PMN were incubated in buffer or primed with 10 ng/ml LPS for 4 h and stimulated with 5 μM nigericin for 2 h or fed SA (MOI 5:1) for 10 min. After phagocytosis, extracelular SA were removed and PMN-SA were incubated for 2 h. Supernatants were collected and LDH assay performed. Bars represent the average from three independent experiments ± SEM. (D) Ultrapure PMN were left in buffer or primed with 10 ng/ml LPS for 4 h and then fed with SA for 10 min. After phagocytosis, uningested SA were removed, and PMN-SA were incubated for 2 h. Samples were spun down, supernatants collected, and pellets lysed with RIPA buffer as described in section “Methods.” IL-1β ELISA was performed. Bars represent the average from three independent experiments ± SEM. P-values were determined using a pair t-test (**P < 0.01 supernatant vs. lysate)

In addition to measuring secreted IL-1β, we subjected lysates of ultrapure PMN to ELISA for IL-1β, thereby quantifying intracellular pro-IL-1β and IL-1β. Although there was no detectable IL-1β-related species in freshly isolated ultrapure PMN (Fig. 1D), simply tumbling PMN at 37°C for 6 h increased pro-IL-1β intracellularly. The intracellular levels of pro-IL-1β did not change with the addition of LPS during the incubation before exposure to SA and secretion of IL-1β required stimulation with SA.

Taken together, these data demonstrate that: (1) ultrapure PMN lacked intracellular stores of pro-IL-1β or IL-1β; (2) tumbling of PMN suspensions at 37°C was sufficient stimulation to promote pro-IL-1β production but with little IL-1β secretion; and (3) IL-1β secretion by PMN-SA required viable SA, was not a generic response to phagocytosis alone, and was not consequence of PMN-SA lysis.

3.2 |. Neutrophils fed S. aureus do not secrete IL-18

Because inflammasome activation typically controls secretion of both IL-1β and IL-18,¹⁴ we assessed IL-18 release from PMN-SA. As expected for an inflammasome-mediated response, nigericin stimulated IL-18 secretion from ultrapure PMN (Fig. 2A). However, PMN-SA did not release IL-18 above levels constitutively secreted by unstimulated PMN (Fig. 2A). Moreover, PMN fed SA did not upregulate IL-18 gene expression in comparison to the level in unstimulated PMN at time zero (Fig. 2B). In contrast, PMN primed with LPS, without or with subsequent stimulation with SA, increased IL-1β mRNA more than 10-fold compared to expression by unstimulated PMN at time zero (Fig. 2B). Given the contrast between PMN-SA responses and those of nigericin-stimulated PMN, which were consistent with activation of the NLRP3 inflammasome,¹³ we examined the role of inflammasome activation in IL-1β secretion by PMN-SA.

FIGURE 2. PMN-SA do not secrete IL-18.

Ultrapure PMN were left in buffer or treated with 10 ng/ml LPS for 4 h prior to stimulation. PMN were then incubated with 5 μM nigericin for 2 h or fed SA (MOI 5:1) for 10 min. After phagocytosis of SA, uningested SA were removed, and PMN-SA incubated for 2 h. (A) Supernatants were collected and ELISA for IL-18 was performed. Bars represent the average from three independent experiments ± SEM. P-values were determined using a one-way ANOVA and Dunnett posttest (**P < 0.01 vs. untreated). (B) Pellets were lysed, RNA isolated, and IL-1β or IL-18 gene expression quantitated. Data were normalized to GAPDH expression and presented as 2^−ΔΔct. Freshly isolated PMN (t = 0 h) mRNA level was set to 1.0, and mRNA from unstimulated or stimulated LPS-primed PMN are shown in relation. Data are presented as mean ± SEM of three independent experiments. P-values were determined using a paired t-test (*P < 0.05 vs. untreated PMN)

3.3 |. IL-1β secretion is not dependent on inflammasome activation

PMN responses to nigericin indicated that ultrapure human PMN possess a functioning inflammasome (Fig. 1 and Bakele et al.¹³). However, the observed discordance between secretion of IL-1β and IL-18 by PMN-SA suggested that PMN that had previously ingested SA released IL-1β independent of inflammasome activation. We focused our attention first on the NLRP3 inflammasome, the most extensively characterized inflammasome in the context of IL-1β secretion by myeloid cells.¹⁵

MCC950, a small molecule inhibitor that blocks assembly of the NLRP3 inflammasome, selectively inhibits activation of NLRP3 without affecting AIM2, NLRC4, or NALP1 inflammasomes.¹⁶ As expected for an NLRP3-dependent response, MCC950 blocked IL-1β secretion by LPS-primed HMDM stimulated with nigericin (Fig. 3A). Although to a lesser extent than the effect on HMDM, MCC950 inhibited IL-1β secretion by PMN stimulated with nigericin (Fig. 3B), consistent with our evidence that the response of human PMN to nigericin was in part dependent on the NLRP3 inflammasome. In contrast to its effects on responses of HMDM or PMN to nigericin, MCC950 did not alter PMN-SA secretion of IL-1β (Fig. 3B). In agreement with an earlier study of inflammasome activation in human PMN,¹³ we did not identify NLRP3 protein in ultrapure PMN, although HMDM expressed detectable NLRP3 protein (Fig. 3C). However, mRNA for NLRP3 increased in LPS-primed PMN stimulated with nigericin but not in PMN-SA (Fig. 3D). However, we detected a modest (2-to 3-fold) increase in mRNA for NLRP3 under all experimental conditions, when compared to levels expressed in freshly isolated PMN at time zero (Fig. 3D). Taken together, these data link the NLRP3 inflammasome to IL-1β production by human PMN stimulated with nigericin and suggest that an inflammasome-independent mechanism may operate in PMN-SA.

FIGURE 3. NLRP3 is not involved in IL-1β secretion by PMN-SA.

HMDM (A and C) or ultrapure PMN (B–D) were treated with buffer or primed with 10 ng/ml LPS for 4 h. Thirty minutes prior to stimulation, cells were incubated in buffer or in 1 μM MCC950. Subsequently, LPS-primed PMN or HMDM were stimulated with 5 μM nigericin or fed SA (PMN-SA) or heat-killed SA (HK SA) and incubated for 2 h as described in section “Methods.” (A and B) Supernatants were collected and IL-1β ELISA performed. Bars represent the average from at least three independent experiments ± SEM. P-values were determined using a paired t-tests (for comparison effect of MCC950 in HMDM + LPS + nigericin samples; or PMN + LPS + nigericin samples; or PMN + LPS + SA samples (*P < 0.05; **P < 0.01). (C) Lysates from untreated and treated PMN or HMDM were analyzed by immunobloting for NLRP3 and actin. Shown is a representative of three separate experiments. (D) Pellets were lysed, RNA isolated and reverse transcribed into cDNA. RT-qPCR was performed. Data were normalized to GAPDH and presented as 2^−ΔΔct. The mRNA level for freshly isolated PMN (t = 0) was set to 1.0, and relative mRNA expression of stimulated PMN is shown. Data are presented as mean ± SEM of three independent experiments

Typically, caspase-1 mediates the proteolytic processing of pro-IL-1β to mature IL-1β.¹⁷ As anticipated for a NLRP3-and caspase-1-dependent response, the caspase-1-specific inhibitor Ac-YVAD blocked secretion of IL-1β (Fig. 4A) and caspase-1 activation (Fig. 4B) by HMDM stimulated with nigericin. Furthermore, Ac-YVAD decreased IL-1β release by nigericin-treated PMN mirrored the inhibition by MCC950, although the extent of the inhibition was not statistically significant. In contrast, Ac-YVAD had a negligible effect on IL-1β secretion by PMN-SA (Fig. 4C). Because human PMN express too little caspase-1 to be detected in immunoblots (Supplemental Fig. 1), we employed a sensitive chemiluminescence-based activity assay to probe PMN for active caspase-1. Whereas the assay detected caspase-1 activity in PMN, increases in nigericin-stimulated PMN or PMN-SA were small and not statistically different from the baseline activity, which contrasted with the Ac-YVAD inhibitable, robust caspase-1 activity seen in nigericin-stimulated HMDM (Fig. 4D). Taken together, these data demonstrate that PMN-SA generated and secreted mature IL-1β through pathways independent of NLRP3 or caspase-1.

FIGURE 4. Caspase-1 is not activated during IL-1β processing from stimulated PMN.

HMDM (A, B, and D) or ultrapure PMN (C and D) were treated with buffer or primed with 10 ng/ml LPS for 4 h. Thirty minutes prior stimulation, PMN or HMDM were incubated with 10 μM Ac-YVAD. Subsequently, cells were stimulated with 5 μM nigericin and incubated for 2 h or fed SA (MOI 5:1) for 10 min. After phagocytosis, uningested SA were removed and PMN-SA were incubated for 2 h as described in section “Methods.” (A and C) Supernatants were collected and IL-1β ELISA performed. Bars represent the average from at least three independent experiments ± SEM. P-values were determined using a paired t-tests (for comparison effect of Ac-YVAD in HMDM + LPS + nigericin samples; or PMN + LPS + nigericin samples; or PMN + LPS + SA samples (*P < 0.05). (B) Lysates from untreated and treated HMDM were analyzed by immunoblot for caspase-1 and actin. Shown is a representative of three separate experiments. (D) Cells suspensions were collected and caspase-1 activity detected using Caspase-Glo reagent and expressed as relative light units (RLU). Bars represent the average from three independent experiments ± SEM

3.4 |. Neutrophil serine proteases process pro-IL-1β in human PMN

In cells rich in proteases, such as PMN and mast cells, serine proteases rather than caspases cleave pro-IL-1β into its mature and active product.¹⁸ The azurophilic granules of human PMN contain four serine proteases at high concentration (reviewed in Kettritz¹⁹), thus providing a ready source for pro-IL-1β processing activity. Treatment of PMN with the irreversible serine protease inhibitor DFP blocked subsequent secretion of IL-1β from SA-or nigericin-stimulated PMN (Fig. 5A and B). In contrast, DFP treatment did not alter IL-1β secretion by stimulated HMDM (Fig. 5C), consistent with processing of IL-1β in these cells being mediated by caspase-1. In contrast to the effect of DFP on IL-1β secretion, serine protease inhibition did not block IL-18 secretion from PMN-SA (Supplemental Fig. 2). These data suggest that neutrophil serine proteases, rather than caspase-1, mediated pro-IL-1β processing in PMN-SA.

FIGURE 5. Inhibition of serine proteases blocks IL-1β secretion from PMN-SA but not from stimulated HMDM.

Ultrapure PMN (A and B) or HMDM (C) were pre-treated in buffer or with 1 mM DFP, primed with 10 ng/ml LPS for 4 h and (A) PMN fed SA (MOI 5:1) for 10 min. After phagocytosis, extracelular SA were removed and PMN-SA were incubated for 2 h. (B) PMN were incubated with nigericin for 2 h. (C) HMDM were stimulated with 500 μg/ml silica for 15 h as described in section “Methods.” Supernatants were collected and ELISA for IL-1β was performed. Bars represent the average from at least three independent experiments ± SEM. P-values were determined using a paired t-test for comparison effect of DFP in stimulated PMN or in stimulated HMDM (**P < 0.01)

3.5 |. RIPK3 is involved in IL-1β secretion

In contrast to macrophages, dendritic cells utilize an atypical pathway for processing pro-IL-1β that depends on active RIPK3.²⁰ In light of our identification of features of IL-1β secretion by PMN-SA that were atypical and contrasted with what appeared to be NLRP3-dependent secretion by nigericin-stimulated PMN or by stimulated HMDM, we examined the effects of RIPK3 inhibition on PMN-SA release of IL-1β. Specific RIPK3 inhibitors, GSK’872 or GSK’843, significantly reduced IL-1β secretion (Fig. 6A) but not IL-18 release (Fig. 6B) from PMN-SA. In contrast, RIPK3 inhibitors had no effect on IL-1β secretion by nigericin-stimulated PMN (Fig. 6C). Neither inhibitor affected phagocytosis of SA by PMN (data not shown). Inhibition of RIPK1, a kinase essential for the activation of RIPK3 in the necroptosis pathway,²¹ did not alter IL-1β secretion by PMN-SA (Supplemental Fig. 3).

FIGURE 6. Secretion of IL-1β from PMN-SA requires active RIPK3.

Ultrapure PMN were incubated with buffer or with 10 ng/ml LPS for 4 h. Thirty minutes prior stimulation, PMN were pre-treated with 50 μM GSK’872 or GSK’843 and subsequently stimulated with SA (MOI 5:1) for 10 min. After phagocytosis, uningested SA were removed and PMN-SA were incubated for 2 h, or PMN were subsequently stimulated with 5 μM nigericin for 2 h as described in section “Methods.” Supernatants were collected and ELISAs for IL-1β (A and C) or IL-18 (B) were performed. Bars represent the average from at least three independent experiments ± SEM. P-values were determined using a one-way ANOVA and Tukey posttest (*P < 0.05; or **P < 0.01; or ***P < 0.001 vs. PMN-SA)

Collectively, these data suggest that IL-1β secretion by ultrapure human PMN that harbor SA in their phagosomes required enzymatically active RIPK3 and serine proteases but was independent of NLRP3 and caspase-1 and thereby distinct from mechanisms employed by HMDM or nigericin-stimulated PMN.

4 |. DISCUSSION

Early responders critical to effective host defense against bacteria, PMN migrate to sites of infection, where they eliminate invading microbes and produce a wide array of cytokines that under optimal conditions orchestrate both the prompt initiation and the timely resolution of the inflammatory response.^1,22 Imbalance in coordinated secretion of pro-and anti-inflammatory cytokines derails reestablishment of tissue homeostasis and instead can promote tissue damage and development of chronic inflammatory disorders.²³ Despite the potent PMN antimicrobial system, ingested CA-MRSA remain viable within PMN phagosomes. The persistence of viable intracellular CA-MRSA leads to lysis of PMN and release of intracellular contents,^7,24 which can serve as DAMPs that stimulate secretion of proinflammatory cytokines. We reasoned that even without lysis, the release of proinflammatory cytokines such as IL-1β from PMN harboring viable SA may contribute to the exuberant inflammation and tissue damage seen clinically.

Our experimental system to study IL-1β secretion by human PMN has several unique features. First, we used only highly purified human PMN (>99.8% purity), thereby avoiding any potential contribution of cytokines from other cells that typically contaminate preparations that rely solely on centrifugation through density gradients to isolate PMN.²⁵ Second, we focused our attention on a unique context for PMN response, namely after phagocytosis. We removed uningested SA after phagocytosis was completed and observed the behavior of PMN laden with viable SA. The few previous studies of ultrapure PMN secretion of IL-1β used only conventional agonists to activate inflammasomes; for example, the bacterial toxin nigericin or ATP.¹³

Typically, multicomponent protein complexes called inflammasomes drive canonical production of the proinflammatory IL-1 family of cytokines, including IL-1β and IL-18.⁹ We confirmed previous reports that nigericin-stimulated human PMN secrete both IL-1β and IL-18. In contrast to PMN responses to nigericin, PMN-SA secretion of IL-1β occurred without concomitant IL-18 release, an unexpected observation. Once assembled and activated, NLRP3, the most extensively studied inflammasome in myeloid cells,¹¹ processes procaspase-1 to active caspase-1, the enzymatically active form that in turn cleaves both pro-IL-1β and pro-IL-18 to yield the mature species secreted extracellularly.^14,17 Given that prior work by Bakele et al.¹³ and our data implicate NLRP3 in IL-1β and IL-18 release by nigericin-stimulated PMN, we probed in greater detail additional determinants of IL-1β secretion from PMN-SA. However, our data do not demonstrate that PMN serine proteases directly process pro-IL-1β. Serine proteases have been implicated in signaling activities in a variety of biological setting (reviewed by Meyer-Hoffert and Wiedow²⁶), and it is possible that proteolytic maturation of pro-IL-1β requires a mediator whose action is serine protease-dependent. In that case, DFP would exact its inhibitory effect upstream of pro-IL-1β processing.

Because phagocytosis of staphylococci by macrophages causes assembly of the NLRP3 inflammasome and activation of caspase-1,²⁷ we asked if IL-1β secretion by PMN fed SA was also dependent on NLRP3 and caspase-1. We used the selective NLRP3 inhibitor, MCC950, which blocks canonical and noncanonical NLRP3 activation at nanomolar concentrations.¹⁶ Consistent with previous studies that demonstrated that MCC950 inhibits IL-1β secretion by stimulated murine or human macrophages,^16,28,29 we observed that MCC950 reduced IL-1β secretion by nigericin-stimulated HMDM (Fig. 3A). MCC950 partially inhibited IL-1β secretion by nigericin-stimulated ultrapure PMN, suggesting that nigericin-stimulated IL-1β release from human PMN depended in part on NLRP3. We detected upregulation of mRNA for NLRP3 in LPS-primed, nigericin-stimulated PMN, relative to levels in freshly isolated PMN, but did not detect NLRP3 protein (Fig. 3C and D), confirming observations by Bakele et al.¹³ and suggesting that the concentration of NLRP3 protein in lysates of human PMN was insufficient to be detected by immunoblot. In contrast to its effect on responses of PMN to nigericin, MCC950 did not alter IL-1β secretion by PMN-SA, thereby suggesting that IL-1β release from PMN-SA occurred independently of the NLRP3 inflammasome (Fig. 3).

Because the assembled inflammasome generates caspase-1 that cleaves pro-IL-1β into active IL-1β, as was observed in macrophages stimulated with nigericin (Laliberte et al.³⁰ and Fig. 4A and B), we asked if caspase-1 could be cleaved and promote IL-1β secretion without inflammasome activation. The specific caspase-1 inhibitor Ac-YVAD did not block IL-1β secretion by PMN-SA (Fig. 4C), thus supporting the notion that IL-1β secretion by human PMN fed SA was independent of caspase-1 activation. Proteolytic cleavage of pro-IL-1β can be mediated by serine proteases as well as by caspase-1.^13,31,32 Because our data showed that IL-1β secretion by PMN stimulated with nigericin or SA was independent of caspase-1, and because the azurophilic granules of human PMN are rich in serine proteases, we hypothesized that one or more PMN serine proteases may play role in IL-1β processing by PMN-SA. The irreversible serine protease inhibitor DFP blocked IL-1β secretion by nigericin-or SA-stimulated PMN but not by stimulated HMDM (Fig. 5), suggesting that human PMN relied on a serine protease for pro-IL-1β cleavage rather than on caspase-1, as employed in HMDM.

The role of RIPK3 in IL-1β secretion by human PMN has not been previously examined, and we are the first to demonstrate that inhibition of RIPK3 blocked IL-1β secretion by human PMN-SA. Recent reports demonstrate that RIPK3, a protein typically associated with programmed cell death and necrosis,³³ participates in cell responses other than necroptosis³⁴ and supports IL-1β secretion by murine dendritic cells and macrophages.^35–37 In contrast, Moriwaki et al.20 showed that inhibition of RIPK3 increases IL-1β secretion and caspase-8 activation in LPS-stimulated murine dendritic cells. These responses seem to be species-or cell-specific, because we did not observe enhanced caspase-8 activity in PMN-SA compare to unstimulated PMN (Supplemental Fig. 4). In contrast to the response of PMN-SA, IL-1β secretion by nigericin-stimulated PMN did not require active RIPK3, underscoring the different mechanisms that support human PMN responses to SA and those to nigericin. Furthermore, inhibition of RIPK3 had no effect on constitutive or nigericin-stimulated IL-18 secretion from human PMN (Fig. 6).

Taken together, our data demonstrate that human PMN differ from HMDM with respect to IL-1β production and, like many other myeloid cells, possess agonist-specific mechanisms to support processing of pro-IL-1β and secretion of IL-1β (Table 1). Unlike HMDM, which employ the canonical NLRP3-caspase-1 cascade to process and secrete IL-1β and IL-18, human PMN utilize at least two different systems to generate IL-1β. IL-1β and IL-18 secretion by nigericin-stimulated PMN relied in part on NLRP3 but also depended on a neutrophil serine protease rather than caspase-1 for proteolytic processing of the precursor forms of the cytokines. In contrast to both HMDM and nigericin-stimulated human PMN, PMN-SA released IL-1β without concomitant secretion of IL-18 above constitutive levels through a mechanism that required enzymatically active RIPK3 and a neutrophil serine protease. How active RIPK3 contributes to IL-1β production is unknown. Previous studies linking RIPK3 and IL-1β secretion demonstrate that kinase activity is not required and that RIPK3 serves as a scaffold to associate with caspase-8, which in turn drives IL-1β processing. For PMN-SA, IL-1β secretion required active RIPK3, was independent of caspase-8 activity (Supplemental Fig. 4), and did not require RIPK1 (Supplemental Fig. 3), the kinase upstream of RIPK3 and essential for activation of the necroptosis pathway.²¹ The substrate for active RIPK3 that drives IL-1β generation in PMN-SA is unknown. Whereas the best characterized target for phosphorylation is the pore-forming protein mixed lineage kinase like (MLKL) in the necroptotic pathway,^38,39 human PMN lack MLKL⁷ and additional substrates have been recently identified, including enzymes in metabolism, such as pyruvate dehydrogenase complex,⁴⁰ or regulators of chemokine expression during viral infection.⁴¹ It seems that RIPK3 serves multiple functions, well beyond driving necroptosis, and in cells such as human PMN may coordinate agonist-specific immune responses.

TABLE 1.

Differences in processing of pro-IL-1β and secretion of IL-1β between PMN and HMDM

	IL-1β secretion	IL-18 secretion	NLRP3	Caspase-1	Serine proteases	RIPK3
HMDM-nigericin	+	+	+	+	–	–
PMN-nigericin	+	+	±	±	+	–
PMN-SA	+	–	–	–	+	+

Processing of pro-IL-1β and secretion of IL-1β from human PMN and HMDM is cell-and agonist-specific mechanism. (+) pathway is involved; (–) pathway is not involved.

Although PMN release far less IL-1β on a per cell basis than do macrophages or dendritic cells, the generation of even small amounts of IL-1β per PMN would be greatly magnified by the large numbers of PMN recruited to an inflammatory site and hence could have significant clinical consequences. Additional work is ongoing to define the effector molecules downstream of RIPK3 that promote processing and secretion of IL-1β in PMN-SA.

Abbreviations:

Ac-YVAD-cmk

N-acetyl-L-tyrosyl-L-valyl-N-[(1S)-1-(carboxymethyl)-3-chloro-2-oxo-propyl]-L-alaninamide

CA-MRSA

Community-associated methicillin-resistant Staphylococcus aureus

DAMPs

Damage-associated molecular patterns

DFP

Diisopropylfluorophosphate

HK SA

Heat-killed Staphylococcus aureus

HMDM

Human monocyte-derived macrophages

HSA

Human serum albumin

IL

Interleukin

LDH

Lactate dehydrogenase

LPS

Lipopolysaccharide from Escherichia coli O111:B4

MLKL

Mixed lineage kinase like

MOI

Multiplicity of infection

OpZ

Opsonized zymosan

PBMC

Peripheral blood mononuclear cells

PMN

Polymorphonuclear neutrophils

PMN-SA

PMN that harbored viable SA

RIPK

Receptor-interacting protein kinase

RLU

Relative light units

SA

Staphylococcus aureus USA300 (LAC)

TSB

Tryptic Soy Broth