Poly(ADP-Ribose) Polymerase-1 Regulates Pyroptosis Independent Function of NLRP3 Inflammasome in Neutrophil Extracellular Trap Formation

Neutrophil extracellular traps (NETs) continue to be at the center of intense scientific inquiry owing to expanding reports on their association with infectious and sterile inflammatory conditions alike (1, 2). NETs constitute decondensed chromatin fibers expelled primarily through a specialized cell death program of neutrophils called “NETosis,” where neutrophils stimulated by external stimuli activate intracellular signaling cascades that converge on nuclear envelope and granular membrane disintegration, resulting in intracellular mixing of DNA and proteolytic granular enzymes (3–5). Studies on the molecular mechanism of NETosis so far have revealed that this form of lytic cell death is distinct from apoptosis or necrosis and can occur in a reactive oxygen species–dependent and –independent manner in response to relevant stimuli (6). Although the signaling mechanisms that coordinate apoptotic and other nonapoptotic lytic cell death in myeloid cells such as macrophages, and to some extent neutrophils, are well described, such mechanisms in NETosis are poorly understood (7, 8).

Inflammasomes are cytosolic multiprotein assemblies that provide critical innate defense against microbial infections (9). NACHT domain-leucine-rich repeat-and pyrin domain-containing protein 3 (NLRP3) is among the most well-characterized inflammasomes that drive maturation and release of inflammatory cytokine IL-1β associated with pyroptotic cell death (10, 11). NLRP3 activation constitutes a two-step process where the first, priming signal, typically TLR4 activation by LPS, leads to NF-κB–dependent transcriptional upregulation of NLRP3 and proproteins caspase-1 and IL-1β. The second signal, resulting in cation efflux and ATP generation, causes NLRP3 inflammasome complex assembly with pyrin and CARD domain-containing protein apoptosis-associated speck-like protein containing a CARD (ASC) that provides a platform for maturation of procaspase-1 to active enzyme caspase-1, which cleaves pro–IL-1β to its mature form. Active caspase-1 also cleaves gasdermin D (GSDMD), a pore-forming protein to N-terminal fragments that oligomerize on plasma membrane to form pores serving as secretory conduits for the release of IL-1β and pyroptotic death of the cell (12). Concomitantly, blockage of GSDMD pores inhibits cytokine release and pyroptosis (13). Canonical (caspase-1–dependent) or noncanonical (caspase 4/11–dependent) inflammasome activation with GSDMD as the common denominator has been shown to promote NETosis (14–16). Inhibition of GSDMD was reported to prevent NET-mediated organ injury in mice (17). In contrast, mice with neutrophil-specific depletion of GSDMD exhibit intact NET formation in vivo and remain susceptible to sepsis (18). Additional evidence has since supported the notion that GSDMD is dispensable for NETosis (19). This suggests that the mechanisms controlling inflammasome activation in neutrophils and the downstream events distinguishing NETosis from pyroptosis machinery remain to be fully understood.

Poly(ADP-ribose) polymerase-1 (PARP-1) is a nuclear enzyme with a central role in DNA repair and DNA damage response (20). Catalytic activity of PARP-1 transfers ADP-ribose moieties from NAD⁺ to a variety of protein substrates, a process called “PARylation” that alters protein–protein and protein–DNA interaction (21). Not surprisingly, PARP-1 was identified as a therapeutic target in cancer, and inhibitors of PARP-1 enzyme activity are U.S. Food and Drug Administration approved for tumors deficient in repair of DNA double-strand breaks (22). However, emerging evidence suggests additional functions of PARP-1 in cell death by mechanisms beyond DNA repair. In response to oxidative stress and inflammatory cytokines, PARP-1 can regulate NF-κB activation, which can establish a feedforward inflammatory circuit by generating heightened reactive oxygen species and inflammation leading to cellular necrosis (23, 24). PARP-1 is also shown to regulate parthanatos (PAR polymerase-1–dependent cell death) via mitochondrial release of apoptosis-inducing factor (25). More recently, PARP-1 was shown to activate NLRP3-mediated IL-1β release and pyroptosis in macrophages (26). These studies have highlighted PARP-1 as a key player in cell death mechanisms; however, the mechanistic function of this enzyme in the regulation of NETosis is largely unknown.

In this study, we examined NLRP3-activated NET formation in response to LPS/nigericin (LPS/nig) and the role of PARP-1 in the regulation of this inflammasome activation. We found an unexpected function of PARP-1 in regulating NLRP3-mediated NETosis independent of IL-1β secretion and pyroptosis via p38 MAPK. PARP-1 inhibition (PARPi) in LPS/nig-stimulated neutrophils abrogated inflammasome-mediated NET formation, but caspase-1 activation and GSDMD processing remained unaffected. To our knowledge, our study identifies a novel role of PARP-1 in controlling a noncanonical, pyroptosis-independent function of NLRP3 in NETosis and provides a rationale for therapeutic targeting of PARP-1 to specifically inhibit excessive NET formation.

LPS from Escherichia coli (catalog no. 297-473-0), nigericin (catalog no. 28643-80-3), and PARP-1 inhibitor 3,4-dihydro-5-[4-(1-piperidinyl)butoxyl]-1(2H)-isoquinolinone (DPQ) (catalog no. D5314) were purchased from Sigma-Aldrich (St. Louis, MO). Additional reagents include p38 MAPK inhibitor SB203580 (catalog no. tlrl-sb20; Invitrogen), SYTOX Green Nucleic Acid Stain (catalog no. S7020; Molecular Probes, Life Technologies), SuperSignal West Femto Substrate (catalog no. 34094, Thermo Fisher Scientific), SuperSignal West Pico PLUS Chemiluminescent Substrate (catalog no. 34580, Thermo Fisher Scientific), disuccinimidyl suberate (catalog no. 21655, Thermo Fisher Scientific), and olaparib (SelleckChem catalog no. S1060). Primary Abs anti-NLRP3 (catalog no. 15101), anti-ASC/TMS1 (catalog no. 67824), anti-caspase-1 (catalog no. 24232), anti-cleaved caspase-1 (catalog no. 89332), anti-IL-1β (catalog no. 31202), anti-cleaved-IL-1β (catalog no. 63124), anti-p38 MAPK (catalog no. 8690), anti-phospho-p38 MAPK (catalog no. 4511), anti-phospho-NF-κβ p65 (catalog no. 3031), anti-NF-κβ (catalog no. 8242), anti-GSDMD (catalog no. 39754), anti-cleaved GSDMD (catalog no. 10137), anti-β-actin (catalog no. 4970), and the secondary Ab anti-rabbit IgG, HRP-linked, were purchased from Cell Signaling Technology. Other Abs included poly(ADP-ribose) (catalog no. AG-20T-0001, AdipoGen Life Sciences), StarBright blue 520 goat anti-rabbit IgG (catalog no. 12005869), StarBright blue 700 goat anti-mouse IgG (catalog no. 12004158), rabbit anti-myeloperoxidase (anti-MPO) Ab (Invitrogen catalog no. PA5-16672), rabbit anti-H3Cit mAb (Abcam 219407), and goat anti-rabbit IgG (H + L) highly cross-adsorbed secondary Ab, Alexa Fluor Plus 594 (Invitrogen, catalog no. A32740). For flow cytometry, Live/Dead Zombie aqua (catalog no. 433101) and Fc Blocker TruStain FcX PLUS (catalog no. 156604) were purchased from BioLegend; CD11b-allophycocyanin-eFluor 780 Ab (catalog no. 47-0112-82) was purchased from Invitrogen, and LY-6G-PE Ab (catalog no. 551461) was purchased from BD Biosciences.

For in vitro studies, neutrophils were recovered from mice 8–12 h following injection of 4% sterile thioglycolate in the peritoneal cavity, as we previously described (27–30). Purity of neutrophils was determined by Ly6G staining and SYTOX Green staining (Supplemental Fig. 1). Cells (10 × 10⁶) from the neutrophil-enriched peritoneal lavage were plated (100 × 20–mm dish) and treated with 5 µM DPQ (a well-characterized PARP-1 inhibitor) for 30 min before priming with LPS to a final concentration of 1.0 µg/ml for 3 h followed by activation with nigericin (15 µM) for 45 min. In some experiments to probe the activation of ASC/TMS1, cells were incubated with 2 mM crosslinker disuccinimidyl suberate for an additional 30 min at 37°C and 5% CO₂, as previously described (31). In the experiments involving p38 MAPK inhibitor SB203580, neutrophils were treated with DPQ for 15 min followed by incubation with 10 µM SB203580 for another 15 min before priming and activation with LPS and nigericin as described above. For experiments involving the caspase-1 inhibitor ac-YVAD-cmk, neutrophils were treated with DPQ for 15 min followed by incubation with 30 µg/ml ac-YVAD-cmk for another 15 min before priming and activation with LPS and nigericin as described above.

Cells were kept at 37°C and 5% CO₂ during all incubations and were used to prepare cell lysate for Western blot analysis or for SYTOX Green staining to quantitate NETs as described below.

The nucleic acid dye SYTOX Green was used to stain and visualize NETs, as previously described by our laboratory (27–29, 32, 33). Briefly, after stimulation, neutrophils were washed with PBS, cytocentrifuged, and fixed with 4% paraformaldehyde for 10 min. The cells were stained using a 100 nM solution of SYTOX Green. The images were captured at a magnification of 20× using an EVOS M5000 microscope. The percentage of NET formation was calculated by dividing the number of NET-forming neutrophils by the total number of cells in 7–10 random microscopic areas and multiplying the results by 100. For some experiments, cytocentrifuged NETs were costained with nuclear stain DAPI (Thermo Scientific, catalog no. 62248) and rabbit anti-MPO polyclonal Ab (Invitrogen, catalog no. PA516672) or with rabbit anti-H3Cit mAb (Abcam 219407) followed by goat anti-rabbit Alexa Fluor 546 Ab. The images were taken using a Nikon Eclipse E600 fluorescence microscope under 400× magnification and analyzed using NIS-Elements BR Imaging software.

Peritoneal neutrophils were harvested following stimulations, washed with ice-cold PBS, and lysed with radioimmunoprecipitation assay buffer supplemented with cOmplete Mini protease and phosphatase inhibitor mixture with 10 μg/ml aprotinin (catalog no. 10236624001), 10 μg/ml leupeptin hemisulfate (catalog no. L2884), 10 μg/ml pepstatin (catalog no. 11359053001), 10 μM 4-(2-aminoethyl)benzenesulfonyl fluoride (catalog no. A8456), 10 mM sodium orthovanadate (catalog no. 450243), 5 mM benzamidine (catalog no. 12072), and 20 mM levamisole chloride (catalog no. 16595-80-5), all from Sigma-Aldrich. The cell lysates were resolved in a 10–20% gradient SDS-PAGE (Thermo Fisher Scientific) and transferred to polyvinylidene difluoride membranes using an iBlot 2 apparatus (Thermo Fisher Scientific). The membranes were washed with TBS containing 0.1% Tween-20, blocked with 5% milk, and incubated overnight with the primary Ab at 4°C. The next day, the membranes were washed and incubated with the HRP-linked secondary Ab for 1 h at room temperature. SuperSignal West Femto or Pico chemiluminescence substrate was used to detect the signals with a LI-COR Odyssey imager. β-Actin was used as a loading control in all blots. Densitometry of individual blots was done using ImageJ software. For dual-protein detection fluorescent immunoblotting, the cell lysate proteins resolved in the SDS-PAGE were transferred to polyvinylidene difluoride membranes, then washed with TBS containing 0.1% Tween-20. The membrane was blocked for 1 h with 5% milk followed by addition and incubation overnight of anti-NLRP3 primary Abs with the membrane at 4°C. The next day, the membrane was washed, then incubated with PARylation Ab overnight. The membrane was washed and incubated in the dark with both fluorescently labeled secondary Abs (StarBright blue 520 goat anti-rabbit IgG, StarBright blue 700 goat anti-mouse IgG) simultaneously for 1 h. The fluorescence images of the membrane were taken using a Bio-Rad ChemiDoc MP imaging system after multiple washes.

The peritoneal lavage supernatant obtained after spinning down the stimulated and nonstimulated cells from different samples was saved for cytokine-level analysis by ELISA using an IL-1β Quantikine ELISA kit (catalog no. MLB00C; Biotechne, R&D Systems). The protocol provided by the manufacturer was used.

Neutrophils were seeded at 10 × 10⁵ cells in 100 × 20–mm dishes and treated with the indicated stimuli as described above. LDH release was quantified in the culture supernatants using the CyQUANT LDH Cytotoxicity Assay kit (InvivoGen, catalog no. C20300) according to the manufacturer’s instructions.

Total RNA was isolated from purified neutrophils using the RNeasy Micro Kit (catalog no. 74004) from Qiagen per the manufacturer’s instructions and used for quantitative RT-PCR as reported by us (27, 34, 35). The primers used in the present study were as follows: NLRP3 (forward) 5′-CTCCCGCATCTCCATTTGT-3′ and (reverse) 5′-GCGTTCCTGTCCTTGATAGAG-3′; and IL-1β (forward) 5′-TGGCAACTGTTCCTGAACTC-3′ and (reverse) 5′-GGAAGCAGCCCTTCATCTTT-3′. Expression levels of these genes of interest were normalized to transcripts of housekeeping gene 18S RNA in the same samples. The fold change was calculated by dividing the normalized value of the gene of interest in stimulated samples by the corresponding normalized value in unstimulated samples.

Neutrophil isolation and in vivo studies were carried out using 6- to 8-wk-old wild-type C57BL/6 mice in the animal facility of the University of Texas MD Anderson Cancer Center. The animals were used in accordance with institutional and federal guidelines. For in vivo studies, all mice were peritoneally infected with 5000 CFU Klebsiella pneumoniae (Kpn) in 100 µl PBS and segregated into two groups: one group called the vehicle group and the other the treated group. At 8 h postinfection (hpi), the treated group was i.p. injected with 10 µg/kg DPQ in 100 µl PBS. The vehicle group control mice was i.p. injected with PBS containing 0.2% DMSO. The survival of mice was monitored for 6 d. For some experiments, the peritoneal neutrophils of some mice of each group were collected 24 hpi as described by us (27). Cells were processed and purified to prepare lysates for Western blot analysis or to quantitate NETs, as described above.

Peritoneal lavage cells from mice infected i.p. with GFP-labeled Kpn(kind gift from Dr. Steven Clegg, University of Iowa) with or without DPQ treatment were isolated 24 hpi and were processed for flow cytometry as previously described by us (29, 35, 36). Cells were preincubated with Fc block in PBS and Live/Dead Zombie Aqua Stain (BioLegend, catalog no. 423101), followed by incubation with fluorochrome‐labeled Abs for neutrophil and macrophage marker CD11b and Ly6G. A BD FACSARIA II flow cytometer was used for data acquisition, and FlowJo software (FlowJo LLC) was used to analyze all data.

Statistical studies of survival were carried out using the Kaplan–Meier log-rank test. The comparison of mean values across various groups was done using the Student t test and one-way ANOVA (GraphPad Prism version 9). A p value <0.05 was considered to be statistically significant.

To examine NLRP3-mediated NET formation in vitro, thioglycolate-elicited mouse peritoneal neutrophils were used, as reported by us previously (27, 28, 32). Cells were primed with LPS for 3 h followed by 45-min exposure to nigericin, a potassium ionophore. This costimulation, which characteristically induces NLRP3 inflammasome assembly and activation, triggered robust NET formation as assessed by SYTOX Green staining and imaging quantitation (Fig. 1A, A′). These NETs exhibited characteristic localization of citrullinated histone (Fig. 1A) and MPO (Supplemental Fig. 2A) with DNA. Importantly, treatment of neutrophils with PARP-1 antagonist DPQ in a similar experimental setting resulted in significant inhibition of NET formation induced by LPS/nig stimulation (Fig. 1A, A′). The induction of NETs by LPS/nig correlated with an increased expression of NLRP3 protein compared with unstimulated neutrophils, which was abrogated by DPQ treatment (Fig. 1B). LPS alone, used as a control, induced moderate NETs with little H3 citrullination and increased NLRP3, which remained largely unaffected by DPQ treatment (Fig. 1A, 1B), suggesting a differential PARP-1 regulation of LPS/nig-stimulated NETs. Furthermore, DPQ inhibited the release of LDH upon LPS/nig, but not LPS-alone stimulation (Fig. 1C), suggesting that specific PARP-1 regulation of NLRP3 activation-induced lytic NETosis NLRP3 activation of bone marrow neutrophils by LPS/nig also induced NET formation that was abrogated similarly upon PARPi (Supplemental Fig. 2B). Moreover, olaparib, a U.S. Food and Drug Administration–approved inhibitor of PARP-1, was also effective in inhibiting NETs and NLRP3 in LPS/nig-stimulated neutrophils (Supplemental Fig. 2C).

Having confirmed a similar response of thioglycolate-elicited and bone marrow neutrophils and the specificity of PARP-1–mediated blockage of NLRP3-stimulated NETosis, all experiments from there on were conducted with peritoneal neutrophils with LPS/nig stimulation and DPQ as a PARP-1 inhibitor. Overall, our findings suggested a regulatory role of PARP-1 in neutrophil NLRP3 activation and NET formation.

Next, we assessed if reduced levels of PARP-1 protein contributed to a NET inhibitory effect of DPQ in vitro. LPS/nig stimulation increased the expression of PARP-1 protein, which was significantly reduced upon DPQ treatment of LPS/nig-stimulated neutrophils compared with those without treatment (Fig. 2A, A′), suggesting that abrogation of PARP-1 correlated with inhibition of NLRP3-mediated NET formation. Canonical activation of the NLRP3 inflammasome by LPS and nigericin constitutes a two-step process where priming by LPS induces an increase in the expression of NLRP3 followed by nigericin-mediated assembly of an inflammasome complex via recruitment of ASC. We found that LPS/nig-stimulated neutrophils exhibited robustly increased NLRP3 transcript (Fig. 2B) as well as protein (Fig. 2C, C′), as compared with unstimulated control cells. Interestingly, PARPi by DPQ reduced the levels of NLRP3 protein, but not the NLRP3 transcript levels. This indicated that PARPi regulated inflammasome activation but not the priming events. In line with this finding, phosphorylation of NF-κB, which is required for transcriptional activation of genes encoding inflammasome components, was significantly increased upon LPS/nig stimulation and remained similarly high despite PARPi (Fig. 2D, D′). On the other hand, LPS/nig-stimulated neutrophils showed a robust increase in ASC dimers compared with unstimulated cells, indicative of inflammasome assembly, which was diminished upon PARPi by DPQ (Fig. 2E, E′).

Inflammasome activation leads to pyroptotic cells death characterized by IL-1β release. Remarkably, PARPi did not exhibit any effect on IL-1β release from neutrophils because similar levels of this cytokine were detected intracellularly (Fig. 3A) as well in culture supernatants (Fig. 3B) of LPS/nig-stimulated neutrophils with or without DPQ treatment. Moreover, in line with the intact priming events, the levels of IL-1β transcript (Fig. 3C) and the unprocessed proprotein (Fig. 3D) remained unaffected by PARPi with DPQ, although the inflammasome stimulation by LPS/nig caused their increase in neutrophils as compared with the unstimulated cells. This suggested that PARP-1 regulation of NLRP3 activation in NETting neutrophils was distinct from pyroptotic cell death pathway. To confirm this, we examined the effect of PARPi on pyroptosis effector proteins caspase-1 and GSDMD. Indeed, no change was observed in the levels of cleaved caspase-1 (Fig. 3E) or N-terminal p30 fragments of GSDMD (Fig. 3F) in LPS/nig-stimulated neutrophils with or without PARPi. Full blots are shown in Supplemental Fig. 3. This strongly suggested that PARP-1 regulation of NLRP3 activation in response to LPS/nig stimulation in neutrophils occurred independently of pyroptosis. This was further confirmed by examining NET formation in neutrophils with pharmacological inhibition and genetic ablation of caspase-1 or GSDMD, two effector proteins essential for IL-1β processing and release during pyroptotic cell death. Indeed, treatment of neutrophils with Ac-YVAD-cmk, a specific caspase-1 inhibitor, failed to inhibit NET formation in response to LPS/nig (Fig. 3G). Similarly, neutrophils from gasdermin^−/− mice remained fully competent to produce NETs in response to LPS/nig, which was inhibited only upon PARPi (Fig. 3H). Overall, these data strongly supported PARP-1 regulation of NLRP3 activation-induced NETosis independent of the canonical pathway of NLRP3-mediated pyroptosis in neutrophils.

We next examined the molecular signaling pathway affected by PARPi downstream of inflammasome activation in neutrophils. Because p38 MAPK has been shown to regulate a wide range of cellular responses, including DNA damage, inflammation, cell death, and survival (37), we sought to determine if PARPi affected activation of this signaling protein in neutrophils. Western blot analysis showed little to no increase in levels of phosphorylated p38 in LPS/nig-activated neutrophils compared with the unstimulated cells (Fig. 4A). Surprisingly, we found a significant upregulation of p38 phosphorylation in neutrophils upon PARP- 1 inhibition (Fig. 4A A′). The levels of total p38 protein and the internal control β-actin remained similar across all samples. This suggested a negative regulatory role of p38 phosphorylation downstream of PARP-1. To test this in NLRP3-activated NETs, we examined if the inhibitory effect of DPQ on NET formation could be reversed by restraining p38 phosphorylation. For this, LPS/nig-stimulated neutrophils were treated with PARP-1 inhibitor in the presence or absence of SB203580, a pharmacological inhibitor of p38 phosphorylation. As shown in Fig. 4, LPS/nig induced NET formation in 67 ± 25.7% of neutrophils, which was significantly reduced to 1.4 ± 0.6% in the presence of PARP-1 inhibitor DPQ. Importantly, treatment of neutrophils with SB203580 reversed this inhibitory effect of DPQ and restored LPS/nig-induced NETs to significantly higher levels than DPQ alone (Fig. 4B, B′). That this PARP-1/p38 regulatory loop orchestrated NET formation via NLRP3 activation was confirmed by Western blot analysis where inhibition of p38 phosphorylation restored NLRP3 protein in LPS/nig-stimulated neutrophils with PARPi to levels seen in LPS/nig alone (Fig. 4C, C′). Taken together, these data showed that the negative regulatory effect of PARPi on NLRP3 inflammasome requires p38 phosphorylation.

To test the physiological relevance of PARPi, DPQ was injected i.p. in mice undergoing infectious peritonitis caused by i.p. infection with Kpn. In this preclinical model, we have previously shown an inflammatory role of NETs (27). We found that purified peritoneal neutrophils from mice undergoing Kpn peritonitis showed substantial NET formation, which was significantly reduced in mice treated with DPQ (Fig. 5A, A′). Reduced NET formation by PARPi in vivo correlated with significantly improved survival of mice (Fig. 5B). Interestingly, despite reduced NET formation, DPQ treatment resulted in improved bacterial clearance (Fig. 5C), which likely contributed to the protective effect. We examined if an overall cytoprotective effect of PARPi caused an increase in the number of cells in the peritoneum available for bacterial clearance. Indeed, there was a significantly higher number of total peritoneal exudate cells recovered from Kpn-infected mice upon DPQ treatment compared with the vehicle control (Fig. 5D). Flow cytometric analysis did not reveal any significant changes in the frequency of phagocytic cells, CD11b⁺ macrophages, or CD11b⁺Ly6G⁺ neutrophils in infected DPQ-treated mice compared with untreated counterparts (Fig. 5E). Intraperitoneal infection with GFP-labeled Kpn also did not highlight any differences in bacterial uptake by peritoneal cells with or without DPQ treatment (Fig. 5E). Although the identity of an immune cell subset contributing to improved bacterial clearance in the absence of NET formation in vivo and a detailed mechanism of overall protective effect of DPQ remain to be determined, our findings confirmed that PARPi leads to inhibition of NETs, which correlates with overall improved survival of mice in the disease setting (i.e., bacterial peritonitis).

NET formation is a form of lytic cell death distinct from other cell death mechanisms that sits at the intersection of inflammation and host defense. Viable methods of controlling NETs in diseases has remained elusive because of limited knowledge about cell signaling pathways distinguishing these mechanisms. Recent studies have established NLRP3 as a common denominator in neutrophil cell death programs of pyroptosis and NETosis. With completely different consequences of these two cell death processes in various disease settings, there is a wide knowledge gap regarding the upstream cellular components regulating the function of this inflammasome. In this study, we show PARP-mediated regulation of NLRP3 activation controlling NET formation in a pyroptosis-independent manner. Our data identified PARP inhibition as a critical regulatory mechanism of NLRP3-mediated NET formation via modulation of p38 MAPK.

We found that pharmacological inhibition of PARP-1 abrogated NETosis in response to NLRP3 activation. PARP-1 is the prototypic member of an evolutionarily conserved PARP family of 17 proteins, which uses nicotinamide as a substrate to add branched ADP-ribose moieties to target proteins to modify their function. Although it is known mainly as a sensor of DNA damage and activation of DNA repair machinery and is targeted for inhibition of cancer cell growth, more recent evidence has supported its wider function in inflammation and immune responses (38). Studies from our group and others have implicated PARP-1 in reactive oxygen species–induced NET formation, albeit via two distinct mechanisms of activating reactive oxygen species sensor ion channel and by chromatin decondensation to facilitate NET release (30, 39). These studies suggest a multifaceted role of PARP-1 in neutrophils. Indeed, our findings of PARP-1 regulation of NLRP3-mediated NETs reported in this study add to the repertoire of DNA repair-independent cellular functions attributed to this protein.

The NET inhibitory effect of PARPi in LPS/nig-stimulated neutrophils was found to be via reduction of NLRP3 protein. Canonical NLRP3 inflammasome activation consists of a priming step, typically TLR4 stimulation by LPS, that is accomplished by activating the NF-κB pathway, which causes changes in post-translational modifications in NLRP3 permitting NLRP3 inflammasome assembly as well as upregulation of NLRP3 and pro–IL-1β proteins. We found that PARPi decreased the protein level but not the transcript of NLRP3. In line with this, NF-κB activation remained unaffected in LPS/nig-stimulated neutrophils upon PARPi. This indicated a different mechanism of PARP function in neutrophils from what has been reported in macrophages, where PARP-1 positively regulates NF-κB activation downstream of TLR activation (23). Reduced protein level of NLRP3 by PARPi indicated post-translational regulation. In this regard, PARP-1 catalyzes the polymerization and attachment of the negatively charged poly(ADP-ribose) complex to target proteins, a process known as PARylation, which regulates protein–protein interaction and complex formation. Although PARylation typically occurs at DNA damage sites and contributes to assembly of DNA repair response proteins, recent evidence has shown PARP-1–mediated PARylation of NLRP3 in macrophages modulating inflammasome activation (26). In contrast, we did not find any difference in PARylation of NLRP3 in neutrophils with PARPi (Supplemental Fig. 4). Although we have not tested other possible modifications that may be contributing to the reduced level of NLRP3 by PARPi, degradation of the protein by ubiquitination is an attractive possibility. This is because recently it has been shown that PARP-1 activity blocks ubiquitin-proteasomal degradation of proteins (40). It is possible that in the absence of PARP-1, NLRP3 is targeted for proteasomal degradation. We are currently investigating this mechanism.

The canonical function of NLRP3 is to facilitate activation of caspase-1 for secretion of inflammatory cytokines IL-1β and IL-18, accompanied by caspase-1–dependent cell death called “pyroptosis” (41). Following an activating step, oligomerized NLRP3 promotes oligomerization of the adaptor protein ASC, which binds and activates procaspase-1. Pro-IL-18 and pro-IL-1β are then cleaved into their mature forms by active caspase-1. Caspase-1 induces the cleavage of GSDMD as well into N-terminal fragments that produce lytic pores and the release of inflammatory cytokines IL-1β and IL-18. Our data showed that in neutrophils, PARPi-mediated inhibition of NLRP3 activated NET formation occurred independent of these events, because IL-1β secretion, caspase-1 activation, and GSDMD processing remained unaffected in PARPi-treated neutrophils. This suggested a noncanonical, pyroptosis-independent function of NLRP3 in NETosis, controlled by PARP-1. Although a noncanonical caspase-1–independent NET formation by inflammasome activation has been shown previously (14), GSDMD processing and IL-1β secretion accompanied NETosis in this report. These contrasting observations could be due to the stimulus used (PMA or cytosolic LPS). In line with this, our data showing lack of PARPi blockage of LPS-induced NLRP3 and LDH release strongly suggests a distinct pathway of NLRP3-activated lytic NETosis that is regulated by PARP-1. To our knowledge, our finding of caspase-1 and GSDMD-independent NET formation downstream of NLRP3, regulated by PARPi, likely represents a novel pathway of pyroptosis-independent function of NLRP3 activation in neutrophils. As such, NLRP3 is increasingly recognized to play inflammasome-independent functions in T lymphocytes (42) and nonimmune cells (43, 44). To the best of our knowledge, ours is the first report of PARP-1 regulation of noncanonical NLRP3 activation in neutrophils, highlighting the potential application of PARP-1– and NLRP3-targeted therapies for neutrophilic acute inflammatory diseases.

An unexpected finding in our study was the requirement of p38 MAPK activation for the NET-inhibitory effect of PARPi. Three distinct subgroups of MAPK signaling pathways constitute cytoplasmic phosphorylation of ERKs, JNK, and the p38 kinases. Of these, activation of the p38 MAPK pathway can lead to inflammation and is also implicated in cell growth, differentiation, and death (37). Evidence exists for an association of PARP-1 signaling with both positive and negative regulation of MAPK signaling in a context-dependent manner (45). One study showed that PARP-1 inhibition protected cardiomyoblasts against hydrogen peroxide–induced cell death by increasing the phosphorylation of ERK1/2 and p38 MAPK and that the effect of PARPi was reversed by ERK1/2 and p38 inhibitors (46). Another study showed that administration of PARPi protected rat hearts from myocardial injury, and this protective effect correlated with increased activation of p-ERK and p-p38 (47). Our results are in line with these findings, and the NET inhibitory effect of PARPi was attributed directly to increased p-p38. Whether this cytoprotective effect of increased p-p38 is a consequence of upregulation of anti-NETosis factors or downmodulation of pro-NETosis factors, if any, remains to be determined. Also of importance to our study, p38 MAPK activity was recently shown to prevent hyperactivation of NLRP3 by regulating intracellular Ca²⁺, and mice with granulocytic deficiency of p38 MAPK exhibited increased susceptibility to LPS-induced septic inflammation (31). This molecular interplay between MAPK signaling, NLRP3, and PARPi raises an exciting possibility of targeting p38 MAPK for suppressing inflammatory and prometastatic NETosis in cancers treated with PARPi that develop resistance to this treatment over time.

Increased PARP-1 activity has been reported in multiple nononcologic diseases, including severe sepsis (48), asthmatic airway inflammation (49), burn injury (50), Chagas disease (51), diabetes (52), and rheumatoid arthritis (53). Not surprisingly, mice with genetic deficiency of PARP-1 or its pharmacological inhibition were protected from many of these inflammatory disease conditions, mainly because of reduced infiltration of immune cells resulting in dampening of inflammation (54). More relevant to our study, PARP-1 knockout mice are resistant to endotoxic shock (24). Although the downregulation of macrophage-mediated inflammatory response was shown to underlie this effect, a specific contribution of neutrophils or the effect of PARP-1 deficiency on neutrophil function was not been determined. In this regard, one study reported downregulation of MPO activity in PARP-1^−/− mice with LPS-induced lung inflammation, which was most likely attributed to reduced neutrophil infiltration as a result of PARP-1 deficiency (55). It is worth noting that MPO activity is shown to be required for NET formation (56), and NLRP3-mediated amplification of neutrophil MPO activity is implicated in inflammation (57). Thus, it is likely that reduced NLRP-3–mediated NET formation contributed to protection of PARP-1–deficient mice in these studies. We found that administration of PARPi protected the mice from acute peritonitis caused by Kpn. This treatment inhibited NETosis, with concomitantly reduced bacterial burden, which appeared paradoxical, considering a proven antimicrobial function of NETs in infectious diseases. However, consistent with a known overall cytoprotective effect of PARPi, we found a significantly increased number of leukocytes in the peritoneal lavage of mice treated with PARPi, which likely made more cells available for bacterial clearance and contributed to significantly improved survival of mice.

Overall, to our knowledge, our study demonstrates a previously unrecognized regulatory function of PARP-1 in controlling NLRP3 activated NET formation. PARPi revealed an unexpected mechanism of NLRP3-triggered NETosis independent of the classical pyroptosis through negative regulation by p38 MAPK. This pathway can uncover previously unappreciated therapeutic targets to control NLRP3-activated NETs without compromising pyroptosis, an important host defense mechanism in infectious diseases.

The authors have no financial conflicts of interest.

We thank Joylise Mitchell and Jay Sayonanh for assistance in NET quantitation.