IL-4Rα-Independent Expression of Mannose Receptor and Ym1 by Macrophages Depends on their IL-10 Responsiveness

IL-4Rα-dependent responses are essential for granuloma formation and host survival during acute schistosomiasis. Previously, we demonstrated that mice deficient for macrophage-specific IL-4Rα (LysMcreIl4ra−/lox) developed increased hepatotoxicity and gut inflammation; whereas inflammation was restricted to the liver of mice lacking T cell-specific IL-4Rα expression (iLckcreIl4ra−/lox). In the study presented here we further investigated their role in liver granulomatous inflammation. Frequencies and numbers of macrophage, lymphocyte or granulocyte populations, as well as Th1/Th2 cytokine responses were similar in Schistosoma mansoni-infected LysMcreIl4ra−/lox liver granulomas, when compared to Il4ra−/lox control mice. In contrast, a shift to Th1 responses with high IFN-γ and low IL-4, IL-10 and IL-13 was observed in the severely disrupted granulomas of iLckcreIl4ra−/lox and Il4ra−/− mice. As expected, alternative macrophage activation was reduced in both LysMcreIl4ra−/lox and iLckcreIl4ra−/lox granulomas with low arginase 1 and heightened nitric oxide synthase RNA expression in granuloma macrophages of both mouse strains. Interestingly, a discrete subpopulation of SSChighCD11b+I-A/I-EhighCD204+ macrophages retained expression of mannose receptor (MMR) and Ym1 in LysMcreIl4ra−/lox but not in iLckcreIl4ra−/lox granulomas. While aaMφ were in close proximity to the parasite eggs in Il4ra−/lox control mice, MMR+Ym1+ macrophages in LysMcreIl4ra−/lox mice were restricted to the periphery of the granuloma, indicating that they might have different functions. In vivo IL-10 neutralisation resulted in the disappearance of MMR+Ym1+ macrophages in LysMcreIl4ra−/lox mice. Together, these results show that IL-4Rα-responsive T cells are essential to drive alternative macrophage activation and to control granulomatous inflammation in the liver. The data further suggest that in the absence of macrophage-specific IL-4Rα signalling, IL-10 is able to drive mannose receptor- and Ym1-positive macrophages, associated with control of hepatic granulomatous inflammation.


Introduction
Schistosomiasis is a severe parasitic disease with more than 200 million people infected worldwide with an estimated 280,000 deaths per annum in sub-Saharan Africa alone [1,2]. In the murine model, mice infected with Schistosoma mansoni develop a severe liver pathology with granulomatous inflammatory responses directed towards the parasite eggs. During chronic infections, Th2type inflammation in the liver results in fibrosis, which leads to portal hypertension, bleeding from collateral vessels and ultimately death [3,4]. At peak egg excretion, Th2-mediated granuloma formation appears to be indispensable for host protection and control of egg-induced inflammation.
Previous studies have demonstrated signalling via interleukin 4 receptor a-chain (IL-4Ra) to be essential for granuloma formation and host survival [5,6,7,8,9]. The cellular contributions of IL-4Ra to the mechanisms conferring protection to the host can be dissected using mice with IL-4Ra expression disrupted on specific cell types. Infection of macrophage/neutrophil-specific (LysM cre Il4ra 2/lox ) and T cell-specific (iLck cre Il4ra 2/lox ) IL-4Ra-deficient mice showed these mice to have high mortality during acute schistosomiasis irrespective of granuloma formation [6,10]. Eight weeks after infection, the absence of IL-4Ra-dependent T cell responses resulted in severe hepatotoxicity while an absence of IL-4Ra-activated macrophages was responsible for the development of severe inflammation and damage in both intestine and liver, with a subsequent increase in serum lipopolysaccharide (LPS) levels and aspartate transaminase levels in the serum, respectively [6]. Though the mechanism(s) explaining the high mortality rates observed in LysM cre Il4ra 2/lox mice is(are) not fully understood, the observed increased Th1/type1 in presence of normal Th2/type 2 cytokine and antibody systemic responses could be detrimental for the host survival [6].
Tissue macrophages were recently classified into three major categories based on their functions [11]. ''Classical'' macrophage activation occurring during Th1 immune responses, essentially via IFN-c signalling is involved in cellular immunity against intracellular pathogens via the production of pro-inflammatory molecules and nitric oxide (NO). Macrophages can also develop a ''deactivated'' phenotype in association with IL-10 and TGF-b signalling. ''Alternative'' macrophage activation occurs via IL-4 and IL-13 signalling through their heterodimeric IL-4R during Th2 immune responses [11,12,13]. Alternative activation of macrophages results in the downstream activation of various molecules/markers among which arginase 1 (Arg-1) has been considered to be decisive for their functional activities [11]. Earlier studies suggested that alternatively-activated macrophages (aaMw) may be important regulators of wound healing via arginase-1 activation. Arg-1 catalyses L-arginine to produce polyamines and proline, an important factor for collagen deposition [14,15,16]. More recent studies in LysM cre Il4ra 2/lox mice showed that aaMw are involved in immuno-modulation, -suppression or -pathology in infectious diseases such as schistosomiasis, cutaneous leishmaniasis and cryptococcosis, as well as in non-infectious diseases such as insulin resistance, EAE, and tissue repair. In these studies, IL-4Radependent aaMw have been shown to influence innate and adaptive immune responses with both beneficial or detrifmental outcomes, depending on the nature of the disease [6,17,18,19,20,21,22]. Arg-1 activation by IL-4Ra signalling is able to inhibit the activation of inducible nitric oxide synthase (iNOS) and therefore NO production, directly repressing classical macrophage activation. In a recent study, mice selectively lacking Arg1 expression in macrophages during schistosomiasis were less susceptible than LysM cre Il4ra 2/lox mice and developed higher Th2 cytokine responses and antigen-specific Th2 cell proliferation [23].
Besides inducing Arg-1, IL-4Ra signalling induces a large number of markers in macrophages [11,24]. Of these, macrophage mannose receptor (MMR, Mrc1), ''resistin-like'' (Relm) (also referred as ''found in inflammatory zone'' (FIZZ)) molecules and chitinase-like molecules (also referred as ''Ym'') have come to be considered as markers of the aaMw phenotype [11,25,26]. MMR, a C-type lectin, is induced in granulomas during schistosomiasis and is involved in the uptake of soluble antigens by antigen presenting cells [25,27,28]. Resistin-like molecule alpha (Retnla/ Relma/FIZZ1) upregulation has been recently shown to play essential roles in suppression of helminth-induced Th2 inflammation [29,30]. Finally, though the chitinase-3-like 3 molecule (Ym1, Chi3l3) is strongly upregulated in macrophages during Th2-type immunity, its role remains unclear during helminth infection [31,32]. It is also apparent that these markers can be upregulated independently of IL-4 and IL-13, with IL-10 in particular being able to induce the expression of MMR along with other factors including glucocorticoids [33,34]. In addition, innate alternative activation of macrophages in response to chitin has also been reported [35]. Expression of these markers might therefore not be exclusively dependent on IL-4/IL-13 responsiveness of macrophages.
Although the absence of IL-4Ra-dependent aaMw and T cell responses during schistosomiasis has been mainly studied at the systemic level [6,10], little is known of their specific role(s) in the liver granuloma microenvironment. As the major alterations of the immune response depending on IL-4Ra might be restricted to the direct environment surrounding the parasite eggs, we focus in this study on the liver granuloma level to investigate how macrophage-or T cell-specific IL-4Ra responses affect the local responses directed against Schistosoma eggs. Our results demonstrate that the Th1/Th2 balance and the cellular composition of liver granulomas are not affected during infection of LysM cre Il4ra 2/lox mice; but are both severely modified in iLck cre Il4ra 2/lox mice developing granuloma cytokine and cellular responses very similar to the global Il4ra 2/2 mice. Although LysM cre Il4ra 2/lox and iLck cre Il4ra 2/lox granuloma macrophages show a global reduction of expression of aaMw markers, we demonstrate that MMR and Ym1 protein expressions remain high in LysM cre Il4ra 2/lox granuloma macrophages. This IL-4Ra-independent MMR + Ym1 + subpopulation was restricted at the periphery of LysM cre Il4ra 2/lox granulomas, in contrast to Il4ra 2/lox control mice where those cells were in close proximity to the parasite eggs. We finally used a parasite egg model with IL-10R blockade, and demonstrate that in absence of macrophage-specific IL-4Ra expression, IL-10 signalling can induce the expression of MMR and Ym1 in macrophages.

Mice
Macrophage/neutrophil-specific IL-4Ra-deficient mice (LysM cre Il4ra 2/lox ) and T cell specific IL-4Ra-deficient mice (iLck cre Il4ra 2/lox ) were generated with hemizygous Il4ra 2/lox mice and homozygous Il4ra 2/2 mice were used as controls, as previously described [6,10]. All mice used were on a BALB/c background. 8-12 week old sex-matched mice were obtained from the University of Cape Town specific-pathogen-free animal facility. All experiments were approved by the University of Cape Town Animal Ethics Committee.

Parasite infections and antigen preparation
Mice were infected percutaneously with 100 cercariae of a Puerto Rican strain of S. mansoni obtained from infected Biomphalaria glabrata snails as described [10]. Snails and S. mansoni eggs isolated from livers of infected mice were obtained from Dr A.P. Mountford (University of York, UK). In some experiments, mice were injected intraperitoneally (i.p.) with 3,000 liver eggs in PBS. Soluble egg antigen (SEA) was prepared from purified eggs (liver) as described [36] and used at 20mg/ml.

Author Summary
Schistosomiasis is a tropical disease caused by one of the species of the parasitic worm Schistosoma which infects over 200 million people worldwide. Signalling via the IL-4 receptor alpha (IL-4Ra), which is the common receptor chain for the ligands IL-4 and IL-13, is essential for inducing protective Type 2 immune response and granuloma formation in response to the parasite eggs. In experimental Schistosoma mansoni infection and egginduced inflammation studies with cell type-specific IL-4Ra deficient mice, the role of IL-4Ra-activated alternative macrophages (aaMw) and IL-4Ra-responsive T cells was investigated with focus on the control of hepatic inflammation and granuloma formation. Interestingly, aaMw were not essential for the cellular composition or the Th1/Th2 cytokine profile in liver granulomas. In contrast, IL-4Ra-dependent T cell responses were important for predominant Th2 and IL-10 responses, as well as the presence of aaMw in the granulomas, avoiding major disruption in the granuloma cell composition. Moreover, a macrophage subpopulation was identified and those cells expressed the two aaMw markers, mannose receptor-and Ym1 in an IL-4Ra-independent but IL-10-dependent manner. These cells might be involved in the control of inflammation.
Liver granuloma-associated leukocyte purification Granulomatous livers were finely cut in small pieces and digested in culture media containing 50mg/ml collagenase type IV (Sigma) at 37uC for 1.5h. Single cell suspensions were further passed through a 100-mm nylon mesh before leukocytes were isolated through a 30% Percoll cushion at 6006g during 15 min. Liver granuloma cells were further treated for 2 min with NH 4 Cl lysis buffer to lyse erythrocytes. Purified granuloma cells were used for ex vivo restimulation or flow cytometry analysis.

Anti-IL-10 receptor treatment
Mice were inoculated intraperitoneally with PBS or PBS containing 4 mg anti-mouse IL-10Ra (R&D Systems, goat IgG) at day 0, 4 and 6 after injection of eggs. Staining specificity was verified with the appropriate isotype-matched antibody controls and compensation performed with single-stained samples before acquiring the multi-coloured samples. Incubations with antibodies were performed in washing buffer (PBS containing 0.1% BSA, 5mM EDTA and 2mM NaN 3 ) supplemented with heat-inactivated rat serum (2%) and rat anti-mouse FcRII/III mAb (2.4G2, 10mg/ml). PerCP or APC-conjugated-streptavidin was used to detect biotinylated mAbs. For detection of intracellular Ym1 liver cells were fixed for 20 min on ice in 2% (wt/vol) paraformaldehyde and permeabilized during 30 min with 0.5% saponin buffer and further stained with biotinylated goat anti-Ym1/ECF-L (R&D systems). mAbs were from BD Pharmingen except where noted otherwise. Acquisition was performed using a FACSCalibur (BD Immunocytometry Systems), and data were analyzed by FlowJo software (Treestar). In some experiments, liver granuloma cells were sorted to high purity ($98%) using a FACSVantage (BD Immunocytometry Systems) cell sorter. Sorted cell populations (SSC high CD11b + I-/I-E 2 F4/80 +/2 Gr-1 int/high granulocytes or CD11b + I-A/I-E high CD204 + F4/80 + macrophages) were either used for RNA extraction or stained with Diff-Quick (Rapidiff set -Clinical Science Diagnostics) after cytospin for direct microscopic examination.

Ex vivo restimulation and cytokine detection
Single-cell suspensions of mesenteric lymph node (mLN) cells or liver granuloma-associated leukocytes were cultured overnight in Iscove's modified Dulbecco medium (IMDM) containing 10% FCS, 2mM L-glutamine, 0.1 mM non essential amino acids, 1 mM Na Pyruvate (Invitrogen) and 50 mM 2-mercaptoethanol (Sigma) in presence or absence of 20 mg/ml SEA before cytokine detection by flow cytometry. IL-4, IL-10 and IFN-c secreting cells were detected with the cytokine-specific mouse secretion assay detection kits (PE) (Miltenyi) according to the manufacturer's instructions. Dead cells were excluded from analysis by using 7amino-actinomycin D (7-AAD, Sigma). For detection of intracellular IL-13, cells were further incubated 4h with monensin, fixed for 20 min on ice in 2% (wt/ol) paraformaldehyde and permeabilized during 30 min with 16 permeabilization buffer (eBioscience) before being stained with anti-IL-13-PE (eBioscience, eBio13a).

Immunohistology
Liver tissue was embedded in OCT (Tissue-Tek, Sakura) before cryopreservation at 280uC. Seven-mm cryosections were cut and mounted onto 3-aminopropyltriethoxysilane (APES) coated slides, dehydrated overnight at 4uC before being fixed in ice-cold acetone. Washes were performed in 0.05% Tween-20 in PBS and staining steps in PBS containing 0.1% BSA. We used the primary antibodies to the following: CD206-Biotin (MMR, 5D3, Serotec), Ym1/ECF-L-Biotin (goat IgG, R&D Systems), CD204-FITC (Scavenger receptor-A, 2F8, Serotec), or iNOS (rabbit IgG, provided by Dr J.Pheilschifter, Germany). PE-conjugated streptavidin (BD Biosciences) was used to detect biotinylated primary antibodies after an avidin/biotin blocking step (Vector). FITCconjugated goat anti-rabbit IgG secondary antibody (Abcam) was used for the detection of iNOS staining. Sections were washed in PBS before coverslipped in anti-fade fluorescent mounting medium (Dako). Staining specificity was verified by using irrelevant antibody controls. Images were taken with a Nikon Eclipse E400 microscope (Nikon), control by a NIS-Elements Basic Research 3.0 imaging system. Exposure times and fluorescence intensities were normalized to appropriate control images. Photomicrographs of liver granuloma focusing on autofluorescent parasite eggs were captured using a Nikon 5.0 mega pixels color digital camera (Digital SIGHT DS-SMc). Images were photographed separately for each fluorescent channels and were merged using Adobe Photoshop 7.0.

Quantitative RT-PCR
Total RNA was extracted and purified with RNeasy Microprep kit (Qiagen) and cDNA synthesised with Transcriptor First Strand cDNA synthesis kit TM (Roche). Real-time PCR was performed using LightcyclerH FastStart DNA Master PLUS SYBR Green I reaction mix (Roche) and the reactions run on a LightcyclerH carousel-based system (Roche). Primers used are listed in Table 1.

Statistical analysis
Statistical analyses were conducted using GraphPad Prism 4 software (www.graphpad.com). One-way ANOVA test was used to determine significant differences, with Bonferroni's multiple comparison post test applied to calculate significance values between samples.

Characterization of macrophagic/granulocytic cell populations within liver granulomas in schistosomiasis
The major leukocytic cell types associated with the liver granuloma in S. mansoni infected wild-type mice are T cells, eosinophils and macrophages [3]. IL-4/IL-13-activated macrophages have been shown to play a crucial role in host survival during acute schistosomiasis [6]. Studies to define tissue macrophages are however complex as these cells can express different markers depending on the organ or tissue, in particular eosinophils are known to express high levels of both CD11b and F4/80 [37].
Cellular composition and cell-specific cytokine responses of liver granulomas in LysM cre Il4ra 2/lox and iLck cre Il4ra 2/lox mice We previously reported that both S. mansoni-infected LysM cre Il4ra 2/lox and iLck cre Il4ra 2/lox mice developed hepatotoxicity and increased granuloma sizes [6,10]. It is however unknown whether macrophage/neutrophil-specific or T cell-specific deletion of IL-4Ra affects the cellular composition or cytokine responses within liver granulomas. At 8 weeks post-infection (p.i.), corresponding to the peak of Th2 responses, leukocytes were isolated from granulomas of infected Il4ra 2/lox , LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox ,and global Il4ra 2/2 mice and analyzed by flow cytometry. Il4ra 2/lox littermates were used as controls. Results obtained in Table 2 show that the relative frequencies of macrophages, neutrophils, eosinophils and lymphocyte subpopulations in granulomas were not affected by the specific impairment of IL-4Ra on macrophages, with eosinophils (CD11b + F4/80 + Gr-1 int Siglec-F + ) dominating granulomatous cell composition in LysM cre Il4ra 2/lox and infected Il4ra 2/lox control mice. We observed however dramatically disrupted granuloma cell populations in iLck cre Il4ra 2/lox mice with increased frequencies of macrophages (CD11b + I-A/I-E high CD204 + ), neutrophils (CD11b + F4/80 2 Gr-1 high ), and cd-T cells. Though we previously showed that these mice developed bigger granulomas with reduced numbers of eosinophils per granuloma [10], iLck cre Il4ra 2/lox mice showed high numbers of eosinophils in their liver (143.1610 5 cells in iLck cre Il4ra 2/lox mice vs. 110.5610 5 cells in Il4ra 2/lox mice). This apparent discrepancy could be explained by the fact that these mice had globally increased cell numbers per liver (data not shown). The cellular changes in iLck cre Il4ra 2/lox mice were very similar to the disruption of the granuloma cell frequencies observed in Il4ra 2/2 mice, also showing increased CD4 + T cells and B cells. The disruption of the granulomas in Il4ra 2/2 mice could be explained by the 4-fold-reduced numbers of eosinophils (25.4610 5 cells in Il4ra 2/2 mice vs. 115.9610 5 and 110.6610 5 cells in LysM cre Il4ra 2/lox and Il4ra 2/lox mice, respectively) ( Table 2). CD4 + CD62L high naive T cell numbers however changed with a 4fold increase in infected iLck cre Il4ra 2/lox mice and a 10-fold increase in infected Il4ra 2/2 mice compared to Il4ra 2/lox control mice overproportionally (28.3610 4 cells in Il4ra 2/2 mice and 11.9610 4 cells in iLck cre Il4ra 2/lox mice vs. 4.3610 4 and 2.9610 4 cells in LysM cre Il4ra 2/lox and Il4ra 2/lox mice, respectively). In contrast, NK (CD11b 2 SSC low TCR ab 2 CD49b + ) and NK-T (CD11b 2 SSC low TCR ab + CD49b + ) cells showed significant lower frequencies in the liver of infected Il4ra 2/2 mice compared to Il4ra 2/lox control mice (10.3% vs 17.3% and 8.2% vs 25.7%, respectively), resulting in a 2-fold cell number reduction of NK-T cells compared to control mice (2.9610 5 cells in Il4ra 2/2 mice vs. 5.8610 5 cells in Il4ra 2/lox mice) ( Table 2). Following ex vivo SEA restimulation of mesenteric lymph node (mLN) cells or liver granuloma-associated leukocytes, cytokine productions of CD4 + T cells and non-CD4 + cells were analyzed by multicolour flow cytometry, either by catch secretion assay (IFN-c, IL-4, IL-10) or intracellular staining (IL-13) (Fig. 2). Cytokine catch secretion assays give information on cytokines secreted by a specific cell-type, gathering therefore advantages of both ELISA and intracellular staining. In both tissues, lymphocytes from LysM cre Il4ra 2/lox mice produced similar levels of IL-13, IL-10 and IFN-c cytokines compared to Il4ra 2/lox control mice with slightly but significant higher levels of IL-4 in liver CD4 + cells compared to Il4ra 2/lox control mice (Fig. 2B). In contrast, lymphocytes from both iLck cre Il4ra 2/lox and Il4ra 2/2 mice produced less Th2 cytokines but more IFN-c revealing a shift towards Th1-type responses within the granulomas, explaining the observed differences in cellular composition.
Mannose receptor and Ym1 expression in liver granuloma-associated macrophages of LysM cre Il4ra 2/lox mice We further compared gene expression levels of markers used to distinguish classical from alternative macrophage activation within liver granulomas during acute S. mansoni infection. Liver granuloma macrophages (SSC high CD11b + I-A/I-E high CD204 + )   were FACS-sorted to high-purity at 8 weeks p.i. ($98%), RNA extraction performed, and gene expression profile analyzed (Fig. 3A). Il4ra (IL-4Ra) mRNA expression, observed in Il4ra 2/lox control macrophages, was impaired in macrophages from LysM cre Il4ra 2/lox mice and global Il4ra 2/2 mice, confirming the efficiency of Cre-loxP mediated DNA deletion in granuloma macrophages. Interestingly, Il4ra expression was also reduced in iLck cre Il4ra 2/lox macrophages, suggesting a predominant classical activation in the granulomas of this mouse strain. Consistent with the IL-4/IL-13/IL-4Ra-and IFN-c/IFN-cR-mediated macrophage activation by direct control of expression of the enzymes catalysing L-arginine, LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox , and global Il4ra 2/2 mice presented classically activated macrophages with heightened Nos2 but reduced Arg1 mRNA expression levels, compared to Il4ra 2/lox control mice. This was in line with the expression pattern of the known alternative activation marker gene Retnla (FIZZ1), significantly reduced in macrophages from LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox and global Il4ra 2/2 mice, compared to Il4ra 2/lox control mice. Of importance, the aaMw marker Chi3l3 and Mrc1 gene expression levels were present although reduced for Mrc1 in LysM cre Il4ra 2/lox macrophages, resulting in Ym1 and MMR surface expression in about 30% of SSC high CD11b + I-A/I-E high CD204 + macrophages from LysM cre Il4ra 2/lox mice. The expression of these two markers was however strongly reduced in both iLck cre Il4ra 2/lox and global Il4ra 2/2 mice, compared to control Il4ra 2/lox mice (Fig. 3B). Mean fluorescent intensities (MFI) values of MMR and Ym1 stainings were also significantly higher in macrophages from LysM cre Il4ra 2/lox compared to both iLck cre Il4ra 2/lox and global Il4ra 2/2 mice (Fig. 3C). The majority (76.3%) of MMR-expressing CD11b + I-A/I-E high granuloma macrophages also expressed Ym1 in the Il4ra 2/lox control mice (Fig. 3D). Absence of surface IL-4Ra expression was further confirmed by 4-colour staining on granuloma macrophages (Fig. 3E). MFI of IL-4Ra staining on gated CD11b + MMR + cells revealed similar expression levels of the receptor between LysM cre Il4ra 2/lox mice and isotype control (MFI = 5.54 and 5.52, respectively), whereas granuloma CD11b + MMR + cells from Il4ra 2/lox mice showed 2-fold increased expression of the IL-4Ra compared to LysM cre Il4ra 2/lox mice and isotype control (MFI = 12.69) (Fig. 3E). Together, these data suggest that MMR and Ym1 can be expressed independently of macrophage-specific IL-4Ra signalling, whereas IL-4Ra-dependent T-cell responses are necessary for driving expression of aaMw markers, including MMR and Ym1.
Peripheral localisation of IL-4Ra-independent MMR-and Ym1-positive macrophages within egg-induced liver granulomas of LysM cre Il4ra 2/lox mice The results detailed in Figure 3 demonstrated that LysM cre Il4ra 2/lox granuloma macrophages have retained expression of MMR and Ym1. This was surprising as we previously showed that liver granuloma from LysM cre Il4ra 2/lox mice had low levels of MMR expression in close proximity to the eggs compared to Il4ra 2/lox control mice [6]. To further determine the localisation of the MMR-and Ym1-positive macrophage subpopulation within the liver granuloma, immunofluorescent stainings were performed on liver cryosections at 8 weeks p.i. (Fig. 4). As expected from the expression data, MMR and Ym1 were highly expressed in granulomas from Il4ra 2/lox control mice; and these cells coexpressing scavenger receptor A (CD204) were in close proximity to the parasite eggs in the centre of the granuloma (Fig. 4A and B). MMR-and Ym1-positive macrophages were restricted to the periphery of the granuloma in LysM cre Il4ra 2/lox mice and nearly undetectable in both iLck cre Il4ra 2/lox and Il4ra 2/2 mice. Instead, iNOS-positive, MMR-and Ym1-negative macrophages were in close proximity around the egg in the granuloma of LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox and Il4ra 2/2 mice (Fig. 4B), confirming previous observations [6,10,41]. These results show that a peripheral IL-4Raindependent MMR-and Ym1-positive macrophage subpopulation surrounds the iNOS-producing classically activated macrophages present in close proximity to the egg within the liver granuloma of LysM cre Il4ra 2/lox mice.
IL-4Ra-independent MMR-and Ym1-positive macrophages depend on IL-10 signalling IL-10 signalling can induce MMR expression in macrophages [34]. We therefore investigated whether IL-10 could drive the MMR and Ym1 expression in macrophages of LysM cre Il4ra 2/lox mice. To test this hypothesis we used a local egg model by injecting purified S. mansoni eggs intraperitoneally to induce Th2 responses and elicit macrophage activation in Il4ra 2/lox , LysM cre Il4ra 2/lox and Il4ra 2/2 mice. At 7 days p.i., corresponding Table 2. Cellular composition of liver granulomas depending on macrophage/neutrophil or T cell-specific IL-4Ra deletion. to the peak of Th2 responses [42], MMR and Ym1 were highly expressed in peritoneal macrophages of Il4ra 2/lox control mice (Fig. 5A). To investigate peritoneal macrophages, we placed a live gate on FSC high SSC low CD11b + F4/80 + cells. Here the association of the CD11b, F4/80 markers and side/forward scatter allows discrimination between eosinophils and macrophages, as previously described [37]. Similar to that observed in macrophages from liver granuloma, a subpopulation of cells expressing MMR and Ym1 was present in peritoneal macrophages of LysM cre Il4ra 2/lox mice but absent in peritoneal macrophages from Il4ra 2/2 mice (Fig. 5A). In vivo blockade of IL-10 signalling via anti-IL-10 receptor antibody treatment was sufficient to significantly reduce the protein expression of both MMR and Ym1 in peritoneal macrophages of LysM cre Il4ra 2/lox mice ( Fig. 5B and C). These results demonstrate that IL-10 drives expression of MMR and Ym1 in macrophages independently of macrophage IL-4Ra.

Discussion
Recent S. mansoni infection studies demonstrated that LysM cre Il4ra 2/lox mice, deficient for IL-4Ra-responsive aaMw succumb to acute S. mansoni infection, due to the development of egg-induced gut inflammation, severe liver damage, and despite the formation of egg-induced granulomas and unaffected fibrosis in the liver [6]. These observations were very similar to the pathology developed by infected global Il4ra 2/2 mice [6], or Il4 2/2 and Il4 2/2 Il13 2/2 mice which developed hepatotoxicity and endotoxaemia due to the degradation of the intestinal barrier [5,9]. Antibiotic treatment in infected LysM cre Il4ra 2/lox mice resulted only in partial protection [6], suggesting that in addition to gut inflammation and systemic leakage of the intestinal content, hepatotoxic lesions could also explain the increased mortality. In addition to these findings, we also recently demonstrated, using iLck cre Il4ra 2/lox mice, that IL-4Ra responsiveness by T cells was required for host survival mainly for controlling liver damage caused by the parasites eggs [10]. As iLck cre Il4ra 2/lox mice did not develop gut injury or endotoxaemia, we concluded that IL-4Ra responsiveness by T cells is not essential for the control of intestinederived sepsis. In this study we focused on the liver granuloma microenvironment and the consequences of impaired IL-4Ra signalling in macrophages or T cells on the local cellular responses directed against the parasite eggs.
Granulomas induced by S. mansoni contain both myeloid cells (mainly eosinophils and macrophages) and lymphoid cells (mainly T cells) [3]. Several markers are readily available for staining murine tissue macrophages but their expression can be highly tissue-dependent [43]. Most of the studies using such markers have been performed with blood, spleen or bone-marrow but little is known on granuloma macrophage markers during schistosomiasis. Liver granulomas are tightly organized, which renders their disruption difficult without affecting cell viability, especially macrophages. Though collagenase can in some conditions cleave surface proteins, comparison between collage-nase-treated or untreated organs gave similar results concerning the expression levels of standard surface markers (including CD11b, F4/80, CD11c, Gr-1, and I-A/I-E), suggesting that collagenase treatment had no major influence on surface marker detection (data not shown). The combination of CD11b and/or F4/80 and MHC-II (I-A/I-E) was sufficient to specifically detect and isolate macrophages from granulomas. These macrophages were defined as SSC high CD11b + (F4/80 + )I-A/I-E high CD204 + cells, showed a macrophage-like morphology and expressed high levels of macrophage-specific markers such as scavenger receptor-A (CD204), mannose receptor (MMR), b-glucan receptor (Dectin-1) and macrosialin (CD68). In parallel, we defined granuloma eosinophils as SSC high CD11b + F4/80 + I-A/I-E 2 Gr-1 int Siglec-F + and neutrophils as SSC high CD11b + F4/80 2 I-A/I-E 2 Gr-1 high Siglec-F 2 .
Despite increased liver granuloma sizes, previously reported in infected LysM cre Il4ra 2/lox mice [6], neither the cellular composition nor the Th1/Th2 cytokine profiles were affected. In contrast, we observed severe changes in granuloma cell composition and morphology of infected iLck cre Il4ra 2/lox and Il4ra 2/2 mice. Interestingly, a recent report showed that infected LysM cre Arg1 2/lox or Tie2 cre Arg1 lox/lox mice (deficient for Arg1-expressing macrophages) did not develop lethal endotoxaemia or hepatotoxicity but showed increased Th2 responses and increased chronic disease [23]. Further, Arg1-expressing aaMw were able to suppress schistosome-specific T cell proliferation, suggesting a possible mechanism for downmodulating granulomatous inflammation and slow the progression of Th2-driven fibrosis during chronic infection [23]. Similarly, studies in Retnla-deficient mice showed that absence of resistin-like molecule alpha (Retnla/FIZZ1) exacerbated Th2 cytokine responses in S. mansoni-egg-induced lung inflammation, suggesting that Retnla/FIZZ1 is able to control effector Th2 responses in S. mansoni egg-induced inflammation and infection [29,30]. As IL-4Ra-sinalling is upstream from the induction of Arg1 and Retnla in aaMw, it might be not surprising that LysM cre Il4ra 2/lox mice have a more severe disease phenotype already detrimental in the acute phase. The relative normal granuloma morphology, cellular composition and Th1/Th2 cytokine responses in the liver of LysM cre Il4ra 2/lox mice but more severe liver pathology in iLck cre Il4ra 2/lox mice, suggest that abrogation of IL-4-promoted Th2 responses in combination with impairment of aaMw negatively affect the control of liver pathology in mice. Together with the chronic phenotype and aaMw-mediated T cell suppression observed in LysM cre Arg1 2/lox mice, aaMw seem to control T helper cell responses in an optimal balance. Future studies need to further investigate the molecular mechanisms of how and which T cell subpopulations are controlled by aaMw.
Interestingly, LysM cre Il4ra 2/lox liver granulomas retained expression of MMR and Ym1 in a subpopulation of macrophages, confirmed in mRNA and protein (Chi3l3) or protein (Mrc1) expression analyses. Interestingly, MMR + Ym1 + macrophages were not observed in iLck cre Il4ra 2/lox mice, suggesting that IL- Figure 2. IL-4Ra-expressing macrophages do not affect cytokine responses in granulomas. Il4ra 2/lox , LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox and Il4ra 2/2 mice were infected with 100 S. mansoni cercariae and killed 8 weeks later. Mesenteric lymph node (mLN) cells or liver granuloma-associated leukocytes (Liver) were then isolated and restimulated overnight with 20 mg/ml SEA before analyzed for ex vivo cytokine production capability. Staining of surface CD4 was followed by detection of IL-4, IL-10 and IFN-c secretion in a cytokine catch assay or detection of intracellular levels of IL-13 after 4h of monensin treatment, as described in Methods. Live gate was placed on lymphocytes according to their forward-and side-scatter by flow cytometry. (A) Representative contour plots of IFN-c, IL-4, IL-10 and IL-13-producing lymphocytes. Quadrant bars were set up on unstimulated cells and numbers represent percent cells in each quadrant. Data represent one out of three independent experiments with similar results. (B) Percentages of IFN-c, IL-4, IL-10 or IL-13-producing cells were determined in gated CD4 + or non-CD4 + cell populations from mLN or liver granulomaassociated lymphocytes (Liver) (mean6SEM, n = 4). Data are representative of three independent experiments. *p,0.05; **p,0.001; ***p,0.001 compared to control Il4ra 2/lox mice. doi:10.1371/journal.pntd.0000689.g002 IL-4R-Independent Macrophage MR and Ym1 Expression www.plosntds.org 4Ra-dependent Th2 responses drive MMR and Ym1 expression independently of IL-4Ra-responsive macrophages in LysM cre Il4ra 2/lox mice. After staining of scavenger receptor-A (CD204) in situ to specifically detect macrophages [44,45], we showed that macrophages were in close proximity to the parasite eggs within the granuloma in all mouse strains. However, macrophages expressing MMR or Ym1 were only found at the periphery of the LysM cre Il4ra 2/lox granulomas, while present in close proximity to the eggs in the Il4ra 2/lox control granulomas. As previously demonstrated, MMR-or Ym1-positive cells were absent of global Il4ra 2/2 granulomas [25] and we observed nearly undetectable MMR + or Ym1 + cells in iLck cre Il4ra 2/lox granulomas. Instead classically activated macrophages, defined by their iNOS expression were centred on the eggs in LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox and global Il4ra 2/2 granulomas (Fig. 4B). Increased iNOS activity in these mice strains could explain the heightened hepatocellular damage observed after S. mansoni infection [6,10]. Together, these results suggest that IL-4Ra-responsiveness in aaMw regulates their ability to migrate in close proximity and/or interact with the parasite eggs. The factors that could be responsible for the interaction with the eggs are not known, but a possible candidate could be MMR. This receptor encodes a C-type lectin, which has been involved in antigen uptake by antigen-presenting cells and components of S. mansoni eggs as well as a fraction of their egg-secreted molecules are ligands of MMR [25]. Furthermore, a study demonstrated that SEA can be internalized by DCs through the C-type lectins dendritic cell specific ICAM-3 grabbing non integrin (DC-SIGN), macrophage galactose-type lectin (MGL) and MMR [28]. We did not look at the expression of other C-type lectins, but the presence of macrophages expressing MMR in the proximity of the parasite eggs could allow IL-4Ra-responsive aaMw to interact with antigen-specific T cells and also play regulatory functions on effector T cells as previously suggested [6,23,29,30,46]. Although Ym1 was described as a chitinase-like secreted protein, it lacks any chitinase activity and S. mansoni eggs do not contain chitin [47,48]. A recent study described Ym1 to digest glycosaminoglycans and might therefore play a role in egg killing and/or antigen processing [49]. It has previously been suggested that Ym1 could encapsulate pathogens [31] and our observation of Ym1expressing macrophages in the proximity of the parasite eggs supports this hypothesis. Encapsulation of the parasite eggs may serve to protect the host from pro-inflammatory molecules and harmful factors released by the eggs. Macrophages isolated from granulomas or macrophages elicited with S. mansoni egg components have previously been proposed to act as ''myeloid suppressive cells (MSCs) by inducing clonal anergy in egg-antigen-specific Th1 cells by unknown mechanisms [46,50,51,52,53,54]. Although we did not detect changes in cellular composition or cytokine production in the liver granulomas of LysM cre Il4ra 2/lox mice, these mice quickly die from increased systemic Th1-type responses and acute infection due to uncontrolled gut inflammation [6]. Recently established MMRdeficient mice [55], generation of Ym1-deficient mice should allow further studies to address the role of such aaMw biomarkers.
Previous studies proposed that MMR and Ym1 expression by macrophages could directly be driven by bacterial endotoxin or chitin [35], and independently of IL-4Ra signalling [56]. Linehan and colleagues (2003) previously described a complete absence of MMR expression in the liver granulomas of global Il4ra 2/2 mice [25], demonstrating that IL-4Ra responsiveness is essential for MMR expression in macrophages recruited to the granulomas. These authors however described resident liver macrophages to retain MMR expression and we also observed MMR expression in resident Küpffer cells (Fig. 4A). As Küpffer cells express no/low levels of CD11b in contrast to inflammatory macrophages [38,39], we can exclude any contamination with these cells in our gating strategy targeting SSC high CD11b + (F4/80 + )I-A/I-E high CD204 + granuloma macrophages.
Our findings demonstrate that direct IL-4Ra signalling in a subpopulation of granuloma macrophages is not essential for MMR and Ym1 expression and these cells are restricted to the periphery of LysM cre Il4ra 2/lox granulomas. IL-10 signalling in macrophages can induce MMR expression [34] and Ym1 expression is downregulated in Il10 2/2 and Il10 2/2 Il4 2/2 deficient mice [32]. These observations, together with reduced IL-10 production and impaired aaMw markers (including MMR and Ym1) in iLck cre Il4ra 2/lox and Il4ra 2/2 liver granulomas, prompted us to consider IL-10 as a good candidate to explain the IL-4Ra-independent expression of MMR and/or Ym1 in macrophages. In order to focus on the egg-induced inflammation in the LysM cre Il4ra 2/lox mice, we chose to use an egg model to induce acute Th2 responses following challenge in the peritoneum. Here, we showed evidence by in vivo IL-10R signalling blockade that IL-10 signalling was responsible for the retained expression of MMR and Ym1 expression in LysM cre Il4ra 2/lox peritoneal macrophages following S. mansoni egg exposure and induction of Th2 cytokine responses, while no significant effect could be observed in macrophages of Il4ra 2/lox control mice. Although elicitation of macrophages in the peritoneal cavity by S. mansoni eggs might not exactly reflect what occurs in the liver granuloma during live infection, data generated with peritoneal macrophages gave similar results as those obtained with granuloma macrophages. Control Il4ra 2/lox mice over-expressed both MMR and Ym1, whereas expression of those markers in Figure 3. Mannose receptor and Ym1 expression in granuloma macrophages of LysM cre Il4ra 2/lox mice. Il4ra 2/lox , LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox and Il4ra 2/2 mice were infected with 100 S. mansoni cercariae, killed 8 weeks later and liver granuloma-associated leukocytes purified. (A) CD11b + I-A/I-E high CD204 + granuloma macrophages were sorted to high purity (.98%) and RNA extracted before analysis of expression levels of genes associated with macrophage activation by real time RT-PCR: IL-4Ra (Il4ra), nitric oxide synthase (Nos2), arginase 1 (Arg1), FIZZ1 (Retnla), macrophage mannose receptor (Mrc1), and Ym1 (Chi3l3). Relative expression levels normalized to r12S house-keeping gene are shown as fold-change over peritoneal macrophages from Il4ra 2/2 mice was strongly impaired. Moreover, LysM cre Il4ra 2/lox peritoneal macrophages also retained expression of MMR and Ym1 after injection of eggs and anti-IL-10R treatment blocked the expression of both markers. These results indicate that IL-10 plays a key role in the retained expression of MMR and Ym1 in the liver granulomas of LysM cre Il4ra 2/lox mice. The macrophages retaining expression of MMR and Ym1 in infected LysM cre Il4ra 2/lox mice could therefore represent a cellular subpopulation with IL-4Raindependent aaMw-related regulatory activities, driven by IL-10. The presence of such macrophage population could explain why LysM cre Il4ra 2/lox liver granulomas do not have disrupted cellular or cytokine responses, in contrast to mice having impaired Th2 responses and showing strongly reduced IL-10 production such as iLck cre Il4ra 2/lox and Il4ra 2/2 mice. Supporting these results, Herbert et al. recently showed that in vivo IL-10R neutralization results in increased hepatocellular damage without affecting gut inflammation in schistosomiasis [57]. Hepatotoxicity was however observed in infected LysM cre Il4ra 2/lox mice [6], suggesting that IL-10 without IL-4/IL-13 signalling in macrophages might not be sufficient to control liver tissue damage. Future work on IL-4Ra-independent MMR + Ym1 + macrophages from liver granulomas may define whether those cells have an ''alternative'' or ''deactivated'' phenotype, which was previously described to be mediated by IL-10 signalling [11].
In conclusion, we demonstrated that IL-4Ra-responsive macrophages are dispensable for the control of cellular and cytokine responses in liver granulomas during acute schistosomiasis, whereas IL-4Ra-responsive T cells are necessary to sustain cellular composition and morphology of the liver granuloma. Though our results suggest that macrophage-specific IL-4Ra expression is necessary for an adequate interaction between . Localisation of mannose receptor and Ym1-expressing granuloma macrophages in close contact with S. mansoni eggs depends on their IL-4Ra signalling. Livers were collected at 8 weeks p.i. from Il4ra 2/lox , LysM cre Il4ra 2/lox , iLck cre Il4ra 2/lox and Il4ra 2/2 mice and immunofluorescent stainings performed on cryosections. (A) Representative micrographs of MMR (panels v-viii), Ym1 (panels xiii-xvi), CD204 (panels xxi-xxiv) and iNOS (panels xxix-xxxii) stainings of liver cryosections as described in Methods. Stainings with secondary antibody (MMR, iNOS), streptavidin alone (Ym1) or isotype-control (CD204) is shown for each corresponding staining (MMR = i-iv, Ym1 = ix-xii, CD204 = xvii-xx, iNOS = xxv-xxviii). Note that MMR + and Ym1 + cells are restricted to the periphery of LysM cre Il4ra 2/lox granulomas. Outlined regions represent the areas magnified in B. (B) Representative micrographs of liver cryosections stained with scavenger receptor (CD204) for macrophages detection (panels i-viii, green) or iNOS for classically activated macrophages detection (panels ix-xvi, green); and co-stained for MMR (panels i-iv and ix-xii, red) or Ym1 (panels v-viii and xiii-xvi, red). Note the low frequency of CD204 + macrophages co-expressing MMR + or Ym1 + cells around the parasite eggs (panels ii and vi) but the high levels of iNOS + cells (panels x and xiv) in LysM cre Il4ra 2/lox mice, suggesting these macrophages to be classically activated. White arrows indicate the parasite eggs. Original magnification: 4006. Data represent one of three independent experiments. doi:10.1371/journal.pntd.0000689.g004 IL-4R-Independent Macrophage MR and Ym1 Expression www.plosntds.org aaMw and the parasite eggs, we showed that MMR and Ym1 expression may be induced via IL-10R signalling in absence of IL-4Ra signalling in macrophages. Taken together, our results suggest that IL-4Ra-derived Th2 cytokine responses are essential to drive aaMw in the liver and demonstrate that in absence of macrophage-specific IL-4Ra, IL-10 signalling induces MMR + Ym1 + macrophages. Future investigation may clarify whether IL-10-driven IL-4Ra-independent MMR + Ym1 + macro-phages have important functions in the control of liver granulomatous inflammation.