Inflammasomes in Intestinal Inflammation and Cancer

Abstract

Inflammasomes are multi-protein complexes that mediate activation of caspase-1, which promotes secretion of the proinflammatory cytokines interleukin (IL)-1β and IL-18 and pyroptosis, a form of phagocyte cell death induced by bacterial pathogens. Members of the Nod-like receptor family (including Nlrp1, Nlrp3, and Nlrc4), the DNA sensor Aim2, the adaptor ASC, and pro-caspase-1 are important components of inflammasomes. Stimulation with specific microbial and endogenous molecules leads to inflammasome assembly and caspase-1 activation. Inflammasomes are believed to mediate host defense against microbial pathogens and tissue homeostasis within the intestine, and their dysregulation might contribute to inflammatory diseases and intestinal cancer. Improving our understanding of inflammasome signaling pathways could provide insights into pathogenesis of many gastrointestinal disorders and the development of therapeutics targets and approaches to treat diseases such as inflammatory bowel diseases and GI cancers.

Introduction

Within the gastrointestinal (GI), innate immune receptors function as an immediate defense mechanism against invasive pathogens. In 2004, Rakoff-Nahoum et al. discovered that pattern recognition receptors (PRRs), which sense commensal bacteria, maintain intestinal homeostasis and resistance to injury¹. This was not surprising because PRRs signal through inflammatory pathways that include factors such as nuclear factor (NF)-κB, mitogen-activated protein kinase (MAPK), and caspase-1. PRRs are not only involved in controlling infection and bacterial colonization, but also in regulating intestinal epithelial barrier function, epithelial repair, and immune homeostasis^1–5. Therefore, in the GI tract, defects in PRR function could be involved in the pathogenesis of disorders such as infectious colitis, inflammatory bowel diseases, and cancer.

There are at least 4 major classes of PRRs that are involved in pathogen recognition. These are the Toll-like receptors (TLRs), the Nod-like receptors (NLRs), the RIG-I-like receptors (RLRs), and the C-type lectin receptors. Generally, these PRRs sense conserved structural motifs or pathogen-associated molecular patterns (PAMPs) in microbes, such as lipopolysaccharide or peptidoglycan, which are in the bacterial cell wall. PRRs are capable of sensing host-derived endogenous damage-associated molecular patterns (DAMPs), which are typically generated during cellular injury or tissue damage⁶. In contrast to the TLRs, which are located on the extracellular surface of cells and on endosomes, NLRs are located within the cytoplasm and recognize PAMPs and DAMPs that have gained access to the intracellular environment. NLRs are defined by a tripartite structure: first, an N-terminal caspase recruitment domain (CARD), pyrin domain (PYD), acidic transactivating domain, or baculovirus inhibitor repeat (BIR) that mediate downstream protein protein interactions; second, a central nucleotide-binding oligomerization (NOD) domain, which mediates self-oligomerization important during activation; and third, C-terminal leucine-rich repeats (LRR) that determines ligand specificity⁷.

The NLR family member Nod2 has been associated with risk for Crohn’s disease. However, a subclass of NLRs that participate specifically in inflammasome signaling has been studied for its role in intestinal inflammation and homeostasis, because of its ability to promote production of the pro-inflammatory cytokines interleukin (IL)-1β and IL-18. We review the characteristics, function, and roles of inflammasomes in inflammation, homeostasis, host defense, and pathologies of the GI tract.

Definition and Components

The inflammasome is multi-protein platform that is characterized by its ability to activate pro-caspase-1, which in turn, proteolytically cleaves pro-IL-1β and pro-IL-18 into their mature active forms⁸. Inflammasomes are named by the PRR that regulates its activity and dictates the nature of the upstream activating stimulus. There are 4 known inflammasomes: the Nlrp1-, Nlrp3-, Nlrc4-, and Aim2-inflammasomes based on their dependence on caspase-1 activation for IL-1β and IL-18 production⁶. Nlrp1, Nlrp3, Nlrc4, and Aim2 are PRRs that are capable of sensing both PAMPs and also DAMPs in the case of Nlrp3. With the exception of Aim2, which is a member of the interferon-inducible HIN-200 protein family, the other known inflammasomes all contain a PRR that belong to the NLR family. As part of the NLR family, these specific members contain either an N-terminal PYD or CARD domain, which can interact with the PYD domain of the adaptor protein ASC and CARD domain of caspase-1 respectively. ASC similarly contains a CARD domain in addition to its PYD domain which can interact with the CARD domain of caspase-1 during activation. Nlrc4, which does not contain an N-terminal PYD domain as the others, contains instead an N-terminal CARD domain which can theoretically directly recruit caspase-1 through a CARD-CARD interaction. Based on genetic, biochemical and NMR studies, the prevailing model for inflammasome assembly and caspase-1 activation is the following^8–10 (Figure 1). Activation of the PRR by PAMPs/DAMPS results in oligomerization through the NOD domain of the PRR. Subsequently, through CARD-CARD and PYD-PYD protein-protein interactions, a large macromolecular complex akin to the apoptosome involved in the activation of caspase-9 via Apaf-1 in apoptosis¹¹, is assembled, which serves as a scaffold for procaspase-1 recruitment and self-cleavage into active caspase-1. Caspase-1 activation then leads to the production of active IL-1β, IL-18 and processsing of other substrates such as pro-caspase-7¹². In addition, caspase-1 activation mediates pyroptosis, a specific form of early cell death induced by intracellular pathogens that promotes cellular lysis and the release of intracellular inflammatory contents to stimulate additional inflammatory signaling pathways^{13, 14}. Thus, cellular events that require IL-1β and IL-18 production or pyroptosis usually involve inflammasome activation; however, there are a few known exceptions where cleavage of pro-IL-1β and pro-IL-18 can occur independently of the inflammasome^{15, 16}.

Figure 1. Inflammasome components and assembly.

With the exception of Aim2 (not shown), all known inflammasomes consist of an NLR, which contain either a PYD domain or a CARD domain that can physically interact with the PYD or CARD domain of ASC and caspase-1, respectively. Both Nlrp1 and Nlrc4 contain a CARD domain that can directly independently of ASC; however, the addition of ASC in the inflammasome assembly can enhance its activity in the case of Nlrp1. The Nlrp3 inflammasome on the other hand may form different inflammasome complexes depending on the requirement for ASC¹¹⁵. Oligomerization of ASC and the NLR results in a macromolecular complex consisting of multiple subunits that are capable of cleaving pro-caspase-1 to its active form, resulting in the cleavage of the pro-forms of IL-1β and IL-18 to their mature, biologically active forms. Activation of caspase-1 also leads to pyroptosis.

Inflammasome Activators

Both microbial and non-microbial stimuli can induce the activation of inflammasomes. Importantly, activation of each NLR or Aim2 inflammasome is induced by specific microbial (Table I) or endogenous molecules. Although the molecular basis of the specificity is not well understood, the stimulus-specific activation of each inflammasome is largely determined by the fact that particular activating molecule(s) are sensed by individual NLRs or Aim2. Unlike Aim2 that directly binds cytosolic double-stranded DNA, there is no conclusive evidence as yet that NLRs involved in the inflammasome interact directly with activating microbial or endogenous molecules. Thus, it is possible that the recognition of upstream molecules by inflammasome NLRs is indirect and mediated through intermediate factors as it has been proposed for other mammalian NLR family members or their plant homologues^{17, 18}. We discuss below the specific stimuli and mechanisms that activate the inflammasomes that are more relevant to gastrointestinal inflammation including Nlrc4, Nlrp1, and Nlrp3.

Table 1.

Pathogen	Microbial Activator	Inflammasome	References
BACTERIAL
Staphylococcus aureus	Hemolysins	Nlrp3	Munoz-Planillo et al. 2010
Vibrio cholera	HlyA and MARTXVc	Nlrp3	Toma et al. 2010
Streptococcus pyogenes	Streptolysin O	Nlrp3	Harder et al, 2010
Chlamydia pneumonia	unknown	Nlrp3	He at. 2010
Neisseria gonorrhea	unknown	Nlrp3	Duncan et al, 2009
Mycobacterium tuberculosis	unknown	Nlrp3, Nlrc4	Koo et al. 2008; Master et al. 2008
Listeria monocytogenes	LLO, flagellin, bacterial DNA	Nlrp3, Nlrc4, Aim2	Mariathasan, 2006; Warren et al. 2008; Sauer et al. 2010; Wu et a. 2010; Tsuchiya K 2010; Meixenberger K 2010
Salmonella typhimurium	Flagellin, PrgJ	Nlrc4, Nlrp3	Franchi et al. 2006; Miao et al. 2006; Miao et al. 2010; Broz et al. 2011
Shigella flexneri	MxiI?	Nlrc4; Nlrp3	Susuki et al. 2007; Miao et al.2010; Willingham SB 2007;
Pseudomonas aeruginosa	Flagellin	Nlrc4	Franchi et al.2007; Galle et al, 2007
Legionella pneumophila	Flagellin		Amer et al.; Lightfield KL 2008; Zamboni DS, 2006
Bacillus anthracis	Lethal toxin	Nlrp1b	Nour et al. 2009, Boyden ED 2006; Terra JK, 2010
Franciscella tularensis	Bacterial DNA	Aim2	Fernandes-Alnemri et al. 2010; Rathinam et a. 2010; Jones et al. 2010
FUNGAL
Candida albicans	unknown	Nlrp3	Gross et al. 2009; Hise et al, Joly S 2009
Aspergillus fumigatus	unknown	Nlrp3	Said-Sadier et al. 2009
VIRAL
Sendai virus	unknown	Nlrp3	Kanneganti et al. 2006
Influenza A	Viral M2, viral RNA?	Nlrp3	Thomas et al. 2009; Allen et al. 2009; Ichinohe et al. 2009
Adenovirus	unknown	Nlrp3	Muruve et al. 2008
Varicella-zoster	unknown	Nlrp3	Nour et al. 2011
Cytomegalovirus	viral dsDNA	Aim2	Rathinam et al. 2010;
Vaccinia virus	viral dsDNA	Aim2	Hornung et al. 2009; Rathinam et al. 2010

NLRC4

Several Gram-negative bacteria, including Legionella pneumophila, Pseudomonas aeruginosa, and the enteric pathogens, Salmonella typhimurium and Shigella flexneri, induce caspase-1 activation via the Nlrc4 inflammasome. The activation of the Nlrc4 inflammasome requires an intact type III secretion system (T3SS) for S. typhimurium, S. flexneri, and P. aeruginosa or type IV secretion system (T4SS) for L. pneumophila^19–22. Upon infection, these secretion systems form pores in host cell membranes that mediate translocation of a wide array of virulence factors (effectors proteins) into the cell cytosol which is critical for pathogen colonization and induced pathology²³. Mutant analyses of S. typhimurium, L. pneumophila and P. aeruginosa revealed that flagellin, a component of the flagellum apparatus that is required for bacterial motility, is critical for activation of caspase-1 via Nlrc4^19–21. Because cytosolic delivery or expression of flagellin is sufficient to trigger the activation of the Nlrc4-inflammasome, small amounts of flagellin leaked via the T3SS or T4SS into the host cytosol appears to be the signal for Nlrc4 activation during bacterial infection. Caspase-1 activation induced by flagellin is independent of TLR5, another PRR that senses this component of the flagellum¹⁹. Thus, flagellin is separately recognized by two different sensors; TLR5 senses extracellular flagellin whereas Nlrc4 recognizes flagellin in the cytosol. In addition to flagellin, there is evidence that PrgJ, a conserved inner rod component of the T3SS of Salmonella, can activate Nlrc4²⁴. Notably, Shigella does not express flagellin, but robustly activates caspase-1 via Nlrc4²⁴. Although the specific molecule produced by Shigella that activates Nlrc4 remains to be identified, it is possible that its conserved PrgJ homologue is involved because expression of a functional T3SS is required for caspase-1 activation.

Recent evidence indicate that the activation of Nlrc4 requires the presence of Naip5, another NLR family member. Naip5 appears to sense the C-terminal portion of flagellin from L. pneumophila and is required for the activation of the Nlrc4 inflammasome in response to this bacterium²⁵. It was proposed that Naip5 physically associates with Nlrc4 to form an inflammasome complex in response to L. pneumophila. In contrast, caspase-1 activation induced by S. typhimurium and P. aeruginosa infection, which is also triggered by cytosolic flagellin, is largely Naip5-independent²⁵. Because L. pneumophila expressing Salmonella flagellin activates Nlrc4 in a Naip5-dependent manner²⁶, the differential requirement of Naip5 is not due to differences in the flagellin molecule between the two bacterial species. Recently, two groups have shown that flagellin directly associates with Naip5 whereas the related Naip2 binds TTSS rod proteins^{27, 28}. These results suggest that activation of the Nlrc4 inflammasome is mediated indirectly through the interaction of specific microbial molecules with distinct Naip proteins.

NLRP1

The human NLRP1 inflammasome was the first caspase-1-activating protein complex to be identified⁸. Furthermore, the NLRP1 was the only inflammasome reconstituted in vitro with purified components, which revealed that NLRP1 oligomerizes with caspase-1 in the presence of muramyl dipeptide (MDP)⁹. Using this in vitro system, it was suggested that caspase-1 is activated via a two-step mechanism whereby microbial MDP induces a conformational change in NLRP1, which, in turn, allows it to bind nucleotide and oligomerize, leading to caspase-1 activation⁹. However, direct evidence that MDP binds NLRP1 is lacking, and therefore, the mechanism that triggers NLRP1 oligomerization remains unclear. Notably, the adaptor molecule ASC enhanced, but was not required for NLRP1-mediated caspase-1 activation in vitro⁹, which is consistent with the observation that ASC is not essential for the activation of caspase-1 mediated by Nlrp1b in mouse macrophages²⁹. Thus, the adaptor ASC is essential for the activation of caspase-1 in the great majority, but not all, inflammasomes. Unlike humans that possess a single NLRP1 gene, three Nlrp1 paralogs, namely Nlrp1a, -b and –c, are present in the mouse genome. The Nlrp1b-inflammasome is activated by lethal toxin (LT), a bipartite toxin secreted by Bacillus anthracis that is composed of Protective Antigen (PA), a pore-forming subunit that delivers Lethal factor, a metalloprotease, into the cytosol of infected cells²⁹. An important role for the Nlrp1b-inflammasome in host defense is suggested by the observation that mice harboring the Nlrp1b “susceptible” allele are more sensitive to LT^{30, 31}. Recent evidence indicate that caspase-1 activation in response to B. anthracis infection is beneficial to the host^{32, 33}. However, the mechanism by which Lethal factor, triggers Nlrp1b-inflammasome activation remains poorly understood.

NLRP3

Unlike Nlrc4 that is activated primarily by cytosolic flagellin, a large array of microbial and non-microbial stimuli have been reported to activate the Nlrp3 inflammasome in macrophages. These include several TLR agonists and the Nod2 agonist MDP in the presence of extracellular ATP^34–36. In addition, Nlrp3 is activated by certain bacterial toxins and particulate matter including urate crystals, silica, asbestos, β-amyloid, and aluminum hydroxide in phagocytes pre-stimulated with microbial ligands such as lipopolysaccharide (LPS)^37–40. The ability of multiple PAMPs to activate the Nlrp3-inflammasome is puzzling because most of these molecules including TLR ligands are structurally unrelated. Recent findings suggest that most, if not all, TLR agonists as well as MDP, do not act as direct Nlrp3 activators^{41, 42}. Instead, these microbial stimuli promote inflammasome activation indirectly through the induction of Nlrp3 via NF-κB signaling whereas ATP provides the direct signal for inflammasome activation^{41, 42} (Figure 2A). Similarly, bacterial toxins and particulate matter activate Nlrp3, but they still require pre-stimulation with TLR ligands which induce Nlrp3 expression^37–40(Figure 2B). Consistently, TNF-α and IL-1 that activate NF-κB are as effective as TLR agonists in promoting caspase-1 activation in response to ATP or silica^{40, 42}

Figure 2. Activation of the Nlrp3 inflammasome requires 2 steps.

A, Priming step: Based primarily on studies of Nlrp3, inflammasome activity requires first the production of pro-IL-1β and Nlrp3 through upregulation of NFκB by either activation of the TLRs or Nod1/2. Inflammatory cytokines such as TNFα or IL-1 can also induce NFκB. In addition, this first signal (i.e., cytokines, TLR or Nod1/2 activation) also results in the transcriptional upregulation of Nlrp3. B, Activation step: It remains unclear how Nlrp3 is activated by diverse signals. Three different mechanisms have been suggested: 1) Intracellular potassium depletion by the opening of a pore via ATP-dependent P2X7R activation or microbial pore-forming toxins, 2) lysosomal membrane damage and release of activated cathepsin B after endocytosis of sterile particulates such as silica, asbestos and cholesterol crystals, and 3) generation of ROS from the mitochondria as a consequence of cellular injury (although ROS may affect inflammasome priming only).

Mechanism of NLRP3 activation

Activation of the Nlrp3 inflammasome requires 2 signals. The first signal is provided by microbial molecules such as TLR ligands or by certain cytokines that induce induction of Nlrp3 via NF-κB (Figure 2A). The second signal directly triggers caspase-1 activation, and can be mediated by several stimuli including extracellular ATP, particulate matter, and certain bacterial toxins (Figure 2B). It is likely that these different pathways culminate in a common step that leads to Nlrp3 activation. However, the identification of a unifying mechanism of Nlrp3 activation remains elusive. Extracellular ATP acting through the ATP-gated P2X7 receptor (P2X7R) and bacterial toxins that activate Nlrp3 in a P2X7R-independent manner induce potent K⁺ efflux, a signal that appears to be required for Nlrp3 inflammasome activation⁴³. Although Nlrp3-inflammasome activation is effectively blocked by high concentrations of extracellular K⁺, it is not clear whether K⁺ efflux is induced by all stimuli that activate Nlrp3 and is sufficient to trigger inflammasome activation. Earlier studies suggested that P2X7R activates Nlrp3 via the formation of a large pore mediated by pannexin-1^{44, 45}. However, recent work clearly showed that Nlrp3 activation proceeds independently of pannexin-1⁴⁶. Lysosomal membrane damage induced by silica and other particles has been suggested to play an important role in Nlrp3 activation through the activation of lysosomal proteases and specifically cathepsin B³⁸. However, caspase-1 activation-induced by ATP is cathepsin B-independent³⁸, suggesting that cathepsin B activation is not the unifying link to Nlrp3 activation. The generation of reactive oxygen species (ROS) was proposed to be the common step in Nlrp3 activation largely based on the observation that caspase-1 activation is blocked by ROS inhibitors^{39, 47, 48}. However, recent studies have challenged the role of ROS in Nlrp3 activation. For example, recent evidence indicate that ROS inhibitors do not block the activation of the Nlrp3 inflammasome, but prevent the upregulation of Nlrp3 and pro-IL-1β, thus acting on the priming of cells⁴⁹. Furthermore, the evidence suggesting a critical role of thioredoxin-interacting protein (TXNIP), an inhibitor of the antioxidant thioredoxin, in Nlrp3 activation was not reproduced⁵⁰. Thus, a critical role for ROS in inducing Nlrp3 inflammasome activation is doubtful. Clearly, further work is needed to understand the mechanism of the Nlrp3 inflammasome.

NLRP6 as a Possible Regulator of the Inflammasome

Nlrp6 is an as yet poorly characterized NLR that contains an N-terminal PYD domain, a central NOD domain, and C-terminal LRRs. Although the nature of the signal which activates Nlrp6 remains unknown, early studies based on the overexpression of NLRP6 suggested its role as an inflammasome participant⁵¹. Specifically, co-expression of NLRP6 with ASC resulted in cooperative production of IL-1β in COS-7L cells that was caspase-1 dependent and required the presence of the PYD domain⁵¹. Furthermore, immunofluorescence studies demonstrated colocalization of NLRP6 with ASC in a characteristic speckled pattern within the cytoplasm⁵¹. Presumably, the PYN domain of NLRP6 may directly interact with the PYD domain of ASC to enable caspase-1 activation, although a direct interaction with ASC or caspase-1 has never been demonstrated. More recently, Nlrp6-deficient mice were shown to have impaired steady-state levels of IL-18 production within the serum and in colon explant cultures compared with wildtype mice⁵², consistent with a role in inflammasome signaling although we have not observed differences in basal IL-18 levels between Nlrp6^−/− and wildtype colon tissue homogenates⁵³.

Until recently, a function for Nlrp6 was unknown. Expression analysis of Nlrp6 on an mRNA level showed that it was highly expressed in the intestine and significantly less so in hematopoietic cells^{51, 53–55}. However, in hematopoietic cells, Nlrp6 was demonstrated to be expressed differentially with greater expression levels detected in granulocytes and lymphocytes as compared with macrophages and dendritic cells⁵¹. Others have reported high expression levels within myofibroblasts, which have important roles in intestinal epithelial repair^{55, 56}, although whether Nlrp6 functions within this group of cells is unknown. In the past year, several reports have confirmed a role for Nlrp6 in promoting intestinal homeostasis (see below). Whether Nlrp6 truly functions as part of an inflammasome however remains to be determined in vivo and will require a better understanding of its upstream signals and whether the production of IL-1β and IL-18 via Nlrp6 is ASC and caspase-1 dependent.

Host Defense

Acute infectious diarrhea is a major cause of morbidity and mortality worldwide especially in developing countries and in children. Pathogens such as Staphylococcus aureus, enteropathogenic Escherichia coli, and Vibrio cholera are capable of producing toxins that cause diarrhea and others such as Salmonella, Campylobacter, Shigella and rotavirus are invasive and can cause severe intestinal inflammation. Inflammasomes play an important role in host defense against enteropathogenic bacteria. For example, there is evidence that activation of the Nlrc4 inflammasome functions as a host defense strategy against pathogenic infections. In the case of S. typhymurium and S. flexneri, Nlrc4-dependent activation of caspase-1 is accompanied by the secretion of IL-1β and the induction of pyroptosis^{19, 57}. Surprisingly, while both Nlrc4 and the adaptor molecule ASC are required for caspase-1 activation and IL-1β secretion, Nlrc4, but not ASC, is critical for the induction of pyroptosis^{21, 22}. There is evidence that the inflammasome is important in host defense against Salmonella infection in vivo. For example, caspase-1-null mice are more susceptible to oral infection which is associated with increased bacterial loads in the spleen and mesenteric lymph nodes^{58, 59}. Caspase-1 plays a role in the systemic phase of the infection because caspase-1 deficient mice were more susceptible than wild-type mice when challenged with Salmonella intraperitoneally⁵⁹. Furthermore, IL-18 was suggested as the critical cytokine responsible for host defense against Salmonella⁵⁹. However, Nlrc4-deficient mice were as susceptible as wild-type mice⁵⁸. The latter results can be explained by the observation that Salmonella down-regulates the expression of flagellin during the systemic phase of the infection⁶⁰. Consistently, Nlrc4 plays an important role in host defense against the bacterium when expression of flagellin in Salmonella is enforced during the systemic phase of the infection⁶¹. Salmonella encodes two T3SS, SPI-1 promotes invasion of intestinal epithelial cells while SP1-2 promotes replication in macrophages ⁶². Recently, it was reported that in SPI-1 deficient Salmonella strains Nlrc4 acts redundantly with Nlrp3 to induce the activation of caspase-1⁶³. The mechanism by which SPI-2 activates the Nlrp3 inflammasome, however, remains unknown.

Infection of macrophages by several bacterial pathogens including Salmonella triggers pyroptosis that relies on caspase-1 activation⁶⁴. Surprisingly, while both Nlrc4 and the adaptor ASC are required for caspase-1 activation and IL-1β secretion, Nlrc4, but not ASC, is critical for the induction of pyroptosis^{21, 22, 65}. Although the precise mechanism remains poorly understood, there is evidence that IL-1β secretion and pyroptosis are mediated by distinct caspase-1 signaling pathways that can be separated by the involvement of ASC⁶⁵. In the intestinal phase, Salmonella expresses SPI-1 and flagellin, leading to activation of caspase-1 via Nlrc4 whereas in the systemic phase, Salmonella expresses neither SPI-1 nor flagellin. However the forced expression of flagellin during the systemic phase of the infection reveals an important role for Nlrc4-dependent pyroptosis in promoting bacterial clearance. Specifically Nlrc4-induced pyroptosis promoted the release of intracellular bacteria from phagocytic cells leading to uptake and killing of the pathogen by neutrophils⁶¹. This process occurs independently of IL-1β and IL-18 possibly explaining why the adaptor ASC is not required to regulate the susceptibility to Salmonella infection^{58, 61}.

Several pathogenic microorganisms including certain viruses, fungi and bacteria induce the activation of the Nlrp3-inflammasome. Among the pathogenic bacteria, S. aureus, S. pyogenes, and V. cholera activate caspase-1 via Nlrp3^66–68. Specifically, pore-forming or membrane-damaging toxins produced by pathogenic bacteria are important in inducing activation of the Nlrp3-inflammasome^66–68. Unlike extracellular ATP, Nlrp3 activation induced by bacterial infection is independent of the P2X7R^66–68. Nlrp3 also regulates IL-1β production in response to several viruses, such as influenza A, fungi, such as Candida albicans and parasites, including Plasmodium^{69, 70}. However, whether Nlrp3 regulates host defense against viruses or helminths that specifically cause pathology in the intestine remains to be determined.

Tissue Repair and GI Homeostasis

IL-1β vs IL-18

Inflammatory bowel disease is a debilitating condition that afflicts approximately 1.4 million people in the United States and 2.2 million in Europe⁷¹. IBD encompasses two different diseases, ulcerative colitis (UC) and Crohn’s disase (CD) in which the intestine becomes chronically inflamed leading to the common clinical presentation of abdominal pain, bloody diarrhea and weight loss. UC and CD are distinguished by clinical and endoscopic criteria as well as histology. In UC, inflammation is restricted to the mucosal layer of the colon typically in the rectum and can extend continuously to involve other areas of the colon, whereas in CD, inflammation is transmural and can occur in a discontiguous fashion (‘skip lesions’) that involve any part of the gastrointestinal tract, most commonly in the distal ileum. Because of the transmural nature of involvement, complications such as bowel strictures and fistula formation can occur that is rarely associated with ulcerative colitis. The pathogenesis of IBD remains unclear, but studies suggest both an environmental and genetic contribution^{72, 73}. Generally, the prevailing model for the development of IBD involves an aberrant immune response to commensal bacteria and/or an imbalance in the structure of the gut microbiota⁷⁴.

An interest in the role of inflammasomes in the pathogenesis of IBD arose from observations that polymorphisms in genes encoding IL-18 and the IL-18 accessory protein were associated with increased susceptibility to Crohn’s disease^{75, 76}. In addition, polymorphisms in specific inflammasome components, particularly Nlrp3, which resulted in impaired IL-1β production by LPS-stimulated monocytes was associated with increased susceptibility to CD⁷⁷ although this observation was not reproducible in different populations⁷⁸. Nonetheless, these initial genetic linkage studies suggest a role for inflammasome signaling and IL-1β and IL-18 production in IBD development.

There are several mouse models of IBD that have been used to understand mechanisms of IBD development, but the one that has been used most frequently to investigate the role of the inflammasome has been the chemically-induced colitis by dextran sulfate sodium (DSS)⁷³. DSS causes direct epithelial injury causing epithelial cell death and increased intestinal permeability, which results in mucosal ulceration and erosion, infiltration of inflammatory cells, and upregulation of proinflammatory cytokines^{79,80, 81}. Lesions typically affect only the mucosal layer primarily in the colon and rectum as in ulcerative colitis, but can develop discontinuously unlike UC. Criticism against the DSS model that is more representative of acute epithelial injury has been the lack of T cell-dependence, an important component of IBD, as DSS-induced colitis can still occur in T-cell deficient animals, although T cells can contribute to its severity^82–85. Nonetheless, the DSS model has been useful in identifying critical factors important in the intestinal epithelial repair and barrier function as well as innate immune responses involved in the development and control of enteric inflammation.

Since IBD is typically associated with upregulation of proinflammatory cytokine induction including IL-1β and IL-18^{86, 87}, one can reasonably hypothesize that downregulation of these proinflammatory cytokines would be associated with decreased disease severity. Indeed, it was initially demonstrated that caspase-1-deficient Balb/c mice were less susceptible to DSS-induced colitis with improved clinical and histologic scores compared to wildtype mice⁸⁸. Consistently, IL-18 overexpressing transgenic mice were associated with increased colitis⁸⁹ and chemical inhibition of IL-18 or caspase-1 was associated with amelioration of disease^{90, 91}.

However, more recent studies have demonstrated a negative regulatory role of inflammasome signaling in the development of colitis. Specifically, both ASC- and caspase-1-deficient B6 mice had increased susceptibility to DSS with impaired epithelial proliferation and restitution, increased intestinal permeability, and greater translocation of commensal bacteria into the colonic mucosa and mesenteric lymph nodes⁹². Although caspase-1 and ASC deficiency was associated with reductions in both IL-1β and IL-18 production, the mechanism behind the increased severity of inflammation in caspase-1-deficient mice was linked to an impairment in IL-18 production, as administration of recombinant IL-18 was sufficient to rescue caspase-1 deficient B6 mice⁹². It was further suggested that the source of IL-18 was in the intestinal epithelial cell based on in vitro cultures of intestinal cells enriched in epithelial versus lamina propria cells from DSS-treated mice⁹². The impairment in IL-18 production in caspase-1 deficient mice during DSS-induced colitis and the ability to rescue with administration of IL-18 was also reproduced by a separate group giving credence to the model that caspase-1dependent IL-18 production, presumably by the intestinal epithelium, is important for epithelial repair and regeneration during DSS-induced colitis⁹³ (Figure 3). However, the source of IL-18 may not solely be from the epithelium because there is evidence that IL-18 is produced by both epithelial and lamina propia cells in IBD ⁹⁴.

Figure 3. IL-18 is important for intestinal epithelial repair.

Nlrp3-, ASC- and caspase-1-deficient mice have increased susceptibility to DSS-induced colitis. DSS causes direct epithelial injury resulting in increased permeability and translocation of bacteria into the breached mucosa leading to an inflammatory response that includes the recruitment of immune cells such as macrophages and neutrophils. The production of IL-18, such as by the Nlrp3-inflammasome, allows the epithelium to be fully restituted, limiting the extent of inflammation. The precise cellular source of IL-18 and how IL-18 promotes epithelial repair (e.g., whether it acts directly on the epithelial cell or indirectly through lamina propria cells) remain unclear.

IL-18 has traditionally been considered pro-inflammatory with importance in inducing Th1 responses, specifically IFN-γ, and therefore the precise mechanism by which IL-18 promotes early tissue repair remains unclear⁹⁵. However, the IL-18 receptor signals through the adaptor protein MyD88, which has been implicated in intestinal epithelial repair through interactions with prostaglanding and Cox-2 signaling pathways^{1, 96, 97}. IL-18 has also been shown to be important in other models of wound repair such as in the skin⁹⁸. Thus, the extent of IL-18 production may dictate the level of inflammation and repair such that any dysregulation in IL-18 production can disrupt this fine balance within the intestine. It remains intriguing why there are discrepant results in the phenotype of caspase-1 deficient mice between the current studies and earlier studies, and it may have been due in part to differences in mouse genetic background (Balb/c versus B6, respectively). However, it has also been recently reported that caspase-1 deficient B6 mice treated with DSS did not demonstrate any difference in inflammation than wildtype mice⁹⁹. Although differences in the length and concentration of DSS treatment (2% DSS for 7 days versus 3% for 5 days) may be one explanation, a more likely and concerning possibility is the difference in the gut microbiota between different facilities.

Role of NLRP3 in GI Homeostasis

Which PRR component of the inflammasome is important for caspase-1 activation and protection against chemically-induced injury and intestinal homeostasis? Given the possible genetic association between Nlrp3 and IBD⁷⁷, much attention has been given to the role of Nlrp3. Three different groups have demonstrated that Nlrp3-deficient mice had increased susceptibility to DSS-induced colitis with increased mortality and weight loss^{93, 100, 101}. Nlrp3-deficient mice had increased mortality, increased rectal bleed and colonic inflammation histologically. The increased colonic inflammation in DSS-treated Nlrp3-deficient mice was associated with increased intestinal permeability and translocation of bacteria to the liver and mesenteric lymph nodes. Similar to what was observed in caspase-1 deficient mice, Nlrp3-deficient mice had an impairment in epithelial proliferation, suggesting a defect in epithelial repair as the mechanism to the increased severity of symptoms⁹³ (Figure 3). However, whether Nlrp3-deficient mice had impaired IL-18 production or whether the administration of recombinant IL-18 is capable of rescuing Nlrp3-deficient mice has not yet been demonstrated. Nonetheless, bone marrow chimera experiments suggested that Nlrp3 function in radioresistant, non-hematopoietic cells is important for protection against DSS-induced colitis, consistent with the model of intestinal homeostasis maintained by intestinal epithelial IL-18 production although a role for other non-hematopoietic cells has not been entirely excluded⁹³(Figure 3). Therefore, IL-18 production to promote intestinal homeostasis may be mediated by the Nlrp3 inflammasome. Another possible mechanism for the protective role of Nlrp3 was suggested by a study by Hirota et al. who demonstrated an impairment in β-defensin production in Nlrp3-deficient mice that was associated with a microbiota composition distinct from that in wildtype mice¹⁰¹. Thus, an interesting possibility yet to be fully explored is that Nlrp3-deficiency results in defective antimicrobial mechanisms, which, in turn, causes bacterial dysbiosis that leads to increased susceptibility to DSS-induced colitis.

However, despite the convincing data that Nlrp3 functions to protect against DSS-induced colitis, there are at least two groups that do not show a negative regulatory effect of Nlrp3 on colitis. Bauer et al., found that their Nlrp3-null mice had less severe colitis when treated with DSS, which was related to decreased IL-1β secretion of DSS-treated Nlrp3-deficient macrophages in vitro¹⁰². Similarly pharmacological inhibition of caspase-1 with pralnacasan also resulted in decreased severity of colitis⁹⁰. A second group has also shown an attenuated colitis pattern in Nlrp3-deficient mice⁵². It is difficult to attribute these differences entirely to differences in protocol. Rather, baseline differences in the microbiota may more likely explain dissimilar phenotypes.

NLRP6 and Regulation of the GI Microbiota

DSS-induced colitis has not been studies with respect to other known inflammasomes such as Nlrc4, Aim2, or Nlrp1. However, as discussed above, Nlrp6 likely participates in inflammasome signaling based on in vitro studies and is also highly expressed in the intestine. Recent reports now confirm Nlrp6 to be another important factor in the regulation of intestinal homeostasis^{52, 53, 55}. We have recently reported that Nlrp6-deficient mice have increased susceptibility to DSS-induced colitis which was also demonstrated by two other groups^{52, 53, 55}. Flavell and colleagues provide a possible mechanism for Nlpr6-mediated protection (Figure 4). Through high throughput culture-independent 16S RNA bacterial gene sequencing, they demonstrated that the gut microbiota of Nlrp6 mice was distinct from that in wildtype mice, and furthermore, the microbiota associated with Nrlp6 mice were colitogenic that was transferrable to wildtype mice with co-housing⁵². This phenomenon was specific to Nlrp6 and did not occur with mice deficient in other inflammasome components such as Nlrc4 and Aim2. Specifically, they showed that there was greater colonization of the bacterial genus Prevotollaceae in Nlrp6-deficient mice⁵². Treatment of Nlrp6-deficient mice with antibiotics resulted in a reduction of these bacteria, improvement in colitis, and reduced transferability of colitis to wildtype mice after cohousing⁵². Interestingly, the authors demonstrate that the gut microbiome of Nlrp6-deficient mice was associated with increased CCL5 production. While CCL5-deficient mice developed DSS-induced colitis that was not significantly different from wildtype mice, CCL5-deficient mice developed colitis that was less severe than wildtype mice after cohousing with Nlrp6-deficient mice despite an equivalent transfer of Prevotellaceae. The authors therefore suggest that the aberrant microbiota of Nlrp6-deficient mice results in greater CCL5 production that predisposes to greater inflammation upon chemically-induced epithelial damage by DSS as Nlrp6-deficient mice do not develop spontaneous colitis⁵² (Figure 4). In addition, Nlrp6 had impaired IL-18 production specifically within the non-hematopoietic compartment, or epithelium⁵², which can further reduce the ability of Nlrp6-deficient mice to recover from DSS-induced epithelial injury⁵². This would be consistent with the proposed model for caspase-1 activation and IL-18 production in intestinal homeostasis; however, whether the administration of IL-18 is capable of rescuing Nlrp6-deficient mice and furthermore, whether Nlrp6 function specifically in the epithelium by generating bone marrow chimeras was not investigated and remains to be determined. It is also worth mentioning that it remains unclear whether Nlrp6 truly regulates the gut microbiota as bacterial sequencing was performed on mice that were not littermates, and therefore differences in microbiota may merely have been due to differences in ancestry. Although antibiotic treatment was associated with decreases in Prevotellaceae, effects on other bacterial groups that contribute to susceptibility to colitis also cannot be excluded. Regardless, these studies have identified potentially colitogenic bacteria that can be targeted for therapeutic purposes. Monocolonization experiments with germfree mice would be helpful to definitively determine the true colitogenic potential of the Prevotellaceae and TM7 species.

Figure 4. Model of Nlrp6-mediated regulation of intestinal and microbial homeostasis.

Nlrp6 maintains intestinal homeostasis by regulating the composition of the gut microbiome and the production of IL-18 (left). In the absence of Nlrp6 (right), there is dysbiosis resulting in the accumulation of colitogenic bacteria, upregulation of inflammatory mediators, such as CCL5, and subclinical inflammation. Upon additional epithelial damage by DSS, rampant inflammation ensues as a result of impaired epithelial repair from decreased IL-18 production..

Tumorigenesis

IL-18 in Colitis-Associated Tumorigenesis

Colorectal cancer is the third most common cancer and is also the third most common cause of cancer-related deaths. There are two major risk factors for the development of colorectal cancer, that is, a genetic predisposition, and inflammatory bowel disease. Even in sporadic colon cancers, 80% is associated with a genetic mutations in the tumor suppressor APC. APC participates in the Wnt signaling pathway which is upregulated in colon stem cells and activated under conditions of cellular proliferation. Mice that harbor a mutation in the Apc gene (ApcMin mice), develop spontaneous small intestinal and colon tumors, although the majority of tumors are in the small intestine. These mice are often used to model spontaneous colon carcinogenesis¹⁰³. Innate immune signaling and commensal bacteria have been shown to be important in this mouse model as ApcMin mice that are also deficient in the MyD88 have decreased tumor potential, and germ-free ApcMin mice have a slight decrease in small intestinal tumorigenesis^{104, 105}. There, however, does not appear to be a role for inflammasomes in spontaneous tumorigenesis in this model as ApcMin mice that were crossed to either IL-1 receptor (IL-1R) or caspase-1 deficient mice had no difference in polyp formation compared to ApcMin mice¹⁰⁶. Furthermore, the administration of the IL-1R antagonist anakinra did not affect the extent of tumorigenesis¹⁰⁶.

On the other hand, in the presence of inflammation, inflammasomes can have a critical role in colon tumorigenesis. Colitis-associated colon cancer (CAC) is a major complication of IBD, and directly correlates with the extent and duration of colitis. The link between chronic inflammation and carcinogenesis can be explained by the production of DNA-damaging oxygen radical species, proinflammatory mediators that promote cellular survival, enhance proliferation and angiogensis and tissue remodeling, resulting in a microenvironment conducive to tumorigenesis¹⁰⁷. A popularly used mouse model for CAC is the AOM/DSS model in which the experimental carcinogen azoxymethane is use to introduce genomic mutations by methylation followed by repeated rounds of DSS to induce a chronic pattern of colitis¹⁰⁸. After three rounds of DSS, mice develop grossly visible adenomas, and if mice are sacrificed at later time points, the proportion of adenocarcinomas increases, suggesting that this model recapitulates events involved in the progression from premalignant adenomas to adenocarcinomas in humans¹⁰⁹. However, this model has been criticized as not fully encapsulating important features of CAC, such as the frequency of APC mutations which occurs early in the AOM/DSS model, but late and less commonly in CAC¹¹⁰, and the predominance of adenomatous polyps typical of the AOM/DSS model as opposed to flat dysplastic lesions more commonly seen in IBD¹¹¹. Regardless, this model has been useful in identifying factors and signaling pathways important in modulating inflammation and subsequent carcinogenesis.

Given the tight association between inflammation and colorectal cancer, it is not surprising that inflammasomes which function to protect the intestine from excessive inflammation after chemically-induced injury would have a role in determining susceptibility to CAC (Figure 5). In the AOM/DSS model, although IL-1R-deficient mice develop a similar number of tumors as wildtype mice, IL-18 and IL-18R-deficient mice both have increased number of tumors, suggesting a role for inflammasome signaling in tumor suppression in the context of chronic inflammation¹¹². Mice deficient in MyD88, which is downstream of IL-18 signaling, mice also had increased tumors in the AOM/DSS model associated with depressed levels of various DNA repair factors in addition to enhanced expression of mitogenic and angiogenic genes.¹¹² Thus, it has been suggested that altered DNA repair mechanisms may contribute to increased tumorigenesis in mice deficient in MyD88, which is downstream of IL-18 receptor signaling.

Figure 5. The inflammasome protects against colitis-associated tumorigenesis through multiple mechanisms.

Mice deficient in inflammasome activity have increased susceptibility to colitis-associated tumorigenesis. Specifically, Nlrp3, Nlrc4, and Nlrp6 all have been demonstrated to negatively regulate colitis-associated tumorigenesis (see text for details). Bone marrow chimera experiments suggest that Nlrp3 and Nlrp6 mediate tumor suppression within the hematopoietic compartment; however, mouse chimera experiments suggest that a non-hematopoietic compartment (epithelial or stromal cells) is important for Nlrc4-mediated tumor suppression. Although the mechanism still remains unclear, IL-18 may be an important factor in the protection against neoplasia by Nlrp3 and Nlrp6. Unlike Nlrp3 and Nlrp6, the absence of Nlrc4 signaling is not associated with increased inflammation. Instead, Nlrc4 appears to limit tumor development by regulating epithelial proliferation and apoptosis either in the intestinal epithelium or within the tumor itself.

NLRP3 in CAC

Upstream of IL-18, Nlrp3 has also been reported to negatively regulate colitis-associated tumorigenesis^{100, 113}. However, the mechanism remains unclear. Kanneganti and colleagues demonstrate that caspase-1-deficient mice have more tumors similar to Nlrp3-deficient mice that was associated with decreased IL-18 production and STAT1 activation¹¹³. Administration of IL-18 restored wildtype levels of STAT1 phosphorylation and therefore, the authors posit that IL-18 signaling is important for IFNγ induction and immune surveillance¹¹³. However, whether normalization of STAT1 signaling with IL-18 administration resulted in less tumors was not demonstrated. Ting and colleagues also demonstrated increased susceptibility of Nlrp3-deficient and consistently ASC-deficient mice to colitis-associated tumors associated with a defect in IL-18 production during the induction of inflammation in the AOM/DSS model¹⁰⁰. Interestingly, in bone marrow chimera experiments, they show that Nlrp3 is important in the hematopoietic compartment rather than in the epithelial compartment, suggesting a role for the inflammasome in tumor suppression that is independent of IL-18 production by the intestinal epithelium. It is also possible that in contrast to inflammation, tumorigenesis requires inflammasome function in the hematopoietic cell compartment as opposed to the epithelial compartment.

NLRP6 suppresses intestinal tumorigenesis in addition to inflammation

Similar to what was observed for Nlrp3, we have also demonstrated an important role for Nlrp6, particularly in the hematopoietic compartment, in protecting against the development of colitis associated tumors⁵³. This result is intriguing given the fact that Nlrp6 is more highly expressed in the epithelium and to a much lesser extent in hematopoietic cells, although it is possible that similar to Nlrp3, Nlrp6 expression is inducible upon activation⁵³. The increased tumorigenesis was associated with increased inflammatory responses, but defective IL-18 production within the colon after AOM/DSS treatment⁵³. How this relates to increased tumor potential remains to be delineated especially given the predominance of IL-18 production by the intestinal epithelium rather than by colon lamina propria or bone marrow-derived cells⁵³. Using gene expression analysis by microarray profiling of tumors in Nlrp6 and WT mice, Chamaillard and colleagues observed differences in the signaling molecules of the pathways important in epithelial proliferation and transformation, such as the Notch and Wnt pathways, and therefore suggest that dysregulation of these pathways in Nlrp6-deficient mice are contributing factors to tumorigenesis⁵⁵. Clearly, additional experiments investigating the relationship between IL-18 production, Nlrp6 function in the hematopoietic cell, and other carcinogenic signaling pathways potentially regulated by Nlrp6 need to be done.

NLRC4 Regulates Tumorigenesis Independently of Inflammation

Nlrc4 has also been demonstrated to be capable of suppressing tumors, although this remains controversial as 2 different groups have shown inconsistent results with one study demonstrating no role for Nlrc4¹⁰⁰ in colitis-associated tumor development and another showing a negative regulatory role for Nlrc4⁹⁹. Specifically, Flavell and colleagues demonstrate that Nlrc4-deficient mice developed increased tumors in the AOM/DSS model, which was unrelated to inflammation, as the severity of colitis in these mice was no different from that observed in wildtype mice⁹⁹. Instead, Nlrc4 deficiency was associated with increased epithelial proliferation and decreased apoptosis in advanced tumors⁹⁹. Although the mechanism behind this phenomenon is unclear, the authors demonstrate in bone marrow chimeric experiments that Nlrc4 function in the non-hematopoietic cell compartment is important for tumor suppression and suggest that Nlrc4 signaling within the epithelium is important for regulating apoptosis and proliferation within tumors^{99, 114}. However, caspase-1 function was required in both hematopoietic and epithelial compartments⁹⁹, perhaps due to the influence of other upstream inflammasome components that function in the hematopoietic compartment for tumor suppression (e.g., Nlrp3 or Nlrp6). Interestingly, Ting and colleagues found no difference in tumor potential between Nlrc4 and wildtype mice using the AOM/DSS model¹⁰⁰, suggesting again that differences in the gut microbiota in different facilities may influence the outcome of Nlrc4 deficiency.

Conclusion

Inflammasomes, as PRRs, are important in the defense against pathogenic organisms that can invade the GI tract. However, an extended role for inflammasomes in maintaining the integrity of the intestinal epithelium and promoting repair has clearly emerged, placing inflammasomes as important players in the pathogenesis of inflammatory bowel disease and cancer. Recent studies have also implicated the inflammasome in regulating the GI microbiome, which in turn, can affect host susceptibility to diseases beyond the GI tract, including obesity and diabetes. As the repertoire of microbial and host-derived signals sensed by inflammasomes become better understood, strategies to modulate inflammasome activity may be extremely useful in the development of therapeutics for a variety of diseases, not just limited to the GI tract.

Abbreviations used in this paper

AOM

azoxymethane

ASC

adaptor apoptosis associated speck-like protein

CAC

colitis-associated colon cancer

CARD

caspase recruitment domain

CD

Crohn’s disease

DSS

dextran sulfate sodium

DAMPs

damage-associated molecular patterns

IBD

inflammatory bowel disease

IL-1β

interleukin-1β

IL-18

interleukin-18

LPS

lipopolysaccharide

LRRs

leucine-rich repeats

MDP

muramyl dipeptide

NLRs

Nod-like receptors

NOD

nucleotide-binding oligomerization domain

PAMPs

pathogen-associated molecular patterns

P2X7R

purinergic P2X7 receptor

PRRs

pattern-recognition receptors

PYD

pyrin domain

RLRs

RiG-I-like receptors

T3SS

type III secretion system

T4SS

type IV secretion system

TLRs

Toll-like receptors

UC

ulcerative colitis