Inflammasomes are cytosolic multi-protein complexes that form in response to infectious or injurious challenges. Inflammasomes control the activity of caspase-1, which is essential for the maturation and release of IL-1β family cytokines. The NLRP1, IPAF and AIM2 inflammasomes recognize specific substances, while the NLRP3 inflammasome responds to many structurally and chemically diverse triggers. Here, we discuss the critical roles of priming and lysosomal damage in NLRP3 inflammasome activation.
Transmembrane and cytosolic signaling receptors of the innate immune system
Our knowledge of the molecular mechanisms by which host cells sense infections and tissue damage has grown considerably in recent years. It has become evident that immune cells and also many non-immune cells express a limited number of germ-line encoded signaling receptors that jointly are able to recognize virtually all forms of infectious agents and alterations in tissue homeostasis 1. Transmembrane signaling receptors of the TLR or C-type lectin receptor families are triggered when the cell exterior contacts stimulatory substances or when these substances are taken up in endocytic compartments. The cytoplasm hosts different sets of signaling receptors such as members of the RIG-I like helicase, the nucleotide-binding domain leucine-rich repeat (NB-LRR)- containing NLR (Nod-like receptor) and the PYRIN domain and HIN200 domain containing PYHIN protein families. These receptors can be activated by a range of microbial-derived substances or altered endogenous molecules that appear under tissue stress situations 2. Most innate immune signaling receptors induce a specific transcriptional response leading to host factors that act against infectious pathogens and that lead to the reconstitution of tissue homeostasis. Some members of the NLR and PYHIN protein family can initiate a proteolytic cascade that controls the activity of proinflammatory cytokines of the IL-1β family. These proteins have been observed to form larger multimolecular signaling complexes as the initial step of activation and were hence termed “inflammasomes” 3.
Inflammasomes and the control of inflammatory caspase-1
To date, the NLR protein family members NLRP1, NLRP3 and NLRC4, and AIM2 (a member of the PYHIN protein family) have been identified as capable of forming inflammasomes 3, 4. Inflammasomes control the activity of the pro-inflammatory caspase-1, which is constitutively expressed as an inactive pro-form in the cytosol. Caspase-1 contains a caspase recruitment domain (CARD) that can either interact directly with the CARD of an NLR (i.e. NLRC4) or with the CARD of the adapter protein apoptosis-associated speck-like protein containing a CARD (ASC). ASC, as a bipartite protein, additionally contains a PYRIN domain that associates with the PYRIN domain of both AIM2 and the NLRP family members. The CARD–CARD interaction results in the autocatalytic processing of pro-caspase-1 and, in turn, the formation of the active form of caspase-1 3. The main substrates for active caspase-1 are the cytokines of the IL-1β family, IL-1β or IL-18. Similar to caspase-1, these proteins are expressed in the cytosol as biologically inactive pro-forms and caspase-1 mediated cleavage leads to their activation and release into the extracellular environment. Thus, in contrast to most other cytokines, the activation of the IL-1 family of cytokines is not only controlled by transcription but also by a proteolytic pathway that is under the control of inflammasomes 5.
NLRP3 inflammasome activation requires priming
It is commonly appreciated that the induction of the substrate for caspase-1, i.e. pro-IL-1β, requires a priming step for its production. However, the fact that the NLRP3 inflammasome also needs a priming step for its activation has not been fully appreciated and this has led to significant confusion in the field. To date, the prevailing notion regarding NLRP3 inflammasome activation has been that a priming signal acts to induce the expression of pro-IL-1β (signal I) and an additional stimulus results in the activation of the NLRP3 inflammasome itself (signal II). Of note, when studying inflammasome activation, by measuring the cleavage or the release of IL-1β, the priming signal (signal I) cannot be dissected from the actual activation signal (signal II), since priming (signal I) is always required to induce pro-IL-1β expression. The upstream proteins caspase-1 and ASC, however, are not expressed at limiting amounts and these genes are also not significantly induced by classical priming stimuli. Thus, by assessing caspase-1 cleavage or ASC speck formation directly one can evaluate the activity of a “signal II stimulus” independently of signal I.
One of the best-studied signals that activate the NLRP3 inflammasome pathway is LPS in combination with ATP, which triggers the release of cleaved, bioactive IL-1β. In this context, LPS primes the transcription of pro-IL-1β and ATP activates NLRP3 by inducing potassium efflux via the P2X7 receptor and possibly by inducing other unknown mechanisms 6. However, it was noted that LPS was not only required for the induction of pro-IL-1β (signal I), but also for the activation of NLRP3 (signal II) when NLRP3 activity was assessed by caspase-1 cleavage, i.e. cells stimulated with ATP alone, in the absence of LPS failed to induce ASC speck formation or cleavage of caspase-1 7–9.
One hypothesis that was put forth to explain this two signal activation requirement was that NLRP3 could act as a direct sensor for LPS and that ATP was required to form a large channel (consisting of P2X7 and pannexin-1) allowing the access of LPS to the cytosol of cells 10. This concept would suggest that LPS (and many other signals leading to cellular priming) would directly engage NLRP3 in the cytosol and that ATP would only facilitate its translocation from the endosome into the cytosol. However, it was also noted that NLRP3 is a transcriptionally regulated protein 11, which opened up the possibility that priming of the NLRP3 inflammasome itself was required for activation. Indeed, recent studies documented that LPS triggered de novo protein synthesis via TLR4, and this was a critical prerequisite for NLRP3 inflammasome activation by ATP, pore-forming toxins and various crystals 8, 9. This requirement for transcriptionally active signaling receptors for NLRP3 activation was not only limited to LPS in combination with an NLRP3 stimulus. In fact all other tested priming agents, such as TLR agonists (e.g. R848), cytokines (e.g. TNF-α or IL-1β) and NLR ligands (MDP) could act as priming agents. In each case, the respective PRR or cytokine receptor was essential for NLRP3 activation when the ligands where applied in combination with ATP, pore-forming toxins or crystals 8.
The mechanisms behind this absolute requirement for priming of the NLRP3 inflammasome in vitro turned out to be quite simple. The expression level of NLRP3 itself appears to be the limiting factor, since heterologous expression of NLRP3 in macrophages can entirely substitute for the priming signals. Cells that overexpress NLRP3 respond to ATP, pore-forming toxins or crystals in the absence of prior priming by a TLR or NLR activator 8. Of note, even cytokines, such as TNF-α or IL-1β, can serve as priming signals (again, via activation of the respective cytokine receptors) enabling the cells to respond to NLRP3 inflammasome activators 8, 9 (Fig. 1).
These studies suggest that the NLRP3 inflammasome is tightly controlled by the activity of a range of transcriptionally active signaling receptors that, by inducing NLRP3 expression, can license the NLRP3 inflammasome to respond to its activators. One should keep in mind that any substance or signaling pathway that is implicated as a “signal II” in NLRP3 activation should be carefully checked for its impact on NLRP3 expression itself.
Tight control of the NLRP3 inflammasome is vital to prevent potentially detrimental systemic inflammasome activation. Clinically this is exemplified by the appearance of systemic inflammation with the induction of fever and inflammation in various tissues in several so-called auto-inflammatory diseases 12. In some patients, mutations in Nlrp3 itself can lead to hyperactive NLRP3 protein with constitutive activity; however, a large proportion of the patients with classical symptoms lack any mutations within Nlrp3. Interestingly, a recent analysis identified patients with Nlrp3 promoter mutations leading to enhanced levels of Nlrp3 gene expression in vitro 13. These data underline the notion that NLRP3 levels are probably also critical for inducing inflammation in vivo. It is tempting to speculate that not only mutations in coding sequences within Nlrp3 but also promoter variants leading to supra-normal expression levels of NLRP3 could induce NLRP3 hyperexcitability producing the clinic symptoms of auto-inflammatory diseases.
Role of lysosomal destabilization in the activation of the NLRP3 inflammasome
Presently, the molecular details of NLRP3 activation are not fully understood. Given the diverse physical and chemical nature of the currently known NLRP3 activators, a direct physical interaction of NLRP3 with the different triggers appears unlikely. Indeed, recent evidence suggests that the activation of the NLRP3 inflammasome involves indirect mechanisms. Two main hypotheses have been formulated, which could explain how a single signaling receptor can be triggered by the effects that are elicited by many different substances. One hypothesis suggests that the common feature of NLRP3 activators is that they induce ROS, which, in turn, are sensed indirectly by NLRP3 leading to its activation 14, 15. Martinon further details the evidence supporting this model of NLRP3 inflammasome in the companion article 16 in this Viewpoint series. In brief, several NLRP3 inflammasome activators, such as crystals, can induce ROS in cells. The produced ROS, in turn, is proposed to lead to the generation of a potential ligand of NLRP3 or, alternatively, to modify NLRP3 or associated proteins directly. This mechanism would most likely be strictly regulated since ROS production downstream of many PRR alone is insufficient to activate the NLRP3 inflammasome itself and conversely, overproduction of ROS inactivates caspase-1 by covalent modification 17.
Another, possibly interrelated model proposes that NLRP3 activators induce lysosomal damage, which is indirectly sensed by the NLRP3 inflammasome. This hypothesis is supported by experiments demonstrating that crystalline or aggregated NLRP3 activators can induce rupture of phago-lysosomes leading to the release of the proteolytic lysosomal contents into the cytosol 18, 19. In agreement with this model, proton pump inhibitors that lead to the neutralization of lysosomal pH and thereby prevent the activation of most lysosomal proteases can effectively inhibit NLRP3 activation in response to crystalline materials. Furthermore, inhibition or lack of single cathepsins significantly, but not completely, reduces NLRP3 activation suggesting a role for the proteolytic enzymes upstream of NRLP3 19, 20. Indeed, experimental disruption of lysosomes by pharmacological or physical means can also induce NLRP3 activation in a partially pH- and protease-dependent manner 19 (Fig. 2). This model places protease activity upstream of NLRP3 and is thus reminiscent of the prevailing mechanisms by which some of the plant orthologue resistance family proteins recognize microbial infection. For example, NB-LRR can be activated after proteolytic cleavage of an NB-LRR inhibitory substrate protein leading to deinhibition. Alternatively, NB-LRR have been shown to sense proteolytic cleavage of cytoplasmic endogenous proteins whose cleavage products act as “ligands” 21. It is likely that the mechanisms of NLRP3 activation are in fact more complicated and could involve any combination of these possibilities. For example, it is conceivable that NLRP3 is inhibited by a protein that is subject to proteolytic degradation and that a ligand generated by the activity of cellular ROS would be required for full activation of the NLRP3 inflammasome (Fig. 2). In either case, NB-LRR do not directly interact with the introduced proteases. Instead, they indirectly sense proteolytic activity by surveying host proteins, a mechanism which has led to the so-called “guard” hypothesis 22.
Concluding remarks and future directions
The recent progress in our understanding of how the pro-inflammatory cytokines of the IL-1β family are post-transcriptionally regulated and thereby activated by inflammasomes has led to an exciting new field of research which could directly impact on the development of novel anti-inflammatory agents. Recent studies on the NLRP3 inflammasome suggest that NLRP3 could be critically involved in the development of a variety of inflammatory diseases. Probably the best evidence stems from auto-inflammatory syndromes, which are characterized by recurrent episodes of systemic inflammation that can induce debilitating pathologies of many organ systems and lead to the early death of affected individuals 12. The realization that the devastating recurrent inflammation is based on genetic alterations in inflammasome loci has changed the therapy and in many cases the patients' lives. While the evidence for the involvement of the NLRP3 inflammasome in other inflammatory diseases remains fragmentary at best, the prospect that targeted therapy could also influence pathology in a diverse range of acute and inflammatory diseases is very attractive.
Funding from the National Institutes of Health (AI-065483 and AI-083713) (to E.L.), the German Research Foundation (SFB704) and the European Research Council ERC-2009-StG 243046 (both to V.H.) is acknowledged.
Conflict of interest: The authors declare no financial or commercial conflict of interest.