NLRP1-Dependent Leads to Acute Lung Injury and Morbidity in Mice Martina Kovarova, Pamela R. Hesker, Leigh Jania, MyTrang Nguyen, John N. Snouwaert, Zhidan Xiang, Stephen E. This information is current as Lommatzsch, Max T. Huang, Jenny P.-Y. Ting and Beverly of September 29, 2021. H. Koller J Immunol published online 29 June 2012 http://www.jimmunol.org/content/early/2012/06/29/jimmun

ol.1201065 Downloaded from

Why The JI? Submit online. http://www.jimmunol.org/ • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 29, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2012 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 29, 2012, doi:10.4049/jimmunol.1201065 The Journal of Immunology

NLRP1-Dependent Pyroptosis Leads to Acute Lung Injury and Morbidity in Mice

Martina Kovarova,*,1 Pamela R. Hesker,†,‡,1 Leigh Jania,†,1 MyTrang Nguyen,† John N. Snouwaert,† Zhidan Xiang,† Stephen E. Lommatzsch,† Max T. Huang,x Jenny P.-Y. Ting,x,{ and Beverly H. Koller*,†

Acute inflammation in response to both exogenous and endogenous danger signals can lead to the assembly of cytoplasmic inflam- masomes that stimulate the activation of -1. Subsequently, caspase-1 facilitates the maturation and release of cytokines and also, under some circumstances, the induction of death by pyroptosis. Using a mouse line lacking expression of NLRP1, we show that assembly of this inflammasome in cells is triggered by a toxin from and that it initiates caspase-1 activation and release of IL-1b. Furthermore, NLRP1 inflammasome activation also leads to cell death, which escalates over 3 d following exposure to the toxin and culminates in acute lung injury and death of the mice. We show that these events are not dependent on production of Downloaded from IL-1b by the inflammasome but are dependent on caspase-1 expression. In contrast, muramyl dipeptide-mediated inflammasome formation is not dependent on NLRP1 but NLRP3. Taken together, our findings show that assembly of the NLRP1 inflammasome is sufficient to initiate pyroptosis, which subsequently leads to a self-amplifying cascade of cell injury within the lung from which the lung cannot recover, eventually resulting in catastrophic consequences for the organism. The Journal of Immunology, 2012, 189: 000–000. http://www.jimmunol.org/

LRP1 was first characterized as a member of the CED-4 component, muramyl dipeptide (MDP), and ASC was not required, family of apoptotic , which are required to ini- though it enhanced caspase-1 activity (6). The protease caspase-1 N tiate in developing Caeno- mediates the production of mature IL-1b and IL-18 cytokines, rhabditis elegans (1–3). Consistent with this relationship, two which stimulate a proinflammatory immune response and protect independent groups showed that overexpression of NLRP1 in the organism from pathogen-induced death. However, a number of mammalian cells leads to cell death through (1, 2). lines of evidence indicate that robust activation of caspase-1 leads to Functional homology with CED-4 was further demonstrated by induced cell death under some circumstances. Cell death that is in vitro studies with macrophages that reported that NLRP1 is inflammation associated and caspase-1 dependent has been termed by guest on September 29, 2021 regulated by the antiapoptotic proteins Bcl-2 and Bcl-XL. Bcl-2 is pyroptosis (7). The downstream effects of pyroptosis are largely homologous to CED-9, an antiapoptotic that binds CED-4 unknown. Recently, pyroptosis was identified as a mechanism to (4). clear intracellular . Following Salmonella typhimurium A possible mechanism by which NLRP1 could mediate cell infection, caspase-1–dependent pyroptosis resulted in release death has been proposed based on its detection in a protein complex of S. typhimurium from infected macrophages, and the bacteria named the inflammasome, which activates inflammatory were subsequently killed by (8). (5). Thus, the first inflammasome complex to be described con- NLRP1 is now recognized as one of a number of proteins that can tained the proteins caspase-1,caspase-5,ASC,andNLRP1.In form complexes capable of supporting the maturation of caspase-1 later work using recombinant proteins, caspase-1 activation was in response to unique, but overlapping, danger and pathogen initiated by the NLRP1 protein alone in response to a microbial signals. To date, these include NLRP1, NLRP3, NLRC4, NLRP6, NLRP7, and absent in melanoma 2 (5, 9–16), although studies of *Pulmonary Division, Department of Medicine, University of North Carolina, Chapel NLRC5 and NLRP7 were performed in human cells with Hill, NC 27599; †Department of Genetics, University of North Carolina, Chapel Hill, knockdown strategies and NLRP1 by assembly in a cell-free ‡ NC 27599; Curriculum of Genetics and Molecular Biology, University of North system. Similar to many of the NLRs capable of forming an Carolina, Chapel Hill, NC 27599; xDepartment of Microbiology and Immunology, University of North Carolina, Chapel Hill, NC 27599; and {Lineberger Comprehen- inflammasome, NLRP1 contains a that can bind sive Cancer Center, University of North Carolina, Chapel Hill, NC 27599 ASC, a nucleotide binding domain (or NATCH) responsible for 1M.K., P.R.H., and L.J. contributed equally to the work. intramolecular folding, and nucleotide triphosphate-dependent Received for publication April 11, 2012. Accepted for publication June 14, 2012. oligomerization, a function to find domain, a CARD domain that This work was supported by National Institutes of Health Grants U19 A1077437 (to binds procaspase-1, and a leucine-rich repeat domain. The leucine- B.H.K. and J.P.-Y.T.) and U54-AI057157 (to J.P.-Y.T.). rich repeat domain is believed to confer the sensitivity of a given Address correspondence and reprint requests to Dr. Beverly H. Koller, University of inflammasome to a particular pathogen or alteration in cellular North Carolina at Chapel Hill, 120 Mason Farm Road, CB 7264, Chapel Hill, NC homeostasis, although direct evidence for this is not available. 27599. E-mail address: [email protected] Although the list of agents that stimulate NLR/absent in mela- Abbreviations used in this article: 7-AAD, 7-aminoactinomycin D; ARDS, acute respiratory distress; BAL, bronchoalveolar lavage; BMDM, bone marrow-derived noma 2-dependent caspase-1 activation continues to expand, the macrophage; HMGB1, high mobility group box 1; i.t., intratracheal(ly); LDH, lactate intracellular trigger for protein oligomerization remains unclear. dehydrogenase; LT, lethal toxin; MDP, muramyl dipeptide; MPO, myeloperoxidase; To date, an in vitro protein reconstitution assay showing forma- PGN, peptidoglycan; TiO , titanium dioxide. 2 tion of the NLRP1 inflammasome complex in response to MDP Copyright Ó 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00 treatment is the best display of stimulus-induced complex for-

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1201065 2 NLRP1-DEPENDENT PYROPTOSIS CAUSES LUNG INJURY mation (6). Much of the information regarding the activation of mann–La Roche, Basel, Switzerland). Peritoneal macrophages were eli- 2/2 2/2 2/2 these human inflammasomes has been garnered from experiments cited from wild-type, Nlrp1b , Nlrp3 , and Casp-1 mice with an carried out in mice lacking the orthologous NLR gene; however, i.p. injection of 4 ml 4% thioglycollate in PBS. Seventy-two hours after injection, the peritoneal cavity was lavaged with DMEM, and cells were this has not been achieved for NLRP1. plated and incubated for 1 h. Adherent cells were then washed twice with In mice, three orthologs of the human NLRP1 gene have been cold PBS and cultured in DMEM containing 10% FBS, 2 mM L-glutamine, identified, termed Nlrp1a, Nlrp1b, and Nlrp1c (17). Using genetic 150 U/ml penicillin, 150 mg/ml streptomycin, and 57.2 mM 2-ME (all from studies, Boyden and Dietrich (17) showed that the encoding Life Technologies). For LT treatments, macrophages were primed over- night with 100 ng/ml ultrapure LPS (Invitrogen Life Technologies), the Nlrp1 confers the sensitivity of some mouse lines to the washed with PBS, and incubated with OPTIMEM plus Glutamax con- lethal toxin of . Anthrax causes rapid tissue taining either PBS or 1 mg/ml LT for 6 h. Cytotoxicity was determined by necrosis, circulatory collapse, and acute respiratory failure be- measuring lactate dehydrogenase (LDH) release in the cell supernatant by cause of fluid retention in the lung (18). Many of these pathologies colorimetric assay following the manufacturer’s protocol (BD Clontech, can be recapitulated in animal models by an exposure to lethal Mountain View, CA). For MDP, peptidoglycan (PGN), titanium dioxide (TiO2), and L18–MDP treatments, cells were primed with 100 ng/ml ul- toxin (LT), which is a combination of lethal factor (LF) and trapure LPS for 6 h, washed with PBS, and incubated in OPTIMEM, 5 mg/ protective Ag (19), indicating that intoxication is mediated, at ml TiO2, 5 mg/ml TiO2 plus 10 mg/ml MDP, 5 mg/ml TiO2 plus 20 mg/ml least in part, by these factors. The mouse haplotype that confers PGN, 10 mg/ml MDP, 20 mg/ml PGN, or 1 mg/ml L18–MDP for 16 h. sensitivity to LT also carries in the Nlrp1a and Nlrp1c MDP treatment of mice genes that render them unlikely to direct the expression of a m m functional protein (17). This suggests that it is the Nlrp1b allele Mice received an i.p. injection of 5 g LPS in 200 l PBS. Three hours after LPS administration, mice were i.p. injected with 50 mg L18–MDP in carried by some mouse strains that initiates the response to LT. 200 ml PBS. Serum and peritoneal lavage fluid was collected 1.5 h post- Downloaded from In this paper, we report the generation of a mouse line lacking injection with L18–MDP. Nlrp1b expression. Using cells from this line, we show that LT, LPS/LT and LPS/ATP treatments of mice but not MDP, can trigger the assembly of the murine NLRP1 inflammasome, leading not only to caspase-1 activation and IL-1b 129+/+ and Nlrp1b2/2 mice were treated with 0.3 mg/kg body weight LPS production, but also to pyroptosis. These events are sufficient to intratracheally (i.t.). Two hours later, mice were treated with vehicle, 20 mg LT, or 50 ml 100 mM ATP i.t. Two hours after the second treatment, mice result in acute lung injury that escalates and leads to morbidity http://www.jimmunol.org/ were sacrificed, and their lungs were lavaged with HBSS. within 3 d of activation of the NLRP1 inflammasome. LT treatment of mice Materials and Methods Nlrp1b2/2, Il1b2/2, Casp-12/2, Nlrp32/2, and appropriate 129+/+ control Mice mice were dosed i.t. with vehicle or 20 mg LT and bronchoalveolar lavage (BAL) fluid was collected at various time points. The total number of cells A vector was designed to inactivate the mouse Nlrp1b locus by removing in the BAL fluid was determined using a hemocytometer. a 16-kb segment DNA extending from the middle of exon 3, which en- codes the ATP-binding NACHT domain, to the promoter region of the gene Protein analyses and replacing it with a neomycin resistance marker. An XbaI restriction r b a digest site was added at the 39 end of the Neo cassette allowing detection IL-1 , IL-1 , KC, MIP-2, and CXCL15/lungkine levels were detected by by guest on September 29, 2021 of the mutant locus by Southern blot analysis. 129 ES cells carrying the ELISA as recommended by the manufacturer (eBioscience, San Diego, correctly modified Nlrp1b allele were mated with 129S6 mice (Taconic CA, and R&D Systems, Minneapolis, MN). Albumin levels were evaluated Farms, Hudson, NY) to maintain the on this inbred genetic by colorimetric assay according to the manufacturer (Bethyl Laboratories, background. Montgomery, TX). BAL cell samples were processed to isolate myelo- The Nlrp32/2 mouse line was generated by targeted mutagenesis in 129 peroxidase (MPO) enzyme, and its catalytic activity was evaluated by ES cells using a replacement type vector, which removed all Nlrp3 exons colorimetric assay as a measure of recruitment into the lungs. and replaced this segment of DNA with the neomycin resistance gene. LDH release was measured by colorimetric assay following the manu- The Casp12/2 mouse line was generated in 129 ES cells using a re- facturer’s protocol (BD Clontech). Results were compared with the total placement type targeting construct. Recombination with the endogenous LDH released from cells treated with 1% Triton-X (100% lysis control). Casp1/4 locus removed the entire coding regions of both genes as well as For Western blot analysis, protein from BAL fluid was concentrated using 8kbDNA59 of Casp1 and 6 kb 39 Casp4. Chimeras generated from the a Microcon YM-30 column (Millipore, Billerica, MA) and subsequently targeted ES cells were bred directly to 129S6 mice to maintain the mu- quantified by bicinchoninic acid assay (Thermo Scientific Pierce, Rock- tation on this genetic background. ford, IL). Twenty micrograms of total protein was separated by SDS- The P2X72/2 mouse line was moved to the BALB/cByJ background by PAGE, transferred onto a polyvinylidene difluoride membrane, and incu- b six consecutive crosses to BALB mice (20). bated with IL-1 (AF-401-NA; R&D Systems), caspase-1 (sc-514; Santa All experiments were conducted in accordance with the National Cruz Biotechnology, Santa Cruz, CA), high mobility group box 1 (HMGB1) Institutes of Health Guide for the Care and Use of Laboratory Animals as (ab12029; Abcam, Cambridge, MA), or GAPDH (no. 3683; Cell Signaling well as the Institutional Animal Care and Use Committee guidelines of the Technology, Danvers, MA) Abs. University of North Carolina at Chapel Hill. Histology RT-PCR Lungs were inflated with 10% formalin via a tracheal cannula, removed 2 2 RNA was isolated from intestine of wild-type 129 and Nlrp1b / mice with from the thoracic cavity, and fixed overnight in formalin. Sections of the left RNABee (Tel-Test, Friendswood, TX). RNA was reverse transcribed with lobe of the lung were stained with H&E. Digital images of the sections were Superscript III (Invitrogen Life Technologies, Grand Island, NY), followed captured under bright-field mode using Olympus BX61 Upright Fluores- by PCR using primer sets specific for Nlrp1a, Nlrp1b, and Nlrp1c. Am- cence Microscope (Olympus, Center Valley, PA) and Velocity software plified products were visualized by gel electrophoresis. (PerkinElmer Life and Analytical Sciences, Waltham, MA). Whole lung sections and at least four images of representative areas of each section Cell culture were captured. For 32 images, a background image was digitally sub- Bone marrow-derived macrophages (BMDM) were prepared by harvesting tracted using ImageJ software (National Institutes of Health, Bethesda, marrow from femurs and tibias of 129+/+ and Nlrp1b2/2 mice. Marrow was MD). cultured for 7 d in IMDM containing 15% L cell-conditioned medium, FACS analysis 10% FBS, 2 mM L-glutamine, 150 U/ml penicillin, 150 mg/ml strepto- mycin, and 57.2 mM 2-ME (all from Life Technologies). BMDM were BAL was collected 2 h after i.t. instillation of 20 mg LT or vehicle. Cells incubated in OPTIMEM plus Glutamax (Life Technologies) containing were stained with fluorescently labeled caspase inhibitors FAM-YVAD- varying concentrations of LT (List Biological Laboratories, Campbell, CA) FMK (caspase-1) or FAM-DVED-FMK (caspase-3) and with 7-amino- for 4 h, and cell viability was assessed with the WST-1 reagent (F. Hoff- actinomycin D (7-AAD; BD Biosciences, Franklin Lakes, NJ), according The Journal of Immunology 3 to the manufacturer’s instructions (Cell Technology, Mountain View, CA). chimeras generated from the modified ES cells directly to 129S6 For whole lung cell suspension, isolated lungs were cut into pieces and mice. The impact of the Nlrp1b mutation was, therefore, studied in incubated for 1 h in digestion medium (RPMI 1640 medium, 12 mg/ml this paper in coisogenic cohorts of animals. No differences were collagenase IV, and 10 mg/ml DNase I) at 37˚C. A single-cell suspension 2/2 was prepared by passing through a 70-mm cell strainer (BD Biosciences). noted in the growth or the behavior or fertility of the Nlrp1b RBCs were removed by incubation in lysis buffer (155 mM NH4Cl, 10 mM mouse line, and expression analysis indicated that the mutation KHCO3, and 0.1 mM Na2EDTA). Cell were incubated with Fc block and generated a null allele. Both gross and histological analysis re- stained for CD45, Mac-I, and GR-1 or CCR-2 (all from BD Biosciences). vealed no lesions and no changes in the development. RNA anal- FACS analysis was done using a Beckman Coulter CyAn ADP Analyzer (Beckman Coulter, Brea, CA). Data were analyzed using FloJo (Tree Star, ysis confirmed the absence of Nlrp1a-c transcripts (Fig. 1C). Ashland, OR) software. Nlrp1b2/2 macrophages are protected from LT-induced cell Statistical analyses death Data are represented as mean 6 SEM. A Student t test, one-way ANOVA Genetic studies indicate that expression of a functional Nlrp1b with Tukey’s posttest, or two-way ANOVA with Bonferroni posttest was allele confers sensitivity of 129S6 cells to anthrax LT (17). If this used to determine the statistical probability of differences between ex- is indeed the case, and if the targeting event created a null allele as , , , # , perimental cohorts: *p 0.05, **p 0.01, ***p 0.001, p 0.0001. expected based on RNA analysis of tissue from the Nlrp1b2/2 line, then macrophages from mice homozygous for the mutant Results allele should be resistant to LT. To test this, we first cultured Generation of mice lacking Nlrp1b macrophages from the bone marrow (BMDM) of 129S6 (129+/+) 2/2 The mouse Nlrp1 locus contains three closely related genes des- and Nlrp1b animals and examined their viability in the pres- Downloaded from ignated Nlrp1b, Nlrp1c, and Nlrp1a. To generate a mouse lacking ence of increasing amounts of LT. A decrease in the viability of NLRP1 function, we took advantage of the fact that, in the 129 129+/+ macrophages was noted even upon exposure to very low mouse, Nlrp1a and Nlrp1c do not encode functional proteins. levels (1 ng/ml) of LT. At doses . 10 ng/ml, virtually all cells Therefore, inactivation of the functional Nlrp1b gene is expected to succumbed to the toxin. In contrast, no decrease in viability of the 2 2 yield an Nlrp12/2 animal. A vector was designed that, upon ho- Nlrp1b / cells was observed following exposure to increasing

mologous recombination at the Nlrp1b locus, results in the re- doses of toxin (Fig. 2A). http://www.jimmunol.org/ placement of the 59 portion of exon 3 as well as exons 1 and 2 and b a the promoter region of the gene with a neomycin resistance marker. LT-mediated cell death and IL-1 and IL-1 release are Using Southern blot analysis, we confirmed the generation and caspase-1 dependent but NLRP3 independent transmission of the mutant Nlrp1b allele (Fig. 1B). The mutant We next determined whether the LT-mediated cell death coincided allele was maintained on the 129 genetic background by breeding with the release of IL-1b and IL-1a. Thioglycolate-elicited peri- by guest on September 29, 2021

FIGURE 1. Generation of a Nlrp1-deficient mouse line. (A) A schematic of the endogenous Nlrp1b locus, the targeting vector designed to disrupt the gene, and the expected structure of the gene after homologous recombination. (B) Southern blot analysis of DNA from ES cells (upper panel) and from tail biopsies of pups generated from the intercross of mice carrying the modified Nlrp1b locus (lower panel) verified correct recombination of the plasmid with the endogenous locus and transmission of this allele through the germline. The expected 9.5-kb XbaI fragment is observed in wild-type ES cells and in mice carrying a wild-type Nlrplb allele. Ho- mologous recombination introduces a novel XbaI site reducing the size of the DNA fragment that binds to the probe to 7.7 kb. (C) RNA was prepared from the jejunum of 129 and Nlrp1b2/2 mice, and expression of all three Nlrp1 genes was examined by PCR using primers specific for each gene. As expected, 129S6 mice express only Nlrp1b, and this expression was not ob- served on analysis of RNA prepared from mice homozygous for the mutant allele. 4 NLRP1-DEPENDENT PYROPTOSIS CAUSES LUNG INJURY

therefore examined the possibility that, although NLRP1b was re- quired for initiating the response to LT, subsequent recruitment of NLRP3 was required for caspase-1 activation and cell death. To test this, an Nlrp3-deficient, 129S6 coisogenic mouse line was generated by homologous recombination. Nlrp32/2 thioglycolate- elicited macrophages were collected and treated with LPS fol- lowed by LT. No significant difference in the survival of LT- treated 129+/+ and Nlrp32/2 cells was observed (Fig. 2B). Pro- duction and release of IL-1b and IL-1a by LT-treated cells were detected by ELISA in supernatants collected from 129+/+, Nlrp1b2/2, Nlrp32/2,andCasp12/2 macrophages. Although levels of IL-1b and IL-1a in the supernatant collected from the LT-treated Nlrp32/2 cells did not significantly differ from LT- treated 129+/+ cells, in Casp12/2 and Nlrp1b2/2 cells, the re- lease of IL-1b and IL-1a was attenuated to level similar to PBS- treated controls (Fig. 2C, 2D). NLRP3, not NLRP1b, is essential for MDP and PGN-mediated b

IL-1 release Downloaded from In vitro assembly of the human NLRP1 inflammasome and acti- vation of caspase-1 can be initiated by exposure of inflammasome components to bacterial PGNs (6). We therefore asked whether NLRP1b was necessary for inflammasome assembly and pro- cessing of IL-1b by mouse macrophages in response to PGN or its

degradation product MDP. Peritoneal macrophages were collected http://www.jimmunol.org/ from 129+/+, Nlrp1b2/2, Nlrp32/2,andCasp12/2 mice. Cells were treated with MDP and PGN alone or in combination with TiO2.TiO2 has been reported to bind to PGNs, increasing their uptake by macrophages (21). Consistent with this report, the re- lease of IL-1b in response to PGN and MDP in the absence of TiO2 was below the level of detection. The addition of TiO2 to- gether with either PGN or MDP resulted in a robust release of IL- 1b that was approximately twice that observed with TiO2 alone. However, this response was unaltered in the Nlrp1b2/2 cells. In- by guest on September 29, 2021 stead, the response to TiO2 alone or TiO2 in combination with MDP or PGN was dependent on NLRP3 expression. As expected, the maturation and release of IL-1b was not observed in Casp12/2 cells (Fig. 3A). To further verify the normal response of Nlrp1b2/2 FIGURE 2. Lethal toxin mediated cell death and IL-1b and IL-1a cy- macrophages to MDP, we examined the response of 129+/+, +/+ 2/2 tokine release from macrophages. (A) BMDMs from 129 and Nlrp1b Nlpr1b2/2, Nlrp32/2, and Casp12/2 macrophages to L18–MDP. mice were incubated with LT for 4 h and cell viability was determined with L18–MDP is a lipidated form of MDP that improves uptake of the the WST-1 reagent. (B–D) Thioglycolate-elicited peritoneal macrophages molecule. Again, the response of the peritoneal macrophages to were primed with 100 ng/ml ultrapure LPS overnight and subsequently m MDP was dependent on NLRP3 but not NLRP1b (Fig. 3B). incubated for 6 h in presence of 1 g/ml LT. Cell death was determined by +/+ 2/2 LDH activity in supernatant and expressed as the percentage of total cel- In vivo studies on 129 and Nlrp1b mice exposed to L18- lular LDH activity (B). IL-1b (C) and IL-1a (D) release to supernatant was MDP also failed to support a role for NLRP1b in the response to detected by ELISA. n = 4 for 129+/+, Nlrp1b2/2, Nlrp32/2; n = 3 for Casp- MDP (Fig. 3C, 3D). 12/2.*p , 0.05, **p , 0.01, ***p , 0.001. In vivo LT-mediated processing and release of IL-1b is Nlrp1b dependent toneal macrophages were collected and treated for 12 h with LPS We next verified that NLRP1b could mediate IL-1b processing to induce expression of cytokines prior to exposure to LT. Six in vivo. To test this, pro–IL-1b expression in the lung was induced hours after the addition of LT, the cell supernatant was collected. by i.t. instillation of LPS at 0 h and followed by LT (LPS/LT) or LT-mediated cell death was assessed by measuring the levels of ATP (LPS/ATP) treatment at 2 h to induce IL-1b processing. BAL the cytoplasmic enzyme, LDH, released into the medium. IL-1b fluid was collected at 4 h, and levels of mature IL-1b were de- and IL-1a release were determined by ELISA. LDH levels in the termined by ELISA. Although no IL-1b was observed in the BAL supernatant collected from 129+/+ cultures corresponded to 75% of fluid of the vehicle-treated animals, treatment with LPS/ATP total cellular LDH content. Only low levels of LDH (0.8%) were resulted in high levels of IL-1b in the lavage fluid of 129+/+ and detected in the supernatant from Nlrp1b2/2 macrophages. Casp- Nlrp1b2/2 animals. LPS/LT treatment resulted in release of IL-1b 12/2 macrophages also released small amounts of LDH compared in 129+/+ mice but not in mice lacking NLRP1b (Fig. 4A). Sim- with 129+/+ cells (9.7%; Fig. 2B). Although the levels of LDH ilarly to IL-1b, the release of IL-1a following LPS/LT treatment released by the Casp-12/2 cultures were higher than those ob- was dependent on NLRP1b (Fig. 4B). To verify that the IL-1b served in the Nlrp1b2/2 cultures, these results indicated that cell present in the BAL fluid represented processed cytokine, proteins death is largely NLRP1b and caspase-1 dependent. NLRP1b has were analyzed by Western blot. Processed IL-1b as a p17 frag- been reported to collaborate with other NLR proteins (21). We ment was detected in the BAL fluid from 129+/+ mice treated with The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/

FIGURE 3. NLRP1b is not required for IL-1b release induced by MDP or PGN. (A) Thioglycolate-elicited peritoneal macrophages were primed with 100 ng/ml ultrapure LPS for 6 h and incubated in OPTIMEM with indicated concentration of TiO2, MDP, PGN, or combinations of those stimuli for 16 h. IL-1b from supernatants was detected by ELISA; n =4.

(B) Thioglycolate-elicited peritoneal macrophages were primed as in (A) by guest on September 29, 2021 follow by treatment with 1 mg/ml L18-MDP, 5 mg/ml TiO2, or 5 mg/ml TiO2 and 10 mg/ml MDP for 16 h. IL-1b was detected by ELISA in supernatants; n =4(C, D) 129+/+ (n) and Nlrp1b2/2 (N) mice were injected i.p. with 5 mg LPS. Three hours later, mice were i.p. injected with 50 mg FIGURE 4. Nlrp1b is required for LT-mediated release of IL-1b and IL- +/+ 2/2 L18–MDP. Serum and peritoneal lavage fluid (PLF) was collected 1.5 h 1a in the lung. 129 (n) and Nlrp1b (N) mice were treated with 0.3 postinjection with L18–MDP. The total number of inflammatory cells in mg/kg LPS or PBS i.t. at 0 h, and with PBS, 1 mg LT, or 50 ml 100 mM peritoneal lavage fluid (60% neutrophils) (C) and IL-1b in the serum, and ATP i.t. at 2 h. BAL fluid was collected at 4 h after the initial treatment. IL- PLF measured by ELISA (D) were determined; n = 13 for 129+/+, 14 for 1b (A) and IL-1a (B) in the BAL determined by ELISA. (C) Processed IL- Nlrp1b2/2, #p , 0.0001. 1b and caspase-1 proteins from concentrated BAL fluid analyzed by Western blot, 20 mg protein per line, two representative results from five independent experiments shown.***p , 0.001, #p , 0.0001, n = 1 for LPS/LT, but not in BAL collected from Nlrp1b2/2 animals (Fig. PBS, n = 10 for LPS/LT, n = 4 for LPS/ATP. 4C). Consistent with this, amount of processed caspase-1 in the 2 2 BAL fluid from the Nlrp1b / animals treated with LT was sub- (Fig. 5F). The failure to observe a progressive increase in cell stantially reduced compared with that observed in the 129+/+ LT- numbers as the health of the mice declined likely reflects a balance treated animals (Fig. 4C). between cell death and cell recruitment. Consistent with this in- terpretation, the BAL fluid collected from animals with more Mice lacking NLRP1 are protected from LT-mediated acute advanced disease contained increasing levels of fragmented cells lung injury that were difficult to categorize based on morphological criteria We next asked whether, in addition to altering the processing and (data not shown). Furthermore, LDH levels in the BAL fluid release of IL-1b, the absence of NLRP1b would protect mice continued to increase throughout the experiment, whereas the from the development of LT-induced acute lung injury. 129+/+ number of intact cells present in the BAL fluid remained largely and Nlrp1b2/2 mice were instilled with LT, and the progression unaltered (Fig. 5G). After 24 h, alterations in lung permeability, of disease was monitored over 24 h. LT induced a rapid and determined by measurement of the levels of albumin in the BAL progressive disease in the 129+/+ mice that was characterized by fluid, became apparent (Fig. 5H). Recruitment of cells into the a dramatic increase in KC, MIP-2, and HMGB1 levels. A small airways of the Nlrp1b2/2 mice was significantly lower than that increase in IL-1b and CXCL15 were also measured in the LT- observed in 129+/+ mice. This was consistent with the failure of treated animals (Fig. 5A–E). A parallel increase in the cellularity LT to lead to increase in levels of KC, MIP-2, IL-1b, CXCL15, or of the BAL fluid was observed in mice that received the toxin. HMGB1. The latter is an indication of cell death and has been However, cell numbers in the BAL fluid peaked at 6 h and then previously observed in Klebsiella pneumonia infected Nlrp32/2 remained largely unchanged throughout the course of the disease mice (22). Levels of LDH in the BAL fluid were significantly 6 NLRP1-DEPENDENT PYROPTOSIS CAUSES LUNG INJURY

FIGURE 5. Nlrp1b2/2 mice are pro- tected from lethal toxin-induced lung in- jury. 129+/+ (n) and Nlrp1b2/2 (N) mice were treated with 20 mg LT i.t. BAL fluid was collected at 0, 2, 6, 16, and 24 h after treatment. MIP-2 (A), KC (B), IL-1b (C), and CXCL15/lungkine (D) in the BAL fluid were detected by ELISA. (E) Western blot analysis of HMGB1 protein in con- centrated BAL fluid; GAPDH was used as a loading control, 20 mg protein per line. Results shown are representative of three independent experiments. (F)Thetotal number of inflammatory cells in BAL fluid. (G) Cell death in BAL fluid assessed by LDH activity. (H) Vascular permeability Downloaded from evaluated as a concentration of albumin in the BAL fluid, measured by colorimetric assay. (I) MPO level in BAL cell pellet collected from 129+/+ and Nlrp1b2/2 mice 48 and 72 h after LT treatment. (J–N) FACS analysis of whole lung cell suspen- http://www.jimmunol.org/ sion from 129+/+ (J, L) and Nlrp1b2/2 mice 16-h post-LT treatment (K, M). CD45+ cells were gated and analyzed for Mac-1 and GR-1 epitope (J, K) or Mac-1 and CCR-2 epitope (L, M). (J)–(M) are repre- sentative of three independent experiments, quantitative analysis of CD45+Mac-1+CCR2+ cells (N). n = 2 per genotype at 0 h, n = 5/genotype at 2 h, n = 6/genotype at 6 h, n = 3/genotype at 16 h, n = 7/genotype by guest on September 29, 2021 at 24–72 h. *p , 0.05, **p , 0.01, ***p , 0.001.

lower than those observed in the 129+/+ animals and did not sig- lungs collected from mice at various time points after LT expo- nificantly increase during the course of the experiments. The sure. Histological changes were assessed by analysis of H&E- permeability of the lungs was not altered, as albumin levels stained tissue sections prepared from the left lobe. As expected, remained low throughout the experiment (Fig. 5A–H). The ma- no difference was observed in lungs collected from vehicle-treated jority of the cells recruited to the lung after exposure to LT are Nlrp12/2 and 129+/+ animals (Fig. 6A, 6E). Exposure to LT in- neutrophils and their recruitment is attenuated in the mice lacking duced diffuse lung injury, and the extent of the damage became NLRP1 as shown by quantitative analysis of MPO levels in cell more apparent in samples collected at later time points, especially pellet obtained from the BAL at 48 and 72 h post-LT treatment 72 h after exposure to LT. Visualization of whole lung sections (Fig. 5I). Similarly, FACS analysis of single-cell suspensions of showed diffuse alveolar damage and evidence of hemorrhaging in whole lung revealed that the NLRP1-deficient mice had signifi- multiple regions of the lobe (Fig. 6B–D). Neutrophils and mac- cantly lower numbers of total lung neutrophils even at this earlier rophages could be detected in the alveolar walls and interstitial time point (16 h) (Fig. 5J, 5K). The recruitment of CCR2+ spaces surrounding the alveolar airways (Fig. 6I). However, con- monocytes has been noted in models of lung injury and inflam- sistent with the cellularity of the BAL fluid, the number of cells mation (23). We therefore examined the recruitment of these cells in the alveoli did not increase notably as the injury progressed. to the lung in single-cell preparations of whole lung using Abs Deposition of proteinous fibers in the airways was observed, specific for this population. As can be seen in Fig. 5L and 5M, consistent with significant trapping of RBCs within the alveolar inflammatory monocyte numbers were significantly reduce in the spaces (Fig. 6J, 6K). Although inflammatory infiltrates were ob- lungs of the NLRP1-deficient animals compare with 129+/+. served in the lymphatics, the airway epithelial cells remained 2/2 +/+ largely intact. Remarkably, lungs of Nlrp1b mice were largely Progression of acute lung injury in Nlrp1 animals spared from injury (Fig. 6F–H). Some increase in inflammatory To further compare the progression of lung disease in 129+/+ and cell numbers was observed at 48 and 72 h but was less extensive Nlrp1b2/2 animals, we evaluated changes in the structure of the compared with 129+/+ mice. The airways and alveolar architecture The Journal of Immunology 7

Casp-1+/+ mice, although the difference did not reach significance by 72 h (Fig. 7C). Levels of the proinflammatory mediators KC and CXCL15 were also significantly lower in Casp-1–deficient mice (Fig. 7D, 7E). In contrast, the lack of IL-1b failed to protect mice from the development of disease. No significant differences in inflammation or disease pathology were observed between Il-1b+/+ and Il-1b2/2 mice (Fig. 7F, 7G; data not shown). LT induces pyroptosis of alveolar macrophages We next examined whether caspase-1 activity was critical in the initiation and progression of lung disease in the 129+/+ LT-treated animals. Mice were treated with LT, and cells present in the air- ways were collected by BAL at 2 h after treatment. The viability of the cells was determined by staining with 7-AAD. Over 50% of the cells collected from 129+/+ animals were nonviable based on this criterion, whereas greater than 95% of the cells collected from the Nlr1b2/2 or Casp-12/2 mice failed to take up the dye (Fig. 8A). Thus, again, Casp-12/2 and Nlrp1b2/2 cells were protected

from cell death induced by LT. To determine whether catalytic Downloaded from activity of caspase-1–mediated cell death in 129+/+ animals, we stained the airway cells collected by BAL with the fluorescently labeled caspase-1 preferring substrate and inhibitor, FAM-YVAD- FMK. This inhibitor binds irreversibly to activated caspase-1. Nearly all the cells collected from the 129+/+ animals were FIGURE 6. Histological analysis of LT-induced lung injury. Micro- stained with the caspase-1 inhibitor, whereas no significant change http://www.jimmunol.org/ graphs of H&E-stained sections of lungs harvested from 129+/+ and in the mean fluorescent intensity was measured in the cells col- 2/2 Nlrp1b mice after treatment with 20 mg LT i.t. Lungs were harvested at lected from similarly treated Nlrp1b2/2 animals. As expected, no A H I L J K M N the indicated times ( – ), at 48 h ( , ), or at 72 h ( , , , ). Scale bars, caspase-1 activity was detected in cells collected in the BAL fluid m A H m I N 3 A H 3 I N 250 m( – ); 10 m(– ). Original magnification 2( – ), 40 ( – ). of Casp-12/2 mice. We next determined whether LT either di- KO, Knockout; WT, wild-type. rectly or indirectly resulted in activation of caspase-3. Only a small shift in the mean fluorescent intensity was observed in 2 2 were maintained, and no hemorrhaging was observed even at 72 h BAL cells from 129+/+ animals, but no cells from Casp-1 / or 2 2 (Fig. 6L–N). Consistent with these histological changes, 129+/+ Nlrp1b / animals were stained with the caspase-3 preferring mice exhibited significant inflammation, compromised of barrier fluorescent inhibitor FAM-DVED-FMK (Fig. 8B). Caspase-1 ac- by guest on September 29, 2021 function, lethargy, decline in body temperature, and displayed tivation in macrophages can occur downstream of cell death in morbidity requiring euthanasia (data not shown). In contrast, mice response to the leakage of some of the intracellular contents from lacking NLRP1b were largely protected from the sequel of events dying cells. For example, the accumulation of ATP and/or uric that followed exposure of the lung to LT. Even at 72 h post- acid in the extracellular milieu stimulates caspase-1 activity in treatment, Nlrp12/2 mice appeared normal and could not be nearby cells, and this activation is dependent on NLRP3 but not distinguished from untreated littermates. NLRP1 (25-27). However, we did not observe any attenuation of lung disease due to the absence of the NLRP3 inflammasome in LT-mediated lung injury is dependent upon caspase-1 Nalp32/2 mice (data not shown), suggesting that, similar to LT activation induced cell death in macrophages in vitro (Fig. 2), LT-induced Because exposure of the mouse airways to LTwas accompanied by pyroptosis in alveolar macrophages is NLRP1b and caspase-1 a modest, NLRP1b-dependent increase in IL-1b measured in the dependent but NLRP3 independent. Similarly, loss of P2X7 did BAL fluid (Fig. 5C), we further assessed the contribution of this not alter the development of severe lung disease after Nlrp1b NLRP1b pathway to the development of acute lung disease. We activation. generated cohorts of mice that expressed the 129+/+ 129S6 Nlrplb gene, which confers the ability to respond to LT, but lacked Casp- Discussion 1 or Il1b genes. The 129S6 mouse also carries a mutant Casp-4 Early studies using human cell free reconstitution showed that (alias Casp-11) (24). Thus all the animals included in the study NLRP1, in the presence of MDP, could form a complex with also lack expression of Casp-4. These mice and their Casp-1+/+ or caspases-1 capable of augmenting autoactivation of the caspase and Il1b+/+ littermates were exposed i.t. to LT, and the development of subsequent cleavage of pro IL-1b into its active form (6). ASC lung disease in the two groups of animals was compared. If the enhanced this activity but was not essential for caspase-1 cleav- lung disease observed in the LT-treated animals is the conse- age. The requirement for MDP to initiate the assembly of the quence of pyroptosis, a caspase-1–mediated cell death, we would complex suggested that at least one of the functions of the NLRP1 expect that the Casp-12/2 mice would be protected from the toxin. inflammasome was intracellular surveillance for this bacterial cell Indeed, the lack of caspase-1 protected the mice from LT-induced wall component. These initial observations linking MDP and the lung inflammation. Inflammatory cell recruitment was reduced in NLRP1 inflammasome were supported by indirect evidence col- Casp-12/2 mice exposed to LT compared with Casp-1+/+ mice lected from a number of studies that followed. Mouse macro- (Fig. 7A). Cell death was also significantly reduced in Casp-12/2 phages lacking NOD2 failed to produce IL-1b in response to mice (Fig. 7B), as expected based on our in vitro analyses of MDP, clearly defining a role for this protein in the response to Casp-12/2 macrophages (Fig. 2B). Correspondingly, vascular MDP (21). This observation was consistent with earlier studies permeability was less severe in Casp-12/2 mice compared with that showed that introduction of NOD2 into cell lines allowed 8 NLRP1-DEPENDENT PYROPTOSIS CAUSES LUNG INJURY Downloaded from http://www.jimmunol.org/

FIGURE 7. LT-induced lung injury is dependent on caspase-1 but not IL-1b.(A–E) Casp-1+/+ (black) and Casp-12/2 (shade) mice were treated with PBS as a vehicle or 20 mg LT i.t. BAL fluid was collected at 72 h after treatment. (A) The total number of inflammatory cells collected in the BAL fluid. (B) Cell death in BAL fluid determined by LDH activity in supernatants and expressed as the percentage of total cellular LDH activity. (C) Vascular permeability evaluated as a concentration of albumin in the BAL fluid, measured by colorimetric assay. KC (D) and CXCL15/lungkine (E) in the BAL fluid were detected by ELISA. n = 2 mice/genotype for vehicle treatment and n = 10 mice/genotype for LT treatment. (F and G) Il1b+/+ (black) and Il1b2/2 (striped) mice were by guest on September 29, 2021 treated as above. (F) The total number of inflammatory cells collected in BAL fluid. (G) The percentage of weight lost 72 h after treatment. n = 2/genotype for vehicle treatment and n = 9 mice/genotype for LT treatment. *p , 0.05, **p , 0.01, #p , 0.0001. these cells to respond to MDP (28, 29). To reconcile this with the studies also failed to define a role for NLRP1 in the response to cell-free studies showing MDP-dependent assembly of the NLRP1 this bacterial wall component. In all of these models, increase in inflammasome, IL-1b production in response to MDP was ex- IL-1b release was entirely dependent on Nlrp3 expression. No amined in the human monocyte line, THP-1, as well as HEK293T significant increase in release of IL-1b was observed from mac- cells transfected with various components of the inflammasome. rophages isolated from the 129 Nlrp12/2 coisogenic mice after In this study, complicity between NOD2 and NLRP1 in IL-1b stimulation with MDP, and Nlrp32/2 mice were protected from release was supported by demonstrating association of NLRP1 elevations in serum IL-1b after MDP exposure. Thus, our results with NOD2 using coimmunoprecipitation strategies after MDP support a model in which NLRP3 rather than NLRP1, alone or in exposure as well as demonstration of a decrease in release of cooperation with NOD2, mediates IL-1b processing and release in IL-1b after RNA interference knockdown of NLRP1 expression. response to MDP. Our findings are consistent with a previous MDP has also been used to stimulate NLRP1 inflammasome as- study examining the response of NOD2 and NLRP3 deficient bone sembly in studies focused on identifying regulators of NLRP1 marrow derived macrophages to MDP (30). No increase in IL-1b activity, again using THP-1 cells. Although NLRP1-dependent release was observed after stimulation of either Nod22/2 or maturation and release of IL-1b is also reported in this study, Nlrp32/2 cells, suggesting a possible collaboration between these interpretation of these studies is complicated by the treatment two NLRs in caspase-1 activation and maturation in response to of cells with ATP, a potent stimulator of NLRP3 inflammasome MDP exposure. These findings together with our results raise the assembly (4). possibility that there are species differences in assignment of Species conservation in the ability of a particular PAMP to functions to inflammasomes: the importance of NLRP3 in the trigger a specific inflammasome has generally been observed, albeit response to MDP being limited to mouse, with NLRP1 carrying in the case of humans these studies are heavily dependent on use out the orthologous function in human cells. However, this in- of the THP-1 cell line. We were, therefore, surprised to observe terpretation is not consistent with previous studies using a human no difference in IL-1b release between wild type and NLRP1- ectopic expression system in which MDP was shown to stimulate deficient cells in response to MDP. This was true even when, as the NLRP3. Furthermore, Martinon et al. (31) show that cells from in previous studies, we used TiO2 to increase uptake of MDP by Muckle–Wells patients, who carry gain of function mutations in peritoneal macrophages. Lack of NLRP1 did not alter the response NLRP3, display increased IL-1b secretion in the presence of MDP. to L18-MDP, a derivative with improved cellular uptake. In vivo Taken together, our study and the differences between previous The Journal of Immunology 9

FIGURE 8. LT activates caspase-1 in the lung independently on P2X7 receptor. (A and B)129+/+, Nlrp1b2/2, and Casp-12/2 mice were treated with 20 mg LT by i.t., BAL fluid was collected 2 h after treatment, and cells were stained with fluorescently labeled caspase inhibitors and with 7-AAD; n = 4/genotype. Cells were analyzed on FACS and compared with untreated controls. (A) Data are expressed as the difference in median Downloaded from fluorescent intensity (MFI) in the LT- treated mice and untreated mice. (B) A representative overlay of the un- treated (gray filled) and LT-treated (black line) cell populations. (C–F) BALB/c or P2X72/2 mice were treated with 20 mg LT i.t. BAL fluid http://www.jimmunol.org/ was collected at 72 h after treatment. (C) The total number of inflammatory cells collected in the BAL fluid. (D) Vascular permeability evaluated as a concentration of protein in the BAL fluid, measured by bicinchoninic acid assay. (E) Cell death in BAL fluid determined by LDH activity in su-

pernatants. (F) KC in the BAL fluid by guest on September 29, 2021 was detected by ELISA, n = 5 mice/ genotype. *p , 0.05, **p , 0.01, ***p , 0.001 compared with WT.

reports, depending on the system used in the study, suggest that activation can lead to macrophage pyroptosis. Activation of this the role of NLRP1 in the response to MDP needs to be re-eval- pathway is also observed in vivo. Exposure of the mouse airways uated. to LT resulted in activation of caspase-1 in alveolar macrophages, Given the failure of MDP to stimulate the inflammasome, we resulting in cell death. Again, the accompanying increase in IL-1b used anthrax LT to define events following NLRP1 inflammasome release was only observed if the animals had been exposed to LPS, assembly in peritoneal macrophages and in vivo in the lung. which stimulates transcription of this cytokine. Consistent with previous observations (32), the most dramatic The cellular events critical for engagement of the NLRP1 in- consequence of exposure of macrophages to LT was cell death as flammasome that follow LT exposure have not been defined, nor monitored by increase in cellular dehydrogenase levels in vitro in is it clear why the Nlrp1b allele carried by a number of inbred culture supernatant. In vivo this was manifested as an increase in mouse lines, including BALB/c and 129, is uniquely activated by LDH in the BAL fluid collected from LT-treated animals. When this toxin. LT is a zinc metalloproteinase, and cleavage of MEK by cells were treated with LPS prior to exposure to LT, IL-1b release this proteinase and subsequent alteration in the activity of ERK, was also observed. Caspase-1–deficient cells and Casp12/2 mice JnK, and p38 pathways has been documented by a number of are largely protected. Because 129 mice lack a functional Casp-4 groups (33–36). These events do not distinguish the LT-resistant (formerly Casp-11) gene, it is clear that activation of caspase-1 by strains such as C57BL/6 from LT-sensitive strains such as 129. LT the NLRP1 inflammasome is sufficient for both IL-1b processing inactivation of MEK occurs equally in both 129 and C57BL/6 and cell death (24). The dependence of cell death on caspase-1 cells, even though the latter cells do not undergo rapid pyropto- and the accompanying release of cytokines indicate that NLRP1b sis, but rather after an extended period of time begin to undergo 10 NLRP1-DEPENDENT PYROPTOSIS CAUSES LUNG INJURY apoptosis (35, 37). Interestingly, it was recently discovered that Although it is well established that IL-1b can induce inflam- the N-terminal domain of NLRP1 from LT-sensitive strain is mation by stimulating resident immune cells to produce chemo- cleaved by LT and that the cleavage is essential for NLRP1 kines (45), the IL-1b observed in this model does not contribute inflammasome activation, caspase-1 activation, IL-1b production, significantly to the development of ARDS-like disease. In con- and pyroptotic macrophage death. This observation suggests that trast, the lungs of the Casp12/2 mice were protected, suggesting allelic differences lead to alterations in the sensitivity of NLRP1 to that caspase-1 activation and the resulting cell death, rather than cleavage by proteases. In this case the 129 and BALB alleles show IL-1b processing, is the critical outcome of NLRP1 activation in dramatic increase in sensitivity to LT, but it is equally likely that this model. We have not yet identified the critical factors/ the many common coding variants in the mouse and human gene cytokines released by the initial NLRP1-dependent pyroptosis alter the sensitivity of NLRP1 to other proteases, thus modifying that lead to escalation in lung injury. ATP is released upon cell the activation of this inflammasome in response to various innate death and could lead to activation of the NLRP3 inflammasome in stimuli (38, 39). It is apparent that, once assembly of the Nlrp1b neighboring cells, inducing further cytokine release and pyropto- inflammasome is triggered, it leads to cellular events that mimic sis. A role for this pathway is not supported by the observation those of other inflammasomes, namely caspase-1 recruitment and that neither the Nlrp32/2 nor the P2X72/2 mice were protected autoactivation and IL-1b maturation and release. Cell death from lung damage. It also makes it unlikely that uric acid released requires caspase-1. However, we cannot rule out the possibility from dying cells contributes to further tissue damage, because this that the other activities of LT leave the cells vulnerable to py- damage-associated molecular pattern has also been shown to rely roptosis. on the NLRP3 inflammasome. HMGB1 released from pyroptotic

Similarly to LT exposure, infection with virulent strains of cells can induced acute inflammation characterized by increased Downloaded from Bacillus anthracis results in caspase-1 activation and IL-1b pro- cytokine production and neutrophil recruitment (46). In animal duction and release. However, contrary to LT treatment, caspase-1 models of systemic injury, inhibition of HMGB1 is reported to as well as IL-1b protects infected animals against B. anthracis attenuate pulmonary endothelial barrier dysfunction. Thus, it is infection (40). Although it was shown that LT is a major virulence possible that release of this intracellular protein contributes to loss factor during B. anthracis infection, other virulence factors are of vascular integrity, marked by hemorrhaging and increase in

probably involved during infection with B. anthracis compare serum proteins in the airways, observed in this NLRP1 dependent http://www.jimmunol.org/ with sterile induction of inflammation by LT. This is supported by model of ARDS. Further studies using specific inhibitors of fac- recent data, showing that mice lacking major components of other tors released upon initial activation of the NLRP1 inflammasome inflammasomes (Nlrc4) and receptor for MDP (Nod2) are critical as well as mice in which NLRP1 expression is limited to specific for protection from virulent B. anthracis infection (40). cell populations should allow further definition of the events that In the lung, the activation of NLRP1 has catastrophic con- lead from initial activation of NLRP1 to acute and progressive sequences, resulting in changes that mimic some aspects of acute lung disease. respiratory distress (ARDS) (41). A notable feature of this in- flammation is that, unlike many other sterile models of ARDS

Acknowledgments by guest on September 29, 2021 disease, NLRP1 activation results in progressive damage to the We thank Anne Latour for embryonic stem cell work, Darren Hursey for lung. In contrast, exposure of mouse lungs to endotoxin results molecular biology, and Rebecca Dye, Peter Repenning, and William Barker in a rapid induction of chemokines followed by an influx of for assistance in breeding lines and genotyping mice. neutrophils/monocytes and leukocytes. However, the inflamma- tion rapidly resolves, and the majority of animals show no evi- Disclosures dence of disease at 72 h after treatment. NLRP1 activation also The authors have no financial conflicts of interest. results in a rapid induction of chemokines, including those most relevant for neutrophil recruitment KC (CXCL1) and MIP-2 (42). Lungkine, (CXCL15) which is expressed at high levels by lung References airway epithelial cells (43) and is protective in bacterial pneu- 1. Hlaing, T., R. F. Guo, K. A. Dilley, J. M. Loussia, T. A. Morrish, M. M. Shi, monia, is also found at higher levels in the inflamed lung (44). C. Vincenz, and P. A. Ward. 2001. Molecular cloning and characterization of DEFCAP-L and -S, two isoforms of a novel member of the mammalian Ced-4 However, unlike LPS induced models of lung injury, damage family of apoptosis proteins. J. Biol. Chem. 276: 9230–9238. continues to accumulate, with visible lesions apparent upon nec- 2. Chu, Z. L., F. Pio, Z. Xie, K. Welsh, M. Krajewska, S. Krajewski, A. Godzik, and J. C. Reed. 2001. A novel enhancer of the Apaf1 involved in cy- ropsy 72 h after NLRP1 activation. It is likely, therefore, that the tochrome c-dependent caspase activation and apoptosis. J. Biol. Chem. 276: initial pyroptosis of the resident macrophages initiates events that 9239–9245. overwhelm the homeostatic mechanisms of the lung. Consistent 3. Yuan, J. Y., and H. R. Horvitz. 1990. The Caenorhabditis elegans genes ced-3 and ced-4 act cell autonomously to cause programmed cell death. Dev. Biol. 138: with this, the dramatic loss in barrier function becomes apparent 33–41. long after the initial activation of the inflammasome by LT. An 4. Bruey, J. M., N. Bruey-Sedano, F. Luciano, D. Zhai, R. Balpai, C. Xu, important clue to the mechanism(s) leading to organ failure is the C. L. Kress, B. Bailly-Maitre, X. Li, A. Osterman, et al. 2007. Bcl-2 and Bcl-XL regulate proinflammatory caspase-1 activation by interaction with NALP1. Cell steady increase in LDH levels in the BAL fluid. This indicates that 129: 45–56. after the initial NLRP1-dependent pyroptosis, cell death in the 5. Martinon, F., K. Burns, and J. Tschopp. 2002. The inflammasome: a molecular platform triggering activation of inflammatory caspases and processing of pro– lungs continues and in fact increases. Susceptible cells may in- IL-b. Mol. Cell 10: 417–426. clude structural cells of the lung and cells recruited by the initial 6. Faustin, B., L. Lartigue, J. M. Bruey, F. Luciano, E. Sergienko, B. Bailly-Maitre, pyroptosis of resident macrophages, neutrophils, and monocytes. N. Volkmann, D. Hanein, I. Rouiller, and J. C. Reed. 2007. Reconstituted NALP1 inflammasome reveals two-step mechanism of caspase-1 activation. Consistent with pyroptosis of resident cells, high levels of HMGB1 Mol. Cell 25: 713–724. are observed in the BAL immediately after exposure to LT. Ele- 7. Galluzzi, L., I. Vitale, J. M. Abrams, E. S. Alnemri, E. H. Baehrecke, vated levels observed at later time points with continued detection M. V. Blagosklonny, T. M. Dawson, V. L. Dawson, W. S. El-Deiry, S. Fulda, et al. 2012. Molecular definitions of cell death subroutines: recommendations of the of IL-1b suggest that pyroptosis continues. In these cases it is Nomenclature Committee on Cell Death 2012. Cell Death Differ. 19: 107–120. possible that cell death is mediated by damage-associated mo- 8. Miao, E. A., I. A. Leaf, P. M. Treuting, D. P. Mao, M. Dors, A. Sarkar, S. E. Warren, M. D. Wewers, and A. Aderem. 2010. Caspase-1–induced lecular patterns released when the cells undergo the NLRP1- pyroptosis is an innate immune effector mechanism against intracellular bacte- dependent pyroptosis in response to LT. ria. Nat. Immunol. 11: 1136–1142. The Journal of Immunology 11

9. Sutterwala, F. S., Y. Ogura, D. S. Zamboni, C. R. Roy, and R. A. Flavell. 2006. terial muramyl dipeptide mediated through NOD2: implications for Crohn’s NALP3: a key player in caspase-1 activation. J. Endotoxin Res. 12: 251–256. disease. J. Biol. Chem. 278: 5509–5512. 10. Franchi, L., A. Amer, M. Body-Malapel, T. D. Kanneganti, N. Ozo¨ren, 29. Girardin, S. E., I. G. Boneca, J. Viala, M. Chamaillard, A. Labigne, G. Thomas, R. Jagirdar, N. Inohara, P. Vandenabeele, J. Bertin, A. Coyle, et al. 2006. Cy- D. J. Philpott, and P. J. Sansonetti. 2003. Nod2 is a general sensor of peptido- tosolic flagellin requires Ipaf for activation of caspase-1 and interleukin 1b in glycan through muramyl dipeptide (MDP) detection. J. Biol. Chem. 278: 8869– Salmonella-infected macrophages. Nat. Immunol. 7: 576–582. 8872. 11. Miao, E. A., C. M. Alpuche-Aranda, M. Dors, A. E. Clark, M. W. Bader, 30. Pan, Q., J. Mathison, C. Fearns, V. V. Kravchenko, J. Da Silva Correia, S. I. Miller, and A. Aderem. 2006. Cytoplasmic flagellin activates caspase-1 and H. M. Hoffman, K. S. Kobayashi, J. Bertin, E. P. Grant, A. J. Coyle, et al. 2007. secretion of interleukin 1b via Ipaf. Nat. Immunol. 7: 569–575. MDP-induced interleukin-1b processing requires Nod2 and CIAS1/NALP3. J. 12. Fernandes-Alnemri, T., J. W. Yu, P. Datta, J. Wu, and E. S. Alnemri. 2009. AIM2 Leukoc. Biol. 82: 177–183. activates the inflammasome and cell death in response to cytoplasmic DNA. 31. Martinon, F., L. Agostini, E. Meylan, and J. Tschopp. 2004. Identification of Nature 458: 509–513. bacterial muramyl dipeptide as activator of the NALP3/cryopyrin inflamma- 13. Hornung, V., A. Ablasser, M. Charrel-Dennis, F. Bauernfeind, G. Horvath, some. Curr. Biol. 14: 1929–1934. D. R. Caffrey, E. Latz, and K. A. Fitzgerald. 2009. AIM2 recognizes cytosolic dsDNA 32. Friedlander, A. M. 1986. Macrophages are sensitive to anthrax lethal toxin and forms a caspase-1-activating inflammasome with ASC. Nature 458: 514–518. through an acid-dependent process. J. Biol. Chem. 261: 7123–7126. 14. Elinav, E., T. Strowig, A. L. Kau, J. Henao-Mejia, C. A. Thaiss, C. J. Booth, 33. Klimpel, K. R., N. Arora, and S. H. Leppla. 1994. Anthrax toxin lethal factor D. R. Peaper, J. Bertin, S. C. Eisenbarth, J. I. Gordon, and R. A. Flavell. 2011. contains a zinc metalloprotease consensus sequence which is required for lethal NLRP6 inflammasome regulates colonic microbial ecology and risk for colitis. toxin activity. Mol. Microbiol. 13: 1093–1100. Cell 145: 745–757. 34. Duesbery, N. S., C. P. Webb, S. H. Leppla, V. M. Gordon, K. R. Klimpel, 15. Khare, S., A. Dorfleutner, N. B. Bryan, C. Yun, A. D. Radian, L. de Almeida, T. D. Copeland, N. G. Ahn, M. K. Oskarsson, K. Fukasawa, K. D. Paull, and Y. Rojanasakul, and C. Stehlik. 2012. An NLRP7-containing inflammasome G. F. Vande Woude. 1998. Proteolytic inactivation of MAP-kinase-kinase by mediates recognition of microbial lipopeptides in human macrophages. Immunity anthrax lethal factor. Science 280: 734–737. 23: 464–476. 35. Park, J. M., F. R. Greten, Z. W. Li, and M. Karin. 2002. Macrophage apoptosis 16. Davis, B. K., R. A. Roberts, M. T. Huang, S. B. Willingham, B. J. Conti, by anthrax lethal factor through p38 MAP kinase inhibition. Science 297: 2048– W. J. Brickey, B. R. Barker, M. Kwan, D. J. Taxman, M. A. Accavitti-Loper, 2051. et al. 2011. Cutting edge: NLRC5-dependent activation of the inflammasome. J. 36. Vitale, G., R. Pellizzari, C. Recchi, G. Napolitani, M. Mock, and C. Montecucco. Downloaded from Immunol. 186: 1333–1337. 1998. Anthrax lethal factor cleaves the N-terminus of MAPKKs and induces 17. Boyden, E. D., and W. F. Dietrich. 2006. Nalp1b controls mouse macrophage tyrosine/threonine phosphorylation of MAPKs in cultured macrophages. Bio- susceptibility to anthrax lethal toxin. Nat. Genet. 38: 240–244. chem. Biophys. Res. Commun. 248: 706–711. 18. Quintiliani, R., Jr., and R. Quintiliani. 2003. Inhalational anthrax and bio- 37. Muehlbauer, S. M., T. H. Evering, G. Bonuccelli, R. C. Squires, A. W. Ashton, terrorism. Curr. Opin. Pulm. Med. 9: 221–226. S. A. Porcelli, M. P. Lisanti, and J. Brojatsch. 2007. Anthrax lethal toxin kills 19. Moayeri, M., D. Haines, H. A. Young, and S. H. Leppla. 2003. Bacillus anthracis macrophages in a strain-specific manner by apoptosis or caspase-1-mediated lethal toxin induces TNF-a–independent hypoxia-mediated toxicity in mice. J. necrosis. Cell Cycle 6: 758–766.

Clin. Invest. 112: 670–682. 38. Levinsohn, J. L., Z. L. Newman, K. A. Hellmich, R. Fattah, M. A. Getz, S. Liu, http://www.jimmunol.org/ 20. Solle, M., J. Labasi, D. G. Perregaux, E. Stam, N. Petrushova, B. H. Koller, I. Sastalla, S. H. Leppla, and M. Moayeri. 2012. Anthrax lethal factor cleavage R. J. Griffiths, and C. A. Gabel. 2001. Altered cytokine production in mice of Nlrp1 is required for activation of the inflammasome. PLoS Pathog. 8: lacking P2X(7) receptors. J. Biol. Chem. 276: 125–132. e1002638. 21. Hsu, L. C., S. R. Ali, S. McGillivray, P. H. Tseng, S. Mariathasan, E. W. Humke, 39. Frew, B. C., V. R. Joag, and J. Mogridge. 2012. Proteolytic processing of Nlrp1b L. Eckmann, J. J. Powell, V. Nizet, V. M. Dixit, and M. Karin. 2008. A NOD2- is required for inflammasome activity. PLoS Pathog. 8: e1002659. NALP1 complex mediates caspase-1–dependent IL-1b secretion in response to 40. Ali, S. R., A. M. Timmer, S. Bilgrami, E. J. Park, L. Eckmann, V. Nizet, and Bacillus anthracis infection and muramyl dipeptide. Proc. Natl. Acad. Sci. USA M. Karin. 2011. Anthrax toxin induces macrophage death by p38 MAPK inhi- 105: 7803–7808. bition but leads to inflammasome activation via ATP leakage. Immunity 35: 34– 22.Willingham,S.B.,I.C.Allen,D.T.Bergstralh,W.J.Brickey,M.T.Huang, 44. D. J. Taxman, J. A. Duncan, and J. P. Ting. 2009. NLRP3 (NALP3, Cryopyrin) 41. Matute-Bello, G., G. Downey, B. B. Moore, S. D. Groshong, M. A. Matthay, facilitates in vivo caspase-1 activation, necrosis, and HMGB1 release via A. S. Slutsky, and W. M. Kuebler. 2011. An official American Thoracic Society

inflammasome-dependent and -independent pathways. J. Immunol. 183: 2008–2015. workshop report: features and measurements of experimental acute lung injury by guest on September 29, 2021 23. Gordon, S., and P. R. Taylor. 2005. Monocyte and macrophage heterogeneity. in animals. Am. J. Respir. Cell Mol. Biol. 44: 725–738. Nat. Rev. Immunol. 5: 953–964. 42. Grommes, J., and O. Soehnlein. 2011. Contribution of neutrophils to acute lung 24. Kayagaki, N., S. Warming, M. Lamkanfi, L. Vande Walle, S. Louie, J. Dong, injury. Mol. Med. 17: 293–307. K. Newton, Y. Qu, J. Liu, S. Heldens, et al. 2011. Non-canonical inflammasome 43. Rossi, D. L., S. D. Hurst, Y. Xu, W. Wang, S. Menon, R. L. Coffman, and activation targets caspase-11. Nature 479: 117–121. A. Zlotnik. 1999. Lungkine, a novel CXC chemokine, specifically expressed by 25. Schroder, K., R. Zhou, and J. Tschopp. 2010. The NLRP3 inflammasome: lung bronchoepithelial cells. J. Immunol. 162: 5490–5497. a sensor for metabolic danger? Science 327: 296–300. 44. Chen, S. C., B. Mehrad, J. C. Deng, G. Vassileva, D. J. Manfra, D. N. Cook, 26. Mariathasan, S., D. S. Weiss, K. Newton, J. McBride, K. O’Rourke, M. Roose- M. T. Wiekowski, A. Zlotnik, T. J. Standiford, and S. A. Lira. 2001. Impaired Girma, W. P. Lee, Y. Weinrauch, D. M. Monack, and V. M. Dixit. 2006. Cry- pulmonary host defense in mice lacking expression of the CXC chemokine opyrin activates the inflammasome in response to toxins and ATP. Nature 440: lungkine. J. Immunol. 166: 3362–3368. 228–232. 45. Ulich, T. R., S. M. Yin, K. Z. Guo, J. del Castillo, S. P. Eisenberg, and 27. Martinon, F., V. Pe´trilli, A. Mayor, A. Tardivel, and J. Tschopp. 2006. Gout- R. C. Thompson. 1991. The intratracheal administration of endotoxin and associated uric acid crystals activate the NALP3 inflammasome. Nature 440: cytokines. III. The interleukin-1 (IL-1) receptor antagonist inhibits endotoxin- 237–241. and IL-1–induced acute inflammation. Am. J. Pathol. 138: 521–524. 28. Inohara, N., Y. Ogura, A. Fontalba, O. Gutierrez, F. Pons, J. Crespo, K. Fukase, 46. Abraham, E., J. Arcaroli, A. Carmody, H. Wang, and K. J. Tracey. 2000. HMG-1 S. Inamura, S. Kusumoto, M. Hashimoto, et al. 2003. Host recognition of bac- as a mediator of acute lung inflammation. J. Immunol. 165: 2950–2954.