Address correspondence and reprint requests to Gunnar Schulte, PhD, Department Physiology & Pharmacology, Receptor Biology & Signaling, Karolinska Institutet, S-17177 Stockholm, Sweden. E-mail: firstname.lastname@example.org
Surveying microglia, the resident macrophage-like cells in the central nervous system, continuously screen their surroundings to sense imbalance in tissue homeostasis. Their activity is tightly regulated in both a pro- and anti-inflammatory manner. We have previously shown that the lipoglycoproteins WNT-3A and WNT-5A drive pro-inflammatory transformation in primary mouse microglia cells, arguing that WNTs have a role in the modulation of the central nervous system immune response. In this study, we address the effects of recombinant WNT-3A and WNT-5A on lipopolysaccharide (LPS)-activated mouse primary microglia to investigate the putative anti-inflammatory modulation of microglia by WNTs. While both WNT-3A and WNT-5A alone induce an up-regulation of cyclooxygenase 2 (COX2), a generic pro-inflammatory microglia marker, LPS exceeds these effects dramatically. However, combination of LPS and WNTs results in a dose-dependent decrease in LPS-induced cyclooxygenase 2 protein and mRNA expression. In conclusion, our data suggest that WNTs have a dual and context-dependent effect on microglia acting in a homeostatic pro- and anti-inflammatory manner.
Microglia, the immunocompetent cells of the central nervous system constantly screen for any homeostatic change in their surroundings (Hanisch and Kettenmann 2007). Upon chemical or structural imbalance, microglia transform into a more active state i.e. the cells change morphology, proliferate, migrate, produce reactive oxygen species and nitric oxide, secrete cytokines and increase expression of pro-inflammatory genes, such as cyclooxygenase-2 (COX2) (Hanisch and Kettenmann 2007; Pocock and Kettenmann 2007; Kettenmann et al. 2011). Many factors modulate microglia activity ranging from protons, neurotransmitters, reactive oxygen species, cytokines, chemokines, ATP, and pathogens to name but a few (Kettenmann et al. 2011). Cytokines, such as tumor necrosis factor α (TNFα), are known to act in a dual manner both acting cytotoxic/pro-inflammatory on surveying microglia as well as protective/anti-inflammatory on pre-activated microglia to maintain and promote tissue homeostasis (Sriram and O'Callaghan 2007).
Wingless/int1 (WNTs), a family of secreted lipoglycoproteins exert their effects through receptors of the Class Frizzled (FZD) and play important roles in neuronal development, synaptogenesis and plasticity, stem cell maintenance, neurodegenerative disease, and neuroinflammation (Inestrosa and Arenas 2010; Clevers and Nusse 2012; Salinas 2012; Marchetti and Pluchino 2013). In previous studies, we have shown that mouse primary microglia cells express a set of FZDs and additional co-receptors rendering microglia responsive to recombinant purified WNT-3A and WNT-5A. Stimulation of microglia cells with WNT-3A or WNT-5A results in differential pro-inflammatory transformation, assessed by expression of pro-inflammatory cytokines, chemokines and other immune-response genes, proliferation and invasion (Halleskog et al. 2011, 2012; Halleskog and Schulte 2013). Furthermore, WNT signaling regulates the capacity of brain microglia to attract tumor cells for colonialization of brain tissue (Pukrop et al. 2010). Studies on peripheral macrophages indicate also that WNT-3A and WNT-5A drive pro-inflammatory transformation (Pereira et al. 2009). WNT-5A is detectable in tumor-associated macrophages and supports macrophage-induced tumor invasiveness (Pukrop et al. 2006). Furthermore, WNT-5A is up-regulated in response to mycobacterial species, lipopolysaccharide (LPS), and interferon γ treatment and plays a critical role in macrophage communication in patients with severe sepsis (Blumenthal et al. 2006; Pereira et al. 2008). On the other hand also anti-inflammatory effects of WNTs were reported in peripheral macrophages: WNT-3A was capable to reduce mycobacterium-induced TNFα through its receptor FZD1 and the WNT/β-catenin pathway (Neumann et al. 2010).
Here, we investigate if WNTs can affect pre-activated microglia cells to limit LPS-induced pro-inflammatory transformation. For this purpose, we examine COX2 protein and mRNA levels in response to LPS alone or in combination with WNT-3A or WNT-5A. As both WNT-3A and WNT-5A efficiently limit LPS-induced changes, we hypothesize that WNTs act as homeostatic modulators of CNS immune function in a context-dependent manner.
Primary cell cultures
As previously described (Halleskog et al. 2012), primary microglia cells were isolated from new born C57Bl six mice (postnatal day 1–3), according to the ethical permit N436/10; local ethical committee Stockholms Norra Djurförsöksetiska Nämnd. Pups (mixed male/female), bred at the departmental animal facility, were decapitated, brains were dissected and kept in ice-cold Hank's buffered salt solution (HBSS). After removal of meninges, tissue was homogenized in 37°C Dulbecco's modified Eagle's medium, penicillin (50 U/mL), streptomycin (50 μg/mL), l-glutamine (2 mM), 10% fetal bovine serum, and Fungizon (0.5 μg/mL) (all from Invitrogen, Carlsbad, CA, USA). The homogenate was centrifuged and the pellet was resuspended in fresh medium and cultured in 160 cm2 flasks (two brains/flask). Medium was changed every fourth day. After 10–12 days microglia cells were separated from the underlying astrocyte monolayer by agitation. Purity (> 95%) of microglia cultures was assessed by immunocytochemistry with an anti-CD11b antibody (Abd Serotec) in combination with anti-GFAP (antibody from DAKO). Cells were stimulated 24 h after plating.
Treatment of cells
Recombinant, carrier free (CF) mouse WNT-3A (cat# 1324-WN/CF) and human/mouse WNT-5A (cat# 645-WN/CF) derived from WNT-overexpressing Chinese hamster ovary cells were purchased from R & D Systems, Minneapolis, MN, USA. According to stimulation paradigms used in earlier studies 30, 100, 300, and 1000 ng/mL WNT-3A and WNT-5A were used (Halleskog et al. 2012; Halleskog and Schulte 2013). LPS (cat# L4391; Sigma, St Louis, MO, USA) was used at 100 ng/mL. Cells were stimulated for 6 h with WNTs and/or LPS for assessment of changes in both protein levels and gene expression.
Immunoblotting was performed as previously described (Halleskog et al. 2012). Briefly, microglia lysates were analyzed with an 8% polyacrylamide gel and electrotransferred onto Immobilon-P membranes (Millipore Corporation, Bedford, MA, USA) . After blocking with 3% milk in Tris-buffered saline-Triton X100, membranes were incubated with primary mouse anti-β-actin (1 : 30.000; Sigma #A5441) or goat anti-COX2 (1 : 500; Santa Cruz # sc-1745) (Santa Cruz Biotechnology, Inc, Dallas, Texas; USA) overnight at 4°C. The proteins were immunodetected with appropriate horseradish peroxidase-conjugated secondary antibodies [goat anti-mouse (Pierce, Rockford, IL, USA) or rabbit anti-goat (Sigma)] and visualized by the enhanced chemiluminescence method (Western-Lightning, PerkinElmer, Waltham, MA, USA).
RNA isolation and Quantitative- PCR
Microglia cells were harvested and seeded in 6-well plates (Costar) at a density of 1 million cells/well. The cells were grown for 24 h, medium was changed to starvation medium for 24 h and after that the cells were stimulated for 6 h with carrier free WNT-3A carrier free WNT-5A or ctrl (phosphate-buffered saline), with and without 100 ng/mL LPS. RNA was isolated using the RNeasy Mini kit (Qiagen, Valencia, CA, USA) and transcribed to cDNA using the high-capacity cDNA Archive kit (Applied Biosystems/Life Technologies Ltd, Paisley, UK). QPCR was performed in triplicates on an ABI Prism 7000 sequence detector with the Taqman gene expression assay (Applied Biosystems) according to the manufacturer's instructions. Primer pairs: TNFα: Mm00443260_g1, IL-6: Mm99999064_m1, and COX2 (Ptgs2): Mm00478374_m1 (all from Applied Biosystems). As an internal reference standard primers for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were used. QPCR was performed in triplicates of each cDNA sample and 2−ΔCt (ΔCt = Ct, target gene - Ct, GAPDH) values were calculated for each condition and each independent experiment based on the mean of the three Ct values. For distinct visualization of the WNT effects on the LPS response, 2−ΔCt values for LPS stimulated cells from each individual experiment were set to 100%. The bar graphs summarize at least four independent, normalized experiments.
All data were statistically analyzed in Graph Pad Prism 5 software (GraphPad Software, Inc., La Jolla, CA, USA). Data were analyzed with one-way anova followed by Bonferroni's multiple comparison post hoc test. All experiments were repeated at least three independent times. Levels of statistical significance: */#p < 0.05; **/##p < 0.01; ***/###p < 0.001.
To investigate the potential of WNT-3A and WNT-5A as pro- and anti-inflammatory regulators, we stimulated naïve and LPS-pretreated mouse primary microglia with increasing doses (0, 30, 100, 300, 1000 ng/mL) of WNTs for 6 h. Cells were then lysed and COX2 expression, which served here as a generic marker of pro-inflammatory transformation (Mitchell et al. 1995), was analyzed in whole cell lysates by immunoblotting. In line with previously published data (Halleskog et al. 2011, 2012), WNT-3A (Fig. 1) and WNT-5A (Fig. 2) increased the level of COX2 expression, supporting the pro-inflammatory action of these two WNTs. LPS, a bacterial cell wall component, which generally promotes microglia activation through action at Toll-like receptor 4 (TLR4) (Lehnardt 2010), exceeded the elevation of COX2 levels upon WNT stimulation substantially. However, in combination, increasing doses of WNT-3A (Fig. 1) and WNT-5A (Fig. 2) reduced the LPS-induced expression of COX2 dose-dependently. WNT-3A led to 66% (300 ng/mL) and 82% reduction (1000 ng/mL), whereas WNT-5A diminished LPS effects on COX2 expression by 62% and 67% at the respective concentrations. The WNT-induced down-regulation of protein levels of COX2 could be achieved through translational or transcriptional regulation. To shed more light on underlying mechanisms, we analyzed mRNA expression of COX2 and other candidate genes as markers of the cell's pro-inflammatory transformation: Quantification of COX2, interleukin 6 (IL-6), and TNFα mRNA by quantitative PCR was performed in cells that were control treated, stimulated with LPS (100 ng/mL) or with LPS and WNT-3A (300 ng/mL) or WNT-5A (300 ng/mL) for 6 h (Fig. 3). WNT-3A led to a reduction of COX2, IL-6, and TNFα mRNA expression to 25%, 35%, and 35% of LPS-induced gene expression, respectively. WNT-5A diminished LPS effects to 36%, 46%, and 57%, respectively. The data indicate on one hand that WNTs can act anti-inflammatory by reducing LPS-induced expression of COX2 and proinflammatory cytokines. On the other hand they argue that WNTs affect LPS-induced gene expression on the transcriptional level.
Our previous studies show that mouse primary microglia cells transform into a pro-inflammatory state in response to WNT-3A and WNT-5A stimulation (Halleskog et al. 2011, 2012). As WNT-3A and WNT-5A generally activate different, even opposing signaling profiles with WNT-3A engaging mainly in β-catenin-dependent signaling and WNT-5A involving β-catenin-independent pathways (Topol et al. 2003; Nemeth et al. 2007; Schulte 2010), it appears surprising that both WNTs affect the LPS-induced transcription of COX2, IL-6, and TNFα in a similar manner. It appears that WNT-3A and WNT-5A employ common mechanisms in primary microglia to regulate LPS-induced gene transcription of this subset of inflammatory genes. Several mechanisms could account for the WNT-mediated reduction of LPS-induced gene transcription. As WNT-3A and WNT-5A both evoke activation of Gαi proteins and heterotrimeric G protein-dependent signaling in primary microglia resulting in the activation of the extracellular signal-regulated kinases1/2 pathway, calcium signaling and reduction in cyclic AMP levels (Halleskog et al. 2012; Halleskog and Schulte 2013), events downstream of heterotrimeric G proteins common for both WNT isoforms provide mechanisms for putative crosstalk between LPS and WNT signaling. The point of convergence of this crosstalk could either be at the receptor level, on intermediate signaling steps or at the transcriptional level. It is known that TLR4 membrane expression, functionality, and signaling can be reduced upon costimulation with diverse pharmacological compounds (Lin et al. 2007; Park et al. 2011; Jung et al. 2013), arguing that a heterologous desensitization of TLR4 itself or downstream signaling events could be the basis of reduced LPS effects in WNT-treated primary microglia. More detailed studies are required to elucidate underlying mechanisms.
Apparently, the bidirectional effects of WNTs as both pro- and anti-inflammatory regulators mirror the dual role of microglia in health and disease providing both supportive and inflammatory cues depending on the physiological context (Hanisch and Kettenmann 2007; Kettenmann et al. 2011). LPS, a cell wall component of bacteria, is a very strong activator of microglia and treatment with LPS should be seen as an experimental in vitro model of microglia pro-inflammatory activation. It remains so far unclear, under which circumstances WNT stimulation is interpreted as either pro- or anti-inflammatory input in vivo. β-catenin levels are dramatically increased in inflammatory microglia in postmortem human Alzheimer disease brains as well as in brains from old mice and APdE9 mice, an animal model for Alzheimer's disease (Halleskog et al. 2011). On the basis of these findings, we have previously argued that β-catenin signaling accompanies a pro-inflammatory transformation. The data provided here, however, strongly argue that WNTs are also capable to diminish pro-inflammatory input irrespective of their abilities to communicate through the WNT/β-catenin pathway.
Thus, we pose that WNTs act on microglia to accomplish homeostatic functions serving tissue protection in the CNS. To clarify the role of WNTs in the regulation of microglia and the CNS immune response, further studies are required to: (i) decipher the underlying molecular mechanisms of WNT/FZD-LPS/TLR4 signaling crosstalk, (ii) identify conditions in vivo where WNTs act either pro- or anti-inflammatory and (iii) characterize the homeostatic role of WNTs on the CNS immune response.
We thank Eva Lindgren for excellent laboratory assistance. The study was financially supported by grants from Karolinska Institutet, the Swedish Research Council (K2008-68P-20810-01-4, K2008-333 68X-20805-01-4, K2012-67X-20805-05-3), the Swedish Cancer Society (CAN 2008/539, 2011/690), Alzheimerfonden, Stiftelsen Lars Hiertas Minne and Signhild Engkvist's Foundation. Authorship credit: CH designed and planned the study, performed all experiments, quantifications, and statistical analysis. CH prepared figures and wrote the manuscript. GS coordinated the study and wrote the manuscript. All authors approved the final version of the manuscript. The authors declare that no conflict of interest exists.
WNT-3A and WNT-5A alone exert pro-inflammatory effects in primary microglia.
WNT-3A or WNT-5A combined with LPS act anti-inflammatory.
WNT-3A and WNT-5A regulates LPS-induced transcription of COX2, IL-6 and TNFα.
We conclude that WNTs serve as homeostatic immunoregulators in the CNS.