MCC950/CRID3 potently targets the NACHT domain of wild-type NLRP3 but not disease-associated mutants for inflammasome inhibition

The nucleotide-binding-domain (NBD)–and leucine-rich repeat (LRR)–containing (NLR) family, pyrin-domain–containing 3 (NLRP3) inflammasome drives pathological inflammation in a suite of autoimmune, metabolic, malignant, and neurodegenerative diseases. Additionally, NLRP3 gain-of-function point mutations cause systemic periodic fever syndromes that are collectively known as cryopyrin-associated periodic syndrome (CAPS). There is significant interest in the discovery and development of diarylsulfonylurea Cytokine Release Inhibitory Drugs (CRIDs) such as MCC950/CRID3, a potent and selective inhibitor of the NLRP3 inflammasome pathway, for the treatment of CAPS and other diseases. However, drug discovery efforts have been constrained by the lack of insight into the molecular target and mechanism by which these CRIDs inhibit the NLRP3 inflammasome pathway. Here, we show that the NAIP, CIITA, HET-E, and TP1 (NACHT) domain of NLRP3 is the molecular target of diarylsulfonylurea inhibitors. Interestingly, we find photoaffinity labeling (PAL) of the NACHT domain requires an intact (d)ATP-binding pocket and is substantially reduced for most CAPS-associated NLRP3 mutants. In concordance with this finding, MCC950/CRID3 failed to inhibit NLRP3-driven inflammatory pathology in two mouse models of CAPS. Moreover, it abolished circulating levels of interleukin (IL)-1β and IL-18 in lipopolysaccharide (LPS)-challenged wild-type mice but not in Nlrp3L351P knock-in mice and ex vivo-stimulated mutant macrophages. These results identify wild-type NLRP3 as the molecular target of MCC950/CRID3 and show that CAPS-related NLRP3 mutants escape efficient MCC950/CRID3 inhibition. Collectively, this work suggests that MCC950/CRID3-based therapies may effectively treat inflammation driven by wild-type NLRP3 but not CAPS-associated mutants.


Introduction
Inflammasomes are a suite of multiprotein complexes that play central roles in innate immune responses through their ability to recruit and activate caspase-1 (Casp1) [1,2]. This cysteine protease cleaves the cytokines interleukin (IL)-1β and IL-18 and drives pyroptosis, a highly inflammatory regulated cell death mode that is induced by cleavage of gasdermin D (GSDMD) [3,4]. Among the different inflammasome pathways, the nucleotide-binding-domain (NBD)and leucine-rich-repeat (LRR)-containing (NLR) family, pyrin-domain-containing 3 (NLRP3) inflammasome responds to the broadest suite of inflammasome agonists that include diverse pathogen-associated molecular patterns (PAMPs); host-derived danger-associated molecular patterns (DAMPs) such as ATP; protein aggregates and β-fibrils such as β-amyloid; a broad range of environmental insults and ionophores such as nigericin; and medically relevant crystals such as alum, calcium pyrophosphate dihydrate (CPPD), monosodium urate (MSU), silica, and asbestos [1,2]. Moreover, the NLRP3 inflammasome is engaged by gramnegative pathogens, lipopolysaccharides (LPS) that are sensed in the cytosol by the non-canonical NLRP3 inflammasome pathway [5][6][7][8]. Through the latter mechanism, cleavage of GSDMD by caspase-11-and its human orthologs caspases 4 and 5-promotes assembly of cytolytic GSDMD pores in the plasma membrane that also activate the NLRP3 inflammasome to drive Casp1-dependent IL-1β and IL-18 maturation [3,5]. Consistent with the chemical diversity of NLRP3 stimuli, activation of the NLRP3 inflammasome is thought to converge on sensing of a secondary messenger or cellular state that is universally induced by NLRP3-activating agents [9].
Aberrant NLRP3 inflammasome activity is thought to contribute to the pathogenesis of many chronic diseases, including inflammatory diseases such as gout and pseudogout; metabolic diseases like atherosclerosis and nonalcoholic fatty liver disease (NAFLD)/nonalcoholic steatohepatitis (NASH); and neurodegenerative diseases like Alzheimer's disease, Parkinson's disease, and multiple sclerosis [10,11]. Moreover, gain-of-function mutations in and around the central NAIP, CIITA, HET-E, and TP1 (NACHT) domain of NLRP3 cause three autosomal dominantly inherited periodic fever syndromes that together are known as cryopyrinassociated periodic syndrome (CAPS). Symptoms span a clinical spectrum, with Familial Cold Autoinflammatory Syndrome (FCAS) being the mildest; Muckle-Wells syndrome (MWS) being of moderate severity; and Neonatal Onset Multisystem Inflammatory Disease (NOMID)/Chronic Infantile Neurological, Cutaneous, and Articular Syndrome (CINCA) being the most severe form of CAPS, featuring systemic inflammation, neurological and sensory impairment, and deforming arthropathy [12].
Selective and potent inhibitors of the NLRP3 inflammasome may have broad therapeutic potential in CAPS and other diseases [10,11]. Early studies with the sulfonylurea drug glyburide provided proof of concept that PAMP-, DAMP-, and crystal-induced activation of the NLRP3 inflammasome pathway may be selectively targeted without interfering with other inflammasome pathways [13]. The related diarylsulfonylurea compound MCC950/Cytokine Release Inhibitory Drug 3 (CRID3) was originally reported as an inhibitor of IL-1β secretion [14] and subsequently shown to potently and selectively inhibit the NLRP3 inflammasome pathway in murine and human macrophages and monocytes with half maximal inhibitory concentration (IC 50 ) values in the low nM range [15,16]. There is significant interest in the discovery and development of diarylsulfonylurea CRIDs such as MCC950/CRID3 for the treatment of CAPS and other diseases based on its ability to curb inflammatory pathology in mouse models of CAPS, the experimental autoimmune encephalomyelitis mouse model of multiple sclerosis, NAFLD/NASH, and many other inflammatory disease models [10,11]. However, the direct molecular target of MCC950/CRID3 in the NLRP3 inflammasome pathway has remained elusive, hampering the rational optimization and development of MCC950/ CRID3-based therapies.
Utilizing two different chemoproteomic strategies, we here demonstrate that the NACHT domain of NLRP3 is the molecular target of diarylsulfonylurea CRIDs. Interestingly, we find photoaffinity labeling (PAL) of the NACHT domain of NLRP3 requires an intact (d)ATPbinding pocket and is substantially reduced for most CAPS-associated NLRP3 mutants. In accordance, NLRP3-driven inflammatory pathology in mouse models of CAPS was not efficiently curbed by MCC950/CRID3. Consistently, MCC950/CRID3 abolished circulating levels of IL-1β and IL-18 in LPS-challenged wild-type mice but not in CAPS mice and ex vivo-stimulated mutant macrophages. These results identify the central NACHT domain of wild-type NLRP3 as the molecular target of MCC950/CRID3 and show that CAPS-related NACHT mutations prevent efficient MCC950/CRID3 inhibition. Collectively, this work suggests that MCC950/CRID3-based therapies may effectively treat inflammation driven by wild-type NLRP3 but not CAPS-associated mutants.

MCC950/CRID3 selectively binds to human and murine NLRP3
As a first approach to identify the molecular target of MCC950/CRID3, we made use of PAL in combination with click chemistry [17]. Guided by limited structure-activity studies, a cell-permeable photoaffinity probe was synthesized (Fig 1A, compound PAL-CRID3) that contains a photoreactive benzophenone group to enable direct covalent labeling of MCC950/CRID3 targets upon exposure to UV light. The alkyne functionality in PAL-CRID3 allowed in situ click reaction with a 5-carboxytetramethylrhodamine (TAMRA) fluorescent reporter to support ingel fluorescence detection of the covalent MCC950/CRID3-protein adduct by SDS-PAGE. A dose-response analysis confirmed that PAL-CRID3 retained the ability to potently inhibit NLRP3 inflammasome activation with submicromolar IC 50 values. PAL-CRID3 inhibited nigericin-induced IL-1β secretion from N-palmitoyl-S-dipalmitoylglyceryl Cys-Ser-(Lys)4 (Pam3CSK4)-primed primary bone-marrow-derived macrophages (BMDMs) (Fig 1B and S1 Data; IC 50 = 731 nM) and estrogen-regulated homeobox protein Hox-B8 (ER-Hoxb8)immortalized macrophages (iMac) (Fig 1C and S1 Data; IC 50 = 453 nM). As a reference, MCC950/CRID3 inhibited nigericin-induced IL-1β secretion with IC 50 values of 9 and 4 nM, respectively (Fig 1B and 1C; S1 Data). We hypothesize that the lesser activity of PAL-CRID3 compared to MCC950/CRID3 is a combination of physical property and binding differences that decrease target occupancy.
To screen candidate targets of MCC950/CRID3, HEK293T cells overexpressing Flag-epitope-tagged fusions of the human inflammasome sensor proteins NLRP3, NLRP1, NLRC4, NLRP6, and MEFV were incubated with PAL-CRID3 and exposed to UV light to allow covalent binding of PAL-CRID3 to potential targets. Following cell lysis, the probe was conjugated to the TAMRA reporter by click chemistry, and protein lysates were separated by SDS-PAGE. Notably, in-gel fluorescence imaging showed significant TAMRA labeling of ectopically expressed NLRP3 but not other inflammasome sensors in the panel (Fig 1D). To further extend these findings, we next confirmed binding to murine Nlrp3 ( Fig 1E). As controls, protein phosphatase PP1Cb and glutathione S-transferase GSTO1, which have been proposed as the target of CRID compounds [14], were not labeled by PAL-CRID3 ( Fig 1E). Moreover, binding of PAL-CRID3 to human NLRP3 and murine Nlrp3 was only observed following UV cross-linking, and labeling was rescued by competition with excess MCC950/CRID3 (Fig 1E), thus ruling out nonspecific cross-linking and validating the specificity of these findings. To further characterize the interaction of PAL-CRID3 with components of the NLRP3 inflammasome, we assessed binding to human never-in-mitosis-gene-a-related kinase 7 (NEK7), apoptosis-associated speck-like protein (ASC), Casp1, and GSDMD. PAL-CRID3 failed to label the NLRP3 inflammasome components in the panel apart from NLRP3 ( Fig 1F). Collectively, these results suggest that CRID compounds, including MCC950/CRID3, inhibit NLRP3 inflammasome signaling by directly binding to NLRP3.
To confirm and extend these findings to endogenous NLRP3, we made use of recently described iBody technology to immobilize MCC950/CRID3 on polymers that enabled immunoprecipitation of MCC950/CRID3 targets [18]. In brief, MCC950/CRID3 was stochastically modified with a photoactivatable phenyldiazirine linker, and the resulting isomeric mixture was conjugated to a hydrophilic N-(2-hydroxypropyl)methacrylamide (HPMA) polymer backbone that is decorated with a biotin affinity tag (for details, please refer to the Materials and Methods section and [18]). The MCC950/CRID3 iBody conjugate is further referred to as iBody U121. The corresponding iBody conjugate lacking MCC950/CRID3 served as a negative control and is referred to as iBody U126.

PAL-CRID3 targets the central NACHT domain of wild-type NLRP3 but not CAPS mutants
NLRP3 consists of an amino-terminal Pyrin domain (PYD), a central NACHT domain, and carboxy-terminal LRRs that are thought to lock NLRP3 in an inactive conformation. To map the NLRP3 region(s) to which MCC950/CRID3 binds, we generated deletion mutants that cover the following three regions of human NLRP3: (i) the N-terminal PYD; (ii) the central NACHT domain (comprising the nucleotide-binding oligomerization domain (NOD) and helical domain [HD] 2); and (iii) the carboxy-terminal LRR region (Fig 2A). Consistent with our previous results (Fig 1), TAMRA analysis showed significant binding of PAL-CRID3 to ectopically expressed full-length NLRP3 in HEK293T cells ( Fig 2B). In addition, we observed binding of PAL-CRID3 to the isolated NACHT (NOD + HD2) region of NLRP3 but not to the PYD and LRR domains ( Fig 2B). Labeling of both full-length NLRP3 and the isolated NACHT region by PAL-CRID3 was rescued by treatment with an excess amount of free MCC950/ CRID3 (Fig 2B), confirming specificity and suggesting PAL-CRID3 and MCC950/CRID3 compete for the same binding site. Binding of ATP/dATP to the Walker A pocket in the NACHT domain is essential for NLRP3 inflammasome assembly and function [22]. Interestingly, mutation of the Walker A motif (GKT 229 /AAA) in full-length Nlrp3 abolished PAL-CRID3 binding (Fig 2C), suggesting that an intact ATP-binding pocket is required for strongest binding of diarylsulfonylurea CRID compounds.
The central NACHT domain of NLRP3 also contains most reported gain-of-function mutations that cause CAPS disease (https://infevers.umai-montpellier.fr/web/index.php). This prompted us to explore PAL-CRID3 binding to CAPS-associated NLRP3 mutants. Unexpectedly, introducing a random selection of 6 different CAPS mutations in the NACHT domain of human NLRP3 that are associated with, respectively, MWS (NLRP3 R325W , NLRP3 R260L , NLRP3 A352V ), FCAS (NLRP3 A439V , NLRP3 L305P ), or NOMID (NLRP3 R260P ) all blunted binding of PAL-CRID3 to full-length human NLRP3 ( Fig 2D). The classical MWS A352V and FCAS L353P mutations in human NLRP3 correspond to the A350V and L351P mutations that have been knocked into the murine Nlrp3 gene to model CAPS disease in mice [23]. Notably, the Nlrp3 A350V mutant did not significantly affect labeling by PAL-CRID3, whereas the Nlrp3 L351P mutation abolished labeling by PAL-CRID3 at 1 μM ( Fig 2E). A dose-response analysis showed that binding to the Nlrp3 L351P mutant was also substantially affected with higher (3 μM and 10 μM) PAL-CRID3 concentrations ( Fig 2F). Together, these results suggest that PAL of the NACHT domain requires an intact (d)ATP-binding pocket and is substantially reduced for most CAPS-associated NLRP3 mutants.

MCC950/CRID3 inhibits the inflammasome in wild-type but not Nlrp3 L351P macrophages
Considering our observation that CAPS-associated Nlrp3 mutants escape PAL-CRID3 binding may have potential implications for treating CAPS with MCC950/CRID3, we sought to functionally validate these results with MCC950/CRID3 in the reported Nlrp3 L351P CAPS model [23]. To this end, mice that were homozygous for the Nlrp3 L351P allele were bred to transgenic mice that hemizygously expressed the tamoxifen-inducible Cre recombinase-estrogen receptor (Cre-ERT2) fusion gene (CreT) [24]. Following tamoxifen treatment and excision of the floxed neomycin resistance cassette, BMDMs of the resulting Nlrp3 L351P/+ CreT + mice expressed Nlrp3 from both the wild-type and Nlrp3 L351P alleles. Macrophages from Cre-ERT2-negative littermates (Nlrp3 L351P/+ CreT − ), which only express Nlrp3 from the wild-type allele, were used as controls in these experiments.
Culture media of LPS-stimulated Nlrp3 L351P/+ CreT + macrophages contained significant levels of IL-1β ( Fig 3A and S2 Data) and IL-18 ( Fig 3B and S2 Data), which were associated with marked maturation of Casp1 and IL-1β in cell lysates ( Fig 3C). As reported [23], these LPSinduced inflammasome responses were driven by the CAPS-associated Nlrp3 L351P allele because Nlrp3 L351P/+ CreT − BMDMs failed to secrete detectable levels of IL-1β ( Fig 3A and S2 Data) and IL-18 ( Fig 3B and S2 Data) and did not contain mature Casp1 and IL-1β in their cell lysates ( Fig 3C). Consistent with our previous results that CAPS-associated Nlrp3 L351P escaped PAL-CRID3 binding, we found that MCC950/CRID3 failed to inhibit the above inflammasome responses driven by the Nlrp3 L351P allele. We confirmed that MCC950/CRID3 was active against wild-type NLRP3 because it abolished levels of LPS+ATP-induced secretion of IL-1β ( Fig 3D and S2 Data) and IL-18 ( Fig 3E and S2 Data) in culture media of Nlrp3 L351P/+ CreT − BMDMs (which are solely driven by wild-type Nlrp3 in this genotype, given the absence of a Cre-ERT2 transgene), as well as the concomitant maturation of Casp1 and IL-1β in the corresponding cell lysates (Fig 3F). Similarly, MCC950/CRID3 abolished LPS + nigericin-induced secretion of IL-1β (Fig 3G and S2 Data) and IL-18 ( Fig 3H and S2 Data) and maturation of Casp1 and IL-1β by Nlrp3 L351P/+ CreT − macrophages (Fig 3I). In marked contrast, LPS + ATPand LPS + nigericin-induced inflammasome responses were insensitive to MCC950/CRID3 blockade in Nlrp3 L351P/+ CreT + BMDMs that express both the Nlrp3 L351P allele and wild-type Nlrp3 (Fig 3D-3I and S2 Data). Together, these results establish that the CAPS-associated Nlrp3 L351P allele is insensitive to MCC950/CRID3 inhibition.

MCC950/CRID3 inhibition of inflammasome responses in Nlrp3 A350V macrophages
Unlike for Nlrp3 L351P , labeling of Nlrp3 A350V by PAL-CRID3 was not significantly impacted by the mutation (Fig 2E and 2F and S3 Data). To address whether this was mirrored by potent inhibition of Nlrp3 A350V -driven inflammasome responses by MCC950/CRID3, mice that were homozygous for the previously reported Nlrp3 A350V allele were bred to the CreT transgenic mice described above. Following tamoxifen treatment and excision of the floxed neomycin resistance cassette, BMDMs of the resulting Nlrp3 A350V/+ CreT + mice expressed Nlrp3 from both the wild-type and Nlrp3 A350V alleles. BMDMs from CreT-negative littermates (Nlrp3 A350V/+ CreT − ), which only express Nlrp3 from the wild-type allele, were used as controls in these experiments.

Inflammasome inhibition in homozygous mutant macrophages
The studies described above were performed in heterozygous macrophages that express both wild-type and CAPS-associated Nlrp3 mutants. Considering that the Nlrp3 NACHT region facilitates Nlrp3 oligomerization, we decided to further assess MCC950/CRID3 responses in macrophages that uniquely express the CAPS-associated Nlrp3 mutants in the absence of wild-type Nlrp3. To do so, we transduced wild-type, Nlrp3 A350V/A350V , and Nlrp3 L351P/L351P BMDMs with Cre-recombinase-expressing lentiviruses to excise the neomycin resistance cassette that is placed upstream of the Nlrp3 mutation and to allow expression of the CAPS-associated Nlrp3 mutants.

In vivo MCC950/CRID3 inhibition of inflammasome activation in CAPS disease models
Myeloid-specific expression of the Nlrp3 L351P and Nlrp3 A350V alleles in knock-in mice was shown to drive systemic inflammation accompanied by, respectively, embryonic and perinatal lethality that in both cases required Nlrp3 inflammasome activation [23]. To seek further validation of the notion that the Nlrp3 L351P mutations escape MCC950/CRID3 inhibition, we next investigated how MCC950/CRID3 treatment impacts on the CAPS phenotype of Nlrp3 L351P/+ CreT + mice. As expected, serum levels of IL-1β, IL-18, and IL-6 were significantly increased in Nlrp3 L351P/+ CreT + mice three days after tamoxifen dosing relative to the basal levels of tamoxifen-treated Nlrp3 L351P/+ CreT − littermate mice (S1A- S1C Fig and S7 Data). Moreover, tamoxifen treatment resulted in Nlrp3 L351P/+ CreT + mice presenting with substantial weight loss and mortality, with all mice being lost or requiring termination because of humane endpoints within 5 days after commencing tamoxifen treatment (Fig 6A and 6B; S5 Data). As a control, tamoxifen administration did not alter body weight or survival of Nlrp3 L351P/+ CreT − mice ( Fig  6A and 6B; S5 Data). Daily intraperitoneal (IP) injection of MCC950/CRID3 did not rescue body weight loss or mortality rates of Nlrp3 L351P/+ CreT + mice, suggesting that MCC950/ CRID3 failed to inhibit in vivo Nlrp3 L351P -induced inflammasome activation (Fig 6A and 6B; S5 Data). In agreement, IP dosing of MCC950/CRID3 failed to reduce tamoxifen-induced levels of IL-1β, IL-18, and IL-6 in serum of Nlrp3 L351P/+ CreT + mice (Fig 6C-6E; S5 Data). Thus, the CAPS-associated L351P mutation renders Nlrp3 insensitive to MCC950/CRID3 inhibition.
Our results in Nlrp3 A350V macrophages showed that this mutation subtly altered the ability of MCC950/CRID3 to inhibit inflammasome activation with potent inhibition seen only in response to LPS but not following "signal 2" triggers such as ATP and nigericin (Fig 4). To determine how this translates to the in vivo disease setting, we analyzed circulating cytokine levels and body weight loss of Nlrp3 A350V/+ CreT + mice following tamoxifen administration. Although there was a trend towards increased serum levels of IL-1β, the low measured concentrations did not reach statistical significance compared to tamoxifen-treated Nlrp3 A350V/+ CreT − littermate mice (S1D Fig and S7 Data). However, serum concentrations of IL-18 and IL-6 were significantly elevated in tamoxifen-treated Nlrp3 A350V/+ CreT + mice relative to Nlrp3 A350V/+ CreT − littermates (S1E and S1F; S7 Data). Nevertheless, levels of the latter cytokines were on average 10-to 20-fold lower than seen in tamoxifen-treated Nlrp3 L351P/+ CreT + mice, a finding that is consistent with the milder pathology and the lack of mortality associated with the Nlrp3 A350V/+ CreT + CAPS model. Consistent with published findings [25], Nlrp3 A350V/+ CreT + mice developed an inflammatory phenotype characterized by a steady weight loss of up

MCC950/CRID3 inhibition of LPS-induced cytokines in CAPS mutant mice
To complement the chronic CAPS disease models described above, we next evaluated the potency of MCC950/CRID3 in inhibiting acute Nlrp3-dependent inflammasome responses by subjecting wild-type and CAPS mutant mice to LPS-induced endotoxemia and probing the effect of MCC950/CRID3 on well-documented Nlrp3-dependent readouts such as LPSinduced elevation of serum levels of IL-1β and IL-18 [26]. As expected, Nlrp3 L351P/+ CreT − control mice presented with increased serum levels of IL-1β and IL-18 3 h after LPS challenge (Fig 7A and 7B and S6 Data). MCC950/CRID3 substantially curbed circulating IL-1β and IL-18 levels in this control group (Fig 7A and 7B), consistent with reported findings in LPS-challenged C57BL/6 mice [15]. Contrastingly, serum levels of IL-1β and IL-18 in LPS-dosed Nlrp3 L351P/+ CreT + mice were not significantly impacted by MCC950/CRID3 relative to PBS-treated Nlrp3 L351P/+ CreT + littermates (Fig 7C and 7D and S6 Data).
LPS challenge increased serum concentrations of IL-1β and IL-18 in Nlrp3 A350V/+ CreT − control mice, which were inhibited by MCC950/CRID3 (Fig 7E and 7F and S6 Data) similarly to its effect on LPS-treated Nlrp3 L351P/+ CreT − control mice (Fig 7A and 7B and S6 Data). This is not unexpected because both genotypes express solely wild-type Nlrp3 in the absence of Cre-recombinase-mediated excision of the neomycin resistance cassette that disrupts expression of the respective CAPS-associated Nlrp3 mutants. However, in marked contrast to Nlrp3 L351P -expressing CreT + mice (Fig 7C and 7D and S6 Data), MCC950/CRID3 markedly curbed circulating concentrations of IL-1β and IL-18 in Nlrp3 A350V -expressing CreT + mice (Fig 7G and 7H). These results confirm that MCC950/CRID3 potently inhibits inflammasome activation by wild-type Nlrp3 and the MWS-associated Nlrp3 A350V mutant but not the FCASassociated Nlrp3 L351P mutant.

Discussion
Activation of the NLRP3 inflammasome has been observed in many diseases, and its central role in driving pathological inflammation renders it an attractive target for therapeutic intervention [10]. Gain-of-function mutations in NLRP3 cause hereditary periodic fever syndromes  Fig 6A-6H can be found in S5 Data. Graphs show mean ± SD and represent two independent experiments. CAPS, cryopyrin-associated periodic syndrome; Cre-ERT2, Cre recombinaseestrogen receptor fusion protein; CreT, Cre-ERT2 fusion gene; CRID, Cytokine Release Inhibitory Drug; IL, interleukin; LRR, leucine-rich repeat; NBD, nucleotide-binding domain; NLR, NBD-and LRR-containing; NLRP3, NLR family, pyrin-domain-containing 3; ns, not significant.
https://doi.org/10.1371/journal.pbio.3000354.g006 NLRP3 disease mutations affect MCC950/CRID3 binding and inflammasome inhibition that are collectively referred to as CAPS, and these patients are currently treated with biologics that target secreted IL-1 [12]. As reported [27], chronic IL-1 blockade increases the risk for fatal infections and sepsis, suggesting that selective targeting of the NLRP3 inflammasome may potentially be a safer and more efficacious therapeutic strategy because it would block production of the central inflammatory mechanisms that contribute to inflammatory pathology while at the same time keeping nontargeted inflammasomes available to produce IL-1β to cope with infections.
Early studies with the sulfonylurea compound glyburide provided proof of concept that small molecules may selectively inhibit the NLRP3 inflammasome pathway without interfering with other inflammasomes [13]. Subsequently, several additional compounds have been reported to specifically inhibit the NLRP3 inflammasome pathway, but the majority of these agents have weak activity against the NLRP3 inflammasome pathway (μM IC 50 concentrations) and may target nuclear factor κB (NF-κB) signaling and other immune pathways (reviewed in [10]). MCC950/CRID3 is structurally related to glyburide [14,15] and is considered the most potent and selective inhibitor of NLRP3 inflammasome signaling reported to date. There is substantial interest in developing this chemical scaffold for the treatment of CAPS and other diseases. However, these efforts are constrained by the lack of insight into the molecular target and mechanism by which MCC950/CRID3 and related sulfonylurea molecules inhibit activation of the NLRP3 inflammasome pathway.
Making use of PAL and iBody technology as complementary chemical biology approaches, we here identified NLRP3 as the physical target of MCC950/CRID3. We further mapped the binding pocket to the central NACHT domain of NLRP3 and showed that PAL required an intact ATP/dATP binding pocket. This suggests that binding may occur at the nucleotidebinding/hydrolysis pocket of NLRP3. This hypothesis is consistent with and extends results from two recent reports that suggest the adjacent Walker B motif as the putative binding site of MCC950/CRID3 [28,29]. However, it cannot be excluded that mutations in the Walker A/B motifs may cause long-distance conformational changes that distort a more remote MCC950/ CRID3 binding pocket elsewhere in the NACHT domain. Notable in this regard is our observation that a randomly selected panel of 6 CAPS-associated gain-of-function mutations in human NLRP3 all failed to be labeled by our PAL probe, further supporting the notion that the sulfonylurea CRID binding pocket is highly sensitive for conformational changes in the protein that are imposed by mutations. Thus, further insight into the MCC950/CRID3 binding pocket and the molecular mechanism by which MCC950/CRID3 inhibits NLRP3 activation will likely require high-resolution structural analysis.
Considering the potential implications of these findings for treating CAPS with MCC950/ CRID3-based therapies, we evaluated the functional impact of MCC950/CRID3 in two reported CAPS models [23]. When knocked into the murine Nlrp3 sequence, the FCAS-associated Nlrp3 L351P mutant (corresponding to human L353P) also failed to bind to PAL-CRID3. NLRP3 disease mutations affect MCC950/CRID3 binding and inflammasome inhibition Consistent with these results, our analysis of both ex vivo-stimulated mutant macrophages and in vivo CAPS and endotoxemia models unequivocally established that MCC950/CRID3 potently inhibits inflammasome activation by wild-type Nlrp3 but not the FCAS-associated Nlrp3 L351P mutant. Surprisingly, however, we found that labeling by PAL-CRID3 to the MWS-linked Nlrp3 A350V mutant (corresponding to A352V in human NLRP3) was not significantly impacted by the mutation. MCC950/CRID3 partially inhibited LPS-induced IL-1β and IL-18 secretion from macrophages that homozygously or heterozygously expressed the Nlrp3 A350V mutant, but inhibition of the Nlrp3 A350V inflammasome was completely lost in response to "signal 2" agents ATP and nigericin. Consequently, MCC950/CRID3 significantly lowered IL-1β and IL-18 levels in serum of LPS-challenged Nlrp3 A350V knock-in mice, whereas it provided only limited protection against chronic CAPS mutation-driven body weight loss. The Cre-ERT2 recombinase expression system used with our CAPS disease models allows for controlled ubiquitous tissue expression of the mutant Nlrp3 knock-in allele upon tamoxifen treatment in adult animals. Expression of the Nlrp3 A350V allele does not induce lethality in adult mice under these conditions. However, another study [15] that relied on lysosome M-Cre-driven expression of the Nlrp3 A350V allele in cells of the myeloid lineage observed that MCC950/CRID3 rescued neonatal lethality, consistent with our observation that the Nlrp3 A350V mutant retained sensitivity to MCC950/CRID3 inhibition. This report also suggested MCC950/CRID3 inhibits LPS-induced IL-1β processing in peripheral blood mononuclear cells (PBMCs) from MWS patients. Another study showed that unlike samples from healthy controls, LPS-and LPS + ATP-induced IL-1β secretion from whole-blood samples of genetically defined CAPS patients resisted MCC950/CRID3 inhibition [30], which is consistent with our results suggesting that MCC950/CRID3-based therapies may effectively treat inflammation driven by wild-type NLRP3 but may be markedly less effective in CAPS patients. To conclude, by identifying the molecular target of MCC950/CRID3 in the NLRP3 inflammasome pathway and by evaluating its ability to inhibit CAPS mutant variants, the findings presented here provide a mechanistic framework for advancing therapeutic development of this chemical scaffold and for understanding its therapeutic potential in patients.

Ethics statement
All animal experiments were conducted with permission of the ethical committee on laboratory animal welfare of Ghent University. All procedures were performed in accordance with the guidelines and regulations associated with protocol number EC2017-090.

Synthesis of PAL-CRID3
Detailed experimental procedures for the synthesis of PAL-CRID3 are provided in the Supplementary Data section (S1 Text).

Synthesis of iBody conjugates
Experimental procedures for the preparation of the iBody conjugates have been described by [18], and details for the production of MCC950/CRID3 iBody conjugates are provided in the Supplementary Data section.

Transfections
HEK293T cells were maintained in DMEM media supplemented with 10% fetal bovine serum. HEK293T cells were transfected in 12-well plates with Lipofectamine 2000 (Thermo Fisher Scientific) according to manufacturer's instructions. All constructs, including NLRP3 cDNAs were synthesized and subcloned into pCDNA3 (+) Zeo (Thermo Fisher Scientific).

In vivo LPS challenge
6-to 12-week-old mice were intraperitoneally injected with PBS or 50 mg kg −1 MCC950/ CRID3, 30 minutes before being challenged with 40 mg kg −1 LPS (Escherichia coli, serotype 0111:B4, L-2630; Sigma-Aldrich). Mice were euthanized 3 h after LPS challenge for blood collection. At the 0 h time point before LPS challenge, blood was collected by retro-orbital bleeding.

iBody immunoprecipitation
12 × 10 6 BMDMs were harvested, and cells were washed in PBS and lysed by three cycles of freeze/thawing in PBS with 0.09% NP40 supplemented with complete protease inhibitor cocktail (4693159001; Roche Applied Science). Cell lysates were clarified by centrifugation at 14,000 rpm for 20 minutes, and supernatants were subsequently incubated with 1 μM of the indicated iBody at RT for 10 minutes. Then prewashed streptavidin conjugated beads were added, followed by overnight incubation at 4˚C. The next day, beads were washed 3 times in PBS with 0.09% NP40 buffer, and biotinylated proteins were eluted in Laemmli buffer and analyzed by western blot.

Cytokine analysis
Cytokine levels in culture medium and serum were determined by magnetic-bead-based multiplex assay using Luminex technology (Bio-Rad, Hercules, CA, USA), IL-1β ELISA (R&D Systems, Minneapolis, MN, USA) and mouse IL-1β tissue culture kit (Meso Scale Discovery, Rockville, MD, USA) according to the manufacturers' instructions.