|Home | About | Journals | Submit | Contact Us | Français|
Yersinia pestis, the causative agent of plague, is able to suppress production of inflammatory cytokines IL-18 and IL-1β, which are generated through caspase-1–activating nucleotide-binding domain and leucine-rich repeat (NLR)-containing inflammasomes. Here, we sought to elucidate the role of NLRs and IL-18 during plague. Lack of IL-18 signaling led to increased susceptibility to Y. pestis, producing tetra-acylated lipid A,and an attenuated strain producing a Y. pseudotuberculosis-like hexa-acylated lipid A. We found that the NLRP12 inflammasome was an important regulator controlling IL-18 and IL-1β production after Y. pestis infection, and NLRP12-deficient mice were more susceptible to bacterial challenge. NLRP12 also directed interferon-γ production via induction of IL-18, but had minimal effect on signaling to the transcription factor NF-κB._ These studies reveal a role for NLRP12 in host resistance against pathogens. Minimizing NLRP12 inflammasome activation may have been a central factor in evolution of the high virulence of Y. pestis.
Inflammasomes are multimolecular complexes consisting of inactive pro-caspase-1 and members of the nucleotide-binding domain-leucine rich repeat (NLR) family of immune system proteins (Latz, 2010) The assembly of an inflammasome leads to proteolytic activation of caspase-1 which in turn cleaves pro-interleukin (IL)-1β and pro-IL-18 into mature forms (Latz, 2010). Active IL-1β and IL-18 are essential members of host defenses towards various pathogens, and may also participate in sterile inflammatory processes. The NLR family has more than 20 members, however, many of these proteins have unknown functions (Martinon et al., 2009), and their relative roles in promoting resistance to infection are in many instances unclear. There is evidence supporting a function in bacterial recognition for several NLRs. These include NOD1/2 (recognizing peptidoglycan fragments) (Martinon et al., 2009), NLRP1 (sensing anthrax lethal toxin) (Averette et al., 2009), NLRP3 (activated by exposure to many pathogens, bacterial RNA, toxins and crystal structures (Davis et al., 2011; Duewell et al., 2010; Halle et al., 2008; Hornung et al., 2008; Kanneganti et al., 2006; Sander et al., 2011), NLRC4 (sensing of Salmonella, intracellular flagellin and bacterial type III secretion rod proteins) (Franchi et al., 2006; Miao et al., 2010) and Naip5 (promoting resistance to Legionella) (Kofoed and Vance, 2011; Molofsky et al., 2006; Ren et al., 2006) Recent results also suggested a role for NLRP6 in maintenance of bacterial homeostasis in the colon and NLRP7 in the recognition of lipoproteins (Khare et al., 2012). NLRP12 (also called Nalp12, Monarch-1 and Pypaf-7) was the first NLR shown in biochemical assays to interact with the adaptor protein Asc to form an active IL-1β-maturing inflammasome (Wang et al., 2002). The role of NLRP12 in innate immunity has remained unclear. Both inflammatory and inhibitory functions have been suggested, as has a role in hypersensitivity (Allen et al., 2012; Arthur et al., 2010; Lich and Ting, 2007; Lich et al., 2007; Wang et al., 2002; Zaki et al., 2011) Interestingly, like for NLRP3, mutations in NLRP12 are linked to hereditary inflammatory disease (Jeru et al., 2008), and mutations may lead to increased Asc speckle formation and caspase-1 activity (Jeru et al., 2011b). It has been reported that patients carrying NLRP12 mutations associated with increased inflammasome activation have been successfully treated with anti-IL-1 therapy, similar to patients containing mutations in NLRP3 (Hawkins et al., 2003; Jeru et al., 2011a; Lachmann et al., 2009). No previous studies have addressed the role of NLRP12 in host resistance to infectious agents.
Evading innate immunity early in infection plays a key role in virulence of many microorganisms including the plague bacillus Yersinia pestis (Cornelis, 2000; Perry and Fetherston, 1997; Stenseth et al., 2008). This pathogen has several means of minimizing immune activation (Lathem et al., 2007; Monack et al., 1998; Mukherjee et al., 2006; Sodeinde et al., 1992; Zhou et al., 2005), with the effect that bacterial replication can proceed with minimal interference by the immune system. As a result, plague is often characterized by very high bacterial numbers in patient sera and organs (Perry and Fetherston, 1997). Major factors neutralizing host defenses by active means include a complex type III secretion system (T3SS) (Cornelis, 2002; Perry and Fetherston, 1997), the plasminogen activator Pla (Lathem et al., 2007; Sodeinde et al., 1992), and a high-affinity iron acquisition system (Perry and Fetherston, 1997). The Yersinia T3SS delivers effector proteins, which disrupt signaling within the host cell to prevent phagocytosis, induce apoptosis, and evade the immune response (Cornelis, 2002). Many Gram-negative bacteria, including Y. pseudotuberculosis, a very close ancestor of Y. pestis, produce a hexa-acylated lipid A andLPS which has the potential of strongly triggering innate immunity via Toll-like receptor 4 (TLR4)-MD-2 signaling (Munford, 2008; Raetz et al., 2007; Rebeil et al., 2004; Therisod et al., 2002). In contrast, Y. pestis generates a tetra-acylated lipid A-LPS that poorly induces TLR4-mediated cellular activation (Kawahara et al., 2002; Knirel et al., 2005; Montminy et al., 2006; Rebeil et al., 2006). We have reported that expression of E. coli lpxL in Y. pestis, which lacks a homologue of this gene, forces the biosynthesis of a hexa-acylated LPS (Montminy et al., 2006), and that this single modification dramatically reduces virulence in wild type mice, but not in mice lacking a functional TLR4. This emphasizes that avoiding activation of innate immunity is important for Y. pestis virulence. It also provides a model in which survival is strongly dependent on innate immune defenses, presenting a unique opportunity for evaluating relative importance of innate immunity signals in protection against bacterial infection.
One implication of TLR4 engagement is the induction of the immature forms of the central pro-inflammatory cytokines IL-1β and IL-18. TLR4 signaling can also promote expression of inflammasome components such as Nlrp3 (Bauernfeind et al., 2009). This establishes links between TLR4 activation and the inflammasome pathways. In this study we have used wild type Y. pestis and attenuated strains expressing a strong TLR4-activating hexa-acylated LPS as a model system to investigate the involvement of NLRP12 in pathogen recognition and IL-18 - IL-1β release.
Here we show that NLRP12 is an inflammasome component that is central in the recognition of Y. pestis and that IL-18 signaling substantially contributes to resistance against bacteria. Compared to wild type mice, NLRP12 deficient animals had higher mortality and increased bacterial loads following infection, correlated with lower amounts of IL-18, IL-1β and IFNγ. We propose a role for NLRP12 in the sensing of microbial pathogens.
We have found that all members of the genus Yersinia other than Y. pestis, and including the very closely related Y. pseudotuberculosis, contain the lpxL gene (S. Paquette et al, unpublished). Absence of lpxL and the resulting production of a tetra-acylated LPS was proposed to be essential for Y. pestis virulence (Montminy et al., 2006). To study the evasion of TLR4 signaling in an evolutionary perspective, we cloned lpxL from the closely related Y. pseudotuberculosis and expressed it in Y. pestis, generating Y. pestis-pYtbLpxL, to determine its effects on virulence. Y. pestis grown at 37°C has a tetra-acylated lipid A (Figure S1A) (Montminy et al., 2006) whereas Y. pseudotuberculosis and Y. pestis-pYtbLpxL have a hexa-acylated lipid A (Figure S1B). Mice infected s.c. with 500 CFU of highly virulent Y. pestis KIM1001 rapidly succumb to infection (Figure 1A). All wild-type mice infected with KIM1001-pYtbLpxL expressing a hexa-acylated Y. pseudotuberculosis-like lipid A survived (Figure 1A), and the animals were protected towards challenge with virulent KIM1001 (Table S1).
Survival of mice was strongly TLR4 dependent (Figure 1A). To determine the pathways responsible for in vivo clearance, mice from several strains deficient in inflammatory cytokines or cytokine receptors were infected s.c. with 500 colony forming units (CFU) of KIM1001-pYtbLpxL (Figure 1B). Interestingly, 100% of the animals lacking IL-18 and IL-18R died, as did the TLR4 deficient mice and 70% of the IL-1R1 deficient mice. Weaker effects were observed in animals lacking IFNαβR, TNFR1, or IL-12p40 (Figure 1B). Resistance to infection in IL-1β and IL-1R1 deficient animals was reduced to a similar degree, with approximately 30% of animals surviving (Figure 1C). However, IL-18 was critically important for resistance to infection in this model, as IL-18 and IL-18R deficient mice developed symptoms of bubonic plague, and rapidly succumbed to disease when infected with KIM1001-pYtbLpxL (Figures 1B and 1D). As inflammasomes are responsible for processing of IL-18 and IL-1β into mature forms, this result indicates that this infection model is well-suited for the study of inflammasome mechanisms and implications of IL-18 release. Mice deficient in MyD88, an adaptor molecule common to TLR, IL-1R and IL-18R signaling pathways were more susceptible to wild type Y. pestis KIM1001 than wild-type C57Bl/6 mice (Figure S1C) and are also highly susceptible to strains expressing lpxL (Montminy et al., 2006). Intravenous (i.v.) infection causes systemic infection even when attenuated bacterial strains are used, hence the inflammatory capacity in tissues for various bacterial strains can better be compared using this route of delivery. We found elevated levels of spleen IL-1β and IL-18 after i.v. infection with Y. pestis, and fully virulent KIM1001 induced lower cytokine levels as compared to KIM1001-pYtbLpxL producing the potent LPS (Figure 1E–F). A similar release pattern could also be seen in vitro using bone marrow derived macrophages (BMDM) (Figure 1G) after stimulation with KIM5 (a pgm mutant attenuated strain used for in vitro experiments) or KIM5-pYtbLpxL. Immunoblot analysis (Figure 1H) indicated that pro- IL-1β was indeed cleaved into mature IL-1β following infection with Y. pestis strains, a sign of inflammasome action. Infection with the Y. pestis-YtbLpxL strain markedly increased levels of pro- and cleaved IL-1β. These results indicate that minimizing inflammasome priming may have been an important implication of lpxL loss during evolution of Y. pestis from Y. pseudotuberculosis.
We next wanted to determine which NLRs were involved in resistance to Y. pestis strains and in IL-18 and IL-1β release. NLRP12 and NLRP3 have both been shown to interact with Asc in generating an IL-1β processing inflamammasome (Agostini et al., 2004; Manji et al., 2002; Wang et al., 2002), but little is known of the role of NLRP12 during infection. We infected both NLRP3-deficient and NLRP12-deficient mice (Figure 2A–B) s.c. with 500 CFU of KIM1001-pYtbLpxL, and found that only 20% of NLRP12-deficient mice survived the infection, whereas approximately 50% of mice lacking NLRP3 survived. This suggests that NLRP12 plays an important role in host defense against some bacterial pathogens. In contrast, NLRP12 deficient mice were resistant to infection with S. typhimurium, whereas TLR4-deficient mice all succumbed to the infectious challenge (Figure 2C). This indicates that NLRP12 deficient animals are not universally more sensitive to infections. The function of NLRP12 is not well understood, but mRNA is detectable in several organs and immune cells (Figure S2A–B), including macrophages, although prolonged macrophage maturation led to a decrease in Nlpr12 levels (Figure S2C). NLRP12-deficient mice (Figure S2D) had a normal composition of cell populations in spleen and bone marrow (Figure S2E). The possible involvement of NLRP12 in maturation of IL-1β and IL-18 led us to perform in vitro experiments with mouse cells to study inflammasome components that promote caspase-1 cleavage and IL-1β - IL-18 release following infection with Y. pestis and modified strains. Neutrophils express more Nlrp12 than macrophages (Fig S2B), but the role of inflammasomes in pathogen-induced neutrophil release of IL-1β and IL-18 is not yet studied in detail for many microbes. We found that thioglycollate-elicited neutrophil-enriched peritoneal cells released IL-1β after Y. pestis infection (Figure 3A). When compared to cells from wild-type mice, the amounts of IL-1β, but not TNF (Figure S3A) released from the neutrophils lacking NLRP12 were markedly reduced after stimulation with Y.pestis strains. Moreover, infected neutrophils from the caspase-1-deficient mice lack IL-1β in the supernatant, suggesting that Y. pestis-induced neutrophil IL-1β release involves caspase-1 inflammasomes, although we cannot rule out a role for other neutrophil proteases (Netea et al., 2010). It is also unclear which role caspase-11 plays relative to caspase-1 in Y. pestis-induced inflammasome activation, as the caspase-1 deficient mice utilized in this study contain the same truncated and apparently non-functional caspase-11 as previously published (Kayagaki et al., 2011 ). Macrophages deficient in NLRP12 or NLRP3 also had a reduced ability to release both IL-18 and IL-1β after infection with parental Y. pestis and Y.pestis-pYtbLpxL (Figure 3B–C). These observations are consistent with the survival data (Fig 2), which indicated that host recognition of Y. pestis involves NLRP12. Cells deficient in Asc and caspase-1 also had decreased IL-18 and IL-1β release (Figure 3B,C). Thus, NLRP12 signaling may occur parallel to or in cooperation with additional inflammasome components, as NLRP12 deficiency did not completely block cytokine release. NLRP12 KO macrophages responded normally to alum, S. typhimurium (Figure 3C), nigericin and poly(dA:dT) (Figure S3B–C)) suggesting that NLRP12 may not participate in NLRP3, AIM2 or NLRC4 inflammasomes formed in response to those stimuli. None of the inflammasome proteins had an impact on TNF release (Figure 3D, Figure S3). Furthermore, NLRP12 deficiency had little impact on the expression of 31 selected macrophage genes, including Il1b, in the absence or presence of bacteria (Figure S3C). Many of those genes are controlled by NF-kB and/or MAP kinases. In a more detailed study, NF-kB signaling measured by IKK kinase assay and I-kB degradation was also largely preserved in NLRP12 deficient cells (Figure S3D–E)._Y. pestis pre-grown at 26°C naturally expresses a hexa-acylated LPS (Montminy et al., 2006), and release of IL-1β in response to infection by 26°C-grown bacteria was also influenced by NLRP12 (Figure S3B)._Upon infection of wild-type and NLRP12 KO BMDM with the human pathogens Y. pseudotuberculosis and Y. enterocolitica, ancestors of Y. pestis (Chain et al., 2004), we observed a reduction in secreted IL-1β from the cells lacking NLRP12 (Figure 3E) while TNF release was normal (Figure S3F). By using KIM6, a derivative of KIM5 that lacks the pCD1 virulence plasmid containing genes for the T3SS (Perry and Fetherston, 1997), we found that the secretion system was necessary for stimulating IL-1β release, even in the presence of a highly stimulatory LPS as found in KIM6-pYtbLpxL (Figure 3F). YopJ may participate in inflammasome activation (Zheng et al., 2011) and the deletion of YopJ or the T3SS translocon protein YopB reduced IL-1β release (Figure 3G). Experiments performed using a strain with the expression of lpxL on a YopJ mutant background suggested that YopJ is a key player controlling IL-1β release, even in the presence of a stimulatory LPS (Figure 3H), although other T3SS-dependent factors may also regulate IL-1β (Brodsky et al., 2010). The data suggest that the ligand(s) responsible for NLRP12 activation are dependent on the Yersinia T3SS. TLR4 plays a critical role in the IL-1β and IL-18 production after infection of the mouse macrophages (Figure 3I, and Figure S3G), although the relative importance of mouse vs. human TLR4 - MD-2 in inducing Y. pestis responses may differ. Rodent cells have higher ability to recognize hypo-acylated lipid A (Lien et al., 2000; Montminy et al., 2006). This may be influenced by a shallow positioning of the hypoacylated lipid A in mouse MD-2 compared to human MD-2, and the enabling of enhanced ionic interactions between hypoacylated lipid A and mouse TLR4, facilitating receptor cluster dimerization and signaling (Meng et al., 2010). Our results indicate a role for both TLR4 and NLRP12 in the pro-inflammatory macrophage response against Y. pestis strains.
Upregulation of Nlrp3 has been suggested to positively affect the activity of the NLRP3 inflammasome (Bauernfeind et al., 2009). We therefore studied expression of Nlrp12 and Nlrp3 (Figure 4A–B) after infection of macrophages with KIM5 or KIM5-pYtbLpxL. Expression of Nlrp12 in BMDM was markedly increased after infection with Y. pestis strains and this may boost host responses to an infection. Treatment with LPS alone induced upregulation of Nlrp12 gene expression (Figure S4A). Furthermore, Y. pestis-induced formation of cleaved and active caspase-1, as measured by an assay showing binding of active caspase-1 to a fluorescent substrate, was also impaired in NLRP12-deficient cells, providing evidence for NLRP12-dependent inflammasome function (Figure 4C–D). Caspase-1 cleavage measured by this assay is also decreased in spleen macrophages or neutrophils from NLRP12 KO mice 24 hours after infection with KIM1001 or KIM1001-pYtbLpxL (Figure 4E–F). Il1b gene expression was similar in infected wild type cells and NLRP12 KO cells infected with Y. pestis (Figure 4G, Figure S3D). The macrophages infected in vitro showed a reduction in caspase-1 and IL-1β processing by immunoblot (Figure 4H), also cells infected at a higher MOI (Figure S4B). Thus, several lines of evidence support the hypothesis that NLRP12 is a component of inflammasomes formed after Y. pestis infection. Macrophage cell death induced by Y. pestis has been reported to be caspase-1 independent (Lilo et al., 2008). We confirmed those data (not shown), and in line with this observation, NLRP12 deficient cells did not show an altered cell death in response to Y. pestis infection (Figure S4C). Cell death may be induced by other mechanisms than pyroptosis in macrophages infected with Y. pestis.
As shown in Figures 1 and and2,2, NLRP12 KO and IL-18 KO mice are more susceptible than wild-type mice to infection with Y. pestis-pYtbLpxL. To monitor changes in IL-18 and IL-1β in tissues during systemic disease, we subjected WT and NLRP12 mice to i.v. infection with fully virulent or attenuated Y. pestis (KIM1001 or KIM1001-pYtbLpxL). At 44 hours post infection with KIM1001, IL-18 cytokine amounts were considerably lower in the NLRP12 KO mice, expressed as both cytokine normalized to the spleen bacterial load in each particular animal (Figure 5A) or simply as cytokine concentration in homogenate (Figure S5A). A decrease of IL-18 and IL-1β in the spleen (Figure 5B–C) and serum (Figure 5D–E) was also observed after KIM1001-pYtbLpxL infection in the NLRP12 KO mice as compared to wild-type mice. Experiments with IL-1R KO, IL-1β KO and IL-18R KO suggested that IL-18 signaling had the greatest impact on resistance to Y. pestis-pYtbLpxL, as 100% of IL-18 and IL-18R KO animals died after infection (Figures 1B and 1D). IL-18R KO mice had reduced IL-18 and IL-1β in the spleens compared to wt mice after infection with KIM1001-pYtbLpxL (Figure S5B), suggesting a positive feed-back loop via IL-18R for IL-1β and IL-18 production.
A reduction of several orders of magnitude in spleen bacterial load was seen when mice were infected i.v. with KIM1001-pYtbLpxL compared to wild-type KIM1001 (Figure 5E), indicating beneficial host responses induced by the presence of the hexa-acylated LPS. These differences in systemic bacterial load between the two bacterial strains were absent in mice lacking NLRP12 or IL-18R. NLRP12 deficient and IL-18R deficient mice also had increased bacterial loads compared to wild-type mice when infected with the virulent Y. pestis KIM1001 (Figure 5E, p=0.01, WT vs. Nlrp12−/−; p<0.001, wt vs. Il18r1−/−). This is important in that it shows that NLRP12 and IL-18R participate in host resistance in vivo towards both virulent and attenuated strains of Y. pestis. Thus, it appears that Y. pestis has an inherent ability to activate NLRP12-dependent recognition, and that the potent LPS found in strains expressing LpxL increases the formation of pro-forms and subsequently mature forms of inflammasome-controlled cytokines such as IL-1β (Figures 1H and and4C).4C). Livers from animals infected with wild type Y. pestis have large extracellular clusters of bacteria (Figure 5F, left panels, marked with *) and remarkably few signs of inflammation, likely reflecting active suppression of immunity combined with stealth via limited initiation of TLR4 signaling. Livers from animals infected with Y. pestis-pYtbLpxL display foci consisting of inflammatory cells (Figure 5F, upper right, arrows) and absence of visible bacterial masses, suggesting that recruitment of phagocytes limits bacterial growth (Montminy et al., 2006). Livers from NLRP12 KO infected with KIM1001-pYtbLpxL had recruitment of inflammatory cells (Figure 5F, arrows). Such masses of inflammatory cells typically contain large number of neutrophils and some mononuclear cells (Montminy et al., 2006), and a calculation of number of recruited cells showed no significant difference between infected wild type versus NLRP12 deficient livers (Figure S5C). However, this cell recruitment did not correlate with suppression of bacterial growth, as bacterial masses were visible (Figure 5F). These results suggest that NLRP12 may not play a major role in the attraction of phagocytes to infected sites in the liver; but is central to the effective anti-bacterial actions they perform. Few, if any inflammatory cells were visible in livers of IL-18R deficient mice (Figure 5F), indicating failures of both cell recruitment and anti-bacterial defenses.
Taken together, the results suggest that NLRP12 and IL-18 contribute to host resistance against Y. pestis and Y. pestis-pYtbLpxL. We also found that NLRP12 KO mice infected with KIM1001-pYtbLpxL had reduced amounts of TNF and the chemokine CXCL12 compared to C57Bl/6 (Figure S5D–F), possibly secondary effects of reduced IL-1β and IL-18 release, as primary cells lacking NLRP12 did not display decreased TNF release in culture (Figure 2). In contrast, NLRP12 deficient mice injected with an alum-LPS mixture did not show decreased serum IL-1β, IL-18, TNFα and CXCL12 (Figure S5G). Furthermore, we found similar recruitment of neutrophils to the peritoneum of wild type mice or NLRP12 deficient mice injected i.p. with sterile thioglycollate (Figure S5H). Movement of neutrophils (Figure 5F, Figure S5C) and DC (Figure S5I) during infection of NLRP12 deficient mice appears to be largely preserved. Differences in survival between NLRP12 deficient or IL-18 deficient mice and wild-type mice after s.c. infection with only 10 CFU of fully virulent KIM1001 were not significant (Figure 5G). This result is of uncertain importance because the very low LD50 of Y. pestis by s.c. infection (less than 10 CFU) makes it difficult to demonstrate reductions in host resistance impacting survival without the use of very large numbers of animals. Tissue bacterial loads (Figure 5E) appear to be more sensitive assays for analyzing host resistance to Y. pestis.
NLRP3 has also been proposed as an inflammasome component recognizing Y. pestis (Zheng et al., 2011) (Figure 2, Figure 3). NLRP3-deficient animals also were less resistant to infection by KIM1001-pYtbLpxL, in that they displayed increased bacterial loads in the spleen (Figure 5H) that correlated with reduced spleen cytokines (Figure 5I). To sum up, NLRP12 and NLRP3 both contribute to the host resistance towards Y. pestis strains.
IL-18 is a known inducer of IFNγ (Okamura et al., 1995), a key protein in many host responses to pathogens. This suggests that signaling via NLRP12 and the IL-18R, resulting in the release of IFNγ, could mediate resistance to Y. pestis-pYtbLpxL. Mice lacking both IFNαβR and IFNγR (dKO) were infected with KIM1001-pYtbLpxL s.c., and we found that all the dKO animals succumbed to the infection (Figure 6A). This phenomenon was largely attributed to IFNγR signaling, as only a few mice lacking IFNαβR died upon infection, whereas almost all mice lacking IFNγR succumbed (Figure 6B). No differences in IFNγ concentrations were observed between spleens of uninfected wt, NLRP12 deficient, and IL18R deficient mice (Figure 6C–D). However, the IFNγ concentrations in spleens from KIM1001-pYtbLpxL infected NLRP12 deficient mice compared to wild-type were drastically reduced (Figure 6D), as was also true for the mice lacking IL-18R (Figure 6D). Thus, we propose a cascade of signals from NLRP12 to IL-18 maturation that in turn mediates IFNγ release following infection with Y. pestis strains.
We propose that recognition of Y. pestis expressing a stimulatory LPS by TLR4 leads to upregulation of NLRP12 and pro-inflammatory cytokines such as IL-18 and IL-1β. NLRP12 then recognizes a ligand produced upon Y. pestis infection and assembles into an inflammasome that processes IL-18 and IL-1β. Although the precise nature of the true NLRP12 ligand is unknown, and may be a host or bacterial protein, the generation of the ligand appears to require the virulence-associated T3SS of Yersinia. Models for activation may include possibilities that cells sense membrane damage associated with the T3SS, secreted effectors or other molecules channeled by the T3SS, and modified host proteins. NLRP3 also contributes to IL-18 - IL-1β release. IL-18 seems to be more critical than IL-1β, and plays a key role in induction of IFNγ.
We show that NLRP12 is an inflammasome component recognizing Y. pestis and contributes to in vivo resistance to infection with Y. pestis strains. To our knowledge. this is the first demonstration of a clear role for NLRP12 in resistance to infection. Our data suggest an inflammasome role for NLRP12 in pathogen recognition, and that the NLRP12 - IL-18 - IFNγ axis is effective in limiting infection with Y. pestis-pYtbLpxL. We also show that the expression of Y. pseudotuberculosis LpxL in Y. pestis increases TLR4-dependent release of IL-18 and IL-1β. This increase correlates with increased resistance to the modified pathogen. In fact, the results indicate that a major consequence of producing LPS with low TLR4-activating potential could be lack of priming necessary for effective synthesis of active IL-1β and IL-18.Therefore, Y. pestis is able to utilize inflammasome-activating components like the T3SS to neutralize the immune response without an effective activation of an inflammatory response. This phenomenon may have played a role in evolution of high virulence in Y. pestis.
These findings support the view that inflammasomes, the cellular protein complexes cleaving IL-18 and IL-1β into mature forms, are fundamental components of the host response to many pathogens. Indeed, several viral, bacterial and fungal microbes have strongly increased ability to induce disease in the absence of IL-1β, IL-18 and inflammasome components (Broz et al., 2010; Davis et al., 2011; Hise et al., 2009; Lamkanfi and Dixit, 2009; Rathinam et al., 2010). In spite of this, only a few mammalian NLRs out of a family of more than 20 members have currently been shown to directly participate in host defenses. Here we show that NLRP12 participates in host responses to wild type Y. pestis and modified Y. pestis strains expressing a potent LPS, although the factor(s) in Y. pestis responsible for directly activating the NLRP12 inflammasome are still unknown.
NLRP12 may also be involved in resisting infections caused by other human pathogens. It is unclear how NLRP12 may interact with other inflammasome components. NLRP12 deficiency did not cause a complete reduction in ability to release IL-18 and IL-1β following exposure to Y. pestis and Y. pestis-pYtbLpxL infection . Also, the increased mortality observed in NLRP12-deficient mice did not appear as great as observed in IL-18 deficient animals , and NLRP3 also plays a role in host defenses. Redundancy between NLRs may occur, and other NLRs may also participate in optimal responses to infection. This may support the idea that NLRs work together for optimal protection of the host (Broz et al., 2010). The generation of animals with combined deficiencies in NLRP12 and other NLRs may clarify how NLRP12 functions in cooperation with other signaling components. NF-kB signaling following bacterial challenge appeared normal in NLRP12-deficient cells.
IL-18, IL-1β and IFNγ are all cytokines active at the interface between innate and adaptive immunity. We have found that Y. pestis strains generating a hexa-acylated LPS could function as effective live vaccines (Montminy et al., 2006). It would be of interest to investigate the role of NLRP12 in promoting the development of adaptive immunity and protection following vaccination with both live and subunit+adjuvant vaccines.
The emerging role of inflammasomes as key players in host defenses during many infections makes them desirable targets for therapeutic intervention and drug development. We note that alum, one of the first components known to activate specific inflammasomes, already is in widespread use as one of the few vaccine adjuvants licensed for human use. However, a delicate balance between pathological effects and enhanced host defenses arising from inflammasome-stimulating treatments will be necessary. Mutations in NLRs are linked to inflammatory diseases (Hawkins et al., 2003; Jeru et al., 2011a), and anti-IL-1 treatment does in fact reduce symptoms in many such patients. More knowledge on the role of NLRs in inflammation and homeostasis is needed in order to fine-tune future NLR-based therapies.
Y. pestis KIM is originally a clinical isolate from a Kurdistan Iran man (Brubaker, 1970; Perry and Fetherston, 1997) Y. pestis strains KIM5, KIM5-pEcLpxL (containing E. coli lpxL, earlier called pLpxL) and KIM1001 were as reported (Montminy et al., 2006). Y. pseudotuberculosis IP2666 (containing a complementation of PhoP/PhoQ deficiency) and Y. enterocolitica 8081 were provided by Joan Mecsas. Strains were grown in tryptose-beef extract (TB) broth with 2.5mM CaCl2 all by shaking at 37°C. lpxL of Y. pseudotuberculosis IP2666 including 480 basepairs upstream and 266 basepairs downstream from coding region was cloned using Pfu Ultra (Stratagene) and was ligated into the BamHI and SalI sites of pBR322, creating pSP::YtbLpxL (or ‘pYtbLpxL’). The resulting plasmid was electroproated into Y. pestis KIM5 (Goguen et al., 1984) or Y. pestis KIM1001 (Sodeinde et al., 1992) and bacteria were selected by growth on TB agar supplemented with 2.5mM CaCl2 in the presence of 100µg/ml of ampicillin. All strains containing plasmids above remained tetracycline sensitive. KIM1001 (pPCP1+, pCD1+, pMT1+) is highly virulent (Perry and Fetherston, 1997) whereas, KIM5 bears the chromosomal deletion ‘Δpgm’ which substantially attenuates virulence. The pgm locus contains no genes thought to affect LPS biosynthesis. KIM6 is a KIM5 derivative lacking the T3SS-containing pCD1 virulence plasmid . KIM5-ΔYopB was provided by Greg Plano (Torruellas et al., 2005). For the generation of KIM5-ΔYopJ, the following method was used. An in-frame deletion removing codons 4–287 was created via allelic exchange. PCR products made with primer sets A (ATAGAGCTCCACTACTGATTCAACTTGGACG), B (TCCGATCATTTATTTATCCTTATTCA) and C (TGAATAAG GATAAATAAATGATC G GATAATGTATTTTG GAAATCTTG CT), D (GGGTCTAGACTGATGTCGTTTATTTCTGGGTAT), respectfully, were used to make a fused product by overlap PCR using primers A and D (Horton et al., 1989). This product was cloned in the allelic exchange vector pRE107 (Edwards et al., 1998) in E.coli K12 strain B2155, transferred to Y. pestis by conjugation, and recombinants selected on TB medium containing 100µg/ml ampicillin but no diaminopimelic acid. Following counter selection with 5% sucrose, deletion mutants were identified by PCR. For in vitro infections, bacteria were grown overnight at 37oC in TB broth with or without ampicillin, diluted 1:4 in fresh media and cultured for three more hours at 37°C, then washed three times with PBS and resuspended in DMEM or RPMI. S. enterica serovar typhimurium strain SL1344 was provided by Mary O’Riordan, and strain M525P by Clare Bryant.
Mouse BMDM (bone marrow derived macrophages) were prepared by maturing fresh bone marrow cells for 5–7 days in the presence of M-CSF containing supernatant from L929 cells. Mouse neutrophils were enriched by injecting 1ml of thioglycolate i.p., peritoneal cells (typically >80% Ly6G-positive cells, (Nilsen et al., 2004) cells were harvested 4 hrs later after flushing with RPMI. Mouse BMDMs were plated at 2×105 per well in 96-well plates for ELISA or 2×106 per well in 12-well plates for immunoblot. Stimulation was for 6 hours and supernatants were collected for cytokine analysis. Three hours after bacterial infections, 50µg/ml of gentamycin was added. Alum was from Pierce, nigericin and poly(dA:dT) was from Sigma. IL-1β p17 and Caspase-1 p10 immunoblots were conducted mainly as described (Hornung et al., 2008) using antibodies from Santa Cruz Biotechnology (caspase-1 p10) and R&D (IL-1β). The antibody against β-actin was from Sigma. ]Q-PCR for Nlrp12 and Nlrp3 in resting or infected BMDM or magnetic bead (StemCell Technologies)-isolated neutrophils was performed by RNeasy Mini Kit (Qiagen), and iScript cDNA Synthesis Kit (BioRad). PCR was performed on transcribed cDNA or mouse tissue cDNA (Clontech) with primers for detection of mouse Nlrp12 (5’-TGCAAGCTTCGAGTCCTGT-3’, 5’-CCTGGTCGGCTTCATTCTG-3’), Nlrp3 (5’ AACCAATGCGAGATCCTGAC 3’, 5’ ATGCTGCTTCGACATCTCCT 3’), or Il1b (5’-GCCCATCCTCTGTGACTCAT-3’, 5’-AGGCCACAGGTATTTTGTCG-3’) using SYBR green (BioRad) according to the manufacturer’s instructions. ELISA kits for IL-1β, TNFα, IL-8, CXCL12, IFNγ (R&D) and IL-18 (MBL) were used for cytokine detection. Reagents for FACS detection of active and cleaved caspase-1 by FLICA-FITC substrate were from Immunochemistry Technologies.
All experiments involving animals were approved by the Institutional Animal Care and Use Committee. ASC (Pycard−/−), NLRP3 (Nlrp3−/−) and NLRP12-deficient (Nlrp12−/−) mice were generated by Millennium Pharmaceuticals and were backcrossed eight to eleven generations to C57BL/6 background. Mice deficient in TLR4 (TLR4−/−) and MyD88 (Myd88−/−) were from S. Akira, and mice lacking caspase-1 (Casp1−/−) were from Michael Starnbach. C57BL/6 and mice deficient in IL-1R1 (Il1r1−/−), IL-18R (II18r1−/−), IL-18 (II18−/−), TNFR1 (Tnfr1−/−), IL-12p40 (II12b−/−), and IFNγR (Ifngr1−/−) were all from Jackson Laboratories. J. Sprent (The Scripps Research Institute) provided the IFNαβR1 (Ifnar1−/−) and IFNγR1 x IFNαβR1 doubly deficient mice. IL-1β (II1b−/−) deficient mice (Horai et al., 1998) were provided by Y. Iwakura. Wild-type (from Jackson Laboratories or bred at UMass) or knock-out mice were infected s.c. in the nape of the neck with Y. pestis and their survival monitored twice a day for 30 days. Infection with S. typhimurium: mice were infected with 1000 CFU of M525P i.p. and survival was monitored as described above. For cytokine and CFU analysis, mice were infected either s.c. or i.v. and sacrificed at the indicated time points. Serum was generated by centrifugation in microtainer tubes (BD), and spleens were homogenized in 0.5 ml PBS using a closed system Miltenyi gentleMACS dissociator and c-tubes to preserve intact cells, subsequently cells/debris were removed by centrifugation. Samples for cytokine analysis were subjected to protease inhibitor (Roche) treatment. Cytokine amounts normalized by bacterial loads were calculated by dividing IL-18 concentrations (ng/ml) by the bacterial load (CFU × 108) for each animal. Hematoxylin and eosin (H&E) staining and microscopy were performed as published (Montminy et al., 2006). In vivo caspase-1 cleavage analysis: mice were infected with 500 CFU of Y. pestis i.p. After 24 hrs, spleens were harvested, homogenized and cell suspensions were stained with caspase-1 FLICA reagent.
In vitro cytokine release was analyzed by two-way ANOVA with Bonferroni post-test. Differences in spleen and serum cytokine concentrations were analyzed by the unpaired t-test. Differences in survival were studied using Kaplan-Meyer analysis and the logrank test. Differences in spleen CFU or cytokine/CFU ratio values between genotypes of mice were evaluated with the Mann-Whitney test, or in more complex comparisons involving multiple mouse genotypes, with a generalized linear regression model of cubic transformed log CFU values (95% confidence interval) to meet normality assumptions. P values <0.05 were considered significant.
We thank A. Cerny, M. Whalen, A. Zacharia H. Ducharme and C. Raskett for animal husbandry, and X. He and members of the Lien, Fitzgerald and Goguen labs for help and discussions. Work was supported by the NIH (grants AI057588-American Recovery and Reinvestment Act and AI075318 to EL, AI64349 and AI083713 to KF, AI095213 to GIV, NERCE fellowships AI057159 to SKV and VAKR), the Research Council of Norway and the Norwegian Cancer Society. The study also utilized core services supported by DERC grant NIH DK32520. We thank those who provided reagents, B. Monks for help with cloning and R. Ingalls for critical reading of the manuscript.