The NLRP1 and CARD8 inflammasomes

Abstract Inflammasomes are multiprotein complexes that activate inflammatory cytokines and induce pyroptosis in response to intracellular danger‐associated signals. NLRP1 and CARD8 are related germline‐encoded pattern recognition receptors that form inflammasomes, but their activation mechanisms and biological purposes have not yet been fully established. Both NLRP1 and CARD8 undergo post‐translational autoproteolysis to generate two non‐covalently associated polypeptide chains. NLRP1 and CARD8 activators induce the proteasome‐mediated destruction of the N‐terminal fragment, liberating the C‐terminal fragment to form an inflammasome. Here, we review the danger‐associated stimuli that have been reported to activate NLRP1 and/or CARD8, including anthrax lethal toxin, Toxoplasma gondii, Shigella flexneri and the small molecule DPP8/9 inhibitor Val‐boroPro, focusing on recent mechanistic insights and highlighting unresolved questions. In addition, we discuss the recently identified disease‐associated mutations in NLRP1 and CARD8, the potential role that DPP9’s protein structure plays in inflammasome regulation, and the emerging link between NLRP1 and metabolism. Finally, we summarize all of this latest research and consider the possible biological purposes of these enigmatic inflammasomes.


| INTRODUC TI ON
Mammals express a number of germline-encoded pattern recognition receptors (PRRs) that detect and mount immune responses to pathogens. 1,2 Several of these PRRs are expressed intracellularly and, upon activation, assemble into large multiprotein complexes called inflammasomes. 3,4 Typically, an inflammasome-forming PRR recognizes a particular pathogen-associated structure or activity, oligomerizes and recruits the adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD). ASC, which consists of a pyrin domain (PYD) and a caspase activation and recruitment domain (CARD) (Figure 1), bridges either a PYD or a CARD of the activated PRR to the CARD of pro-caspase-1 (pro-CASP1). Next, pro-CASP1 undergoes proximity-induced autoproteolysis to generate an active enzyme (CASP1) that cleaves and activates inflammatory cytokines (ie IL-1β and IL-18) and gasdermin D (GSDMD). [5][6][7] The N-terminal fragment of GSDMD oligomerizes and forms pores in the cellular membrane, triggering an inflammatory form of programmed cell death called pyroptosis. It should be noted that ASC is not always required for inflammasome formation, as some CARD-containing PRRs can directly recruit pro-CASP1. [8][9][10] ASC-independent inflammasomes still cleave GSDMD and induce pyroptosis, but do not efficiently cleave and activate the inflammatory cytokines. NLRP1 (nucleotide-binding domain leucine-rich repeat pyrin domain containing 1) was the first PRR discovered to form an inflammasome. 11 In this landmark study, Martinon et al 11 described the spontaneous assembly of an "inflammasome" complex containing NLRP1 and ASC that activated pro-CASP1 in immune cell extracts.
Since that report, at least five distinct mammalian inflammasomes have been identified, and extensive research has delineated many key aspects of their activation mechanisms. 3,4 Despite being the first identified inflammasome-forming PRR, however, NLRP1 remained poorly characterized for many years, and only recent research has started to illuminate its activation mechanism and biological purpose. Here, we review the recent insights into the biology of NLRP1 inflammasome and highlight the key mysteries that remain unsolved.
Unlike the other NLRPs, hNLRP1 has a C-terminal extension containing a function-to-find domain (FIIND) and a CARD. The FIIND consists of ZU5 (found in ZO-1 and UNC5) and UPA (conserved in UNC5, PIDD and Ankyrin) subdomains and undergoes post-translational autoproteolysis after its ZU5 subdomain to generate two non-covalently associated polypeptide chains. [12][13][14] As described in detail below, FIIND autoproteolysis is required for NLRP1 inflammasome activation. 13,14 It should be noted that only a fraction (~50%) of the total NLRP1 protein undergoes autoproteolysis, 13 although it is unknown whether the remaining full-length protein, which cannot form an inflammasome, has a specific function. The C-terminal CARD, and not the N-terminal PYD, recruits ASC to form an inflammasome. 13 Interestingly, although some CARD domains can directly recruit pro-CASP1 independent of ASC, [8][9][10] the hNLRP1 CARD absolutely requires ASC to bridge the interaction with pro-CASP1. 7 CARD8 is the only other human protein with a FIIND. 12 CARD8 and hNLRP1 have similar FIIND-CARD regions, but CARD8 lacks the structured N-terminal domains found in NLRP1 (Figure 1). Like hNLRP1, CARD8 undergoes FIIND autoproteolysis 12 and the C-terminal CARD domain can form an inflammasome. 15 However, unlike hNLRP1, the CARD8 CARD directly interacts with the CARD of pro-CASP1 and does not form an ASC-containing platform. 7 The similarities and differences between hNLRP1 and CARD8 are discussed in detail below.
Rodents express homologs of NLRP1 but not CARD8. The mouse genome contains three paralogs of Nlrp1 (Nlrp1a,b,c), although Nlrp1c is predicted to be a pseudogene. 16,17 Unlike hNLRP1, mouse NLRP1A (mNLRP1A) and NLRP1B (mNLRP1B) both lack the N-terminal pyrin domain ( Figure 1) and can recruit pro-CASP1 either with or without ASC. 10,18,19 mNLRP1B is extraordinarily polymorphic, with at least five considerably different alleles present in common inbred mouse strains. 17 mNLRP1B alleles 3 and 4 are non-functional due to defective autoproteolysis and truncation F I G U R E 1 Domain architecture of the NLRP1 inflammasome proteins. NLRP1 and CARD8 protein have FIIND and CARD domains and undergo autoproteolysis between the ZU5 and UPA subdomains that comprise the FIIND. NLRP1 proteins have NACHT and LRR domains preceding the FIIND, and human NLRP1 also has an N-terminal PYD. Some rodent NLRP1 proteins are cleaved by lethal factor (LF) near their N-termini. ASC contains a PYD and a CARD, and pro-CASP1 contains a CARD preceding its catalytic p20 and p10 subunits

PYD CARD
prior to the CARD, respectively. 14,17,20 The rat genome contains one Nlrp1 gene, which, like the mouse Nlrp1 genes, does not encode an N-terminal PYD (Figure 1). At least five distinct Nlrp1 alleles exist in inbred rat strains, although polymorphisms are largely in the first 100 amino acids preceding the NACHT domain and all of the encoded proteins are functional. 20,21 To our knowledge, the role that ASC plays in the assembly of the rat NLRP1 (rNLRP1) inflammasome has not been established.

| ANTHR A X LE THAL TOXIN
Anthrax lethal toxin (LT) is a bipartite toxin consisting of the poreforming protein protective antigen (PA) and the zinc metalloprotease lethal factor (LF). PA transports LF into the host cell cytosol, where it cleaves a number of host proteins, including the mitogenactivated protein kinase kinases (MAPKKs). 22,23 In the 1990s, LF protease activity was found to trigger rapid cell death in rodent macrophages. [24][25][26][27] Notably, proteasome and N-end rule pathway inhibitors blocked LT-induced macrophage death, [27][28][29][30] indicating that the degradation of at least one protein by the N-end rule machinery was required for cell death to occur.
Interestingly, LT killed macrophages derived from some inbred rodent strains, while macrophages from other rodent strains and humans were completely resistant (Table 1). [24][25][26][27]31 The susceptibilities of mouse and rat macrophages to LT-induced death were mapped to the Nlrp1b 17 and Nlrp1 21 genes, respectively. Specifically, LT killed macrophages expressing mNlrp1b alleles 1 and 5 and rNlrp1 alleles 1 and 2, suggesting that LT was either directly or indirectly activating only those NLRP1 proteins. It should be noted that macrophage pyroptosis was found to be beneficial to the host, as LT-sensitive Nlrp1 alleles provided resistance to Bacillus anthracis infection. 32,33 As the rat NLRP1 alleles mainly differ in their first 100 amino acids, the identity of these residues appeared to be responsible for conferring susceptibility to LT. 21 Indeed, LF was soon thereafter found to directly cleave the sensitive rNLRP1 allele 2, but not the resistant rNLRP1 allele 5, in this N-terminal region 34 (Figure 1). Consistent with direct cleavage stimulating inflammasome assembly, mutation of the LF cleavage site abolished both LF proteolysis and caspase-1 activation. LF was subsequently found to directly cleave the sensitive mNLRP1B alleles, but not the resistant mNLRP1B alleles, mNLRP1A or hNLRP1. 20

| TOXOPL A S MA G OND II
Toxoplasma gondii (T gondii) is an obligate intracellular parasite that infects a wide range of warm-blooded animals, although susceptibility to infection varies widely between species and even among individuals of the same species. Generally speaking, rats and humans are far more resistant to T gondii infection than mice, 45 Toxoplasma gondii was discovered to induce rapid pyroptosis in LEW and spontaneously hypertensive (SHR) rat bone marrow-derived macrophages (BMDMs), both of which express Nlrp1 allele 5 (Table 1). 45,49,50 In contrast, T gondii induced far less pyroptosis in BN and Sprague Dawley (SD) rat BMDMs, which express Nlrp1 allele 1, and in Fischer (CDF) rat BMDMs, which express Nlrp1 allele 2 ( Table 1). It should be emphasized that T gondii-resistant rats have T gondii-sensitive macrophages and vice versa, indicating that macrophage pyroptosis is protective against infection. As anticipated by the genetics, T gondii-induced pyroptosis was indeed dependent on NLRP1, as siRNA knockdown of Nlrp1 in LEW BMDMs reduced cell death and overexpression of the Nlrp1 allele 5 in CDF macrophages increased cell death. 45 Overall, these data identified T gondii as the F I G U R E 2 LF activation of the NLRP1B inflammasome. NLRP1B undergoes post-translational autoproteolysis after its ZU5 subdomain to generate N-and C-terminal fragments that remain non-covalently associated. LF cleaves between residues K44 and L45 in the N-terminal fragment, generating an destabilized N-terminal residue. The N-end rule E3 ligase UBR2 recognizes and ubiquitinates this neo-N-terminus, inducing its proteasome-mediated degradation. The non-covalently bound C-terminal fragment is then freed to recruit and activate pro-CASP1 second pathogen-associated activator of the NLRP1 inflammasome after LT. Intriguingly, the rat Nlrp1 alleles that confer susceptibility to T gondii are precisely the opposite of those that confer susceptibility to LT, although it is unknown whether this is biologically meaningful or simply coincidence.
As noted above, mice are considerably more susceptible to T gondii infection than rats. Consistent with Nlrp1-mediated macrophage pyroptosis playing a critical role in restricting T gondii infection, T gondii induces far less (and in some assays undetectable) cell death in mouse BMDMs than in rat BMDMs. 45 and Nlrp1 −/− mice had reduced survival and higher parasite loads than control mice in response to T gondii challenge. 49,51 Thus, T gondii activates at least a few mouse NLRP1 alleles, albeit less strongly than the rat alleles. It should be noted that multiple inflammasomes can signal simultaneously, and T gondii infection was also reported to activate the NLRP3 inflammasome in mice. 51 However, the mechanistic basis of T gondii-induced NLRP3 activation has not been extensively studied.
Preliminary data suggest that the NLRP1 inflammasome may also play a role in restricting T gondii infection in humans. Perhaps most notably, polymorphisms in the human NLRP1 gene were found to be associated with congenital toxoplasmosis. 52 Surprisingly, however, this same study counterintuitively reported that shRNA-mediated knockdown of NLRP1 increased the amount of T gondii-induced cell death in MonoMac6 cells. A subsequent study found that T gondii induced inflammasome activation in human THP-1 monocytes, although the contribution of NLRP1 to this response was not evaluated. 53 The role of NLRP1, as well as CARD8, in the response to T gondii infection in humans warrants future study.
The molecular mechanism of T gondii-induced NLRP1 activation is unknown. One possibility is that T gondii secretes an effector protein that destroys NLRP1 N-terminus like LF protease or S flexneri ase. An alternative possibility is that a T gondii activity manipulates the host cell in some way, and NLRP1 senses this perturbation in host cell state ( Figure 4). Interestingly, although the NLRP1 alleles vary considerably in their sensitivity to T gondii-induced activation, all alleles tested appear to detect T gondii to some degree, unlike LF or IpaH7.8. Thus, it may be likely that T gondii triggers a more "universal" activation mechanism than the pathogen effectors that directly act on the NLRP1 protein itself. As described in detailed below, the relative responsiveness of NLRP1 alleles to T gondii infection and DPP8/9 inhibition is remarkably similar, suggesting a possible shared activation mechanism.

| THE D IPEP TIDYL PEP TIDA S E S 8 AND 9 (D PP8/9)
The small molecule Val-boroPro (VbP, Figure 4) was discovered to stimulate anti-cancer immune responses in syngeneic mouse models more than 15 years ago. 54 VbP is a non-selective inhibitor of the post-proline cleaving serine proteases, with particularly potent activity against the dipeptidyl peptidases DPP4, DPP7, DPP8 and DPP9. 58  The identity of the VbP-activated inflammasome remained unknown. However, ASC was reported to be dispensable for VbPinduced pyroptosis in human THP-1 cells and mouse RAW 264.7 macrophages, 56 suggesting that DPP8/9 inhibition activates a CARD-containing PRR that can directly recruit pro-CASP1 in those cell types. 56 As mentioned above, mNLRP1A and mNLRP1B do not require ASC to bridge to pro-CASP1, 10,18,19 making these likely candidates. 60  inflammasomes. 20 Intriguingly, the various rat NLRP1 alleles have remarkably different sensitivities to VbP, perfectly mirroring their relative sensitivities to T gondii (Table 1). Thus, it seems possible that T gondii and VbP may share the same activation mechanism, as discussed further below.
Interestingly, Johnson et al discovered that VbP induced caspase-1-dependent pyroptosis in a number of acute myeloid leukaemia (AML) cancer cell lines, including MV4;11 and OCI-AML2 cells, in addition to THP-1 cells. 15 Surprisingly, CARD8, and not hNLRP1, was found to mediate VbP-induced pyroptosis in these cells, 15 identifying CARD8 for the first time as an inflammasome-forming PRR.
Although VbP did not activate NLRP1 in these cancer cell lines, perhaps due to low NLRP1 expression, VbP was subsequently found to induce NLRP1-dependent pyroptosis in keratinocytes. 62 As such, VbP not only activates all functional rodent alleles, but also hNLRP1 and CARD8, and thus appears to be a universal NLRP1 activator. 20 The mechanistic basis of NLRP1 and CARD8 activation by DPP8/9 inhibition is an area of active research. The ectopic expression of CARD8 or mNLRP1 together with CASP1 and GSDMD renders HEK 293T sensitive to VbP, indicating that all of the other key proteins needed to execute this pyroptotic pathway are endogenously present in HEK 293T cells. hNLRP1 also requires the co-expression of ASC in order to bridge to CASP1. 7,13 Like LF, VbP induces the proteasome-dependent N-terminal degradation of the sensitive PRRs, releasing their C-terminal fragments to form inflammasomes. 15,39 Unlike LF but like T gondii, VbP does not appear to cause the direct cleavage of the N-terminal fragments. 15 Moreover, the N-terminal sequence of CARD8's neo-C-terminal fragment is NH 2 -Ser-Leu, which is not a preferred DPP8/9 substrate.
Taken together, these data indicate that DPP8/9 do not restrain the NLRP1 and CARD8 inflammasomes by direct cleavage.
It is possible that some, as yet unknown, DPP8/9 substrate(s) regulate NLRP1 and CARD8 activation. 56,62 Intriguingly, DPP9 substrate profiling studies using terminal amine isotopic labelling of substrates (TAILS) in SKOV3 cells 66 and CHOPS in THP-1 cells 65 identified very few potential protein substrates. Instead, CHOPS analysis indicated that DPP8/9 preferentially cleaved after proline residues in unstructured peptides, rather than globular proteins. Consistent with this result, DPP9 has been reported to catalyse the rate-limiting step in the catabolism of proline-containing peptides generated by the proteasome. 67 Based on these data, it is tempting to speculate that inhibition of DPP8/9-mediated peptide cleavage induces a cellular perturbation (perhaps the same one that T gondii induces) that is indirectly sensed by NLRP1 and CARD8 (Figure 4). 20 Future studies are needed to explore this hypothesis and clarify the function of DPP8/9's catalytic activity.
In addition to its catalytic activity, DPP9's protein structure also appears to play a role in restraining inflammasome activation. Notably, DPP9 (as well as DPP8) was recently discovered to directly bind the FIINDs of both hNLRP1 and CARD8. 62,68 The catalytic activity of DPP9 was not required for these interactions, as the catalytically dead S759A mutant DPP9 still bound to both hNLRP1 and CARD8. Despite these similarities, however, the hNLRP1-DPP9 and CARD8-DPP9 interactions appear to be remarkably distinct (Table 2). First, ectopically expressed wildtype and autoproteolysis-defective mutant CARD8 bound equally well to DPP9 in HEK 293T cells. 62,68 In stark contrast, autoproteolysis-defective mutant hNLRP1 exhibited significantly impaired binding in this assay. Thus, autoproteolysis is required only for the hNLRP1-DPP9 interaction. Second, DPP8/9 inhibitors partly reduced the binding of DPP9 to hNLRP1, but not to CARD8 62,68 ( Figure 5). These data strongly suggest that the hNLRP1 binding interface involves at least some surface that is proximal to the DPP9 active site, whereas the CARD8 binding interface appears to be entirely spatially distant from the DPP9 active site. Consistent with this premise, an extended activity-based probe (MW = 1087) was found to react with the catalytic serine of DPP9 without displacing the CARD8-DPP9 interaction. 68 Interestingly, it should be noted that mNLRP1B was also reported to be a direct binding partner of DPP9. 68 Surprisingly, however, the mNLRP1B-DPP9 interaction more closely resembled the CARD8-DPP9 interaction than the hNLRP1-DPP9 interaction, as it did not require FIIND autoproteolysis and was not disrupted by VbP. 68 Although the functional importance of these binding interactions has not yet been full established, preliminary data suggest that DPP9 binding restrains hNLRP1 inflammasome formation. For example, the expression hNLRP1 and ASC in DPP8/9 −/− , but not wild-type, HEK 293T cells caused the formation of spontaneous ASC specks. In this system, stable expression of wild-type DPP9, which both binds to hNLRP1 and has catalytic activity, completely abrogated speck formation. Notably, the stable expression of DPP9 S759A, which binds to hNLRP1 but is catalytically inactive, partially rescued speck formation. 62 Consistent with these data, a hNLRP1 P1214R mutant protein (discussed below) that is unable to bind to DPP9 spontaneously assembles into an inflammasome. 62,69 Additional research is needed to elucidate how DPP9 binding prevents inflammasome activation at a molecular level, but it seems plausible that DPP9 may either stabilize the association of the two NLRP1 fragments or help regulate non-inflammatory, basal NLRP1 protein turnover.
Regardless, based on these data, it appears that DPP8/9 inhibitors likely activate hNLRP1 by both inhibiting DPP8/9 activity and disrupting the DPP9-hNLRP1 interaction ( Figure 5).

| ME TABOLIC INHIB ITOR S
Mogridge and co-workers, noting the link between infectious disease and host metabolism, investigated the relationship between energy stress and NLRP1B activation. T gondii and DPP8/9 inhibitors, 2DG plus sodium azide activated NLRP1B without direct N-terminal cleavage. 70,71 Although the molecular details leading to NLRP1B activation remain largely unknown, the authors speculated that 2DG plus sodium azide activates NLRP1B via depletion of cytosolic ATP. Supporting this hypothesis, cellular ATP levels were inversely correlated with inflammasome activation, and other perturbations that lower ATP, including hypoxia and glucose-free media, also activated NLRP1B in the HT1080 system. As NACHT domains bind ATP, 72 the authors postulated that NLRP1B might directly sense ATP levels. Intriguingly, mutations in the Walker A site of the NACHT domain, which plays a critical role in ATP binding, generated a constitutively active NLRP1B protein. 70 Thus, it is possible that a loss of ATP in the NACHT domain results in inflammasome formation. However, it is also possible that the Walker A mutation simply destabilizes the N-terminus of the protein, leading to its increased degradation by the proteasome.
Recently, the Mogridge lab reported that high doses of 2DG, Listeria monocytogenes and S flexneri activate the NLRP1B inflammasome in RAW 264.7 cells. 61 Like 2DG, L monocytogenes and S flexneri reduced cytosolic ATP levels, and the authors thus speculated that these stimuli may all activate NLRP1B via ATP depletion.
However, it is possible that other bacteria-associated processes, including, for example, direct ubiquitination of NLRP1B by S flexneri IpaH7.8, 40 are triggering this inflammasome assembly. Nevertheless, these findings suggest that NLRP1B may sense a metabolic disturbance, and perhaps this same disturbance is induced by DPP8/9 inhibition and T gondii (Figure 4).

| NLRP1 AND C ARD8 IN HE ALTH AND DISE A SE
The first hyperactivating NLRP1 mutation was discovered in mice. 19 In this study, Masters  Consistent with this risk, PBMCs from subjects with this haplotype released greater amounts of processed IL-1β relative to the reference haplotype. 82 Interestingly, NLRP1 M1184V underwent FIIND autoprocessing to a greater extent than wild-type NLRP1 when overexpressed in HEK 293T cells, suggesting a potential molecular mechanism for increased activity. 13 However, it should be noted that this polymorphism is not sufficient to cause disease on its own, and thus, additional factors also contribute to the development of autoimmune disease in these individuals.
A number of rare gain-of-function mutations in the human NLRP1 gene were discovered to cause skin inflammatory and cancer susceptibility syndromes in 2016. 38  seems likely that these PYD mutations destabilize the N-terminal fragment, increasing its susceptibility to degradation by protein quality control pathways. 44 The molecular basis of inflammasome activation by the FKLC deletion has not been extensively studied, but it might similarly destabilize the N-terminal fragment or in some other way weaken its autoinhibitory activity.
A subsequent analysis of three patients from two unrelated families presenting with autoinflammation with arthritis and dyskeratosis (AIADK) identified two additional mutations in NLRP1 (R726W and P1214R). 69 In addition, a homozygous gain-of-function mutation in NLRP1 (T755N) was found in siblings with a syndromic form of juvenile-onset recurrent respiratory papillomatosis (JRRP). 83 As mentioned above, the P1214R mutation was found to abrogate binding to DPP9. 62 Importantly, these data strongly suggesting that DPP9 binding indeed serves to stabilize the autoinhibited form of NLRP1 and that disruption of this binding interaction leads to inflammasome activation ( Figure 6B). The mechanistic basis of the NLRP1 R726W and T755N mutations have not been extensively studied, but N-terminal destabilization or weakened autoinhibitory activity are likely to be involved. anti-cancer potential. As mentioned above, VbP itself induces anti-cancer responses in syngeneic mouse models. 54,55 It is possible that a more selective DPP8/9 inhibitor, or specific combinations of DPP8/9 inhibitors with other agents, will further increase the efficacy of this immuno-oncology strategy. In addition, DPP8/9 inhibitors directly kill cancer cell expressing the key inflammasome components. For example, VbP induces CARD8-mediated pyroptosis in AML cells in vivo, slowing cancer progression. 15 Thus, the anti-cancer potential of NLRP1 and CARD8 inflammasome activation also warrants further study. and CARD8 proteins detect. Interestingly, these two groups suggest starkly different biological purposes of the NLRP1 inflammasome.

| CON CLUS I ON S AND FUTURE DIREC TIONS
As described above, the direct activators raise the possibility that NLRP1 exists as a "molecular decoy" for other innate immune receptors ( Figure 3). This model proposes that a number of distinct pathogen effectors have evolved to degrade host-derived proteins, including the NLR protein family, that normally inhibit pathogen replication. However, these effectors also accidently destroy NLRP1's N-terminus, which closely resembles the intended targets, and trigger immune responses. This model also offers a possible explanation for highly polymorphic nature of NLRP1, as a decoy protein that detects a wide variety of constantly evolving effectors would be under intense selection pressure. It should be noted that decoy receptors have been observed in plants, 96,97 and thus, this mechanism is not entirely unprecedented. However, more research is needed to confirm that NLRP1 really acts as a decoy. Most importantly, there is no evidence yet that the pathogen effectors that directly degrade NLRP1 also degrade other NLR proteins (ie the intended targets), as this model predicts. In fact, LF itself cleaves NLRP1 in an unstructured region that is not present in other NLRs, arguing that LF was not evolved to destroy NLRs. Of course, it is possible that the intended targets include proteins in addition to NLRs and that the relationships between these targets and the NLRP1 decoy have not yet been discovered.
Alternatively, the indirect activators suggest that NLRP1's principal function might be to monitor cellular homeostasis ( Figure 4).
Specifically, VbP and T gondii potentially disturb homeostasis in the same way, which in turn activates an unknown host E3 ligase to degrade the NLRP1 N-terminus. 20 Given the established relationships between NLRP1, 75 DPP enzymes 98 and T gondii 99 with metabolism, a provocative possibility is that interference with some specific aspect of cell metabolism initiates inflammasome assembly. Notably, the reports that glycolysis and oxidative phosphorylation inhibitors activate mNLRP1B in at least some contexts further supports this idea. 61,70,71 More research is needed to identify this potential perturbation and the responsive E3 ligase. On that note, it will also be important to determine the molecular features of the very different NLRP1 and CARD8 N-termini that mediate E3 ligase recognition. It is tempting to speculate that the general recognition features are similar for both proteins, but that the domains of NLRP1 modulate its accessibility. As NLRP1 forms an ASC-containing inflammasome that likely generates a more intense immune reponse, 7 it is not unlikely that more elements regulate NLRP1 activation than CARD8 activation.
In summary, a number of recent studies have significantly advanced our understanding of the NLRP1 and CARD8 inflammasomes. However, several unresolved mysteries remain, including the most important one of all: the biological purpose of these proteins. Intriguingly, some lines of evidence have suggested that NLRP1 may act as a molecular decoy to guard other innate immune receptors, while others have indicated that NLRP1 might monitor the cell's metabolic state. It will be of great interest to further explore these possibilities and ultimately define the role that these inflammasomes play in host defence.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
DAB, CYT, and ARG wrote this review.