Caspase-1/ASC Inflammasome-Mediated Activation of IL-1β–ROS–NF-κB Pathway for Control of Trypanosoma cruzi Replication and Survival Is Dispensable in NLRP3−/− Macrophages

In this study, we have utilized wild-type (WT), ASC−/−, and NLRP3−/− macrophages and inhibition approaches to investigate the mechanisms of inflammasome activation and their role in Trypanosoma cruzi infection. We also probed human macrophages and analyzed published microarray datasets from human fibroblasts, and endothelial and smooth muscle cells for T. cruzi-induced changes in the expression genes included in the RT Profiler Human Inflammasome arrays. T. cruzi infection elicited a subdued and delayed activation of inflammasome-related gene expression and IL-1β production in mφs in comparison to LPS-treated controls. When WT and ASC−/− macrophages were treated with inhibitors of caspase-1, IL-1β, or NADPH oxidase, we found that IL-1β production by caspase-1/ASC inflammasome required reactive oxygen species (ROS) as a secondary signal. Moreover, IL-1β regulated NF-κB signaling of inflammatory cytokine gene expression and, subsequently, intracellular parasite replication in macrophages. NLRP3−/− macrophages, despite an inability to elicit IL-1β activation and inflammatory cytokine gene expression, exhibited a 4-fold decline in intracellular parasites in comparison to that noted in matched WT controls. NLRP3−/− macrophages were not refractory to T. cruzi, and instead exhibited a very high basal level of ROS (>100-fold higher than WT controls) that was maintained after infection in an IL-1β-independent manner and contributed to efficient parasite killing. We conclude that caspase-1/ASC inflammasomes play a significant role in the activation of IL-1β/ROS and NF-κB signaling of cytokine gene expression for T. cruzi control in human and mouse macrophages. However, NLRP3-mediated IL-1β/NFκB activation is dispensable and compensated for by ROS-mediated control of T. cruzi replication and survival in macrophages.


Introduction
Chagas disease affects 11-18 million people world-wide [1]. Upon exposure to Trypanosoma cruzi (T. cruzi or Tc), infected individuals exhibit an acute phase of Chagas disease that lasts for a couple of months and is characterized by symptoms such as fever, fatigue, body aches, diarrhea, and vomiting. After the control of parasitemia, a majority of infected patients enter an indeterminate chronic phase that is marked by a lack of clinical symptoms of the acute phase. Ten to thirty years after initial infection, 30-40% of indeterminate phase patients progress to develop chagasic cardiomyopathy [2].
The studies in experimental models have shown that macrophages (mQs), as well as dendritic and natural killer cells, play an important role in control of T. cruzi infection [3][4][5]. The interaction of T. cruzi with mQs and other cell types involved in the innate immune response are mediated by pattern recognition receptors (PRRs) such as toll-like receptors (TLRs). Upon recognition of pathogen-assoCiated molecular patterns (PAMPs), TLRs transmit the signal via cytoplasmic domains for the recruitment of cytosolic adaptor molecules, including myeloid differentiation primary-response protein 88 (MyD88), and subsequently induce nuclear factor kB (NFkB) activation, leading to the production of inflammatory cytokines and linking an innate response to an adaptive immune response (reviewed in [4]). T. cruzi-derived glycosylphosphatidylinositols and mucins have been shown to serve as PAMPs in engaging TLR signaling of a cytokine response. Others have demonstrated that TLR4 and TLR9 are engaged by parasite-derived glycosylinositol phospholipids and DNA, respectively, during the activation of host innate immune response leading to regulation of infection [6,7]. T. cruzi also expresses cruzipain, a kinin-releasing cysteine protease, which induces dendritic cells maturation via activation of bradykinin (BK) B 2 receptors (B 2 R) [8,9].
A newly discovered family of PRRs is named Nucleotidebinding oligomerization domain (NOD) like receptors (NLRs) [10,11]. NLRs have a tripartite domain structure and are characterized by the presence of a central nucleotide-binding oligomerization domain (NOD), also called NACHT domain, present in neuronal apoptosis inhibitor proteins (NAIP) and a Cterminal leucine-rich repeats (LRRs) domain of variable length (20-29 amino acids). The N-terminal effector binding region consists of a protein-to-protein interaction domain, i.e., Pyrin domain (PYD), a caspase recruitment domain (CARD), or baculovirus inhibitor of an apoptosis protein repeat (BIR) domain. Based upon the presence of PYD, CARD and BIR effector domains, NLRs are classified as NLRP, NLRC, and NAIP, respectively [11,12]. Currently known members of the NLR family in humans include seven NLRCs (NLRC1-NLRC5, NLRX, and CIITA or NLRA), fourteen NLRPs (NLRP1-NLRP14), and seven NAIPs (NAIP1-NAIP7). The multi-meric protein macromolecules formed by NLRs are named inflammasomes. The most studied NLRP1 and NLRP3 inflammasomes recruit ASC (apoptosis-assoCiated, speck-like protein containing a CARD domain) and caspase-1 proteins. The ASC-dependent cleavage and activation of caspase-1 results in the formation of an active complex responsible for converting to active forms of pro-IL-1b (31 kDa to 17 kDa) and pro-IL-18 (24 kDa to 18 kDa) [13] and the activation of the inflammatory cytokine response.
In the context of pathogens invading the heart, it is recognized that besides innate immune cells, both endothelial and vascular smooth muscle cells (VSMCs) can also sense and respond to pathogens (or PAMPs) [14][15][16]. CardiomyoCytes, the main type of cells in the heart, and heart resident fibroblasts also express TLRs and/or NLRs [17,18].
In this study, we have utilized wild type (WT), ASC 2/2 and NLRP3 2/2 mQs and inhibitory approaches to investigate the mechanisms of inflammasome activation and their role in the context of T. cruzi infection. We also probed the RT Profiler PCR Array System to identify the inflammasome-related changes induced by T. cruzi infection of human mQs and analyzed the published microarray datasets from T. cruzi-infected fibroblasts, and smooth muscle and endothelial cells for the change in expression of the 84 genes included in the inflammasome arrays. Our data demonstrate that T. cruzi infection, in comparison to treatment with LPS, elicits a subdued activation of inflammatory gene expression and IL-1b production in mQs. Yet, caspase-1/ ASC inflammasome-dependent activation of the IL-1b -reactive oxygen species (ROS) -NF-kB pathway played an important role in control of T. cruzi replication in mQs. Further, NLRP3 controlled the ROS levels in mQs, and NLRP3 deficiency resulted in a potent increase in ROS-mediated parasite killing in infected mQs. To the best of our knowledge, this is the first study demonstrating a double-edged role of NLRP3 in determining mQ activation of ROS and cytokine response, both of which are required for clearance of T. cruzi infection.

Ethics statement
All animal experiments were performed according to the National Institutes of Health Guide for Care and Use of Experimental Animals and approved by the UTMB's Animal Care and Use Committee (protoCol # 08-05-029).
Mice, parasites, and cells C57BL/6 female mice (6-8-weeks old) were purchased from Harlan Labs (Indianapolis, IN). NLRP3 2/2 and ASC 2/2 mice In some experiments, ATP was added during last 30 min of incubation (C&D). IL-1b release in supernatants was determined by ELISA. (E-G) IL-1b contributes to parasite control in mQs. THP-1 mQs were incubated with SYTO H 11-labeled T. cruzi in the presence or absence of anti-IL-1b antibody for 18 h. (E) SYTO H 11 fluorescence as an indicator of parasite uptake (shown by arrows) was determined by using an Olympus BX-15 microscope equipped with a digital camera (magnification 40X). (F) Quantitative PCR analysis of parasite burden in infected mQs by using Tc18SrDNA-specific oligonucleotides (normalized to human GAPDH). (G) Addition of anti-IL-1b antibody depletes secreted IL-1b levels in T. cruziinfected mQs. In all figures, data are representative of three independent experiments and presented as mean 6 SD. Significance is shown by *normal versus infected and # treated/infected versus infected (* ,# p,0.05, ** ,## p,0.01, and *** ,### p,0.001). doi:10.1371/journal.pone.0111539.g001 Novel Role of Inflammasome in T. cruzi Infection PLOS ONE | www.plosone.org Table 1. Inflammasome-related differential gene expression in THP-1 macrophages in response to T. cruzi infection (6 ATP) in comparison to normal controls. Single-cell suspensions of bone marrow (BM) -derived monoCytes from WT, ASC 2/2 and NLRP3 2/2 mice (C57BL/ 6 background) were added to petri dishes (10 6 cells/ml) in complete RPMI media containing 20 ng/ml murine macrophagecolony stimulating factor (M-CSF, eBioscience, San Diego, CA) and incubated at 37uC in 5% CO 2 for 10 days to support differentiation to mQs. The differentiated BM mQs were maintained in the presence of 5 ng/ml M-CSF during experimental use. THP-1 human monoCytes were differentiated into mQs by overnight incubation with 50 ng/ml phorbol-12-myristate-13acetate (PMA), and then rested at 37uC/5% CO 2 for 48 h in RPMI complete media containing 10% FBS. RAW 264.7 murine mQs were routinely cultured in DMEM with 10% FBS.

Probing the RT Profiler Human Inflammasome PCR Arrays
THP-1 cells, with and without T. cruzi infection and treatments, were harvested with 500 ml Bio-Rad cell lysis/RNA extraction buffer. Total RNA was extracted by using an Aurum DNA-free RNA isolation kit (Bio-Rad, Hercules, CA) and measured at 260 and 280 nm for determination of purity and concentration. The cDNA probes were generated by reverse transcription of 5 mg total RNA by using the Bio-Rad iScript cDNA synthesis kit.
The 96-well RT Profiler Human Inflammasome PCR Arrays (SA Biosciences/Qiagen, Valencia, CA) containing primer pairs for 84 key genes involved in the function of inflammasomes and NLR signaling were probed with 2 ml of cDNA template in the presence of 6.5 ml of dNTPs, MgCl 2 , and stabilizers (iQ SYBR Supermix, Bio-Rad), and PCR was carried out on a LightCycler 480 Multiple Plate System. A total of 42 arrays were probed with 14 research samples in triplicate, and datasets were analyzed by Web-based PCR Array Data Analysis software (SA Biosciences) for threshold cycle (Ct) value determination.
The Ct values from qPCR data were analyzed by using the open source HTqPCR v.1.12 software package [19]. Briefly, all array data were imported into HTqPCR, normalized by the quantile method and then filtered to exclude genes that exhibited Ct values .35. The relative expression level of each target gene in infected cells was calculated by using the formula, fold change = 2 2DDC t , where DC t represents the C t (Tc-infected or LPS-treated sample) -C t (control). The LimmaCt routine in the HT-qPCR package was employed for contrast analysis of all groups included in the experiment and for the identification of genes that were overall differentially expressed. LimmaCt utilizes the Linear Model for Microarray data (limma) R package to fit linear models for analyzing designed experiments and for the assessment of differential expressions in microarray data to perform contrast analysis between the different experimental groups. LimmaCt uses Empirical Bayesian methods from the eBayes function in limma to provide stable results even when the number of arrays is small. The basic statistics used for significance analysis consists of moderated t-statistics with the same interpretation as ordinary tstatistics computed for each gene and contrasts, except that the standard errors are shrunk towards a common value by using a Bayesian model. The eBayes function computes moderated F- statistics which combines all of the t-statistics from all the contrasts to calculate the significance of a gene(s). This F-statistic determines if gene(s) are differentially expressed across any contrast. The pvalue is calculated based on the moderated t-statistics and Fstatistics followed by an FDR adjustment. LimmaCt with the top table module was used to gain an appreciation of the change in differential expression of a particular gene over time (or treatment), and significance in changes in gene expression was accepted at p,0.05.
In some experiments, quantitative RT-PCR was performed for IL-1b, IL-18, CXCL1, and TNF-a mRNA levels and house-keeping genes (GAPDH and b-actin) with gene-specific primer pairs (Table S1) and iQ SYBR Supermix on a C1000 Touch Thermal Cycler (Bio-Rad). The relative expression level of each target gene was calculated by using the formula, fold change = 2 2DDC t , where DC t represents the C t (infected sample) -C t (control).

Functional analysis
Datasets for differential expression of the genes included in RT Profiler Human Inflammasome PCR Arrays in human foreskin fibroblasts (HFF), human microvascular endothelial cells Figure 2. Venn diagram of inflammasome-related differential gene expression in mQs infected with T. cruzi (±ATP). THP-1 mQs were incubated with T. cruzi or LPS (6 ATP treatment) as in Fig. 1. Total RNA was isolated, and cDNA was used as a template to probe the expression of 95 genes (including house-keeping genes) in the RT 2 Profiler Inflammasome PCR Arrays. The differential mRNA level was captured by quantitative RT-PCR, normalized to housekeeping genes, and HTqPCR was employed to attain the statistically significant differential expression in treated-versuscontrol samples (Table 1 and Table S2). Shown are Venn diagrams of comparative analysis of gene expression in T. cruzi-infected mQs at 3 h versus 18 h (A), effect of ATP stimulus on gene expression at 3 h (B) and 18 h (C) pi, and comparative effect of ATP stimulus on gene expression in LPStreated mQs at 3 h (D) and 18 h (E). Differential up-regulation (green) and down-regulation (red) of genes with respect to controls is presented. Genes presenting as red with green arrow in B-F showed decreased expression without ATP but were up-regulated by ATP treatment (and vice versa). doi:10.1371/journal.pone.0111539.g002 Table 2. Ingenuity iReport analysis of inflammasome-related datasets in THP-1 macrophages infected by T. cruzi for 3 h or 18 h (6 ATP).

Pathways
Gene name FoCus mols P value

T. cruzi infection of THP-1 mQs vs control (3h)
Top biological and molecular functions  [20]. The selected differentially expressed gene datasets from Tc-infected HFF, HMVEC, and VSMC cells were submitted to Ingenuity iReport Analysis (Ingenuity Systems, Redwood city, CA). The iReport retrieves a set of biological information such as gene name, sub-cellular loCation, tissue specificity, function, assoCiation with disease, and integrates into networks and signaling pathways with biological meaning and significance. An e-value was calculated by estimating the probability of a random set of genes having a frequency of annotation for that term greater than the frequency obtained in the real set, and a threshold of e value ,10 23 was set to retrieve significant molecular functions and biological processes.

Parasite infectivity and replication in mQs
T. cruzi trypomastigotes were labeled with 5 mM SYTO H 11 (binds DNA, Molecular Probes-Invitrogen, Eugene, OR) or 5 mM carboxyfluorescein succinimidyl ester (CFSE, binds amines, Invitrogen) for 20 min at 37oC. THP-1-or BM-derived mQs were infected and incubated with labeled T. cruzi trypomastigotes,  Total DNA from normal and infected cells was isolated by using TRIzol reagent (Life Technologies, Grand Island, NY). Total DNA (100 ng) was used as a template in a quantitative PCR (qPCR) on an iCycler thermal cycler with SYBR Green Supermix (Bio-Rad) and oligonucleotide pairs specific for Tc18S ribosomal DNA (Table S1). Data were normalized to murine or human GAPDH, and fold change calculated as 2 2DDC t , where DDC t represents the C t (sample) -C t (control). Micrographs of DCF or DHE fluorescence were visualized on an Olympus BX-15 microscope, and images were captured by using a mounted digital camera (magnification 40X).
ROS release in supernatants was determined by an Amplex red assay. Briefly, 50 ml of supernatants from infected mQs were added in triplicate to 96-well, flat-bottomed plates, and mixed with a similar volume of 100 mM 10-acetyl 3, 7-dihydroxyphenoxazine (Amplex Red, Life Technologies) and 0.3 U/ml horseradish peroxidase. The H 2 O 2 -dependent oxidation of Amplex red to red fluorescent resorufin (Ex 563nm /Em 587nm ) was recorded as above (standard curve: 50 nM -5 mM H 2 O 2 ) [21].
Culture supernatants from mQs incubated with T. cruzi (6 inhibitors) were also utilized for measuring IL-1b and IL-18 release by using OptEIA ELISA kits (eBioscience), according to the manufacturer's instructions.
Transient transfection and luciferase assay RAW 264.7 cells were plated in 6-well plates, and, when at . 70% confluency, transfected with pGL4.NF-kB-Luc reporter plasmid (3 mg/well, Promega, San Diego, CA) by using JetPEI transfection reagent (Polyplus transfection, New York, NY), according to instructions provided by the manufacturer. The pREP7-Rluc plasmid (500 ng) expressing renilla luciferase was cotransfected into RAW macrophages and used as an internal control reporter. After 30 h of transfection, cells were washed, replenished with complete medium, and infected with T. cruzi (6 inflammasome inhibitors) for 18 h (positive control: 10 ng/ml recombinant TNF-a for 6 h). The relative NF-kB transcriptional activity was detected by using a Steady-Glo luciferase assay system (Promega) and recorded on a luminometer (Turner Biosystems, Sunnyvale, CA).

Statistical Analysis
All experiments were conducted at least twice with triplicate observations per sample per experiment. All data were analyzed by using Graph Pad InStat ver.3 software and expressed as mean 6 SD. Data were analyzed by the Student's t test (comparison of 2 groups) and 1-way analysis of variance (ANOVA) with Tukey's post-hoC test (comparison of multiple groups). Significance is shown by * ,# p,0.05, ** ,## p,0.01, *** ,### p,0.001 (*normalversus-infected; # infected versus infected/treated).

Results
MQs elicit subdued IL-1b response to T. cruzi infection THP-1 mQs incubated with T. cruzi trypomastigotes (1:3, cell: parasite ratio) or Tc-lysate exhibited a ,2-fold increase in IL-1b release at 3 h pi that was consistently increased at 6 h and 12 h pi (data not shown) and maximized to a.9.8-fold increase by 18 h pi (Fig.1A&B). Incubation with higher number of parasites (1:4, 1:5 or 1:6, cell: parasite ratio) did not result in a further increase in IL-1b release at 3 h and 18 h pi (data not shown). LPS treatment (100-ng/ml) triggered a substantially higher level of IL-1b release in THP-1 mQs than was observed with T. cruzi infection, the maximal difference being noted at 3 h (Fig.1A&B). Exogenous addition of ATP, the K + flux agent that can trigger caspase-1 cleavage and inflammasome activation in response to PAMPs, elicited a 2-fold and no increase in IL-1b release in Tc-infected and LPS-treated cells at 3 h (Fig.1A&C). No significant effect of exogenous ATP on IL-1b release was observed at 18 h postincubation (Fig.1B&D) To determine if IL-1b is required for parasite control, mQs were infected with SYTO H 11-labeled T. cruzi and incubated in the presence or absence of anti-IL-1b antibody for 18 h (Fig.1E). The intracellular SYTO H 11 fluorescence (indicates parasite presence) was significantly increased in anti-IL-1b antibody -treated THP-1 mQs (Fig.1E.d&f). The qPCR estimation of parasite burden confirmed the microscopic findings and showed a 2-fold increase in Tc18SrDNA signal in infected cells treated with anti-IL-1b antibody (Fig.1F). Antibody efficacy in depleting secreted IL-1b is shown in Fig.1G. Together, these data suggested that a) mQs respond to T. cruzi infection, in comparison to LPS treatment, by a subdued IL-1b release; b) IL-1b release can be enhanced by ATP at an early time-point pi; and c) IL-1b is required for controlling intracellular T. cruzi.

Inflammasome-related gene expression in mQs infected by T. cruzi
We investigated the mRNA expression levels of various genes involved in the function of inflammasomes and NOD-like receptor (NLR) signaling. THP-1 mQs were incubated with T. cruzi or LPS for 3 h and 18 h (6 ATP treatment) and probed for the expression of 95 genes (including house-keeping genes) utilizing the RT 2 Profiler Inflammasome PCR Arrays. The differential mRNA level was captured by qRT-PCR, and HTqPCR software was employed to attain the statistically significant differential expression in Tc-infected versus control mQs (6 ATP) and LPS-treated versus control mQs at 3 h and 18 h (Table 1 & Table S2). Venn diagrams of comparative analysis of gene expression in Tc-infected and LPS-treated mQs at 3 h and 18 h (6 ATP) are shown in Fig.2. Of the 95 genes that were examined, 63 genes exhibited differential expression in one of the studied groups (p,0.05). We noted the differential expression of 18 (8 up-regulated and 10 down-regulated) and 29 (9 up-regulated and 20 down-regulated) genes in Tc-infected mQs at 3 h and 18 h, respectively, in comparison to that noted in controls ( Fig.2A, Table 1). When Tcinfected mQs were treated with exogenous ATP during the last 30 min of incubation, 22 (15 up-regulated and 7 down-regulated) and 35 (11 up-regulated and 24 down-regulated) of the inflammasome-related genes were differentially expressed at 3 h and 18 h pi, respectively (Fig. 2B & C, Table 1). Only seven and eight genes were uniquely expressed in Tc/ATP-treated mQs when compared to Tc-infected mQs at 3 h and 18 h, respectively. However, other genes, e.g., BCL2, PSTPIP1, and CIITA at 3 h and NAIP1 at 18 h, that were down-regulated by T. cruzi infection, were up regulated when ATP was provided exogenously (Fig.2B&C, Table 1). In LPS-treated mQs, we noted differential expression of 31 (12 up-regulated and 19 down-regulated) and 27 (11 up-regulated and 16 down-regulated) of the inflammasomerelated genes at 3 h and 18 h pi, respectively, and the LPSinduced gene expression profile was not changed by exogenous addition of ATP (Fig.2D&E, Table S2). These data suggest that very few of the inflammasome-related genes are up regulated in response to T. cruzi infection. Exogenous ATP was effective in enhancing the inflammasome-related gene expression at 3 h, but not at 18 h pi. In comparison, LPS served as a potent activator, as evidenced by a significant up regulation of inflammasome-related gene expression within 3 h post-treatment. These data support the results presented in Fig.1 and suggest that T. cruzi is a silent invader that elicits low level of inflammatory response in macrophages.

Functional characterization of differentially expressed genes in Tc-infected cells
The top biological and molecular functional analysis of the differentially expressed gene datasets in Tc-infected mQs by Ingenuity iReport is presented in Table 2. In infected THP-1 mQs at 3 h pi (6 ATP), 13-16 of the differentially expressed genes were indicative of an increase in inflammatory responses and control of macrophages' cell death (p,0.001, z score: 21.195 to 22.333, p,0.001). Exogenous addition of ATP had no significant effect on the top biological functions altered in response to T. cruzi infection. Toxicity analysis indicated that changes in mitoChondrial membrane potential that would likely result in ROS generation would be a key event in inflammatory and cell remodeling/cell death responses (p,0.01). By 18 h pi (6 ATP), 14-21 of the differentially expressed genes were involved in inflammation, activation of leukoCytes, and/or migration of phagoCytes (z score: 1.78 to 2.6) and 24-26 of the differentially expressed genes implicated in cell death were down-regulated (z score: 21.23 to 21.25). The top canonical pathway, toxicity, and upstream regulators analysis suggested the macrophage's attempt to balance inflammation with cellular protection via PPAR-a signaling (p,0.001).

Caspase-1-mediated IL-1b activation is ROS-dependent, and plays a role in control of T. cruzi replication in wild-type mQs
To determine if inflammasomes play a role in IL-1b activation in mQs infected by T. cruzi, we pre-treated the THP-1 mQs with selective inhibitors of inflammasome activation for 2 h and continued the inhibition pressure during the infection period. The IL-1b release induced at 3 h and 18 h pi was decreased by 100% and 75-95%, respectively, when mQs were treated with cycloheximide (inhibits protein synthesis and thus the inducible arm of the inflammasome pathway), glibenclamide (bloCks the caspase-1 and pro-IL-1b maturation), Ac-YVAD-CHO (inhibits caspase-1 activity), or KCl (inhibits K + efflux required for caspase 1 activation) (Fig.3A&B). Importantly, glibenclamide and KCl also bloCked the mQs' ability to control parasite replication, resulting in a 50-70% increase in intracellular Tc18SrDNA, as determined by sensitive qPCR (Fig.3C).
Ingenuity iReport analysis of differentially expressed genes in this study (Table 2) and our previously published reports [22] have led us to suggest that infection by T. cruzi would elicit ROS by changes in mitoChondrial MPT or NOX2 activation in mQs, and that ROS may serve as a 2 nd signal for inflammasome activation [11,23]. We, therefore, determined whether ROS is induced and plays a direct role in IL-1b production in infected mQs. THP-1 mQs exhibited 3.4-fold and 2.7-fold increases in DCF fluorescence (detects intracellular ROS) at 3 h and 18 h pi, respectively (Fig.4A). When NOX2 inhibitors (DPI or apoCynin) were added during incubation with T. cruzi, we noted a 21-44% and 68-80% decline in ROS (Fig.4A&B) and 37% and 45% decline in IL-b release (Fig.4C&D) at 3 h and 18 h pi, respectively. These data suggested that NOX2, at least partially, regulates ROS-dependent IL-1b activation in infected mQs. The observation of a moderate, but a significant (35-45%, p,0.05) decline in DCF fluorescence in infected mQs treated with anti-IL-1b antibody (Fig.4E&F) implied that IL-1b also contributes to activation of ROS production. However, we found no increase in intracellular parasite burden in THP-1 mQs treated with NOX2 inhibitors (Fig.4G). Together, the data presented in Figs.3&4 suggest that a feed-back cycle of caspase-1/IL-1b and ROS activation oCcurs in response to T. cruzi infection in THP-1 mQs and is required for control of intracellular parasite replication. While direct inhibition of IL-1b (Fig.1F) or caspase-1 (Fig.3C) affected the WT mQs ability to control T. cruzi, inhibition of ROS-dependent IL-1b was compensated for, likely by activation of other immune defenses capable of controlling the intracellular parasite replication and survival (Fig.4G).

IL-1b signaling of NFkB in Tc-infected mQs
Because gene expression analysis has identified several of the signaling (e.g. TRAF6, MYD88, NFKBA) and cytokine (e.g. CCL2, CCl5, CXCL1, TNF) molecules involved in inflammatory responses that were activated in Tc-infected mQs (Tables 1 & 2), we determined if IL-1b signals the activation of the nuclear factor kB (NF-kB) pathway of cytokine gene expression in mQs. The mRNA levels for IL-1b, TNF-a and CXCL1 in THP-1 mQs were increased by 1.6 fold, 2-fold, and 2.3-fold, respectively, at 3 h pi. At 18 h pi, IL-1b and CXCL1 were increased by 1.6-fold and 5fold, respectively, while no increase was observed in TNF-a mRNA level (Fig.5A.a-c). Treatment of infected mQs with anti-IL-1b antibody abolished the increase in IL-1b mRNA in infected mQs; possibly indicating that the IL-1b engagement of surface receptors induced the IL-1b mRNA expression.
We performed a luciferase reporter assay to verify the role of IL-1b in signaling cytokine gene expression via NFkB in Tc-infected mQs. RAW mQs transiently transfected with luciferase reporter plasmid pNF-kB-luc with 36NF-kB binding site and pREP7-Rluc (expresses renilla luciferase) were infected with T. cruzi for 18 h, and NF-kB-dependent luciferase activity (normalized to renilla luciferase) was monitored. We observed a .5-fold and 15-fold increase in luciferase activity, respectively, when mQs were infected with T. cruzi or treated with recombinant TNF-a (Fig.5B). Treatment with anti-IL-1b antibody resulted in 70% decline in Tc-induced luciferase activity (Fig.5B, p,0.01). Together, the data presented in Fig.5 suggested that IL-1b signals the NF-kB activation of cytokine gene expression in mQs infected by T. cruzi.
ASC -/-mQs were compromised in ROS-dependent IL-1b activation and NF-kB-dependent cytokine gene expression, and exhibited pronounced T. cruzi replication and survival To determine if ASC is involved in caspase-1-dependent inflammasome formation and activation of inflammatory proCesses for parasite control, we utilized primary BM-derived mQs from ASC -/mice and matched controls. The ASC -/-mQs infected by T. cruzi exhibited a significantly compromised IL-1b release (67% and 40% decline at 3 h and 18 h, respectively) when compared to that noted in matched WT controls (Fig.6A&B, p,0.01). Treatment with anti-IL-1b antibody and ROS scavengers (NAC or apoCynin) normalized the IL-1b production in Tc-infected ASC -/and WT mQs to control levels. Further, the expression of NF-kB-inducible pro-IL-1b and TNF-a mRNAs, that were significantly increased in Tc-infected WT mQs, were completely abolished in Tc-infected ASC -/-mQs (Fig.6C&D). Subsequently, ASC -/-mQs exhibited a 3.2-fold and 2-fold increase in intracellular parasite burden at 3 h and 18 h pi, respectively, when compared to that noted in matched WT infected mQs (Fig.6E&F). As previously observed in THP-1 mQs (Fig.4G), ASC -/-mQs exhibited no further increase in parasite burden upon treatment with ROS scavengers (Fig.6). These data confirmed that the caspase-1/ ASC inflammasomes play an important role in parasite control through ROS-dependent IL-1b activation and expression of other inflammatory cytokines in mQs.

NLRP3 -/-mQs exhibited increased ROS-dependent control of T. cruzi
To specifically determine if NLRP3 inflammasome activation by ASC/caspase-1 is the primary source of protective IL-1b in the context of Tc infection, we conducted further studies in primary bone marrow-derived mQs from NLRP3 -/mice and matched controls. As expected from studies in ASC -/cells, NLRP3 -/-mQs lacked the ability to elicit IL-1b activation in response to T. cruzi infection (Fig.7A&B, p,0.01). Treatment with anti-IL-1b antibody and ROS scavengers (NAC or apoCynin) normalized the IL-1b production in Tc-infected WT mQs to control levels (Fig.7B). Further, NF-kB-inducible pro-IL-1b, TNF-a and CXCL1 mRNA levels that were increased in Tc-infected WT mQs, were almost absent in Tc-infected NLRP3 -/-mQs (Fig.7C&D). However, to our surprise, NLRP3 -/-mQs exhibited a 12-fold decline in intracellular parasite burden at 18 h, when compared to that noted in matched WT infected mQs (Fig.7E).
To determine if NLRP3 -/-mQs were simply refractory to T. cruzi, we incubated the NLRP3 -/and matched WT mQs with CFSE-labeled trypomastigotes for 3 h and analyzed for CFSE fluorescence by flow cytometry. We observed no statistically significant difference in intracellular mean fluorescence intensity (MFI) indicative of the number of parasites/cell (Fig.8A) and the percentage of CFSE + cells indicative of the number of Tc-infected NLRP3 -/and WT mQs (Fig.8B). Representative micrographs showing efficient uptake of CFSE-labeled parasites in NLRP3 -/and WT primary mQs are presented in Fig.8C (panels a&b). These data suggested that NLRP3 -/-mQs were not refractory to T. cruzi and efficiently phagoCytized parasites in a manner similar to that of the WT mQs. Instead, NLRP3 -/-mQs, in comparison to WT mQs, exhibited 8.5-fold higher basal ROS levels (Fig.8D). The ROS production in NLRP3 -/-mQs was enhanced upon T. cruzi infection and not inhibited by the addition of anti-IL-1b antibody (Fig.8D&E). The increase in ROS production in NLRP3 -/-mQs in response to T. cruzi infection was also evidenced by a significant increase in DHE fluorescence (red, detects intracellular ROS, Fig.8C.c-f). Together, the data presented in Figs.7&8 suggested that NLRP3/ASC/caspase-1 inflammasome mediates IL-1b activation and expression of other inflammatory cytokines in mQs infected by T. cruzi; however, NLRP3 deficiency is compensated for by increased ROS levels capable of preventing parasite replication and intracellular survival in NLRP3 -/-mQs.

Discussion
The innate immune response to T. cruzi infection is mediated by diverse PRRs, including receptors of the TLR family [24,25] and ASC-containing inflammasomes (e.g. NOD1 [26], and NLRP3 [27,28]). Genetically modified mice deficient in MYD88 (interacts with TLR-2, -4, -6), TLR-4, NOD1, and ASC exhibited an increased susceptibility to T. cruzi, thus pointing to these PRRs as critical determinants of host resistance to T. cruzi infection [24,26,28,29]. In this manuscript, we have utilized cultured and primary mQs and employed inhibitory approaches to investigate the mechanisms of caspase-1/ASC inflammasome activation and their role in the context of T. cruzi infection. We found that mQs respond to T. cruzi infection with suboptimal activation of inflammasome-related gene expression and IL-1b production. Functional analysis of the differentially expressed gene datasets in Tc-infected mQs was indicative of an increase in inflammatory responses and control of macrophages' cell death. The IL-1b production in Tc-infected mQs was ROS-dependent and could be enhanced by the exogenous addition of ATP. Studies in WT mQs treated with specific inhibitors and ASC -/-mQs suggested that caspase-1/ASC inflammasome played a role in activation of the IL-1b-ROS-NF-kB pathway, that when inhibited, resulted in a compromised inflammatory cytokine response and increase in T. cruzi replication and survival in macrophages. However, NLRP3 -/-mQs, despite an inability to elicit IL-1b activation and inflammatory cytokine gene expression, were capable of parasite control. Thus, our data allow us to conclude that caspase-1/ASC inflammasomes play a significant role in the activation of IL-1b/ ROS and NF-kB signaling of cytokine gene expression for T. cruzi control in human and mouse mQs. However, NLRP3 balances the mQs' activation of ROS and NFkB/cytokine response, and its deficiency shifted the mQs' responses towards increased ROSdependent control of T. cruzi. To the best of our knowledge, this is the first study demonstrating a double-edged role of NLRP3 in determining macrophage activation of ROS and cytokine response, both of which are required for clearance of T. cruzi infection.
The NLRs are expressed in most cell types of the immune system, but are also reported to be expressed in other tissues.
Based upon the expression profile of components of NLRP1, NLRP3, and NLRC1; blood, placenta and thymus are shown to constitutively express inflammasomes. Other tissues (e.g., heart, vascular tissue, bone marrow) require up regulation of one or two components in order to assemble functional inflammasomes [30]. Our data demonstrated that inflammasome-related gene expression is induced in mQs exposed to T. cruzi infection (Table 1) or LPS treatment (Table S2), as well as in non-phagoCytic, human vascular smooth muscle, fibroblast, and microvascular endothelial cells (Table S3). Functional analysis of the gene expression profile indicated that in mQs, an inflammatory response to control T. cruzi infection was assoCiated with significant efforts to prevent cell death (Table 2). Infected mQs also exhibited suppression of several of the genes involved in PPARa signaling that induces apoptosis following activation with TNF-a/IFN-c [31,32]. In non-phagoCytic cells, TLR/MYD88 signaling of NFkB-dependent cytokines' (IL-6, IFNa/b, IL-12) gene expression and caspase-1/ NLRP3-mediated activation of IL-1b gene expression was noted. Canonical analysis, as in mQs, also indicated PPAR-a regulation of gene expression in HMVEC and HFF cells infected by T. cruzi (Table S4). Our observations allow us to surmise that phagoCytic and non-phagoCytic cells responded to T. cruzi infection by induction of diverse inflammasome-related gene expression. Eventually, all cell types appeared to be overwhelmed by the inflammatory signal and tended to switch to signaling events related to prevention of cell death.
Several studies indicate that ROS are an essential secondary messenger for signaling caspase-1/ASC inflammasome activation [11,33]. The use of ROS scavengers controlled IL-1b activation by virtually all agonists of the NLRP3 inflammasome [33,34]. Besides ROS, extracellular ATP, through activation of the P2X7 (purogenic ionotrophic ATP-gated cation channel), triggers K + efflux, which, in turn, triggers pore formation by pannexin, thereby allowing the delivery of pathogen products into the cytosol, resulting in caspase-1/ASC inflammasome activation [35]. In agreement with the literature, we also observed ROSdependent IL-1b activation mediated by caspase-1/ASC inflammasome in mQs exposed to T. cruzi (Figs.1-4). The exogenous addition of ATP resulted in a higher level of IL-1b activation in infected mQs at 3 h pi, as also noted by others [28]. The lack of an effect of exogenous ATP on the extent of IL-1b activation in infected mQs at 18 h pi may mean that other signals were generated. We speculate that either Tc kDNA released by dying T. cruzi within mQs or mtDNA released due to bystander damage in infected mQs served as a secondary signal in activating caspase-1/ ASC inflammasomes at 18 h pi. This notion is supported by the observation of similar levels of IL-1b activation in mQs incubated with live as well as dead T. cruzi and requires further investigation.
Importantly, we made a novel observation and noted a feedback cycle of IL-1b signaling of ROS activation in infected mQs (Figs.3&4). The molecular mechanism for the ROS production by IL-1b remains to be elucidated. In previous studies, IL-1b was found to stimulate phospholipase A 2 , promoting release of arachidonic acid. Since arachidonic acid can activate NAPDH oxidase to produce superoxide, it is possible that this fatty acid may serve as an intermediate in the IL-1b-induced activation of enzymes, leading to the production of ROS [36,37]. However, several non-NADPH oxidase-dependent sources, including mito-Chondrial electron transport and arachidonate metabolism, may also be involved in the cytokine-induced ROS generation, to be investigated in future studies.
It is intriguing that both ASC -/and NLRP3 -/-mQs were equally invaded by T. cruzi, lacked IL-1b production, and elicited ineffective, NF-kB-mediated cytokine gene expression (Figs.6-8); yet only ASC -/-mQs were restrictive in their capacity to control T. cruzi infection (Fig. 6). ASC -/and caspase-1 -/mice have been doCumented to exhibit a higher incidence of mortality, cardiac parasitism and heart inflammation, meaning that ASC/caspase-1 inflammasomes are critical determinants of host resistance to infection with T. cruzi [28]. Like ASC -/mice and in vitro cultured ASC -/-mQs, NOD1 -/mice and BM-derived derived mQs from NOD1 -/mice also showed an impaired induction of NF-kBdependent products and failed to restrict T. cruzi infection [26]. Our finding that a deficiency of NLRP3 did not affect mQs ability to control parasites allows us to surmise that the caspase-1/ASC requirement for effective control of T. cruzi is delivered via formation and activation of inflammasomes with other NLRs, e.g., NLRP1, AIM2 or NLRC4. The ability of NLRP3 -/-mQs to efficiently manage parasite killing via enhanced ROS release (Figs.7&8) suggest that NLRP3 suppresses NOX2-dependent ROS production in mQs. Though it is shown that activation of NLRP3 containing inflammasome is not dependent on the function of NOX1-4 [38], our data provide the first evidence that NLRP3 dysregulates the NOX2 function at the superoxide production level and suppresses the mQs' ability to kill T. cruzi. Caspase-l, upon activation by NLRP3/ASC, loCalizes to phagosomes and disturbs NOX2 control of pH, thereby triggering acidification and microbicidal activity of phagosomes in mQs infected by Staphy-loCoCcus aureus [39]. We surmise that caspase-1 loCalization to phagosome and suppression of NOX2/ROS enhances the bactericidal activity of mQs. However, NLRP3 interaction with NOX2 resulting in low level of ROS production likely maintains the alkalization of the phagosomal lumen that prevents the mQs' ability to directly kill T. cruzi, but plays a critical role in allowing mQs to function as specialized phagoCytes adapted to proCess antigens for cross presentation and elicitation of adaptive immunity.
Other investigators have shown a severe defect in nitric oxide (NO) production and impairment in mQ-mediated T. cruzi killing in NLRP3 -/mice and isolated mQs [27]. In this study, we utilized the BM-derived monoCytes from NLRP3 -/mice that were differentiated to mQs by M-CSF treatment, and infected with T. cruzi (SylvioX10) for 3 h and 18 h. Others utilized peritoneal mQs obtained 4 days after intra-peritoneal injection of 1% starch solution in NLRP3 -/mice and infected these mQs with the Y strain of T. cruzi for 48 h for all the studies [27]. We propose that NLRP3 deficiency is compensated for by over activation of NOX2/ROS that effectively controlled the early invasion and replication of T. cruzi in mQs, as observed in this study. However, a lack of NF-kB-mediated cytokine response (Fig.7) and iNOS/ NO activation [27] prevented the NLRP3 -/-mQs from sustained, long-term control of the parasites, resulting in increased susceptibility. It is also plausible that the differences in the source of mQs, parasite isolates, and the time course of infection may explain the observed differences in the ability of NLRP3 -/-mQs to control T. cruzi infection in our and published studies.
In summary, we have demonstrated that T. cruzi interfered with the potent activation of caspase-1/ASC inflammasome-related gene expression and cytokine response in mQs to ensure its survival. We found that caspase-1/ASC inflammasomes played a role in the activation of the IL-1b-ROS-NF-kB pathway, that when inhibited, resulted in an increase in T. cruzi replication and survival in mQs. However, NLRP3 -/-mQs were compensated for by increased NOX2/ROS activation capable of parasite killing. Our data suggest that the NLRP3/caspase/ASC inflammasome balances the mQs' activation of ROS and NFkB/cytokine response and provide the first evidence for NLRP3 regulation of NOX2 function as an effector mechanism contributing to parasite persistence.

Supporting Information
Table S1 Oligonucleotides used in this study.

(DOCX)
Table S2 Inflammasome-related differential gene expression in THP-1 macrophages in response to LPS treatment (± ATP) in comparison to normal controls. (DOC)