Article Text

Download PDFPDF

Inflammasomes and rheumatic diseases: evolving concepts
  1. P I Sidiropoulos,
  2. G Goulielmos,
  3. G K Voloudakis,
  4. E Petraki,
  5. D T Boumpas
  1. Department of Rheumatology, Clinical Immunology and Allergy, University Hospital, Medical School, University of Crete, Heraklion, Greece
  1. Prodromos Sidiropoulos, Department of Internal Medicine, Division of Rheumatology, Clinical Immunology and Allergy, University of Crete, Medical School, Voutes 71500, Heraklion, Greece; sidiropp{at}


The realisation that the production of inflammatory cytokines in inflammatory rheumatic diseases may be induced by non-infectious endogenous signals has encouraged researchers to explore mechanisms of innate immunity and their contribution to the pathogenesis of these diseases. The nucleotide-binding and oligomerisation domain (NOD)-like receptors (NLRs) sense pathogens, products of damaged cells or endogenous metabolites and could potentially be involved in the initiation, amplification and progression of the inflammatory response in rheumatic diseases. NLRs are involved in the regulation of innate immune responses with some of them promoting the activation of inflammatory caspases within multiprotein complexes, called inflammasomes. A typical inflammasome consists of a sensor, an NLR protein, an adaptor protein such as ASC (for apoptosis-associated speck-like protein containing a caspase recruitment domain (CARD)) and an effector protein that is a caspase that activates pro-inflammatory cytokines such as interleukin (IL)1β and IL18. Recent data suggest a role of the inflammasome in the pathogenesis of autoinflammatory as well as inflammatory rheumatic diseases such as juvenile chronic arthritis, adult onset Still disease, rheumatoid arthritis and gout. Modulation of these pathways may be a potential therapeutic target for inflammatory rheumatic diseases.

View Full Text

Statistics from

The role of the immune system is to defend the host from foreign invaders. Although once thought to be a non-specific immune response, recently it was found to perform with substantial specificity. Innate immune recognition is mediated through a limited number of receptors evolved to recognise products born by pathogens, referred as pathogen-associated molecular patterns (PAMPs).1 PAMPs are molecules vital for microbial survival with preserved structure such as the bacterial lipopolysaccharide (LPS) and peptidoglycans (PGNs) or viral RNA. Organisms have developed a group of receptors that recognise these PAMPs, referred as pathogen-recognition receptors (PRRs). Activation of PRRs triggers inflammatory reactions of the innate immune system, but in mammals also leads to the release of mediators that activate the adaptive immune system.

Two theories have been put forward for the interpretation of the initiation of immune responses: the pattern recognition theory and the danger theory. The former suggests that microbial non-self induces an innate immune response; this in turn, triggers an adaptive immune response.2 The danger theory holds that the host’s injured cells release alarm signals that activate antigen-presenting cells (APCs).3 In the case of PRRs, both propositions are operant. Innate immunity receptors may recognise PAMPs and abnormal self or danger signals such as DNA, RNA, or uric acid, which should normally not be present outside the cells or at certain locations within the cell.46


Knowledge of PRRs has evolved in recent years and they comprise extracellular and intracellular components, namely the TLRs, the NLRs and the RLRs. Among them, the TLR family is the most studied.7 8 TLRs survey the extracellular fluids and endosomal compartments; upon stimulation by specific bacterial, viral and fungal components, they trigger gene products that control innate immune responses and further direct the development of antigen-specific acquired immunity, mainly by regulating the function of dendritic cells (DCs). Stimulation of TLRs results in the activation of different intracellular signalling cascades, which generally activate NFκB and activated protein-1 (AP-1) or type I interferon (IFN) synthesis.9 In addition to recognising molecular patterns associated with different classes of pathogens, TLRs may also recognise a number of self proteins and endogenous nucleic acids. Data originating predominantly from animal models of autoimmune diseases and circumstantial data from patients suggest that inappropriate activation of TLR pathways by endogenous or exogenous ligands may lead to the initiation and/or perpetuation of autoimmune responses and tissue injury. In MRL-lpr mice, TLR-3, TLR-7 and TLR-9 agonists may exacerbate pre-existing immune complex glomerulonephritis,10 while immunocomplexes containing IgG can stimulate rheumatoid factor producing B cells through concomitant TLR-9 and B cell receptor stimulation.11 In patients with systemic lupus erythaematosus (SLE), we have recently reported that an increased proportion of B cells and monocytes expressed TLR-9 among patients with active compared to patients with inactive disease; increased percentages of TLR-9 expressing B cells correlated with the presence of anti-dsDNA antibodies.12 Additional data demonstrate that innate immune responses mediated by TLRs may regulate inflammation in rheumatoid synovial tissue. Mice deficient for the adaptor molecule MyD88 are resistant to streptococcal cell wall arthritis, and TLR2 deficient mice have reduced disease severity.13 In patients with rheumatoid arthritis (RA), TLRs are abundantly expressed in the synovial tissue and TLR-2 activation of synovial fibroblasts through bacterial peptidoglycans results in upregulation of integrins, matrix metalloproteinases and inflammatory cytokines (interleukin (IL)6, IL8).14 Finally, ds-RNA released from necrotic cells and heat shock proteins expressed in the synovial tissue may activate TLR-3 and TLR-4 respectively.

The NLRs are intracellular receptors that sense pathogens or danger signals and mount an inflammatory response.5 They comprise a group of 22 proteins that can be divided into 2 major subfamilies, the NOD and the NTPases implicated in apoptosis and multihistocompatability complex transcription (NACHT) leucine-rich repeat protein (NALP) group; class II transactivator (CIITA), the IL1B-converting enzyme (ICE)-protease activating factor (IPAF) and the neuronal apoptosis inhibitor protein (NAIP) are the other NLR group members. NALP, the major subfamily, has 14 members (NALP1 to NALP14), while the NOD subfamily consists of 5 members (NOD1 to NOD5). They are also known as CATERPILLER proteins or NACHT-leucine-rich repeat (LRR) proteins. NLRs have three structural domains (fig 1).15

Figure 1 Nucleotide-binding and oligomerisation domain (NOD)-leucine-rich repeat (LRR) and inflammasome structures. NOD-like receptors (NLRs) have three structural domains.15 A. The LRR domain at the C-terminus. B. The intermediary NTPases implicated in apoptosis and multihistocompatability complex transcription (NACHT) domain. C. The N-terminal domain that can be a pyrin domain (PYD), a caspase recruitment domain (CARD) or a baculovirus inhibitor-of-apoptosis protein repeat domain (BIR). The LRR domain is considered as the ligand-sensing motif, thus involved in the interaction with pathogen-associated molecular patterns (PAMPs), in analogy to Toll-like receptors (TLRs). The NACHT domain is responsible for the oligomerisation and activation of NLRs. The PYD or CARD domain of NLR is the link to downstream adaptors (such as apoptosis-associated speck-like protein containing a CARD (ASC)) or effectors (such as caspase-1). The BIR domain is proposed to act as caspase inhibitor.99 During NACHT leucine-rich repeat protein (NALP)3 and NALP1 inflammasome activation, NALP3 or NALP1 interact through PYD–PYD homotypic interactions with ASC, resulting in its activation.25 Subsequently, the CARD domain of ASC interacts with the CARD domain of caspase-1 and mediates its activation. Of note, NALP1 may also activate directly the caspase-5 through its C-terminal CARD domain. In contrast, NALP3 does not simultaneously activate caspase-5, but NALP3 can recruit a second capsase-1 through the CARD domain of CARD inhibitor of nuclear factor κB activating ligand (CARDINAL), a component of the NALP3 inflammasome.40 Interestingly, interleukin 1B-converting enzyme (ICE)-protease activating factor (IPAF), that can on its own sense PAMPs possesses a CARD domain at the N-terminal and thus may directly activate caspase-1 without ASC recruitment (“IPAF inflammasome”).

Not surprisingly, there is significant crosstalk between TLR and NLR signalling pathways. To give an example, TLR stimulation results in pro-IL1β and pro-IL18 production; these are the main substrates for active caspase-1, which is produced NLR-dependently upon inflammasome activation, culminating in the production of mature IL1β. More recently, it has been shown that caspase-1 is essential for TLR2 and TLR4 signalling, through cleavage of MyD88 adaptor-like (Mal), an adaptor protein downstream of the TLR2 and 4 signalling cascade.16

While the main viral sensors on DCs are the antiviral TLRs (TLR3, 7, 8 and 9), on cells other than DCs these appear to be the RLRs.17 18 RIG-1 and melanoma differentiation-associated gene-5 (MDA5) are two members of the RLRs that sense intracellular dsRNA. RLRs have similarities with TLRs and NLRs. Following activation, RLRs signal downstream through their CARD domains to a CARD-containing adaptor protein interferon β promoter stimulator (IPS1), leading to IFNα/β production.


Caspases are a group of aspartate-specific proteases involved in apoptotic or inflammatory pathways. Inflammatory caspases (known as group I caspases) are caspase-1, caspase-4 and caspase-5. Caspases are produced in cells as catalytically inactive zymogens and undergo proteolytic processing during activation. Caspase-1 mediates pro-IL1β maturation,19 20 as well as pro-IL18 and possibly IL33.21 Caspase-5 is a component, together with caspase-1, of the NALP-1 inflammasome that cleaves pro-IL1β (see below). Caspase-5 may also have a regulatory role in tumourigenesis. Caspase-5 cleaves Max, a component of the Myc/Max/Mad network of transcription factors that is frequently deregulated in tumours.22 Frameshift mutations of caspase-5 are frequently found in endometrial carcinoma.23

The puzzle of the machinery involved in the activation of caspases has recently been expanded with the characterisation of certain of NLRs proteins, which are involved in the formation of caspase activating complexes referred as inflammasomes (fig 2).24 NLR proteins that, upon specific stimuli engagement, promote the assembly of inflammasome include NALP1, NALP3, IPAF and NAIP. A typical inflammasome consists of a NLR protein serving as a sensing protein, one or more adaptor proteins (ie, apoptosis-associated speck-like protein containing a CARD (ASC)) and one or more inflammatory caspases acting as effectors (fig 1).25 The main adaptor molecule in the interface of NLRs and caspase-1 activation is ASC, a bipartite molecule containing a PYD and a CARD domain. Upon ligand sensing (table 1), NLRs are activated and expose the effector domains; CARD or PYD. Those domains, through homotypic interactions recruit molecules containing CARD or PYD, bringing them in close proximity to each other, ultimately activating them. These molecules may be adaptors, like ASC protein, or effector caspases.

Figure 2 Activation of inflammasome by exogenous or endogenous signals results in caspase activation and interleukin (IL)1β or IL18 production. Two signals are essential for mature IL1 secretion.100 In signal 1, immune cells stimulated with Toll-like receptor (TLR) ligands, produce pro-IL1β and other inducible components of inflammasome (caspase-11). These cells are primed to receive the second signal, which through inflammasome will activate caspase-1 to cleave pro-IL1β to mature IL1β. Upon ligand sensing (signal 2), nucleotide-binding and oligomerisation domain (NOD)-like receptors (NLRs) are activated and expose the effector domains, caspase recruitment domain (CARD) or pyrin domain (PYD), which, through homotypic interaction, recruit molecules containing CARD or PYD, bringing them in close proximity to each other leading thus to their activation. These molecules may be adaptors, such as apoptosis-associated speck-like protein containing a CARD (ASC) protein, or effector caspases.
Table 1 Ligands of inflammasomes

Three human inflammasomes have been described, named from the NLR protein involved: the NALP1 inflammasome, activates caspase-5 and caspase-1; and the NALP3 and IPAF inflammasomes, both of which activate caspase-1 (fig 1). Of note, there is no evidence to support whether ligands of microbial origin bind directly to NLRs or, as has been proposed for the plant NLR homologues, this is an indirect sensing.26 27 Human studies have shown that NALP1 and NALP3 have distinct and separate expression profiles in human tissues.28 Granulocytes, T cells, B cells and dendritic cells express NALP1 and NALP3. NALP1 is present in glandular epithelial structures (stomach, gut, lung) while NALP3 is expressed mainly in non-keratinising epithelia (oropharynx, oesophagus and ectocervix).

Major questions regarding NLRs to be clarified include: (1) what are the exact subcellular localisation and trafficking of NLRs; (2) do NLRs act as direct receptors of PAMPs or do they sense PAMPs bound to adaptor proteins? and (3) how they may be involved in mounting adaptive immune responses?


Although IL1β is essential for the control of infections or self danger signals, its uncontrolled production could be harmful. Thus, delineation of the regulators of the inflammasome could be of therapeutic potential. To this end, different proteins have been proposed to interfere with inflammatory caspases activation. The first type of proteins considered as inflammasome regulators are characterised by the presence of a CARD domain highly similar to that of caspase-1.29 These proteins are thought to prevent recruitment and activation of the caspase by the adaptor ASC or by IPAF, through CARD–CARD interactions. This group includes the decoy caspase-1 genes present in the human caspase-1 locus, such as iceberg, INCA, COP and caspase-12.3033

Other inhibitors of the inflammasome are characterised by the presence of PYD domain and are thought to interfere with PYD–PYD interactions between ASC and NALPs. Pyrin, the protein involved in familial Mediterranean fever (FMF), is considered as one of inflammasome regulators. “Pyrin-only” proteins (POP), are attractive negative regulators of PYD-mediated functions. Recently Johnston et al reported that M13L-PYD a POP from myxoma virus, inhibits caspase-1 dependent IL1 production and NFκB activity.34 Deletion of M13L inhibited virus replication in vivo.


Inflammasome activation has been examined so far in the context of the innate immunity in a variety of experimental systems of infections4 6 3539 and autoinflammatory diseases.40 Its role in adaptive immune system reactions is less clear. Recently reported data support a critical role for NALP3-inflammasome in mounting contact hypesensitivity reactions, a T cell mediated cellular immune response to repeated epicutaneous exposure of contact allergens, that can be divided in two phases, the sensitisation and elicitation. The roles of IL1β and caspase-1 in the sensitisation phase41 42 have been described earlier.43 Recent data suggest that NALP3–/– and ASC–/– animals elicit significantly impaired hypersensitivity response to the hapten trinitrophenylchloride compared to wild type animals.44 Furthermore, by the use of bone marrow chimera experiments, it was shown that NALP3 and ASC are essential in the sensitisation phase. Key components of the inflammasome are present in human keratinocytes; contact sensitisers such as trinitrochlorobenzene, induce ASC/caspase-1 dependent IL1β and IL18 processing and secretion.45

It has been known for several years that IL1β may have an adjuvant capacity; in mice immunised with protein antigens together with IL1, serum antibody production was enhanced.46 By contrast, it has recently been shown that adjuvants used in humans act through inflammasome. Thus, aluminium hydroxide (Alum), an adjuvant approved for routine use in humans, may act through activation of caspase-1.47 Human peripheral blood mononuclear cells (PBMCs) and DCs treated with a combination of various TLR ligands and aluminium, activated caspase-1 and produced large amounts of IL1β and IL18. Alum-induced IL1β and IL18 production was not due to enhanced TLR signalling but rather reflected caspase-1 activation; experiments with MyD88 deficient mouse DC showed that the Alum effect was MyD88 independent, further supporting that Alum mediated its signalling independently of TLR.

Signalling through TLRs, such as for example TLR4, activates the MyD88-dependent pathway, resulting in inflammatory gene transcription, and the Trif-dependent cascade resulting in interferon regulatory factor 3 (IRF3) activation and type I IFN production.9 An interplay of type I IFN and inflammasome activation has been reported recently;48 upon infection with cytosolic bacteria adequate production and signalling of type I IFN is required for effective inflammasome activation. Given the broad role of type I IFN in immunity to viruses and autoimmunity, this complex interplay needs further investigation.


The role of IL1β in the pathophysiology of rheumatic diseases such as RA is well established.49 IL1β is present in inflamed synovium of mice with antigen-induced and collagen-induced arthritis,50 51 while intra-articular ex vivo gene transfer of IL1β in rabbits resulted in a highly aggressive arthritis.52 In synovial biopsies from patients with RA, IL1β has been found in areas of macrophages and fibroblasts.53 In spite of strong evidence of IL1β involvement in RA pathophysiology, the clinical efficacy of IL1 blockade targeting IL1 receptor antagonist was rather disappointing. A novel agent that blocks IL1b, IL1 TRAP54 is currently under investigation in patients with RA or cryopyrin-associated periodic syndromes.

In addition to its multiple pro-inflammatory properties, IL1β also promotes autoreactivity by several mechanisms. In immune complex disease such as lupus, IL1β may participate in renal tissue injury because it is essential for the production of monocyte chemotactic protein-1 (MCP-1) by resident renal cells.55 IL1β/IL1R signalling may affect adaptive immune responses and promote autoreactivity affecting: (1) production of cytokines by DCs and thus T cell priming;56 (2) induction of CD40 ligand and OX40 expression on T cells and thus T and B cell interaction;57 and (3) affecting directly autoreactive effector T cells.58

By contrast, IL18 is thought to be involved in the pathogenesis of many autoimmune diseases in several animal models, such as the non-obese diabetic (NOD) mouse,59 the collagen-induced arthritis;60 61 moreover in lupus prone MRL lpr/lpr animals that were repetitively vaccinated with IL18 cDNA coding the murine IL18 precursor were protected from developing the disease.62 In humans, elevated levels of IL18 have been found in affected tissues in patients with Crohn disease 63 64 and RA.65 66


FMF and tumour necrosis factor receptor-associated periodic syndrome (TRAPS) are the two prototypes of a group of diseases referred as systemic autoinflammatory diseases. These represent a group of inherited disorders characterised by unprovoked inflammation and absence of autoantibodies or antigen-specific T cells.67 FMF is an autosomal recessively inherited disease caused by mutations in the MEFV gene that encodes for a protein known as the pyrin (or marenostrin) (table 2). To date over 117 MEFV mutations have been reported. Pyrin consists of four domains and all of them can facilitate protein interactions: the PYD domain, the B-box zinc finger domain, a coiled-coil domain and a C-terminal 30.2/rfp/SPRY domain (also known as PRYSPRY domain).

Table 2 Inflammatory diseases associated with inflammasome and related proteins

Pyrin seems to be one of the regulators of inflammasome activation, with most data supporting an inhibitory effect on IL1β production. More specifically, pyrin interacts through its PYD domain with PYD domain of ASC, and negatively regulates inflammasome by competing for ASC. The murine RAW monocytic cell line transfected by full-length mouse pyrin had reduced IL1β secretion, while mouse peritoneal macrophages expressing a truncated pyrin exhibited increased IL1β compared to wild type controls.68 Of note, a high percentage of pyrin mutations associated with FMF are located in the C-terminal B30.2 domain, whose functional importance has just begun to be investigated. Chae et al found that human pyrin can directly (ASC independently) inhibit caspase-1; this interaction is mediated through the B30.2 domain and interaction results in reduced IL1β secretion.69 Supporting the hypothesis of pyrin as a negative regulator of inflammasome, FMF-associated pyrin mutants located in the B30.2 domain could not reduce IL1β secretion. By contrast, Yu et al showed that in transfected 293T human kidney embryonic cells, pyrin may actually assemble an inflammasome complex and activate caspase-1 and IL1β, thus acting as a proinflammatory protein.70 Finally, Papin et al have recently found that the SPRY domain of pyrin interacts with NALP3, caspase-1 and pro-IL1 and inhibits caspase-1 dependent IL1b production,71 adding more complexity to the pyrin involvement in IL1 homeostasis.

Blockade of IL1β with IL1 receptor antagonist (anakinra) has been tried in patients with FMF. Although there are case reports showing clinical benefit,69 72 there are resistant cases. An ongoing phase-II clinical trial of IL1 TRAP, a novel IL1 inhibitor, sponsored by the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), ( identifier NCT00094900) explores the clinical benefit of IL1β blockade in patients with FMF.


The group of autoinflammatory diseases has been expanded to include disorders associated with mutations in the NALP3 protein (or cryopyrin), also called CAPS. CAPS comprises of familial cold autoinflammatory syndrome (FCAS), Muckle–Wells syndrome (MWS), neonatal onset multisystem inflammatory disease (NOMID) also called chronic infantile neurological cutaneous and articular syndrome (CINCA).73 Missense mutations in the NACHT domain of NALP3 are involved in all three diseases.74 75 Since the original genetic identification74 more than 30 missense mutations have been identified.76 These mutations confer a gain of function to the NALP3 protein that results in constitutively active NALP3, that results in aberrant maturation of IL1β.40 The fundamental role of IL1β in these diseases has been shown by the profound therapeutic efficacy of IL1β inhibition through IL1 receptor antagonist (IL1Ra, anakinra) in patients with MWS, FCAS or NOMID.7779


Gout is one of the most acute inflammatory arthritides, characterised by tissue deposition of monosodium urate (MSU) crystals. Recently it has been shown that MSU inflammation is inflammasome-dependent, while MyD88-dependent IL1 receptor signalling is essential for inflammatory responses. NALP3 inflammasome is essential for IL1β production through ASC and caspase-1 recruitment upon MSU challenge. Martinon et al4 showed that mouse macrophages from NALP3–/–, ASC–/– or caspase-1–/– mice produced significantly less mature IL1β, compared to wild type animals when challenged with MSU. The same was true when calcium pyrophosphate dehydrate (CPPD) crystals were used as inflammasome trigger. By contrast, none of the 11 TLRs, the extracellular sensors of innate immunity, had any involvement in sensing MSU crystals.80 Nevertheless, one cannot exclude TLR involvement in promoting IL1β precursor synthesis. The IL1β released then mediates IL1R signalling and MyD88-dependent NFκB activation, resulting in the transcription of neutrophil-recruiting chemokines such as IL8, S100 and macrophage inflammatory protein-2 (MIP-2). However, the mechanism of downregulation of acute gouty attacks remains poorly understood. The role of IL1β in acute gout has been reinforced by the clinical effectiveness of IL1β receptor blockade in a small open label study.81 In this study, 10 patients with gout who could not tolerate or had failed standard anti-inflammatory therapies were treated for 3 consecutive days with 100 mg/day of anakinra; all 10 patients responded well with no adverse events.


SoJIA, is a form of juvenile idiopathic arthritis characterised by prominent systemic symptoms. IL1β is a key inflammatory cytokine in disease pathophysiology. Pascual et al reported that PBMCs from SoJIA upon stimulation secrete high amounts of IL1β compared to control, but not TNFα or IL6.82 Microarray analysis of healthy PBMCs after incubation with serum from patients with active SoJIA , showed that among the genes that were upregulated was the IL1β gene (by a mean of 8.2-fold) as well as other IL1 cytokine/cytokine receptor family genes and other innate immunity genes (chemokines, fibronectin etc). The clinical efficacy of recombinant ILRa confirms the significance of IL1β in the inflammatory cascade in SoJIA.83


As already discussed, inhibition of the end product of inflammasome activation, namely IL1β, has been proved clinically effective in diseases such as CAPS,77 78 adult onset Still disease83 and more recently in gout .81 Although direct IL1β inhibition is appealing, alternative ways to inhibit inflammasome activation could be of interest. It is apparent that there are multiple proteins and regulators that participate in the activation of inflammasome; this offers the opportunity for intervention for inhibition of inflammatory caspases activation. For example, novel regulators of inflammasome activation could be applied to inhibit caspase activation. Moreover, caspase inhibitors, either pan-caspase inhibitors or specific caspase-1 inhibitors, have been developed and assessed by different pharmaceuticals. For example, caspase-1 inhibitor Z-YVAD-FMK inhibits in vitro IL1β production from monocytes from patients with familial cold autoinflammatory syndrome.84 Other caspase inhibitors have been tested in animal models or even in clinical trials of rheumatic diseases, such as RA.85 Since inflammasome activation is the second signal for IL1β production, the first one is mediated through TLR4 signalling (resulting in pro-IL1β production), it is conceivable that the combination of inhibitors in TLR4 signalling cascade could be important for suppressing IL1β production.


The interplay between different pathogen sensors on the cell surface (TLRs) and in the cytoplasm (NLRs) has revealed the complex mechanisms used by the organisms to mount effective immune responses to pathogens and endogenous danger signals. Inflammasome products are also essential components of the adaptive immune responses. Together these data support the notion that a diverse spectrum of inflammatory rheumatic diseases irrespective of their aetiology (autoimmune, autoinflammatory or crystal-induced) may share common inflammatory pathways. Improved understanding of the mechanisms of inflammasome activation and regulation will provide insights in the pathophysiology of many chronic inflammatory diseases and may uncover novel therapeutic targets.


View Abstract


  • Competing interests: None declared.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.