Background
Microglia represent the resident immune cells of the central nervous system (CNS) and account for approximately 12% of all cells in the brain
[
1]. As counterparts of peripheral macrophages, microglia sense the brain parenchyma for perturbations resulting from injury or pathological conditions. Several CNS neurodegenerative pathologies including Alzheimer’s disease (AD)
[
2‐
4], multiple sclerosis (MS)
[
5,
6] and Parkinson’s disease (PD)
[
7,
8] are characterised by a strong microglia reaction that is, at least partially, responsible for the progressive nature of these diseases.
Increasingly, studies have demonstrated that microglia, like peripheral macrophages, exhibit two entirely functional different activation states that are referred to as classical and alternative activation. The classical activation of microglia (M1) is induced by Th1 cytokines, such as IFNγ, IL1β, IL12, and IL6 as well as lipopolysaccharide (LPS) and results in production and release of pro-inflammatory cytokines such as tumour necrosis factor-alpha (TNFα), IL6, matrix metalloproteinase (MMP)-9, nitric oxide (NO) and reactive oxygen/nitrogen species (ROS)
[
9‐
11] which are involved in inflammation-mediated neurotoxicity
[
12‐
14]. Whereas alternative activation of microglia (M2) initiated by Th2 cytokines, such as IL4 and IL13, as well as IL10 and TGFβ, results in up regulation of arginase-1 (Arg1), Chitinase 3 like 3 (Ym1) and found in inflammatory zone-1 (Fizz1), which are primarily associated with tissue repair and extracellular matrix composition
[
3,
11]. As hypothesised by Town
et al. these differently activated microglia likely exist as a dynamic continuum
in vivo, with functions ranging from deleterious to beneficial
[
15,
16]. All these notions suggest that modulating microglia activation states might be a potential therapeutic approach to different types of neurodegenerative diseases including AD, MS, and PD.
Based on the M1 and M2 activation states, a more detailed categorization of macrophage activation states has recently been discussed. As suggested by Gordon and colleagues
[
11,
17,
18], alternative activation is only limited to macrophages treated with IL4 or IL13 and is primarily associated with injury resolution including tissue repair and extracellular matrix reconstruction. While IL10 and TGFβ promote a macrophage phenotype characterised by inflammation resolution including inhibition of pro-inflammatory cytokine production, modification of inflammatory signalling pathways and increased expression of scavenger receptors, thereby, promoting debris clearance. This type of activated macrophage has been termed acquired deactivation
[
19‐
23]. Whereas IL10 and TGFβ induce acquired deactivation, the acquired deactivation macrophages can further produce IL10 and TGFβ in an autocrine manner
[
3,
24,
25]. Regarding the immunoregulatory function of IL10 and TGFβ produced by the acquired deactivation macrophages, this phenotype of macrophages has been also named as a regulatory macrophage by Mosser and Edwards
[
24]. Although these studies promoted the knowledge of microglia/macrophage activation phenotypes and their distinct functions, there is still no general agreement in the field on the nomenclature and, more importantly, the interaction between different activation states of microglia/macrophages is still poorly understood.
TGFβ is a multifunctional cytokine involved in a variety of physiological and pathological conditions
[
26]. TGFβs bind to the TGFβ receptor type II, which recruits and phosphorylates a type I receptor. The type I receptor then phosphorylates Smad2/3, which further bind to Smad4 to form a heteromeric complex that translocates into the nucleus to regulate the expression of target genes
[
27,
28]. Next to the canonical Smad-dependent pathway, TGFβs also signal via Smad-independent signalling cascades, including mitogen-activated protein kinase signalling (MAPK) pathways
[
29].
In this study we used the microglial cell line BV2 and primary microglia to investigate the role of TGFβ in IL4-induced alternative activation, thereby illustrating the interaction between different microglia/macrophage activation states. For the first time, we provide evidence that although TGFβ1 treatment alone is not able to induce microglia, alternative activation, treated together with IL-4, strongly enhances IL4-induced alternative microglia activation. Arg1 and Ym1 expression was significantly increased after co-treatment with IL4 and TGFβ1. To our surprise, Arg1 and Ym1 expression induced by IL4 treatment alone was significantly impaired in the presence of the TGFβ receptor type I inhibitor. Further investigation revealed that IL4 treatment alone increased microglial TGFβ2 expression and secretion, which in turn might promote IL4-induced Arg1 and Ym1 expression. Moreover, we found TGFβ1 treatment resulted in up regulation of the IL4 receptor alpha (IL4Rα). Finally, we provide evidence that the Mitogen-activated protein kinase (MAPK) pathway is essential for TGFβ-mediated enhancement of Arg1 expression after IL4 treatment in microglia.
Methods
Cytokines and reagents
All reagents for cell culture, namely Trypsin-EDTA 1×, Hank’s balanced salt solution (BSS), Dubecco’s modified Eagle medium (DMEM)-Ham’s F12, penicillin/streptomycin (P/S) 100×, and fetal calf serum (FCS) were purchased from PAA Laboratories (Cölbe, Germany). MAPK/ERK (MEK) inhibitor PD98059 and poly-D-lysine were purchased from Sigma-Aldrich (Deisenhofen, Germany). Recombinant murine IL4 and recombinant human TGFβ1 were purchased from PeproTech (Hamburg, Germany). TGFβ receptor type I kinase inhibitor (TβKI) was obtained from Merck Chemicals (Darmstadt, Germany). Primary antibodies: anti-Arginase1, anti-IL4Rα, anti-TGFβ2 and anti-Smad1/2/3 were purchased from SantaCruz (Heidelberg, Germany). Phospho-Smad2-Ser465/467 and phospho-Stat6-Tyr641 were obtained from New England Biolabs (Frankfurt, Germany), Ym1 antibody was from StemCell Technologies (Grenoble, France). GAPDH was purchased from Abcam (Cambridge, UK). Goat anti-mouse Cy3, goat anti-rabbit Cy3 were from Dianova (Hamburg, Germany).
BV2 cell culture
The murine microglia cell line BV2 was maintained in DMEM/F12 (PAA) supplemented with 10% heat-inactivated FCS and 1% P/S. Cultures were kept at 37°C in 5% CO2/95% humidified air atmosphere. Prior to treatment cells were washed with PBS and serum-free medium was added.
Primary microglia cultures
Whole brains obtained from P0/1 C57BL/6 mice were washed twice with Hank’s BSS solution and vessels and meninges were removed from brain surfaces under the microscope. Cleaned brains were collected and enzymatically dissociated with Trypsin-EDTA (1×) for 15 minutes at 37°C. An equal amount of ice-cold FCS, together with DNase I (Roche Diagnostics, Mannheim, Germany) at a final concentration of 0.5 mg/ml was added prior to dissociation with wide- and narrow-bored polished Pasteur pipettes. Cells were then washed and single cells were centrifuged, collected and suspended with (DMEM)-Ham’s F12 medium containing 10% fetal bovine serum (FBS) and 1% Penicillin/Streptomycin. Cell suspensions were transferred to poly-D-lysine-coated tissue culture flasks with a density of 2 brains/75 cm2 flask. Cultures were maintained in a humidified 5% CO2/95% air atmosphere at 37°C. At day in vitro (DIV) 2 and 3, cultures were washed twice with pre-warmed phosphate- buffered saline (PBS) and fresh culture medium was added. After 10 to 14 days in culture, microglia were shaken off from adhesive grown astroglia by shaking at approximately 250 to 300 rpm for 1 hour. Isolated microglia were plated into various dishes or plates and treated with proper factors, according to different experimental purposes.
Immunocytochemistry
Microglia were plated on glass coverslips and were fixed after treatment with 4% paraformaldehyde (PFA) for 15 minutes at room temperature. After blocking with PBS containing 10% normal goat serum and 0.1% TritonX-100 (Roche, Mannheim, Germany) for 1 hour at room temperature, cells were incubated with primary antibodies at 4°C overnight, followed by an incubation with corresponding Cy3-conjugated secondary antibodies (goat anti-mouse Cy3 1:100, goat anti-rabbit Cy3 1:100). Nuclei were counterstained using 4′,6-diamidino-2-phenylindole (DAPI, Roche). Phase contrast and fluorescence images were captured using the Leica AF6000 imaging system (LEICA, Wetzlar, Germany).
Protein isolation and western blotting
Total proteins were isolated from primary microglia and BV2 cells after washing with PBS and incubation with ice-cold mammalian protein extraction reagent (M-PER, Thermo Scientific, Bonn, Germany) plus Complete Protease Inhibitor (Roche) with gentle up and down shaking for 5 minutes. The supernatant as well as debris were collected, and centrifuged at 14,000 rpm for 8 minutes to obtain the supernatant, which contains proteins. After determination of protein concentrations, 10–15 μg total proteins were loaded onto 10 to 12% SDS gels. After electrophoresis, proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Immobilon, Milipore, Schwalbach, Germany). Blots were blocked with 5% non-fat dry milk in Tris-buffered saline containing 0.05% Tween-20 (TBST) for 1 h and incubated with primary antibodies overnight at 4°C. Primary antibodies against Arginase-1 (rabbit polyclonal, 1:1000, SantaCruz), phospho-Smad2 (Ser465/467) (rabbit, 1:1000, Cell Signaling), phospho-Stat6 (Tyr641) (rabbit, 1:1000, Cell Signaling), IL4Rα (mouse, 1:1000, SantaCruz), TGFβ2 (rabbit polyclonal, 1:1000, SantaCruz), Ym1 (rabbit polyclonal, 1:1000, StemCell Technologies) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (mouse monoclonal, 1:10,000, Abcam, Cambridge, UK) were used. After incubation with goat anti-rabbit or goat anti-mouse IgG horseradish peroxidase (HRP)-linked antibodies (1:10,000, Cell Signaling), blots were developed using Western Lightning® Plus-ECL, Enhanced Chemiluminescence Substrate (Perkin-Elmer, Rodgau, Germany). Signals were captured on Amersham Hyperfilm™ECL (GE Healthcare, München, Germany). Band intensities were evaluated using the software FlourChem 8800 (Alpha Innotech, Biozym, Olendorf, Germany).
RNA isolation and quantitative RT-PCR
RNA was isolated from BV2 and primary microglial cells with the RNeasy kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. RNA was reverse transcribed to cDNA with the GeneAmp RNA PCR Core Kit (Applied Biosystems, Darmstadt, Germany). Quantitative RT-PCR (qRT-PCR) analysis was performed with the MyiQ™ (BIO-RAD, München, Germany) and the Quantitect SYBR Green PCR Kit (Applied Biosystems) with 1 μl of cDNA template in a 25 μl reaction mixture. Results were analysed with the Bio-Rad iQ5 Opitcal System Software and the comparative CT method. Data are expressed as 2-ΔΔCT for the experimental gene of interest normalized to the housekeeping gene (GAPDH) and presented as fold change relative to control. The following primers were used: TGFβ1for: 5′-TAATGGTGGACCGCAACAACG-3; TGFβ1rev: 5′-TCCCGAATGTCTGACGTATTGAAG-3 [NM_011577.1, NCBI]; TGFβ2for: 5′-AGAATCGTCCGCTTTGATGTCTC-3′, TGFβ2rev: 5′-ATACAGTTCAATCCGCTGCTCG-3′ [NM_009367.3, NCBI]; TGFβ3for: 5′-GCCCTGGACACCAATTACTGC-3; TGFβ3rev: 5′-CCTTAGGTTCGTGGACCCATTTC-3´ [NM_009368.3, NCBI]; Arg1for: 5′-TCATGGAAGTGAACCCAACTCTTG-3′, Arg1rev: 5′-TCAGTCCCTGGCTTATGGTTACC-3′ [NM_007482.3, NCBI]; Ym1for: 5′-AGACTTGCGTGACTATGAAGCATTG-3′; Ym1rev: 5′-GCAGGTCCAAACTTCCATCCTC-3′ [NM_009892.2, NCBI]; IL4for: 5′-ATTTTGAACGAGGTCACAGGAGAAG-3′; IL4rev: 5′-ACCTTGGAAGCCCTACAGACGAG-3′ [NM_021283.2, NCBI]; IL4Rαfor: 5′-GAACTCAGACCCACCCAAAAGC-3′; IL4Rαrev: 5′-AAGTGGCAAGTGAGGGACGAG-3′ [NM_001008700.3, NCBI]; Gapdhfor: 5′-ATGACTCTACCCACGGCAAG-3′; Gapdhrev: 5′-GATCTCGCTCCTGGAAGATG-3′ [NM_008084.2, NCBI].
Characterisation of TGFβ secretion
Primary microglia were treated with or without IL4 (10 ng/ml) in serum-free DMEM-Ham’s F12 medium for 24 hours. Conditioned medium was collected for the mink lung epithelial cell (MLEC) assay and ELISA. The MLEC assay is widely used to measure the amount of TGFβs in conditioned medium. The principal is that the MLECs containing a luciferase reporter under the control of a TGFβ-responsive truncated plasminogen activator inhibitor (PAI)-promoter are able to generate luciferase with a TGFβs dose-dependent manner. Since the MLECs only response to the activated TGFβs, in order to detect the latent part of TGFβs, the conditioned medium has to be acidification first to convert latent TGFβs into activated ones. To evaluate the levels of released TGFβs after IL4 treatment, the MLEC assay was performed as described by Abe
et al.[
30]. Briefly, MLECs were placed into 96-well plates at the density of 1.5 × 10
4 cells per well and treated with collected conditioned medium either with or without acidification with 1 M HCL and pH adjustment with NaOH (to activate latent TGFβs) as well as the standard mediums containing different contractions of recombinant TGFβ for 16 hours. Cells were washed with PBS and total proteins were extracted using lysis buffer (Tropix, Applied Biosystems). The luciferase activity was analysed in duplicates using a luminometer (LumatB5076, Berthold, Bad Wildbad, Germany).
Direct enzyme-linked immunosorbent assay for TGFβ2 detection
Conditioned media collected from primary microglia after treatment with and without IL4 (10 ng/ml), as well as diluted recombinant human TGFβ2 (2000, 1000, 500, 250, 125, 62.5 and 31.25 pg, R&D Systems, Wiesbaden-Nordenstedt, Germany) standard solutions, were added into ELISA plates (NUNC, Wiesbaden, Germany) and incubated overnight at 4°C. After washing with washing buffer (0.05% Tween-20 in PBS) and blocking with 1% BSA in PBS for 2 hours at 37°C, the plates were incubated with anti-TGFβ2 (Santa Cruz) primary antibodies overnight at 4°C. Followed by washing and incubation with Biotin-linked anti-rabbit secondary antibodies for 2 hours at 37°C, plates were incubated with ABC-solution (Vectashield, BIOZOL, Eching, Germany). Colour reaction was performed using 2, 2′-azino-bis (3-ethylbenzothiazoline-6-sulphonic acid) substrate (ABTS, Sigma-Aldrich, Deisenhofen, Germany) for 30 minutes in the dark. Finally, absorbance was detected using an FC Multiskan plate reader (Thermo Fischer, Bonn, Germany) at the absorption of 405 nm.
Cytokine array
For the analysis of IL4 release, supernatant from untreated and TGFβ1-treated primary microglia was analysed using the Proteome Profiler™ Array Mouse Cytokine Array Panel A (R&D Systems, Wiesbaden-Nordenstedt, Germany) according to the manufacturer’s instructions. Briefly, equal amounts of primary miroglia cells were incubated for 24 hours in the presence or absence of TGFβ1 and media were collected. Cytokine array membranes were incubated with cell culture supernatants at 4°C overnight with gentle shaking. Membrane signals were developed using Western Lightning® Plus-ECL, Enhanced Chemiluminescence Substrate (Perkin-Elmer, Germany) and signals were captured on Amersham Hyperfilm™ ECL (GE Healthcare).
Statistical procedures
The data were expressed as means ± standard error (SE). Statistical significance between multiple groups was compared by one-way analysis of variance (ANOVA) followed by an appropriate multiple comparison test. Two-group analysis was performed using the Student’s t-test. P-Values < 0.05 were considered statistically significant. All statistical analyses were performed using GraphPad Prism4 software (GraphPad Software Inc.).
Discussion
In this study we demonstrate for the first time that TGFβ enhances the IL4-induced alternative activation of microglia. Using Arg1 and Ym1 as established markers for alternative activation
[
3,
11] we provide evidence that IL4-mediated up-regulation of
Arg1 and
Ym1 is significantly enhanced in the presence of TGFβ1. Further, IL4 treatment resulted in increased expression and secretion of TGFβ2, whereas TGFβ treatment of microglia increased the expression of the IL4Rα. Moreover, blocking the TGF-β receptor type I resulted in significantly impaired Arg1 and Ym1 up-regulation after IL4 treatment. Finally, we demonstrate that TGFβ-mediated enhancement of Arg1 expression in microglia is dependent on the MAP kinase pathway.
In parallel to transcriptional regulation of microglia markers, the morphological changes are used to discriminate between different activation states
in vivo and
in vitro. In the resting or inactive state, microglia present a ramified morphology with several processes, while stimulation with classical activation factors such as LPS or IFNγ results in retraction of microglial processes and development of an amoeboid phenotype
[
33,
34]. Although changes in morphology also suggest changes in the functional states of microglia, the morphology alone cannot be used to predict a functional outcome. Therefore, we analysed Arg1 and Ym1 as markers for macrophage and microglia alternative activation. Arg1 has been shown to be localised in the cytoplasm of hepatocytes where it is involved in nitrogen elimination by catalysing arginine hydrolysis to urea and ornithine
[
11,
35]. Unlike the constitutively expressed Arg1 in the liver, Arg1 in macrophages and microglia is induced by exogenous stimuli including the Th2 cytokines IL4 and IL13
[
36,
37]. Arg1 inhibits NO production by competing with the inducible nitric oxide synthase (iNOS) for the common substrate L-arginine
[
38]. On the other hand, the production of ornithine can be used to generate polyamines, glutamate, and proline, the latter being a substrate for the formation of extracellular matrix proteins such as collagen
[
38‐
40]. Interestingly, apart from involvement in the regulation of wound healing and fibrosis
[
41,
42], Arg1 can directly support neuron survival
[
43]. Next to Arg1, Ym1 is another established marker for microglia alternative activation
[
2,
44]. Ym1 is a heparin/heparin sulphate-binding lectin that is transiently expressed during inflammation
[
44] and although the precise functions of Ym1 remain elusive, recent reports have suggested an involvement in tissue remodelling and regulation of inflammation
[
45,
46].
TGFβ has been shown to either up-regulate Arg1 expression or increase the Arg1 enzymatic activity in a cell type-dependent manner. Whereas TGFβ treatment results in increased expression of Arg1 in fibroblasts and epithelial cells
[
47‐
49], TGFβ strongly increases enzyme activity in macrophages
[
50,
51]. In this study we demonstrate that TGFβ1 alone is not able to significantly increase expression of Arg1 and Ym1 in microglia. However, in the presence of IL4, TGFβ1 significantly enhanced IL4-induced Arg1 and Ym1 expression, which indicates the potential role of TGFβ in CNS tissue repair and neurorestoration by modulating alternative activation of microglia.
To understand the mechanisms behind this phenomenon we addressed the question of whether TGFβ interferes with the IL4 signalling pathway. We demonstrated that TGFβ1 treatment alone up regulated the common receptor for IL4 and IL13, IL4Rα, both at mRNA and protein levels in a time-dependent manner. Opposite to the effect of TGFβ1 on IL4Rα expression, IL4 treatment alone reduced IL4Rα expression with treatment time (data not shown). Further, we observed that TGFβ1 was able to enhance the expression of Arg1 induced by IL13 (data not shown) and IL4. These data indicate that the synergistic effect of TGFβ1 and IL4 on the expression of Arg1 might partially be mediated by enhanced IL4Rα expression after TGFβ1 treatment, thereby increasing the sensitivity of microglia for IL4. IL4 signalling is further propagated by phosphorylation of the transcription factor Stat6
[
11,
31]. Analysis of Stat6 phosphorylation revealed that TGFβ1 failed to induce Stat6 phosphorylation. Moreover, IL4 was not able to induce Smad2 phosphorylation in microglia, indicating that the synergistic effect of TGFβ1 and IL4 on the expression of Arg1 could not be explained by the direct interaction of the TGFβ/Smad and IL4/Stat6 signalling pathways. Therefore, we further investigated if a TGFβ-induced Smad-independent pathway, the MAPK pathway, is involved in this synergistic effect. By performing a pharmaceutical blockage of the TGFβ-induced MAPK pathway using the MEK1/2 inhibitor PD98059, we could show that the Arg1 protein expression, not only induced by TGFβ1 and IL4 co-treatment but also induced by IL4 treatment alone, were significantly inhibited in the presence of PD98059. These data clearly demonstrate that TGFβ signalling is involved in IL4-induced microglia alternative activation and an essential role of TGFβ-mediated MAPK pathway in the enhancement of IL-4 induced microglia alternative activation by TGFβ signalling.
We observed that IL4 treatment of microglia lead to up regulation of TGFβ2, whereas the mRNA levels of TGFβ1 and TGFβ3 were not changed after IL4 treatment. TGFβ2 levels were significantly increased in the supernatants of IL4-treated microglia. Although most of the secreted TGFβ2 was in a latent and inactive form, a small proportion of bioactive TGFβ2 seems to be sufficient to support IL4-induced up regulation of Arg1 and Ym1. However, microglia express several factors and enzymes that are capable of activating latent TGFβs, such as integrins, plasminogen, MMP2 and thrombospondin-1 (unpublished data). Moreover, microglia also express extracellular matrix components in vitro that might bind TGFβ. This amount of bound, and probably activated, TGFβ will escape all analyses of the microglial supernatant, but is likely to activate TGFβ signalling in these cells.
It is widely accepted that TGFβ is involved in the down regulation of microglia classical activation. TGFβ1 reduces reactive oxygen species (ROS) induced by LPS and suppresses the IFNγ-induced expression of MHC II and the production of cytokines, IL1, IL6, and TNF-alpha production in activated microglia
[
52,
53]. TGFβ also prevents IL1β-induced microglial activation
[
54]. Although the anti-inflammatory role of TGFβ has been widely accepted, it is still quite ambiguous whether this effect is beneficial or detrimental in terms of different CNS diseases. Whereas TGFβ1 has protective and beneficial functions in cerebral ischaemia
[
55], it promotes the deposition of amyloid-beta plaques in models of Alzheimer’s disease
[
56]. Interestingly, Town and colleagues have demonstrated that blocking of TGFβ/Smad signalling almost completely abrogated the plaque formation in transgenic mice overexpressing mutant human amyloid precursor protein
[
57]. These results underline the importance of a tight temporal and spatial regulation of innate immune responses and further demonstrate the necessity to enhance our knowledge of the pathological conditions under which TGFβ-mediated regulation of inflammation is beneficial or detrimental.
Whereas TGFβ induces acquired deactivation, the acquired deactivation macrophages also produce TGFβ in an autocrine manner
[
3,
24,
25]. Next to down regulating the classical activation of microglia, here we show, for the first time, that the TGFβ also enhances IL4-induced microglia alternative activation
in vitro, which broadens the knowledge of interactions among different microglia activation states. Similar functions have been shown for another immunoregulatory cytokine, IL10. For example, IL10 is able to impair IFNγ-induced macrophage classical activation
[
58], increase arginase activities
[
59], and further enhance IL4-induced Arg1 expression, probably by increasing IL4Rα expression
[
60]. Findings of this work and previous studies suggest an interaction and dynamic change between different microglia activation states. TGFβ might serve as a gatekeeper to inhibit classical activation and promote alternative activation of microglia. The data presented throughout this study confirm the role of TGFβ as an anti-inflammatory molecule and broaden its functions as an enhancer of microglia alternative activation, thereby regulating microglia-mediated neuroregeneration and neurorestoration in inflammatory CNS diseases.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
XZ conceived of the study idea. XZ carried out all the experiments. BS participated in the experiments. XZ, BS and KK were involved in conception and design of the study as well as in the manuscript preparation. All authors have read and approved the final manuscript.