Coactivation of Syk Kinase and Myd88 Adaptor Protein Pathways by Bacteria Promotes Regulatory Properties of Neutrophils

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Immunity Article

Coactivation of Syk Kinase and MyD88 Adaptor Protein Pathways by Bacteria Promotes Regulatory Properties of Neutrophils

Xiaoming Zhang,1,2 Laleh Majlessi,1,2 Edith Deriaud,1,2 Claude Leclerc,1,2 and Richard Lo-Man1,2,* 1Institut Pasteur, Unite´ Re´ gulation Immunitaire et Vaccinologie, 25 rue du Dr. Roux, 75724 Paris Cedex 15, France 2Inserm, U883, 75724 Paris, France *Correspondence: [email protected] DOI 10.1016/j.immuni.2009.09.016

SUMMARY has improved in recent years, and they include RNA helicases, intracellular sensors for viruses (Kawai and Akira, 2006), and Neutrophils are one of the first lines of defense Nod-like receptors (NLRs) that recognize mostly bacterial against microbial pathogens and are rapidly recruited components (Fritz et al., 2006). C-type lectin receptors (CLRs) at the infection site upon inflammatory conditions. make up another large family of cell-surface molecules with We show here that after bacterial stimulation, and in a carbohydrate-recognition domain and contribute mainly to contrast to monocytes and macrophages, murine phagocytosis of microbes, and most of them lack reported neutrophils contributed poorly to inflammatory re- signaling functions. A number of CLR can signal directly or indirectly through an immunoreceptor tyrosine-based activation sponses; however, they secreted high amounts of motif (ITAM). Clec2 (Suzuki-Inoue et al., 2006) and Clec7A (Dec- the anti-inflammatory cytokine IL-10 in a DAP12 tin-1) (Rogers et al., 2005) can activate cells directly through an adaptor-Syk kinase and MyD88 adaptor-dependent ITAM-like motif in their cytoplasmic region. Alternatively, CLR- manner. Cotriggering of TLR-MyD88- and C-type lec- mediated cell activation may occur indirectly through the interac- tin receptor (CLR)-Syk-dependent pathways led to tion of their charged residues in the transmembrane region with a quick and sustained phosphorylation of p38 MAP the ITAM-containing DAP12 for Clec5A (MDL-1) (Bakker et al., and Akt kinases in neutrophils. In vivo, both Gram- 1999) or the FcRg chain for Clec4e (Mincle) (Yamasaki et al., negative bacteria and mycobacteria induced the 2008) and DCAR (Kanazawa et al., 2003). In all cases, it activates recruitment of neutrophils secreting IL-10. In acute spleen tyrosine kinase (Syk)-dependent signaling pathways in mycobacterial infection, neutrophil-derived IL-10 innate cells. Clec7A (Dectin-1) recognizes b-glucans in fungal controlled the inflammatory response of dendritic cell walls and is important for antifungal immunity (Saijo et al., 2007; Taylor et al., 2007). Clec4e (Mincle) also recognizes fungi cells, monocytes and macrophages in the lung. Malassezia (Yamasaki et al., 2009), but also Candida albicans, During a chronic infection, neutrophil depletion and contributes to the protection against candidiasis (Wells promoted inflammation and decreased the mycobac- et al., 2008). Clec5A (MDL-1) has been recently shown to recog- terial burden. Therefore, neutrophils can have a previ- nize Dengue virus and to play an important role in the resistance ously unsuspected regulatory role during acute and to infection (Chen et al., 2008). chronic microbial infections. Most known human TLRs are expressed by neutrophils (Hay- ashi et al., 2003). Stimulation of human neutrophils with TLR2, TLR4, TLR5, TLR7, TLR8, and TLR9 agonists induces rapid INTRODUCTION activation as evidenced by L-selectin (CD62L) shedding, CD11b integrin upregulation, IL-8 secretion, and a respiratory Neutrophils are the first cells to migrate to the site of infection to burst (Hayashi et al., 2003). Murine neutrophils express TLR2, eliminate microorganisms. After activation, neutrophils sequen- TLR4, and TLR9 mRNAs (Tsuda et al., 2004) and respond to tially discharge granules, which contain a large panel of antimi- LPS by shedding L-selectin and upregulation of CD11b (Andone- crobial agents. They also recruit other innate cells including gui et al., 2003). Several studies have identified murine neutro- monocytes and dendritic cells (DCs) at infection sites through phils as a major source of cytokines, including TNF-a (Bennouna the release of chemokines and of antimicrobial peptides with and Denkers, 2005) and IL-12 (Romani et al., 1997; Bliss et al., chemotactic properties (Scapini et al., 2000). 2000), after stimulation by LPS or Toxoplasma gondii. The mAb After pathogen encounter, innate immune cells, including RB6-8C5 (Tepper et al., 1992) has been extensively used for neutrophils, are activated via a limited number of germline-en- studying murine neutrophil functions in many murine disease coded pattern-recognition receptors (PRRs). Among the best- models. RB6-8C5 binds to Gr-1 (Tepper et al., 1992), corre- documented PRRs are Toll-like receptors (TLRs) (Akira et al., sponding to both Ly6G, a specific granulocyte surface marker, 2006). Myeloid differentiation factor-88 (MyD88)-dependent and Ly6C, an antigen broadly expressed on immune cells, TLR-triggering leads to cell activation and the release of proin- including neutrophils, eosinophils, monocytes, dendritic cells, flammatory mediators. Our understanding of non-TLR PRRs T cells, and NK and NKT cells. However, RB6-8C5 can be

Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. 761 Immunity Bacteria Induce IL-10 Producing Neutrophils

used for in vivo depletion of neutrophils and also to deplete plas- macytoid DCs, T cells, or monocytes, clearly showing that RB6- 8C5 is inappropriate for the study of murine neutrophil functions. In this work, we used the Ly6G-specific antibodies 1A8 (Fleming et al., 1993) and NIMP-R14 (Lopez et al., 1984) for a systematic analysis of the innate responses of murine neutrophils to bacteria and to various agonists for PRRs. We demonstrate that highly purified murine neutrophils can be activated through TLR or CLR pathways, but produced minimal amounts of proinflammatory cytokines. In contrast, neutrophils produced large amounts of the anti-inflammatory IL-10 after microbial stimulation due to coactivation of MyD88 and Syk pathways. In the context of mycobacterial infections, neutrophils negatively regulated local lung inflammation in vivo through IL-10 and coun- teracted the control of the mycobacterial burden during the chronic phase of the infection. Our findings thus reveal an unex- pected and unappreciated role of neutrophils in downmodulat- ing immune responses.

RESULTS

Murine Neutrophils Produce Large Amounts of IL-10 in the Context of Bacterial Infection The study of murine neutrophils has been confused by the use of Gr-1 antibodies (clone RB6-8C5) to identify neutrophils: these antibodies recognize both Ly6G and Ly6C antigens (Fleming et al., 1993), and the latter are expressed by monocytes and dendritic cells. The CD11b+Gr-1hi cells in mouse bone marrow (BM), blood, spleen, and peritoneum can be clearly separated into two populations according to Ly6G expression (Figure 1A and Figure S1 available online): after FACS-sorting, May-Gru¨ n- wald-Giemsa staining clearly showed that only Gr-1hiLy6Ghi cells are true neutrophils, whereas most Gr-1hiLy6GÀ cells show monocyte morphology (Figure 1A and Figure S2). Further analysis showed that both Gr-1hiLy6Ghi and Gr-1hiLy6GÀ cells were CD11b+, but Gr-1hiLy6Ghi cells were CD115À/low, whereas Gr-1hiLy6G- cells were strongly CD115+ (Figure 1A and Figure S1). We next assessed the capacity of various TLR agonists to activate purified Gr-1hiLy6Ghi neutrophils. Pam3, LPS, and flagellin readily activated neutrophils in a MyD88-dependent manner, as shown by their extended morphology, upregulation of CD11b, downregulation of CD62L, and generation of reactive oxygen and nitrite intermediates (Figure 1B and Figure S3). We tested the relative capacity of neutrophils and monocytes to produce proinflammatory cytokines. Stimulated BM cells were Figure 1. TLR-Activated Neutrophils Are Weak Producers of Proinflammatory Cytokines stained for intracellular TNF-a and IL-12: it was clear that among hi (A) C57BL/6 mouse BM cells were stained with anti-Ly6G, anti-Gr-1, Gr1 cells, monocytes, but not neutrophils, were the major anti-CD11b, and anti-CD115 and analyzed by flow cytometry. The Gr1hi, Gr1hi source of cytokine production in response to agonists for Ly6Ghi, and Gr1hiLy6GÀ populations were further sorted by FACS and stained TLR2, TLR4, TLR5, TLR7, and TLR9 (Figure 1C and Figure S4A). with May-Gru¨ nwald-Giemsa dye. We then purified neutrophils and monocytes to 99% purity from hi (B) Purified BM Ly6G neutrophils were loaded with DHR123 and stimulated BM and found that neutrophils did not produce IL-12 p40 with the indicated TLR agonists for 1 hr to allow detection of respiratory bursts by flow cytometry. (Figure 1D), IL-1b, IL-6, or CXCL1 (Figure S4B) in response to (C) BM cells were stimulated with various TLR agonists for 6 hr, then intracel- any of the TLR agonists; only minute amounts of TNF-a were lular TNF-a and IL-12 were detected by gating on Gr1hi cells. detected when neutrophils were stimulated with TLR2 and (D) BM neutrophils (105) and monocytes (5 3 104) were stimulated with various TLR4 agonists. In contrast, Gr-1hiLy6GÀCD115+ monocytes TLR agonists or 50 ng/ml PMA for 24 hr. TNF-a and IL-12 p40 in culture produced large amounts of TNF-a, IL-12, IL-6, and CXCL1 in supernatants were detected by ELISA. Results are shown as mean of response to all TLR agonists (Figure 1D and Figure S4B). These duplicates ± SD. results were confirmed with neutrophils purified from blood,

762 Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. Immunity Bacteria Induce IL-10 Producing Neutrophils

Figure 2. Bacterial Infections Promote an Anti-inflammatory Cytokine Response from Neurophils In Vitro and In Vivo (A) BM neutrophils and monocytes were stimulated with various TLR agonists. The supernatants were collected 24 hr later and tested for IL-10. (B) BM neutrophils were stimulated with live or heat killed (HK) M. bovis BCG at an MOI of 10, and then the respiratory burst, CD11b expression level and IL-10 production were measured. Shaded, medium alone; red, live BCG; and blue, heat-killed BCG. (C) BM neutrophils and monocytes were stimulated with live BCG for 24 hr. Supernatants were tested for IL-10, TNF-a, IL-1b, IL-6, IL-12 p40, IL-12 p70, CCL2, CXCL1, and CXCL2. (D) Neutrophils purified from BM, blood, spleen, or from peritoneum of thioglycolate-treated (Thio- Peri) mice were stimulated with BCG, E. coli DH5a, S. flexneri, or Pam3 for 24 hr, then assayed for IL-10. Results are expressed as multiples of the IL-10 response relative to the response obtained after Pam3 stimulation. (E and F) Mice were i.p. infected for 8 hr with BCG (5 3 107), E. coli DH5a (107)orS. flexneri (107), and monensin was injected for the last 3 hr. Afterward, peritoneal Ly6G+ cells were directly analyzed for intracellular IL-10 (E). Alternatively, peritoneal neutrophils were purified from mice 6 hr after infection, and further cultured for 24 hr. Subse- quently, supernatants were tested for IL-10 (F). Results are shown as mean of duplicates ± SD. spleen, and peritoneum (Figure S5). These various results next assessed the neutrophil response in vivo after the infection show that in the context of TLR activation, murine neutrophils of mice by mycobacteria, Escherichia, or shigella. Using intracel- are efficiently activated, but contribute very much more weakly lular staining, we found that neutrophils recruited in the perito- than monocytes to proinflammatory cytokine production. neum produced IL-10 after bacterial infection (Figure 2E). We We also found that only Pam3 and LPS stimulated Gr-1hiLy6- purified neutrophils from the infected mice and confirmed their Ghi neutrophils to produce IL-10, and the production was low capacity to release IL-10 (Figure 2F). Therefore, after bacterial (Figure 2A). Mycobacterium bovis BCG, which contains agonists stimulation, monocytes express a strong proinflammatory for TLR2, TLR4, and TLR9, rapidly activated neutrophils (Fig- signature, whereas the neutrophil response is mostly anti- ure 2B and data not shown), and neutrophils stimulated with inflammatory. either live or heat killed (HK) BCG produced large amounts of IL-10. The production of IL-10 induced by bacteria was at least MyD88-TLR2 Is Essential but Not Sufficient 10- to 50-fold more than that in response to synthetic TLR for IL-10 Production by Neutrophils agonists alone (Figure 2 and Figure S6). Monocytes produced We next addressed the issue of the signaling pathways required large amounts of proinflammatory mediators TNF-a, IL-1b, for the substantial production of IL-10 by murine neutrophils after IL-6, IL-12 p40 and p70, CCL2, and CXCL1 but less IL-10 bacterial stimulation. TLR agonists are major components of (Figure 2C) in response to BCG. Interestingly, monocytes, but bacterial cell walls, and accordingly, there was no IL-10 produc- not neutrophils, from Il10À/À mice showed a 10-fold increase in tion in neutrophils from Myd88À/À mice in response to E. coli, the production of proinflammatory cytokines in response to S. flexneri, and M. bovis BCG (Figure 3). TLR2, but not TLR4 BCG (data not shown), indicating that an autocrine IL-10 regula- and TLR9, was involved in the neutrophil-derived IL-10 in tory loop controls monocyte response but does not inhibit the response to bacteria. Therefore, IL-10 secretion by bacteria- proinflammatory cytokine response of neutrophils. IL-10 pro- stimulated neutrophils was fully dependent on the TLR2 and duction by neutrophils was partially dependent on an IFN-a MyD88 pathways, despite the synthetic TLR2 agonist Pam3 and -b autocrine loop (data not shown). When BM neutrophils being unable to produce the response induced by whole bacteria were stimulated with Gram-negative bacteria (Escherichia coli (Figures 2D and 2F and Figure S6). IL-10 induction by the TLR2 and Shigella flexneri), IL-10 production was greater than that agonists, lipomannan (LM), and lipoarabinomannan (LAM) from after stimulation with Pam3 (Figure 2D). Similar results were M. smegmatis either individually or in combination represented obtained with neutrophils isolated from blood and spleen, and 10% of the induction by M. smegmatis (Figure S6). Synergy to a lesser extent with thioglycolate-elicited peritoneal neutro- between different TLR pathways has been described for proin- phils (Figure 2D). The production of IL-10 by neutrophils was flammatory responses of DC, but we failed to find any combina- not influenced by the binding of the Ly6G mAb (Figure S7). We tion of various TLR agonists that enhanced the IL-10 response of

Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. 763 Immunity Bacteria Induce IL-10 Producing Neutrophils

curdlan did not modify the cytokine and chemokine profile of the neutrophil response, but—similar to bacteria—potently drove neutrophils to produce substantial amounts of IL-10 and CXCL2 (Figure 4C and Figures S8A and S9B). LPS, but not TLR7 or TLR9 agonists, also synergized with curdlan to induce IL-10 production by neutrophils (data not shown). The NOD2 agonist MDP combined with curdlan had no effect on IL-10 production by neutrophils (Figure S8A). We also tested A. fumigatus as another source of b-glucans to stimulate neutrophils. Again, fungi alone induced minute amounts of IL-10, and induction was greatly enhanced by the presence of a TLR2 agonist (Figure S8C). Very recently, the mycobacterial factor trehalose-6,6 dimycolate (TDM) was shown to stimulate the innate responses of DCs and macrophages in a FcRg chain-dependent manner (Werninghaus et al., 2009). When we stimulated neutrophils with Pam3 together with TDM from M. tuberculosis, they readily produced large amounts of IL-10 (Figure 4D). The influence of the FcRg pathway was further confirmed with immune complexes that substantially enhanced IL-10 production in Pam3-stimulated neutrophils (Figure S8D). Altogether, these results demonstrate that activa- Figure 3. MyD88-TLR2 Signaling Is Essential for IL-10 Production tion through selective TLR together with Clec5A and Clec7A path- by Neutrophils after Bacterial Stimulation BM neutrophils from C57BL/6 (A–C), BALB/c (A), Myd88À/À (A and C), Tlr2À/À ways or with carbohydrate compounds can switch neutrophils to (B and C), Tlr4À/À (B), or Tlr9À/À (B) mice were stimulated with BCG (A and B), become high IL-10 producers. E. coli DH5a, S. flexneri, and influenza virus PR8 (C). After 24 hr stimulation, supernatants were tested for IL-10. Results are shown as mean of dupli- FcRg, DAP12, and Syk Signaling Contribute cates ± SD. to Bacterial Induction of IL-10 by Neutrophils To discriminate between the different CLR signaling pathways neutrophils (data not shown). Similarly, NOD and TLR pathways that may contribute to IL-10 production after bacterial stimula- synergize for innate responses (Fritz et al., 2007), but peptidogly- tion, we tested the response of neutrophils from FcRg-deficient cans did not modulate the Pam3-induced IL-10 response of (Fcer1gÀ/À) mice (Figure 4E) and from DAP12KDY75 mice neutrophils (Figure S8A and data not shown). In conclusion, (Figure 4F), a loss-of-function mutant of DAP12 (Tomasello TLR2 is necessary but not sufficient for inducing the anti- et al., 2000). We first stimulated WT and Fcer1gÀ/À neutrophils inflammatory response of neutrophils in the context of bacterial with Pam3 alone or in combination with glycans. Fcer1gÀ/À stimulation. neutrophils failed to respond to mycobacterial TDM glycolipid; however, they responded well to b-glucans or DAP12-dependent C-Type Lectin Agonists Synergize with TLR2 to Promote Clec5A stimulation. Under the same conditions, DAP12KDY75 the Anti-inflammatory Signature of Neutrophils neutrophils failed to respond to Clec5A triggering, whereas their Another important large family of PRRs, C-type lectins, plays an production of IL-10 was strongly stimulated in response to cur- important role in innate defenses. We first screened for the dlan and TDM. After bacterial stimulation, DAP12 was mainly, expression of various lectins and scavenger receptors by neutro- but not solely, involved in the response to S. flexneri and E. coli, phils (Figure 4A). CD204, CD205, Clec7A (Dectin-1), and Clec5A whereas the lack of FcRg did not affect the response. The (MDL-1) are expressed by neutrophils. We coated antibodies response to BCG was mainly dependent on DAP12, and FcRg specific for all these lectins and assessed the capacity of cross- made a minor contribution, indicating that TDM is not the major linked CLR to activate neutrophils. Under these conditions, only active mycobacterial component. Clec7A (Dectin-1) has been anti-Clec5A activated neutrophils as judged by phenotypical implicated in the interaction between mycobacteria and both changes (data not shown), but they failed to induce IL-10 dendritic cells and macrophages (Yadav and Schorey, 2006); (Figure 4B and Figure S8B). However, Clec5A antibodies, either however, IL-10 production in response to M. bovis BCG was monoclonal or polyclonal, combined with Pam3 induced strong only slightly inhibited by laminarin, a low MW b-glucan with IL-10 production, as bacteria do. Given that Clec7A antibodies Clec7A-blocking properties (Gantner et al., 2003)(Figure 4G). In block ligand binding without cell signaling, we stimulated neutro- contrast, laminarin inhibited activation due to S. flexneri by phils with curdlan (a high MW b1-3glucan from agrobacteria), 50%. FcRg and DAP12 signaling, as well as Clec5A- (Bakker a natural agonist of Clec7A. Like TLR agonists, curdlan strongly et al., 1999) and Clec7A-mediated activation (Rogers et al., activated neutrophils to upregulate CD11b, to shed CD62L, and 2005), involve Syk tyrosine kinase, which is essential for down- to generate a respiratory burst (Figure S9A), but induced only stream inflammatory cytokine responses. We found that the minimal production of cytokines and chemokines, including Syk tyrosine kinase inhibitor piceatannol and Syk inhibitors III IL-10, TNF-a, and CXCL2 (Figure 4C and Figures S8A and halved IL-10 production by neutrophils after stimulation by S9B). Curdlan-activated monocytes produced TNF-a, IL-6, S. flexneri, E. coli, or Pam3 combined with curdlan or anti-Clec5A IL-12, CCL2, CXCL1, and CXCL2, and the production was further (Figure 4G). It is unclear whether the partial inhibition observed increased by the addition of Pam3 (Figure S9B). Pam3 plus was due to the limited efficacy of these pharmacological

764 Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. Immunity Bacteria Induce IL-10 Producing Neutrophils

Figure 4. C-Type Lectins Synergize with TLR2 but Differentially Require FcRg, DAP12, or Syk Pathways to Promote the Anti-inﬂammatory Activity of Neutrophils (A) Analysis of the expression proﬁle for C-type lectins and scavenger receptors on BM neutrophils by FACS (thin line). (B–F) BM neutrophils were stimulated with immo- bilized anti-Clec5A mAb (B), curdlan (C), and TDM from M. tuberculosis (D) in the presence or absence of Pam3. BM neutrophils from WT (E–G), Fcer1gÀ/À (E), and DAP12KDY75 (F) were stimulated as indicated, and the supernatants were tested for IL-10. Results are shown as mean of duplicates ± SD. (G) BM neutrophils were pretreated with 500 mg/ml laminarin or Syk inhibitors at 37C for 1 hr, then stimulated as indicated. Results for the IL-10 responses in the presence of inhibitors are expressed as percentages of that in the absence of inhibitors. Results are shown as mean of three independent experiments ± SD. Not done, N.D.

kinase inhibitor SB203580 strongly inhibited the IL-10 production by neutrophils in various experimental conditions (Figure 5A). In contrast, SB203580 had a less obvious effect on TNF-a production (data not shown). The phosphatidylinosi- tol-3 kinase (PI3K)-Akt pathway can also contribute to IL-10 production (Pengal et al., 2006): we found that LY294002, an inhibitor of PI3K, inhibited neutrophil production of IL-10 in response to bacteria as well as to single TLR and CLR agonists (Figure 5B). The doses of pharmacological inhibitors used did not affect neutrophil viability (Figure S10). To explore further the kinase pathways, we monitored the phosphorylation of p38 and Akt kinases during neutrophil activation. After 30 min, p38 and Akt phosphorylation was strongly induced inhibitors and would need further investigation with Syk-deficient by Clec5A and Clec7A agonists, whereas only a modest phos- neutrophils. These inhibitors fully inhibited activation by M. bovis phorylation was observed after TLR2 activation (Figures 5C BCG, but no inhibition was observed after Pam3 stimulation. This and 5D). Interestingly, kinetic studies revealed that CLR-induced series of experiments provides strong evidence that the anti- p38 and Akt activation quickly receded in neutrophils, but was inflammatory properties of neutrophils are engaged by synergy maintained in the presence of Pam3 stimulation (Figure 5D). It between the MyD88-TLR and Syk pathways. remains to be determined whether Pam3 stimulation inhibits dephosphorylation of the proteins phosphorylated by CLR TLR2 and CLR-Syk Sustain Phosphorylation signaling or induces new phosphorylation. These observations of the p38 and Akt for IL-10 Production by Neutrophils show that activation of p38 MAP and PI3-Akt kinases is essential The regulation of IL-10 transcription is a complex process, and for IL-10 production by neutrophils after TLR and/or CLR stimu- most relevant work has been with macrophages or GMCSF- lation and that sustained kinase activation may contribute to the derived DC. These studies found that the activation of MAP high amount of IL-10 secretion. kinase pathways, particularly p38 and ERK kinases, are essential for IL-10 production by innate cells (Lucas et al., 2005; Yi et al., Neutrophils Control Local Lung Inflammation 2002). We therefore used inhibitors specific for p38, ERK, and In Vivo after Mycobacterial Infection JNK kinases to test whether MAP kinases are also important To investigate directly the role of neutrophils in the regulation of for IL-10 production by neutrophils. We found that only the p38 innate responses in vivo, we studied the influence of neutrophils

Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. 765 Immunity Bacteria Induce IL-10 Producing Neutrophils

Figure 5. The p38 MAPK and Akt Kinase Pathways Are Essential for IL-10 Production by Neutrophils (A and B) BM neutrophils were pretreated with inhibitors for p38 (SB203580), ERK (U0126), JNK (SP600125), and PI3 kinase (LY294002) or with DMSO at 37C for 1 hr and then stimulated as indicated. Twenty-four hours later, supernatants were tested for IL-10. The results for the IL-10 responses in the presence of inhibitors are expressed as percentages of that in the absence of inhibitors. Results are shown as mean of three independent experiments ± SD. (C and D) BM neutrophils were stimulated as indicated for 30 min (C) or for different time periods (D); then, intracellular phosphorylated p38 MAPK and Akt kinases (bold lines) were detected by ﬂow cytometry. Shaded histograms correspond to Ig isotype controls.

compensate the increase of inflammation resulting in a neutral effect on the control of the acute phase of infection. To test this hypothesis, we depleted neutrophils in Il10À/À mice prior BCG infection. In these conditions, lack of neutrophils led to an increase of mycobacterial load, indicating the sole loss of neutrophil bacterial killing activity (Figure 6D). These results underline the dual role of neutrophils in acute infection, a direct antimicrobial activity counterbalanced by anti-inflammatory properties. To confirm the latter point and to clarify the impact of neutrophil-derived IL-10, we transferred WT or Il10À/À neutrophils into Il10À/À recipient mice and then challenged these mice with BCG. Transfer of Il10À/À, but not WT, neutrophils decreased the mycobacterial load as compared to the controls (Figure 6E). These data indicate that in the absence of anti-inflammatory properties, transferred neutrophils contribute in the development of lung inflammation induced by BCG. Intra- directly to mycobacterial elimination. In contrast, when WT nasally, BCG infection in mice led to a rapid recruitment of neutrophils are transferred, their dual influence on direct bacte- neutrophils into the lung alveolar cavity (as early as 6 hr postin- rial elimination and on inflammation leads to a neutral effect for fection; Figure 6A). These in vivo-activated neutrophils secreted mycobateria control, in accordance with the neutral effect IL-10 after 24 hr ex vivo culture. Neutrophils from the alveolar observed in neutrophil-depleted WT mice (Figure 6C). Impor- cavity produced more IL-10 than did neutrophils from lung tantly, WT and Il10À/À transferred neutrophils were similarly parenchyma (Figure 6B). recruited in the lungs after infection (Figures 6F and 6G). IL-10 To test the influence of neutrophils on the inflammatory was only detected in mice receiving WT neutrophils (Fig- response in the lung, we first depleted neutrophils by using ure 6H). The production of TNF-a by macrophages, monocytes, Ly6G mAb and then administered BCG intranasally: neutrophil and DCs was much lower than that untreated control mice infiltration into the lungs was strongly reduced, without affecting in Il10À/À mice that had received wild-type, but not Il10À/À, the number of monocytes, macrophages, and DCs (Figure S11). neutrophils (Figure 6I). In an identical experimental system, Neutrophil depletion led to a reduced amount of IL-10 and BAL was assayed for proinflammatory cytokines. The BCG-in- increased TNF-a produced by lung cells, without affecting the fected mice given WT neutrophils produced much less TNF-a, mycobacterial load (Figure 6C). One possible explanation would IL-1b, and IL-6 than those given Il10À/À neutrophils (Figure 6J). be that the loss of direct killing activity of neutrophil could These results clearly show that neutrophil-derived IL-10 can

766 Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. Immunity Bacteria Induce IL-10 Producing Neutrophils

Figure 6. Neutrophils Control Local Lung Inflammation in an IL-10-Dependent Manner after Intranasal Inoculation with M. bovis BCG (A) Kinetics of neutrophil infiltration into the alveolar cavity after intranasal (i.n.) inoculation of B6 mice (n = 3–6) with 8 3 106 CFU of BCG. Results are shown as mean of each group of mice ± SD. (B) Neutrophils were purified from BAL or lung parenchyma, 6 hr after intranasal administration of BCG, cultured for 24 hr, and tested for IL-10 production. Total-lung neutrophils from PBS-treated mice were used as controls. Results are shown as mean of duplicates ± SD. (C) B6 mice were PMN depleted with Ly6G mAb 1 day before i.n. inoculation with BCG. Whole-lung cells were collected and further cultured for 24 hr and tested for IL-10 and TNF-a production; BCG CFU lungs were counted at day 3 postinfection. (D) Il10À/À mice were PMN depleted or not before BCG challenge, and CFUs were counted in the lungs. (E–J) A total of 5 3 106 B6 or Il10À/À neutrophils or PBS were i.n. transferred into Il10À/À mice before BCG challenge. As shown in (E), CFUs were counted in the lungs. (F) shows tissue sections of lungs (nuclei in blue) after transfer of PE-labeled neutrophils (in red) and after BCG challenge or with no treatment (inset). (G) shows analysis of neutrophil migration to the lungs after transfer of a 1:1 mixture of B6 and Il10À/À PE-labeled neutrophils. (H) shows IL-10 detection in lungs from mice administered WT or Il10À/À neutrophils. Results are shown as mean of duplicates ± SD. As shown in (I), lung macrophages (MF), monocytes (Mo), and DCs were tested for intracellular TNF-a. As shown in (J), BALs were analyzed for TNF-a, IL-1b, and IL-6. Six mice were used for each group. Results are shown as mean of each group of mice ± SD. *p < 0.05 compared to other groups. downmodulate local lung inflammation induced by myco- neutrophils infected with H37Rv produced, in a MyD88-depen- bacteria. dent manner (data not shown), large amounts of IL-10 (Figure 7A). We then infected mice with an aerosol of M. tuberculosis H37Rv Neutrophils Influence Chronic Infection (Mtb, 100 CFU per mouse). Five weeks later, we purified and by Virulent Mycobacteria cultured neutrophils without additional stimulation. Under these Finally, we evaluated the role of neutrophils in the context of conditions, lung neutrophils isolated from both BAL and paren- a chronic infection by virulent M. tuberculosis H37Rv. In vitro, chyma of Mtb-infected mice produced IL-10 (Figure 7B). To

Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. 767 Immunity Bacteria Induce IL-10 Producing Neutrophils

Figure 7. Neutrophils Lower Immunity during M. tuberculosis Chronic Infection (A) BM neutrophils from BALB/c mice were stimulated with live M. tuberculosis H37Rv for 24 hr, and supernatants were tested for IL-10. (B) BALB/c mice were aerosolly infected with a M. tuberculosis H37Rv. Five weeks later, 2 3 105 neutrophils puriﬁed from lungs and BM were cultured ex vivo for 24 hr. Alternatively, Ly6G+ cells were puriﬁed independently from BAL and lung parenchyma of infected mice (n = 4). The supernatants were tested for IL-10. (C) BALB/c mice were infected with M. tuberculosis H37Rv by the aerosol route. Five weeks later, mice received two intravenous injections of anti- Ly6G mAb (NIMP-R14) or Ig control 3 days apart. (C and D) One day after each injection, lungs cells were analyzed phenotypically and for IL-10 and TNF-a release (n = 3 mice/group). (E) Alternatively, whole-lung homogenates after two injections were tested for IL-6, IL-17, and IFN-g (n = 6 mice/group). (F) Infected mice (n = 10/group) received three injections of depleting and control antibody, then H37Rv CFU were counted in the lungs. Results are representative of at least two experiments. *p < 0.05.

isms; second, stimulation with bacteria and combined TLR/CLR-Syk activation trigger a predominantly anti-inflammatory response in neutrophils involving substantial IL-10 production; and third, identify the effects of neutrophils on inflammation and Mtb infec- neutrophils negatively regulate lung inflammation induced by tion, we depleted neutrophils from chronically infected mice by mycobacterial infection in vivo. injection of Ly6G mAb. After this treatment, the number of mono- Murine neutrophils were readily activated in a MyD88-depen- cytes slightly decreased, whereas inflammatory DCs increased dent manner by TLR2, TLR4, and TLR5, but not by TLR3, TLR7, (Figure 7C). When we cultured cells from infected lungs or TLR9 agonists, but only produced minute amounts of proin- ex vivo, we found a significant (p < 0.05) reduction of IL-10 flammatory cytokines; this finding is not consistent with previous release and slight increase of TNF-a in depleted mice as reports (Bennouna and Denkers, 2005; Bliss et al., 2000). Anti- compared to control infected mice (Figure 7D). In addition, Gr-1 (Ly6G and C) has been routinely used for identifying murine concentrations of IL-6 and IL-17, but not IFN-g, were higher in neutrophils in numerous studies, but Gr-1 is also expressed on whole-lung homogenates from neutrophil-depleted infected many other cells; particularly relevant is a population of Gr-1hi mice than those from infected control isotype mice (Figure 7E). inflammatory monocytes (Geissmann et al., 2003), which, as Concommitantly, 1 week after neutrophil depletion, the number we demonstrated here, have a substantial capacity to produce of mycobacterial CFU in the lung was 2.5-fold lower than in proinflammatory cytokines. Therefore, our work suggests that controls (Figure 7F). These experiments indicate that virulent many of the functions of murine neutrophils described in early mycobacteria can activate neutrophils to produce IL-10 and studies should be interpreted with caution and need to be recon- that neutrophils contribute to the persistence of a high mycobac- sidered. We show here that neutrophils preferentially produce terial burden during chronic infection by M. tuberculosis H37Rv. IL-10, rather than inflammatory cytokines, in response to bacteria and do so in a MyD88- and Syk-dependent manner. DISCUSSION Coligation of multiple PRRs corresponds to the physiologically relevant innate cell activation by microorganisms, and syner- In this study, we systemically studied the innate responses of gistic effects have been observed between different TLR murine neutrophils in response to single agonists for TLR, agonists and between TLR and other PRRs. The TLR7 or TLR8 NLR, and CLR, as well as in response to whole bacteria (Myco- agonist synergizes with TLR3 or TLR4 agonists to induce strong bacteria, E. coli, and S. flexneri), fungus (A. fumigatus), and production of IL-12 p70 by DCs (Napolitani et al., 2005). The TLR virus (influenza virus). We report several interesting findings: and NLR pathways also synergize for inflammatory responses First, murine neutrophils give no or weak proinflammatory (Fritz et al., 2006). However, we did not observe any such responses to stimulation with single agonists or microorgan- synergy for IL-10 production by neutrophils. The relationship

768 Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. Immunity Bacteria Induce IL-10 Producing Neutrophils

between TLR and ITAM-activating receptors is less clear, with In the context of bacterial infection, this phenomenon counter- respect to either synergistic or inhibitory outcomes. Macro- balances the high proinflammatory activity of other innate cells phages deficient for DAP12 show an enhanced response to such as monocytes, macrophages, and DCs. We confirmed TLRs probably because of basal ITAM-Syk activation in these this finding by showing that neutrophils control local lung innate cells (Hamerman et al., 2005). Ligation of DAP12-associated inflammation after mucosal M. bovis BCG infection. In the acute receptors in myeloid cells leads to the production of proinflam- phase of the infection, the global activity of neutrophils is matory cytokines (Bouchon et al., 2001; Turnbull et al., 2005), a balance between the positive (direct bacterial elimination) but may also dampen the TLR response of macrophages versus the negative (anti-inflammatory) properties of neutrophils (Hamerman et al., 2006). Fcer1gÀ/À and DAP12KDY75 neutrophils without much influencing the mycobacterial growth. However, in showed a slightly greater IL-10 response to TLR2 activation than the chronic phase of M. tuberculosis infection, i.e., with a high WT neutrophils, but a lower response to bacteria. Therefore, the and stable mycobacterial load, we found that neutrophil deple- net effect of TLR2 and CLR-Syk coactivation or bacterial stimu- tion was beneficial to the host and was associated with lation was a strong enhancement of the IL-10 response. The increased amounts of IL-6 and IL-17, but not IFN-g in the collaboration of TLR and Clec7A (Dectin-1) pathways has been infected lungs. IL-6 is upstream from Th17 cell differentiation shown to induce an increased inflammatory response in both (Dong, 2008) and both Th1 and Th17 cells are associated with macrophages and DCs (Gantner et al., 2003), although Clec7A M. tuberculosis control (Khader et al., 2007); this type of (Dectin-1) together with CD40 also potently induces IL-10 from phenomenon may also be relevant for other bacterial infections. DCs (Rogers et al., 2005). In our study, bacteria induced strong Interestingly, one of the rare descriptions of IL-10 production IL-10 production by neutrophils, and this phenomenon was by human neutrophils concerns those infiltrating UV-irradiated dependent on TLR2 and Syk coactivation. Clearly, different skin (Piskin et al., 2005). Because we found that TLR2 and CLR PAMPs and PRRs are involved in the Syk activation process coactivation leads to IL-10 production by neutrophils, it is because the role of DAP12 and FcRg was dependent on the possible that endogenous ligands for these receptors may be bacteria tested. Full inhibition of mycobacteria-induced neutro- involved in this phenomenon. In addition to C. albicans recogni- phil-derived IL-10 was achieved with Syk inhibitors, but only tion (Wells et al., 2008), Clec4e (Mincle), a DAP12-dependent 50% inhibition was obtained in the context of shigella stimula- CLR, recognizes SP130 nuclear protein from damaged cells tion. In addition, half of the IL-10 response was dependent on (Yamasaki et al., 2008). Likewise, extracellular matrix compo- DAP12 for all bacteria tested, whereas a role for FcRg was nents, such as hyaluronan (Scheibner et al., 2006) and versican only evidenced for mycobacteria. Coengagement of TLR2 and (Kim et al., 2009), can activate innate inflammation through Clec5A also promoted regulatory neutrophils, and recognition TLR2. Therefore, endogenous TLR2 ligands may act in concert of Dengue virus by Clec5A mainly involves fucose residues with CLR ligands released by damaged cells to activate anti- (Chen et al., 2008). This phenomenon was also successfully inflammatory neutrophils. Similar phenomena may contribute mimicked by a combination of TLR2 and b-glucans. All these to the activity of Gr1-positive myeloid-derived suppressor cells stimuli preferentially induced substantial IL-10 production by that are involved in tumor immune suppression (Bronte et al., neutrophils, with only small or undetectable amounts of proin- 2003). flammatory cytokines. This is in sharp contrast with monocytes In summary, we have identified a unique regulatory role for that display a strong proinflammatory profile in the same condi- neutrophils triggered by mycobacteria and Gram-negative tions. Importantly, the proinflammatory response of monocytes bacteria that is based on co-activation through the MyD88 and is under the control of an autocrine IL-10 regulatory loop, CLR-DAP12-Syk pathways; this regulation involves the produc- whereas neutrophils are not sensitive to IL-10, and thus have tion of large amounts of IL-10. IL-10-producing neutrophils are a true anti-inflammatory phenotype. This indicates that, in addi- able to temper lung inflammation upon mycobacterial infection tion to their microbicidal activities, neutrophils can turn into in vivo and may play a similar role in other bacterial infections. a regulatory cell type upon encountering a pathogen. DAP12- deficient mice are more susceptible to endotoxin shock and EXPERIMENTAL PROCEDURES more resistant to Listeria infection (Hamerman et al., 2006). In view of the innate anti-inflammatory response of neutrophils, Mice the lack of DAP12 signaling may also play a role in inflammation C57BL/6 and BALB/c adult mice were purchased from Janvier, and B10D2 and resistance to bacterial infection in vivo. from Jackson Laboratories. Mice deficient for IL-10, TLR2, TLR4, TLR9, type I IFN receptor, and FcRg were all on a C57BL/6 background and bred in our Neutrophils have been described as supporters of proinflam- D animal facilities. DAP12 Y75 mice, loss-of-function mutant of DAP12, were matory responses (Ricevuti, 1997) and as able to exert their on a B10D2 background (Tomasello et al., 2000). The mice were housed effects directly on DCs (Bennouna et al., 2003; van Gisbergen on-site in specific pathogen-free conditions and used when they were et al., 2005). However, activated neutrophils inhibit T cell func- 6–10 weeks old. Animal studies were approved by the Institut Pasteur safety tions and suppress cytokine production through the generation committee in accordance with French and European guidelines. of hydrogen peroxide (Schmielau and Finn, 2001). Neutrophils purified from mice suffering of systemic inflammatory response Cell Purification and FACS Analysis syndrome display regulatory features that can convert resident Neutrophils and monocytes were purified from BM, lungs, spleen, blood, or thioglycolate-treated peritoneum. For neutrophils, cells from organs were macrophages into alternatively activated macrophages that stained with biotin- or PE-anti Ly6G, stained with anti-biotin or anti-PE beads usually contribute to chronic infections (Tsuda et al., 2004). Our (Miltenyi Biotec), and positively selected on an automated magnetic cell sorter work unequivocally demonstrates a central regulatory role of (AutoMACSpro; Miltenyi Biotec). In some experiments, BM cells were directly neutrophils in dampening the proinflammatory response. stained with FITC-anti Ly6G, PE-anti CD115, and APC-anti CD11b, then

Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. 769 Immunity Bacteria Induce IL-10 Producing Neutrophils

neutrophils (Ly6GhiCD11b+CD115À) and monocytes (CD115+CD11b+Ly6GÀ) for S. flexneri, R. Brosch for M. smegmatis, and C. Fayolle for NIMP-R14 mAb. were sorted by flow cytometry on a FACSAria apparatus (BD Biosciences). X. Z. was supported by the European Commission (Cellprom-NMP4-CT-2004- FACS analysis was performed with Flowjo software (Tree Star). 500039) and by Institut Pasteur (PTR260). This work was supported by a grant from Agence Nationale de la Recherche (MIME 2006) and by the Ligue du Neutrophil Responses Cancer. A total of 105 or 2 3 105 neutrophils and/or the indicated number of monocytes were cultured in RPMI-1640 complete medium in 96-well flat-bottom plates Received: April 3, 2009 and stimulated with various stimuli for 24 hr. The supernatants were collected Revised: June 20, 2009 and assayed for cytokines. IL-10, TNF-a, IL-1b, IL-6, IL-12 p40, and IL-12 p70 Accepted: September 9, 2009 were determined by standard sandwich ELISA with appropriate Ab pairs (all Published online: November 12, 2009 from BD Biosciences). Chemokines CCL2, CXCL1, and CXCL2 were detected with a multiplex kit (Invitrogen). For intracellular cytokine staining for TNF-a, IL-12, and IL-10, cells were stimulated as indicated in figure legends, brefeldin REFERENCES A was added for the last 3 hr during stimulation, and the cells were processed in accordance with the manufacturer’s protocol (BD Biosciences). For detect- Akira, S., Uematsu, S., and Takeuchi, O. (2006). Pathogen recognition and ing reactive oxygen intermediates, purified neutrophils were loaded with 5 mM innate immunity. Cell 124, 783–801. Dihydrorhodamine 123 (Invitrogen) for 15 min at 37 C and then stimulated with Andonegui, G., Bonder, C.S., Green, F., Mullaly, S.C., Zbytnuik, L., Raharjo, E., various stimuli for the indicated time. For monitoring intracellular phosphory- and Kubes, P. (2003). Endothelium-derived Toll-like receptor-4 is the key lated kinases, purified neutrophils were stimulated as indicated, then fixed molecule in LPS-induced neutrophil sequestration into lungs. J. Clin. Invest. with BD Phosflow Lyse/Fix Buffer I at 37 C for 10 min, permeabilized with 111, 1011–1020. BD Phosflow Perm Buffer III for 30 min on ice, and stained. Bakker, A.B., Baker, E., Sutherland, G.R., Phillips, J.H., and Lanier, L.L. (1999). For inhibition experiments, purified BM neutrophils were pretreated with Myeloid DAP12-associating lectin (MDL)-1 is a cell surface receptor involved in 500 mg/mL laminarin (Sigma) or 1 mM of the various kinase inhibitors at 37C the activation of myeloid cells. Proc. Natl. Acad. Sci. USA 96, 9792–9796. for 1 hr. Afterward, the cells were stimulated with various agents for 24 hr and the supernatants were tested for IL-10. The kinase inhibitors used were Bennouna, S., Bliss, S.K., Curiel, T.J., and Denkers, E.Y. (2003). Cross-talk in the following: SB203580 (p38), U0126 (ERK), SP600125 (JNK), LY294002 the innate immune system: Neutrophils instruct recruitment and activation of (PI3K), piceatannol, and Syk inhibitor III (Syk), all from Calbiochem Merck. dendritic cells during microbial infection. J. Immunol. 171, 6052–6058. Bennouna, S., and Denkers, E.Y. (2005). Microbial antigen triggers rapid Mycobacterial Infections mobilization of TNF-alpha to the surface of mouse neutrophils transforming 6 The mice were inoculated intranasally with 8 3 10 CFU M. bovis BCG and the them into inducers of high-level dendritic cell TNF-alpha production. J. Immu- first BAL wash in 1 ml PBS was tested for cytokines. The lung parenchyma was nol. 174, 4845–4851. digested with collagenase and DNase I at 37C for 30 min so that single-cell Bliss, S.K., Butcher, B.A., and Denkers, E.Y. (2000). Rapid recruitment of suspensions could be obtained. Cells from BAL fluids and lung parenchyma neutrophils containing prestored IL-12 during microbial infection. J. Immunol. were further incubated for 3 hr in the presence of brefeldin A for intracellular 165, 4515–4521. TNF-a detection. Alternatively, neutrophils were purified from lung cell suspensions for tests for IL-10 production. In brief, cell suspensions were loaded onto Bouchon, A., Facchetti, F., Weigand, M.A., and Colonna, M. (2001). TREM-1 Percoll gradient media (80%–65%–55%) and after centrifugation, cells at the amplifies inflammation and is a crucial mediator of septic shock. Nature 410, 80%–65% and 65%–55% interfaces were collected. Lung neutrophils were 1103–1107. purified as described above with PE anti-Ly6G and anti-PE beads. Neutrophil Bronte, V., Serafini, P., Mazzoni, A., Segal, D.M., and Zanovello, P. (2003). depletion was achieved by intravenous (i.v.) injection of 200 mg anti-Ly6G L-arginine metabolism in myeloid cells controls T-lymphocyte functions. (either clone 1A8 or NIMP-R14, with similar results) mAb or isotype control Trends Immunol. 24, 302–306. antibody 1 day before BCG administration. Purified 1A8 mAb was obtained from BioXcell, and NIMP-R14 mAb was prepared from the eponymous cell Chen, S.T., Lin, Y.L., Huang, M.T., Wu, M.F., Cheng, S.C., Lei, H.Y., Lee, C.K., clone kindly given by G. Millon (Institut Pasteur). Chiou, T.W., Wong, C.H., and Hsieh, S.L. (2008). CLEC5A is critical for For M. tuberculosis H37Rv infection, BALB/c mice were infected via the aero- dengue-virus-induced lethal disease. Nature 453, 672–676. sol route with a customized apparatus for delivering a retained inhaled dose Dong, C. (2008). TH17 cells in development: An updated view of their molec- of 100 ± 10 CFU. For neutrophil depletion, mice received two to three injections ular identity and genetic programming. Nat. Rev. Immunol. 8, 337–348. of 200 mg anti-Ly6G (NIMP-R14) mAb or isotype control antibody as indicated in Fleming, T.J., Fleming, M.L., and Malek, T.R. (1993). Selective expression of the legends. Five to six weeks after infection, lungs were collected and either Ly-6G on myeloid lineage cells in mouse bone marrow. RB6–8C5 mAb to gran- cells were recovered for FACS analysis and culture or whole-lung homoge- ulocyte-differentiation antigen (Gr-1) detects members of the Ly-6 family. nates were prepared for cytokine detection and H37Rv CFU counting. Alterna- J. Immunol. 151, 2399–2408. tively, lung neutrophils were purified and cultured for 24 hr. Fritz, J.H., Ferrero, R.L., Philpott, D.J., and Girardin, S.E. (2006). Nod-like Statistical Analysis proteins in immunity, inflammation and disease. Nat. Immunol. 7, 1250–1257. Unpaired t test was used for comparisons between two groups (data from Fritz, J.H., Le Bourhis, L., Sellge, G., Magalhaes, J.G., Fsihi, H., Kufer, T.A., such experiments are presented as mean values ± SD). p values of < 0.05 Collins, C., Viala, J., Ferrero, R.L., Girardin, S.E., and Philpott, D.J. (2007). were considered statistically significant. Nod1-mediated innate immune recognition of peptidoglycan contributes to the onset of adaptive immunity. Immunity 26, 445–459. SUPPLEMENTAL DATA Gantner, B.N., Simmons, R.M., Canavera, S.J., Akira, S., and Underhill, D.M. (2003). Collaborative induction of inflammatory responses by dectin-1 and Supplemental Data include Supplemental Experimental Procedures and Toll-like receptor 2. J. Exp. Med. 197, 1107–1117. 11 Figures and can be found with this article online at http://www.cell.com/ immunity/supplemental/S1074-7613(09)00456-7. Geissmann, F., Jung, S., and Littman, D.R. (2003). Blood monocytes consist of two principal subsets with distinct migratory properties. Immunity 19, 71–82. ACKNOWLEDGMENTS Hamerman, J.A., Jarjoura, J.R., Humphrey, M.B., Nakamura, M.C., Seaman, W.E., and Lanier, L.L. (2006). Cutting edge: Inhibition of TLR and FcR We are grateful to E. Vivier and M. Dalod for DAP12 KDY75 mice, and to responses in macrophages by triggering receptor expressed on myeloid cells P. Bruhns for Fcer1gÀ/À mice. We thank O. Granet for A. fumigatus, G. Sellge (TREM)-2 and DAP12. J. Immunol. 177, 2051–2055.

770 Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. Immunity Bacteria Induce IL-10 Producing Neutrophils

Hamerman, J.A., Tchao, N.K., Lowell, C.A., and Lanier, L.L. (2005). Enhanced Schmielau, J., and Finn, O.J. (2001). Activated granulocytes and granulocyte- Toll-like receptor responses in the absence of signaling adaptor DAP12. Nat. derived hydrogen peroxide are the underlying mechanism of suppression of Immunol. 6, 579–586. T-cell function in advanced cancer patients. Cancer Res. 61, 4756–4760. Hayashi, F., Means, T.K., and Luster, A.D. (2003). Toll-like receptors stimulate Suzuki-Inoue, K., Fuller, G.L., Garcia, A., Eble, J.A., Pohlmann, S., Inoue, O., human neutrophil function. Blood 102, 2660–2669. Gartner, T.K., Hughan, S.C., Pearce, A.C., Laing, G.D., et al. (2006). A novel Kanazawa, N., Tashiro, K., Inaba, K., and Miyachi, Y. (2003). Dendritic cell Syk-dependent mechanism of platelet activation by the C-type lectin receptor immunoactivating receptor, a novel C-type lectin immunoreceptor, acts as CLEC-2. Blood 107, 542–549. an activating receptor through association with Fc receptor gamma chain. Taylor, P.R., Tsoni, S.V., Willment, J.A., Dennehy, K.M., Rosas, M., Findon, H., J. Biol. Chem. 278, 32645–32652. Haynes, K., Steele, C., Botto, M., Gordon, S., and Brown, G.D. (2007). Dectin-1 Kawai, T., and Akira, S. (2006). Innate immune recognition of viral infection. is required for beta-glucan recognition and control of fungal infection. Nat. Nat. Immunol. 7, 131–137. Immunol. 8, 31–38. Khader, S.A., Bell, G.K., Pearl, J.E., Fountain, J.J., Rangel-Moreno, J., Cilley, Tepper, R.I., Coffman, R.L., and Leder, P. (1992). An eosinophil-dependent G.E., Shen, F., Eaton, S.M., Gaffen, S.L., Swain, S.L., et al. (2007). IL-23 and mechanism for the antitumor effect of interleukin-4. Science 257, 548–551. IL-17 in the establishment of protective pulmonary CD4+ T cell responses after Tomasello, E., Desmoulins, P.O., Chemin, K., Guia, S., Cremer, H., Ortaldo, J., vaccination and during Mycobacterium tuberculosis challenge. Nat. Immunol. Love, P., Kaiserlian, D., and Vivier, E. (2000). Combined natural killer cell and 8, 369–377. dendritic cell functional deficiency in KARAP/DAP12 loss-of-function mutant Kim, S., Takahashi, H., Lin, W.W., Descargues, P., Grivennikov, S., Kim, Y., mice. Immunity 13, 355–364. Luo, J.L., and Karin, M. (2009). Carcinoma-produced factors activate myeloid cells through TLR2 to stimulate metastasis. Nature 457, 102–106. Tsuda, Y., Takahashi, H., Kobayashi, M., Hanafusa, T., Herndon, D.N., and Suzuki, F. (2004). Three different neutrophil subsets exhibited in mice with Lopez, A.F., Strath, M., and Sanderson, C.J. (1984). Differentiation antigens on different susceptibilities to infection by methicillin-resistant Staphylococcus mouse eosinophils and neutrophils identified by monoclonal antibodies. Br. J. aureus. Immunity 21, 215–226. Haematol. 57, 489–494. Turnbull, I.R., McDunn, J.E., Takai, T., Townsend, R.R., Cobb, J.P., and Lucas, M., Zhang, X., Prasanna, V., and Mosser, D.M. (2005). ERK activation Colonna, M. (2005). DAP12 (KARAP) amplifies inflammation and increases following macrophage FcgammaR ligation leads to chromatin modifications mortality from endotoxemia and septic peritonitis. J. Exp. Med. 202, 363–369. at the IL-10 locus. J. Immunol. 175, 469–477. Napolitani, G., Rinaldi, A., Bertoni, F., Sallusto, F., and Lanzavecchia, A. (2005). van Gisbergen, K.P., Sanchez-Hernandez, M., Geijtenbeek, T.B., and van Selected Toll-like receptor agonist combinations synergistically trigger a T Kooyk, Y. (2005). Neutrophils mediate immune modulation of dendritic cells helper type 1-polarizing program in dendritic cells. Nat. Immunol. 6, 769–776. through glycosylation-dependent interactions between Mac-1 and DC- SIGN. J. Exp. Med. 201, 1281–1292. Pengal, R.A., Ganesan, L.P., Wei, G., Fang, H., Ostrowski, M.C., and Tridandapani, S. (2006). Lipopolysaccharide-induced production of inter- Wells, C.A., Salvage-Jones, J.A., Li, X., Hitchens, K., Butcher, S., Murray, R.Z., leukin-10 is promoted by the serine/threonine kinase Akt. Mol. Immunol. 43, Beckhouse, A.G., Lo, Y.L., Manzanero, S., Cobbold, C., et al. (2008). The 1557–1564. macrophage-inducible C-type lectin, mincle, is an essential component of the innate immune response to Candida albicans. J. Immunol. 180, Piskin, G., Bos, J.D., and Teunissen, M.B. (2005). Neutrophils infiltrating 7404–7413. ultraviolet B-irradiated normal human skin display high IL-10 expression. Arch. Dermatol. Res. 296, 339–342. Werninghaus, K., Babiak, A., Gross, O., Holscher, C., Dietrich, H., Agger, E.M., Ricevuti, G. (1997). Host tissue damage by phagocytes. Ann. N Y Acad. Sci. Mages, J., Mocsai, A., Schoenen, H., Finger, K., et al. (2009). Adjuvanticity of 832, 426–448. a synthetic cord factor analogue for subunit Mycobacterium tuberculosis vaccination requires FcR{gamma}-Syk-Card9-dependent innate immune Rogers, N.C., Slack, E.C., Edwards, A.D., Nolte, M.A., Schulz, O., Schweigh- activation. J. Exp. Med. 206, 89–97. offer, E., Williams, D.L., Gordon, S., Tybulewicz, V.L., Brown, G.D., and Reis e Sousa, C. (2005). Syk-dependent cytokine induction by Dectin-1 reveals Yadav, M., and Schorey, J.S. (2006). The beta-glucan receptor dectin-1 func- a novel pattern recognition pathway for C type lectins. Immunity 22, 507–517. tions together with TLR2 to mediate macrophage activation by mycobacteria. Blood 108, 3168–3175. Romani, L., Mencacci, A., Cenci, E., Del Sero, G., Bistoni, F., and Puccetti, P. (1997). An immunoregulatory role for neutrophils in CD4+ T helper subset Yamasaki, S., Ishikawa, E., Sakuma, M., Hara, H., Ogata, K., and Saito, T. selection in mice with candidiasis. J. Immunol. 158, 2356–2362. (2008). Mincle is an ITAM-coupled activating receptor that senses damaged Saijo, S., Fujikado, N., Furuta, T., Chung, S.H., Kotaki, H., Seki, K., Sudo, K., cells. Nat. Immunol. 9, 1179–1188. Akira, S., Adachi, Y., Ohno, N., et al. (2007). Dectin-1 is required for host Yamasaki, S., Matsumoto, M., Takeuchi, O., Matsuzawa, T., Ishikawa, E., defense against Pneumocystis carinii but not against Candida albicans. Nat. Sakuma, M., Tateno, H., Uno, J., Hirabayashi, J., Mikami, Y., et al. (2009). Immunol. 8, 39–46. C-type lectin Mincle is an activating receptor for pathogenic fungus, Malasse- Scapini, P., Lapinet-Vera, J.A., Gasperini, S., Calzetti, F., Bazzoni, F., and zia. Proc. Natl. Acad. Sci. USA 106, 1897–1902. Cassatella, M.A. (2000). The neutrophil as a cellular source of chemokines. Yi, A.K., Yoon, J.G., Yeo, S.J., Hong, S.C., English, B.K., and Krieg, A.M. Immunol. Rev. 177, 195–203. (2002). Role of mitogen-activated protein kinases in CpG DNA-mediated Scheibner, K.A., Lutz, M.A., Boodoo, S., Fenton, M.J., Powell, J.D., and IL-10 and IL-12 production: Central role of extracellular signal-regulated Horton, M.R. (2006). Hyaluronan fragments act as an endogenous danger kinase in the negative feedback loop of the CpG DNA-mediated Th1 response. signal by engaging TLR2. J. Immunol. 177, 1272–1281. J. Immunol. 168, 4711–4720.

Immunity 31, 761–771, November 20, 2009 ª2009 Elsevier Inc. 771