GSDMD drives canonical inflammasome‐induced neutrophil pyroptosis and is dispensable for NETosis (original) (raw)

Abstract

Neutrophils are the most prevalent immune cells in circulation, but the repertoire of canonical inflammasomes in neutrophils and their respective involvement in neutrophil IL‐1β secretion and neutrophil cell death remain unclear. Here, we show that neutrophil‐targeted expression of the disease‐associated gain‐of‐function _Nlrp3_A350V mutant suffices for systemic autoinflammatory disease and tissue pathology in vivo. We confirm the activity of the canonical NLRP3 and NLRC4 inflammasomes in neutrophils, and further show that the NLRP1b, Pyrin and AIM2 inflammasomes also promote maturation and secretion of interleukin (IL)‐1β in cultured bone marrow neutrophils. Notably, all tested canonical inflammasomes promote GSDMD cleavage in neutrophils, and canonical inflammasome‐induced pyroptosis and secretion of mature IL‐1β are blunted in GSDMD‐knockout neutrophils. In contrast, GSDMD is dispensable for PMA‐induced NETosis. We also show that Salmonella Typhimurium‐induced pyroptosis is markedly increased in Nox2/_Gp91_Phox‐deficient neutrophils that lack NADPH oxidase activity and are defective in PMA‐induced NETosis. In conclusion, we establish the canonical inflammasome repertoire in neutrophils and identify differential roles for GSDMD and the NADPH complex in canonical inflammasome‐induced neutrophil pyroptosis and mitogen‐induced NETosis, respectively.

Synopsis

The alternative text for this image may have been generated using AI.

Canonical inflammasomes in neutrophils promote caspase‐1‐dependent cleavage of proIL‐1β and GSDMD. GSDMD mediates secretion of IL‐1β and canonical inflammasome‐induced neutrophil pyroptosis, whereas GSDMD is dispensable for NETosis induction by the phorbol ester PMA.

Similar content being viewed by others

Introduction

Inflammasomes are intracellular multiprotein complexes that recruit and activate inflammatory caspases. In particular, human procaspases 4 and 5 and murine procaspase‐11 are integral components of the non‐canonical inflammasome that responds to the cytosolic presence of lipopolysaccharides (LPS) of certain Gram‐negative bacterial pathogens, whereas procaspase‐1 operates as the universal effector protease of canonical inflammasomes (Lamkanfi & Dixit, 2014; Van Opdenbosch & Lamkanfi, 2019). Canonical inflammasome sensor proteins—whose role in the recognition of pathogen‐associated molecular patterns (PAMPs) and danger‐associated molecular patterns (DAMPs) have been firmly established in macrophage studies—include nucleotide‐binding domain and leucine‐rich repeat‐containing (NLR) family pyrin domain‐containing −1 and −3 (NLRP1 and NLRP3), NLR family CARD domain‐containing protein 4 (NLRC4), absent in melanoma‐2 (AIM2) and Pyrin (Lamkanfi & Dixit, 2014; Van Opdenbosch & Lamkanfi, 2019). These canonical inflammasome sensors respond to diverse PAMPs, such as microbial nucleic acids, bacterial secretion systems and components of microbial cell wall components; as well as to environmental stressors and endogenous DAMPs that report damage to host cells in the context of sterile inflammation such as uric acid crystals, high concentrations of extracellular adenosine triphosphate (ATP) and dislocated mitochondrial and nuclear nucleic acids in the cytosolic compartment (Lamkanfi & Dixit, 2014; Van Opdenbosch & Lamkanfi, 2019). Once active, caspase‐1 cleaves the pro‐inflammatory proteins interleukin (IL)‐1β and IL‐18 into bioactive cytokines (Lamkanfi & Dixit, 2014; Van Opdenbosch & Lamkanfi, 2019). In parallel, inflammatory caspases instruct the cleavage of the pore‐forming protein GSDMD, the liberated N‐terminal domain of which incorporates as multimers in the plasma membrane and elicits the formation of large GSDMD pores that result in plasma membrane permeabilization (PMP) and pyroptotic cell lysis of activated macrophages (Vande Walle & Lamkanfi, 2016; Van Opdenbosch & Lamkanfi, 2019).

Many studies characterized the functional inflammasome repertoire in macrophages and monocytes, and described how their activation in these myeloid cell types contributes to inflammatory diseases and host defence against microbial infections. However, relatively little is known about the functional inflammasome repertoire in neutrophils, although they are a highly prevalent myeloid cell type in the bloodstream and frequently the first immune cell type to be recruited to acutely infected tissues and during (sterile) inflammatory episodes (Papayannopoulos, 2018). Neutrophils are capable of a specialized neutrophil cell death programme termed NETosis that is centrally driven by the NADPH oxidase complex, degranulation and the formation of neutrophil extracellular traps (NETs) (Fuchs et al, 2007; Papayannopoulos et al, 2010; Papayannopoulos, 2018). A recent report showed that caspase‐11 activation in neutrophils may elicit GSDMD‐dependent NETs (Chen et al, 2018), and neutrophil elastase‐mediated GSDMD cleavage was also suggested to drive mitogen‐induced NETosis by phorbol 12‐myristate 13‐acetate (PMA) (Sollberger et al, 2018). Several studies have shown that the canonical NLRP3 and the NLRC4 inflammasomes in neutrophils promote maturation and secretion of IL‐1β (Mankan et al, 2012; Chen et al, 2014; Karmakar et al, 2015, 2016, 2020; Perez‐Figueroa et al, 2016; Hassane et al, 2017; Monteleone et al, 2018). However, a systematic analysis of the repertoire of functional canonical inflammasomes in neutrophils has not been reported. Moreover, the role of GSDMD in the context of neutrophil canonical inflammasome activation and PMA‐induced NETosis is unclear.

Here, we show that neutrophil‐targeted (MRP8/S100A8‐Cre‐driven) expression of the CAPS disease‐associated _Nlrp3_A350V mutant in mice fully recapitulated the reported perinatal lethality and skin inflammatory phenotype of full‐body mutant mice (Brydges et al, 2009), suggesting that neutrophil‐intrinsic NLRP3 hyperactivation suffices to promote inflammasome‐mediated inflammatory pathology and tissue damage in vivo. Additionally, a systematic analysis of inflammasome‐deficient bone marrow neutrophils (BMNs) confirmed the reported (Mankan et al, 2012; Chen et al, 2014; Karmakar et al, 2015, 2016, 2020; Perez‐Figueroa et al, 2016; Hassane et al, 2017; Monteleone et al, 2018) activity of the NLRC4 and NLRP3 inflammasomes in promoting maturation and secretion of IL‐1β from neutrophils, and also demonstrated maturation and secretion of IL‐1β by the canonical NLRP1b, Pyrin and AIM2 inflammasomes in neutrophils. Moreover, we showed that canonical inflammasomes induce GSDMD cleavage in neutrophils, and that caspase‐1‐ and GSDMD‐mediated PMP, pyroptotic cell lysis and secretion of mature IL‐1β were all blunted in GSDMD‐knockout neutrophils. In marked contrast, PMA stimulation failed to induce GSDMD cleavage, and PMA‐induced NETosis was unaffected in GSDMD‐deficient neutrophils. Consistent with published results (Fuchs et al, 2007; Ermert et al, 2009), PMA‐induced NETosis was abolished in neutrophils from Nox2/_Gp91_Phox‐deficient mice (that lack NADPH oxidase activity). Contrastingly, Salmonella Typhimurium‐induced pyroptosis was substantially increased in Nox2/_Gp91_Phox‐deficient neutrophils, whereas the central NETosis mediators PAD4, neutrophil elastase, neutrophil proteinase 3 and cathepsin G were all dispensable. In conclusion, we demonstrate that neutrophil‐intrinsic mutant NLRP3 activation suffices for autoinflammatory disease in vivo; established the canonical inflammasome repertoire in neutrophils; and identified differential roles for GSDMD and the NADPH complex in canonical inflammasome‐induced neutrophil pyroptosis and mitogen‐induced NETosis respectively.

Results

_NLRP3_A350V expression in neutrophils drives cutaneous erythema and perinatal lethality in mice

Gain‐of‐function mutations in the NLRP3 gene that render the protein constitutively active cause a spectrum of dominantly inherited cryopyrin‐associated periodic fever syndromes (CAPS) in patients (Hoffman et al, 2001; Van Gorp et al, 2019). Contrary to wild‐type macrophages, bone marrow‐derived macrophages (BMDMs) of knock‐in mice that heterozygously express the Muckle–Wells syndrome (MWS)‐associated NLRP3A350V mutant secrete substantial amounts of IL‐1β in their culture medium in response to LPS stimulation alone (Brydges et al, 2009). To determine whether this is also true for neutrophils of the disease‐associated _Nlrp3_A350V mutant mice, we bred mice that are homozygous for the _Nlrp3_A350V allele to transgenic mice that hemizygously expressed the tamoxifen‐inducible Cre‐ERT2 fusion gene (CreT) (Ventura et al, 2007). Following tamoxifen treatment and excision of the floxed neomycin resistance cassette, neutrophils of the resulting _Nlrp3_A350V/wtCreT+ mice express NLRP3 from both the wildtype and the mutant _Nlrp3_A350V alleles. Neutrophils from Cre‐ERT2‐negative littermates (_Nlrp3_A350V/wtCreT−), which only express NLRP3 from the wildtype allele, were used as controls in these experiments. As expected, secreted IL‐1β was not detected in culture media of vehicle (i.e. PBS)‐treated _Nlrp3_A350V/wtCreT− neutrophils (Fig 1A). Culture media of vehicle‐treated _Nlrp3_A350V/wtCreT+ neutrophils contained low but detectable levels of IL‐1β (Fig 1A), indicating limited basal activation of NLRP3A350V in naïve _Nlrp3_A350V/wtCreT+ neutrophils. Secreted IL‐1β levels were significantly increased upon LPS stimulation of _Nlrp3_A350V/wtCreT+ neutrophils, whereas LPS‐stimulated _Nlrp3_A350V/wtCreT− neutrophils failed to secrete IL‐1β (Fig 1A). These results indicate that, unlike wildtype NLRP3, the NLRP3A350V mutant promotes IL‐1β secretion from LPS‐stimulated neutrophils. To further corroborate inflammasome activation, we separately analysed cell lysates and culture media by Western blotting. As expected, LPS stimulation upregulated pro‐IL‐1β expression levels in lysates of both _Nlrp3_A350V/wtCreT− and _Nlrp3_A350V/wtCreT+ neutrophils (Fig 1B). However, cleaved IL‐1β (p17) was only detected in culture media of LPS‐stimulated _Nlrp3_A350V/wtCreT+ neutrophils (Fig 1B). Notably, culture media of LPS‐stimulated _Nlrp3_A350V/wtCreT+ neutrophils also contained robust amounts of mature caspase‐1 (p20) (Fig 1B), indicating that NLRP3A350V inflammasome activation in neutrophils promotes the concomitant release of mature IL‐1β and active caspase‐1 into the extracellular milieu.

Figure 1

Figure 1

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Neutrophil‐intrinsic NLRP3A350V activation promotes auto‐inflammatory pathology in mice

A, B Neutrophils isolated from tamoxifen‐treated _Nlrp3_A350V/wt CreT+ and _Nlrp3_A350V/wt CreT− mice were incubated with LPS (3 h) before culture media were analysed for secreted IL‐1β (A), and cell lysates and culture media were immunoblotted for β‐actin, caspase‐1 and IL‐1β (B). Cytokine values represent mean ± SEM of n = 4 biological repeats. Statistical significance was analysed by two‐way ANOVA and _P_‐values were corrected for multiple comparisons (Bonferroni). **P ≤ 0.01.

C Representative pictures of _Nlrp3_A350V/wt MRP8‐CreTg and _Nlrp3_A350V/wt MRP8‐Cre− mice on days 4 and 8 after birth. Left side in each picture: _Nlrp3_A350V/wt MRP8‐Cre− mice; right side in each picture: _Nlrp3_A350V/wt MRP8‐CreTg. Abscesses (day 4) and scaling erythema (day 8) are observed in _Nlrp3_A350V/wt MRP8‐CreTg mice.

D Survival curves of _Nlrp3_A350V/wt MRP8‐CreTg (n = 22) and littermate _Nlrp3_A350V/wt MRP8‐Cre− (n = 21) mice. Statistical significance was analysed by the log‐rank Mantel–Cox test. ****P ≤ 0.0001.

E Representative H&E‐stained sections of the spleen (top, scale bar: 100 μM) and liver (bottom, scale bar: 200 μM) of _Nlrp3_A350V/wt MRP8‐CreTg and littermate _Nlrp3_A350V/wt MRP8‐Cre− mice. Spleens of _Nlrp3_A350V/wt MRP8‐Cre− control mice (top left) display characteristic white pulp (WP) regions (example shown surrounded by dashed line) composed of lymphoid follicles and periarteriolar lymphoid sheaths (PALS), which at this magnification are recognized by dark‐purple staining, large aggregates of lymphocytes surrounding the central artery (CA) and their distinctive separation from the red‐staining red pulp (RP). WP was largely absent in spleens of _Nlrp3_A350V/wt MRP8‐CreTg (top middle). Megakaryocytes (MK) are indicated for easier recognition of areas of extramedullary haematopoiesis (EMH). Most of the spleen parenchyma of _Nlrp3_A350V/wt MRP8‐CreTg animals was replaced by EMH of mostly myeloid lineage, which can be recognized by the medium‐sized cells with more cytoplasm, giving a lighter appearance (inset). In both the overview picture (top middle picture) and inset (top right) of _Nlrp3_A350V/wt MRP8‐CreTg spleen, a small amount of EMH of the erythroid lineage is present that can be recognized by small aggregates of darker‐stained, smaller cells (asterisk, *). Liver sections of both _Nlrp3_A350V/wtMRP8‐Cre− control mice and diseased _Nlrp3_A350V/wt MRP8‐CreTg pups displayed comparable EMH of the erythroid lineage (bottom left picture, indicated with yellow arrows). Liver sections of _Nlrp3_A350V/wt MRP8‐CreTg pups also showed a modest increase in EMH of the myeloid lineage (bottom middle picture and inset (scale bar: 100 μM) in the bottom right, indicated with red arrowheads), as well as hepatocyte enlargement.

F–I H&E‐stained histological sections of the spleen (F, G) or liver (H, I) of _Nlrp3_A350V/wt MRP8‐Cre− (Cre−, n = 8) and _Nlrp3_A350V/wt MRP8‐CreTg (CreTg, n = 6) mice were assigned a score of 0–5 for lymphoid depletion (F) or EMH (G, H) or hepatocyte enlargement (I) by a board‐certified pathologist. Values represent mean ± SEM of n = 6–8 biological repeats. Statistical significance was analysed by the Mann–Whitney _U_‐test. ***P ≤ 0.001.

J–L Cytokine secretion analysis (Luminex) of serum samples obtained on day 7 from _Nlrp3_A350V/wt MRP8‐Cre− (Cre−, n = 8) and _Nlrp3_A350V/wt MRP8‐CreTg (CreTg, n = 6) pups. Values represent mean ± SEM of n = 6–8 biological repeats. Statistical analysis of serum cytokine levels was performed by the Welch's _t_‐test. *P ≤ 0.05; **P ≤ 0.01.

Knock‐in mice that express the NLRP3A350V mutant in myeloid lineage cells (based on lysosome M‐promoter‐driven expression of Cre recombinase) were previously shown to present with skin abscesses, scaling erythema and hyperkeratosis, growth retardation and perinatal lethality, with all animals succumbing within 14 days after birth (Brydges et al, 2009). Having established that the MWS‐associated NLRP3A350V mutant is active in LPS‐stimulated neutrophils, we next sought to examine the pathophysiological role of neutrophil‐restricted NLRP3A350V inflammasome activation in vivo. To this end, we bred mice that were homozygous for the _Nlrp3_A350V allele to transgenic mice that hemizygously express the neutrophil‐targeted MRP8/S100A8‐Cre recombinase (MRP8‐CreTg) (Abram et al, 2014). _Nlrp3_A350V/wtMRP8‐Cre− littermate control pups gained weight and flourished, whereas _Nlrp3_A350V/wtMRP8‐CreTg pups presented with a reduced body weight starting at postnatal day 4, and developed skin inflammatory abscesses and scaling erythema that were clearly visible by postnatal day 8 (Fig 1C). Unlike their _Nlrp3_A350V/wtMRP8‐Cre− littermate control pups, the survival rate of _Nlrp3_A350V/wtMRP8‐CreTg pups gradually declined in the first days after birth, with approximately 90% of _Nlrp3_A350V/wtMRP8‐CreTg pups having died within the first 2 weeks after birth (Fig 1D). Haematoxylin and eosin (H&E) staining of the spleens harvested at postnatal day 7 showed the typical lymphoid follicles and periarteriolar lymphoid sheaths (PALS) that compose the white pulp regions of the healthy spleens of _Nlrp3_A350V/wtMRP8‐Cre− control mice (Fig 1E, upper left). However, the white pulp regions were nearly absent in spleens of the transgenic pups expressing the _Nlrp3_A350V mutant allele (Fig 1E, upper middle, and Fig 1F). Instead, most of the spleen parenchyma in _Nlrp3_A350V/wtMRP8‐CreTg pups was replaced by a substantial increase in extramedullary haematopoiesis (EMH) of mostly myeloid lineage and to a more limited extent of erythrocyte lineage (Fig 1E, right and Fig 1G). Liver sections of both _Nlrp3_A350V/wtMRP8‐Cre− control mice and diseased _Nlrp3_A350V/wt MRP8‐CreTg pups displayed EMH of the erythroid lineage. Compared to the liver of _Nlrp3_A350V/wtMRP8‐Cre− littermate controls (Fig 1E, lower left), liver samples of _Nlrp3_A350V/wt MRP8‐CreTg pups also showed a modest increase in EMH of the myeloid lineage (indicated with arrowheads) and hepatocyte enlargement (Fig 1E, lower middle and right, and Fig 1H and I). In addition, mild focal or multifocal granulocytic infiltrates were present in the H&E‐stained heart sections of _Nlrp3_A350V/wtMRP8‐CreTg pups (Fig EV1A and B), whereas the small intestine and colon showed few or no abnormalities when compared to tissue sections of _Nlrp3_A350V/wtMRP8‐Cre− littermate controls (Fig EV1A). A multiplexed analysis of systemic cytokines and chemokines in serum of postnatal day 7 pups showed that _Nlrp3_A350V/wtMRP8‐CreTg pups had a significant elevation in the serum concentration of several inflammatory cytokines, growth factors and chemokines, among which the inflammasome‐dependent cytokine IL‐18, the inflammatory cytokines IL‐6 and TNF‐α and the chemokine MCP‐1 (Figs 1J–L and EV1C–Q). Together, these results show that NLRP3A350V inflammasome activation in neutrophils is sufficient to elicit systemic inflammatory CAPS pathology and tissue damage in vivo.

Canonical inflammasomes driving IL‐1β maturation and secretion in neutrophils

NLRP3 and several other key inflammasome components are expressed in resting neutrophils at the mRNA and protein levels (Mankan et al, 2012; Bakele et al, 2014). Several studies have shown that the canonical NLRP3 and NLRC4 inflammasomes in neutrophils promote the maturation and secretion of IL‐1β (Mankan et al, 2012; Chen et al, 2014; Karmakar et al, 2015, 2016, 2020; Perez‐Figueroa et al, 2016; Hassane et al, 2017; Monteleone et al, 2018). However, little is known about the functional activity of other canonical inflammasome pathways in neutrophils. Using flow cytometry, we first confirmed that our experimental procedure for negative selection of BMNs from C57BL/6 (B6) and inflammasome knockout mice resulted in a highly enriched population of CD11b+Ly6C+Ly6G+ neutrophils with a mean purity of 96% ± 2.1% (Fig EV2A–C). We next used the isolated neutrophil population to systematically assess which canonical inflammasome pathways can induce IL‐1β maturation and secretion in the culture media of LPS‐primed neutrophils, as has been extensively characterized in macrophages (Van Opdenbosch et al, 2017; de Vasconcelos et al, 2019).

To examine whether the NLRP1b inflammasome is functional in neutrophils, we isolated BMNs from transgenic C57BL/6J mice carrying a Bacillus anthracis lethal toxin (LeTx)‐responsive NLRP1b allele under the control of its endogenous promoter (hereinafter referred to as B6NLRP1b+) (Boyden & Dietrich, 2006). As expected, LPS priming alone failed to induce IL‐1β secretion from B6NLRP1b+ BMNs (Fig 2A). Unlike in B6 neutrophils, LeTx stimulated extracellular release of IL‐1β from B6NLRP1b+ neutrophils (Fig 2A), indicating that the NLRP1b inflammasome is functional in neutrophils. Western blot analysis confirmed that pro‐IL‐1β was converted into the mature cytokine in LeTx‐intoxicated B6NLRP1b+ neutrophils (Fig 2B). A LeTx‐sensitive NLRP1b allele is required because pro‐IL‐1β cleavage was not observed in B6 neutrophils (Fig 2B). Moreover, B6NLRP1b+ neutrophils that lack the expression of the inflammatory caspases 1 and 11 (B6NLRP1b+_C1_−/−_C11_−/−) failed to induce maturation and secretion of IL‐1β in response to LeTx stimulation (Fig 2A and B). Together, these results establish the requirement for a functional NLRP1b inflammasome to induce IL‐1β maturation and secretion from LeTx‐intoxicated neutrophils.

Mouse macrophages and human monocytes respond to RhoA GTPase inactivation by the Clostridium difficile toxin A (TcdA) with Pyrin inflammasome activation (Gao et al, 2016; Van Gorp et al, 2016). We show that TcdA also potently induced maturation and secretion of IL‐1β from murine neutrophils (Fig 2C and D). TcdA‐induced IL‐1β processing and secretion were abolished in _C1_−/−_C11_−/− neutrophils and in neutrophils isolated from _Mefv_−/− mice (which lack the inflammasome sensor Pyrin) (Fig 2C and D), establishing that the Pyrin inflammasome is functional in neutrophils.

Francisella tularensis activates the AIM2 inflammasome in macrophages (Fernandes‐Alnemri et al, 2010; Jones et al, 2010; Rathinam et al, 2010). We also observed IL‐1β maturation and secretion from neutrophils infected with Francisella tularensis (Fig 2E and F). Notably, IL‐1β maturation and secretion were fully abolished in infected _C1_−/−_C11_−/− neutrophils, whereas _Aim2_−/− neutrophils showed a trend of a partial decrease in extracellular IL‐1β levels compared to _F. tularensis_‐infected wildtype neutrophils (Fig 2E and F). These results suggest that the AIM2 inflammasome contributes to _F. tularensis_‐induced IL‐1β secretion from infected neutrophils, and that an additional inflammasome accounts for the residual IL‐1β secretion observed in _Aim2_‐deficient neutrophils.

Activation of the NLRP3 inflammasome was previously shown in nigericin‐ and streptolysin‐stimulated BMNs (Mankan et al, 2012; Karmakar et al, 2015, 2016; Hassane et al, 2017). We confirmed that extracellular ATP also engages the NLRP3 inflammasome for IL‐1β maturation and secretion in wildtype neutrophils, whereas these responses were abolished in neutrophils from _C1_−/−_C11_−/− and _Nlrp3_−/− mice (Fig EV3A and B).

Figure 2

Figure 2

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Canonical inflammasomes driving IL‐1β maturation and secretion in neutrophils

A–H LPS‐primed neutrophils of indicated genotypes were left untreated (Mock) or stimulated with LeTx (3 h) (A, B), TcdA (3 h) (C, D), infected with F. tularensis (18 h) (E, F) or stimulated with FlaTox (3 h) (G, H). IL‐1β secretion levels were determined in culture media (A, C, E, G); and combined cell lysates and culture media (Lys+Sup) were immunoblotted for cleavage of pro‐IL‐1β (~39 kDa) into mature IL‐1β (p17) (B, D, F, H). Cytokine values represent mean ± SEM of n = 3 biological repeats and immunoblots are representative of n = 3 biological repeats. Statistical significance was analysed by two‐way ANOVA: *P ≤ 0.05; **P ≤ 0.01. _P_‐values were corrected for multiple comparisons (Bonferroni).

The recombinant Legionella pneumophila flagellin‐derived NLRC4 stimulus FlaTox (de Vasconcelos et al, 2019) potently induced IL‐1β maturation and secretion from wildtype neutrophils, but not in _Nlrc4_−/− and _C1_−/−_C11_−/− neutrophils (Fig 2G and H). With the notable exception of TcdA, IL‐1β secretion induced by FlaTox and other inflammasome stimuli was substantially higher in macrophages than in neutrophils (Appendix Fig S1A–E). As reported (Chen et al, 2014), S. Typhimurium‐infected neutrophils also secreted significant levels of IL‐1β that were largely blunted in _Nlrc4_−/− and _C1_−/−_C11_−/− neutrophils (Fig EV3C and D). S. Typhimurium‐infected mouse neutrophils required caspase‐1, but not caspase‐11, for IL‐1β secretion (Fig EV3E). Human blood neutrophils infected with S. Typhimurium also secreted IL‐1β, and the caspase‐1 inhibitor VX‐765 markedly inhibited this response (Fig EV3F). Collectively, these results demonstrate that the canonical NLRP1b, Pyrin, AIM2, NLRP3 and NLRC4 inflammasome pathways are functional and promote the secretion of mature IL‐1β in neutrophils.

Canonical inflammasomes induce plasma membrane permeabilization in neutrophils

In macrophages, activation of canonical inflammasomes promotes pyroptotic cell lysis (Kayagaki et al, 2015; Shi et al, 2015; Vande Walle & Lamkanfi, 2016). To explore whether canonical inflammasome activation in neutrophils induces PMP and pyroptosis, we first measured incorporation of the cell‐impermeant DNA dye SYTOX Green by isolated neutrophils from wildtype, _C1_−/−_C11_−/− and relevant inflammasome sensor‐deficient mice over time. Consistent with neutrophils being short‐lived cells (8–20 h), we observed a spontaneous gradual increase in SYTOX Green incorporation in untreated (mock) neutrophils of all analysed genotypes (Fig EV4A–F), suggesting that spontaneous time‐dependent neutrophil cell death occurs through inflammasome‐independent mechanisms. Notably, LPS priming markedly diminished spontaneous background cell death levels in neutrophils (Fig EV4A–F), prompting us to examine canonical inflammasome‐induced SYTOX Green incorporation in LPS‐primed BMNs.

Upon LeTx‐induced activation of the NLRP1b inflammasome, B6NLRPb+ neutrophils displayed increased SYTOX Green incorporation over background levels at about 180 min post‐stimulation, which continued to rise over the next hours (Fig 3A). In contrast, SYTOX Green incorporation did not increase in concomitantly stimulated B6 and B6NLRP1b+_C1_−/−_C11_−/− neutrophils (Fig 3A), demonstrating that LeTx‐induced neutrophil PMP requires a functional NLRP1b inflammasome. Similarly, activation of the Pyrin inflammasome with TcdA selectively induced SYTOX Green incorporation in B6 BMNs, but not in _C1_−/−_C11_−/− and _Mefv_−/− neutrophils (Fig 3B); and ATP‐induced SYTOX Green staining in LPS‐primed neutrophils required caspase‐1 (Fig 3C). A functional NLRC4 inflammasome was also required to induce PMP in FlaTox‐stimulated neutrophils (Fig 3D). It should be noted that although these inflammasome stimuli induced consistent and time‐dependent SYTOX Green incorporation in neutrophils, the observed signal appeared less prominent than usually observed in macrophages (Appendix Fig S1F–I). Furthermore, Western blot analysis showed that the 37 kDa cytosolic protein marker glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) was released in culture media of FlaTox‐stimulated as well as in those of S. Typhimurium‐infected neutrophils (Fig 3E and F), indicating that NLRC4 inflammasome activation by these agents induces pyroptotic cell lysis and extracellular release of cytosolic proteins in neutrophils, in addition to promoting maturation and secretion of IL‐1β.

Figure 3

Figure 3

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Canonical inflammasomes promote plasma membrane permeabilization (PMP) in neutrophils

A–D LPS‐primed neutrophils of the indicated genotypes were left untreated (Mock) or stimulated with LeTx (A), TcdA (B), ATP (C) or FlaTox (D). PMP was assessed by measurement of SYTOX Green incorporation over time. Data information: Error bars represent mean ± SEM of n = 3 biological repeats. Statistical significance was analysed by two‐way ANOVA: *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. _P_‐values were corrected for multiple comparisons (Bonferroni).

E LPS‐primed B6 BMNs were left untreated (LPS) or subsequently stimulated with FlaTox (LPS+FlaTox) for 3 h. Data are representative of n = 3 biological repeats.

F B6 BMNs were left untreated (Mock) or infected with S. Typhimurium at MOI 25. Culture media (Sup) and whole‐cell lysates (Lys) were immunoblotted for the release of cytosolic GAPDH into culture media. Data are representative of n = 3 biological repeats.

GSDMD drives canonical inflammasome‐induced neutrophil cytotoxicity and IL‐1β secretion

Having established that canonical inflammasome activation in neutrophils induces cytotoxicity, we next sought to characterize the nature of this regulated cell death response. In macrophages, pyroptosis is driven by inflammatory caspase‐dependent cleavage of GSDMD and the formation of multimeric GSDMD pores that perforate the plasma membrane (Vande Walle & Lamkanfi, 2016). To examine the role of GSDMD in neutrophils, we first assessed GSDMD cleavage in response to a set of canonical inflammasome stimuli. A prototypic 30 kDa amino‐terminal GSDMD cleavage fragment was detected in immunoblots of TcdA‐ or FlaTox‐stimulated wildtype BMNs (Fig 4A and B). This immunoreactive band was not observed in lysates of TcdA‐ or FlaTox‐stimulated GSDMD‐deficient neutrophils (Fig 4A and B), demonstrating the specificity of these findings and indicating that Pyrin and NLRC4 inflammasome activation in neutrophils induce cleavage of GSDMD into its amino‐terminal pore‐forming domain. Robust GSDMD cleavage was also observed in immunoblots of LPS+nigericin‐stimulated wildtype neutrophils (Fig 4C). Caspase‐1 drove nigericin‐induced GSDMD maturation because GSDMD cleavage was abolished in caspase‐1‐deficient neutrophils (Fig 4C). Consistent with our previous results (Fig 3), TcdA, nigericin and FlaTox all induced SYTOX Green incorporation in LPS‐primed _Gsdmd_‐sufficient neutrophils (Fig 4D–F). However, SYTOX Green incorporation by these canonical inflammasome stimuli was blunted in _Gsdmd_‐deficient neutrophils (Fig 4D–F). Together, these results demonstrate that canonical inflammasome activation in neutrophils triggers caspase‐1‐mediated GSDMD cleavage and results in GSDMD‐dependent pyroptosis.

Figure 4

Figure 4

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

GSDMD drives canonical inflammasome‐induced PMP and IL‐1β secretion in neutrophils

A–I LPS‐primed neutrophils of the indicated genotypes were left untreated (Mock) or stimulated with TcdA (3 h) (A, D, G), FlaTox (3 h) (B, E, H) or nigericin (100 min) (C, F, I).

A–C Whole‐cell lysates (Lys) and culture media (Sup) were immunoblotted separately for cleavage of pro‐IL‐1β (~39 kDa) into mature IL‐1β (p17) and cleaved GSDMD (p30). * With Western blots indicates non‐specific cross‐reactivity of the antibody. Data are representative of n = 3 (A, B) or n = 2 (C) biological repeats.

D–F PMP was assessed by measurement of SYTOX Green incorporation over time. Data information: Values represent mean ± SEM of n = 3 (D, E) or n = 4 (F) biological repeats. Statistical significance was analysed by two‐way ANOVA: **P ≤ 0.01; ***P ≤ 0.001. _P_‐values were corrected for multiple comparisons (Bonferroni).

G–I IL‐1β secretion levels were determined by Luminex assay. Values represent mean ± SEM of n = 3 (G, I) or n = 4 (H) biological repeats. Statistical significance was analysed by two‐way ANOVA: *P ≤ 0.05; **P ≤ 0.01.

GSDMD activation by caspase‐1 represents a major mechanism of IL‐1β secretion from _ex‐vivo_‐stimulated macrophages, as well as in mice with hyperactivation of the canonical Pyrin and NLRP3 inflammasomes respectively (Kayagaki et al, 2015; Shi et al, 2015; Kanneganti et al, 2018; Xiao et al, 2018). To probe whether GSDMD promotes canonical inflammasome‐induced IL‐1β secretion from neutrophils, we measured the levels of IL‐1β in culture media of _Gsdmd_‐deficient neutrophils following canonical inflammasome stimulation. TcdA‐induced Pyrin inflammasome activation in wildtype neutrophils triggered robust secretion of IL‐1β, whereas culture media of _Gsdmd_−/− neutrophils contained significantly less extracellular IL‐1β (Fig 4G). Paralleling these results, IL‐1β secretion in _Gsdmd_−/− neutrophils was blunted in response to stimulation with FlaTox (Fig 4H) and nigericin (Fig 4I). Complementing these cytokine measurements, our immunoblotting analyses show that secretion of mature IL‐1β was blunted in culture media of stimulated _Gsdmd_−/− and _caspase‐1_−/− neutrophils (Fig 4A–C). Collectively, these results show that GSDMD activation in neutrophils promotes canonical inflammasome‐induced PMP and IL‐1β secretion.

Differential roles of GSDMD and NADPH oxidase in neutrophil pyroptosis and PMA‐induced NETosis

NETosis is a cell death mode in neutrophils that was first described to occur following stimulation with the phorbol ester PMA (Takei et al, 1996), and subsequently shown upon infection with certain microbial pathogens (Fuchs et al, 2007; Papayannopoulos, 2018). Consistent with PMA inducing NETosis independently of caspases (Takei et al, 1996), PMA failed to secrete IL‐1β in culture media of LPS‐primed murine BMNs (Figs 5A and EV5A). Furthermore, PMA‐induced NETosis was unaffected in caspase‐1‐deficient BMNs as assessed by confocal microscopy and kinetic analysis of SYTOX Green incorporation (Fig EV5B and C). Moreover, PMA‐induced SYTOX Green incorporation and extracellular LDH release from human neutrophils resisted inhibition by the caspase‐1 inhibitor VX‐765 (Fig EV5D and E). Neutrophil elastase‐mediated GSDMD cleavage was proposed to play a vital role in PMA‐induced NETosis (Sollberger et al, 2018). However, whereas canonical inflammasome stimuli induced prominent GSDMD cleavage in BMNs (Fig 4A–C), we failed to detect GSDMD cleavage in immunoblots of PMA‐stimulated wildtype and _caspase‐1_−/− BMNs (Fig 5A). Moreover, PMA‐induced SYTOX Green incorporation proceeded unabated in _Gsdmd_‐deficient neutrophils (Fig 5B), suggesting that GSDMD may not be required for PMA‐induced NETosis. As a comparison and consistent with PMA‐induced NETosis being driven by NADPH oxidase activity (Fuchs et al, 2007; Ermert et al, 2009), we confirmed that PMA‐induced NETosis was abolished in _Nox2/Gp91_Phox‐deficient neutrophils as assessed by confocal microscopy and kinetic analysis of SYTOX Green incorporation (Appendix Fig S2A–C). As reported (Vong et al, 2014), culture media of PMA‐stimulated BMNs failed to produce a signal in LDH activity assays (Appendix Fig S2D), suggesting that LDH activity assays may not be a suitable readout for neutrophil lysis. However, immunoblotting showed that culture media of PMA‐stimulated BMNs contained levels of extracellular β‐actin and GAPDH that were significantly lower than even the basal background levels of mock‐treated BMNs (Fig 5C). These marked reductions were also observed with PMA‐stimulated BMNs from mice lacking expression of caspase‐1 or GSDMD (Fig 5C and D), indicating that the mechanism involved is not regulated by GSDMD or canonical inflammasomes. Unlike PMA, nigericin stimulation potently increased extracellular concentrations of β‐actin and GAPDH in culture media of LPS‐primed BMNs (Fig 5E). Further support for the notion that nigericin induces canonical inflammasome‐induced pyroptosis in neutrophils is provided by the findings that nigericin‐induced extracellular release of GAPDH and β‐actin (Fig 5E and F) and nigericin‐induced SYTOX Green incorporation (Figs 4F and 5G) were abolished in caspase‐1‐ and GSDMD‐deficient BMNs. Similarly, we confirmed that GSDMD‐mediated PMP in TcdA‐stimulated neutrophils (Fig 4D) was accompanied by GSDMD‐dependent extracellular GAPDH and β‐actin release (Fig 5H). Together, these results indicate that several canonical inflammasome stimuli induce GSDMD‐dependent pyroptotic cell lysis in neutrophils.

Figure 5

Figure 5

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

PMA‐induced NETosis requires NADPH oxidase activity, but not GSDMD

A–H LPS‐primed neutrophils of the indicated genotypes were left untreated (Mock), stimulated with PMA (3 h) (A–D), nigericin (100 min) (E–G) or TcDA (3 h) (H). Whole‐cell lysates (Lys) and culture media (Sup) were immunoblotted for cleavage of pro‐IL‐1β (~39 kDa) into mature IL‐1β (p17) and cleaved GSDMD (p30) (A), and extracellular release of β‐actin and GAPDH (C–F, H). * With Western blots indicates non‐specific bands. (B, G) PMP was assessed by measurement of SYTOX Green incorporation over time. Values represent mean ± SEM of n = 3 biological repeats. Statistical significance was analysed by two‐way ANOVA: ***P ≤ 0.001; **P ≤ 0.01; ns = non‐significant. _P_‐values were corrected for multiple comparisons (Bonferroni).

We next sought to further explore the cell death mechanism of S. Typhimurium‐infected neutrophils (Fig 3F). S. Typhimurium‐infected wildtype neutrophils displayed a prominent induction of SYTOX Green incorporation of about 20–30% above background levels by 3 h post‐infection (Appendix Fig S3A). S. Typhimurium‐induced PMP was abolished in infected _C1_−/−_C11_−/− and _Nlrc4_−/− neutrophils (Appendix Fig S3A), suggesting a key role for the canonical NLRC4 inflammasome. In agreement, caspase‐1 or GSDMD deficiency was sufficient to blunt S. Typhimurium‐induced SYTOX Green incorporation (Appendix Fig S3B and C). Confocal micrographs of S. Typhimurium‐infected wildtype and caspase‐1‐deficient BMNs confirmed these findings (Fig 6A). Moreover, S. Typhimurium‐infected neutrophils generally had a less rounded appearance compared to the more ballooning morphology of PMA‐treated BMNs (Fig 6B and Appendix Fig S2B). Immunoblotting analysis showed GSDMD cleavage and release of mature IL‐1β in culture media of wildtype BMNs, but not in caspase‐1‐deficient neutrophils that have been infected with S. Typhimurium for 1 or 3 h (Fig 6C). Moreover, S. Typhimurium‐infected wildtype neutrophils had released considerable amounts of GAPDH and β‐actin in their culture media by 3 h post‐infection, whereas infected caspase‐1‐deficient neutrophils failed to do so (Fig 6D). Although caspase‐1‐independent LDH release from S. Typhimurium‐infected neutrophils may occur under different experimental conditions, we found that S. Typhimurium‐induced LDH release was blunted in caspase‐1‐deficient BMNs (Fig 6E). IL‐1β secretion was also abolished in culture media of GSDMD‐deficient BMNs (Fig 6F), consistent with GSDMD promoting IL‐1β secretion downstream of caspase‐1 in S. Typhimurium‐infected BMNs. Of note, a time kinetic analysis of IL‐1β secretion showed that a lower S. Typhimurium infection load (MOI 10) delayed IL‐1β secretion without affecting its reliance on NLRC4 and GSDMD (Appendix Fig S3D). Neutrophil elastase was previously shown to cleave GSDMD, and neutrophil serine proteases were also suggested to directly cleave IL‐1β under certain conditions (Lindemann et al, 1999; Sollberger et al, 2018). However, unlike NLRC4, caspase‐1 and GSDMD, we found that the NETosis mediators neutrophil elastase, neutrophil proteinase 3, cathepsin G and protein arginine deiminase 4 (PAD4) were all dispensable for S. Typhimurium‐induced pyroptosis (Appendix Fig S3E and F). Moreover, whereas NADPH oxidase activity was required for PMA‐induced NETosis (Appendix Fig S2A and C) (Ermert et al, 2009), S. Typhimurium‐induced PMP was significantly increased and kinetically accelerated in _Nox2/Gp91_Phox‐deficient neutrophils (Fig 6G and H). This suggests that NADPH oxidase‐produced reactive oxygen species promote PMA‐induced NETosis but suppress S. Typhimurium‐induced pyroptosis in neutrophils. Collectively, these results indicate that caspase‐1‐mediated GSDMD maturation drives S. Typhimurium‐induced neutrophil pyroptosis, which is mechanistically distinct from PMA‐induced NETosis.

Figure 6

Figure 6

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Caspase‐1 and GSDMD drive S. Typhimurium‐induced neutrophil pyroptosis

A–H Neutrophils of the indicated genotypes were left untreated (Mock) or infected with S. Typhimurium at MOI 25. (A) Representative confocal micrographs taken at different time points were stained with DAPI (dark blue nuclei) and SYTOX Green (light blue). Scale bar: 20 μM. (B) Representative confocal micrographs of untreated (Mock), S. Typhimurium‐infected or PMA‐stimulated wildtype BMNs stained for neutrophil elastase (Red) and SYTOX Green (Green) (Scale bar: 10 μM). (C, D) Cell lysates (Lys) and culture media (Sup) were immunoblotted for the indicated proteins at 1 and 3 h post‐infection. (E) Cell death was measured by LDH assay 3 h post‐infection. (F) IL‐1β secretion levels were determined by Luminex assay 1 h post‐infection. (G) Representative confocal micrographs of neutrophils stained with DAPI (dark‐blue nuclei) and SYTOX Green (light blue) taken at different time points post‐infection (scale bar: 20 μM). (H) PMP was assessed by measurement of SYTOX Green incorporation over time.

Data information: In E, F and H, values represent mean ± SEM of n = 3 biological repeats. Statistical significance was analysed by two‐way ANOVA: **P ≤ 0.01; ***P ≤ 0.001. _P_‐values were corrected for multiple comparisons (Bonferroni). Other data are representative of n = 3 (A, B, G) or n = 2 (C, D) biological repeats. S. Typh.: S. Typhimurium.

Discussion

Inflammasome activation is a key component of the innate immune response that contributes critically to pathogen control and clearance, as well as to detrimental inflammation and immune‐mediated tissue damage in autoinflammatory and autoimmune diseases (Lamkanfi & Dixit, 2014; Van Gorp et al, 2019; Van Opdenbosch & Lamkanfi, 2019). The repertoire and signalling mechanisms of canonical inflammasomes have been extensively characterized in macrophages and monocytes. Recent studies showed that caspase‐11 activation in neutrophils may elicit GSDMD‐dependent NETosis in response to LPS transfection, infection with flagellin‐deficient S. Typhimurium mutants or upon Citrobacter rodentium infection (Chen et al, 2018). Moreover, inflammasome‐independent (but neutrophil elastase‐mediated) cleavage of GSDMD was suggested to promote PMA‐induced NET formation (Sollberger et al, 2018). Other studies have shown that the canonical NLRP3 inflammasome promotes IL‐1β secretion from BMNs stimulated with the NLRP3 agonists ATP and nigericin (Mankan et al, 2012; Karmakar et al, 2016, 2020), and from lung neutrophils of mice that had been infected with the respiratory bacterial pathogen Streptococcus pneumoniae (Karmakar et al, 2015; Hassane et al, 2017). NLRC4 inflammasome activation in neutrophils by S. Typhimurium infection was also shown to promote maturation and secretion of IL‐1β (Chen et al, 2014, 2018; Monteleone et al, 2018); and NLRC4 activation in neutrophils in vivo was shown to drive systemic inflammatory disease (Kitamura et al, 2014; Nichols et al, 2017). However, a systematic analysis of the repertoire of functional canonical inflammasomes beyond NLRP3 and NLRC4 in neutrophils has not been reported. Moreover, the role of GSDMD in the context of neutrophil canonical inflammasome activation and PMA‐induced NETosis is unclear.

Here, we demonstrated the in vivo relevance of neutrophil‐intrinsic inflammasome activation by showing that neutrophil‐targeted activation of the NLRP3 inflammasome in a mouse model of CAPS disease suffices to recapitulate the previously reported clinical phenotype of mice that ubiquitously express the MWS‐associated NLRP3A350V mutant (Brydges et al, 2009). Our findings show that neutrophil‐intrinsic NLRP3 inflammasome activation is an important driver of auto‐inflammatory tissue damage and inflammatory pathology, highlighting that an in‐depth analysis of inflammasome signalling in neutrophils is warranted. During the preparation of this manuscript, a complementary report was published (Stackowicz et al, 2021) showing that neutrophils are an important source of IL‐1β in the skin of NLRP3A350V mutant mice and that neutrophil‐intrinsic NLRP3A350V mutant expression promotes CAPS in vivo. Our studies extend this report by demonstrating that LPS stimulation is sufficient to induce the extracellular release of mature IL‐1β and caspase‐1 from neutrophils that express the NLRP3A350V mutant. Moreover, we performed an extensive histological analysis showing that neutrophil‐intrinsic NLRP3A350V mutant expression triggers a substantial increase in extramedullary haematopoiesis in the spleen and liver, along with mild focal or multifocal granulocytic infiltrates in the heart, and increased systemic levels of a broad suite of inflammatory cytokines and chemokines.

Previous studies have shown that the canonical NLRP3 and NLRC4 inflammasomes in neutrophils promote the maturation and secretion of IL‐1β (Mankan et al, 2012; Chen et al, 2014; Karmakar et al, 2015, 2016; Perez‐Figueroa et al, 2016; Hassane et al, 2017). We further established the repertoire of inflammasome sensors that is active in neutrophils by systematically profiling IL‐1β maturation and secretion in wildtype and inflammasome‐deficient neutrophils. These analyses provided unambiguous genetic evidence that in addition to NLRC4 and NLRP3, also the canonical NLRP1b, Pyrin and AIM2 inflammasomes are fully functional and capable of promoting maturation and secretion of IL‐1β from _ex vivo_‐stimulated neutrophils. We also showed that canonical inflammasome activation induces caspase‐1‐dependent GSDMD cleavage in neutrophils. Furthermore, the presented experiments in neutrophils of caspase‐1‐deficient and _Gsdmd_−/− mice provided clear genetic evidence that canonical inflammasome activation promotes GSDMD‐dependent PMP, IL‐1β secretion and extracellular release of intracellular proteins such as GAPDH and β‐actin from neutrophils. A previous report relied on an LDH activity assay to suggest that caspase‐1 does not induce pyroptosis in neutrophils (Chen et al, 2018). Here, we assessed neutrophil cell death markers by relying on a combination of sensitive kinetic SYTOX Green incorporation assays, LDH activity assay and immunoblotting of conditioned culture media to assess extracellular release of GAPDH, β‐actin and mature IL‐1β and caspase‐1 from wildtype and inflammasome‐deficient neutrophils. These experiments clearly identified GSDMD activation by canonical inflammasomes as a cytotoxic mechanism inducing neutrophil pyroptosis.

Unlike with inflammasome‐activating stimuli, we showed that classical NETosis induced by the phorbol ester PMA occurs in the absence of GSDMD cleavage. Moreover, PMA‐induced PMP was unaffected in _Gsdmd_‐deficient neutrophils, suggesting that GSDMD may be dispensable for PMA‐induced NETosis. GSDMD has been proposed to mediate PMA‐induced NETosis based on the pharmacological characterization of LDC7559, a compound that was first suggested to inhibit PMA‐induced NETosis by targeting GSDMD (Sollberger et al, 2018). However, LDC7559 was subsequently shown not to target GSDMD but to inhibit PMA‐induced NETosis by acting as a potent agonist of the glycolytic enzyme phosphofructokinase‐1 liver type (PFKL) (Amara et al, 2021). Together with our findings, this implies that the role of GSDMD in neutrophil cell death is confined to inflammasome‐induced cues. Conversely, we showed that deficiency in NADPH oxidase—which is critical for PMA‐induced NETosis (Fuchs et al, 2007; Ermert et al, 2009)—increased the sensitivity of _Nox2/gp91_phox‐deficient neutrophils to S. Typhimurium‐induced pyroptosis, further demonstrating that phorbol esters and inflammasome activation induce neutrophil cell death through mechanistically distinct mechanisms. These findings are consistent with a previous report showing that also _Pseudomonas‐_infected neutrophils undergo caspase‐1‐mediated pyroptosis, which is increased in a _Nox2/gp91_phox‐deficient background (Ryu et al, 2017). We hypothesize that NADPH‐induced reactive oxygen species (ROS) in wildtype neutrophils may target caspase‐1 for oxidation and glutathionylation to suppress caspase‐1 activation, as has been reported in SOD1‐deficient macrophages (Meissner et al, 2008). This suppressive environment may be alleviated in _Nox2/gp91_phox‐deficient neutrophils, giving rise to higher caspase‐1 activation levels in an NADPH‐deficient genetic background. Work is ongoing in our laboratories to test this hypothesis.

Relative to macrophages, neutrophils on a per‐cell basis appear less potent in inducing inflammasome‐mediated IL‐1β secretion and pyroptosis. However, as neutrophils can fast and massively infiltrate inflammatory tissues, their impact on pathophysiological outcomes is not to be underestimated. The notion that inflammasome activation in a small subset of monocytes and granulocytes suffices for clinical manifestation of inflammatory disease is highlighted by the discovery of low‐level mosaicism in a subset of CAPS patients. According to initial reports (Rowczenio et al, 2017; Labrousse et al, 2018), mosaic CAPS patients can present with a full‐blown clinical phenotype despite having as few as 8–12% of their monocyte and granulocyte population expressing a mutant NLRP3 allele.

In conclusion, we showed that neutrophil‐targeted expression of the disease‐associated NLRP3A350V mutant triggers systemic autoinflammatory disease and tissue pathology in mice, suggesting that NLRP3 inflammasome activation in neutrophils may contribute importantly to autoinflammatory disease pathology. Furthermore, by establishing the repertoire of canonical inflammasomes that is functional in neutrophils, and by showing that neutrophil‐intrinsic GSDMD and NADPH oxidase activity are differential effectors of canonical inflammasome‐induced pyroptosis and mitogen‐induced NETosis, this work reveals new mechanisms by which neutrophils contribute to anti‐microbial host defence and inflammasome‐associated diseases.

Materials and Methods

Bacteria and stimuli

A colony from a blood agar plate of Salmonella enterica serovar Typhimurium SL 1344 (S. Typhimurium) and a glycerol stock of Francisella tularensis ssp. novicida strain U112 (F. tularensis) were cultured overnight at 37°C in aerobic conditions in Luria broth or trypticase soy broth supplemented with 0.2% L‐cysteine respectively. On the experimental day, cultures were diluted in a fresh medium and grown to reach an OD600 of 0.7–0.8 before dilution in cell culture medium for infection. Phorbol 12‐myristate 13‐acetate (PMA), tamoxifen and corn oil were purchased from Sigma; ultrapure LPS from Salmonella minnesota was from InvivoGen; tamoxifen‐containing diet from Envigo RMS; adenosine‐5′‐triphosphate (ATP) from Roche; Clostridium difficile toxin A (TcdA) from Enzo Life Sciences; and recombinant anthrax lethal factor (LF) from Quadratech. Recombinant expression and purification of B. anthracis protective antigen (PA) and recombinant LFn‐FlaA were performed as described in von Moltke et al. (2012).

Mice

_C1_−/−_C11_−/− (Kuida et al, 1995), _Mefv_−/− (Van Gorp et al, 2016), _Aim2_−/− (Jones et al, 2010), B6Nlrp1b+ (Boyden & Dietrich, 2006), B6Nlrp1b+_C1_−/−_C11_−/− (Van Opdenbosch et al, 2014), _C1_−/− (Van Gorp et al, 2016), _C11_−/− (Kayagaki et al, 2011), _Nlrp3_−/− (Mariathasan et al, 2006), _Nlrc4_−/− (Mariathasan et al, 2004), Pad4_−/− and NE_−/−_CatG_−/−_Pr3_−/− (Yan et al, 2016) have been described. Nox2 y/− (B6.129S‐_Cybb tm1Din/J) (Pollock et al, 1995), Nlrp3 A350V/wt (B6.129‐Nlrp3tm1Hhf/J) (Brydges et al, 2009), MRP8‐Cre (B6.Cg‐Tg(S100A8‐cre,‐EGFP)1Ilw/J) (Passegue et al, 2004) and CreT (B6.129‐_Gt(ROSA)26Sor tm1(cre/ERT2)Tyj/J) (Ventura et al, 2007) mice on a C57BL/6 background were originally obtained from the Jackson Laboratories and bred in‐house. Cre expression in male and female Nlrp3 A350V/wt CreT + mice was induced by intragastric administration of tamoxifen (5 mg dissolved in 100 μl corn oil) on days 1 and 2. On day 2, mice were given tamoxifen‐containing diet ad libitum for 6 days; on day 8, mice were sacrificed for bone marrow neutrophils (BMN) isolation, regardless of the gender. In experiments involving BMN isolation from different mouse lines, adult animals were sacrificed regardless of gender or age. All animals were bred in‐house in individually ventilated cages under specific pathogen‐free conditions, and all studies were conducted under protocols approved by the Ghent University Committee on the Use and Care of Animals (ECD 14/92 and EC2017/090). Neither randomization nor blinding was applied for animal studies.

Histology and histologic grading

Pups were sacrificed at postnatal day 7, and tissues were harvested and fixed in 10% neutral buffered formalin and embedded in paraffin. The specimens were cut transversally at 5 μM on a microtome and stained with haematoxylin and eosin (H&E). Tissues were assessed by a board‐certified pathologist and findings were given severity grades 0–5, where 0 represents the absence of the finding, 1 is minimal, 2 is mild, 3 is moderate, 4 is marked and 5 is severe.

Cell culture and inflammasome activation assays

BMNs were purified from 6‐ to 12‐weeks‐old male and female mice of the indicated genotypes by negative MACS selection with the Neutrophil Isolation Kit (Miltenyi Biotec) according to the manufacturer's instructions. Upon purification, BMNs were directly seeded in Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% foetal bovine serum (FBS) and 1% non‐essential amino acids (Gibco). In other experiments, BMNs were seeded in serum‐free OPTI‐MEM medium (Gibco), and for PMA stimulations, cells were seeded in Hanks' Balanced Salt Solution (HBSS) supplemented with 1% fetal bovine serum (FBS). BMNs were seeded at a density of 8 × 105 cells/ml for cell death assays, and at a density of 3.2 × 106 cells/ml for cytokine analysis and confocal microscopy. BMNs were primed with 250 ng/ml LPS for 2 h prior to treatment with LeTx (= PA (2.5 μg/ml) and LF (2.5 μg/ml)), FlaTox (= PA (2.5 μg/ml) and LF‐Flagellin (2.5 μg/ml)), ATP (5 mM), TcdA (5 μg/ml), nigericin (6.7 μM) and PMA (100 nM). There was no priming for BMNs infected with S. Typhimurium or F. tularensis at MOI 25. Gentamycin (10 µg/ml) was added 1 h post‐infection with S. Typhimurium and 2 h post‐infection with F. tularensis. After incubation, cell‐free supernatants were harvested for cytokine quantification and cell extracts were prepared for Western analysis. For experiments involving bone marrow‐derived macrophages, bone marrow cells were cultured in L929‐cell‐conditioned IMDM supplemented with 10% FBS, 1% non‐essential amino acids and 1% penicillin–streptomycin for 6 days in a humidified atmosphere containing 5% CO2 before cells were detached and seeded for experiments.

Flow cytometry

Fluorescence‐activated cell sorter (FACS) buffer consisted of PBS (Gibco) supplemented with 2% FBS and 2 mM EDTA (Sigma‐Aldrich); the samples were stained with the following antibodies: CD11b BV605 (Biolegend), Ly6G Per‐CP Cy5.5 (BD Pharmingen) and Ly6C FITC (BD Pharmingen). Cell viability was assessed by incubating the samples with the Zombie Violet™ Fixable Viability Kit (Biolegend). FACS data were obtained using a BD LSRFortessa (BD Biosciences) and analysed with FlowJo.

Cytokine analysis

Cytokine levels in the cell culture medium were determined by a magnetic bead‐based multiplex assay using Luminex technology (Bio‐Rad) according to the manufacturer's instructions.

Western blotting

Unless stated otherwise, all Western blot samples were prepared after 3 h of stimulations. Cell lysates and culture media were incubated in cell lysis buffer (20 mM Tris HCl [pH 7.4], 200 mM NaCl and 1% NP‐40) and denatured in Laemmli buffer for combined lysate and supernatant samples. Samples for individual lysates and supernatants were prepared as follows: Neutrophils were plated in 24 well plate and stimulated in 400 μl of medium (with 1% FBS). After stimulation, supernatants were gently collected in 1.5 ml Eppendorf and centrifuged for 500 g for 5 min to remove residual cells and then transferred into fresh tubes. It was followed by precipitation by methanol–chloroform extraction, as described in Jakobs et al. (2013). Pellet was resuspended in 90 μl of Laemmli buffer. After removal of supernatants, lysates were prepared by direct lysing of cells in 120 μl Laemmli buffer on the plate. Protein samples were boiled at 95°C for 10 min and separated by SDS–PAGE. Separated proteins were transferred to nitrocellulose membranes. Blocking, incubation with antibody and washing of the membrane were done in PBS supplemented with 0.05% Tween 20 (v/v) and 3% (w/v) non‐fat dry milk. Immunoblots were incubated overnight with primary antibody against IL‐1β (GTX74034, GeneTex), GAPDH (clone 14C10, 2118, Cell Signalling), GSDMD (ab209845 (cleaved) or ab219800 (full length), Abcam) and Caspase‐1 (Casper‐1) (Adipogen, AG‐20B‐0042‐C100), β‐actin‐HRP (Santa Cruz, sc‐47778‐HRP). Horseradish peroxidase‐conjugated goat anti‐rabbit (111–035‐144, Jackson ImmunoResearch Laboratories) and anti‐mouse IgG‐HRP (CST, #7076) secondary antibody were used to detect proteins by enhanced chemiluminescence (ThermoScientific). Some immunoblots were contrast enhanced in a linear fashion applied equally across the entire image.

Cell death assays

A plate‐based fluorometric assay (FLUOstar, Omega BMG Labtech) was used to quantify cell permeabilization (2.5 mM SYTOX Green). Briefly, stimulated cells were incubated in the presence of SYTOX Green in a CO2‐ and temperature‐controlled environment that allowed measurement of fluorescence over time. A Triton X‐100 incubated well served as a maximum fluorescent signal. The ratio of fluorescence over maximum fluorescence was used as a measure for cell death. Alternatively, a Clariostar plate reader equipped with a 37°C incubator or the IncuCyte Zoom system (Essenbio) were used to acquire and analyse cell death rates over a defined period of time. For LDH activity assays, at least 0.6 million cells were plated in 24 well plates and the assay was performed according to manufacturer's guidelines (CyQUANT LDH Cytotoxicity Assay, Thermo #C20300). Values were normalized to triton X as 100% lysis control.

Confocal microscopy

BMNs were seeded and imaged in a μ‐Slide 8‐well Glass Bottom chamber (ibidi). In some experiments, cells were incubated in the presence of Hoechst and SYTOX Green in a CO2‐ and temperature‐controlled environment that allowed the measurement of fluorescent signals over time. Cell death was induced just before imaging and was monitored by the increase in SYTOX Green positivity. Live‐cell imaging was performed on an Axio Observer Z1 (Zeiss, Germany) equipped with a CSU‐X1 spinning‐disk head (Yokogawa) and AxioCam MRm (Zeiss), with a Plan‐Apochromat 40X/1.4 oil immersion objective. Images were acquired every 10 min for 6 h. Alternatively, cells were fixed in 4% PFA, permeabilized with 0.2% Triton X‐100, blocked with 5% BSA in PBS and stained for neutrophil elastase (Abcam, cat. Ab68672) with a Dylight 650‐conjugated donkey anti‐rabbit secondary antibody (Thermo‐Fisher, cat. SA5‐10041) and SYTOX Green. Data analysis and image reconstruction were performed with ImageJ (NIH).

Human blood neutrophil isolation

Neutrophils were isolated from healthy consenting donors obtained from the Belgian Red Cross as previously described (Desai et al, 2016). In short, fresh blood was gently mixed in equal volume with PBS containing 3% dextran (MP Biomedicals, cat. 101514) and left for 40 min to incubate at room temperature (RT). Light‐orangish upper layer mixture was transferred to a fresh 50 ml falcon, followed by centrifugation on 1,500 rpm at RT for 5 min. The resulting cell pellet was resuspended in 30 ml PBS followed by the addition of 10–12 ml of Ficoll (GE healthcare, Cat. 17‐1440‐02) by putting a pipette at the bottom of the falcon and was allowed to settle. The mixture was centrifuged at 2,000 rpm at RT for 20 min without brakes. Next, PBMCs and Ficoll mixture were gently removed, and the cell pellet was resuspended in ice‐cold 36 ml of water for 15–20 s before mixing 4 ml of 10X PBS. Cells were centrifuged at 1,800 rpm at 4°C for 5 min. The resulting neutrophils were resuspended in plain RPMI medium (without supplement).

Human neutrophil stimulation

Cells were seeded at a density of 3 × 105 cells/ml for cytokine analysis. All conditions were primed with 100 ng/ml of LPS for 2 h followed by the addition of VX‐765 or blank medium in the last 30 min of priming. Cells were infected S. Typhimurium at indicated MOI for 1 h followed by the addition of gentamycin (10 μg/ml), and incubated for the indicated time periods.

Statistics

All ex vivo experiments were run with at least two biological replicates and included additional technical replicates. GraphPad Prism 7.0 software was used for data analysis. Data are represented as mean with SEM. For mouse survival curves, statistical significance was determined by log‐rank (Mantel–Cox) test. For serum cytokine analysis, data were analysed using the Welch's _t_‐test. For histological damage score, statistical analysis was determined by Mann–Whitney _U_‐test. Cytokine analysis in supernatants of different mouse genotypes and cell death kinetics was analysed by using two‐way ANOVA with Bonferroni's multiple‐comparison test. Human cytokines analysis was performed using ordinary one‐way ANOVA with Bonferroni's multiple‐comparison test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns = non‐significant.

Data availability

All data to understand and assess the conclusions of this research are available in the main text and Supplementary Materials. No primary datasets have been generated and deposited.

References

Download references

Acknowledgements

We thank the VIB Bioimaging Core Facility for technical support. We thank Vishva M. Dixit (Genentech), Massimiliano Mazzone (VIB and KU Leuven) and Maria Walter and Katrin Schröder (Goethe‐University) for mutant mice. This work was supported by the Research Foundation Flanders (FWO) grant G011315N, and European Research Council Grant 683144 (PyroPop) to ML.

Author information

Author notes

  1. These authors contributed equally to this work.

Authors and Affiliations

  1. Janssen Immunosciences, World Without Disease Accelerator, Pharmaceutical Companies of Johnson & Johnson, Beerse, Belgium
    Dhruv Chauhan, Annalisa Zecchin, Nina Van Opdenbosch & Filip Van Hauwermeiren
  2. Department of Internal Medicine and Paediatrics, Ghent University, Ghent, Belgium
    Dieter Demon, Lieselotte Vande Walle, Oonagh Paerewijck, Silvia Ribo, Amelie Fossoul, Hanne Van Gorp, Nina Van Opdenbosch, Filip Van Hauwermeiren, Andy Wullaert & Mohamed Lamkanfi
  3. VIB‐UGent Center for Inflammation Research, VIB, Ghent, Belgium
    Dieter Demon, Lieselotte Vande Walle, Silvia Ribo, Amelie Fossoul, Amanda Gonçalves, Hanne Van Gorp, Nina Van Opdenbosch, Filip Van Hauwermeiren & Andy Wullaert
  4. Nonclinical Safety, Janssen Research & Development, Pharmaceutical Companies of Johnson & Johnson, Beerse, Belgium
    Leslie Bosseler
  5. Institute of Pharmacology and Structural Biology (IPBS), University of Toulouse, CNRS, Toulouse, France
    Karin Santoni, Rémi Planès & Etienne Meunier
  6. VIB BioImaging Core, Ghent, Belgium
    Amanda Gonçalves
  7. Laboratory of Protein Chemistry, Proteomics and Epigenetic Signalling, Department of Biomedical Sciences, University of Antwerp, Antwerp, Belgium
    Andy Wullaert

Authors

  1. Dhruv Chauhan
  2. Dieter Demon
  3. Lieselotte Vande Walle
  4. Oonagh Paerewijck
  5. Annalisa Zecchin
  6. Leslie Bosseler
  7. Karin Santoni
  8. Rémi Planès
  9. Silvia Ribo
  10. Amelie Fossoul
  11. Amanda Gonçalves
  12. Hanne Van Gorp
  13. Nina Van Opdenbosch
  14. Filip Van Hauwermeiren
  15. Etienne Meunier
  16. Andy Wullaert
  17. Mohamed Lamkanfi

Contributions

Dhruv Chauhan: Conceptualization; data curation; formal analysis; validation; investigation; visualization; methodology; writing – original draft; project administration; writing – review and editing. Dieter Demon: Conceptualization; data curation; formal analysis; validation; investigation; visualization; methodology; writing – original draft; project administration. Lieselotte Vande Walle: Formal analysis; investigation; visualization; methodology. Oonagh Paerewijck: Formal analysis; validation; investigation; visualization; methodology. Annalisa Zecchin: Conceptualization; data curation; formal analysis; validation; investigation; visualization; methodology; writing – original draft; writing – review and editing. Leslie Bosseler: Formal analysis; validation; investigation; visualization; methodology. Karin Santoni: Conceptualization; formal analysis; validation; investigation; visualization; methodology. Rémi Planès: Conceptualization; formal analysis; validation; investigation; visualization; methodology. Silvia Ribo: Conceptualization; formal analysis; validation; investigation; visualization; methodology. Amelie Fossoul: Investigation; methodology. Amanda Gonçalves: Investigation; visualization; methodology. Hanne Van Gorp: Formal analysis; investigation; visualization; methodology. Nina Van Opdenbosch: Formal analysis; investigation; visualization; methodology. Filip Van Hauwermeiren: Formal analysis; investigation; visualization; methodology. Etienne Meunier: Conceptualization; resources; formal analysis; supervision; funding acquisition; investigation; methodology. Andy Wullaert: Conceptualization; resources; formal analysis; supervision. Mohamed Lamkanfi: Conceptualization; resources; supervision; funding acquisition; validation; methodology; writing – original draft; project administration; writing – review and editing.

Corresponding author

Correspondence toMohamed Lamkanfi.

Ethics declarations

DC, AZ, LB, NVO and FVH are employees of Janssen Pharmaceutica. ML is an editorial advisory board member of EMBO Reports. This has no bearing on the editorial consideration of this article for publication. The other authors declare that they have no competing interests.

Additional information

EMBO reports (2022) 23: e54277

Supplementary Information

Appendix (download PDF )

Expanded View Figures PDF (download PDF )

Source Data for Expanded View (download ZIP )

Review Process File (download PDF )

PDF+ (download PDF )

Source Data for Figure 1 (download PDF )

Source Data for Figure 2 (download PDF )

Source Data for Figure 3 (download PDF )

Source Data for Figure 4 (download PDF )

Source Data for Figure 5 (download PDF )

Source Data for Figure 6 (download PDF )

EV1 (download PNG )

Figure

A Representative H&E‐stained sections of the heart (scale bar: 200 μM), small intestine (scale bar: 200 μM) and colon (scale bar: 100 μM) of _Nlrp3_A350V/wt MRP8‐CreTg and littermate _Nlrp3_A350V/wt MRP8‐Cre‐ mice. A mild EMH (arrows) is present in the heart sections of _Nlrp3_A350V/wt MRP8‐CreTg mice.

B An expert pathologist graded the presence of EMH in H&E‐stained histological sections of the heart of _Nlrp3_A350V/wt MRP8‐Cre‐ (Cre‐, n = 8) and _Nlrp3_A350V/wt MRP8‐CreTg (CreTg, n = 6) pups using a score of 0–5. Values represent mean ± SEM of n = 6–8 biological repeats. Statistical analysis of histological damage score was determined by Mann–Whitney _U_‐test. ***P < 0.001.

C–Q Cytokine and chemokine concentrations in serum samples obtained on day 7 from _Nlrp3_A350V/wt MRP8‐Cre‐ (Cre‐, n = 8) and _Nlrp3_A350V/wt MRP8‐CreTg (CreTg, n = 6) pups. Values represent mean ± SEM of biological repeats. Statistical analysis of serum cytokine levels was performed using Welch's _t_‐test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

EV2 (download PNG )

Figure

A–C Representative flow cytometry plots of neutrophils isolated from the bone marrow of C57Bl6/N (A), B6Nlrp1b+ (B) and AIM2 KO (C) mice. Left plots: identification of CD11b+ enriched population; right panels: Ly6G+/Ly6C+ neutrophils on gated CD11b+ cells. Data are representative of n = 3 biological repeats.

EV3 (download PNG )

Figure

A, B LPS‐primed neutrophils of the indicated genotypes were left untreated (Mock) or stimulated with ATP (1 h). IL‐1β secretion levels were determined in culture media (A); and combined cell lysates and culture media (Lys+Sup) were immunoblotted for cleavage of pro‐IL‐1β (~39 kDa) into mature IL‐1β (p17) (B). Cytokine values represent mean ± SEM of n = 3 biological repeats. Statistical significance was determined by two‐way ANOVA: **P ≤ 0.01. _P_‐values were corrected for multiple comparisons (Bonferroni).

C–E Neutrophils of the indicated genotypes were left untreated (Mock) or infected with S. Typhimurium (3 h). IL‐1β secretion levels were determined in culture media (C, E), and combined cell lysates and culture media (Lys+Sup) were immunoblotted for cleavage of pro‐IL‐1β (~39 kDa) into mature IL‐1β (p17) (D). Cytokine values represent mean ± SEM of n = 3 biological repeats. Statistical significance was determined by two‐way ANOVA: *P ≤ 0.05; **P ≤ 0.01; ns = non‐significant. _P_‐values were corrected for multiple comparisons (Bonferroni).

F LPS‐primed human blood neutrophils were left untreated (Mock) or infected with S. Typhimurium (MOI 0.5) for 3 h in the presence or absence of VX‐765 before IL‐1β secretion levels in culture media were determined. Cytokine data represent mean ± SEM of n = 4 donors. Statistical significance was determined by one‐way ANOVA: **P ≤ 0.01. _P_‐values were corrected for multiple comparisons (Bonferroni).

EV4 (download PNG )

Figure

A–F Neutrophils of the indicated genotypes were left untreated (Mock) or stimulated with LPS. PMP was assessed by measurement of SYTOX Green incorporation over time (A–F). Values represent mean ± SEM of three independent biological repeats (A–C, E, F). Statistical significance was determined by non‐parametric wilcoxon two‐tailed _t_‐test: *P ≤ 0.05; **P ≤ 0.01. (D) Data are depicted as individual data points with a mean of n = 2 biological repeats.

EV5 (download PNG )

Figure

A–C LPS‐primed neutrophils of the indicated genotypes were left untreated (Mock) or stimulated with PMA (3 h) before IL‐1β was measured in culture media (A). (B) Representative confocal micrographs of Mock‐ and PMA‐stimulated neutrophils of the indicated genotypes stained with DAPI (dark‐blue nuclei) and SYTOX Green (light blue) taken at the indicated time points (scale bar: 20 μM). (C) PMP was assessed by measurement of SYTOX Green incorporation over time.

D, E LPS‐primed human blood neutrophils were left untreated (Mock) or stimulated with PMA in the presence or absence of VX‐765. PMP was determined by SYTOX green uptake over time (D) and cell death was measured by LDH release in culture media 3 h post‐stimulation (E).

Data information: Error bars represent mean ± SEM of n = 3 (A) or n = 4 (C–E) biological repeats. (C–E) Statistical significance was determined by two‐way ANOVA (C, D) or one‐way ANOVA (E): *P ≤ 0.05; **P ≤ 0.01; ns = non‐significant. _P_‐values were corrected for multiple comparisons (Bonferroni).

Rights and permissions

Copyright: The Author(s)

Reprints and permissions

About this article

Cite this article

Chauhan, D., Demon, D., Vande Walle, L. et al. GSDMD drives canonical inflammasome‐induced neutrophil pyroptosis and is dispensable for NETosis.EMBO Rep 23, EMBR202154277 (2022). https://doi.org/10.15252/embr.202154277

Download citation

Keywords