Lack of PPARγ in Myeloid Cells Confers Resistance to Listeria monocytogenes Infection

The peroxisomal proliferator-activated receptor γ (PPARγ) is a nuclear receptor that controls inflammation and immunity. Innate immune defense against bacterial infection appears to be compromised by PPARγ. The relevance of PPARγ in myeloid cells, that organize anti-bacterial immunity, for the outcome of immune responses against intracellular bacteria such as Listeria monocytogenes in vivo is unknown. We found that Listeria monocytogenes infection of macrophages rapidly led to increased expression of PPARγ. This prompted us to investigate whether PPARγ in myeloid cells influences innate immunity against Listeria monocytogenes infection by using transgenic mice with myeloid-cell specific ablation of PPARγ (LysMCre×PPARγflox/flox). Loss of PPARγ in myeloid cells results in enhanced innate immune defense against Listeria monocytogenes infection both, in vitro and in vivo. This increased resistance against infection was characterized by augmented levels of bactericidal factors and inflammatory cytokines: ROS, NO, IFNγ TNF IL-6 and IL-12. Moreover, myeloid cell-specific loss of PPARγ enhanced chemokine and adhesion molecule expression leading to improved recruitment of inflammatory Ly6Chi monocytes to sites of infection. Importantly, increased resistance against Listeria infection in the absence of PPARγ was not accompanied by enhanced immunopathology. Our results elucidate a yet unknown regulatory network in myeloid cells that is governed by PPARγ and restrains both listeriocidal activity and recruitment of inflammatory monocytes during Listeria infection, which may contribute to bacterial immune escape. Pharmacological interference with PPARγ activity in myeloid cells might represent a novel strategy to overcome intracellular bacterial infection.


Introduction
The peroxisome proliferator-activated receptor-gamma (PPARc) is a member of the nuclear hormone-receptor superfamily of ligand-activated transcription factors [1]. It is expressed in different cell types of the immune system, e.g. macrophages/ monocytes [2] and lymphocytes [3]. Upon ligand binding, PPARc heterodimerizes with the retinoid X receptor (RXR) and binds to PPAR response elements located in the promoter region of metabolic target genes [4]. Besides its well-studied role in metabolism and cellular differentiation, PPARc is a negative regulator of inflammatory gene expression and macrophage activation [5,6]. PPARc exerts its anti-inflammatory effects in part by trans-repression, i.e. negative interaction with proinflammatory transcription factors like NFkB, or by stabilization of co-repressor complexes such as SMRT or NCoR on promoters of target genes [5]. Among the genes targeted by the inhibitory function of PPARc are pro-inflammatory cytokines and chemokines but also lineage-determining transcription factors, such as RORct that promotes differentiation of pro-inflammatory TH17cells [7]. PPARc is known to have cross-regulatory function in inhibiting innate immune stimulation elicited through ligand binding to Toll like receptors [8]. Moreover, we and others have demonstrated a negative influence of PPARc on the immune stimulatory capacity of dendritic cells (DC) [9,10]. Inhibition of PPARc in myeloid cells led to induction of systemic inflammation even in the absence of challenge with an infectious agent [11]. There is a large number of endogenous ligands for PPARc that primarily are derived from arachidonic acid metabolism and are in part induced by immune mediators such as IL-4 and IL-13 [12]. Pharmacological stimulation of PPARc has been shown to lead to an increased occurrence of bacterial infections in patients [13], suggesting that PPARc plays a key role in anti-bacterial defense. Also, certain gram-positive bacteria such as Mycobacterium tuberculosis have been shown to increase the expression of PPARc [14,15]. However, there is no study on the relevance of PPARc in myeloid cells for the outcome of bacterial infection, although it is clear that myeloid cells are the key cell population in anti-bacterial defense.
Listeria monocytogenes is a facultative intracellular Gram-positive bacterial pathogen. Infection of humans and animals can lead to serious, often fatal disease. In humans, disease is most common among pregnant women, newborns, and immune compromised individuals [16]. Murine listeriosis is used widely as a model to study the immune response against intracellular bacterial infection [17]. Early after infection with L. monocytogenes, neutrophils and inflammatory monocytes phagocytose and kill invading bacteria. These phagocytic cells secret TNF and IL-12 [18,19,20], that activate NK cells to produce IFNc, which in turn activates bactericidal effector functions of other phagocytes, such as oxidative burst and production of nitric oxide (NO) [21,22]. Accordingly, mice lacking these essential inflammatory mediators and mice lacking p47phox oxidase or iNOS, which are deficient in ROS or NO production, all display enhanced susceptibility to L. monocytogenes infection [23,24]. Recruitment of inflammatory monocytes to the site of infection, that is driven by the chemokines MCP1 (CCL2) and MCP3 (CCL7), is an important mechanism supporting development of innate immunity against L. monocytogenes infection [25]. Mice lacking CCL2 or CCR2 exhibit diminished inflammatory monocyte recruitment to infection sites resulting in enhanced bacterial growth and overwhelming infection [26].
It has remained unclear whether PPARc with its potent regulatory activity on innate immune functions in myeloid cells plays a regulatory role during infection with L. monocytogenes. Here we address this question and provide clear evidence that PPARc restricts innate immunity in myeloid cells against Listeria infection. As infection with L. monocytogenes leads to increased expression of PPARc our results reveal a so far unrecognized regulatory network in myeloid cells that may be abused by L. monocytogenes to counter innate immunity.

Mice
Wild-type C57BL6/J and PPARc flox/flox mice [27] were purchased from Charles River, LysM-Cre, IFNc 2/2 , TNF 2/2 and CCR2 2/2 transgenic mice in C57BL6/J background were previously described [28,29,30,31]. All animal studies were approved by local authorities. Animals were bred and kept under specific pathogen-free conditions and used at 8-12 weeks of age in accordance with local animal experimentation guidelines. LysMspecific PPARc knockout mice (LysM-PPARc KO mice) were generated by crossing PPARc flox/flox mice with LysM-Cre +/2 transgenic mice expressing Cre-recombinase under control of the LysM promoter. These mice display no alterations in immune cell frequencies.

Infection of mice and cells
If not indicated otherwise, mice were infected intraperitonealy with 2610 4 CFU of wild type Listeria monocytogenes (EGDe strain). Cells were infected in antibiotic free medium with Listeria monocytogenes at a MOI of 10 for one hour followed by one step wash with PBS and were further cultured in medium containing 50 mg/ml Gentamycin.

Uptake of fluorochrome labeled Listeria
Murine BMDM or PEC (10 6 cells) were infected with FITClabeled listeria (MOI 5 or 10) 1 ml medium. After 30 minutes of incubation, cells were washed with PBS followed by passage through a 30% sucrose layer to get rid of extracellular bacteria. Uptake efficiency was assessed by flow cytometry following staining with anti-Listeria antibody to discriminate between intracellular and extracellular Listeria. To determine the infection efficiency of the infected BMDM and PEC, cell lysates were plated on BHI plated and incubated over night at 37uC and CFU/cell was measured. FITC-labeled bacteria were obtained by incubation of wild type Listeria monocytogenes for 30 min at 37uC in PBS containing 5 mM FITC (5,6)-fluorescein isothiocyanate mixed isomer (Thermo Scientific)) at a density of 1610 9 CFU/ml followed by three times wash with PBS.

Cells and cell culture
Peritoneal exudates cells (PEC) are collected from the peritoneal cavity of mice by peritoneal lavage using PBS. To achieve macrophage-rich exudates, mice were injected intraperitoneally with 1 ml of 3% thioglycolate and PEC were collected 72 hrs later. Cells were then let to adhere for 1 hr and non-adherent cells were collected. Bone marrow derived macrophages from wild-type or mutant mice were generated by collecting cells from mice tibiae and femurs and culture them in RPMI 1640 medium supplemented with FCS, Glutamate, bME and 30% L929 cells conditioned medium for 7 days. Cells were cultured and infected as previously described [32]. Importantly, for all experiments BMDM were cultured in antibiotic-free medium prior to infection or adoptive transfer studies to eliminate remaining antibiotics that may interfere with bacterial growth. The purity of BMDM generated by the protocol used in this study was routinely .95% as shown by a representative analysis of CD11b and F4/80 expression (Fig. S1).
Transfer of BMDM from LysM-PPARc WT and LysM-PPARc KO BMDM were generated as above from LysM-PPARc WT and LysM-PPARc KO . On days 6, medium was replaced with fresh medium lacking antibiotics. On day 7 of differentiation, 5610 6 cells were injected into the peritoneal cavity of IFNc2/2, TNF2/2 or CCR22/2 mice. Two hours later mice were infected with 2610 4 CFU/ml of Listeria monocytogenes.

Immunohistochemistry and immunofluorescence microscopy
Murine BMDM (5610 5 ) were seeded on glass coverslips in 24well plates overnight in antibiotic free medium. Cells were then infected with L. monocytogenes at MOI of 10 for 30 min. Cells were then washed with pre-warmed medium and culture further in medium containing 50 mg/ml Gentamycin (Sigma). At indicated time points after infection, cells were washed three times with PBS and were fixed with 4% paraformaldehyde, permeablized with 0.1% Saponin. Cells were incubated with anti-PPARc in PBS containing 10% FCS and 0.1% Saponin for 1 h. After three washes with PBS, cells were incubated for 1 h at room temperature with FITC-labeled secondary antibody. After that, cells were washed three times with PBS and mounted for fluorescence microscopy. Cells were viewed with a 60 times oil objective lens. Microscopy analysis was performed with an Olympus FluoView IX71Microscope (Olympus).

Cell culture and infection prior gene expression profiling
Human macrophages were generated from peripheral blood mononuclear cells (PBMC) obtained by Pancoll (PAN-Biotech, Aidenbach, Germany) density centrifugation from buffy coats of healthy donors. CD14 + monocytes were then isolated from the PBMC using CD14-specific MACS beads (Miltenyi Biotec) according to the manufacturers protocol (routinely .95% purity).
CD14 + monocytes were cultured in 6-well plates in RPMI1640 medium containing 10% FCS and differentiated into immature macrophages using GM-CSF (500 U/ml; Immunotools, Friesoythe, Germany) for 72 hours. Only cultures with .95% CD14 + CD68 + macrophages purity were used for further analysis. Cells were infected with wild type Listeria monocytogenes (EGDe strain) using a MOI of 10. 24 hours after infection cells were washed twice and lysed in Trizol (Invitrogen Life Technologies, USA) prior RNA isolation. Only cultures with a cell viability of .80% were used for further analysis.
Microarray procedure RNA sample amplification, labeling and hybridization on Illumina HT12 Sentrix BeadChips V4 were performed according to the manufacturer's instructions using an Illumina HiScan SQ. (Table S1) summarizes the performed microarray experiments.

Bioinformatics Analysis
Data analyses were performed using Illumina Genome Studio and Partek Genomics Suite (PGS). Datasets were generated in Illumina and exported to PGS following the manufacturer's instruction. In PGS the quantile method was used for data normalization. Differentially expressed genes were determined using ANOVA. To identify PPARc target genes regulated during infection of macrophages with Listeria a publically available dataset (GSE21314) assessing genome-wide PPARc binding to DNA by ChIP-Seq was downloaded from GEO. A table of genes most closely located to PPARc binding sites in macrophages was  Table S2). The analysis was done with the ChIP-Seq data analysis tool HOMER (Hypergeometric Optimization of Motif EnRichment). This list of genes was used to filter genes differentially regulated in macrophages infected with Listeria in contrast to non-infected macrophages (Table S3). For visualization of the most differentially expressed PPARc target genes (n = 80) hierarchical clustering was performed using PGS and plotted as a heatmap. The complete dataset is available at GEO (GSE34103).

Serum alanine aminotransferase (ALT) determination
Serum ALT was analyzed from whole blood using ALT strips from Roche according to the manufacturers instructions. Measurement was performed in a Reflovet machine from SCIL animal care.

Flow cytometry
We used the following fluorochrome-labeled monoclonal antibodies from e-Bioscience: Anti-CD11b-PerCp-cy5.5 (M1/ 70), F4/80-FITC (BM8), Ly6C-APC (HK1.4), CD11b-Pacific Blue (M1/70) and CD3-PE (G4.18) and the live/dead-APC-Cy7 (near-red) cell staining from Invitrogen. To determine the absolute cell numbers fixed numbers of CaliBRITEH Beads (BD) were added to each sample before analysis as internal reference. Analysis, was performed on a FACSCanto (BD) and analyzed data with Flow-JoH software (Tree Star). For flow cytometric analysis of intrahepatic leukocytes, livers were perfused with 20 mL 400 mg/ ml collagenase type-IV (Sigma) in Hank's balanced salt solution (HBSS), minced with scissors and subsequently digested for 15 minutes with 400 mg/ml collagenase type-IV in HBSS at 37uC. Digested extracts were pressed through 70-mm cell strainers to gain single-cell suspensions. Liver single cell suspension was subjected to density gradient centrifugation (25% vs 50% PercollH (GE Healthcare)) at 2000 rpm for 20 minutes at 25uC. Leukocytes were collected from the interphase after centrifugation, washed twice with HBSS containing 2% bovine serum albumin and subjected to antibody staining for FACS analysis.

Determination of cytokines, NO (Nitrite) and ROS
IFNc, TNF, IL-6, IL-12 were determined in supernatants of cell culture or sera of mice by ELISA using purified and biotinylated antibodies to IFNc, TNF, IL-6 and IL-12 (e-Bioscience). Nitrite concentrations were determined in the supernatant of cells 24 hrs prior to infection using the Griess Reagent Kit (Invitrogen) according to the manufacturer's protocol. Reactive oxygen species (ROS) production was determined using OxyBURSTH Green H2DCFDA (Life Technologies) according to the manufacturer's protocol.

Quantitative real-time PCR analysis
RNA from macrophages was extracted using the RNeasy mini kit (QIAGEN). Reverse transcription of RNA into cDNA was performed using SuperScript III (Invitrogen). Quantitative Real Time PCR (TaqMann) for PPARc, IFNc, TNF, IL-6, IL-12, MCP-1, MCP3, CCR2 and GAPDH expression were performed using pre-designed primers and probes (Gen Expression Assay) from Applied Biosystems on an ABiPrism 7900 HT cycler (Applied Biosystems) according to the manufacturer's instructions. All gene expression data are presented as relative expression to GAPDH.

Ethics statement
The animal experiments within this manuscript were performed according to the guidelines for animal care enforced by the EU (EU Directive 86/609/EEC) and the state Northrhein Westphalia. The study protocol was approved by the local authorities of the state Northrhein Westphalia (8.87-50.10.31.09.033). Ethics approval for obtaining peripheral blood from healthy volunteers was given by the local ethics committee at the University Hospital Bonn and blood was only drawn from healthy donors after written consent was obtained.

Enhanced expression of PPARc following Listeria infection in macrophages
As PPARc is known to be up-regulated during infection of macrophages with Mycobacteria spp. or epithelial cells with Salmonella spp. [14,33], we wondered whether infection of macrophages with Listeria monocytogenes would also increase PPARc expression. We observed a rapid and pronounced increase in the expression of PPARc within 30 minutes after infection, as shown by immunoblot and immunofluorescence analysis of infected human monocytes (Fig. 1A). To address the question whether such increase in PPARc levels were accompanied by transcriptional changes indicative of PPARc activation, we analyzed the entire transcriptome of Listeria-infected human monocytes at a later time point (24 hrs p.i.) compared to non-infected monocytes. We mapped target genes of PPARc in monocytes recently published as ChIPseqencing data [34] onto the transcriptome of macrophages infected with L. monocytogenes for a longer time period. Of the 891 genes with significant increase in gene expression (FC.2, p,0.05) after infection, 125 genes were PPARc target genes (Table S2). Of the 883 significantly suppressed genes after L. monocytogenes infection 100 genes were previously identified to be PPARc targets. The most differentially expressed genes are shown as a heatmap after hierarchical clustering (Fig. 1B). We confirmed these findings for bone marrow derived murine macrophages (BMDM). Already 30 minutes after infection with L. monocytogenes we observed an increase in PPARc expression that was detected by Western Blot and immunohistochemistry (Fig. 1C,D and suppl. Fig. 2D). Taken together, these results strongly suggested that Listeria infection not only led to increased expression of PPARc but also increased activity as transcriptional regulator.
Myeloid cell-specific PPARc-ablation augments survival of mice after L. monocytogenes infection Given the early increase of expression of the anti-inflammatory factor PPARc as well as the differential expression of important PPARc target genes in Listeria-infected macrophages, we next characterized the functional role of PPARc in myeloid cells in host defense against infection with L. monocytogenes. To this end, we generated a transgenic mouse with a myeloid cell-specific knockout of PPARc by crossing LysM-Cre mice with mice carrying loxP sites within the PPARc gene (LysMCre6PPARc fl/fl = LysM-PPARc KO ) ( Fig. S2A-C). The LD 50 for Listeria infection in wildtype littermates (LysM-PPARc WT ) was identical to that for wildtype C57Bl/6 mice (2610 4 CFU, data not shown). LysM-PPARc WT and LysM-PPARc KO mice were challenged with this dose of L. monocytogenes and survival was determined. LysM-PPARc KO mice showed an improved survival compared to wildtype littermates ( Fig. 2A), indicating that absence of PPARc in myeloid cells enhances innate immune defense against Listeria infection. Almost 80% of LysM-PPARc KO mice even resisted infection with 20-fold higher numbers of bacteria, a dose where all wildtype littermates succumbed to infection ( Fig. 2A). However, when infected with very high numbers of bacteria (200 fold LD 50 ), LysM-PPARc KO mice succumbed faster than their wildtype littermates ( Fig. 2A), revealing a limitation of the gain of protective myeloid cell function upon ablation of PPARc.
Next, we determined whether the improved resistance to L. monocytogenes infection in LysM-PPARc KO mice was associated with an improved clearance of infecting bacteria. Time kinetic analysis showed already at d1 after infection (2610 4 CFU), that LysM-PPARc KO mice better controlled bacterial infection than LysM-PPARc WT (Fig. 2B and C). In LysM-PPARc WT mice, little if any decline in bacterial numbers was observed until d4 p.i. In contrast, in LysM-PPARc KO mice the numbers of L. monocytogenes in liver and spleen decreased over the entire time period investigated. At d3 p.i., the bacterial burden in the spleen of LysM-PPARc KO mice was 100 fold lower compared to LysM-PPARc WT mice (Fig. 2C). Interestingly, increased clearance of bacteria was not accompanied by an exaggerated immune reponse against infected tissue, because only a mild elevation of serum ALT levels was observed in LysM-PPARc KO mice infected with L. monocytogenes (LD50 and LD80) (Fig. S3). Taken together, these Loss of PPARc in myeloid cells enhances bactericidal activity and expression of pro-inflammatory mediators after L. monocytogenes infection in vitro As we observed improved bacterial clearance in LysM-PPARc KO mice we next investigated the influence of PPARc on molecular mechanisms known to be critical for control of bacterial infection, such as phagocytosis and generation of listeriocidal mediators like reactive NO and ROS [16,21]. We did not detect any differences in the extent of phagocytic uptake or the percentage of phagocytic macrophages isolated from bone marrow or peritoneal cavity of either LysM-PPARc KO or LysM-PPARc WT mice ( Fig. 3A and B). Similarly, no difference in phagocytic uptake of fluorescently labeled L. monocytogenes was observed ( Fig. 3A and B), thus excluding that increased phagocytosis by PPARc KO macrophages was responsible for improved bacterial clearance. Instead, PPARc KO macrophages more efficiently controlled the growth of intracellular L. monocytogenes than PPARc WT macrophages (Fig. 3C). Along this line, we observed that macrophages from LysM-PPARc KO produced significantly higher level of the listeriocidal mediators ROS and NO2-after infection with L. monocytogenes compared to macrophages from LysM-PPARc WT (Fig. 3D and E). As ROS production was more pronounced in PPARc-deficient macrophages already early during bacterial infection (Fig. 3D) this may explain the improved subsequent early control of intracellular bacterial growth. These results were corroborated by the observation that iNOS expression was also enhanced at early time points in Listeria-infected PPARc-deficient macrophages (Fig. 3F), findings that are in line with reports that PPARc controls expression of pro-inflammatory and bactericidal mediators by interfering with NF-kB signaling [4,6,35].

Absence of PPARc in myeloid cells augments production of pro-inflammatory mediators upon L. monocytogenes infection
Myeloid cells secrete pro-inflammatory mediators, which are essential for innate immune responses against Listeria infection [36,37]. To further characterize the role of PPARc in myeloid cells during L. monocytogenes infection, we determined expression of relevant pro-inflammatory cytokines after L. monocytogenes infection in vivo. We observed increased expression of IFNc, TNF, IL-12 and IL-6 in liver and spleen of infected mice (Fig. 4A and B) and cytokine expression was found to be significantly higher in organs of PPARc KO mice at d1-d3 p.i. when compared to PPARc WT mice ( Fig. 4A and B). In contrast, at d4 p.i. there was a pronounced reduction of IFNc and TNF in PPARc KO compared to PPARc WT mice ( Fig. 4A and B), which can be explained by the rapid clearance of L. monocytogenes from LysM-PPARc KO mice as demonstrated before. The early increased induction of proinflammatory cytokines was also observed in peritoneal macrophages isolated from PPARc KO mice (Fig. S4), thus confirming the control function of PPARc for these cytokines.
To provide further evidence for the role of PPARc in myeloid cells, we analyzed the production of inflammatory cytokines in L. monocytogenes infected macrophages derived from PPARc KO or PPARc WT mice. Following L. monocytogenes infection in vitro, macrophages from PPARc KO mice produced more IFNc, TNF, IL-6 and IL-12 compared to macrophages from PPARc WT mice (Fig. 5A). These results demonstrate an important cell-intrinsic control function of PPARc for pro-inflammatory cytokine production in myeloid cells after infection with L. monocytogenes. To investigate whether PPARc-mediated control of macrophagederived inflammatory cytokines was relevant for bactericidal activity of neighboring cells, macrophages from wildtype C57BL/6 mice were infected with L. monocytogenes and cultured in the presence of conditioned medium from Listeria-infected peritoneal macrophages isolated from LysM-PPARc KO or LysM-PPARc WT mice. While wildtype macrophages incubated in conditioned medium from infected PPARc KO macrophages were able to restrict the growth of intracellular Listeria immediately after infection, macrophages cultured in conditioned medium from wild type macrophages did not rapidly control Listeria growth (Fig. 5B).
These results indicate that ablation of PPARc in myeloid cells also functioned in a paracrine fashion to increase anti-bacterial immunity in neighboring cells by proinflammatory cytokines.

Deletion of PPARc in myeloid cells augments recruitment of inflammatory monocytes to sites of L. monocytogenes infection in vivo
The recruitment of inflammatory monocytes that produce TNF and NO for control of Listeria infection has been demonstrated to be crucial for anti-bacterial innate immune defense [38]. Clearly, the recruitment of Ly6C high CD11b + inflammatory monocytes to sites of infection was increased from d2 p.i. in LysM-PPARc KO mice compared to LysM-PPARc WT mice ( Fig. 6A and Fig. S5). The increased number of inflammatory monocytes in LysM-PPARc KO mice was not caused by reduced apoptosis as measured by expression of activated Caspase 3/7, bax, bcl2 by Western blotting (Fig. S6). As recruitment of these cells requires chemokines such as CCL2, CCL7 [23,25] and adhesion molecules such as CD54 [39], we investigated the expression of these molecules in LysM-PPARc KO mice. We detected a similar upregulation of these genes in lysates from liver (Fig. 6B) and spleen (Fig. 6C) of LysM-PPARc KO mice. Consistent with this finding, in myeloid cells isolated from the peritoneal cavity after L. monocytogenes infection we found a significant increase in gene expression for CCL2 and CCL7 as well as the corresponding chemokine receptor CCR2, which was most pronounced at d1 and d2 after infection (Fig. S7). As entry of inflammatory monocytes into infected liver requires CD54 expression on liver endothelium [39], we exposed primary liver sinusoidal endothelial cells in vitro to supernatants of Listeria-infected macrophages from PPARc KO or LysM-PPARc WT mice. There was a significant induction of CD54 expression on LSEC when exposed to supernatants of Listeria-infected but not non-infected macrophages from PPARc KO mice (Fig. 6D), which suggests that macrophages from PPARc KO mice secrete mediators such as TNF and IFNc that act on LSEC to upregulate CD54 expression. These results indicate that PPARc in myeloid cells controls both, chemokine as well as adhesion-molecule driven recruitment of inflammatory monocytes to sites of infection.  PPARc ablation renders myeloid cells more competent to control L. monocytogenes infection in vivo Finally, we elucidated the relevance of PPARc-mediated regulation in myeloid cells to clear L. monocytogenes infection in vivo. As the soluble mediators IFNc and TNF are crucial to control infection of mice deficient for either of these mediators that rapidly succumb to infection [40,41], we chose to adoptively transfer macrophages from LysM-PPARc KO or LysM-PPARc WT into the peritoneal cavity of IFNc 2/2 or TNF 2/2 mice prior to infection of the mice. All knockout mice succumbed on day 2 p.i. and those mice that received wildtype macrophages succumbed on day 3 p.i. (Fig. 7A). In contrast, 75-80% of IFNc 2/2 or TNF 2/2 mice that received PPARc KO macrophages survived until the end of the observation period (Fig. 7A). Direct comparison of the bacterial load in organs from these mice at d2 p.i. revealed that only IFNc 2/2 or TNF 2/2 mice receiving PPARc KO macrophages had lower levels of bacteria in liver and spleen compared to mice receiving PPARc WT macrophages (Fig. 7B). These results clearly demonstrate that ablation of PPARc renders myeloid cells more competent to control infection in vivo presumably through the increased expression of soluble mediators as transfer of these cells restored resistance towards Listeria infection in cytokine-deficient animals. To address the question whether the gain of function upon PPARc ablation in macrophages is already sufficient to improve control of Listeria infection in the absence of inflammatory monocytes, we adoptively transferred PPARc-deficient macrophages into CCR2 2/2 mice [25]. Under these experimental conditions, we also observed reduced numbers of bacteria in the liver and spleen (Fig. 7C), which indicates that the improved function of PPARc-deficient macrophages alone is already sufficient to control infection. Collectively, we report here the discovery of a so far unknown regulatory network in myeloid cells that is governed by PPARc and which plays a key role in the control of innate immune responses to L. monocytogenes infection.

Discussion
The immune response against infection with intracellular bacteria is orchestrated by myeloid cells [16,36,37]. These cells initiate and amplify innate immune responses leading to immediate reduction of bacteria or infected cells, and also serve as important link for the induction of adaptive immunity to achieve long-lasting protective immunity. Here, we have investigated the role of the potent anti-inflammatory transcription factor PPARc in regulating the early innate immune response of myeloid cells against infection with L. monocytogenes. Our results identify a so far unknown central role of PPARc in the coordination of listeriocidal functions and recruitment of inflammatory monocytes to the site of infection.
The expression of PPARc has been shown to be augmented in macrophages after infection with pathogenic bacteria such as Mycobacterium tuberculosis and Salmonella enteridis [14,15,33]. Here, we demonstrate that also the intracellular bacterium L. monocytogenes increases PPARc expression rapidly within 30 minutes after infection of primary murine macrophages. The molecular mechanism underlying increased expression of PPARc remains unclear and will require further investigation. It is important to note, however, that this induction of a regulatory transcription factor by Listeria infection occurs immediately upon infection and before an inflammatory response was mounted. This is clearly different from immunoregulatory mediators such as IL-10, which typically are expressed after inflammation has been induced and serves in many infections to restrict detrimental immunopathology [42]. Although there is evidence that autoactivation may occur if large numbers of PPARc are present in a cell [43], PPARc is a ligand-activated receptor raising the question whether increased expression levels were also associated with PPARc-activation. Gene expression analysis of Listeria monocytogenes infected macrophages revealed regulation of PPARc responsive gene sets indicating that PPARc has likely undergone ligand-activation after infection. It is possible that other cell populations such as mast cells through production of IL-4 may contribute to increased production of endogenous PPARc ligands in vivo [44]. However, our data demonstrating increased bactericidal function of PPARc KO compared to PPARc WT macrophages against Listeria infection in vitro suggest that there is also induction of endogenous PPARc ligands in macrophages themselves. It appears likely that L. monocytogenes infection induced increased expression of endogenous ligands as we have previously reported that following Listeria infection of myeloid cells COX2 expression was induced [45], that leads to generation of prostanoids that can serve as PPARc ligands.
Given the potent anti-inflammatory effects of PPARc by transrepression of NF-kB and stabilization of co-repressor complexes on promoters of pro-inflammatory genes [1], such induction of PPARc in macrophages upon bacterial infection may give rise to escape from innate immune responses acting on various immune effector mechanisms, because PPARc targets many proinflammatory genes [8]. We investigated this question by generating transgenic mice lacking PPARc in myeloid cells. Indeed, myeloid-cell specific knockout of PPARc rendered mice more resistant to infection with L. monocytogenes supporting the assumption that PPARc expression can help bacteria to evade the early phases of innate immunity. Lack of PPARc in myeloid cells resulted in increased expression of essential inflammatory mediators such as TNF, IFNc and IL-12, which in a paracrine fashion stimulated further IFNc expression by NK cells (data not shown). Our findings of increased expression in PPARc-deficient myeloid cells are consistent with earlier reports that pro-inflammatory mediators such as TNF and IFN are target genes of PPARc [34]. The lack of PPARc in myeloid cells also increased expression of iNOS and production of the listeriocidal NO, which most likely occurred indirectly through the augmented production of proinflammatory mediators, because iNOS does not belong to PPARc-regulated genes [34].
In addition to increased bactericidal activity, PPARc ablation in myeloid cells resulted in enhanced recruitment of inflammatory monocytes to the site of infection. We observed increased expression of the key chemokines relevant for inflammatory monocyte recruitment, i.e. CCL2 and CCL7 [46] as well as their receptor CCR2, in the absence of PPARc in myeloid cells. CCL2 and CCR2 are target genes of PPARc [34], which again supports our notion that PPARc regulates anti-bacterial immunity in myeloid cells in a cell-intrinsic fashion. As inflammatory monocytes differentiate into dendritic cell subset, so-called tipDCs, within inflamed tissues [26], it is possible that PPARc also controls the numbers of tipDCs at the site of infection. This assumption is supported by a recent publication demonstrating that pharmacologic PPARc activation reduced chemokine-driven recruitment and local proliferation of tipDCs during viral infection [47]. While chemokine-expression is sufficient for the recruitment of CCR2 + monocytes into the spleen [46], expression of CD54 (ICAM-1) was Figure 7. Transfer of PPARc-deficient macrophages increases the resistance of IFNc 2/2 , TNF 2/2 and CCR2 2/2 mice to Listeria infection. 5610 6 BMDM from LysM-PPARc WT or LysM-PPARc KO mice were adoptively transferred into the peritoneal cavity of IFNc 2/2 , TNF 2/2 or CCR2 2/2 mice and two hours later mice (n = 10) were infected with 2610 4 CFU of Listeria. (A) Survival of IFNc 2/2 and TNF 2/2 mice after infection with Listeria. Knockout mice not receiving any cells by adoptive transfer served as controls. Significance for results in IFNc 2/2 and TNF 2/2 mice were p,0.0001. (B and C) On day 2 after infection, CFU were determined in the homogenates of spleen and liver. One representative out of three independent experiments is shown. doi:10.1371/journal.pone.0037349.g007 shown to be required for the recruitment of these cells to the liver [39]. Here, we show that ablation of PPARc in myeloid cells also led to release of mediators that subsequently increased expression of CD54 on liver sinusoidal endothelial cells, the cell population that is responsible for recruitment of immune cells from blood passing through the liver [39,48]. The recruitment of inflammatory monocytes is of key importance in immune defense against bacterial, fungal and parasite infections [38,49]. Thus, our results demonstrate that PPARc in myeloid cells has a central role in controlling both, the orchestration of bactericidal activity and immune cell recruitment to the site of infection.
PPARc has been reported to have anti-inflammatory effects in many immune cell populations. Recently, we have shown that lack of PPARc in T cells facilitates RORct-mediated development of pro-inflammatory T H17 cells and thereby promotes central nervous system autoimmunity [7]. Activation of PPARc in dendritic cells impairs their ability to elicit T cell mediated immunity [10]. As augmented expression of PPARc and increased expression of endogenous ligands generated by the enzyme 12/15lipoxygenase are observed under inflammatory conditions [44], it is possible that PPARc plays a role in the prevention of overzealous immunity within inflamed tissues, similar to the expression of co-inhibitory molecules like B7H1 acting on PD1 on T cells to prevent organ immunopathology [50]. In fact, PPARc in intestinal epithelial cells is essential to maintain absence of inflammation from the gut [33]. Our findings raise the question whether absence of PPARc from myeloid cells, which improves early anti-bacterial innate immunity, comes at the price of increased immunopathology to infected organs. However, we did not observe any increase in liver injury in mice infected with an LD 50 of Listeria monocytogenes lacking PPARc in myeloid cells (data not shown), indicating that improved anti-bacterial immunity was not accompanied by increased immunopathology. Even at higher doses of bacteria (LD 80 ) we did not observe immunopathology to infected organs such as the liver in LysM-PPARc KO mice. Only when challenged with very high numbers of bacteria (1006 LD 50 ) this resistance to infection was broken and mice lacking PPARc in myeloid cells then died even more rapidly than their wildtype littermates. The balance between pro-and antiinflammatory signals, such as IL-10, TGFb and factors released from regulatory T cells, decides about the outcome between immunity and immunopathology to viruses [51], and generation of regulatory macrophages restricts immunity to bacterial infection [52]. Whereas all these regulatory mechanisms act during the late innate or early adaptive phases of the immune response against infectious pathogens, our results reveal a so far unrecognized early control of anti-bacterial innate effector mechanisms through PPARc activation restricting inflammation, which are not coupled to immunopathology once inactivated. As survival of L. monocytogenes in myeloid cells such as dendritic cells is critical for the successful establishment of infection [53,54], it appears likely that induction of PPARc by L. monocytogenes is a critical step during infection and may serve as early escape from innate immunity.
Taken together our results support the notion that PPARc represents a valuable target for pharmacologic intervention that when neutralized during bacterial infection would lead to preferential enhancement of immune-mediated clearance of intracellular bacteria such as L. monocytogenes without accompanying immunopathology to infected organs like the liver.