The CD20 homolog Ms4a8a integrates pro- and anti-inflammatory signals in novel M2-like macrophages and is expressed in parasite infection

Authors

  • Astrid Schmieder,

    Corresponding author
    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    • Full Correspondence: Dr. Astrid Schmieder, Department of Dermatology, Venereology and Allergology, University Medical Center Mannheim, University of Heidelberg; Theodor-Kutzer Ufer 1–3, 68167 Mannheim, Germany.

      Fax: +49 621 383 3815

      e-mail: astrid.schmieder@umm.de

    Search for more papers by this author
    • These authors contributed equally to this work.

  • Kai Schledzewski,

    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Julia Michel,

    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    Search for more papers by this author
  • Kathrin Schönhaar,

    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    Search for more papers by this author
  • Yannick Morias,

    1. Myeloid Cell Immunology Laboratory, Brussels, Belgium
    2. Cellular and Molecular Immunology Unit, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Tom Bosschaerts,

    1. Myeloid Cell Immunology Laboratory, Brussels, Belgium
    2. Cellular and Molecular Immunology Unit, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Jan Van den Bossche,

    1. Myeloid Cell Immunology Laboratory, Brussels, Belgium
    2. Cellular and Molecular Immunology Unit, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Pierre Dorny,

    1. Unit of Veterinary Helminthology, Prins Leopold Institute for Tropical Medicine Antwerp, Antwerp, Belgium
    Search for more papers by this author
  • Andrea Sauer,

    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    Search for more papers by this author
  • Carsten Sticht,

    1. Center for Medical Research, Medical Faculty Mannheim, University of Heidelberg, Germany
    Search for more papers by this author
  • Cyrill Géraud,

    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    Search for more papers by this author
  • Zoe Waibler,

    1. Junior Research Group Novel Vaccination Strategies and Early Immune Responses, Paul-Ehrlich-Institut, Langen, Germany
    Search for more papers by this author
  • Alain Beschin,

    1. Myeloid Cell Immunology Laboratory, Brussels, Belgium
    2. Cellular and Molecular Immunology Unit, Vrije Universiteit Brussel, Brussels, Belgium
    Search for more papers by this author
  • Sergij Goerdt

    1. Department of Dermatology, Venereology and Allergology, University Medical Center and Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany
    Search for more papers by this author

Abstract

Recently, we identified the CD20 homolog Ms4a8a as a novel molecule expressed by tumor-associated macrophages that directly enhances tumor growth. Here, we analyzed Ms4a8a+ macrophages in M2-associated infectious pathologies. In late-stage Trypanosoma congolense and Taenia crassiceps infections, Ms4a8a expression was detected in hepatic and peritoneal macrophages respectively. Innate immunity in these infections is modulated by Toll-like receptor (TLR) signaling and TLR2/4/7 agonists strongly induced Ms4a8a expression in bone marrow derived macrophages (BMDMs) treated with M2 mediators (glucocorticoids/IL-4). LPS/dexamethasone/IL-4-induced Ms4a8a+ BMDMs were characterized by strong expression of mRNA of mannose receptor (Mmr), arginase 1, and CD163, and by decreased iNOS expression. Coinduction of Ms4a8a by M2 mediators and TLR agonists involved the classical TLR signaling cascade via activation of MyD88/TRIF and NF-κB. Forced overexpression of Ms4a8a modulated the TLR4 response of RAW264.7 cells as shown by gene expression profiling. Upregulation of Hdc, Tcfec, and Sla was confirmed both in primary LPS/dexamethasone/IL-4-stimulated Ms4a8a+ BMDMs and in peritoneal macrophages from late-stage Taenia crassiceps infection. In conclusion, we show that TLR signaling skews the typical alternative macrophage activation program to induce a special M2-like macrophage subset in vitro that also occurs in immunomodulatory immune reactions in vivo, a process directly involving the CD20 homolog Ms4a8a.

Introduction

Macrophages have an extraordinary ability to rapidly adapt to new environmental stimuli by altering their gene expression profile and functions [1]. These enormous variations enable macrophages to participate in different biological and pathological processes such as the maintenance of tissue homeostasis, orchestration of defense responses, and healing processes in tissues. The phenotypic plasticity and functional heterogeneity of macrophages renders their classification difficult. Thus, a simplified approach was put forward in analogy to the Th1/Th2 dichotomy [2, 3]. Accordingly, M1 or classically activated macrophages differentiate from monocytes under the predominant influence of pro-inflammatory cytokines such as IFN-γ, IL-12, or TNF-α, while M2, or alternative, macrophage activation was primarily described as a response to Th2 cytokines such as IL-4, IL-13, and IL-10, as well as to anti-inflammatory mediators such as glucocorticoids (GCs) [4-7]. As these macrophage phenotypes influence the tissue environment, for example, via secretion of active mediators, a solid characterization of macrophages, especially in pathological processes such as cancer or chronic inflammation, is needed to uncover unique key cellular regulators [8].

In a previous study, the CD20 homolog Ms4a8a was found to be expressed by M2-like tumor-associated macrophages (TAMs) in murine mammary carcinoma and malignant melanoma [9]. For the induction of Ms4a8a expression in these tumor models, simultaneous treatment of bone marrow derived macrophages (BMDMs) in vitro both by the M2 mediators dexamethasone (dexa)/IL-4 and by tumor-conditioned media (TCM) was necessary, resulting in the activation of the glucocorticoid receptor (GR) and of the p38α/β MAPK signaling pathway. These TAM-related Ms4a8a+ macrophages displayed a typical M2 marker profile including mRNA expression of the mannose receptor (Mmr), arginase 1, and CD163, and considerably enhanced the tumor growth rate of subcutaneous transplant tumors when coinjected with TS/A mammary carcinoma cells in mice. Functionally, forced overexpression of Ms4a8a in a macrophage-like cell line induced a special gene signature (Tcfec, Spink5, Sla) involved in immune regulation [9].

Since the establishment of the M1/M2 macrophage concept, parasite infections have been the most frequently used model systems for studying alternatively activated macrophages and their functions in vivo [10, 11]. Although different parasite models and different organs were analyzed, and although in vivo crosstalk at all levels shapes macrophage function and gene regulation, the expression of M2 marker genes such as Mmr, arginase 1, Ym1, Cdh1, or CD163 remained remarkably consistent in most of these models [12-14]. These findings prompted us to study the presence of Ms4a8a in macrophages in parasitic infections, in particular in Trypanosoma congolense and Taenia crassiceps infections in C57BL/6 mouse models, which are characterized by a clear shift from M1 (early stage) to M2 (late stage) predominance [12, 15]. Moreover, as we have shown that the induction of Ms4a8a in BMDMs in vitro required M2 mediators and a second, as yet unidentified molecular signal [9], and as there is ample evidence that innate immunity in Trypanosoma and Taenia infections is modulated by Toll-like receptor (TLR) signaling [16-19], we asked here whether TLR activation in BMDMs provides such a second signal for the induction of Ms4a8a mRNA and protein expression. In addition, we tested whether Ms4a8a might in turn modulate TLR signaling.

Data from these experiments suggest that Ms4a8a is important in integrating pro- and anti-inflammatory stimuli in macrophages and may thus be involved in modulating infectious immune reactions.

Results

Identification of Ms4a8a+ macrophages in models of trypanosomiasis and cysticercosis

Previously, the novel CD20 homolog Ms4a8a was shown to be expressed by M2-like TAM in murine mammary carcinoma and malignant melanoma in vivo. To test the Ms4a8a macrophage expression in other disease models known to induce alternative activated macrophages, C57BL/6 mice were infected with Trypanosoma congolense or Taenia crassiceps intraperitoneally. In the M2-dominated late stage of trypanosomiasis (80 days), Ms4a8a+ macrophages amounted to approximately 30% of the CD68+ macrophage infiltrate and they mainly populated the periportal areas in the liver (Fig. 1A and B). In the M1-dominated early stage of Trypanosoma congolense infection (6 days) and in uninfected animals, Ms4a8a+ macrophages were either absent or detectable only in very limited numbers as compared with the total number of CD68+ macrophages (Fig. 1A). In addition, Ms4a8a expression was detected in peritoneal exudate cells (PEMs) from Taenia crassiceps infected mice (Fig. 1C). These data indicate that Ms4a8a+ macrophages are induced in infections dominated by M2- but not M1-associated inflammation.

Figure 1.

CD68-positive macrophages coexpress Ms4a8a only in late-stage Trypanosoma congolense infected livers and in PEMs from late-stage (16 weeks) cysticercosis. (A) Immunohistochemical stainings of acetone-fixed cryostat sections of control livers, early-stage Trypanosoma congolense infected livers (6 days), and late-stage Trypanosoma congolense infected livers (80 days) with anti-mouse Ms4a8a or anti-mouse CD68 antibody. Positive cells appear in red. Scale bars: 50 μm. (B) Immunofluorescent double staining of acetone-fixed cryostat sections of Trypanosoma congolense infected livers in early (6 days) and late (80 days) stage with anti-mouse Ms4a8a (green) and anti-mouse CD68 antibody (red). Images were acquired using confocal microscopy. Scale bars: 47.62 μm. (A and B) Data are representatives of three independent experiments, each performed with n = 3 mice. (C) Ms4a8a mRNA expression by PEMs isolated 16 weeks after intraperitoneal Taenia crassiceps inoculation or from uninfected mice was analyzed by quantitative real time PCR (qRT-PCR). Data are shown as mean and SEM of n = 3, relative to β-actin and are given as the fold-induction over control PEMs. ***p < 0.001, Student's t-test. Data are representative of two independent experiments, each performed with n = 3 mice.

Synergistic induction of Ms4a8a expression in BMDMs in vitro by TLR agonists and M2 mediators

In our previous work, Ms4a8a was shown to be expressed by BMDMs when M2 mediators such as GC and IL-4 and TCM were combined. As the second signal present in TCM has not yet been defined molecularly and as Trypanosomes and Taenia have been described to modulate the innate immune response by activating the TLR signaling pathway [16-19], different combinations of cytokines were tested in vitro on BMDMs including the TLR4 agonist LPS. Only a combined stimulation with the TLR4 agonist LPS together with dexa/IL-4 was able to significantly induce Ms4a8a mRNA expression in BMDMs (Fig. 2A). In the combined stimulation by LPS/dexa/IL-4, dexa/IL-4 could not be replaced by IL-10 (Fig. 2B).

Figure 2.

Ms4a8a expression is dependent on the synergistic induction by pro-inflammatory TLR-agonists (Lipoteichoic acid (LTA), Imiquimod (IMQ), or LPS) in combination with anti-inflammatory (dexa/IL-4) signaling pathways in BMDMs. (A–E) BMDMs were treated for 72 h as indicated. (A–C) mRNA expression levels of Ms4a8a were assessed by quantitative real time PCR (qRT-PCR) after induction with (A) dexa, IL-4, LPS, or combinations thereof, (B) LPS, dexa, IL-4, IL-10, and combinations thereof, and (C) dexa, IL-4, LTA, IMQ, LPS, and combinations thereof. Data are shown as mean and SEM (n = 3) and are representative of three independent experiments. The p-values were (**p < 0.01; ***p < 0.001) according to ANOVA with Bonferroni algorithm. (D) Western blot analysis (E) and immunohistochemical analysis of Ms4a8a expression in BMDMs treated with TLR agonists in combination with or without dexa/IL-4. Scale bars: 20 μm. Data shown are representative of three independent experiments.

Since in parasitic infections other TLRs such as TLR2 and TLR7 have been described to play an important role in disease progression [20], other TLR agonists such as lipoteichoic acid (LTA, TLR2 agonist) and Imiquimod (IMQ, TLR7 agonist) were tested together with dexa/IL-4. Concomitant stimulation of BMDMs by LTA or IMQ together with dexa/IL-4 resulted in a considerable increase of Ms4a8a mRNA (Fig. 2C) and Ms4a8a protein expression (Fig. 2D and E), but the strongest induction on the protein level as well as on mRNA level remained the combined stimulation with LPS/dexa/IL-4 (Fig. 2C, D, and E). As none of the TLR agonists alone was sufficient to induce Ms4a8a expression in BMDMs on mRNA (Fig. 2C) or protein level (Fig. 2D and E), these results indicate that a complex regulation by pro- and anti-inflammatory mediators is required to induce strong Ms4a8a expression.

Ms4a8a induction is found upon simultaneous TLR and GC signaling and is blocked by Mifepristone

As LPS in combination with dexa/IL-4 induced a strong and significant increase in Ms4a8a mRNA and protein expression in BMDMs, we further investigated the signaling pathway downstream of LPS that classically comprises TLR4, MyD88/TRIF, and NF-κB. Induction of Ms4a8a mRNA and protein expression in BMDMs by LPS/dexa/IL-4 was strongly reduced in TLR4−/− mice in comparison to wild-type mice (Fig. 3A and B). Similar results were found in TRIF−/− and MyD88/TRIF−/− mice. Combined MyD88/TRIF deficiency resulted in an additional decrease in Ms4a8a mRNA and protein expression even in the dexa/IL-4 stimulated group (Fig. 3C and D). Finally, when NF-κB activation was inhibited, LPS-mediated induction of Ms4a8a mRNA and protein expression was reduced to the dexa/IL-4-induced expression level (Fig. 3E and F). In summary, the classical TLR4/MyD88-TRIF/NF-κB signaling pathway is necessary to boost Ms4a8a expression in dexa/IL-4-induced M2 macrophages.

Figure 3.

The LPS-dependent expression of Ms4a8a in BMDMs costimulated with dexa/IL-4 is mediated by the canonical TLR4, Myd88/TRIF, and NF-κB signaling pathway. (A and B) Ms4a8a mRNA and protein expression in BMDMs of TLR 4−/− and wild-type mice stimulated as indicated were analyzed by (A) quantitative real time PCR (qRT-PCR) and (B) by western blotting. Data are shown as mean and SEM of n = 3 and are representative of three independent experiments. The p-value was (**p < 0.01) according to Student's t-test. (C and D) Ms4a8a expression levels of BMDMs derived from wild-type, TRIF−/−, and MyD88/TRIF−/− mice stimulated as indicated were assessed by (C) qRT-PCR and (D) western blot. Data are shown as mean and SEM of n = 3 and are representative of three independent experiments. The p-value was (***p < 0.001) according to ANOVA with Bonferroni algorithm. (E and F) The effect of NF-κB on Ms4a8a expression levels in BMDM stimulated as indicated was analyzed by (E) qRT-PCR and (F) immunohistochemical staining. Scale bars: 20 μm. BMDMs were treated either with control (DMSO 1/2000 diluted in PBS) or 5 μM NF-κB Activation Inhibitor (diluted 1/2000 in PBS from DMSO stock solution, 10 mM). Data are shown as mean and SEM of n = 3 and are representative of three independent experiments. The p-value was (***p < 0.001) according to Student's t-test. (G and H) LPS/dexa/IL-4 induced Ms4a8a expression in BMDMs was analyzed in the presence of increasing concentrations of the GR antagonist Mf. Ms4a8a expression was determined by (G) qRT-PCR and (H) western blot analysis and shown as mean and SEM of n = 3 and are representative of three independent experiments. The p-value was (*p < 0.05) according to ANOVA with Bonferroni algorithm. (A, C, D, and G) Values from qRT-PCR represent the relative induction of Ms4a8a normalized to the housekeeping gene β-actin and are given as the fold-induction over control BMDMs set at 1. Protein loading was controlled by probing western blots with (B) β-actin or (D and H) GAPDH.

The dependency of Ms4a8a gene expression on GR activation was tested with different concentrations of the GR antagonist Mifepristone (Mf). Here, a significant inhibition of LPS/dexa/IL-4-induced Ms4a8a expression was dependent on increasing concentrations of Mf as demonstrated on mRNA (Fig. 3G) and protein levels (Fig. 3H).

As a complex stimulation of several factors is necessary for Ms4a8a gene activation, we performed a promoter analysis using bioinformatics to assess possible transcription factor bindings sites in the promoter region of Ms4a8a (Supporting Information Fig. 1A). There were several GC response elements and one STAT6 binding element, indicating a possible direct transcriptional regulation by GC and IL-4. However, no NF-κB response elements could be detected making direct promoter activation by TLR signaling highly unlikely. In addition, a time course experiment was performed in which LPS was added at different time points (last 6 h, 24 h, 48 h, and 72 h). As significant Ms4a8a induction was observed only after 72 h of LPS stimulation, Ms4a8a cannot be considered as a direct LPS target gene (Supporting Information Fig. 1B).

Ms4a8a+ BMDMs induced by TLR agonists and M2 mediators express M2 markers and cytokines

To assess whether costimulation of BMDMs by a combination of M1 mediators such as LPS and M2 mediators such as dexa/IL-4-induced preferential M1 or M2 macrophage activation, expression of M2 marker genes such as arginase 1, CD163, and the Mmr as well as iNOS, a gene upregulated in M1, was analyzed in LPS/dexa/IL-4-induced BMDMs by quantitative real time PCR (qRT-PCR). Mmr and arginase 1 mRNA expression levels did not show any significant differences (Fig. 4A and C). CD163 mRNA expression, however, was increased in LPS/dexa/IL-4-stimulated BMDMs as compared with control and dexa/IL-4-induced BMDMs (Fig. 4B). In contrast, iNOS expression was reduced in LPS/dexa/IL-4-stimulated BMDMs as compared with LPS-stimulated BMDMs (Fig. 4D).

Figure 4.

LPS dexa/IL-4 treated BMDMs display an M2-like expression profile with enhanced expression of M2 marker genes and reduced secretion of M1 cytokines. (A–D) Expression levels of M2 macrophage differentiation markers (A) Mmr, (B) CD163, (C) ARG1, and (D) the M1 differentiation marker iNOS in BMDMs stimulated as indicated were measured by quantitative real time PCR (qRT-PCR). Data are shown as mean and SEM (n = 3) relative to β-actin and are given as the fold-induction over levels measured for control-induced BMDMs. The p-value was (*p < 0.05) according to ANOVA with Bonferroni algorithm. (E–G) Secretion of (E) IL-10, (F) TNF-α, and (G) IL-1β in culture supernatants of BMDMs stimulated as indicated for 72 h was determined by cytometric bead array. Data are shown as mean and SEM (n = 3) and are representative of three independent experiments. The p-values were (**p < 0.01; ***p < 0.001) according to ANOVA with Bonferroni algorithm.

M2-macrophages have been described to secrete high amounts of IL-10 and low amounts of pro-inflammatory cytokines such as TNF-α and IL-1β. We therefore measured the cytokine secretion profile of BMDMs upon stimulation with LPS, dexa/IL-4, and LPS/dexa/IL-4. Compared with control (66.32 ± 17.61 pg/mL, n = 12), IL-10 showed a significant increase upon dexa/IL-4 (377.60 ± 76.44 pg/mL, n = 12, p < 0.001) and LPS/dexa/IL-4 (324.60 ± 47.06 pg/mL, n = 12, p < 0.001) stimulation. Upon LPS stimulation, an IL-10 concentration of 236.24 ± 27.17 pg/mL, n = 12, (ns) was detected (Fig. 4E). TNF-α was measured at highest concentration upon LPS stimulation (9166.0 ± 795.4 pg/mL, n = 3) and was significantly higher compared with control (43.05 ± 12.83 pg/mL, n = 3, p < 0.001) and dexa/IL-4 (30.11 ± 12.81, n = 3, pg/mL, p < 0.001). Interestingly, values for LPS/dexa/IL-4 stimulation (5400.0 ± 703.8 pg/mL, n = 3, p < 0.001) (Fig. 4F) were significantly higher than upon dexa/IL-4 stimulation and significantly lower than in the LPS only group. Regarding IL-1β (Fig. 4G), the highest level of IL-1β as compared with control (0 ± 0 pg/mL, n = 6) and dexa/IL-4 (19.08 ± 9.19 pg/mL, n = 6, ns) was detected upon LPS stimulation (456.6 ± 99.1 pg/mL, n = 6, p < 0.001) that was almost reduced to baseline upon LPS/dexa/IL-4 stimulation (38.19 ± 6.60 pg/mL, n = 6, p < 0.001). In contrast to TNF-α, no significant increase was observed for IL-1β comparing dexa/IL-4 and LPS dexa/IL-4. These data indicate that there is no M2/M1 shift in dexa/IL-4-induced M2 macrophages upon LPS stimulation in six of seven molecules examined, but rather a trend toward M2 enhancement (four of seven). Vice versa, there was a strong M1/M2 shift in LPS-induced M1 macrophages upon dexa/IL-4 stimulation (six of seven) that reached statistical significance regarding CD163, iNOS, TNF-α, and IL-1β expression. In summary, these results show that LPS/dexa/IL-4 BMDMs do not behave much differently from dexa/IL-4 M2 macrophages regarding typical M2 marker expression and cytokine secretion.

The LPS/Ms4a8a gene signature of RAW264.7 cells is also found in Ms4a8a+ BMDMs and PEMs

As Ms4a8a expression in vitro is subjected to complex regulatory processes integrating pro- and anti-inflammatory signals, we next investigated the functional role played by Ms4a8a in macrophages during TLR4 activation. Using a transgenic model of stably Ms4a8a-transfected RAW264.7 cells (Supporting Information Fig. 1A), gene expression profiling using mouse genome-wide microarray was performed. Four groups of RAW264.7 cells were compared: (i) MMs4a8a, (ii) MEV, (iii) MMs4a8a/LPS, and (4) MEV/LPS (Supporting Information Fig. 2A). Gene expression data for groups 1 and 2 have already been published and are publicly accessible and are used here only for control purposes [9]. Among the genes upregulated by Ms4a8a in combination with LPS, statistically significant coinduction of two genes, that is, Hdc and Tcfec, could be confirmed by qRT-PCR (Supporting Information Fig. 2B and C). Tcfec has already been reported to show basal induction by Ms4a8a overexpression in RAW264.7 cells [9] and was shown here to be super-induced by LPS, while Hdc showed basal induction by LPS and super-induction by Ms4a8a.

In order to test whether genes coinduced by Ms4a8a and LPS in RAW264.7 cells were similarly regulated in Ms4a8a expressing primary macrophages in vitro and in vivo, LPS/dexa/IL-4-induced Ms4a8a+ BMDMs and Ms4a8a+ PEMs from mice suffering from late-stage cysticercosis were analyzed by qRT-PCR. Treatment of BMDMs by LPS alone did not show any effect on expression of either Hdc, Tcfec, or the previously described Ms4a8a-dependent gene Sla [9] when compared with control BMDMs while dexa/IL-4-treated BMDMs showed a slight, nonsignificant increase (Fig. 5A–C). On the contrary, simultaneous stimulation of BMDMs with LPS and dexa/IL-4 strongly induced expression of Hdc, Tcfec, and Sla (Fig. 5A–C) reaching statistical significance in Hdc.

Figure 5.

MS4a8a-associated gene profile assessed in LPS-stimulated Ms4a8a-transgenic RAW264.7 cells is also found in Ms4a8a expressing BMDMs stimulated with LPS dexa/IL-4 and in Ms4a8a-expressing PEMs isolated from mice suffering from late-stage cysticercosis. (A–C) Expression analysis of (A) Hdc, (B) Tcfec, and (C) Sla in BMDMs stimulated as indicated assessed by quantitative real time PCR (qRT-PCR) (n = 3). The p-value was (*p < 0.05) according to ANOVA with Bonferroni algorithm. (D–F) Expression analysis of (D) Hdc, (E) Tcfec, and (F) Sla of PEMs isolated from murine late-stage cysticercosis assessed by qRT-PCR (n = 3). (A–F) Data are shown as means and SEM relative to β-actin and are given as the fold-induction over levels measured for (A–C) control BMDMs or (D–F) control PEMs. The p-values were (*p < 0.05; **p < 0.01) according to Student's t-test. At least three independent experiments were performed.

In line with the data obtained in vitro, Ms4a8a+ PEMs from Taenia crassiceps infected mice showed a tenfold induction of Hdc (Fig. 5D) and Tcfec (Fig. 5E) expression, and a 25-fold induction of Sla gene expression (Fig. 5F) in vivo. Results were statistically significant in these samples for all three genes. Taken together these data indicate that Ms4a8a exerts a regulatory effect on Hdc, Tcfec, and Sla gene expression in macrophages under complex stimulatory immunological conditions.

Discussion

A dynamic microenvironment composed of a mixture of pro- and anti-inflammatory factors influences the development of pathological processes in tumor growth and chronic inflammation by priming innate as well as adaptive immunity [11, 21-23]. Therefore, macrophages are regularly exposed to a multitude of different signals, and conversely, they actively participate in the crosstalk with other cells of the immune system [3]. Among the molecules used by macrophages in these interactive processes, we have previously identified Ms4a8a, a CD20 homolog that is expressed by a special subset of Stabilin-1+, Lyve-1+ tumor-associated M2-like macrophages [9, 24]. When expressed by TAM, Ms4a8a integrates tumor-derived and anti-inflammatory immune signals, thereby enhancing the tumor growth rate and the tumor end weight of subcutaneous transplant cancers in vivo [9]. In this study, we analyzed whether Ms4a8a might also act in integrating pro- and anti-inflammatory signals in other complex immunological microenvironments like infectious diseases.

Infection of mice by Trypanosoma congolense and Taenia crassiceps are suitable animal models for sequential changes in the immunological micromilieu. Both infectious disease models show a similar shift from a M1-dominated immune response in the early stage of infection to a M2-dominated environment in the late stage of infection [12, 15]. Ms4a8a+ macrophages were clearly restricted to the M2-dominated late-stage trypanosomiasis and cysticercosis in C57BL/6 mice. These results support our previous findings that Ms4a8a+ macrophages could play a pathophysiological role in M2-dominated pathological environments. As previously shown, however, M2 mediators were not sufficient to induce Ms4a8a expression in macrophages.

Trypanosomes activate both adaptive and innate immune responses in their hosts. Activation of the innate immune response by Trypanosomes and Taenia is preferentially mediated by activation of pattern-recognition receptors, especially TLRs [16-19]. In this study, we could demonstrate that Ms4a8a expression in BMDMs in vitro was strictly dependent on the activation of TLRs (TLR2, TLR4, TLR7) in addition to the M2 mediators dexa/IL-4. TLR4 activation by LPS exerted its effects on Ms4a8a expression in combination with M2 mediators via the classical TLR signaling cascade including TLR4, MyD88/TRIF, and NF-κB. Although these macrophages were coactivated by pro-inflammatory TLR signals, they displayed a clear-cut M2 macrophage marker profile comprising high arginase 1, Mmr, and CD163, and low iNOS expression. In addition, upon stimulation with LPS/dexa/IL-4, the pro-inflammatory cytokines TNF-α and IL-1β were significantly reduced when compared with that of LPS stimulation of BMDMs.

Analysis of the promoter region of Ms4a8a showed transcription factor binding sites for the GR and IL-4/STAT6, but not for NF-κB. In addition, a time course experiment in which dexa/IL-4 was kept constant and LPS was added for different time points clearly showed that significant Ms4a8a expression in BMDMs stimulated with LPS/dexa/IL-4 was achieved only after 72 h of LPS stimulation indicating an indirect rather than a direct induction by the TLR/MyD88/TRIF/NF-κB signaling pathway. Such indirect TLR-mediated gene regulatory effects causing M2 differentiation may be due to the epigenetic DNA control mechanisms. For example, the TLR-regulated demethylase Jmjd3 is involved in transcriptional regulation of the M2-associated genes Arginase 1 and Mmr in helminth infections [25].

To characterize the function exerted by Ms4a8a upon TLR activation, Ms4a8a expressing macrophage-like RAW 264.7 cells were stimulated with the TLR4 agonist LPS and gene expression profiling was performed. Simultaneous Ms4a8a and LPS stimulation induced several genes, including Hdc and Tcfec associated with M2 activation. As it has been shown recently that other Ms4a family members are capable of transmitting cellular signals by binding to regulatory proteins [26], Ms4a8a could similarly cluster with TLR or with as yet unidentified surface receptors and modulate their activation threshold or signaling behavior, especially in GC-treated macrophages.

To confirm the results obtained by using Ms4a8a+ RAW264.7 transfectants in a more physiological context, we used primary BMDMs and PEMs from Taenia crassiceps infected mice. In BMDMs, Hdc was synergistically induced by M2 mediators dexa/IL-4 in combination with LPS while Tcfec and Sla showed a strong trend for coinduction that did not reach statistical significance. In vivo, all three genes were significantly upregulated in Ms4a8a+ PEMs from Taenia crassiceps infected mice as compared with control PEMs. As Ms4a8a activated these genes in transgenic RAW264.7 cells, we hypothesize that expression of Hdc, Sla, and Tcfec in BMDMs upon coinduction by LPS and dexa/IL-4 as well as in PEMs from Taenia crassiceps infected mice may depend, at least in part, on regulatory effects downstream of Ms4a8a. To finally prove this hypothesis, however, knockdown approaches for Ms4a8a and/or Ms4a8a cross-linking experiments may be required.

The genes upregulated in Ms4a8a+ M2 macrophages have previously been associated with M2 differentiation, T-helper cell skewing or immunosuppression. The transcription factor Tcfec, for example, is important for IL-4-inducible gene expression in macrophages [27] and has been associated with important differentiation processes [28]. The adaptor protein Sla has been reported to suppress T-cell responses by downregulating TCR signaling. In osteoclasts, Sla binds to CSF-1R as an adaptor molecule indicating a role in cellular differentiation [29]. Hdc-deficient mice exhibit a predominantly T(h)1-polarized cytokine pattern. Hdc is the rate-limiting enzyme in histamine synthesis; histamine has been described to activate suppressor T lymphocytes and to inhibit cytotoxic T lymphocytes [30, 31]. In macrophages, histamine inhibits the production of pro-inflammatory cytokines, while enhancing IL-10 secretion [32].

Taken together, we have identified Ms4a8a as a novel protein expressed by a special subset of M2-like macrophages induced in vitro by synergistic activation of BMDMs by pro-inflammatory (TLR signaling, NF-κB) and anti-inflammatory (dexa/IL-4) signaling cascades. These findings fit well with the concept of stepwise activation of macrophages proposed recently by Varin and Gordon [14]. In this concept, M-CSF/GM-CSF-mediated differentiation of hematopoietic stem cells and/or monocytes into macrophages is followed by priming of macrophages with either M1 (IFN-γ) or M2 (IL-4) cytokines. Full functional activation of M1 or M2 macrophages, however, requires a third stimulus delivered by TLR or some other pattern recognition receptors [8]. In applying this concept to our findings, priming of BMDMs would be mediated directly by dexa/IL-4 inducing expression of typical M2 markers. Full activation of M2 macrophages would then be induced by LPS stimulation/TLR signaling mediating Ms4a8a expression and expression of further downstream effector genes of Ms4a8a such as Hdc, Tcfec, and Sla. In this context, Ms4a8a represents a new lead molecule indicating and regulating development of fully activated M2 macrophages. Integrative signaling as demonstrated here for M2 macrophages may play an important role in chronic inflammatory diseases in general and may induce activation of adequate gene expression programs. Proof for a synergy between pro- and anti-inflammatory signaling pathways in macrophage activation will enable us to better understand the multitude of macrophage immune responses during diverse immunological processes and may open new avenues to improved therapeutic strategies targeting macrophages.

Materials and methods

Mice and parasitic models

C57BL/6 and BALB/c wild-type mice were purchased from Elevisier Janvier. TLR4-deficient mice (TLR4−/−) were provided by Shizuo Akira [33, 34]. TRIF-deficient mice (TRIF−/−) were provided by Bruce Beutler [35], MyD88/TRIF double-knockout mice are as described [36, 37]. All knockout mice are backcrossed to C57BL/6 background and housed under specific pathogen-free conditions at the animal facility Mannheim. To induce trypanosomiasias, C57BL/6 mice were inoculated intraperitoneally with 2 × 103 Trypanosoma congolense Tc13 according to the published procedures [38, 39]. Animal experimental protocols were approved by the animal ethics committee (Regierungspräsidium Karlsruhe, Az: 35–9185.81/G-115/07). To study helminth infection, C57BL/6 mice were inoculated intraperitoneally with ten Taenia crassiceps metacestodes; peritoneal cells were collected after 16 weeks (late stage) and macrophages obtained as described before [40].

Materials

The following reagents were used: M-CSF, IL-4, IL-10 (Peprotech), dexa (Sigma), LPS, LTA, and IMQ (Invitrogen); anti CD68 (Acris), rat-anti mouse β-actin, rabbit-anti mouse GAPDH, goat-anti rabbit Alexa488, goat-anti rat Cy3, goat-anti-rabbit IgG-HRP (Santa Cruz Biotechnology). Custom-made rabbit-anti mouse Ms4a8a serum was produced as previously described [9].

Cells

RAW264.7 cells (ATCC CRL-6323) were obtained from American Type Culture Collection. Ms4a8a-transfected RAW264.7 cells and control clones were as described [9]. Authentication of cell lines was assured by regular morphology checks and growth curve analyses. Cells were regularly tested for mycoplasma infection by PCR. BMDMs were generated as previously described [9] and stimulated with IL-4 (10 ng/mL), IL-10 (10 ng/mL), dexa (5 × 10−7 M), LPS (1 μg/mL), IMQ (10 μg/mL), LTA (1 μg/mL) for 72 h if not otherwise indicated. For the signaling pathway analysis, the following inhibitors were used: NF-κB Activation Inhibitor for NF-κB (Calbiochem) at 5 μM and 1, 10, or 100 nM of the GR antagonist Mf (Sigma).

Immunohistochemistry

Cells grown on glass cover slips and cryostat liver sections were air-dried and acetone-fixed. Specimens were blocked with 5% FCS in PBS and incubated with the first antibody. The appropriate HRP-labeled secondary antibody was used. Pictures were taken with a Leica DCRE microscope, Leica DC500 camera, and software system (Leica). Images were arranged using Adobe Photoshop 6.0 software.

Immunofluorescence

Acetone-fixed cryostat sections were blocked with 3% BSA, incubated with the first antibody, and after that with the appropriate secondary antibody. Specimens were analyzed with Leica TCS SP2 laser scanning spectral confocal microscope.

Western blot analysis

Proteins were obtained by treating whole cells with RIPA-P buffer (150 mM NaCl, 15 NP40, 0.5% sodium desoxycholate, 0.1% SDS, 50 mM Tris/HCL [pH 8.0], 10 mM NaF, 1 mM Na3OV4, 2 mM EDTA [all Sigma]) enriched with a protease inhibitor cocktail (Roche). Equal amounts of protein were carried on a 12% SDS polyacrylamide gel. After blotting onto nitrocellulose (Schleicher&Schüll), membranes were incubated overnight with rabbit-anti mouse Ms4a8a, rat-anti mouse β-actin, or rabbit-anti mouse GAPDH antibody, respectively, followed by incubation with the secondary anti-rabbit HRP-labeled antibody. For signal detection, ECL PLUS Western Blotting Detection System (GE Healthcare) was used.

Analysis of cytokine concentration in cell culture supernatants of BMDMs

Culture supernatants of BMDMs were collected after 72 h stimulation, snap frozen, and stored at –20°C. For quantification of TNF-α, IL-1β, and IL-10, the kits (mIL-1β FlowCytomix Simplex [BMS86002FF], mIL-10 FlowCytomix Simplex [BMS8614/2FF], and mTNFα FlowCytomix Simplex [BMS8607/2FF, all ebioscience]) were used according to the manufacturer's instructions. Experimental data were quantified by BD FACSCanto II.

RT-PCR and qRT-PCR analysis

For RNA extraction, the Total RNA Kit I (Omega bio-tek) was used according to the manufacturer's instructions. For cDNA synthesis, 1 μg RNA was used for reverse transcription with RevertAid H Minus M-MuLV Reverse Transcriptase (Fermentas) using Oligo (dT)18 Primer following the manufacturer's instructions. For the RNA extraction of PEMs from Taenia-infected mice, TRIzol was used followed by reverse transcription with Superscript II (Invitrogen).

For qRT-PCR 1 μL of cDNA was amplified using SyBRGreen PCR Master Mix (Applied Biosystems) under standard conditions with an MX3000P sequence detection system (Stratagene). All Primers used are listed in Supporting Information Table 1.

cDNA microarray analysis and statistical procedures

For microarray analysis, Ms4a8a-transfected (K8) and control DNA-transfected RAW264.7 (K3) cells were incubated (or not incubated) for 6 h with 5 μg/mL LPS in DMEM medium supplemented with 10% FCS, 100 IU penicillin, 100 μg/mL streptomycin, 1% pyruvic acid, 1% nonessential amino acids. Gene expression profiling was performed using mouse genome 430 2.0 DNA arrays (Affymetrix), according to the recommendations of the manufacturer. A Custom CDF Version 14 with Entrez-based gene definitions was used to annotate the arrays. Differential gene expression was analyzed based on loglinear mixed model ANOVA [41] using a commercial software package SAS JMP7 Genomics, version 4.0, from SAS (SAS Institute). A false-positive rate of a = 0.05 with Holm correction was taken as the level of significance. Full data are deposited in the Gene Expression Omnibus database.

Promoter analysis

The murine Ms4a8a promoter was defined as 2-kb sequence upstream of the translation start codon region and identified on mouse chromosome 19, position 11157593-11155594 (ENSEMBL database). The selected sequence was analyzed for transcription factor binding sites with MATINSPECTOR 8.0.5 software package (Genomatix). By using a general core promoter element (threshold 0.75) and vertebrate promoter element (threshold 0.75) matrix, a total of 596 consensus sites have been identified and displayed.

Statistical analysis

GraphPad Prism software was used to perform Student's t-test or one-way ANOVA using Bonferroni algorithm. A p-value < 0.05 was regarded as statistically significant. Level of significance was indicated by asterisks (***p < 0.001; **p < 0.01; and *p < 0.05). Error bars show SEM, if not other indicated, of each experiment. Experiments were performed at least in triplicate.

ACKNOWLEDGEMENTS

This work was supported in part by grants of Deutsche Forschungsgemeinschaft SFB405, project B12 to S.G.; SFB-TR23, project B1 to S.G.; and SFB938, project H to S.G.

Conflict of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations
BMDM

bone marrow derived macrophage

dexa

dexa-methasone

GC

glucocorticoid

GR

glucocorticoid receptor

IMQ

imiquimod

LTA

Lipoteichoic acid

Mf

Mifepristone

Mmr

mannose receptor

MyD88/TRIF−/−

MyD88/TRIF knockout mouse

PEM

peritoneal exudate cell

qRT-PCR

quantitative real time PCR

TAM

tumor-associated macrophage

TCM

tumor-conditioned medium

TLR4−/−

Toll-like receptor 4 knockout mouse

TRIF−/−

TRIF knock out mouse

Ancillary