Neutrophil-specific gain-of-function mutations in Nlrp3 promote development of cryopyrin-associated periodic syndrome.

Gain-of-function mutations in NLRP3 are responsible for a spectrum of autoinflammatory diseases collectively referred to as "cryopyrin-associated periodic syndromes" (CAPS). Treatment of CAPS patients with IL-1-targeted therapies is effective, confirming a central pathogenic role for IL-1β. However, the specific myeloid cell population(s) exhibiting inflammasome activity and sustained IL-1β production in CAPS remains elusive. Previous reports suggested an important role for mast cells (MCs) in this process. Here, we report that, in mice, gain-of-function mutations in Nlrp3 restricted to neutrophils, and to a lesser extent macrophages/dendritic cells, but not MCs, are sufficient to trigger severe CAPS. Furthermore, in patients with clinically established CAPS, we show that skin-infiltrating neutrophils represent a substantial biological source of IL-1β. Together, our data indicate that neutrophils, rather than MCs, can represent the main cellular drivers of CAPS pathology.

Introduction

The NLRP3 inflammasome is a protein complex responsible for caspase-1–dependent release of IL-1β and IL-18 (Jo et al., 2016; Latz et al., 2013; Tschopp and Schroder, 2010). Gain-of-function mutations in NLRP3 are responsible for a spectrum of autoinflammatory diseases collectively referred to as “cryopyrin-associated periodic syndromes” (CAPS) or “NLRP3-associated autoinflammatory diseases” (Aksentijevich et al., 2002; Dowds et al., 2004; Hoffman et al., 2001a; Louvrier et al., 2020). These mutations in the NLRP3 gene cause constitutive activation of the NLRP3 inflammasome, leading to enhanced activation of caspase-1 and secretion of IL-1β and IL-18 (Agostini et al., 2004). Treatment of CAPS patients with IL-1–targeted therapies is effective (Fenini et al., 2017; Hoffman et al., 2004; Hoffman et al., 2012; Hoffman et al., 2008; Lachmann et al., 2009), confirming a central pathogenic role for IL-1β. Previous reports using mouse models of CAPS indicate that mutation of Nlrp3 in myeloid cells is responsible for disease pathogenesis (Brydges et al., 2009; Meng et al., 2009). However, the specific myeloid cell population(s) exhibiting inflammasome activity and sustained IL-1β production in CAPS remains elusive.

CAPS encompasses a continuum of disease severity, characterized by fever, urticaria-like skin rashes, and systemic inflammation (Aksentijevich et al., 2007; Neven et al., 2004), and includes familial cold autoinflammatory syndrome (FCAS), Muckle-Wells syndrome (MWS), and neonatal-onset multisystem inflammatory disease (Neven et al., 2004). The urticaria-like skin rash observed in CAPS is similar to that associated with common urticaria, a disorder thought to depend largely on histamine released by mast cell (MC) activation. It has previously been reported that MC deficiency can reduce skin inflammation in mice expressing the Nlrp3 mutation, corresponding to the R260W mutation found in some MWS patients (Nakamura et al., 2012). In this model of CAPS, Nlrp3 mutant mice developed skin inflammation over several weeks (Meng et al., 2009). Additional CAPS mouse models have been reported, with different mutations in Nlrp3 inducing more severe pathology, characterized by the development of skin inflammation within a few days after birth, poor growth, and perinatal mortality (Brydges et al., 2009). However, because human CAPS represents a spectrum of diseases of different severity, the relative contributions of MCs might potentially depend on the nature of the Nlrp3 mutation and/or on the severity of the pathology.

Both monocytes/macrophages and dendritic cells (DCs) have been used extensively to study the mechanism of activation of the NLRP3 inflammasome (Kool et al., 2008; Martinon et al., 2002; Martinon et al., 2006; Sharp et al., 2009). In addition, macrophages from CAPS patients were shown to spontaneously secrete IL-1β (Agostini et al., 2004; Camilli et al., 2020). However, whether macrophages and/or DCs play a key role in CAPS remains to be fully determined.

Blood neutrophilia and neutrophil infiltration in several tissues are hallmarks of CAPS (Hoffman and Broderick, 2016; Hoffman et al., 2001b; Ley et al., 2018). It was reported that levels of neutrophil secretory proteins are significantly elevated in the plasma of mice harboring a CAPS-associated Nlrp3 mutation (Johnson et al., 2017). However, the functions of neutrophils in CAPS remain largely unknown. Here we report that, in mice, Nlrp3 and Nlrp3 mutations, which mirror two clinically established mutations associated with MWS (A352V) and FCAS (L353P), respectively (Brydges et al., 2009), restricted to neutrophils but not MCs, are sufficient to trigger severe CAPS. Furthermore, in patients with clinically established CAPS, we show that skin-infiltrating neutrophils represent a substantial biological source of IL-1β. Together, our data indicate that neutrophils, rather than MCs, can represent the main cellular drivers of CAPS pathology.

Results and discussion

MCs are not required for the development of CAPS driven by Nlrp3 A350V or L351P mutations in mice

Mutations of Nlrp3 in myeloid cells are responsible for CAPS pathogenesis in mice (Brydges et al., 2009; Meng et al., 2009). To achieve expression of NLRP3 mutant proteins in myeloid cells, we crossed Nlrp3 mice with transgenic mice expressing Cre recombinase under the control of the lysozyme 2 promoter (Lys). To test whether MCs played a role in this myeloid cell–induced CAPS model, we crossed Nlrp3 Lys-Cre mice with Kit MC-deficient mice to generate triple-transgenic littermate mice expressing the Nlrp3 mutation under the control of the Lys promoter in either MC-deficient (Kit) or MC-sufficient backgrounds (Kit). As reported previously, Nlrp3 Lys-Cre mice had severe growth delay and died within 2 wk of birth (Brydges et al., 2009; Fig. 1, A–C). As expected, MC numbers were markedly reduced in the skin of Kit Nlrp3 Lys-Cre mice as compared with Kit Nlrp3 Lys-Cre mice (Fig. 1 A). However, no noticeable differences in skin pathology, body weight, or survival were observed in Nlrp3 mutant MC-sufficient versus MC-deficient mice, and all mice died within 2 wk after birth (Fig. 1, A–C). Nlrp3 Lys-Cre mice had slightly increased numbers of MCs in the skin as compared with Cre-negative controls (Fig. S1, A and B). However, <20% of these MCs in the skin of Nlrp3 Lys-Cre mice expressed IL-1β (Fig. S1, A and C).

Figure 1.

MCs are not required for the development of severe CAPS in mice. Nlrp3 Lys-Cre mice were crossed with MC-deficient Kit mice to generate littermate mice expressing the Nlrp3 mutation in an MC-sufficient (Kit) or an MC-deficient (Kit) background. (A) Toluidine blue staining (assessing MCs) and H&E staining (assessing leukocyte infiltration) of skin sections from Kit and Kit or Lys-Cre control mice. Black arrows indicate MCs. (B) Body weight at the indicated time points after birth in Nlrp3 control mice (data are pooled from Kit and Kit backgrounds; n = 40 at day 0), MC-sufficient Kit mice (n = 6 at day 0), and MC-deficient Kit mice (n = 7 at day 0). Data are presented as mean ± SEM. (C) Survival of Nlrp3 control mice (n = 29), Kit mice (n = 9), and MC-deficient Kit mice (n = 42). ***, P < 0.001 using the Mantel-Haenszel log-rank test. (D) Birth table of Kit pups. Shown are total live pups at birth and total pups on days 1 and 2 after birth. (E) Toluidine blue staining (assessing MCs) and H&E staining (assessing leukocyte infiltration) of skin sections from 8-wk-old Nlrp3 (n = 18) and Nlrp3 (n = 21) mice. Black arrows indicate MCs. (F) Survival of Nlrp3 and Nlrp3 mice. N.S., P > 0.05. Scale bars: 50 µm.

Figure S1.

Phenotype of MCs in Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–IL-1β antibodies, and sulforhodamine (SRho)-coupled avidin (which stains MCs). Scale bars: 50 µm. (B) Quantification of avidin+ MCs in skin sections. Results show values from individual mice, with bars indicating mean ± SEM. (C) Percentage of IL-1β+ and IL-1β− MCs (avidin+) in skin sections. (D) Responses of Nlrp3 (n = 8) and responses of Nlrp3 (n = 7) mice in a model of IgE-mediated passive systemic anaphylaxis (PSA). Mice were sensitized i.p. with anti-DNP IgE and challenged i.p. 24 h later with DNP–human serum albumin (HSA). Data are pooled from two independent experiments and show changes in body temperature (Δ°C [mean ± SEM]) after challenge. (E) CD34+ progenitors were purified by positive selection from the blood of healthy donors and CAPS patients and differentiated in vitro for 12 wk into PBCMCs. (F) IgE-sensitized PBCMCs were stimulated with 2 µg/ml anti-IgE in the presence of SRho-coupled avidin (Av.SRho). Data show mean time-lapse sequence ± SEM of SRho-avidin quantification (mean fluorescence intensity) from pooled single-cell analyses (n = 16–19) from PBCMCs from three healthy donors and three CAPS patients. *, P < 0.05 (two-way ANOVA). (G) IgE-sensitized PBCMCs were stimulated with anti-IgE at the indicated concentration. MC degranulation was quantified using β-hexosaminidase release (β-Hex) assay. Data are presented as mean ± SEM and are pooled from PBCMCs from three healthy donors and three CAPS patients. *, P < 0.05; **, P < 0.01 (Mann-Whitney U test). (H) Quantification of IL-1β release at 16 h upon stimulation with LPS (100 ng/ml) or PBS (“vehicle” as a control) in PBCMCs from healthy donors or CAPS patients. Results show values from individual donors.

We used the same approach with the FCAS-associated Nlrp3 mutant mouse model. When crossed to Lys-Cre mice, the resulting mutants displayed a more severe phenotype, because very few mice survived within 1 or 2 d of birth (Brydges et al., 2009). We observed the same phenotype in the MC-deficient Kit Nlrp3 Lys-Cre mice: Among the 25 mice born, only 3 survived until day 1, and all died by day 2 (Fig. 1 D). Altogether, our data suggest that MCs are not required for the development of pathology induced by myeloid cells in these two severe CAPS models.

Previous reports indicate that mouse MCs expressing CAPS-associated Nlrp3 mutations can produce IL-1β (Nakamura et al., 2012; Nakamura et al., 2009). However, Lys-Cre mice do not express the Cre recombinase in MCs (Abram et al., 2014). Thus, to study further the in vivo effects of a CAPS-associated Nlrp3 mutation in MCs, we generated a mouse model in which the pathogenic Nlrp3 mutation was restricted to MCs (i.e., by crossing Nlrp3 mice with Mcpt5-Cre mice, which express Cre recombinase under the control of the connective tissue MC-specific MC protease 5 [Mcpt5] promoter; Dudeck et al., 2011). Interestingly, Nlrp3 Mcpt5-Cre mice did not develop any observable signs of skin inflammation and survived normally (mice were monitored for up to 1 yr after birth; Fig. 1, E and F). This supports the conclusion that a MC-restricted Nlrp3 L351P mutation is not sufficient to drive CAPS in mice. In addition, Nlrp3 Mcpt5-Cre and Nlrp3 Mcpt5-Cre mice had similar numbers of skin MCs (Fig. 1 E) and similar responses in a model of IgE-mediated passive systemic anaphylaxis (PSA; Fig. S1 D) that largely depends on MCs (Finkelman, 2007; Reber et al., 2017b). This indicates that, under the conditions tested, the Nlrp3 mutation in MCs had no significant effect on IgE-mediated effector functions.

Our results differ significantly from those of Nakamura et al., who demonstrated that MCs harboring an R258W gain-of-function Nlrp3 mutation importantly contributed to CAPS pathogenesis in mice (Nakamura et al., 2012). Several factors might account for such differences. For example, the A350V or L351P Nlrp3 mutations that we studied confer a more severe phenotype than the R258W mutation in mice (Brydges et al., 2009; Nakamura et al., 2012). Indeed, we and others reported that, in several disease models, MCs can amplify inflammation in models of moderate severity, whereas other cells may mask any contributions of MCs in more severe models (Reber et al., 2012). In addition, Nakamura et al. reported that the microbiome can play an important role in the development of CAPS in Nlrp3 mice (Nakamura et al., 2012). Therefore, it is possible that differences in the microbiota might account, at least in part, for the different results obtained.

We also derived MCs in vitro from circulating CD34+ hematopoietic progenitors from the blood of different CAPS patients and healthy donors (Fig. S1 E). Although we observed a slight increase in IgE-mediated degranulation dynamics of MCs from CAPS patients compared with healthy donors (Fig. S1, F and G), none of these MCs released detectable amounts of IL-1β upon classical stimulation of the NLRP3 pathway with LPS (Fig. S1 H). Altogether, these data suggest that MCs do not represent an important source of IL-1β in CAPS.

Expression of Nlrp3 and Il1b genes among immune cells

To identify which cell type(s) exhibited mRNA associated with increased inflammasome activity and IL-1β production in mice, we mapped the expression of Nlrp3 and Il1b genes among various subpopulations of mouse immune cells, using publicly available RNA-sequencing data (Heng et al., 2008). This analysis showed that neutrophils, DCs, monocytes, and basophils, but not MCs, represent potential major sources of NLRP3 in mice and that neutrophils are likely to represent the main source of IL-1β among the 21 purified mouse immune cell populations analyzed (Fig. S2 A). In line with these findings, we also found that human granulocytes, DCs, and monocytes express high levels of NLRP3 and IL1B mRNA in similar RNA microarray data (Shay et al., 2013; Fig. S2 B).

Figure S2.

Gene expression analysis of mouse and human NLRP3 and IL1B, expression of caspase-1 and IL-18 in neutrophils from Publicly available RNA-sequencing data of Nlrp3, Il1b, Tpsg1 (tryptase gamma, expressed in MCs), and Ly6g (expressed in neutrophils) in different mouse immune cell subpopulations (http://rstats.immgen.org/Skyline/skyline.html; Heng et al., 2008); data are shown using a heat map of mRNA expression levels. NK, natural killer; DP, double positive; SP, single positive. (B) Publicly available microarray data of NLRP3 and IL1B in different human immune cell populations (http://rstats.immgen.org/comparative/comparative_search.php; Shay et al., 2013); data are shown using a heat map of mean-centered expression values (red/blue color bar on the bottom). (C) Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–caspase-1 antibodies, and anti-Ly6G antibodies (which stains neutrophils). Scale bars: 30 µm. (D) Percentage of caspase-1+ and caspase-1− neutrophils (Ly6G+) in skin sections. (E) Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI, anti–IL-18 antibodies, and anti-MPO antibodies. Scale bars: 20 µm. (F) Percentage of IL-18− and IL-18+ cells in the MPO+ neutrophil population versus other cells in skin sections. Scale bar: 30 µm. (G–I) Neutrophils were purified by negative selection from the blood of five control subjects and four CAPS patients (G). For three patients, neutrophils were purified at two different occasions 6 mo apart (experiments 1 and 2). Neutrophils were incubated with LPS (100 ng/ml) or vehicle (PBS) for 16 h. IL-1β (H) and IL-18 (I) concentrations in the medium were quantified by ELISA. (J) Skin tissue sections from two CAPS patients stained with DAPI, anti–IL-1β, and anti-MPO antibodies. Scale bars: 50 µm. Lower panels are enlargements of the boxed areas. Patient 1 is depicted in Fig. 3 E.

Figure 3.

Neutrophils are an important source of IL-1β in the skin of Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–IL-1β antibodies, and anti-Ly6G antibodies (which stain neutrophils). Lower panels are enlargements of the boxed areas. Scale bars, 100 µm. (B) Quantification of Ly6G+ neutrophils in skin sections. (C) Quantification of Ly6G+CD11b+ neutrophils in the blood from 7-d-old Nlrp3 and Nlrp3 mice (n = 10/group). Results in B and C show values from individual mice, with bars indicating mean ± SEM. *, P < 0.05; **, P < 0.01 (Mann-Whitney U test). (D) Percentage of IL-1β+ and IL-1β− neutrophils (Ly6G+) in skin sections. (E) Confocal microscopy showing staining of skin tissue sections from one healthy donor and one CAPS patient stained with DAPI (to indicate nuclei), anti–IL-1β antibodies, and anti-MPO, which is mainly produced by neutrophils. Scale bars, 50 µm. Lower panels are enlargements of the boxed areas. (F) Quantification of MPO+ neutrophils in the skin from CAPS patients and healthy donors. (G) Percentage of IL-1β+ and IL-1β− neutrophils (MPO+) in skin sections from three CAPS patients. Representative skin sections from CAPS patients 2 and 3 are depicted in Fig. S3 E.

Contribution of DCs and macrophages to IL-1β production and CAPS pathogenesis

Nlrp3 Lys-Cre mice had a marked increase in numbers of skin macrophages compared with Cre− controls (Fig. 2, A and B). However, <10% of these macrophages were positive for IL-1β (Fig. 2 C). By contrast, between 20% and 40% of skin CD11c+ DCs stained positive for IL-1β (Fig. 2, D–F).

Figure 2.

Evidence that DCs participate in CAPS-like syndrome induced by the Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–IL-1β antibodies, and anti-F4/80 antibodies (which stain macrophages). Scale bars, 20 µm. (B) Quantification of F4/80+ macrophages in skin sections. Data represent values from individual mice, with bars indicating mean ± SEM. ***, P < 0.001 (Mann-Whitney U test). (C) Percentage of IL-1β+ and IL-1β− macrophages (F4/80+) in skin sections. (D) Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–IL-1β antibodies, and anti-CD11c antibodies (which stain DCs). Scale bars, 20 µm. (E) Quantification of CD11c+ DCs in skin sections. Data represent values from individual mice with bars indicating mean ± SEM. (F) Percentage of IL-1β+ and IL-1β− DCs (CD11c+) in skin sections. (G) Body weight at the indicated time points after birth in Nlrp3 (n = 12 at day 0), Nlrp3 (n = 6 at day 0), Nlrp3 (n = 10 at day 0), and Nlrp3 mice (n = 10 at day 0). (H) Survival of Nlrp3 (n = 20 at day 0), Nlrp3 (n = 19 at day 0), Nlrp3 (n = 19 at day 0), and Nlrp3 mice (n = 17 at day 0). (I) Body weight at the indicated time points after birth in Nlrp3 (n = 3 at day 0), Nlrp3 (n = 7 at day 0), Nlrp3 (n = 9 at day 0), and Nlrp3 mice (n = 6 at day 0). (J) Survival of Nlrp3 (n = 7 at day 0), Nlrp3 (n = 9 at day 0), Nlrp3 (n = 57 at day 0), and Nlrp3 mice (n = 21 at day 0).

On the basis of these findings, we decided to use Cd11c-Cre mice to assess the potential contributions of DCs and macrophages. Cd11c-Cre mice were reported to express Cre in DCs (85–90% of conventional DCs and plasmacytoid DCs in the spleen) and also in macrophages (>95% alveolar macrophages and ∼70% red pulp macrophages), but not in MCs, neutrophils, or other granulocytes (Abram et al., 2014). In marked contrast with our data in Lys-Cre mice, we found that Cd11c-Cre pups gained weight and survived normally, indicating that restricting the A350V mutation to macrophages and DCs is not sufficient to induce a CAPS-like phenotype in mice (Fig. 2, G and H). We also generated Cd11c-Cre mice. Interestingly, these mice did develop a CAPS-like phenotype, with reduced body weight and perinatal mortality (Fig. 2, I and J). However, the phenotype was less severe than that observed in Lys-Cre mice, because all Lys-Cre mice died within 2 d after birth (Brydges et al., 2009; Fig. 1 D), whereas some Cd11c-Cre mice survived for up to 22 d (Fig. 2 J).

Evidence that neutrophils represent an important source of IL-1β in CAPS

A previous report indicated that neutrophil secretory proteins, including myeloperoxidase (MPO), are significantly elevated in the plasma of mice with a tamoxifen-inducible Nlrp3 mutation (Johnson et al., 2017). This observation suggested that the pathology in this model might be associated with neutrophil activation. In line with these results, we observed elevated numbers of neutrophils in the blood and skin of Nlrp3 Lys-Cre mice (Fig. 3, A–C). Importantly, a large proportion of the skin neutrophils were strongly positive for IL-1β (Fig. 3 D). These results indicate that neutrophils, rather than MCs, represent an important source of IL-1β in this severe CAPS model. Because the antibody we used recognizes both pro–IL-1β and active IL-1β, we also stained skin samples with an antibody recognizing pro–caspase-1 and active caspase-1 (upon cleavage at Asp210). Almost all caspase-1+ cells were neutrophils, confirming that neutrophils might be a major source of inflammasome activity in Nlrp3 Lys-Cre mice (Fig. S2, C and D). Activation of the NLRP3 inflammasome also leads to IL-18 release, and IL-18 contributes to the CAPS-like phenotype in both Nlrp3 Lys-Cre and Nlrp3 Lys-Cre mice (Brydges et al., 2013). We found that ∼45% of neutrophils stained positive for IL-18 in the skin of Nlrp3 Lys-Cre mice. However, >90% of all other cells were also IL-18+, indicating that this cytokine is potentially produced by many cell populations in Nlrp3 Lys-Cre mice (Fig. S2, E and F).

Figure S3.

Expression of caspase-1 in skin neutrophils and effect of treatment with anti–IL-1β or anti–IL-18 antibodies in Representative confocal microscopy showing staining of skin tissue sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–caspase-1 antibodies, and anti-Ly6G antibodies (which stain neutrophils). Scale bars: 30 µm. (B) Percentage of caspase-1+ and caspase-1− neutrophils (Ly6G+) in skin sections. (C–E) Treatment of Nlrp3 mice with anti–IL-1β or anti–IL-18 antibodies. Protocol outline (C), body weight (D), and survival (E) at the indicated time points. In D, data are presented as mean ± SEM.

In accord with these results, neutrophils freshly purified from the peripheral blood of CAPS patients released IL-1β at levels slightly (but not significantly) higher than neutrophils from healthy donors (Fig. S2, G and H). By contrast, none of these cells released detectable amounts of IL-18 (Fig. S2 I). We also observed substantial infiltration of neutrophils in skin biopsies from CAPS patients (Fig. 3, E and F; and Fig. S2 J), and ∼60% of these neutrophils were strongly positive for IL-1β (Fig. 3, E and G; and Fig. S2 J). Altogether, our data suggest that neutrophils are major IL-1β producers in CAPS and thus might play a key role in CAPS pathology.

Neutrophil-intrinsic Nlrp3 mutation triggers development of CAPS

To further investigate the role of neutrophils in CAPS, we crossed Nlrp3 mice with neutrophil-specific MRP8-Cre mice (Passegué et al., 2004; Reber et al., 2017a), generating mice with a gain-of-function mutation in Nlrp3 restricted to neutrophils. Indeed, previous reports indicate that MRP8-Cre mice express at baseline Cre recombinase in ∼90% of blood neutrophils and ∼80% of neutrophils in the spleen and bone marrow (Abram et al., 2014; Reber et al., 2017a). Nlrp3 MRP8-Cre mice exhibited a severe skin inflammatory phenotype and substantial reduction in body weight, exhibited blood neutrophilia, and died within 2 wk of birth (Fig. 4, A–D). This phenotype was very similar to that observed in myeloid-restricted Nlrp3 Lys-Cre mice (Fig. 1, A–C).

Figure 4.

Neutrophil-restricted gain-of-function mutation in Representative picture of 7-d-old Nlrp3 and Nlrp3 mice. (B) Body weight at the indicated time points after birth in Nlrp3 (n = 11 at day 0) and Nlrp3 mice (n = 17 at day 0). (C) Survival of Nlrp3 (n = 16) and Nlrp3 mice (n = 18). (D) Quantification of Ly6G+ CD11b+ neutrophils in the blood of 7-d-old Nlrp3 mice (n = 11) and Nlrp3 controls (n = 17). Results show values from individual mice, with bars indicating mean ± SEM. (E) GFP expression (mean fluorescence intensity [MFI]) among Ly6G− and Ly6G+ blood leukocytes in Nlrp3 mice and Nlrp3 controls (n = 6/group). (F) Representative confocal microscopy showing staining of skin sections from Nlrp3 mice and Nlrp3 controls stained with DAPI (which stains nuclei), anti–IL-1β antibodies, and anti-Ly6G antibodies (which stain neutrophils). Lower panels are enlargements of the boxed areas. Scale bars: 50 µm. (G) Quantification of Ly6G+ neutrophils in skin sections. Results show values from individual mice, with bars indicating mean ± SEM. (H) Percentage of IL-1β+ and IL-1β− neutrophils (Ly6G+) in skin sections. (I) Representative picture of 4-d-old Nlrp3 and Nlrp3 mice. (J) Representative toluidine blue staining (assessing MCs) and H&E staining (assessing leukocyte infiltration) of skin sections from 7-d-old Nlrp3 and Nlrp3 mice. Black arrows indicate MCs. (K) Survival of Nlrp3 (n = 12) and Nlrp3 mice (n = 13). ***, P < 0.001 using the Mantel-Haenszel log-rank test (in C and K) or Mann-Whitney U test (in D, E, and G). Scale bars: 100 µm.

MRP8-Cre mice have an internal ribosome entry site–GFP reporter that can be used to track Cre-expressing cells (Passegué et al., 2004; Reber et al., 2017a). We confirmed that GFP expression was restricted to Ly6G+ neutrophils in the blood of Nlrp3 MRP8-Cre mice (Fig. 4 E). Nlrp3 MRP8-Cre mice also had a strong infiltration of neutrophils in the skin (Fig. 4, F and G). In line with our previous observations in Lys-Cre mice (Fig. 3), most skin neutrophils stained positively for IL-1β (Fig. 4, F and H) and caspase-1 (Fig. S3, A and B).

To assess the respective contributions of IL-1β versus IL-18 to the CAPS-like phenotype observed in Nlrp3 MRP8-Cre mice, we treated these mice, from days 2–12 after birth, every 2 d with neutralizing antibodies against IL-1β or IL-18 (Fig. S3 C). Treatment with anti–IL-18 antibodies had no detectable effect, because all mice gained no more than ∼2 g of body weight and died within 12 d (Fig. S3, D and E), a phenotype very similar to that observed in untreated Nlrp3 MRP8-Cre mice (Fig. 1, B and C). Mice treated with the anti–IL-1β antibody gained slightly more weight, and 50% of the mice were still alive at day 14 (Fig. S3, D and E). However, all of these mice developed a clear CAPS-like phenotype. This suggests that, in mice with the Nlrp3 A350V mutation restricted to neutrophils, either anti–IL-1β antibody did not fully block IL-1β in newborns or additional mediators also significantly contributed to the disease. Among these potential mediators, TNF was reported to contribute importantly to the CAPS-like phenotype in Nlrp3 Lys-Cre mice (McGeough et al., 2017). Besides activation of IL-1β and IL-18, caspase-1 also cleaves and activates gasdermin D, a pore-forming protein that induces pyroptosis (Chen et al., 2018; Sollberger et al., 2018). Gasdermin D–induced pyroptosis thus may also contribute to the CAPS-like phenotype in Nlrp3 mutant mice. In addition, the skin phenotype in Nlrp3 MRP8-Cre mice is already evident at birth. Therefore, it is possible that starting anti–IL-1β or anti–IL-18 therapy at day 2 might be too late to permit full rescue of the phenotype in this model.

Finally, severe CAPS also developed in Nlrp3 MRP8-Cre mice, with a strong infiltration of leukocytes, and increased numbers of MCs, in the skin (Fig. 4, I and J). However, although all Nlrp3 Lys-Cre mice died within 2 d after birth (Brydges et al., 2009; Fig. 1 D), some Nlrp3 MRP8-Cre mice survived for up to 1 wk after birth (Fig. 4 K). These results demonstrate that the neutrophil-intrinsic Nlrp3 mutation is sufficient to trigger the development of the pathological features associated with these models of CAPS, but they also confirm that the Nlrp3 mutation within additional myeloid cell population(s) such as macrophages and DCs (as revealed using Cd11c-Cre mice in Fig. 2, G and H) may contribute to the more severe phenotype associated with the Nlrp3 mutation.

It remains to be fully determined why gain-of-function mutations in Nlrp3 in neutrophils, and to a lesser extent in DCs or macrophages, but not in MCs, are sufficient to drive CAPS pathology. One potential explanation is that MCs lack expression of one or more genes required for optimal activation of the NLRP3 inflammasome. Furthermore, although neutrophils are virtually absent from healthy skin, these cells are the first leukocytes to be attracted at sites of inflammation and clearly represent the main immune cell population in the inflamed skin of CAPS patients and in our mouse models. This might explain why neutrophils represent the main driver of CAPS pathology, although DCs, macrophages, and likely additional cell populations can release IL-1β and display NLRP3 inflammasome activity.

Altogether, our data suggest that neutrophils may represent a novel therapeutic target in CAPS. Although fully depleting neutrophils would likely expose CAPS patients to unacceptably high risks of infection, blocking neutrophil recruitment could represent an efficient strategy to reduce skin inflammation in CAPS. CXCR2 is considered to be the dominant neutrophil chemokine receptor in humans, and several CXCR2 antagonists have already been tested in clinical trials (Németh et al., 2020). Although future studies are needed, these new therapeutics may represent a promising treatment for neutrophilic skin disease and may have beneficial effects in CAPS.

Materials and methods

Mice

MRP8-Cre/iresGFP mice (B6.Cg-Tg(S100A8-cre,-EGFP)1Ilw/J; Passegué et al., 2004; Reber et al., 2017a) were obtained from Irving Weissman (Stanford University, Stanford, CA) and Clifford Lowell (University of California, San Francisco, San Francisco, CA). Kit mice (B6.Cg-Kit/HNihrJaeBsmGlliJ), Lys-Cre mice (B6.129P2-Lyz2/J), and Cd11c-Cre mice (B6.Cg-Tg(Itgax-cre)1-1Reiz/J) were obtained from The Jackson Laboratory. C57BL/6-Mcpt5-Cre mice (Dudeck et al., 2011) were provided by A. Roers. NLRP3 (B6.129-Nlrp3/J) and Nlrp3 mice (B6N.129-Nlrp3/J) have been described previously (Brydges et al., 2009). Mice were bred in the Institut Pasteur; University of California, San Diego; or Stanford University specific pathogen–free animal facilities. All animal care and experimentation were conducted in compliance with the guidelines of the National Institutes of Health and with the specific approval of the institutional animal care and use committees of Stanford University and the University of California, San Diego, and/or of the Committee for Ethics in Animal Experimentation (Institut Pasteur, Paris, France) registered under #C2EA-89.

Survival and growth assessment

Pup survival was calculated as the percentage of pups surviving from birth. Growth gain was determined by weighing the mice every day from birth.

Flow cytometry

We used flow cytometry to identify and enumerate neutrophils in the peripheral blood. RBCs were lysed by treatment with RBC lysis buffer (BD Biosciences). Neutrophils were stained on ice for 30 min with anti–Ly6G-BV412 (clone 1A8, 1:200; BioLegend) and anti–CD11b-APC Vio770 (clone M1/70.15.11.5, 1:50; Miltenyi Biotec) antibodies. In some experiments, we used the FITC channel for analysis of GFP expression. Dead cells (stained with propidium iodide, 1:1,000) were not included in the analysis. Data were acquired using a MACSQuant flow cytometer (Miltenyi Biotec) and analyzed with FlowJo software (BD Biosciences).

Human samples

Skin biopsies from symptomatic CAPS patients were obtained from the Centre de Référence des Maladies AutoInflammatoires et des Amyloses Inflammatoires at Tenon Hospital, Paris, France. Control skin biopsies were obtained from Genoskin SAS (https://www.genoskin.com/). Genoskin has obtained all legal authorizations necessary from the French Ministry of Higher Education, Research and Innovation (AC-2017-2897) and the Personal Protection Committee (2017-A01041-52). Blood samples from three additional CAPS patients and healthy donors were obtained with prior written consent under protocols approved by the University of California, San Diego, Human Research Protections Program, and they were used to purify neutrophils or derive MCs from CD34+ progenitors, as detailed below. Clinical characteristics of all CAPS patients enrolled in the study are described in Table 1.

Table 1.

Clinical characteristics of patients with CAPS

Syndrome	Sex	Mutation	Age (yr)	Treatment	Figure
MWS	M	T436A	32	Canakinumab	Fig. S1, E–H
FCAS	F	L353P	86	Rilonacept	Fig. S1, E–H
FCAS	F	L353P	32	Rilonacept	Fig. S1, E–H
NOMID	M	G326E	13	Anakinra	Fig. S2, G–I
MWS	F	R260W	42	Canakinumab	Fig. S2, G–I
FCAS	F	L353P	74	Canakinumab	Fig. S2, G–I
FCAS?	F	V198M	46	–	Fig. S2 J
FCAS	M	A439V	48	Canakinumab	Fig. 3 E
MWS	F	Q567K	70	–	Fig. S2 J

F, female; M, male; NOMID, neonatal-onset multisystem inflammatory disease.

Histological analysis and immunofluorescence

For histological analysis of MCs and leukocytes in Kit mice and Kit mice, skin tissue was fixed in 10% (vol/vol) buffered formalin and embedded in paraffin and then cut into 4-μm-thick sections. Skin sections were stained with 0.1% (vol/vol) toluidine blue, pH 1, for detection of MCs or with H&E for detection of leukocytes. For immunofluorescence staining of MCs, macrophages, DCs, or neutrophils and IL-1β, IL-18, or caspase-1 in Nlrp3 or Nlrp3 mice, skin tissue was fixed with 1% paraformaldehyde, dehydrated with increasing sucrose concentrations (10%, 20%, and 30%), and embedded in optimal cutting temperature compound (Tissue-Tek). Sections measuring 8 μm in thickness were cut using a cryostat. Human skin biopsies were fixed and embedded in paraffin, and 4-µm-thick sections were cut using a microtome. Frozen mouse skin sections or paraffin-embedded human skin biopsies pretreated using a heat-induced epitope retrieval method (in 10 mM sodium citrate buffer, pH 6.0) were permeabilized for 30 min in PBS supplemented with 0.5% BSA and 0.1% saponin. Permeabilized skin sections were incubated overnight at 4°C with primary antibodies: mouse Ly6G/Ly6C antibody (RB6-8C5, catalog no. 14-5931-82, 1:50; eBioscience), mouse F4/80 antibody (BM8, catalog no. 14-4801-81, 1:50; eBioscience), mouse CD11c antibody (N418, catalog no. 53-0114-82, 1:50; eBioscience), sulforhodamine-coupled avidin (catalog no. S7635-50MG, 1:1,000; Sigma-Aldrich), mouse IL-1β/IL-1F2 antibody (polyclonal, catalog no. AF-401-NA, 1:50; Bio-Techne), mouse IL-18 antibody (YIGIF74-1G7, catalog no. BE0237, 1:50; Bio X Cell), mouse caspase-1 antibody (polyclonal, catalog no. PA5-38099, 1:50; Thermo Fisher Scientific), human MPO antibody (polyclonal, catalog no. AF3667, 1:50; Bio-Techne), and human IL-1β antibody (GT289, catalog no. GTX634188, 1:50; Euromedex). They were then extensively washed and incubated with appropriate secondary antibodies for 2 h at room temperature in the dark. Skin sections were later washed in PBS and mounted between slide and coverslip. 512 × 512–pixel images were acquired using a Zeiss LSM710 Meta inverted confocal laser-scanning microscope and were processed using ImageJ software. For quantification of cell numbers and identification of caspase-1–, IL-1β–, or IL-18–positive cells in sections of mouse skin, images of three to five consecutive microscopic fields from each mouse were obtained with a 20× objective. For each image, neutrophils, MCs, macrophages, or DCs were counted, and numbers of caspase-1–, IL-1β–, or IL-18–secreting cells were determined. Statistical analyses were performed using GraphPad Prism software (GraphPad Software).

Peripheral blood mononuclear cell–derived human MCs (PBCMCs)

Human subject protocols were approved by the institutional review boards of the National Institute of Arthritis and Musculoskeletal and Skin Diseases and Stanford University. CD34+ precursor cells were isolated from peripheral blood mononuclear cells of CAPS patients and healthy donors who gave written consent (EasySep Human CD34 Positive Selection Kit; STEMCELL Technologies). CD34+ cells were maintained for 1 wk under serum-free conditions using StemSpan medium (STEMCELL Technologies) supplemented with recombinant human IL-6 (50 ng/ml; PeproTech), human IL-3 (10 ng/ml; PeproTech), and 3% supernatant of Chinese hamster ovary transfectants secreting murine stem cell factor (SCF; 3% [corresponding to ∼50 ng/ml SCF]; a gift from Dr. P. Dubreuil, Marseille, France). Thereafter, the cells were maintained in IMDM GlutaMAX I, sodium pyruvate, 2-ME, 0.5% BSA, insulin-transferrin selenium (all from Invitrogen), ciprofloxacin (10 μg/ml; Sigma-Aldrich), IL-6 (50 ng/ml), and 3% supernatant of Chinese hamster ovary transfectants secreting mouse SCF. Before their use in experiments, PBCMCs were tested for phenotype by flow cytometry (anti-CD117 antibody [E-3, catalog no. sc-365504, 1:100; Santa Cruz Biotechnology], anti–Fcε receptor I antibody [AER-37, catalog no. 334639, 1:100; BioLegend]). PBCMCs were ready for experiments after ∼10 wk in culture. (MCs [positive for CD117 and Fcε receptor I] represented >95% of all cells.) In Fig. S1 D, PBCMCs were incubated in culture medium with or without pure LPS (100 ng/ml; Alexis Biochemicals) for 16 h, and levels of IL-1β in the culture supernatants were measured by ELISA (Invitrogen).

Single-cell analysis of PBCMC degranulation dynamics

The degranulation dynamics of single PBCMCs was analyzed as previously described (Gaudenzio et al., 2016). 5 × 104 human IgE-sensitized or nonsensitized PBCMCs were placed into poly-D-lysine–coated (5 µg/ml in water, catalog no. P6407; Sigma Aldrich) Nunc Lab-Tek 1.0 borosilicate cover glass system, eight chambers (catalog no. 155411; Thermo Fisher Scientific) in Tyrode’s buffer supplemented with 5 µg/ml avidin-sulforhodamine 101 (catalog no. A2348; Sigma-Aldrich), as previously described (Gaudenzio et al., 2016). Stimuli were added, and then fluorescence was recorded every 2.3 s in a controlled atmosphere (using a Zeiss stagetop incubation system with objective heater, 37°C, and 5% humidified CO2) using a Zeiss LSM710 or a Zeiss LSM780 Meta inverted confocal laser-scanning microscope, 20×/0.8 NA (working distance = 0.55) M27 objective, and electronic zoom 1 (8 bits/pixel, 512 × 512 pixels) for single-cell avidin-sulforhodamine fluorescence monitoring, and 63×/1.40 NA oil differential interference contrast M27 objective and electronic zoom 3 (dimensions: x = 512; y = 512; scaling: x = 0.264 µm; y = 0.264 µm) for high-resolution single-cell analyses. Mean fluorescence intensity was quantified using the measurement function of ImageJ software on randomly selected fields and untreated image sequences.

β-Hexosaminidase release assay

PBCMCs were incubated in culture medium with or without human IgE (1 µg ml–1; Sigma-Aldrich) overnight at 37°C. The cells were then washed and distributed in 96-well, flat-bottom plates at a density of 105 cells in 50 µl of Tyrode’s buffer at 37°C. 40 min later, cells were treated with 50 µl of prewarmed stimuli diluted in Tyrode’s buffer for 45 min at 37°C. PBCMCs were then stimulated with different concentrations of anti-IgE. β-Hexosaminidase release into the supernatants was measured as previously described (Gaudenzio et al., 2016).

Peripheral blood human neutrophils

Human subject protocols were approved by the institutional review boards of the National Institute of Arthritis and Musculoskeletal and Skin Diseases and Stanford University. Peripheral blood neutrophils of CAPS patients and healthy donors who gave written consent were purified by negative selection using a commercially available kit (EasySep Human Neutrophil Isolation Kit; STEMCELL Technologies). Neutrophils were incubated with or without pure LPS (100 ng/ml; Alexis Biochemicals) for 16 h, and levels of IL-1β and IL-18 in the culture supernatants were measured by ELISA (Invitrogen).

IgE-mediated PSA

IgE-dependent PSA was induced as described previously (Lilla et al., 2011). Briefly, mice were sensitized passively with IgE by i.p. injection of 20 µg of DNP-specific IgE (clone ɛ26 [Liu et al., 1980], kindly provided by Dr. Fu-Tong Liu, Department of Dermatology, University of California, Davis, Davis, CA) in 100 µl of PBS and then challenged i.p. the next day with 1 mg DNP–human serum albumin (A6661; Sigma-Aldrich) in 100 µl of PBS. Immediately before and at intervals after antigen challenge, body temperature was measured with a rectal thermometer (Physitemp Instruments).

Treatment with anti–IL-1β and anti–IL-18 neutralizing antibodies

Nlrp3 pups were treated s.c. with Armenian hamster IgG anti-mouse IL-1β (clone B122, catalog no. BE0246, 10 µg/g; Bio X Cell) mixed with rat IgG2aκ isotype controls (clone 2A3, catalog no. BE0089, 10 µg/g; Bio X Cell) in PBS (20 µl/g) or rat IgG2aκ antimouse IL-18 (clone YIGIF74-1G7, catalog no. BE0237, 10 µg/g; Bio X Cell) mixed with an Armenian hamster IgG isotype control (catalog no. BE0091, 10 µg/g; Bio X Cell) in PBS (20 µl/g) every other day beginning on day 2 of life for a total of six doses, as outlined in Fig. S3 C.

Statistical analysis

Data were analyzed for statistical significance using the Mantel-Haenszel log-rank test, unpaired Mann-Whitney U test, or two-way ANOVA, as indicated in the figure legends. P < 0.05 was considered statistically significant.

Online supplemental material

Fig. S1 shows numbers of MCs and IL-1β expression in MCs from Nlrp3 mice, responses of Nlrp3 mice in a model of IgE-mediated anaphylaxis, and degranulation upon stimulation with IgE/anti-IgE and IL-1β production upon stimulation with LPS in MCs derived from peripheral blood of healthy donors and CAPS patients. Fig. S2 shows the gene expression analysis of Nlrp3 and Il1b among major immune cell populations in humans and mice, caspase-1 and IL-18 expression in neutrophils from Nlrp3 mice, IL-1β and IL-18 release from neutrophils purified from the blood of CAPS patients, and the staining of IL-1β and neutrophil MPO in the skin of two CAPS patients. Fig. S3 shows caspase-1 expression in neutrophils from Nlrp3 mice and body weight and survival of Nlrp3 MRP8-Cre mice treated with anti–IL-1β or anti–IL-18 antibodies.