Nucleotide-binding oligomerization domain (NOD)-containing protein-like receptors (NLRs) are a recently discovered class of innate immune receptors that play a crucial role in initiating the inflammatory response following pathogen recognition. Some NLRs form the framework for cytosolic platforms called inflammasomes, which orchestrate the early inflammatory process via IL-1β activation. Mutations and polymorphisms in NLR-coding genes or in genetic loci encoding inflammasome-related proteins correlate with a variety of autoinflammatory diseases. Moreover, the activity of certain inflammasomes is associated with susceptibility to infections as well as autoimmunity and tumorigenesis. In this review, we will discuss how identifying the genetic characteristics of inflammasomes is assisting our understanding of both autoinflammatory diseases as well as other immune system-driven disorders.
The ability of innate immune cells to sense dangers to the host, such as microbial infections or a disruption in tissue homeostasis, depends upon a plethora of receptor families. The Toll-like receptor (TLR) family is the best-characterized of these,1 but in the last decade new intracellular receptor families have been identified. They include nucleotide-binding oligomerization domain (NOD)-containing protein-like receptors (NLRs), retinoic acid-inducible gene-like receptors and C-type lectin receptors.2, 3, 4 Some members of the NLR family have been deeply investigated for their central role in the formation of functional platforms, named inflammasomes, which mediate the initiation of the inflammatory process through IL-1β and IL-18 secretion.
All these receptors have generated a huge amount of interest recently due to their prominent role in a number of human autoinflammatory diseases. Linkage studies of inherited autoinflammatory conditions identified mutations or polymorphisms in a group of genes named hereditary periodic fever syndromes-associated genes encoding NLR proteins or their inflammasome-related partners. Since then, further genome-wide association studies have characterized specific genetic polymorphisms of NLR genes in relation to autoinflammatory diseases.
Alongside the involvement of NLR inflammasomes with spontaneous inflammatory disorders, other investigations have revealed inflammasome roles in protection against infective agents,5 as well as in the pathogenesis of autoimmunity and tumors.6, 7, 8, 9 How genetic variations influence inflammasome structure and function leading to disease onset is the challenging target for future investigations.
The NLR family
NLRs are fundamental components of the innate immune system that are evolutionarily conserved, having been detected in species as simple as sea urchins. NLR-like proteins functionally and structurally similar to mammalian ones have even been observed in plants.10, 11, 12 This fascinating example of convergent evolution hints at NLR-like structures as an efficient and successful tool of innate immunity against a broad variety of pathogens.
NLRs are cytosolic receptors that are expressed predominately by macrophages and dendritic cells. To date, 23 NLRs have been described in humans and 33 identified in mice. In general, NLRs are composed of three domains: a C-terminal region leucine-rich repeat (LRR), the central nucleotide domain NACHT (also known as NOD) and the N-terminal containing either a caspase recruitment domain (CARD) or pyrin domain (PYD) (Figure 1). LRRs are domains of 20–30 residues forming alpha helix and beta strands, similar to plant disease-resistance genes.13 Indeed, similarly to TLRs, LRR structures in NLRs are important for the recognition of pathogen-associated molecular patterns, although direct evidence of an interaction with known NLR-agonists has not yet been demonstrated. The central domain NACHT shares structural similarities with the central motif of apoptotic protease-activating factor 1, which promotes formation of oligomeric structures. Therefore, an analogous mechanism has been proposed for the formation of the NLR inflammasome platform.14, 15
The N-terminal CARD and PYD are the key regions linking the NLR to its downstream functions via oligodimerization with an analogous domain harbored by specific adaptor proteins. NLRs with N-terminal domains that contain CARD usually activate signaling cascades leading to transcription of proinflammatory genes (Figure 1). Alternatively an N-terminus PYD will recruit pro-inflammatory caspases.
Based on their molecular structure and function, NLRs can be divided into three subfamilies: NLRP, NOD and ICE-protease-activating factor (IPAF)/neuronal apoptosis inhibitory protein (NAIP) (Figure 1).12 NLRP, also known as NALP (NACHT domain-, LRR- and PYD-containing protein), is the largest subfamily, having 14 members in human and 12 in mouse. The prototypical members are NLRP1 and NLRP3. NLRP3 contains N-terminal PYD, a central NATCH domain responsible for oligomerization and activation, and a NACHT-associated domain followed by the highly conserved LRR region of 171 nucleotides at its C-terminus (Figure 1).16 NLRP1 has a similar structure to NLRP3 except that at the C-terminus of the LRR region, there is a FIIND (domain with function to find) followed by a C-terminal CARD for caspase recruitment. The two also differ in their expression patterns; the NLRP3 gene is strongly expressed in macrophages, dendritic cells, granulocytes and osteoblasts, whereas NLRP1 has a wider expression pattern, including brain, lung, placenta, small intestine, colon, kidney, liver, muscle, testes and epithelial cells. The functions and the tissue and cellular distribution of the other NLRP proteins have not yet been fully characterized.
The NOD family of NLRs includes NOD1–5 and the class II transactivator (Figure 1). An N-terminal CARD is characteristic of the NOD family. NOD1 and NOD2 are the two best-studied members. The natural ligands of NOD1 and NOD2 are moieties of peptidoglycan, a polymer of the glycan chains within bacterial cell walls; specifically, NOD1 detects mesodiaminopimelic acid, while NOD2 recognizes muramyl dipeptide (MDP).17 Upon activation, NOD1 and NOD2 recruit the receptor-interacting protein-2 through homodimerization of CARDs, which leads to the activation of NF-κB.18 NOD1 and NOD2 also engage CARD9, which mediates the activation of mitogen-activated protein kinase cascades responsible for nuclear translocation of Jun and ATF2 transcription factors.19 NF-κB and mitogen-activated protein kinase-mediated pathways trigger a vigorous inflammatory response characterized by secretion of proinflammatory cytokines and chemokines.
The third NLR subfamily is evolutionarily separate from other NLRs and consists of IPAF and NAIP (Figure 1). IPAF contains a CARD, whereas NAIP includes three baculovirus inhibitor domains that are probably involved in inhibition of apoptosis.20 Both IPAF and NAIP are involved in the formation of inflammasomes. Whole bacteria, such as Salmonella typhimurium, Pseudomonas aeruginosa or Legionella pneumophila, as well as bacterial components including flagellin, also promote IPAF oligomerization resulting in recruitment and activation of caspase-1.21 IPAF seems to cooperate with NAIP to induce caspase-1-mediated secretion of IL-1β by macrophages. However, the exact mechanisms underlying IPAF/NAIP detection of bacterial compounds and activation of caspase-1 remain largely unclear.
Inflammasomes are the multiprotein cytoplasmic complexes that facilitate activation of proinflammatory caspases.22 To date, three inflammasomes have been described according to the NLR protein they engage: the NLRP1, NLRP3 and IPAF inflammasomes.
NLRP3 is one of the best-characterized inflammasomes. The framework of the NLRP3 inflammasome structure seems to form through homotypic reactions of NLRP3's PYD with the PYD of an adaptor protein called apoptotic-associated speck-like protein containing a caspase recruitment domain (ASC) (Figure 2). ASC then recruits procaspase-1 through CARD–CARD interactions. Similarly, NLRP1 can directly recruit both procaspase-1 and -5 via CARD homotypic interactions. The ability to form inflammasome platforms is not limited to NLRP subfamily members; IPAF can recruit caspase-1 directly via CARD–CARD interactions and therefore the IPAF platform is referred as a third inflammasome prototype.23, 24 There is some evidence that that NAIP can assemble with IPAF to form heterocomplex inflammasomes.25, 26 NLRP1 can also associate with NOD2 to mediate caspase-1-dependent IL-1β secretion in response to Bacillus anthracis and MDP.27, 28
Recently a fourth type of inflammasome has been identified, involving the protein absent in melanoma 2 (AIM2) (Figure 1).29 Despite the lack of NACHT domain, AIM2 can undergo oligodimerization and recruit ASC via PYD in response to double stranded cytosolic DNA through its C-terminal HIN domain. AIM2 is generating a strong interest as a sensor of DNA and therefore its putative involvement in viral surveillance, as well as onset of autoimmunity against self-DNA, as seen in systemic lupus erythematosus.30, 31, 32 Since AIM2 does not belong to the NLR family, it will not be discussed further.
Activators of the NLRP3 inflammasome
Similar to the TLR family, much work has focused on identifying the stimuli recognized by NLRs. Activation of the inflammasomes can be triggered by exposure to whole live bacterial, fungal or viral pathogens (Salmonella typhimurium, Shigella flexneri, Legionella pneumonia, Pseudomonas aeruginosa, Candida albicans, Saccharomyces cerevisiae, Sendai virus, adenovirus and influenza virus),30, 33, 34, 35 or their components (MDP flagellin),36, 37, 38 but also by toxins required for pathogen entry into the host cell, such as membrane-disrupting and pore-forming compounds (maitotoxin, nigericin, aereolysis, anthrax lethal toxin and alpha-toxin of Staphylococcus aureus) (Figure 2).39, 40, 41, 42, 43 Interestingly, NLRP3 also contributes to the ability of dendritic cells and macrophages to detect endogenous danger molecules released by dying cells. NLRP3 can signal the presence of adenosine triphosphate (ATP), crystals of monosodium urate (MSU) or calcium pyrophosphate crystals,39, 44 and other large particulates of non-microbial origin, such as alum, silica and asbestos (Figure 2).45, 46, 47 The NLRP3 inflammasome has also been proposed as a sensor of adenoviral DNA, double-stranded RNA and viral RNA.30, 35
Proposed mechanisms of inflammasome activation
The NLRP3 inflammasome is the best-characterized, but the precise mechanism of NLRP3 activation is still debated. At steady state, inflammasome components are present in the cytosol and their assembly is prevented by auto-inhibitory mechanisms mediated by chaperone proteins.48 Inflammasome assembly is promoted by the application of specific stimuli. However there is no evidence that inflammasome-activating molecules would interact directly, as specific ligands, with the NLR proteins themselves. Moreover, it is still puzzling how many heterogeneous molecules, which differ broadly in biological properties and structures, could trigger NLRs, such as NLRP3. Multiple mechanisms of NLRP3 activation have been proposed, which may also cooperate in a unique and complex way (Figure 2).
Using cell-free systems it was shown that the NLRP3 inflammasome can spontaneously assemble if potassium levels are lowered below the physiological intracellular concentration of 70 mM.22, 49, 50 This observation suggested that inflammasome formation might also be triggered following disruption of cellular integrity that would similarly cause potassium levels to drop. Therefore, it was suggested that NLRP3 activators might act indirectly through pathways able to lower potassium concentration. This mechanism has been demonstrated for inflammasome activation driven by extracellular ATP, which induces inflammasome assembly and ASC-mediated IL-1β secretion.33, 51 ATP binds to the purinergic receptor P2X7, resulting in the opening of the protein channel pannexin-1.52 This allows potassium efflux, inflammasome activation and ATP release.53 However, it has been proposed that the pore formation following pannexin-1 recruitment could allow cytosolic access to microbial products, such as lipopolysaccharide (LPS), which can directly activate the inflammasome.52 Although it is unclear whether the pore width would be sufficient to allow entrance of such microbial-derived molecules.
The inflammasome activators MSU, alum, asbestos and silica possess a crystal or particulate shape. Inhibition of cytoskeletal rearrangement with colchicine or cytochalasin D blocks IL-1β secretion.44 These observations led to the second model of NLRP3 activation. The attempts to engulf large particulates by phagocytes via a mechanism known as frustrated phagocytosis can generate NADPH-dependent formation of reactive oxygen species (ROS).45, 46, 54 Recently, the thioredoxin-interacting protein was identified as an NLRP3-binding protein linking ROS generation to assembly and activation of the NLRP3 inflammasome.55 In support of this model, antioxidants and NAPDH inhibitors prevent inflammasome assembly.34, 46, 47, 49, 54, 56 Although ROS production is required for NLRP3 activation, several ROS-inducing agents do not promote inflammasome formation.57, 58
A third mechanism of NLRP3 activation proposed by Hornung and colleagues was that, in the case of large particulates, inflammasome activation occurs as a result of lysosomal swelling and leakage.57 In contrast to the hypothesis of frustrated phagocytosis, this model proposes that upon cellular uptake silica and aluminum salts cause lysosome acidification with consequent release of acidic contents into the cytosol. Indeed, inhibition of the lysosomal protease cathepsin B leads to a substantial decrease in activation of the NLRP3 inflammasome induced by silica.57
Inflammasome-activated cytokines: IL-1β and IL-18
The main outcome of NLRP3 inflammasome activation is the recruitment of caspase-1, inducing the cleavage and secretion of the proinflammatory cytokines IL-1β and IL-18. IL-1 family members are potent modulators of both innate and adaptive immunity and are important for host protection against infections. Notably, IL-1β is involved in the early pathogenesis and the sustained severity of a broad pattern of diseases, including many arthritic diseases and septic shock.59 Blocking the IL-1R is a successful treatment for several immune-related disorders.60, 61
IL-1β production is controlled by a two-step mechanism. TLR or NOD agonists transcribe the immature form, pro-IL-1β, via the NF-κB pathway followed by caspase-1-mediated cleavage of pro-IL-1β upon inflammasome activation.59 The NF-κB pathway is also required for transcription of NLRP3.62 To bypass this limitation, most in vitro studies have used inflammasome agonists in combination with other pathogen-associated molecular patterns, most often LPS, to ‘license’ inflammasome formation and activation.33, 44, 57, 63, 64, 65 It is important to consider, however, that LPS used as priming signal in several studies could contribute to NLR activation through additional mechanisms beyond pro-IL-1β and NLRP3 transcription.
NLRP3 activation also leads to IL-18 maturation, although this cytokine can also be activated by other caspase-1-independent mechanisms.66, 67 The immature form of IL-18 is constitutively expressed in some cells but only those with an active inflammasome, such as macrophages, dendritic cells and Kupffer cells, have the ability to release the mature form.68
In the last decade, a number of groups have focused on the study of inflammasomes as initiators of immune disorders mediated by IL-1β and IL-18. Thus, many genetic associations between members of NLR family and immune diseases have been discovered. Moreover, a correlation between NLR members and non-immune mediated disorders has also been proposed, suggesting that this protein family has as yet undefined functions crucial for the balanced physiology of the organism. For these reasons, involvement of genetic mutations of inflammasomes in several diseases will be discussed in detail with a particular focus on IL-1-mediated disorders.
Genetics of inflammasomes in autoinflammatory diseases
NLRP3 mutations are associated with a group of rare hereditary autoinflammatory diseases called cryopyrin-associated periodic syndromes (CAPS), which includes familial cold auto-inflammatory syndrome (FCAS), Muckle–Wells syndrome (MWS), and chronic infantile neurological, cutaneous, and articular (CINCA) syndrome, also known as neonatal onset multisystem inflammatory disorder.69 These pathologies are listed in order of increasing disease severity and life-threatening potential, and are characterized by recurrent episodes of fever, rush, arthropathies and various degrees of neurological complications.70, 71, 72 Genetic studies of CAPS identified more than 90 disease-associated genetic variants of the NLRP3 gene, the majority of which are autosomal dominant missense point mutations located in exon 3, encoding the NATCH domain (Figure 3 and Table 1).73, 74 Despite the familial recurrence, some genetic variants inducing CAPS phenotypes result from de novo mutations. Interestingly, identical de novo mutations were independently found in different patients and often overlapped with hereditary mutations, suggesting the presence of hot-spot loci within the NLRP3 gene that exhibit high mutation susceptibility.75
It is presently unclear how NLRP3 point mutations affect the functional properties of the inflammasome at a molecular level, leading to the high expressivity and penetrance observed in several CAPS-related genetic variances. Exon 3 missense mutations result in continuous IL-1β secretion caused by constitutive inflammasome assembly and activation (Figure 3). NLRP3 hyperactivation bypasses the need for an NLR agonist. Indeed, blood monocytes from CINCA patients spontaneously release IL-1β upon stimulation by LPS alone without any inflammasome activator.76, 77, 78 This hypothesis is also supported by knock-in mouse models carrying the point mutations R258W and A350V in the exon 3 of NLRP3 gene, which correspond respectively to the human NLRP3 R260W and A352V mutations associated with MWS (Figure 3 and Table 1).75, 77, 79 R258W and A350V mutations result in spontaneous IL-1β release due to aberrant inflammasome structural conformation and hyperactivation in the absence of NLRP3 agonists. Intriguingly, mouse macrophages carrying the L351P mutation, analogous to the human L353P mutation in FCAS patients, showed increased IL-1β secretion when incubated at lower temperatures, mimicking the cold-sensitive phenotype observed in FCAS subjects.79 The pathologies seen in knock-in mouse models of CAPS-associated mutations are similar to human disease syndromes, suggesting that IL-1β hyperactivation is sufficient to initiate and sustain several important inflammatory events. Indeed, therapies blocking the action of IL-1β using an IL-1R antagonist or humanized anti-IL-1β monoclonal antibody are successfully used to treat CAPS and other IL-1-mediated disorders.
Several CAPS-associated mutations and polymorphisms have been identified, but only a few have been analyzed for their effect on the molecular and cellular function. Using an in silico modeling prediction of the NACHT domain structure, Neven and coworkers investigated how the single residue substitutions in NLRP3 frequently associated with CAPS could affect secondary and tertiary protein configurations.75 According to the putative tertiary structure, all the substitutions clustered close to the nucleotide binding cleft, suggesting that nucleotide binding site, located in proximity of the NACHT domain in a specific site outside the structure core, could be crucial for NLRP3 oligomerization. Hence, mutations affecting this region were also suggested to influence the quaternary structure determining aberrant inflammasome formation.75
Mutations in the C-terminal region of the NACHT domain correlate with the most severe CINCA syndrome phenotypes and are associated with significant neurological impairment (Figure 3 and Table 1). LPS-treated monocytes expressing the Y570C mutation showed an increased necrotic-like cell death associated with lysosomal leakage and other cellular metabolic alterations.80, 81 Moreover, monocytes transfected with Y570C exhibited higher NF-κB activity compared to healthy controls.82, 83
A second important aspect is that some CINCA patients with milder symptoms apparently lack NLRP3 mutations but carry a latent low-level mosaicism. This implies that some missense mutations in the NLRP3 gene can exert a dominant phenotype leading to CAPS even when NLRP3 is expressed by less than 25% of the cells in mosaic patients.81, 84 This could be due to the positive feedback loop sustained by the IL-1/IL-1R axis.
Recently disease-correlated missense mutations in exons other than exon 3 have been identified. Exon 6 encodes the LRR domain and the Y859C mutation has been associated with a different CAPS phenotype lacking the typical urticarial rush, while other symptoms were overlapping with late on-set phenotypes of MWS and CINCA (Figure 3).83 Similar to mutations in the NACHT domain, Y859C mutations lead to spontaneous inflammasome formation, ASC recruitment and IL-1β secretion. Remarkably, treatment of patients carrying the Y859C mutation with an IL-1R antagonist was effective, suggesting that a complete screening of the whole NLRP3 gene should be recommended, especially for patients showing non-classical autoinflammatory symptoms.83
Two additional NLRP3 variants, G755R and G755A of exon 4, also result in a CAPS phenotype probably through enhancement of NF-κB activation (Figure 3 and Table 1). However, the role of NLRP3 in regulating the NF-κB pathway is still widely debated. A hint comes from studying the homology among NLRs. Indeed, the missense mutation R334W in the NOD2 gene associated with Crohn's disease exhibits a strong positional and structural analogy to R260W in the NLRP3 gene, which results in aberrant NF-κB activation. Interestingly, an association between Crohn's disease and NLRP3 polymorphisms leading to impaired IL-1β secretion has been proposed.85, 86, 87 However, recent studies in a larger patient group failed to replicate this observation.88 Further work is needed to clarify the possible association of inflammasome genetics to multifactorial diseases.
Although several missense mutations in NLRP3 inflammasomes strongly correlate with CAPS, a significant number of patients with CAPS-like phenotype do not present any mutation in NLRP3 gene.75, 89 Some autoinflammatory syndromes resembling CAPS have been correlated with mutations in inflammasome or IL-1 pathway regulators. NLRP12 mutations are associated with hereditary periodic fever syndromes, and while NLRP12 shares strong homology with NLRP3, it is not involved in inflammasome formation but regulates the NF-κB cascade.90, 91 A nonsense mutation causing defective splicing and introducing a premature stop codon in NLRP12 causes a dominant phenotype due to the lack of regulation of the inflammatory pathway.
Despite its typical inflammasome-promoting structure, another member of NLRP family, NLRP7, acts as a negative regulator of caspase-1 activation.92, 93 Mutations in NLRP7 have mainly been associated with aberrant development of the reproductive tract, such as endometrial cancer tissues and familial recurrent hydatidiform moles.94
Mutations in genes encoding pyrin and proline-serine-threonine phosphatase-interacting protein 1 are associated with two autoinflammatory diseases with hereditary periodic fever syndrome-like symptoms: familial mediterranean fever and pyogenic arthritis pyoderma gangrenosum and acne syndrome, respectively. Pyrin and proline–serine–threonine phosphatase-interacting protein 1 interacts with and modulates pyrin function, which affects inflammasome activity. However, it is still unknown whether pyrin acts as an ASC activator, inducing inflammasome-independent IL-1β release or as negative inflammasome regulator able to dampen NLRP3 function by competing for ASC.95, 96 In both hypotheses pyrin and proline–serine–threonine phosphatase-interacting protein 1 mutations are thought to induce dysregulation of IL-1β secretion.97
A similar hypothesis was formulated for caspase-12, which is expressed in several mammals but commonly present as a null allele in humans, where a mutation leads to a truncated non-functional form of the protein. Caspase-12 is a negative regulator of the NF-κB pathway and for this reason it can increase susceptibility to sepsis following infection. It is likely that in humans the null allele of caspase-12 has been selected for under evolutionary pressure in areas subject to persistent infections. However, in some populations of Africa, South Asia and South America, a caspase-12 mutation causing expression of a full-length protein is responsible for a dominant negative autoinflammatory phenotype.
The genetics of inflammasomes have been deeply investigated in relation to autoinflammation, where missense point mutations in NLR-coding genes determine with high penetrance the pathogenesis of autoinflammatory processes (Figure 4). This discovery allowed better characterization of the genotype of patients exhibiting autoinflammatory symptoms, in order to design a specific anti-IL-1β therapy. Moreover, transgenic animal models bearing these pathological mutations helped in defining the molecular mechanisms mediating aberrant inflammasome activation. However, several genome-wide screenings have recently uncovered the association between genetic variants of NLRs and other diseases beyond autoinflammation. These findings suggest that specific polymorphisms in NLR genes could contribute, albeit with lower penetrance, to the susceptibility to more common multifactorial diseases such as autoimmune disorders and cancer (Figure 4). The study of inflammasome genetics is also important to help in the diagnostic process and in the design of effective therapeutic strategies. Further investigations aimed to define the genetics of inflammasomes in human diseases are urgently needed to reveal the potent mechanisms mediating the pathology of rare autoinflammatory disorders, but also to depict the complexity of widespread multifactorial diseases, where NLR polymorphisms may play an important hidden role.
We thank Lucy Robinson for critically reviewing the manuscript. This work was supported by Agency for Science, Technology and Research (A*STAR) of Singapore.