LRH-1 mediates anti-inflammatory and antifungal phenotype of IL-13-activated macrophages through the PPARγ ligand synthesis

Liver receptor homologue-1 (LRH-1) is a nuclear receptor involved in the repression of inflammatory processes in the hepatointestinal tract. Here we report that LRH-1 is expressed in macrophages and induced by the Th2 cytokine IL-13 via a mechanism involving STAT6. We show that loss-of-function of LRH-1 in macrophages impedes IL-13-induced macrophage polarization due to impaired generation of 15-HETE PPARγ ligands. The incapacity to generate 15-HETE metabolites is at least partially caused by the compromised regulation of CYP1A1 and CYP1B1. Mice with LRH-1-deficient macrophages are, furthermore, highly susceptible to gastrointestinal and systemic Candida albicans infection. Altogether, these results identify LRH-1 as a critical component of the anti-inflammatory and fungicidal response of alternatively activated macrophages that acts upstream from the IL-13-induced 15-HETE/PPARγ axis.

M acrophages orchestrate innate immune responses by initiating and resolving inflammatory signalling programmes. Emerging evidence indicates that the state of macrophage polarization plays a critical role in the regulation of these inflammatory processes. Two different programmes of macrophage activation, the classical (M1) and the alternative differentiation, classify polarized macrophages with either persistence or resolution of inflammation [1][2][3] . M1 macrophages express high levels of opsonic receptors, involved in the production of pro-inflammatory effector molecules such as reactive oxygen and nitrogen intermediates and proinflammatory cytokines (interleukin (IL)-1b, tumour-necrosis factor alpha (TNFa), IL-6 and IL-12). These macrophages contribute to inflammation, microbial killing, regulation of cell proliferation and apoptosis. Alternatively activated macrophages are characterized by abundant levels of the anti-inflammatory cytokine IL-10 and non-opsonic receptors, such as C-type lectin receptors and scavenger receptors (CD36), and resolve inflammation by increasing CD36-mediated efferocytosis and secretion of tissue remodelling/repair mediators 3,4 .
The balance of macrophage differentiation in favour of alternatively activated macrophages can be shifted by the activation of the nuclear receptor peroxisome proliferator-activated receptor gamma (PPARg) (refs 5,6). PPARg expression and activity in macrophages is negatively regulated during inflammatory processes 7,8 . In addition, activated PPARg transrepresses many inflammation-activated transcription factors, including nuclear factor-kappaB (NF-kB), signal transducers and activators of transcription (STATs), activator protein 1 (AP1) and nuclear factor of activated T-cells NFAT), resulting in pro-inflammatory mediator inhibition 9 . PPARg is activated by endogenous ligands derived from the metabolism of arachidonic acid (AA) 9 . Among these ligands, 15-deoxy-D 12,14 PGJ 2 (15d-PGJ2), metabolized through the COX1/COX2 cyclooxygenases, and the 12-and 15hydroxyeicosatrienoic acids (HETEs), metabolized through 5 and 12/15 lipoxygenases, are essential for PPARg endogenous activation 5,10,11 . In addition to cyclooxygenases and lipoxygenases, cytochrome P450 (CYP) enzymes are also considered to be critical for the metabolism of AA in epoxy (EETs) and in hydroxy (HETEs) derivatives 10,11 . Within the CYP family, the CYP1 family is mainly involved in the generation of 12-and 15-HETEs through CYP1A1 and CYP1B1 (refs 12,13).
Liver receptor homologue-1 (LRH-1, NR5A2) is a nuclear receptor highly expressed in the intestine, liver, pancreas and ovary 14,15 . Although LRH-1 has been recognized as an orphan receptor, phospholipids, including the phosphatidyl inositol second messengers, and more recently the 12C-fatty acyl-containing phospholipid, dilauroyl phosphatidylcholine (DLPC), have been described to bind the ligand-binding pocket and to act as LRH-1 agonists [16][17][18] . LRH-1 plays important roles in embryonic development, cholesterol and bile acid homeostasis 14,15 and promotes hepatic glucose sensing through the regulation of the glucokinase enzyme 19 . Several lines of evidence also support a role for LRH-1 in the control of the inflammatory response. While pro-inflammatory factors such as TNFa and lipopolysaccharide (LPS) decrease LRH-1 expression in murine models of human colon tumorigenesis, deficiency of LRH-1 in the intestinal epithelium predisposes mice to intestinal inflammation as a result of a defect in local glucocorticoid production. In the colon from patients with inflammatory bowel disease, inflammation is inversely correlated with the expression of LRH-1 (refs 20,21). In line with these reports, Venteclef et al. 22,23 identified a role for LRH-1 in the negative modulation of the hepatic acute-phase response by inhibiting IL-6-and IL-1b-stimulated haptoglobin, serum amyloid A gene expression in hepatocytes and inducing antiinflammatory IL-1ra expression. Despite the numerous studies documenting the anti-inflammatory properties of LRH-1 in the liver and gut, no studies so far have focused on the role of LRH-1 in macrophages.
In the present study, we identify LRH-1 as an important regulator of the inflammatory response in macrophages. We demonstrate that LRH-1 is induced by IL-13 via a STAT6dependent mechanism, which in turn induces the transcriptional activation of CYP1A1 and CYP1B1, two enzymes involved in the generation of 15-HETE PPARg ligand. Finally, we also demonstrate the importance of intact LRH-1 signalling in the anti-inflammatory and antifungal functions of alternatively activated macrophages, indicating that modulators of LRH-1 activity may have therapeutic potential to restrain infectious and inflammatory diseases.

Results
IL-13-mediated LRH-1 gene expression is dependent on STAT6. The anti-inflammatory properties of LRH-1 are well established in the liver and gut 24 . To elucidate whether LRH-1 also participates in regulating the inflammatory response in macrophages, gene expression profiling was performed. In situ hybridization and reverse transcriptase-quantitative PCR (RT-qPCR) revealed that LRH-1 (encoded by the Nr5a2 gene), known to be expressed in the colon and liver, is also expressed in macrophages but not in B and T immune cells (Fig. 1a). Consistent with the gene expression data, LRH-1 protein was also detected in macrophages ( Supplementary Fig. 1b). We next analysed the impact of pro-and anti-inflammatory factors on Nr5a2 gene expression in primary macrophages. As depicted in Fig. 1b, pro-inflammatory challenges, such LPS and IFNg exposure, but not IL-6, significantly reduced or abolished Nr5a2 mRNA expression. Conversely, IL-13, IL-4 and IL-10 cytokines significantly enhanced Nr5a2 mRNA level in macrophages. Similar to findings in the murine model, NR5A2 mRNA levels were significantly increased by IL-13 treatment in human monocytes (Fig. 1c). These results suggest that LRH-1 could be part of the transcriptional network mediating alternative activation of macrophages. To test this hypothesis, we analysed the downstream signalling components of IL-13 in more detail ( Fig. 1d-i). STAT6, a transcription factor known to be activated by IL-13 is part of the signalling pathway that governs alternative activation 25 . Interestingly, exposure of macrophages with AG490, a Jak-2/STAT6 inhibitor, prevented the IL-13-mediated induction of LRH-1 (Fig. 1d). Consistent with these observations, IL-13 failed to increase Nr5a2 mRNA and protein levels in macrophages deficient for STAT6 (Fig. 1e,f), suggesting that STAT6 mediates the transcriptional regulation of LRH-1. We then performed transient transfection assays in primary macrophages to assess the effect of IL-13 and STAT6 on Nr5a2 promoter activity. While 4 h of IL-13 exposure was already sufficient to induce Nr5a2 promoter activity in wild-type macrophages (Fig. 1g,h), chemical inhibition of STAT6 by AG490 (Fig. 1g) or genetic deletion of STAT6 (Fig. 1h) attenuated or even abolished this response.
To evaluate whether LRH-1 is subject to direct transcriptional control by STAT6, we performed an in silico analysis of the Nr5a2 promoter region. Scanning of the Nr5a2 promoter sequence for the STAT6 response element (STAT6-RE) canonical motif revealed four putative STAT6-RE (Fig. 1i). Chromatin immunoprecipitation (ChIP) analysis of macrophage DNA from C57BL/6 mice revealed specific recruitment of STAT6 to site 1 at À 541, which is most proximal to the transcription initiation site of the gene (Fig. 1i).
To explore the functionality of this site, we next modified by in vitro mutagenesis its sequence and we evaluated the mutated Nr5a2 reporter construct activity on IL-13 exposure (Fig. 1j).
Mutation of the STAT6-RE abolished the activity of the Nr5a2 reporter construct in response to IL-13 in Stat6 þ / þ and Stat6 M À / À macrophages (Fig. 1j). These results demonstrate that STAT6 directly controls the transcription of LRH-1 in response to IL-13.
LRH-1 is involved in IL-13-induced macrophage activation. In order to assess the role of LRH-1 in IL-13-induced alternative macrophage differentiation, we generated mice in which the Nr5a2 gene was selectively disrupted in myeloid-derived cells. To generate these animals, mice carrying floxed Lrh-1 alleles were crossed with transgenic mice that express the Cre recombinase under the control of the mouse phagocyte-selective lysozyme promoter 21,26 . Compared with control (Lrh-1 M þ / þ ) macrophages, LRH-1 mRNA and protein levels were almost undetectable in macrophages derived from the myeloid cellspecific LRH-1-deficient (Lrh-1 M À / À ) mice ( Supplementary  Fig. 1a-c). Furthermore, the disruption of LRH-1 could not be detected in other LRH-1-expressing tissues, such as the liver and the colon ( Supplementary Fig. 1d,e).
We then evaluated the expression of specific markers of classical and alternative activation in untreated or IL-13-treated  Nr5a2 mRNA expression in macrophages from C57BL/6 mice (b) and in human macrophages (c) treated with the indicated cytokines for 4 h, determined using RT-PCR. The results were represented in fold induction relative to the untreated control or wild-type littermate. (d,e) Nr5a2 mRNA expression in macrophages from C57BL/6 mice pretreated with AG490 and stimulated with IL-13 (d) and in macrophages from Stat6 þ / þ and Stat6 À / À mice stimulated with IL-13 for 4 h (e), determined using RT-PCR. The results were represented in fold induction relative to the untreated control or wild-type littermate. (f) Immunoblot analysis of the nuclear expression of LRH-1 and TBP (Tata-binding protein) in macrophages from Stat6 þ / þ and Stat6 À / À mice stimulated with IL-13 for 24 h. (g,h) Luciferase activity in macrophages from C57BL/6 mice transfected with LRH-1 (LRH-1-luc) promoter construct pretreated with AG490 (g) or from Stat6 þ / þ and Stat6 À / À mice (h), and treated with IL-13 for 24 h. The results were represented in fold induction relative to the untreated control or wild-type littermate. (i) Schematic presentation of the four putative STAT6 response elements in the mouse Nr5a2 promoter identified by Genomatix algorithm and assessment of STAT6 recruitment to site 1 and to the Arg1 promoter determined with the ChIP analysis using genomic DNA from C57BL/6 macrophages treated with IL-13 for 4 h. (j) Luciferase activity in macrophages from Stat6 þ / þ and Stat6 À / À mice transfected with LRH-1 promoter constructs and treated with IL-13 for 18 h. The results were represented in fold induction relative to the untreated control. Results correspond to mean ± s.e.m. of triplicates. Data are representative of three independent experiments. *Po0.05 **Po0.01 compared with the respective untreated control and x Po0.05, xx Po0.01 compared with IL-13-treated wild-type littermate. P values were determined using Bonferroni-Dunnett method.

15-HETE prodcution
Consistent with these findings, a strong decrease in Alox15, Cyp1a1 and Cyp1b1 expression could be observed in both untreated and IL-13-treated Stat6 À / À macrophages (Fig. 3h), further supporting the importance of STAT6 in the regulation of these genes.
To further explore whether STAT6 controls the expression of Alox15, Cyp1a1 and Cyp1b1 directly or indirectly through the induction of LRH-1, we performed an in silico analysis of Alox15, Cyp1a1 and Cyp1b1 promoters ( Supplementary Fig. 3a). This analysis revealed one putative LRH-1 and two putative STAT6-RE in the Alox15 promoter, with more than 95% of similarity to the consensus REs. Scanning of the Cyp1a1 and Cyp1b1 promoter sequences indicated the presence of conserved LRH-1 REs in both promoters, while no conserved STAT6 REs (matrix similarity o0.8) could be identified in these regulatory regions ( Supplementary Fig. 3a). Consistent with these findings, DLPC treatment increased Alox15, Cyp1a1 and Cyp1b1 gene expression in Lrh-1 M þ / þ macrophages, but not in Lrh-1 M À / À macrophages (Fig. 3i). These data confirm the importance of LRH-1 in the regulation of Alox15, Cyp1a1 and Cyp1b1.
To define whether Cyp1a1 and Cyp1b1 are direct transcriptional targets of LRH-1, transfection assays in Lrh-1 M þ / þ and Lrh-1 M À / À macrophages were performed using a luciferase reporter containing ± 1.2 kb of the promoter of the Cyp1a1 and Cyp1b1 genes. IL-13 exposure of Lrh-1 M þ / þ macrophages

Cyp1b1wt
Cyp1b1mt Cyp1a1 wt resulted in an eightfold increase in reporter activity of both Cyp1a1 and Cyp1b promoters (Fig. 4a,c). Interestingly, genetic deletion of LRH-1 abolished this response, demonstrating that Cyp1a1 and Cyp1b1 promoters are directly activated by LRH-1.
To identify the critical LRH-1 REs in the Cyp1a1 and Cyp1b1 promoters, we mutagenized the putative RE that were found by in silico analysis ( Supplementary Fig. 3a), and their response to LRH-1 on IL-13 exposure was compared (Fig. 4a-c). For the Cyp1a1 promoter, mutation of the first LRH-1 RE (site 1) abolished the activity of the reporter construct in response to IL-13, whereas mutation of site 2 was still responsive in Lrh-1 M þ / þ macrophages (Fig. 4a). Furthermore, whole inhibition of mutated reporter construct activities in Lrh-1 M À / À macrophages established that site 1 is the principal site transmitting the effect of LRH-1 on the Cyp1a1 promoter. Thus, this result identified specific recruitment of LRH-1 to site 1, which is most distal to the transcription initiation site in the Cyp1a1 promoter.
For the Cyp1b1 promoter, IL-13 treatment failed to increase the activity of the mutated Cyp1b1 reporter in both Lrh-1 M þ / þ or Lrh-1 M À / À macrophages (Fig. 4c), indicating that LRH-1 binds and activates the Cyp1b1 promoter through a unique sequence between À 742 and À 728 bp upstream of the transcription initiation site of the gene. Finally, ChIP assays were performed. IL-13 enhanced the recruitment of LRH-1 on both Cyp1a1 and Cyp1b1 sites in Lrh-1 M þ / þ macrophages, but not in Lrh-1 M À / À macrophages ( Fig. 4b-d). Altogether, these results demonstrate that LRH-1 directly binds Cyp1a1 and Cyp1b1 promoters and hence controls the transcription of Cyp1a1 and Cyp1b1 genes in response to IL-13.
LRH-1/CYP1-dependent 15-HETE release induces PPARc activation. To further determine whether the generation of 15-HETE metabolites through LRH-1 are involved in PPARg activation, we assessed whether supplementation of 15-HETE can rescue the loss of PPARg activation in Lrh-1 M À / À macrophages. In contrast to IL-13, which could not induce PPARg activation in Lrh-1 M À / À macrophages, addition of exogenous 15-HETE efficiently restored the induction of both a PPRE luciferase reporter (Fig. 5a) and of PPARg target genes such as Mrc1, Clec7a and Cd36 (Fig. 5b) in Lrh-1 M À / À macrophages, indicating that the PPARg activation through LRH-1 is critically dependent on 15-HETE production.
To confirm that 15-HETE production through the LRH-1/ CYP1 axis induces PPARg activation, we evaluated PPARg activation in macrophages silenced for Cyp1a1 and Cyp1b1. Interestingly, PPARg activity as determined by the induction of a PPRE luciferase reporter (Fig. 5c) and the induction of PPARg target genes (Fig. 5d) by IL-13 were totally inhibited in macrophages deficient for Cyp1a1 and Cyp1b1. Moreover, the induction of a PPRE luciferase reporter (Fig. 5c) and of PPARg target genes (Fig. 5d) was still significantly enhanced by IL-13 in Alox15 À / À macrophages, showing that the 12/15 lipoxygenase is not required for PPARg activation mediated by LRH-1. These data are in support of a critical role of CYP1A1 and CYP1B1 in LRH-1-mediated PPARg activation through 15-HETE synthesis.
IL-13 activation of macrophages requires STAT6/LRH-1/PPARc. To determine whether STAT6 controls both directly the transcription of markers of IL-13-mediated alternative activation and indirectly through the activation of the LRH-1/PPARg axis, we studied the mRNA level of alternative activation markers in STAT6-deficient macrophages. IL-13-augmented induction of Arg1, Chi3l3 (YM1), Retnla (Fizz1), MR, Clec7a and CD36 was detected in Stat6 þ / þ macrophages but not in Stat6 À / À macrophages (Fig. 5e). The lack of IL-13-augmented induction of alternative markers was associated with a failure of Stat6 À / À macrophages to produce 15-HETE in response to IL-13 (Fig. 5f). Interestingly, the addition of exogenous 15-HETE restored the induction of alternative polarization markers in Stat6 À / À macrophages and not in Pparg M À / À macrophages ( Fig. 5e and Supplementary Fig. 3c). These data suggest that STAT6 is required for induction of macrophage-alternative activation markers and further support the existence of a PPARg-dependent mechanism in the regulation of these genes.
Moreover, induction of Arg1, Retnla (Fizz1) and Chi3l3 (YM1) in response to IL-13 was slightly decreased in Pparg M À / À macrophages, whereas the induction of Mrc1 (MR), Clec7a and CD36 was completely abrogated in Pparg M À / À macrophages (Fig. 5g). These results indicate the existence of distinct regulatory mechanisms involving either STAT6 with a modest contribution of PPARg or predominantly controlled by the LRH-1/PPARg axis. In line, the overexpression of Mrc1, Clec7a and Cd36 after treatment with DLPC in Pparg M þ / þ macrophages was not detected in Pparg M À / À macrophages (Fig. 5h), clearly establishing that LRH-1 acts upstream from PPARg in the signalling cascade leading to the PPARg-dependent gene expression.

IL-13-induced fungicidal properties of macrophages via LRH-1.
Previous work from our laboratory established the importance of PPARg in the fungicidal functions of alternatively activated macrophages 27 . On the basis of the current findings suggesting a role for LRH-1 in PPARg-mediated alternative polarization following IL-13 stimulation, we next investigated whether deletion of LRH-1 in macrophages could have an impact on the outcome of Candida albicans infection. The severe systemic infection of mice with C. albicans resulted in a significantly lower survival rate of Lrh-1 M À / À mice compared with Lrh-1 M þ / þ mice (Po0.001; Fig. 6a), supporting a role for LRH-1 in antifungal defence. To further explore the exact function of LRH-1 in the pathophysiology of fungal infection, we evaluated the fungal burden in the intestinal tract and the macrophage microbicidal functions in a murine experimental model of gastrointestinal candidiasis. Lrh-1 M À / À mice infected with C. albicans had more severe gastrointestinal infection than their wild-type littermates and showed worsened fungal burden in the caecum (Fig. 6b). Remarkably, IL-13, 15-HETE, as well as DLPC, diminished C. albicans gastrointestinal colonization in Lrh-1 M þ / þ mice. However, these effects were lost in Lrh-1 M À / À mice treated with IL-13 or DLPC, but not when the PPARg ligand, 15-HETE, was administered to the animals (Fig. 6b).
To investigate whether LRH-1 in macrophages has any relevant microbicidal phenotype, we evaluated the capacity of Lrh-1 M þ / þ and Lrh-1 M À / À macrophages to kill yeasts in vitro. Compared with Lrh-1 M þ / þ macrophages, Lrh-1 M À / À macrophages showed a defect in their ability to kill C. albicans, demonstrating the contribution of LRH-1 in macrophage-intrinsic antifungal activity (Fig. 6c). Consistent with our observation, Lrh-1 M À / À macrophages were less efficient in engulfing C. albicans and producing reactive oxygen species (ROS) after fungal challenge (Fig. 6d,e). Moreover, the defect of Lrh-1 M À / À macrophages to exert their antifungal activity was correlated with lower MR and Dectin-1 protein levels after C. albicans challenge ( Supplementary  Fig. 3d). As expected, treatment with IL-13 of Lrh-1 M þ / þ macrophages increased the killing and the phagocytosis of C. albicans and also ROS production in response to C. albicans. These inductions were abrogated in Lrh-1 M À / À macrophages, underscoring the importance of LRH-1 in these fungicidal functions (Fig. 6c-e). Similar effects were obtained when macrophages were stimulated with DLPC (Fig. 6c-e).
To unequivocally establish that the LRH-1/CYP1/HETE axis is involved in macrophage-intrinsic antifungal activity of IL-13, we evaluated the ability of macrophages silenced for Cyp1a1 and Cyp1b1 (Cyp1) to kill C. albicans. Interestingly, the increase in C. albicans killing by IL-13 and DLPC was inhibited by the simultaneous gene silencing for Cyp1a1 and Cyp1b1 (Cyp1), but not after 15-HETE stimulation (Fig. 6g). Taken together, these data provide in vivo evidence that LRH-1 is involved in the PPARg-dependent antifungal functions elicited by IL-13 through CYP1-induced 15-HETE production.

Discussion
The nuclear receptor PPARg is essential for IL-13-induced alternative differentiation of macrophages 6,28,29 . We have previously demonstrated that IL-13, via the cPLA 2 signalling pathway, induced AA mobilization associated with the nuclear localization of 15d-PGJ2, an endogenous PPARg ligand 5 . Once activated, PPARg induces the transcription of Dectin-1, MR and CD36, three genes characteristic of the alternative activation 5,30,31 . Therefore, the processes leading to PPARg activation, such as AA release and its subsequent metabolic conversion, could be important aspects of alternative polarization because they are limiting factors for PPARg ligand synthesis.
AA can be metabolized by the COX1/COX2 cyclooxygenases to PGH2, which in turn is transformed by the PGD synthase into 15d-PGJ2 (refs 32,33). AA can also be directly metabolized to 12-and HETEs, other endogenous PPARg ligands, through 12/15 lipoxygenases 34 . A third pathway of AA metabolism leading to endogenous PPARg ligand production is associated with its conversion by the enzymes of the CYP family [35][36][37] . The CYP enzymes generate two biological and active classes of eicosanoids, the epoxy (EETs) and hydroxy (HETEs) derivatives 10,11 . The CYP1 family is mainly involved in the formation of mid-chain HETEs, such as 12-and 15-HETEs, through CYP1A1 and CYP1B1 (refs 12,13). Cd36

15-HETE production
Relative mRNA expression level Here we report that the nuclear receptor LRH-1 is expressed in macrophages and in response to IL-13 directly binds CYP1A1 and CYP1B1 promoters to positively regulate their transcription. Moreover, 15-HETE production following IL-13 stimulation is impaired in macrophages deficient for LRH-1 and not in macrophages lacking 12/15 lipoxygenase, indicating that LRH-1 drives the generation of 15-HETE metabolites through its impact on CYP1 gene expression. Consistently, our findings showing that the concurrent gene silencing of Cyp1a1 and Cyp1b1 in macrophages abolishes the generation of 15-HETE, provide evidence that its production through the LRH-1/CYP1s axis is crucial in PPARg activation. This is corroborated by the findings that PPARg activation on IL-13 stimulation is lost in macrophages silenced simultaneously for CYP1A1/CYP1B1 and restored by the addition of exogenous 15-HETE in macrophages lacking LRH-1. Consistent with these observations, treatment of macrophages with the LRH-1 agonist, DLPC, increased the expression of CD36, MR and Dectin-1 PPARg target genes in wild-type macrophages but not in macrophages lacking PPARg. Altogether, these results establish that PPARg activation by IL-13 is dependent on the LRH-1/CYP1/15-HETE pathway. Another endogenous activator to consider in PPARg activation is 15d-PGJ2. Although we have previously shown that IL-13 generates 15d-PGJ2 production and its nuclear localization in macrophages 5 , the results in this study suggest that it is not sufficient to activate PPARg. This is supported by previous reports showing that 15d-PGJ2 concentration required to stimulate PPARg is in the mM range, in contrast to other prostaglandins that are normally active at low nM concentrations 38,39 . Thus, the levels generated in vivo are not sufficient to be compatible with a role for this metabolite as an endogenous PPARg ligand 38,40 .
Despite the growing knowledge with regard to the biological function of LRH-1, little is known about how LRH-1 is controlled at the transcriptional level. We identified STAT6 as a transcriptional regulator of LRH-1. This was evidenced by the induction of LRH-1 promoter activity by binding of STAT6 to its RE in the LRH-1 promoter and by the decrease in LRH-1 mRNA and protein levels in macrophages lacking STAT6. On the basis of the established role of STAT6 in PPARg activation and macrophage polarization 41 , these findings identify LRH-1 as a critical component in the signalling cascades that drive PPARgmediated alternative macrophage activation. This was further highlighted by the fact that macrophages lacking LRH-1 present an increase in pro-inflammatory cytokines and the simultaneous expression of other M1 markers. The involvement of LRH-1 in anti-inflammatory responses was supported by the robust reduction of LRH-1 gene expression in response to Th1 cytokines and conversely by the upregulation by Th2 cytokines. Interestingly, LRH-1 was also induced in human macrophages in response to the Th2 cytokine IL-13 via a mechanism that is most likely also STAT6-dependent, given the presence of several conserved STAT6 REs in the human LRH-1 promoter (Supplementary Fig. 3a). Our findings may further explain why during Crohn's disease, characterized by a Th1 cytokine profile, mRNA expression levels of LRH-1 are lower than in ulcerative colitis, characterized by a Th2 immune response 21 . Consistent with the anti-inflammatory role of LRH-1, IL-13-induced alternative activation was impaired in macrophages lacking LRH-1. Indeed, on IL-13 treatment, the induction of several signature genes of alternative activation, including Arginase 1, YM1, IL-1 receptor antagonist (IL-1ra), MR, Dectin-1 and CD36, was significantly impaired in macrophages lacking LRH-1. This is in agreement with reports showing that LRH-1 controls the expression of anti-inflammatory IL-1ra and the scavenger receptor class B type I, two markers specific of alternatively activated macrophages 22,23,42 .
In addition to the key role of LRH-1 in the acquisition of alternative activation of macrophages, this study also provides mechanistic insight into the hierarchy between STAT6, LRH-1 and PPARg to achieve this phenotype. Our findings showing that loss of induction of alternative activation markers in Stat6 À / À macrophages can be restored by exogenous 15-HETE support the notion that STAT6 is required for macrophage-alternative activation through PPARg-dependent mechanism. Moreover, the use of Pparg M À / À macrophages provides evidence for the existence of distinct mechanisms in the transcriptional regulation of genes characteristics of alternative activation. Our results demonstrate that the transcriptional regulation of Arginase 1, Fizz 1 and YM1 involves directly STAT6 with a modest contribution of PPARg and that Dectin-1, MR and CD36 are regulated indirectly by STAT6 through the LRH-1/PPARg axis. These observations are not only consistent with the requirement of STAT6 to induce the majority of PPARg target genes 41 but also with the identification of PPARg as a positive regulator of alternative activation 6 .
Consistent with the involvement of the LRH-1/PPARg pathway in inducing MR and Dectin-1 expression during IL-13mediated alternative activation, loss of LRH-1 and PPARg in macrophages also severely compromised their capacity to kill, to engulf C. albicans and to produce ROS. This is in line with the fact that LRH-1 is upstream from PPARg in the signalling pathway leading to the induction of MR and Dectin-1, two C-type lectin receptors strongly involved in the antifungal functions of macrophages against C. albicans 5,27,31 . LRH-1 deficiency in myeloid cells also rendered the mice highly susceptible to gastrointestinal and systemic C. albicans infection, highlighting LRH-1 of myeloid lineage as a key effector of host fungicidal functions. Although we have not characterized the role of neutrophils in this infectious context, our in vitro and in vivo results identify LRH-1 as a nuclear receptor indispensable for alternative activation of macrophages and for its associated antifungal functions.
In conclusion, we have shown that loss of LRH-1 in macrophages prevents IL-13-induced alternative activation of macrophages, demonstrating the pivotal role of LRH-1 in the differentiation of macrophages towards an anti-inflammatory and antifungal phenotype. In response to IL-13, LRH-1 expression is increased in macrophages through STAT6 and controls the expression of CYP1A1 and CYP1B1 enzymes, which catalyses the generation of 15-HETE PPARg ligand. Altogether, these results establish that the alternative polarization of macrophages by IL-13 is dependent on the STAT6/LRH-1/CYPs/15-HETE/ PPARg axis (Fig. 7). Finally, deletion of LRH-1 in myeloid cells renders mice susceptible to gastrointestinal and systemic C. albicans infection, highlighting LRH-1 as a critical factor for antifungal functions. Synthetic agonists of LRH-1 activity may, hence, constitute promising compounds for the treatment of anti-infectious and anti-inflammatory diseases.

Methods
Mice. Male mice aged 10-12 weeks on C57BL/6 background were used for in vitro and in vivo experiments. Mice were bred and handled by following protocols approved by the Conseil Scientifique du Centre de Formation et de Recherche Experimental Médico Chirurgical and the ethics board of the Midi-Pyrénées ethic committee for animal experimentation (Experimentation permit number 31-067, approval no. B3155503). All cages were changed twice weekly, and all manipulations of the animals were carried out in a laminal blow hood under aseptic conditions. The photoperiod was adjusted to 12-h light and 12-h dark. C57BL/6 mice were purchased from Janvier (France) and Stat6 À / À mice and ALox15 À / À mice were purchased from Jackson Laboratories. Pparg M À / À mice deleted for Pparg specifically in macrophages have been described earlier 30,43 . Nr5a2 (encoding LRH-1) macrophage specific knockout mice (referred as Lrh-1 M À / À mice) were obtained by crossing mice carrying floxed Lrh-1 alleles with transgenic mice expressing the Cre recombinase under the control of the mouse phagocyte-selective lysozyme promoter 21,26 . For Lrh-1 M À / À and Pparg M À / À mice, the corresponding floxed littermates were used as controls throughout all the experiments. Corresponding littermates were used as controls for Stat6 À / À and ALox15 À / À mice.
For the in vivo experiments, a gastrointestinal infection with the C. albicans strain was established by gavage with 50 Â 10 6 C. albicans per mouse (n ¼ 10 per group). Mice were treated or not intraperitoneally (i.p.) with IL-13 (Clinisciences), DLPC (Sigma) or 15-HETE (Cayman). For IL-13 treatment, injections of 4 mg per mouse were performed 1 day before and 3 days after the infection with C. albicans (two injections). For DLPC (300 mg per 10 g of mouse) and 15-HETE (28 mg per 10 g of mouse), i.p. injections were realized 1 day before the day of the infection with C. albicans and then every 2 days (five injections). Control groups received saline solution only with DMSO. After 6 days of infection, the ceca were removed aseptically for the experiments.
For C. albicans systemic infection, yeasts were administered i.p. (100 Â 10 6 yeasts per mouse). Survival studies were conducted using 32 mice per group and were repeated twice.
Human macrophages. Monocytes were obtained from healthy blood donors (Etablissement Franc¸ais du Sang, EFS Toulouse). Written informed consents were obtained from the donors under EFS contract no. 21/PVNT/TOU/UPS04/2010-0025. Following articles L1243-4 and R1243-61 of the French Public Health Code, the contract was approved by the French Ministry of Science and Technology (agreement no. AC 2009-921). Human peripheral blood mononuclear cells were isolated from the blood of healthy volunteers by a density gradient centrifugation method on Lymphoprep (Abcys). Monocytes were isolated by adherence to plastic for 2 h in SFM (Gibco) at 37°C, 5% CO 2 . The macrophages were obtained after 3 days of culture only in SFM medium.
Preparation of mouse resident peritoneal macrophages. After being killed, resident peritoneal cells were harvested by washing the peritoneal cavity with 5 ml of sterile NaCl 0.9%. Collected cells were centrifuged at 1,500 r.p.m. for 10 min and the cell pellet was suspended in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with glutamine (Invitrogen), penicillin, streptomycin (Invitrogen) and 5% heat-inactivated fetal calf serum. Cells were allowed to adhere for 2 h at 37°C and 5% CO 2 . Nonadherent cells were then removed by washing with PBS.
In situ hybridization. In situ hybridization was performed with digoxigeninlabelled RNA probe (Plasmid pBSSK Lrh-1) as previously described 44 . Briefly, this manual nonradioactive method allows to detect specific complementary mRNA sequences at the cellular level using digoxigenin-labelled probes in a five-step procedure: hybridization of the probe to pretreated tissue at 65°C; post-hybridization stringent washes; blocking steps to prepare for the immunodetection; primary antibody anti-DIG-AP incubation; and colorimetric enzymatic detection. The detection step lasts for 2-3 days.
Western blot analysis. Nuclear protein extracts were prepared, and lysates were subjected to western blotting as described previously 45 . Briefly, nuclear protein lysates were extracted following standard procedures. Protein extracts were separated using SDS-PAGE. After protein transfer, membranes were incubated overnight at 4°C with either a rabbit anti-Lrh-1 (ref. 21; 1/1,000), a rabbit polyclonal anti-Tbp (Abcam, ab63766, ½,000), a rabbit anti-Cyp1b1 (Santa Cruz, sc-133490, 1/200) or a Actin (Santa Cruz, sc-1615, 1/1,000) and then for 1 h at room temperature with a peroxidase conjugated secondary antibody. Membranes were washed, and proteins were visualized with the SuperSignal West Pico Chemiluminescent Substrate (ThermoScientific). Images have been cropped for presentation. Full-size images are presented in Supplementary Fig. 4.
For siRNA experiments, mouse Cyp1a1 and Cyp1b1 and control siRNA were purchased from Origene. Macrophages were incubated with 20 nM of control siRNA or Cyp1a1 and Cyp1b1 siRNA and with Lipofectamine 2000 (Invitrogen) for 18 h according to the manufacturer's instructions. Cells were then stimulated with IL-13, DLPC or 15-HETE for 18 h.
Reporter assays and site-directed mutagenesis. Genomic DNA was extracted from mouse kidneys and the corresponding Cyp1a1 and Cyp1b1 promoter fragments were amplified by PCR with primers containing KpnI and XhoI restriction sites. The PCR products were then cloned into pENTR-D/TOPO (Invitrogen/ Lifetechnologies) plasmids, which were digested with KpnI and XhoI and then ligated into pGL4.12 (Promega) reporter plasmids.
Mutagenesis was carried out with the GeneArt mutagenesis kit (Invitrogen/ Lifetechnologies) according to the manufacturer's instructions. The sequences of reporter constructs were analysed and confirmed.
Flow cytometry. The analysis was performed on nonadherent macrophages 47 harvested by washing the peritoneal cavity with 5 ml of sterile NaCl 0.9%. Collected cells were centrifuged at 1,500 r.p.m. for 10 min and the cell pellet was suspended in PBS medium supplemented with 1% fetal calf serum (FCS). Surface expressed Dectin-1 or CD36 was detected, respectively, using fluoroscein isothiocyanate (FITC)-Dectin-1 monoclonal antibody (mAb; Serotec MCA2289F, 1/100) or PE-CD36 mAb (Santa Cruz, sc-13572, 1/100) and was compared with an irrelevant appropriate isotype control. To evaluate the mannose receptor (MR) surface expression, we have used MR-specific ligand conjugated with FITC (Sigma A7790, 1 mg ml À 1 ). All stainings were performed on PBS-1% FCS medium. A population of 10,000 cells was analysed for each data point. All analyses were carried out in a Becton Dickinson FACScalibur using the CellQuestPro software.
ELISA Cytokine titration and EIA quantification of 15-HETE. Peritoneal macrophages were stimulated with IL-13 for 18 h and challenged with nonopsonized C. albicans at a yeast-to-macrophage ratio of 3:1 for 8 h. The production of TNF-a, IL-1b and TGF-b in the cell supernatants was determined with a commercially available OptiEIA kit (BD Biosciences) according to the manufacturer's instructions.
For 15-HETE quantification, the macrophages were stimulated with IL-13 for 18 h and 15-HETE were measured using EIA as recommended by the manufacturer's protocol (15(S)-HETE EIA kit, Cayman).
AA mobilization. Peritoneal macrophages were prelabelled with [ 3 H]AA (1 mCi per well, Perkin Elmer) for 18 h. The prelabelled macrophages were then treated with IL-13 (50 ng ml À 1 ) for 1 h. The cellular lipids were extracted twice with hexane/isopropanol (3:2, v/v) and the [ 3 H]AA content in membrane phospholipids was quantified by measurement of the radioactivity by beta liquid scintillation counting, as described with minor adaptations 48 .
C. albicans strain. The strain of C. albicans used throughout these experiments was isolated from a blood culture of a Toulouse-Rangueil Hospital patient 5 . Fluorescent C. albicans was prepared by adding C. albicans to FITC (Sigma) dissolved in sodium carbonate buffer (pH 9.5) at room temperature for 3 h and washed by centrifugation three times in sodium carbonate buffer before storage in aliquots of water at 4°C.
Phagocytosis assay and ROS quantification. For analysis of phagocytosis of C. albicans, cultured macrophages were pretreated or not with IL-13, 15-HETE or DLPC for 18 h and then challenged with six FITC-labelled yeasts per macrophage. Phagocytosis was initiated at 37°C with 5% CO 2 and stopped after 1 h by washing the macrophages with ice-cold PBS. The number of C. albicans engulfed by macrophages was determined with fluorescence quantification using the Envision (Perkin Elmer) fluorimetry-based approach.
The oxygen-dependent respiratory burst of macrophages (ROS production) was measured by chemiluminescence in the presence of 5-amino-2,3-dihydro-1,4phthalazinedione (luminol) using a thermostatically (37°C) controlled luminometer (Wallac 1420 Victor2). The generation of chemoluminescence was monitored continuously for 1 h after incubation of the cells with luminol (66 mM) and pretreatment with IL-13, 15-HETE or DLPC for 18 h and challenge with C. albicans (yeast-to-macrophage ratio: 3:1). Statistical analysis was performed using the area under the curve expressed in counts Â seconds.
Killing assay. The killing assay was performed as previously described 49 . Cells were allowed to interact for 30 min at 37°C with C. albicans (at a ratio of 0.3 yeast per macrophage) and unbound yeasts were removed by four washes with medium. Macrophages were then incubated at 37°C for 4 h. Control plates were kept at 4°C to provide a measure of live C. albicans in the wells. After incubation, the medium was removed and cells were lysed by incubation for 5 min at 25°C with water at a pH of 11. An excess of PBS was used to neutralize the lysis buffer, and CFU C. albicans was determined by plating on Sabouraud plates and incubation overnight at 37°C.
Quantification of C. albicans in the caecum. Cell lysis and DNA extraction. After mouse infection, ceca were aseptically removed and then crushed using lysing matrix tubes (MP Biomedicals). Tissue sample homogenate (250 ml) was resuspended in 200 ml of lysis buffer for 2 h at 65°C and DNA was then extracted with isopropanol and eluted with an elution buffer (High Pure PCR Template preparation kit, Roche Diagnostics).
Light cycler-based PCR assay. The Light Cycler PCR and detection system (Roche Diagnostics) was used for amplification and online quantification. PCR analysis was performed as described previously 27 . Serially diluted samples of genomic fungal DNA obtained from C. albicans cultures (40 Â 10 6 cells) were used as external standards in each run. Cycle numbers of the logarithmic linear phase were plotted against the logarithm of the concentration of template DNA to evaluate the number of yeast cells present in each tissue sample homogenate.