[1] 10w Limiting inflammation-the negative regulation of NF-κB and the NLRP3 inflammasome

[1] 129w A properly mounted immune response is indispensable for recognizing and eliminating danger arising from foreign invaders and tissue trauma. However, the 'inflammatory fire' kindled by the host response must be tightly controlled to prevent it from spreading and causing irreparable damage. Accordingly, acute inflammation is self-limiting and is normally attenuated after elimination of noxious stimuli, restoration of homeostasis and initiation of tissue repair. However, unresolved inflammation may lead to the development of chronic autoimmune and degenerative diseases and cancer. Here, we discuss the key molecular mechanisms that contribute to the self-limiting nature of inflammatory signaling, with emphasis on the negative regulation of the NF-κB pathway and the NLRP3 inflammasome. Understanding these negative regulatory mechanisms should facilitate the development of much-needed therapeutic strategies for treatment of inflammatory and autoimmune pathologies.

[1] 180w Although inflammation is essential for the initiation of protective immunity, uncontrolled inflammation can wreak havoc by leading to tissue destruction and the development of inflammatory and autoimmune diseases. Complex multilayered negative autoregulatory loops have evolved to restrain runaway inflammation 1 . The initiation of inflammation depends on pattern-recognition receptors, whose engagement by pathogen-associated molecular patterns or certain host-cell-derived molecules, termed alarmins or danger-associated molecular patterns, leads to the production of key inflammatory cytokines such as IL-1 and TNF family members, whereas antiinflammatory cytokines, including IL-10 and TGF-β, produced by regulatory T cells and myeloid-derived suppressor cells, negatively control inflammation 2 . Inflammatory cytokines can also be neutralized by receptor antagonists or decoy receptors 3 . Beyond the intricate checks and controls mediated by cell-cell interactions, the intracellular signaling networks that drive inflammation are also subject to negative regulation, which is frequently self-triggered. In this Review, we focus on the negative regulation of signaling involving the transcription factor nuclear factor-κB (NF-κB) and Nod-like receptor pyrin domain containing 3 (NLRP3) inflammasome, which are the two archetypical molecular drivers of the inflammatory response.

[2] 29w The roles of other molecular mechanisms (such as those of lipoxins, resolvins, galectins and neuromodulators) that mediate the resolution phase of inflammation have been extensively reviewed elsewhere 1,4 .

[3] 123w Proinflammatory NF-kB signaling Pattern-recognition receptors initiate tightly regulated signaling cascades culminating in the activation of specific transcriptional networks 5 . Here, we focus on the NF-κB pathway as a paradigm of inflammatory signaling, but several other signaling pathways, such as the mitogen-activated protein kinase or JAK-STAT signaling pathways, also play important roles in orchestrating inflammatory responses 6,7 . NF-κB-mediated gene expression regulates many cellular processes, including production of inflammatory mediators, cell proliferation and survival, differentiation of effector and regulatory T cells and dendritic-cell maturation. Unsurprisingly, dysregulation of NF-κB signaling plays a major role in many inflammatory and autoimmune diseases as well as cancer 8,9 . Consequently, the NF-κB pathway has been the focus of intensive research exploring its molecular mechanisms and pharmacological targeting.

[4] 314w The NF-κB family of transcription factors consists of five Relhomology-containing proteins (cRel, RelA, RelB, NF-κB1 and NF-κB2) that form numerous homo-and heterodimers, which are normally retained in the cytoplasm through binding to inhibitors of NF-κB (IκB). Although the details of upstream signaling events differ among individual receptors, pattern-recognition receptor-, cytokineor antigen-receptor-mediated signaling cascades eventually culminate in activation of the IκB kinase (IKK) complex, which subsequently phosphorylates the inhibitory cytoplasmic NF-κB chaperone IκBα and related proteins. This phosphorylation results in K48-linked IκBα ubiquitination and proteasome-mediated degradation, thus allowing NF-κB dimers to translocate to the nucleus and initiate specific gene R e v i e w transcription 8 (Fig. 1). Notably, NF-κB limits its activation by upregulating the expression of several NF-κB inhibitors, including IκBα. This phenomenon is nicely demonstrated by the observation that mutation of κB enhancer sites in the IκBα promoter in mice results in the development of autoimmunity, dysregulated T cell development and hypersensitivity to endotoxic shock, owing to runaway production of inflammatory mediators 10 . This negative feedback mechanism is further fine-tuned by other IκB family members, which have varying binding affinities toward NF-κB targets. For example, stimulus-specific induction of IκBε, owing to prolonged IKK stimulation, has a role distinct from that of IκBα in regulating B cell proliferation and survival 11 . Pathogen-associated molecular patterns induce expression of the IRAK-M kinase, which is believed to inhibit NF-κB signaling by binding the kinase IRAK4 and preventing its interaction with downstream signaling proteins 12 (Fig. 1). Additionally, IRAK-M initiates a second wave of NF-κB activation through the kinase MEKK3, thus resulting in expression of IκBα and other negative regulators of inflammation 13 . Importantly, glucocorticoids suppress nontypeable Haemophilus influenzae-induced pulmonary inflammation by enhancing IRAK-M expression 14 . These findings indicate that new insights into the negative feedback regulatory mechanisms of NF-κB signaling may provide a path to the development of novel anti-inflammatory strategies.

[5] 114w Beyond transcriptional upregulation of inhibitors of NF-κB signaling, alternative splicing of mRNAs encoding several NF-κB signaling proteins, including MyD88 (ref. 15), IRAK2 (ref. 16) and IKKε (ref. 17), also contribute to the negative regulation of NF-κB activation by disrupting key protein functional domains, protein-protein interactions and post-translational modifications. For example, lipopolysaccharide (LPS)-induced alternative splicing of MyD88 results in the generation of a protein (MyD88s) lacking a short intermediate domain, which can still bind the Toll-like receptor TLR4 but can no longer bind IRAK4 and therefore exerts a dominant-negative effect leading to the termination of downstream NF-κB signaling 15 (Fig. 1). Alternative splicing of MyD88 may thus be important in the context of endotoxin tolerance.

[6] 147w Inflammatory NF-κB signaling is also influenced by negative feedback mechanisms regulating the stability and translation of certain mRNA transcripts encoding key signaling molecules (Fig. 1). The microRNA (miRNA) miR-146a is induced by LPS in an NF-κB-dependent manner in human monocytes and subsequently downregulates the signaling proteins TRAF6 and IRAK1 (ref. 18). Mice deficient in miR-146a are sensitive to LPS-induced shock and, with age, develop autoimmunity, progressive myeloproliferation and hematopoietic tumors that are dependent on NF-κB 19,20 . The miRNA miR-302b is strongly upregulated in alveolar macrophages after infection with Gram-negative bacteria 21 and inhibits inflammatory responses by targeting IRAK4 and blocking NF-κB activation both in vitro and in vivo. Furthermore, injection of a miR-302b mimetic protects mice against Pseudomonas aeruginosainduced mortality 21 , thus demonstrating a potential for therapeutic intervention. The roles of miRNAs in regulating inflammatory signaling have been described in other comprehensive reviews 22,23 .

[7] 27w mRNA transcripts are also destabilized by the actions of several RNA-binding proteins that recruit the nucleolytic machinery to specific mRNAs and consequently induce transcript degradation (Fig. 1).

[8] 209w Tristetraprolin is an RNA-binding protein that is induced by various inflammatory stimuli, such as LPS, as well as by anti-inflammatory cytokines. Tristetraprolin binds AU-rich elements in the 3′ untranslated regions of specific mRNAs, including those encoding the cytokines TNF and IL-23, and consequently promotes mRNA decay 24,25 . Furthermore, genetic deletion of tristetraprolin in mice results in the development of arthritis, cachexia and autoimmunity, all of which can be prevented by administration of TNF-neutralizing antibodies or by IL-23 deficiency 25,26 . The RNA-binding protein Roquin mediates decay of TNF mRNA, as well as mRNAs for co-stimulatory receptors, cytokines and transcription factors, and it regulates follicular T helper cells and represses autoimmunity 27,28 . The endonuclease Regnase-1 (also known as MCPIP-1) shares an overlapping set of target mRNAs with Roquin but functions in distinct subcellular locations and specifically targets translationally active mRNAs, whereas Roquin controls translationally inactive mRNAs 27 . In agreement with the critical anti-inflammatory functions of these proteins, Roquin-or Regnase-1-deficient mice develop severe inflammatory phenotypes [28][29][30] . Interestingly, antibiotic treatment greatly improves the survival of Regnase-1-deficient mice, thus suggesting that the microbiota of mucosal surfaces activates an inflammatory immune response that cannot be restrained in the absence of Regnase-1 (ref. 30). In contrast, Regnase-1-haploinsufficient mice are healthy and

[9] 279w TRADD TRAF2 cIAP1/2 IκBα IRAK-M TRAF1 miR-146a miR-302b A20 TTP Regnase-1 ... ... Regnase-1 Roquin MyD88s IRAK-M TRAF1 TTP IκBα TNFR1 IL-1 IL-6 TNF IL-23 RIPK1 IRAK4 AAA IL6 AAA IL23A AAA TNF AAA miR-302b TRAF6 AAA miR-146a IRAK1 TLR/IL-1R AAA MyD88 IRAK1/2/4 TRAF6 TRAF6 IKKγ IKKα IKKβ NF-κB NF-κB TCR PKC CARD11 P Bcl10 MALT1 P Debbie Maizels/Springer Nature Toll-like receptor (TLR) and IL-1R, or TCR results in the formation of receptor-specific multiprotein signaling complexes that ultimately converge on the recruitment and activation of the IKK complex, consisting of the catalytic subunits IKKα and IKKβ, as well as a regulatory subunit NEMO/IKKγ (for simplicity, some intermediate steps are omitted). IKKβ in turn phosphorylates IκBα, tagging it for ubiquitination and proteasomal degradation, thus allowing NF-κB dimers to translocate to the nucleus and initiate expression of several inflammatory mediators such as IL-1, IL-6, TNF and IL-23. NF-κB also induces expression of multiple proteins that mediate negative feedback regulation of NF-κB signaling at multiple levels: IκBα, which sequesters NF-κB; IRAK-M, which inhibits the MyD88 signalosome; and TRAF1, which interferes with the recruitment of the IKK complex. Additionally, NF-κB induces the expression of miRNAs and RNA-binding proteins (tristetraprolin (TTP), Regnase-1 and Roquin) that regulate the stability or translation of mRNAs encoding specific signaling adaptors of the NF-κB pathway, as well as numerous proinflammatory mediators. In the case of TCR signaling, the activity of Regnase-1 and Roquin is itself regulated via proteolytic cleavage of Regnase-1 and Roquin by MALT1. Furthermore, LPS-and IL-1-induced alternative splicing of MyD88 results in the formation of MyD88s, which competes with MyD88 for TLR4 binding and can no longer bind IRAK4, thus resulting in a dominant-negative effect. P, phosphate.

[10] 172w Phosphorylation of key NF-κB signaling molecules often positively mediates signal transduction by inducing protein conformational changes that release (auto)inhibition and enable enzymatic activity (for example, phosphorylation of the IRAK and IKK kinases) or interaction with other proteins (for example, phosphorylation of IκBα or the caspase-recruitment-domain protein CARD11). Consequently, several phosphatases (for example, protein phosphatase 4, tyrosine phosphatase 1B and Wip1) have been shown to dampen the inflammatory cascade by dephosphorylating specific signaling proteins and transcription factors [35][36][37] . Notably, NF-κB-activating kinases may also exert an inhibitory effect, as illustrated by the IKKα-mediated phosphorylation of RelA and c-Rel, thereby increasing their turnover 38 . In agreement with this finding, genetic inactivation of IKKα in mice results in increased inflammation and susceptibility to septic shock accompanied by enhanced macrophage activation and resistance to apoptosis 38 . IKKα also phosphorylates the ubiquitin-binding protein TAX1BP1, thus enabling assembly of the A20 ubiquitin-editing complex, which negatively regulates NF-κB signaling 39 (described below). Thus, IKKα-mediated phosphorylation represents an additional layer of negative feedback regulation of the NF-κB pathway.

[11] 349w Ubiquitination has also emerged as a crucial regulatory mechanism in signal transduction in diverse biological processes, including inflammatory signaling, and dysregulated ubiquitination events are associated with several inflammatory diseases 40 . Like phosphorylation, ubiquitination is a reversible reaction, and it is tightly controlled by the opposing actions of E3 ubiquitin ligases and DUBs. Research on NF-κB signaling has led to the description of novel types of polyubiquitination and the elucidation of the enzymes and mechanisms involved. Ubiquitin moieties on NF-κB signaling proteins can serve either as a docking platform for other proteins with specific ubiquitinbinding domains (for example, recruitment of the NF-κB modulator NEMO to proteins that are modified by K63-or M1-linked polyubiquitin chains) or as a signal targeting the ubiquitinated protein for proteasomal degradation (for example, K48 ubiquitination of IκBα (ref. 41)). It is therefore unsurprising that interference with protein ubiquitination is a major mechanism for negative regulation of NF-κB signaling. As such, the adaptor protein TRAF1 inhibits TLR4-induced NF-κB signaling by binding to the linear ubiquitin-chain assembly complex (LUBAC), the E3 ubiquitin ligase complex that mediates M1-linked ubiquitination and consequently interferes with M1 ubiquitination and recruitment of NEMO to the TLR4 signaling complex 42 (Fig. 1). Moreover, a TRAF1 single-nucleotide polymorphism that decreases TRAF1 expression in monocytes is associated with increased risk of rheumatoid arthritis in humans 42 . Interestingly, mice carrying a mutation in the Sharpin gene, encoding one of the LUBAC components, develop dermatitis, multiorgan inflammation and immunological dysregulation 43 , although NF-κB activation induced by the stimulatory molecules TNF, IL-1β or CD40L is compromised 44 . This seemingly paradoxical observation is explained by the high sensitivity of Sharpin-deficient cells to TNF-induced cell death, which potentially stimulates the initiation of an inflammatory cascade 44 . However, LUBAC has been shown to coordinate the development and homeostasis of regulatory T cells in an NF-κB-dependent manner, thus indicating how inflammatory signaling pathways, perhaps somewhat counterintuitively, can also attenuate the inflammatory response 45 . Together, these examples illustrate the intertwined complexity of cellular pathways regulating inflammation and the inherent difficulties that this complexity may pose in therapeutic intervention.

[12] 131w OTULIN is a DUB that removes M1-linked ubiquitin chains generated by LUBAC in response to various inflammatory stimuli [46][47][48] . OTULIN knockdown enhances NF-κB activation, whereas its overexpression suppresses NF-κB signaling induced by LUBAC, TNF or polyinosinic-polycytidylic acid 47 . The underlying molecular mechanisms are still largely unclear but appear to involve the interaction of OTULIN with HOIP, the catalytic component of LUBAC 49 (Fig. 2). Moreover, OTULIN has been proposed to remove M1-linked ubiquitin chains from several NF-κB signaling proteins. Conditional deletion of Otulin in myeloid cells in mice results in severe TNFdependent inflammation, whereas, in humans, hypomorphic homozygous mutations of OTULIN that result in defective M1 deubiquitination and increased inflammatory signaling lead to a severe autoinflammatory condition that can be managed by treatment with antibodies to TNF 50,51 .

[13] 125w Another DUB that has been implicated in the negative regulation of NF-κB signaling is CYLD, which removes K63-and M1-linked polyubiquitin chains from several signaling mediators and thus dampens NF-κB-dependent gene expression [52][53][54] (Fig. 2). Similarly to OTULIN, CYLD interacts with HOIP. In the case of CYLD, this interaction requires SPATA2, a protein that links CYLD to HOIP and activates its DUB activity [55][56][57][58] . Importantly, interaction of CYLD and OTULIN with HOIP is mutually exclusive 55 , and the exact interplay between both DUBs in regulating M1-polyubiquitin-dependent inflammatory signaling remains to be further investigated. Various CYLD genetic mouse models and the association of CYLD polymorphisms with irritable bowel disease in humans illustrate the important regulatory function of CYLD in innate and adaptive immunity 59 .

[14] 523w The most extensively studied DUB in the context of inflammation is A20 (also known as TNFAIP3). TNFAIP3 is an NF-κB-responsive gene that is induced by many inflammatory stimuli and contributes to the negative feedback regulation of NF-κB signaling 60 . Moreover, A20 DUB activity is further enhanced by its IKKβ-mediated phosphorylation 61 (Fig. 2). In some cell types, A20 also inhibits TNF-induced cell death. A20 single-nucleotide polymorphisms are associated with susceptibility to multiple inflammatory pathologies, such as psoriasis, inflammatory bowel disease, systemic lupus erythematosus and rheumatoid arthritis 62 . Importantly, haploinsufficiency of A20 due to lossof-function mutations has been described in people with early-onset autoinflammatory disease resembling Behçet's disease 63 . Cells isolated from these people have increased NF-κB activation and inflammatory signaling 63 . The key anti-inflammatory role of A20 is further illustrated in A20-deficient mice, which die prematurely, owing to severe multiorgan inflammation and cachexia 64 . Additionally, multiple studies with conditionally A20-deficient mice have illustrated the cellspecific role of A20 in controlling immunological homeostasis and R e v i e w inflammation 65 . A20 deficiency in lung epithelial cells sensitizes mice to allergen-induced asthma, and prior exposure of wild-type mice to low-dose endotoxin or farm dust suppresses allergen-induced asthma by inducing A20 in lung epithelial cells 66 . However, it remains to be demonstrated whether all these autoimmune or inflammatory phenotypes in A20-deficient mice are due to excessive activation of NF-κB rather than to other mechanisms, such as A20-mediated restriction of inflammation driven by the RIPK3 kinase 67,68 . In humans, several TNFAIP3 single-nucleotide polymorphisms abolish the protective effect of farm dust 66 , thus correlating with the intriguing observation that growing up on a dairy farm protects children against the development of asthma or allergies. Notably, mice with a conditional A20 deletion in lung epithelial cells are less sensitive to influenza A virus infection than their wild-type littermates, an effect correlating with a dampened pulmonary cytotoxic T lymphocyte response and strongly suppressed expression of the chemokine CCL2 during late stages of infection 69 . A20 deficiency in myeloid cells leads to enhanced NF-κB and inflammasome signaling and the development of an arthritis-like pathology 70,71 . Combined deletion of A20 in myeloid and intestinal epithelial cells results in colitis, owing to the enhanced sensitivity of A20-deficient intestinal epithelial cells to TNF-induced apoptosis and the disruption of the intestinal barrier against intestinal microbes, a condition that further amplifies inflammation by triggering Toll-like receptor-induced signaling 72 . These results appear paradoxical because enhanced NF-κB signaling should attenuate TNF-induced apoptosis 73 , thus suggesting that A20 may have NF-κB-independent roles. Indeed, despite A20's antiinflammatory function, transgenic overexpression of A20 in intestinal epithelial cells in IL-10-deficient mice results in early onset of severe colitis 74 , an effect that may be explained by altered expression of mucosal antimicrobial peptides and microbial colonization of the inner mucus layer, which in turn activates an IL-10-dependent anti-inflammatory process. As such, the role of A20 in inflammation is dependent on cellular context and signals, and genetic associations between A20 and human disease should not be considered a priori as an indication of decreased A20 expression or function.

[15] 191w The molecular mechanism of action of A20 also remains a matter of debate. A20 has a unique dual ubiquitin-editing enzymatic activity that can mediate the removal of K63-linked ubiquitin chains from several key substrates; in some cases (such as for the kinase RIPK1), this removal is followed by the addition of K48-linked ubiquitin and proteasomal degradation 75 (Fig. 2). However, DUB-defective or E3 ubiquitin ligase-defective A20 knock-in mice, unlike A20-deficient mice, do not develop multiorgan inflammation and perinatal lethality 61,76,77 . Moreover, cells from these mice, as compared with cells from wildtype mice, show only modest 61,76 or even no 77 differences in TNF-or LPS-induced NF-κB activation. These observations further indicate the involvement of other mechanisms in the immunoregulatory function of A20, not all of which are NF-κB related. Notably, A20 binds M1-linked polyubiquitin via its zinc-finger 7 motif, which has been suggested to mediate A20 binding to M1-ubiquitinated NEMO, thus preventing further recruitment of LUBAC or other M1binding signaling proteins 78 (Fig. 2). In line with such a mechanism, people with NEMO C-terminal-deletion mutations that prevent A20 binding experience skin and intestinal inflammation as well as ectodermal dysplasia 79 .

[16] 142w Given the involvement of DUBs in the negative regulation of inflammatory signaling, their activity is expected to be tightly regulated. For example, the miRNAs miR-486 and miR-19b directly suppress expression of CYLD and A20, respectively 80,81 . In activated lymphocytes, activation of CYLD and A20 is regulated by proteolytic cleavage mediated by MALT1, thus leading to increased antigen-receptorinduced gene expression 82,83 . Moreover, MALT1 also cleaves A20 or CYLD in nonimmune cells, such as keratinocytes and endothelial cells 84,85 , the latter of which have been associated with endothelial-cell retraction, thus leading to an acute permeability response. Pharmacologic inhibition of MALT1 may therefore have a multilevel effect by targeting several disparate steps in the overall inflammatory response 86 . The function, mechanism of action and regulation of A20 in comparison with CYLD and OTULIN has been reviewed in detail elsewhere 59 .

[17] 129w Several other ubiquitin-regulatory proteins have been implicated in the negative regulation of other inflammatory signaling steps. The ubiquitin ligase Itch targets the transcription factor RORγt, which is essential for IL-17 production, for degradation 87 and A20, respectively, thus enabling the formation and activation of the A20 ubiquitin-editing complex. A20 removes K63-linked polyubiquitin from RIPK1 and decorates it with K48-linked polyubiquitin, thus leading to the proteasomal degradation of RIPK1 and the disassembly of RIPK1 signaling complexes. A20 binds M1 polyubiquitin on IKKγ, thereby preventing M1-polyubiquitin-mediated protein-protein interactions. HOIP brings another K63-specific DUB, CYLD, to the signaling complex via its interaction with the SPATA2 adaptor protein. HOIP binds the M1specific DUB OTULIN, which in turn further regulates NF-κB signaling by removing M1-linked polyubiquitin from LUBAC and other signaling proteins. Ub, ubiquitin.

[18] 41w mice develop spontaneous colitis and increased susceptibility to colon cancer 87 . TRABID, a DUB related to A20, removes ubiquitin chains and stabilizes the histone demethylase Jmjd2d, which further regulates expression of the interleukins IL-12 and IL-23 in lymphocytes 88 .

[19] 14w Clearly, maintaining control of ubiquitination is one of the major gatekeeping mechanisms in inflammation.

[20] 559w The NLRP3 inflammasome NLRP3, which is activated by structurally and chemically diverse activators, including microparticles, ATP, cholesterol and microbial toxins 89,90 , is a key sensor of tissue damage and therefore plays a major role in activation of sterile inflammation. After stimulation, NLRP3 interacts with the adaptor ASC, which recruits and activates the effector pro-caspase-1 (Fig. 3). This interaction results in the formation of a large cytosolic, protein complex, the NLRP3 inflammasome 91 , within which pro-caspase-1 undergoes self-cleavage and activation and becomes capable of processing the interleukin precursors pro-IL-1β and pro-IL-18 into their mature and secreted forms, which are able to initiate inflammation 91 . Although NLRP3 inflammasome activation allows for restoration of homeostasis after traumatic tissue injury by stimulating damage clearance, tissue repair and regeneration 92,93 , its dysregulation drives the progression of many inflammatory, metabolic, degenerative and aging-related diseases 94,95 . It is therefore critical for NLRP3 inflammasome activity to be kept in check to prevent the development of such pathologies. NLRP3 inflammasome activation and subsequent production of bioactive IL-1β require two sequential steps: priming and activation 91,96 . Priming entails the recognition of danger-associated molecular patterns or pathogen-associated molecular patterns by Toll-like receptors, thereby resulting in activation of NF-κB, induction of pro-IL-1β and increased synthesis of NLRP3 (Fig. 3). During the more mysterious activation step, diverse extracellular stimuli, which are often referred to as NLRP3 agonists, trigger assembly of the inflammasome complex and eventual activation. However, NLRP3 agonists are structurally and functionally diverse 93 , and none of them directly bind NLRP3, thus prompting the hypothesis that all NLRP3 agonists may act via a common intermediate, with which they communicate through induction of membrane damage, potassium efflux and elevation of intracellular calcium. Most, if not all, NLRP3 agonists induce a certain type of mitochondrial damage resulting in the generation of mitochondrial signals that activate NLRP3. For instance, mitochondrial reactive oxygen species (mtROS) play a critical role in NLRP3 inflammasome activation, because chemicals that inhibit mitochondrial respiration and mtROS production, as well as The NLRP3 inflammasome and its regulation. The NLRP3 inflammasome assembly results in self-cleavage and activation of caspase-1, which in turn processes pro-IL-1β and pro-IL-18 into their mature forms that initiate inflammation. Mitochondrial damage and subsequent release or exposure of mitochondrial contents, such as mtDNA, mtROS and cardiolipin, are critical for NLRP3 inflammasome assembly. Additionally, ion fluxes (potassium efflux and calcium influx) are thought to be required for optimal activation of the NLRP3 inflammasome. Post-translational modifications of the NLRP3 inflammasome components, including BRCC3-mediated deubiquitination of NLRP3, LUBAC-dependent ubiquitination of ASC and Syk-mediated phosphorylation of ASC, are crucial for its assembly and activation. p62 is induced by LPS stimulation, which requires IKK-dependent activation of NF-κB. NLRP3-agonist-induced mitochondrial damage results in Parkin recruitment to mitochondria by the kinase PINK1, where it ubiquitinates multiple components of the mitochondrial outer-membrane proteins, thereby leading to recruitment of p62 to damaged mitochondria and delivery of damaged mitochondria to the autophagosome for eventual degradation. NLRP3 is regulated by miR-233, which decreases NLRP3 mRNA abundance and thereby inhibits NLRP3 inflammasome activation. Type I IFN binds to its receptor IFNAR and inhibits pro-IL-1β expression and/or induction of NLRP3 nitrosylation, and neurotransmitters can trigger the degradation of NLRP3. MSU, monosodium urate; ox-mtDNA, oxidized mtDNA; [Ca 2+ ] i , intracellular calcium concentration; D1R, dopamine D1 receptor; iNOS, inducible nitric oxide synthase; NO, nitric oxide.

[1] 269w The molecular mechanisms that negatively regulate inflammatory signaling are only beginning to be understood. Post-transcriptional and post-translational mechanisms that regulate NF-κB and inflammasome signaling play key roles in maintaining immunological specificity and homeostasis. The emerging field of nondegradative ubiquitination has markedly increased understanding of the complexity of cellular signal transduction and has offered new insights into negative control mechanisms. Aberrations in E3 ligase and DUB activities may have profound consequences for the host, thus resulting in severe inflammatory disease. The negative regulators themselves are regulated by a complex network of interactions that fine-tune their effects on signaling. Moreover, the same proteins may simultaneously control different R e v i e w signaling cascades or even transmit both pro-and anti-inflammatory messages. These effects are very well illustrated by NF-κB, which not only activates proinflammatory and antiapoptotic gene expression but also is involved in the cross-talk among inflammasome activation, autophagy and cell death. Thus, therapeutic inhibition of NF-κB may result in unexpected and unwanted outcomes, and these outcomes are presumably one of the main reasons why therapeutic targeting of the NF-κB pathway has failed in clinical settings. Further insight into the complex biochemical, functional and pathophysiological interplay between NF-κB and other signaling pathways may help to avoid such failures in the future. In addition, strategies to restore key endogenous cellular brakes on inflammatory signaling, such as pharmacological control of A20 expression 115 , may be safer than directly inhibiting the main drivers, such as IKKs. Further investigations in the exciting field of NF-κB and inflammasome signaling are expected to identify novel regulatory mechanisms that may offer additional opportunities for therapeutic intervention.

[1] 153w fertile, but they show increased resistance to fungal infection and an exacerbation of experimental autoimmune encephalomyelitis, both of which are dependent on signaling by the cytokine IL- 17 (ref. 31). Both Roquin and Regnase-1 collectively repress differentiation of IL-17producing helper T cells, and this effect is lost after their proteolytic inactivation by the paracaspase MALT1 in response to antigen stimulation 32,33 . MALT1 has dual activities: first, it functions as an adaptor protein that is crucial for NF-κB signaling in response to several surface receptors in immune and nonimmune cells, including T and B cell antigen receptors, and second, it possesses proteolytic activity that further fine-tunes inflammatory-gene expression by cleaving not only Roquin and Regnase but also other negative regulators of inflammatory signaling, such as the deubiquitinases (DUBs) A20 and CYLD 34 . Collectively, these data illustrate the essential roles of several transcriptional and posttranscriptional mechanisms in the negative regulation of inflammatory signaling.

[2] 123w depletion of mitochondria, prevent NLRP3 inflammasome activation 97 . However, mtROS affect a number of cellular processes and targets 96 , and whether they act directly on NLRP3 remains elusive. Furthermore, NLRP3 does not have structural similarity to redox sensors, which usually contain at least one or several highly reactive cysteine residues or an iron (or copper) center that can undergo redox cycling. One possible mechanism underlying mtROS-induced NLRP3 activation involves opening of the mtROS-sensitive ion channel TRPM2, thus resulting in calcium influx and the establishment of an ionic environment that favors NLRP3 inflammasome assembly 98 . Moreover, elevated cytosolic calcium further damages healthy mitochondria and enhances mtROS production, thereby establishing a feed-forward loop that maximizes mtROS production and mitochondrial damage 98 .

[3] 151w Another important mitochondrial signal that may be key for NLRP3 inflammasome activation is mitochondrial DNA (mtDNA), which is released by damaged mitochondria and is likely to undergo oxidation by mtROS 99 , thereby becoming distinct from nonmethylated prokaryotic or viral DNAs that do not cause NLRP3 inflammasome activation 99,100 . Oxidized mtDNA, but not its reduced form, appears to directly bind and activate NLRP3, at least during apoptosis 99 ; however, further studies are needed to conclusively validate the roles of oxidized versus normal mtDNA in NLRP3 activation in nonapoptotic cells. Additionally, the mitochondrial lipid cardiolipin, which translocates from the inner to the outer mitochondrial membrane after stimulation with NLRP3 agonists, is thought to provide a docking site for NLRP3 translocated from the endoplasmic reticulum to the mitochondria after macrophage stimulation 101 . In conclusion, all of these studies suggest that mitochondria are a key intermediate that signals NLRP3 inflammasome activation.

[4] 111w Because the NLRP3 inflammasome is a key tissue-damage sensor and activator of sterile inflammation 93 , it is of utmost importance to keep its activation in check. Because NLRP3 inflammasome activation is an 'all-or-none' event 102 , one simple way to regulate this process is to keep mitochondrial damage and/or release of mitochondrial signals under control. This task is carried out by the autophagy machinery, which, through clearance of damaged mitochondria in a process called mitophagy, prevents excessive NLRP3 inflammasome activation 97,100,103 . Although it has long been speculated that autophagy promotes clearance of damaged mitochondria, the precise molecular events involved in this process have only recently been described 104 .

[5] 268w NF-κB is the critical transcription factor inducing pro-IL-1β and upregulating NLRP3 synthesis 91 . Unexpectedly, however, pharmacologic or genetic inhibition of NF-κB exacerbates NLRP3-dependent inflammation in preclinical animal models and humans 105 . This unexpected outcome has led to the termination of several drug-development programs aiming at blocking IKK-driven NF-κB activation to prevent inflammation. Intriguingly, the autophagy adaptor p62 (also known as SQSTM1), whose expression is strongly induced during LPS priming of macrophages in an NF-κB-dependent manner 104 , functions as a multifunctional signaling hub that controls cellular homeostasis and detoxification of reactive oxygen species 106,107 . After exposure to diverse NLRP3 activators, macrophage p62 forms aggregates that are located next to mitochondria 104 (Fig. 3). Exposure to the NLRP3 inflammasome activators triggers the Parkin-ligasedependent ubiquitination of damaged mitochondria, the binding of p62 to polyubiquitin chains via its ubiquitin-association domain, p62 interaction with the autophagosome docking protein LC3 via its LC3-interacting domain and shuttling of damaged mitochondria that are coated with p62 into newly formed autophagosomes 104 . In agreement with the notion that damaged mitochondria are delivered to autophagosomes by p62, macrophages deficient in Parkin, p62 or the E1 enzyme Atg7 exhibit excessive, long-lasting NLRP3 inflammasome activation, and mice with myeloid-specific p62 or Atg7 deficiencies develop severe IL-1β-mediated immunopathologies 104 . Importantly, elimination of mitochondrial signals (for example, mtDNA) prevents excessive IL-1β production, thereby confirming the importance of mitochondria in NLRP3 inflammasome activation and establishing the NF-κB-p62-mitophagy axis as the key negative regulatory mechanism that keeps NLRP3 inflammasome activation in check to produce a beneficial inflammatory response that promotes pathogen clearance and favors tissue repair 104 .

[6] 140w In addition to the cell-intrinsic mechanisms described above, NLRP3 inflammasome activity is also subject to cell-extrinsic negative regulation. Effector and memory CD4 + T cells can directly inhibit NLRP3 inflammasome activation, probably through a CD40-ligandmediated inhibitory mechanism dependent on cell-cell contact 108 . Type I interferon (IFN) also regulates IL-1β production 109 . Sensing of type I IFN by macrophages before their exposure to priming signals, such as LPS, inhibits NLRP3-inflammasome activation by decreasing pro-IL-1β expression via a pathway involving IL-10, the IL-10 receptor and the signaling molecule STAT3 (ref. 109). Moreover, type I IFN-mediated induction of 25-hydroxycholestorolase, which converts cholesterol to 25-hydroxycholesterol, suppresses pro-IL-1β transcription 110 . Finally, activation of inducible nitric oxide synthase by IFN-β may also contribute to type I interferon-mediated inhibition of NLRP3-inflammasome activation via nitric oxide-mediated thiol nitrosylation of NLRP3 and/or mitochondrial stabilization 111 .

[7] 142w Finally, NLRP3 is also subject to regulation at the mRNA level. The miRNA miR-223 inhibits NLRP3-inflammasome activation by decreasing NLRP3 expression 112 . Interestingly, miR-223 is expressed in a cell-type-specific manner among different types of innate immune cells; the lowest expression occurs in dendritic cells, and the highest expression occurs in neutrophils 112 , thus potentially explaining why dendritic cells are more sensitive to NLRP3 activating stimuli, whereas neutrophils largely rely on NLRP3-inflammasome-independent mechanisms for pro-IL-1β maturation 113 . In addition to components of the immune system, dopamine, a neurotransmitter, negatively regulates NLRP3 inflammasome activity 114 . Signaling through the dopamine D1 receptor in macrophages activates cAMP, which binds NLRP3 and promotes its degradation via ubiquitination mediated by the E3 ubiquitin ligase MARCH7 (ref. 114). It remains to be determined whether other neurotransmitters with immunological modulating capabilities may also regulate NLRP3-inflammasome activity.