The Journal of Immunology

MyD88 in Macrophages Is Critical for Abscess Resolution in Staphylococcal Skin Infection Reinhild Feuerstein,*,† Maximilian Seidl,*,‡ Marco Prinz,x,{ and Philipp Henneke*,‖ When Staphylococcus aureus penetrates the epidermis and reaches the dermis, polymorphonuclear leukocytes (PMLs) accumulate and an abscess is formed. However, the molecular mechanisms that orchestrate initiation and termination of inflammation in skin infection are incompletely understood. In human myeloid differentiation primary response gene 88 (MyD88) deficiency, staphylococcal skin and soft tissue infections are a leading and potentially life-threatening problem. In this study, we found that MyD88dependent sensing of S. aureus by dermal macrophages (Mf) contributes to both timely escalation and termination of PMLmediated inflammation in a mouse model of staphylococcal skin infection. Mfs were key to recruit PML within hours in response to staphylococci, irrespective of bacterial viability. In contrast with bone marrow–derived Mfs, dermal Mfs did not require UNC-93B or TLR2 for activation. Moreover, PMLs, once recruited, were highly activated in an MyD88-independent fashion, yet failed to clear the infection if Mfs were missing or functionally impaired. In normal mice, clearance of the infection and contraction of the PML infiltrate were accompanied by expansion of resident Mfs in a CCR2-dependent fashion. Thus, whereas monocytes were dispensable for the early immune response to staphylococci, they contributed to Mf renewal after the infection was overcome. Taken together, MyD88-dependent sensing of staphylococci by resident dermal Mfs is key for a rapid and balanced immune response, and PMLs are dependent on intact Mf for full function. Renewal of resident Mfs requires both local control of bacteria and inflammatory monocytes entering the skin. The Journal of Immunology, 2015, 194: 2735–2745.

S

taphylococcus aureus is part of the resident flora in humans and many other mammalian and nonmammalian species. Up to 30% of humans are asymptomatic mucocutaneous carriers, with the nasal mucosa and moist areas of the skin being preferred colonization sites. In contrast, S. aureus is the leading cause of acute bacterial skin and skin structure infections (ABSSSIs), which are associated with significant morbidity and mortality (1). In the last two decades, a surge in hospital contacts of patients with ABSSSI has been documented (2). Increasing resistance against b-lactam antibiotics is further complicating the situation. Patients with defects in cellular innate immunity, such as chronic granulomatous disease and STAT-3 and NEMO deficiency, typically present with ABSSSI. A striking example for innate immunodeficiency disorders manifesting in ABSSSI as leading complication are human myeloid differentiation primary response gene 88 (MyD88) and IRAK4 deficiencies (3). It seems intriguing that loss of MyD88

*Center for Chronic Immunodeficiency, University Medical Center, University of Freiburg, 79106 Freiburg, Germany; †Faculty of Biology, University of Freiburg, 79104 Freiburg, Germany; ‡Institute of Pathology, University Medical Center, University of Freiburg, 79106 Freiburg, Germany; xInstitute of Neuropathology, University Medical Center, University of Freiburg, 79106 Freiburg, Germany; {Centre of Biological Signalling Studies, University of Freiburg, 79106 Freiburg, Germany; and ‖ Center for Pediatrics and Adolescent Medicine, University Medical Center, University of Freiburg, 79106 Freiburg, Germany Received for publication October 8, 2014. Accepted for publication January 9, 2015. Address correspondence and reprint requests to Prof. Philipp Henneke, Center for Chronic Immunodeficiency and Center for Pediatrics and Adolescent Medicine, University Medical Center, Breisacherstrasse 117 - 2.OG, 79106 Freiburg, Germany. E-mail address: [email protected] The online version of this article contains supplemental material. Abbreviations used in this article: ABSSI, acute bacterial skin and skin structure infection; BMDM, bone marrow–derived Mf; hf, heat-fixed; i.d., intradermal; Dlgt, lgt-deficient strain SA 113; Mf, macrophage; MyD88, myeloid differentiation primary response gene 88; p.i., postinfection; PML, polymorphonuclear leukocyte; qRT-PCR, quantitative real-time PCR; wt, wild type. Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1402566

and IRAK4 function, which affects the myeloid cell response against virtually all important human pathogens in vitro and in many animal models, leads to a relatively selective susceptibility for staphylococci and streptococci. These bacterial species account for approximately two thirds of all infections in the affected patients (4). The infectious disease–associated lethality in these patients approximates 50%. Accordingly, MyD88 and its interaction partner IRAK4 are dedicated and limiting factors in controlling S. aureus at the colonization sites. However, the myeloid cell subset–specific role of MyD88-related signals in sensing staphylococci and instructing the intercellular response at the site of bacterial invasion are incompletely understood. Abscess formation, that is, the spatial containment of bacteria by a tissue reaction involving both resident and incoming myeloid cells in the dermis and hypodermis, is commonly viewed as the early immune response, which prevents the bacterial spread to deeper tissues and systemic dissemination via the bloodstream (5, 6). Yet abscess formation may already be viewed as failure of the resident immune cells to eliminate staphylococci that spuriously penetrate the epidermis and get into contact with the dermis. Macrophages (Mfs) constitute the major resident myeloid cell type in the dermis. It is well established that they are involved in inflammation in the skin (7, 8). Mfs are furthermore known to largely depend on the TLR system in recognizing staphylococci (9). However, some doubt has been recently shed on TLRs in Mfs as key elements in the cellular skin response to bacteria. First, important Mf-independent sensing and central networking functions have been assigned to both gd-T cells and polymorphonuclear leukocytes (PMLs) (10). Next, the NLR system has been found to be essential for staphylococcal sensing (11, 12). Moreover, whereas a significant body of data is available on TLR-sensing of staphylococci in bone marrow–derived Mfs (BMDMs) (13, 14), relatively little is known on S. aureus sensing and functional polarization of dermal Mfs, both as resting cells in immediate proximity to staphylococci colonizing the skin and during ABSSSI.

2736

MyD88 IN MACROPHAGES IS CRITICAL FOR ABSCESS RESOLUTION

Another area of uncertainty in dermal Mf biology is their cellular origin and renewal. Mf types have been discriminated based on their specific localization in steady-state dermis as perivascular Mfs, Mfs associated with lymphoid vessels, and Mfs within the intervascular space (15). Recent compelling evidence assigns specific functions to these subsets. However, it is unclear whether dermal Mf in staphylococcal infection are exclusively replenished from blood monocytes, or whether they are able to self-renew and differentiate at site without contribution of bone marrow and blood-derived mononuclear cells (16). Dermal Mfs are expressing a wide range of MyD88-dependent TLRs to detect bacterial components like LPS or bacterial nucleic acids and to produce proinflammatory cytokines and chemokines (15, 17–19). Thus, they contribute to a healthy skin and immune cell homeostasis. In this article, we aimed at clarifying the role of dermal Mfs in abscess formation and resolution in S. aureus skin infection. We define a very distinct role of MyD88 in Mfs in the earliest stages of staphylococcal infection, which determines the fate of the local infection in the coming days. Dermal Mfs deliver the initial signal for the recruitment of PML and monocytes after sensing staphylococci in a MyD88-dependent, but TLR2- and endosomal TLR–independent fashion. Although delayed, PMLs are recruited in large numbers and are potently activated in staphylococcal infection in MyD88 deficiency or if Mfs are missing. However, these activated PMLs fail to timely clear the infection, which underlines the importance of dermal Mfs in the recruitment, activation, and regulation of these first-line effector cells. Once the bacteria are contained, MyD88-independent recruitment of Ly6Chigh inflammatory monocytes entering the skin substantially contributes to local Mf renewal.

Materials and Methods Animals and cell lines All mice used are on C57BL/6J or C57BL/6N genetic background. Mice lacking MyD88 were previously described (13). Mice with an UNC-93BH412R (3D) mutation were kindly provided by Marina Freudenberg (University of Freiburg) (13). CCR22/2 mice and CX3CR1GFP/+ mice were a kind gift of Steffen Jung (Rehovot, Israel), CD11b.DTR mice, Nr4a12/2 mice, C57BL/6J, and C57BL/6N mice were purchased from The Jackson Laboratory and from Charles River Laboratories (Sulzfeld, Germany). HEK 293 cells and HEK-TLR2 cell were kindly provided by Douglas T. Golenbock (Worcester, MA).

Bacterial strains If not indicated otherwise, S. aureus strain Newman (wild type [wt]) was used. In some experiments, S. aureus SA 113 was used. S. aureus SA 113 and its isogenic lipoprotein diacylglyceryl transferase-deficient strain SA 113 (Dlgt) were kindly provided by Fritz Goetz (Tubingen, Germany). Bacteria were grown in Luria-Bertani broth medium to exponential growth phase, washed with PBS, and resuspended in Dulbecco’s PBS. Bacterial concentrations were determined with an optical spectrophotometer (600 nm, OD600). CFU of the inoculum were verified by serial dilutions on blood agar plates (Columbia Agar with 5% Sheep Blood; bioMe´rieux). In some experiments, bacteria were heat-fixed (hf; 80˚C, 45 min) after adjusting the cell number to 1010 CFU/ml as previously described (13).

Mouse model of skin infection

proximately 10 ml liposomal clodronate or control liposomes were injected i.d. into the ear pinna of anesthetized mice. In addition, CD11b.DTR mice were treated with diphtheria toxin (25 ng/g body weight unnicked, from Corynebacterium diphtheriae; Calbiochem). In both cases, live or hf S. aureus was injected i.d. into the same ear pinna 2 d later. Depletion efficiency was verified by flow cytometric analysis of digested ear skin and histologic analysis of skin cryosections.

Immune cell phenotyping of skin tissue Mouse ears were subjected to enzymatic digestion by dispase (1 mg/ml; STEMCELL Technologies), collagenase II (2 mg/ml; PAA), and DNase I (0.8 mg/ml, Roche) in PBS for 2 h at 1400 rpm shaking and 37˚C. After digestion, the samples were filtered with a 40-mm cell strainer (BD), washed with PBS, and stained with the indicated Abs. The following Abs were used: anti-mouse CD45 eFluor450 (eBioscience), anti-mouse CD45 PerCP-Cy5.5 (eBioscience), anti-mouse CD11b PE-Cy7 (eBioscience), anti-mouse Ly6G FITC (BD Pharmingen), anti-mouse Ly6C PerCP-Cy5.5 (BD Pharmingen), anti-mouse F4/80 allophycocyanin (AbD Serotec), anti-mouse F4/80 PE (eBioscience), anti-mouse CD36 Alexa Fluor 488 (BioLegend), anti-mouse CD169 PE (BioLegend), anti-mouse CD68 allophycocyanin (BioLegend), anti-mouse CD64 PerCP-Cy5.5 (BioLegend), antimouse CD115 allophycocyanin (eBioscience), anti-mouse CD207 PE (eBioscience), anti-mouse CD11c FITC (BD Biosciences), and anti-mouse MHC class 2 eFluor450 (eBioscience). Cell samples were analyzed with a 10-laser flow cytometer (Gallios; Beckman Coulter), and data were analyzed with the Kaluza software (version 1.2; Beckman Coulter).

Immune-cell phenotyping of mouse blood Mouse blood was collected from the submandibular venous plexus of anesthetized mice. After red cell lysis (13 RBC Lysis Buffer Solution; eBioscience), leukocytes were washed with PBS and stained with the indicated Abs. The following Abs were used: anti-mouse CD45 eFluor450 (eBioscience), anti-mouse CD11b PE-Cy7 (eBioscience), anti-mouse Ly6G FITC (BD Pharmingen), anti-mouse Ly6C PerCP-Cy5.5 (BD Pharmingen). Cell samples were analyzed with a 10-laser flow cytometer (Gallios), and data were analyzed with the Kaluza software (version 1.2, Beckman Coulter). Mouse inflammatory monocytes were characterized as CD45highCD11bhighLy6GlowLy6Chigh. Mouse PMLs were characterized as CD45highCD11bhighLy6Ghigh.

Tissue embedding and staining At various time points ear skin was collected and bisected; ear halves were embedded in Tissue-Tek O.C.T. compound (Sakura Finetek Europe B.V.) and subsequently frozen in liquid nitrogen. Eight-micrometer cryosections were taken with a Leica CM 1850 (Leica Biosystems, Nussloch, Germany) and stained with H&E, immunohistochemistry, or immunofluorescence. Immunohistochemistry was performed using the avidin-biotin complex method detected by an alkaline phosphatase catalyzed red chromogen reaction (Dako REALTM Detection System kit; Dako, Glostrup, Denmark). Photographs were taken on a Zeiss Axioskop 50 (Zeiss, Jena, Germany) with the Olympus LC20 camera (Olympus Germany, Hamburg, Germany) with magnifications as depicted. Immunofluorescence was performed using DAPI, FITC, Alexa 488, PE, and/or allophycocyanin fluorophores. Photographs were taken on the confocal laser microscope Zeiss LSM 710 with a 203 objective lens and analyzed with the Zeiss software Zen 2012. The following Abs were used: anti-mouse CD3e (clone 145-2C11, dilution 1:100; eBioscience), anti-mouse gd TCR (clone UC713D5, dilution 1:100; eBioscience), anti-mouse F4/80 (clone MCA497, dilution 1:30; AbD Serotec, Kidlington, U.K.), and goat anti-rat (secondary Ab, polyclonal; Lot-No 103938, dilution 1:100; Jackson Immunoresearch, Suffolk, U.K.).

All procedures were approved by the Regional Council of Baden, W€ urttemberg. Approximately 107 CFU S. aureus in 10 ml PBS were intradermally injected (30-gauge needle, U-100 insulin syringe) into the ear pinna of anesthetized mice (i.p. injection of ketamine, 100 mg/g body weight, and xylazine [Rompun], 20 mg/g body weight). In some experiments, hf bacteria were used for intradermal (i.d.) inoculation (∼108 hf bacteria in 10 ml PBS per ear pinna). Groups of three to four mice were used per experiment, followed by at least one repetition to confirm the results. Lesion morphology was documented by digital photographs (Canon PowerShot A650 IS) of mice ears and analyzed by the software program Adobe Photoshop CS4.

ELISA

Depletion of dermal Mfs in vivo

Mice were sacrificed; then ears were cut off at the hairline and homogenized using a Tissue Lyser (Qiagen). CFU in the homogenized skin tissue were determined by serial dilutions on blood agar plates (Columbia Agar with 5% Sheep Blood; bioMe´rieux).

Resident dermal Mfs were depleted with liposomal clodronate (Clodrosome; Encapsula Nanosciences) in the indicated mouse strains. Ap-

Skin tissue. Ears were mechanically homogenized by Tissue Lyser (Qiagen). IL-1b concentrations in the homogenized skin tissue were quantified by ELISA kits according to the manufacturer’s instructions (R&D Systems). HEK cells. IL-8 levels in HEK_293 or HEK_TLR2 culture medium after 48-h stimulation with SA 113_WT or SA 113_Dlgt were quantified by ELISA kits according to the manufacturer’s instructions (R&D Systems).

Quantification of CFU

The Journal of Immunology In vitro blood cell stimulation and intracellular TNF-a staining Blood was collected from the retrobulbar venous plexus of anesthetized mice. After red cell lysis (13 RBC Lysis Buffer Solution; eBioscience), leukocytes were washed twice and resuspended in RPMI 1640 medium supplemented with 10% FBS with antibiotics (ciprofloxacin, 10 mg/ml). After plating (500,000 cells/well, 24-well plate; Corning Costar), cells were treated with BD GolgiPlug Protein Transport Inhibitor (BD Biosciences) and stimulated for 5 h at 37˚C with hf S. aureus (5 3 107/ml). Next, cells were harvested and cell-surface Ags were stained with Abs against CD45, CD11b, Ly6G, and Ly6C. Then cells were fixed, permeabilized (Fixation/Permeabilization Solution, BD Biosciences), and stained for intracellular TNF-a (anti-mouse TNF-a allophycocyanin; BD Biosciences). TNF-a–producing cell populations were calculated by flowcytometric analysis. Blood inflammatory monocytes were characterized as CD45highCD11bhighLy6GlowLy6Chigh. Blood PMLs were characterized as CD45highCD11bhighLy6Ghigh.

Differentiation of BMDMs Eight- to 10-wk-old donor mice were sacrificed, sprayed with 70% ethanol. Then, both femurs and tibias were rinsed with sterile PBS, separated, and opened. Bone marrow cells were flushed with a 27-gauge needle and passed through a 40-mm cell strainer. Pelleted cells were resuspended in RPMI 1640 medium supplemented with 10% FBS, antibiotics (ciprofloxacin, 10mg/ml), and GM-CSF (50 ng/ml), and plated in T75 cell culture flasks (Greiner Cellstar). BMDMs were used for experiments after 10–12 d of differentiation. After 5 d, the resulting cells are CD45high, CD11bhigh, F4/80high, and CD64high.

In vitro BMDM stimulation and intracellular TNF-a staining Differentiated BMDMs were plated in 24-well plates (400,000 cells/well; Corning Costar), treated with BD GolgiPlug Protein Transport Inhibitor (BD Biosciences), and stimulated for 5 h at 37˚C with hf S. aureus (107 or 108/ml). Next, cells were harvested and cell-surface Ags were stained with Abs against CD45 and CD11b. Then cells were fixed, permeabilized (Fixation/Permeabilization Solution; BD Biosciences), and stained for TNF-a (anti-mouse TNF-a allophycocyanin; BD Biosciences) intracellularly. TNF-a–producing cell populations were calculated by flow cytometric analysis.

In vitro BMDM stimulation, RNA preparation, and quantitative real-time PCR Differentiated BMDMs were plated in 24-well plates (400,000 cells/well; Corning Costar) and stimulated for 3 h at 37˚C with hf S. aureus (107 or 108/ml). Total RNA was extracted using the RNeasy mini kit, according to instructions manual (Qiagen). Quantitative real-time PCR (qRT-PCR) was performed as previously described (20).

Ex vivo stimulation of dermal Mfs, RNA preparation, and qRT-PCR Skin of wt, MyD882/2, or UNC-93B (3D) mice was sterilized with 70% ethanol and subjected to enzymatic digestion by sterile skin digestion solution as described earlier. After 2 h incubation at 37˚C and 1400 rpm, samples were filtered, washed with PBS, and stained with anti-mouse CD45 eFluor450 (eBioscience), anti-mouse CD11b PE-Cy7 (eBioscience), and anti-mouse F4/80 allophycocyanin (AbD Serotec). CD45high CD11bhighF4/80high cells were sorted by FACS (MoFlo Astrios), resuspended in RPMI 1640 medium with 10% FBS plus antibiotics (ciprofloxacin, 10 mg/ml), and plated in 48-well plates (30,000 cells/well, Corning Costar). The next day, adherent Mfs were washed once and stimulated with medium containing hf S. aureus (5 3 107/ml) for 2 h at 37˚C. Total RNA was extracted using the RNeasy micro kit, according to instructions manual (Qiagen). qRT-PCR was performed as previously described (20).

Cytokine production by dermal Mfs and dermal PMLs after stimulation in vivo Wt and MyD882/2 mice were infected i.d. with ∼107 CFU S. aureus. After 24 h, the skin was subjected to enzymatic digestion as described earlier. Then samples were filtered, washed with PBS, and stained with anti-mouse CD45 eFluor450 (eBioscience), anti-mouse CD11b PE-Cy7 (eBioscience), anti-mouse Ly6G FITC (BD Pharmingen), and anti-mouse F4/80 allophycocyanin (AbD Serotec). CD45highCD11bhighLy6GlowF4/80high dermal Mfs, as well as CD45highCD11bhighLy6GhighF4/80low dermal PMLs, were sorted by FACS (MoFlo Astrios). Immediately after sorting, total RNA was

2737 extracted with the RNeasy micro kit according to instructions manual (Qiagen). qRT-PCR was performed as previously described (20).

Results MyD88 is essential for the skin immune response to S. aureus First, we analyzed the phagocyte subset–specific contribution of MyD88 to development and resolution of a defined i.d. staphylococcal infection. Dose-finding experiments revealed 107 CFU S. aureus to induce a localized inflammation followed by spontaneous healing of the skin in 7–9 d in C57BL/6 mice. Sequential analysis of infiltrating PMLs, as well as staphylococci, revealed lower numbers of PMLs in MyD882/2 mice as compared with wt mice at day 2 postinfection (p.i.; Fig. 1A). Yet, whereas PML decreased in wt mice from day 3 onward, MyD882/2 mice showed increasing PML numbers until day 6 (Fig. 1A). Notably, despite excessive PML infiltration, MyD882/2 mice failed to clear the bacteria (Fig. 1B). Accordingly, we wondered whether MyD882/2 PMLs were recruited, but not activated. To this end, we analyzed IL-1b levels of wt and MyD882/2 mice at various time points of infection, because IL-1b has been shown to be predominantly formed by PMLs in soft tissue infections in mice (21) (Fig. 1C). Three days p.i., IL-1b concentrations in the wt skin were strongly upregulated and correlated with PML numbers. Surprisingly, 6 d p.i., levels of IL-1b in infected MyD882/2 mice exceeded levels found in wt mice. These levels correlated with PML numbers, confirming that PMLs are the primary IL-1b source in the infected skin. The spatial relationship of Mf, PML, and staphylococci was further studied by immunohistochemistry. One, 2, and 6 d after i.d. infection of wt and MyD882/2 mice with 107 CFU S. aureus, the skin was bisected at the infection site, embedded, and the cryosections were stained (Fig. 1D, 1E, and Supplemental Fig. 1). Both wt and MyD882/2 mice showed abscesses with granulocytes surrounded by Mfs (Fig. 1D, higher magnifications). Abscesses were generally larger in wt as compared with MyD882/2 mice; however, the latter showed a much increased bacterial burden (Fig. 1D, arrowheads highlighting the bacterial front) as compared with wt mice. Immunofluorescence staining showed F4/80high Mfs at the abscess margin in both wt and MyD882/2 mice (Fig. 1E). In contrast, CD3high T cells were found intraepidermally but were not associated with bacteria either in wt or in MyD882/2 mice (Fig. 1E). Accordingly, abundant and highly activated PMLs failed to timely kill staphylococci and to clear the infection in MyD882/2 mice. This suggested that MyD88 in other cells than PMLs determined the immune response to S. aureus and that the MyD88dependent defect in these cells could not be fully compensated for by highly activated PMLs. Next, we aimed at disconnecting bacterial viability and proliferation from PML recruitment by challenging mice i.d. with 108 fixed S. aureus. Notably, at very early stages of infection (4 h p.i.), fixed S. aureus were as potent as viable bacteria in recruiting PMLs to the skin (Supplemental Fig. 2A). As compared with wt mice, MyD882/2 mice showed a substantially delayed PML recruitment (Fig. 1F). To confirm that delayed PML recruitment was not a strain-specific characteristic of the Newman strain, we infected wt and MyD882/2 mice i.d. with 107 CFU of the distinct serotype 8 S. aureus strain SA 113. We found a similar MyD88 dependence of PMLs infiltrating after 24 h as described with strain Newman (Fig. 1G). Accordingly, MyD88 is important for the early PML recruitment to the site of infection independent of bacterial proliferation and toxin formation. We hypothesized that the skin phenotype in MyD882/2 mice, and potentially that in humans, is largely due to defects in resident cells and cannot be

2738

MyD88 IN MACROPHAGES IS CRITICAL FOR ABSCESS RESOLUTION

FIGURE 1. MyD88 is essential for the skin immune response to S. aureus wt, and MyD882/2 mice were i.d. inoculated with 107 CFU S. aureus (Newman) into the left ear pinna and analyzed after 2, 3, and 6 d for (A) skin PML numbers (Ly6Ghigh skin cells gated on CD45highCD11bhigh cells) by FACS for (B) CFU per lesion (dilution series plated from homogenized skin tissue) and for (C) skin IL-1b levels. (D and E) Eight-micrometer skin cryosections from wt and MyD882/2 mice stained with either H&E (D) or immunofluorescence (E) 1 d after i.d. infection with 107 CFU S. aureus. (F) wt and MyD882/2 mice were i.d. inoculated with 108 hf S. aureus, and skin PML numbers were analyzed after 2 and 4 h by FACS. (G) wt and MyD882/2 mice were i.d. inoculated with 107 CFU S. aureus (SA 113) into the left ear pinna and analyzed after 24 h for skin PML numbers (Ly6Ghigh skin cells gated on CD45highCD11bhigh cells) by FACS. Groups of three to four mice were used per experiment followed by at least one repetition to confirm the results. All the data are mean 6 SEM. Original magnification 320 (D and E). *p , 0.05, **p , 0.01 (two-tailed unpaired t test).

compensated for, or may even be aggravated, by highly active PMLs. Ly6Chigh monocytes are involved in the cellular skin immunity to S. aureus Next, we wondered whether monocytes were essentially involved in mediating inflammation in staphylococcal skin infection. To address this question, we exploited the fact that the CCR2 is essential for bone marrow emigration of inflammatory monocytes, and thus indispensable for their recruitment and effector function in the periphery (22). CCR22/2 mice have significantly reduced numbers of Ly6Chigh inflammatory monocytes (23). Accordingly, wt and CCR22/2 mice were challenged i.d. with S. aureus. After 4, 24, and 48 h, skin and blood Ly6Chigh inflammatory monocytes were analyzed (Fig. 2A, 2B). In wt mice, inflammatory monocytes infiltrated the skin as early as 4 h p.i., with increasing numbers until 24 h p.i. Thereafter, monocytes disappeared in parallel with PMLs (Fig. 2B). In peripheral blood, Ly6Chigh monocytes were increasing in number over the entire course of infection (Fig. 2B), indicating an overall important role in the immune response to S. aureus, in particular at late stages of infection. In contrast with wt mice, both circulating blood Ly6Chigh monocytes and skininfiltrating Ly6Chigh monocytes were almost absent in CCR22/2

mice. CCR22/2 mice showed slightly lower numbers of F4/80highresident dermal Mfs, yet no differences could be detected in the recruitment or activation of PMLs (as indicated by high IL-1b tissue levels), neither at very early time points, when fixed staphylococci were used (Fig. 2C), nor at later time points with live staphylococci (Fig. 2D and data not shown). Furthermore, bacterial clearance was normal in these mice (data not shown). These findings indicate that, although inflammatory monocytes are rapidly recruited to the site of staphylococcal infection, they do not substantially contribute to PML recruitment, bacterial killing, and cytokine secretion. Resident dermal Mfs are important for recruitment and regulation of PMLs Mfs are sentinel cells that can be assumed to contribute to the functional sensing body for invading staphylococci. To analyze the overall role of resident Mfs in the defense against S. aureus, we depleted them by application of 10 ml liposomal clodronate i.d. (Supplemental Fig. 3A, 3B). In our hands, this method depletes 70–80% of all F4/80high Mfs in the skin without affecting other bone marrow–derived cells like PMLs or keratinocytes (data not shown). After depletion of dermal Mfs, infection with S. aureus resulted in significantly increased PML influx (Fig. 3A), reduced

The Journal of Immunology

2739

FIGURE 2. Ly6Chigh monocytes are involved in the cell immune response to S. aureus. (A) Exemplary FACS blots gated on wt CD45highCD11bhigh cells before (0 h) and after (48 h) i.d. infection with 107 CFU S. aureus. (B) Proportion of skin and blood Ly6Chigh inflammatory monocytes in the course of skin infection in wt and CCR22/2 mice. (C) wt and CCR22/2 mice were i.d. inoculated with 108 hf S. aureus, and skin PML numbers (Ly6Ghigh skin cells gated on CD45highCD11bhigh cells) were analyzed after 4 h by FACS. (D) wt and CCR22/2 mice were i.d. inoculated with 107 CFU S. aureus and analyzed after 2 and 5 d for skin PML numbers. Groups of three to four mice were used per experiment followed by at least one repetition to confirm the results. All data are mean 6 SEM. **p , 0.01, ***p , 0.001 (two-tailed unpaired t test).

bacterial clearance (Fig. 3B), and higher local IL-1b concentrations (Fig. 3C) as compared with skin treated with control liposomes. Moreover, PML recruitment was substantially delayed, when clodronate-treated mice were challenged with fixed bacteria (Fig. 3D). As an alternative approach, dermal Mfs were depleted in transgenic CD11b.DTR mice with diphtheria toxin (25 ng/g body weight) 48 h before S. aureus infection. We found this procedure to deplete F4/80high Mfs by 60–70%, whereas circulating PMLs and monocytes were not affected (Supplemental Fig. 3C). Notably, CD11b.DTR mice developed increased PML numbers in the blood after DT treatment (data not shown). In striking analogy with the situation after clodronate treatment, skin PML numbers were reduced early after staphylococcal challenge (Fig. 3E). Thus, very similar to the situation in MyD882/2 mice, mice devoid of dermal Mfs show a delayed, but then overwhelming PML response that could not control the infection in a timely fashion. The similarity between Mf-depleted and MyD882/2 mice in S. aureus skin infection suggests that dermal Mfs are recruiting blood PMLs to the site of infection via a MyD88dependent, and Mf-specific, signaling pathway. To substantiate this model, we prepared bone marrow chimeric mice (wt→wt and

wt→MyD882/2), challenged them with 108 fixed S. aureus per mouse ear, and analyzed skin PML numbers after 4 h (Fig. 3F). MyD882/2 mice transplanted with wt bone marrow mimicked conventional MyD882/2 mice with respect to delayed PML recruitment to the site of infection (p = 0.0566). Of note, only mice with a blood cell chimerism of .90% were used for infection experiments. Taken together, these data strongly suggest that sensing of S. aureus by resident dermal Mfs via the TLR adaptor protein MyD88 determines the course of PML recruitment and bacterial killing over days p.i. Mfs and monocytes, but not PMLs, are sensing S. aureus in an MyD88-dependent fashion MyD88 is a well-established and response-limiting adaptor protein for all TLRs except TLR3 (24), as well as the IL-1Rs and IL-18Rs (25, 26). Thus, MyD88 represents a key component in the immune response against a wide range of microbial pathogens (27). Moreover, TLR2 and MyD882/2 mice have been found to be more susceptible to local and systemic S. aureus infections than wt mice (28–30). However, the specific role of MyD88 on the myeloid cell subset level is not well understood. To investigate

2740

MyD88 IN MACROPHAGES IS CRITICAL FOR ABSCESS RESOLUTION

FIGURE 3. Resident skin Mfs are recruiting PMLs and regulate the control and elimination of S. aureus infecting the skin. Before i.d. infection with 107 CFU S. aureus, resident skin Mfs were depleted by liposomal clodronate. After 2 and 5 d of infection, skin PML numbers (Ly6Ghigh skin cells gated on CD45highCD11bhigh cells) (A), CFU per lesion (B), and skin IL-1b levels (C) were analyzed. (D) After depletion of skin Mfs by liposomal clodronate and i.d. injection of 108 hf S. aureus, skin-infiltrating PMLs were analyzed after 4 h by FACS. (E) wt and CD11b.DTR mice were i.d. treated with diphtheria toxin (25 ng/g body weight). After successful depletion of CD11b+ skin cells, 108 heat-fixed S. aureus were i.d. injected and skin cell composition analyzed after 4 h by FACS. (F) wt and MyD882/2 chimeras transplanted with 5 3 106 wt bone marrow cells were i.d. treated with 108 hf S. aureus and skin PML numbers analyzed after 4 h by FACS. Groups of three to four mice were used per experiment, followed by at least one repetition to confirm the results. All data are mean 6 SEM. *p , 0.05, **p , 0.01, ***p , 0.001 (two-tailed unpaired t test). F4/80+, F4/80high resident skin Mfs.

this, we stimulated primary blood leukocytes from wt and MyD882/2 mice with S. aureus, and TNF-a production was analyzed by FACS (Fig. 4A). Surprisingly, Ly6Chigh inflammatory monocytes, but not Ly6Ghigh PMLs, showed an MyD88 dependency of the S. aureus–induced TNF-a response. This suggested that inflammatory monocytes, as well as blood PMLs, were important producers of TNF-a in response to S. aureus. With respect to the role of MyD88 in the response to staphylococci, inflammatory monocytes behaved like BMDMs, where MyD88 was absolutely required for the TNF response (Fig. 4B). Given the variable role of MyD88 in the myeloid subset–specific TNF response to staphylococci, we next analyzed primary dermal Mfs in this context. We purified CD45highCD11bhighF4/80high Ly6Glow Mfs from wt and MyD882/2 mice by FACS sorting (Fig. 4D). We found dermal Mfs to be CD64high, CD36high, CD68high, CD169high, CD115high, cx3cr1pos, MHC class IIhigh, CD207low, and CD11clow (Fig. 4E). Stimulation of dermal Mfs with S. aureus ex vivo (Fig. 4C) revealed a strict MyD88 depen-

dency of transcriptional inflammatory cytokine activation (as determined by qRT-PCR for pro–IL-1b and TNF-a). To assess the activation of myeloid cell subsets in vivo, we sorted dermal Mfs and skin-infiltrating PMLs from wt and MyD882/2 mice, which had been challenged for 1 d with S. aureus, and performed qRTPCR for cytokine transcripts (Supplemental Fig. 2C, 2D). We found that staphylococci induce transcription of both TNF and IL1b in dermal Mfs and PMLs infiltrating the dermis. In contrast with PMLs purified from mouse blood and challenged in vitro, the in vivo cytokine response of infiltrating PMLs appeared to be MyD88 dependent, although the differences did not reach statistical significance and the role of MyD88 was larger in dermal Mfs (Fig. 4C and Supplemental Fig. 2C, 2D). Thus, resident dermal Mfs are likely to regulate the activity of infiltrating PMLs in an MyD88-dependent fashion in trans. Notably, the important role of MyD88 in the Mf response to staphylococci was not restricted to the Newman strain, but was similarly apparent when strain SA 113 was used (Supplemental Fig. 4).

The Journal of Immunology

2741

FIGURE 4. Skin Mfs, unlike blood PMLs, show a specific MyD88 dependency in their cytokine response to S. aureus. (A) RBC-lysed blood cells from wt and MyD882/2 mice were stimulated with hf S. aureus (5 3 107/ml) and stained for CD45, CD11b, Ly6G, and Ly6C. Cells were fixed, permeabilized, and stained for TNF-a intracellularly (i.c.). TNF-a–producing cell populations were calculated by FACS. (B) BMDMs from wt and MyD882/2 mice were stimulated with hf S. aureus (107 or 108/ml) and stained for CD45, CD11b, and TNF-a i.c. mean fluorescence intensity (MFI) of TNF-a was calculated by FACS. (C) FACS-sorted CD45highCD11bhighLy6GlowF4/80high skin Mfs from wt and MyD882/2 mice were stimulated with hf S. aureus (108/ml) for 2 h, total RNA was prepared, and qRT-PCR for pro–IL-1b and TNF-a was performed. (D) Digested skin mouse Mfs were characterized as CD45highCD11bhigh F4/80highLy6Glow. (E) Phenotype analysis of resident skin Mfs, gated on CD45highCD11bhighF4/80highLy6Glow cells. Histograms represent staining for CD64, CD36, CD68, CD169, CD115, CX3CR1-gfp, MHC class II, CD207 (langerin), and CD11c (dark gray) versus fluorescence minus one (FMO) control (light gray). All data are mean 6 SEM. *p , 0.05, **p , 0.01, ***p , 0.001 (two-tailed unpaired t test). M, Ly6Chigh inflammatory monocytes; N, Ly6Ghigh PMLs; SA, heat-fixed S. aureus.

Resident dermal Mfs are sensing S. aureus independently of endosomal TLRs It has been reported previously that ssRNA from S. aureus, as well as many other Gram-positive bacteria, is an essential effector for the activation of inflammatory cytokines in BMDMs (13, 31). Endosomal TLRs are most likely key RNA sensors in Mfs. This assumption was based on the exclusive role of the endoplasmic reticulum membrane protein UNC-93B, which is crucial for endosomal localization of nucleotide-sensing TLRs. In UNC-93B (3D) mice, a single point mutation (H412R) abolishes signaling via TLR3, TLR7, and TLR9 (32). In concordance with these data, S. aureus–induced expression of pro–IL-1 was largely UNC-93B dependent in BMDMs (Fig. 5A). In contrast, inflammatory monocytes and PMLs mounted a robust TNF response to S. aureus in an UNC-93B–independent fashion (Fig. 5B). Moreover, and even more important in the context of this study, dermal Mfs from wt and UNC-93B (3D) mice were indistinguishable in the cytokine response to S. aureus ex vivo (Fig. 5C). This finding suggested that Mfs receive an imprinting by their microenvironment that determines the usage of distinct TLRs when in contact with bacteria. To further address whether sensing of nucleotide acids by endosomal TLRs plays a role in PML recruitment and bacterial

killing in S. aureus skin infection, we infected wt and UNC-93B (3D) mice with either fixed S. aureus (Newman or SA 113) or viable S. aureus. We analyzed PML numbers after 4 h (fixed bacteria, Fig. 5D and Supplemental Fig. 4D) or after 2 and 5 d (viable bacteria; Fig. 5E). Moreover, we analyzed the bacterial load and the IL-1b response in these mice (data not shown). UNC93B–deficient mice showed a slightly faster PML recruitment and the same bacterial killing capacity, as well as IL-1b production, as compared with wt mice. This suggests that sensing of nucleic acids via endosomal TLRs does not play an important role in innate skin immunity to S. aureus. Again, this observation was true for both the Newman strain and the strain SA 113. These data are consistent with the notion that dermal Mfs are sensing S. aureus independently of endosomal TLRs. TLR2 is dispensable for the early immune response of dermal Mfs against S. aureus In view of the redundancy of endosomal TLRs in staphylococcal sensing, we wondered whether TLR2 may be important in mediating the early MyD88-dependent response. Thus, we made use of S. aureus with a genetic deletion of the lipoprotein transferase lgt. In these bacteria, the putative TLR2-activating diacylated

2742

MyD88 IN MACROPHAGES IS CRITICAL FOR ABSCESS RESOLUTION

FIGURE 5. Resident skin Mfs are sensing S. aureus independently of TLR2 and endosomal TLRs. (A) BMDMs from wt and UNC-93B (3D) mice were stimulated with hf S. aureus (107 or 108/ml) for 3 h, total RNA was prepared, and qRT-PCR for pro–IL-1b was performed. (B) RBC-lysed blood cells from wt and UNC-93B (3D) mice were stimulated with hf S. aureus (5 3 107/ml) and stained for CD45, CD11b, Ly6G, and Ly6C. Cells were fixed, permeabilized, and stained for TNF-a i.c. TNF-a–producing cell populations were calculated by FACS. (C) FACS-sorted CD45highCD11bhighLy6GlowF4/80high skin Mfs from wt and UNC-93B (3D) mice were stimulated with hf S. aureus (108/ml) for 2 h, total RNA was prepared, and qRT-PCR for pro–IL-1b and TNF-a was performed. (D) wt and UNC-93B (3D) mice were i.d. inoculated with 108 hf S. aureus and skin PML numbers (Ly6Ghigh skin cells gated on CD45highCD11bhigh cells) were analyzed after 4 h by FACS. (E) wt and UNC-93B (3D) mice were i.d. inoculated with 107 CFU S. aureus and analyzed after 2 and 5 d for skin PML numbers by FACS. (F) RBC-lysed wt blood cells were stimulated with hf SA 113 (wt) or Dlgt (5 3 107/ml) and stained for CD45, CD11b, Ly6G, and Ly6C. Cells were fixed, permeabilized, and stained for TNF-a intracellularly. TNF-a–producing Ly6Chigh inflammatory monocytes were calculated by FACS. (G) FACS-sorted CD45highCD11bhighLy6GlowF4/80high wt skin Mfs were stimulated with hf SA 113 (wt) or Dlgt for 2 h, total RNA was prepared, and qRT-PCR for pro–IL-1b and TNF-a was performed. (H) wt or UNC-93B (3D) mice were i.d. inoculated with 10 ml PBS or 10 ml PBS containing 108 CFU hf SA 113 (wt) or Dlgt, and skin PML numbers (Ly6Ghigh skin cells gated on CD45highCD11bhigh cells) were analyzed after 4 h by FACS. Groups of three to four mice were used per experiment, followed by at least one repetition to confirm the results. All data are mean 6 SEM. *p , 0.05, **p , 0.01, ***p , 0.001 (two-tailed unpaired t test). M, Ly6Chigh inflammatory monocytes; N, Ly6Ghigh PMLs; SA, hf S. aureus.

lipoproteins are not formed (33). First, we compared fixed Dlgt S. aureus with the isogenic wt strain for its TLR2 activation in epithelial cells and found deletion of lgt to abrogate the TLR2specific IL-8 response (Supplemental Fig. 2B). Next, we analyzed whether early (4 h) recruitment of PMLs differed between Dlgt S. aureus and the isogenic wt strain. Importantly, we used fixed S. aureus, which induce an early dermal inflammatory response that is very similar to that induced by live bacteria (Supplemental Fig. 2). Thereby we excluded TLR2-independent effects of the lgt disruption, for example, those related to iron acquisition and bacterial persistence (34). We found that Dlgt S. aureus and isogenic wt were equally potent in cytokine induction in inflammatory monocytes and dermal Mfs (Fig. 5F, 5G). Using Dlgt S. aureus offered the opportunity to test whether TLR2 and endosomal TLRs were redundant in their recognition of S. aureus, thereby compensating for the lacking engagement of the other receptor group. Accordingly, we abrogated both TLR2 and endosomal TLR activation by infecting UNC-93B–deficient mice with Dlgt S. aureus (Fig. 5H). Notably, we did not observe a difference between wt mice infected with wt S. aureus, UNC-93B–

deficient mice infected with Dlgt S. aureus, and all other combinations. Accordingly, both TLR2 engagement by lipoproteins and endosomal TLR activation by staphylococcal nucleic acids cannot explain the important role of MyD88 in the defense against staphylococci. Skin infection leads to expansion of resident dermal Mfs As outlined earlier, our model of a localized staphylococcal skin infection is associated with dramatic quantitative and qualitative changes in cellular innate immunity at site. After the infection has been controlled, immunity has to contract to allow for reestablishment of cellular homeostasis. Our model revealed that the kinetics of this resolution are heavily dependent on MyD88. Moreover, Mfs appear to orchestrate the cellular composition. In this article, we addressed whether and how MyD88 signaling quantitatively affects the Mf population. In wt mice, PMLs increased until day 2 p.i. and then decreased, closely following the bacterial load at site. Ly6Chigh inflammatory monocytes initially behaved in a similar fashion, however, in comparatively low numbers and remaining constant until day 7. In contrast, F4/80high

The Journal of Immunology dermal Mfs initially decreased but expanded in the later course of infection (Fig. 6A, 6B). Notably, MyD88 deletion did not affect Mf numbers over the entire course of infection, in contrast with PML and monocyte numbers (Fig. 6C). In sharp contrast, CCR22/2 mice failed to replace and expand Mfs until day 7 of infection (Fig. 6E). Therefore, Ly6Chigh inflammatory monocytes contribute to the expansion of dermal Mfs, but this has no apparent functional consequences for the host response to S. aureus. In contrast, Nr4a12/2 mice, which are deficient in circulating Ly6Clow patrolling monocytes (35), showed normal numbers of Mfs during the course of infection (Fig. 6D). These data indicate that resident dermal Mfs are expanding after staphylococcal infection without contribution of circulating Ly6Clow patrolling monocytes and the TLR adaptor protein MyD88, but in a clear dependency on Ly6Chigh inflammatory monocytes.

Discussion In staphylococcal skin infection, resident dermal Mfs are sensing skin invading staphylococci in a strictly MyD88-dependent, but TLR2- and endosomal TLR–independent, fashion. Proper Mf

2743 function is essential for activation and regulation of skininfiltrating PMLs. Re-establishment of myeloid cell homeostasis at the site of staphylococcal infection requires early Mf-specific, MyD88-dependent sensing, timely containment of bacteria, and Mf renewal by incoming Ly6Chigh inflammatory monocytes. The TLR system plays an important role in host resistance against staphylococci. Mice that lack the TLR adaptor protein MyD88 are highly susceptible to systemic infections (29). They develop larger skin lesions with higher bacterial counts and show defective PML recruitment and cytokine production (30). Nevertheless, the cell type and sensing mechanisms involved in the early staphylococcal recognition in the skin remained unclear. Our data uncover resident dermal Mfs as initial sensors of skin-infiltrating S. aureus. Surprisingly, conventional MyD882/2 mice and mice with local dermal Mf deficiency exhibit an almost indistinguishable phenotype in staphylococcal infection. The prominent role of MyD88 in Mf signaling is furthermore highlighted by the myeloid lineage specificity, that is, PMLs sense staphylococci in a largely MyD88-independent fashion.

FIGURE 6. Expansion of resident skin Mfs. Proportion of wt skin CD45highCD11bhighLy6GlowF4/80high cells (% of total skin cells) before (day 0) and after (days 2, 5, and 7) i.d. injection of 107 CFU S. aureus. (A) Exemplary FACS blots gated on CD45highCD11bhigh cells. (B) Bar graph demonstrating the proportions of wt myeloid skin cells. (C) Proportion of skin Mfs from wt and MyD882/2 mice before (day 0) and after (day 6) i.d. injection of 107 CFU S. aureus. Proportion of skin Mfs from wt and Nr4a12/2 mice (D) and CCR22/2 mice (E) before (day 0) and after (days 2, 5, 6, and 7) i.d. infection with 107 CFU S. aureus. All data are mean 6 SEM. Groups of three to four mice were used per experiment, followed by at least one repetition to confirm the results. **p , 0.01, ***p , 0.001 (two-tailed unpaired t test). F4/80high, resident skin Mfs; Ly6Chigh, inflammatory monocytes.

2744

MyD88 IN MACROPHAGES IS CRITICAL FOR ABSCESS RESOLUTION

In contrast with MyD88, the established S. aureus–sensing TLRs were dispensable for a proper dermal Mf response against staphylococci. This seems noteworthy given the relatively large body of literature on TLR2 and endosomal TLRs in the Mf response to staphylococci, both in vitro and in vivo (13, 29–31, 36, 37). We made use of staphylococci that lack protein acylation and, therefore, do not activate TLR2. This strain allowed us to analyze lacking TLR2 activation in the context of deficient nucleic acid TLR sensing, which is dependent on the endoplasmic reticulum protein UNC-93B. The experiments revealed that both systems are dispensable without compensating for each other. How can differences to other studies with respect to the engagement of specific TLRs by staphylococci be explained? The answer may lie in the specific animal model that is used. We took greatest care to fine-tune our model in a way that the infection remained localized and self-limiting in 6 d in wt mice. Notably, we did not detect any systemic effects of the infection, such as elevated inflammatory cytokines or bacterial spread, and the mice seemed unaffected by the lesion (with respect to activity and weight gain). It is conceivable that more aggressive skin models like those used by Miller et al. (30) are associated with circulation of TLR2-activating lipopeptides from staphylococci. Furthermore, spreading of staphylococci to other body sites but the skin likely impacts on TLR usage by myeloid cells, which is specific for site of residence or origin. This notion is supported by our observation that BMDMs are strictly dependent on UNC-93B in the recognition of staphylococci, whereas this molecule is dispensable for activation of dermal Mfs. The validity of our model as a reflection of the everyday event in humans, where colonizing bacteria enter the skin through minor lesions in the epidermis, receives support by the fact that patients with MyD88 deficiency, but not those with UNC-93B deficiency or TLR2/TIRAP deficiency, present with staphylococcal skin infections as a leading problem (38). The recent and striking finding that S. aureus stimulates granulopoiesis in skin wounds by engaging TLR2 in hematopoietic stem and progenitor cells (39) further underlines the role of cell lineage-specific preference of TLR2 by staphylococci. In addition, our findings constitute a cautionary note in interpreting data generated in BMDMs, which are the workhorses in many studies addressing cellular innate immunity to bacteria. This study further demonstrates that Mf-dependent sensing of staphylococci controls PMLs at the site of infection. PMLs exit the bone marrow in response to inflammatory stimuli and contribute to the fast clearance of bacteria via an elaborate antimicrobial machinery (40, 41). PML-derived IL-1b contributes to abscess formation in S. aureus infection (21). However, we show in this article that, when dermal Mfs fail to immediately recruit PMLs to the site of staphylococcal invasion, the finally arriving and fully activated PMLs fail to timely control the infection. Notably, PMLs respond to S. aureus in a largely MyD88independent fashion. Thus, although PMLs are clearly essential for the clearance of staphylococcal skin infection, they need to be alerted and directed by resident dermal Mfs for proper function. Resident dermal Mfs are important components of the skin immune system (7, 8) arising either from the hematopoietic system or the fetal liver and yolk sac (42–44). Although the primary source and replacement of most tissue resident Mfs has been well appreciated, the origin of dermal Mfs, as well as their expansion potential in staphylococcal infection, remains incompletely understood. It is further unclear whether these dermal Mfs constitute a homogenous terminally differentiated cell population, which is arising and replenished exclusively from blood monocytes, or whether they are able to self-renew and differentiate at

site without contribution of blood-derived mononuclear cells (16). The CCR2 is essential for bone marrow emigration of Ly6Chigh monocytes, and thus indispensable for the recruitment and function of these inflammatory myeloid cells (22). However, whereas we found circulating inflammatory monocytes to increase over the entire course of active staphylococcal infection, they did not contribute essentially to initiation or clearance of the infection. It appears that Ly6Chigh inflammatory monocytes are predominantly responsible for dermal Mf renewal and expansion in the healing process starting late in infection. Indeed, it seems that the increase in Mfs is not directly linked to the resolution of infection, because CCR22/2 mice do not show an overt defect in the staphylococcal skin model. Accordingly, it appears that the recovery of Mf numbers rather accompanies than induces termination of the infection. It is further tempting to speculate that a rapid reestablishment of tissue cellularity is critical for the response to a subsequent insult. In contrast with inflammatory Ly6Chigh monocytes, we did not identify any role for circulating so-called patrolling Ly6Clow monocytes, neither in controlling inflammation nor in the replacement of dermal Mfs. In summary, this study highlights the essential and specific role of early MyD88-dependent sensing by resident dermal Mfs in the timely control and clearing of staphylococcal infection accompanied by the re-establishment of the myeloid cell homeostasis in the skin.

Acknowledgments We are indebted to Antigoni Triantafyllopoulou (University of Freiburg) for helping with Mf phenotyping, Marina Freudenberg (University of Freiburg) for the provision of knockout mice, and Bernhard Kremer, Monika Haeffner, Anita Imm, and Jan Bodinek-Wersing for outstanding technical assistance.

Disclosures The authors have no financial conflicts of interest.

References 1. Noskin, G. A., R. J. Rubin, J. J. Schentag, J. Kluytmans, E. C. Hedblom, C. Jacobson, M. Smulders, E. Gemmen, and M. Bharmal. 2007. National trends in Staphylococcus aureus infection rates: impact on economic burden and mortality over a 6-year period (1998-2003). Clin. Infect. Dis. 45: 1132–1140. 2. Moran, G. J., F. M. Abrahamian, F. Lovecchio, and D. A. Talan. 2013. Acute bacterial skin infections: developments since the 2005 Infectious Diseases Society of America (IDSA) guidelines. J. Emerg. Med. 44: e397–e412. 3. Picard, C., J. L. Casanova, and A. Puel. 2011. Infectious diseases in patients with IRAK-4, MyD88, NEMO, or IkBa deficiency. Clin. Microbiol. Rev. 24: 490– 497. 4. Picard, C., H. von Bernuth, P. Ghandil, M. Chrabieh, O. Levy, P. D. Arkwright, D. McDonald, R. S. Geha, H. Takada, J. C. Krause, et al. 2010. Clinical features and outcome of patients with IRAK-4 and MyD88 deficiency. Medicine (Baltimore) 89: 403–425. 5. Borregaard, N. 2010. Neutrophils, from marrow to microbes. Immunity 33: 657– 670. 6. Kielian, T., E. D. Bearden, A. C. Baldwin, and N. Esen. 2004. IL-1 and TNFalpha play a pivotal role in the host immune response in a mouse model of Staphylococcus aureus-induced experimental brain abscess. J. Neuropathol. Exp. Neurol. 63: 381–396. 7. Chen, G. Y., and G. Nun˜ez. 2010. Sterile inflammation: sensing and reacting to damage. Nat. Rev. Immunol. 10: 826–837. 8. Abtin, A., R. Jain, A. J. Mitchell, B. Roediger, A. J. Brzoska, S. Tikoo, Q. Cheng, L. G. Ng, L. L. Cavanagh, U. H. von Andrian, et al. 2014. Perivascular macrophages mediate neutrophil recruitment during bacterial skin infection. Nat. Immunol. 15: 45–53. 9. Fournier, B., and D. J. Philpott. 2005. Recognition of Staphylococcus aureus by the innate immune system. Clin. Microbiol. Rev. 18: 521–540. 10. Mo¨lne, L., A. Corthay, R. Holmdahl, and A. Tarkowski. 2003. Role of gamma/ delta T cell receptor-expressing lymphocytes in cutaneous infection caused by Staphylococcus aureus. Clin. Exp. Immunol. 132: 209–215. 11. Deshmukh, H. S., J. B. Hamburger, S. H. Ahn, D. G. McCafferty, S. R. Yang, and V. G. Fowler, Jr. 2009. Critical role of NOD2 in regulating the immune response to Staphylococcus aureus. Infect. Immun. 77: 1376–1382. 12. Hruz, P., A. S. Zinkernagel, G. Jenikova, G. J. Botwin, J. P. Hugot, M. Karin, V. Nizet, and L. Eckmann. 2009. NOD2 contributes to cutaneous defense against

The Journal of Immunology

13.

14.

15. 16. 17. 18. 19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

Staphylococcus aureus through alpha-toxin-dependent innate immune activation. Proc. Natl. Acad. Sci. USA 106: 12873–12878. Deshmukh, S. D., B. Kremer, M. Freudenberg, S. Bauer, D. T. Golenbock, and P. Henneke. 2011. Macrophages recognize streptococci through bacterial singlestranded RNA. EMBO Rep. 12: 71–76. Lembo, A., C. Kalis, C. J. Kirschning, V. Mitolo, E. Jirillo, H. Wagner, C. Galanos, and M. A. Freudenberg. 2003. Differential contribution of Toll-like receptors 4 and 2 to the cytokine response to Salmonella enterica serovar Typhimurium and Staphylococcus aureus in mice. Infect. Immun. 71: 6058– 6062. Tay, S. S., B. Roediger, P. L. Tong, S. Tikoo, and W. Weninger. 2013. The skinresident immune network. Curr. Dermatol. Rep. 3: 13–22. Sieweke, M. H., and J. E. Allen. 2013. Beyond stem cells: self-renewal of differentiated macrophages. Science 342: 1242974. Nestle, F. O., P. Di Meglio, J. Z. Qin, and B. J. Nickoloff. 2009. Skin immune sentinels in health and disease. Nat. Rev. Immunol. 9: 679–691. Malissen, B., S. Tamoutounour, and S. Henri. 2014. The origins and functions of dendritic cells and macrophages in the skin. Nat. Rev. Immunol. 14: 417–428. Niebuhr, M., C. Lutat, S. Sigel, and T. Werfel. 2009. Impaired TLR-2 expression and TLR-2-mediated cytokine secretion in macrophages from patients with atopic dermatitis. Allergy 64: 1580–1587. Deshmukh, S. D., S. M€ uller, K. Hese, K. S. Rauch, J. Wennekamp, O. Takeuchi, S. Akira, D. T. Golenbock, and P. Henneke. 2012. NO is a macrophage autonomous modifier of the cytokine response to streptococcal single-stranded RNA. J. Immunol. 188: 774–780. Cho, J. S., Y. Guo, R. I. Ramos, F. Hebroni, S. B. Plaisier, C. Xuan, J. L. Granick, H. Matsushima, A. Takashima, Y. Iwakura, et al. 2012. Neutrophil-derived IL-1b is sufficient for abscess formation in immunity against Staphylococcus aureus in mice. PLoS Pathog. 8: e1003047. Boring, L., J. Gosling, S. W. Chensue, S. L. Kunkel, R. V. Farese, Jr., H. E. Broxmeyer, and I. F. Charo. 1997. Impaired monocyte migration and reduced type 1 (Th1) cytokine responses in C-C chemokine receptor 2 knockout mice. J. Clin. Invest. 100: 2552–2561. Serbina, N. V., and E. G. Pamer. 2006. Monocyte emigration from bone marrow during bacterial infection requires signals mediated by chemokine receptor CCR2. Nat. Immunol. 7: 311–317. Biondo, C., A. Midiri, L. Messina, F. Tomasello, G. Garufi, M. R. Catania, M. Bombaci, C. Beninati, G. Teti, and G. Mancuso. 2005. MyD88 and TLR2, but not TLR4, are required for host defense against Cryptococcus neoformans. Eur. J. Immunol. 35: 870–878. Adachi, O., T. Kawai, K. Takeda, M. Matsumoto, H. Tsutsui, M. Sakagami, K. Nakanishi, and S. Akira. 1998. Targeted disruption of the MyD88 gene results in loss of IL-1- and IL-18-mediated function. Immunity 9: 143–150. Fitzgerald, K. A., and L. A. O’Neill. 2000. The role of the interleukin-1/Toll-like receptor superfamily in inflammation and host defence. Microbes Infect. 2: 933– 943. Akira, S., and K. Takeda. 2004. Toll-like receptor signalling. Nat. Rev. Immunol. 4: 499–511. Hoebe, K., P. Georgel, S. Rutschmann, X. Du, S. Mudd, K. Crozat, S. Sovath, L. Shamel, T. Hartung, U. Za¨hringer, and B. Beutler. 2005. CD36 is a sensor of diacylglycerides. Nature 433: 523–527.

2745 29. Takeuchi, O., K. Hoshino, and S. Akira. 2000. Cutting edge: TLR2-deficient and MyD88-deficient mice are highly susceptible to Staphylococcus aureus infection. J. Immunol. 165: 5392–5396. 30. Miller, L. S., R. M. O’Connell, M. A. Gutierrez, E. M. Pietras, A. Shahangian, C. E. Gross, A. Thirumala, A. L. Cheung, G. Cheng, and R. L. Modlin. 2006. MyD88 mediates neutrophil recruitment initiated by IL-1R but not TLR2 activation in immunity against Staphylococcus aureus. Immunity 24: 79–91. 31. Oldenburg, M., A. Kr€uger, R. Ferstl, A. Kaufmann, G. Nees, A. Sigmund, B. Bathke, H. Lauterbach, M. Suter, S. Dreher, et al. 2012. TLR13 recognizes bacterial 23S rRNA devoid of erythromycin resistance-forming modification. Science 337: 1111–1115. 32. Brinkmann, M. M., E. Spooner, K. Hoebe, B. Beutler, H. L. Ploegh, and Y. M. Kim. 2007. The interaction between the ER membrane protein UNC93B and TLR3, 7, and 9 is crucial for TLR signaling. J. Cell Biol. 177: 265–275. 33. Stoll, H., J. Dengjel, C. Nerz, and F. Go¨tz. 2005. Staphylococcus aureus deficient in lipidation of prelipoproteins is attenuated in growth and immune activation. Infect. Immun. 73: 2411–2423. 34. Schmaler, M., N. J. Jann, F. Go¨tz, and R. Landmann. 2010. Staphylococcal lipoproteins and their role in bacterial survival in mice. Int. J. Med. Microbiol. 300: 155–160. 35. Hanna, R. N., L. M. Carlin, H. G. Hubbeling, D. Nackiewicz, A. M. Green, J. A. Punt, F. Geissmann, and C. C. Hedrick. 2011. The transcription factor NR4A1 (Nur77) controls bone marrow differentiation and the survival of Ly6Cmonocytes. Nat. Immunol. 12: 778–785. 36. Li, X. D., and Z. J. Chen. 2012. Sequence specific detection of bacterial 23S ribosomal RNA by TLR13. eLife 1: e00102. 37. Muller, S., A. Faulhaber, C. Sieber, D. Pfeifer, T. Hochberg, M. Gansz, S. D. Deshmukh, S. Dauth, K. Brix, P. Saftig, C. Peters, P. Henneke, and T. Reinheckel. 2014. The endolysosomal cysteine cathepsins L and K are involved in macrophage-mediated clearance of Staphylococcus aureus and the concomitant cytokine induction. FASEB J. 28: 162–175. 38. Casrouge, A., S. Y. Zhang, C. Eidenschenk, E. Jouanguy, A. Puel, K. Yang, A. Alcais, C. Picard, N. Mahfoufi, N. Nicolas, et al. 2006. Herpes simplex virus encephalitis in human UNC-93B deficiency. Science 314: 308–312. 39. Granick, J. L., P. C. Falahee, D. Dahmubed, D. L. Borjesson, L. S. Miller, and S. I. Simon. 2013. Staphylococcus aureus recognition by hematopoietic stem and progenitor cells via TLR2/MyD88/PGE2 stimulates granulopoiesis in wounds. Blood 122: 1770–1778. 40. Rigby, K. M., and F. R. DeLeo. 2012. Neutrophils in innate host defense against Staphylococcus aureus infections. Semin. Immunopathol. 34: 237–259. 41. Cain, D. W., P. B. Snowden, G. D. Sempowski, and G. Kelsoe. 2011. Inflammation triggers emergency granulopoiesis through a density-dependent feedback mechanism. PLoS ONE 6: e19957. 42. Schulz, C., E. Gomez Perdiguero, L. Chorro, H. Szabo-Rogers, N. Cagnard, K. Kierdorf, M. Prinz, B. Wu, S. E. Jacobsen, J. W. Pollard, et al. 2012. A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336: 86–90. 43. Davies, L. C., M. Rosas, S. J. Jenkins, C. T. Liao, M. J. Scurr, F. Brombacher, D. J. Fraser, J. E. Allen, S. A. Jones, and P. R. Taylor. 2013. Distinct bone marrow-derived and tissue-resident macrophage lineages proliferate at key stages during inflammation. Nat. Commun. 4: 1886. 44. Murray, P. J., and T. A. Wynn. 2011. Protective and pathogenic functions of macrophage subsets. Nat. Rev. Immunol. 11: 723–737.