Background
Neuron and glial cells are in close association with each other and maintain physiological function in the central nervous system (CNS). When their finely controlled interactions are impaired by inflammation and stress conditions, neuronal networks are damaged, which results in the pathogenesis of several neurodegenerative diseases [
1‐
3]. It has been proposed that apoptotic cells or degenerating neurons release various signals to surrounding glial cells. These signals have been recently classified as ‘find-me’, ‘help-me’, and ‘eat-me’ signals [
4‐
8].
Microglia are resident immune cells in the CNS and express many versatile receptors [
9]. Therefore, they are considered the main recipient of various signals from degenerating neurons. Moreover, microglia exhibit early and rapid responses to various stimuli; for instance, activated microglia accumulate at pathological lesions [
10]. The rapid and accurate migration of microglia to lesions is predominantly mediated by various chemokines [
11]. In addition to chemokines, fibroblast growth factor (FGF)-2 regulates cellular migration in developing brain and in zebra fish [
12‐
15]; however, FGF-2 has not been directly implicated in microglial migration.
Fibroblast growth factor, purified from pituitary extracts, has a variety of functions, including inducing the proliferation and differentiation of various cell types, such as fibroblasts. Twenty-two types of FGF have been identified in human beings, as well as in mice. FGF-2 (basic FGF), one of the most common FGFs, has attracted attention for its widespread activity, such as cell proliferation, carcinoma cell invasion, neoangiogenesis, osteogenesis, and differentiation of developmentally staged constituent cells of the CNS [
16‐
19]. FGF-2 is expressed in various tissues at low levels, but its concentration is much higher in the brain. Five types of FGF receptor (FGFR: FGFR1 to 5) have been identified to date [
20], but their detailed expression levels in individual cells and mode of action in the CNS have not been elucidated. However, the expression levels of FGF-2 and FGFR have been shown to be up-regulated in CNS injury [
21]. Furthermore, several reports show that astrocytes, but not neuronal cells, are the dominant FGF-2-producing cells in the CNS [
16‐
19].
FGF-2 plays important roles in various cells in the CNS. Indeed, morphological change in glial cells and reactivity
in vivo[
22] have been demonstrated with FGF-2 injection into the cerebrospinal fluid. The best known FGF receptor–related signaling is MAPK, which is the common downstream signaling pathway of all FGFR subtypes. FGF-2 is known to induce Wnt/β-catenin signaling in human endothelial cells and developing the zebra fish brain [
12,
23,
24], but it is unclear whether FGF-2 also regulates Wnt/β-catenin signaling in microglia under neurodegenerative conditions.
In this study, we found that FGF-2 was secreted by glutamate or oligomeric amyloid β (oAβ) from damaged neurons, but not from astrocytes or microglia. Degenerating neurons produce signaling molecules that attract surrounding cells including microglia. Among these signaling molecules, we revealed FGF-2 as a predominant coordinator of microglial migration. FGF-2 induced microglial neuroprotection, migration and phagocytosis of neuronal debris via FGFR3. Furthermore, downstream signaling of FGF-2, especially through the FGFR3-extracellular signal-regulated kinase (ERK) signaling pathway, led to microglia-mediated neuronal survival. Wnt signaling directly induced this ERK phosphorylation and microglial migration, which were each enhanced by FGF-2 stimulation. Together, our results demonstrate that FGF-2 could be a key signaling molecule for crosstalk between degenerating neurons and microglia, and that the FGFR3/ERK/Wnt signaling pathway contributes to the induction of microglial neuroprotection.
Methods
Reagents
L-glutamate and goat immunoglobulin G (IgG), mouse IgG, and rat IgG were purchased from Sigma (St. Louis, MO, USA). Mouse recombinant FGF-2, mouse recombinant fractalkine (FKN; the chemokine domain), CCL21, and the FGFR (FGFR2-5) neutralizing antibodies were obtained from R & D Systems (Minneapolis, MN, USA). The MAPK inhibitors (U0126 (MEK1/2 inhibitor), SB203580 (p38 inhibitor), and SP600125 (JNK inhibitor)), PI3K inhibitor wortmannin, FGFR antagonist (PD173074 (pan-FGFR blocker), SU11652 (selective FGFR1 blocker)), and IWR-1-endo (Wnt antagonist) were purchased from Calbiochem (Gibbstown, NJ, USA). FGF-2 neutralizing antibody (aFGF-2) was purchased from Millipore (Billerica, MA, USA), and FKN neutralizing antibody (aFKN) was purchased as previously described [
25].
Preparation of Aβ solutions
Aβ1-42 solution was prepared as previously described [
26]. Briefly, synthetic human Aβ1-42 (Peptide Institute, Osaka, Japan) was dissolved to 1 mM in 100% 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP). The HFIP was dried and resuspended to a concentration of 5 mM in DMSO. To form oligomers, amyloid peptide was diluted to a final concentration of 100 μM with Ham’s F-12, incubated at 4°C for 24 h, and then immediately added to cultures at a final concentration of 5 μM.
Cell culture
The protocols for animal experiments were approved by the Animal Experiment Committee of Nagoya University. Primary neuronal cultures were prepared from the cortices of C57BL/6 mice embryos at embryonic day 17 (E17) as described previously [
27]. Briefly, cortical fragments were dissociated into single cells in dissociation solution (Sumitomo Bakelite, Akita, Japan), and resuspended in nerve culture medium (Sumitomo Bakelite). Neurons were seeded onto 12 mm polyethylenimine-coated glass coverslips (Asahi Techno Glass Corp., Chiba, Japan). The purity of the cultures was greater than 95%, as determined by NeuN-specific immunostaining [
28].
Microglia were isolated from primary mixed glial cell cultures prepared from newborn C57BL/6 mice at day
in vitro (DIV) 14 using the ‘shaking off’ method, which has been described previously [
29]. The purity of the cultures was 97 to 100% as determined by immunostaining for the Fc receptor. Cultures were maintained in DMEM supplemented with 10% fetal calf serum, 5 μg/ml bovine insulin, and 0.2% glucose. Astrocytes were purified from primary mixed glial cultures by three or four repetitions of trypsinization and replating. The purity of astrocytes was greater than 95%, as determined by GFAP-specific immunostaining [
30].
Measurement of FGF-2 levels
Secreted FGF-2 from mouse primary astrocytes, cortical neurons, and microglia were measured using an ELISA kit (RayBiotech, Inc., Norcross, GA, USA). Neurons were treated with L-glutamate (20 μM) or oAβ (5 μM) for 6 to 24 h at 37°C. Supernatants were then collected and assessed for FGF-2 levels.
Western blotting
Microglial cell lysates were boiled after the addition of sample buffer (1 M Tris-HCl, 20% sodium dodecyl sulfate (SDS), and 2.5% glycerol). Fifty micrograms of total protein were separated on a 5 to 20% Tris-glycine SDS-polyacrylamide gel and blotted onto Hybond-P polyvinylidene difluoride (PVDF) membranes (GE Healthcare UK, Buckinghamshire, UK). Membranes were blocked with 1% skim milk in Tris-buffered saline containing 0.05% Tween 20 for 1 h at room temperature. Primary antibodies to detect phosphorylated and total MAPK (Cell Signaling, Danvers, MA, USA) were applied at the concentrations recommended by the manufacturers. The secondary antibody was horseradish peroxidase-conjugated anti-rabbit IgG (GE Healthcare), which was used at a dilution of 1:1000. SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, Rockford, IL, USA) was used according to the manufacturer’s instructions. The intensities of the bands were calculated using the CS Analyzer 1.0 (Atto Corporation, Tokyo, Japan).
HEK293T cells were seeded one day before transfection by FuGENE HD (Promega, Madison, WI, USA) with a luciferase reporter vector from the Cignal TCF/LEF Reporter (luc) kit (Wnt promoter assay system), which was purchased from SABiosciences (Qiagen KK, Tokyo, Japan). After drug treatment, cells were lysed and luciferase reporter activity was measured using the Dual luciferase reporter assay kit (Promega) and a Wallac 1420 ARVOMX (PerkinElmer Japan, Yokohama, Japan).
Evaluation of microglial phagocytosis
A microglial phagocytosis assay was performed as previously described [
25]. Briefly, primary mouse cortical neurons in 24-well plates were labeled on DIV 14 with 1 μM CM-DiI (Molecular Probes), and treated with 20 μM glutamate overnight at 37°C. After changing the culture medium, microglia were added to these neuronal cultures (1:2 ratio for neurons to microglia) with or without FGF-2 for 24 h. Cells were subsequently fixed in 4% paraformaldehyde. Microglia were stained with Cy5-conjugated rat anti-mouse CD11b monoclonal antibodies prior to fixation. Phagocytic uptake of neuronal debris by microglia was estimated based on the detection of DiI-stained neuronal debris [
31] in CD11b-positive microglia (green); the phagocytosis index was calculated as the percentage of red staining that overlapped with green staining (shown in yello
w) among all of the microglia.
Immunocytochemistry
Cells were fixed with 4% paraformaldehyde, blocked, and permeabilized. Neurons were stained with mouse polyclonal anti–MAP-2 antibody (1:1000; Chemicon, Temecula, CA, USA) and secondary antibody conjugated to Alexa 488 (1:1000; Invitrogen). Astrocytes were stained with mouse monoclonal anti-GFAP antibody (Sigma) and secondary antibody conjugated to Alexa 647 (1:1000; Invitrogen). Microglia were stained with Cy5-conjugated rat anti-mouse CD11b monoclonal antibody (1:300, BD Pharmingen) prior to fixation. Images were analyzed using a deconvolution fluorescence microscope system (BZ-8000; Keyence Corporation, Osaka, Japan). The other primary antibodies included FGFRs, which were purchased from R & D systems and used according to the manufacturer’s instructions.
Surviving neurons were identified based on their cytoskeletons as previously described [
28]. Viable neurons were strongly stained with anti-MAP-2 antibodies, whereas damaged neurons showed weaker staining. MAP-2-positive neurons were counted in representative areas in each well. Using five independent trials, more than 200 neurons were evaluated in each well by a scorer who was blind to the experimental conditions. The number of viable neurons in untreated cultures was set as 100%.
Measurement of CCL3 (MIP-1a), NO, and glutamate levels
Supernatants from microglia were assessed using the chemokine (C-C motif) ligand 3 (CCL3) ELISA kit (R & D Systems), and a Griess reaction for nitric oxide (NO) detection. To measure glutamate levels, a colorimetric assay kit (Yamasa Corporation, Tokyo, Japan) was used, as previously described [
25].
MTS assay
To evaluate the viability of the cells, we used the CellTiter 96 Aqueous One Solution Cell Proliferation Assay kit (Promega) and followed the manufacturer’s instructions.
Microglial migration assay
Microglial migration was performed using Transwell plates with 3 μm pore polyethylene terephthalate (PET) membrane filters (BD Biosciences). We placed 800 μl of neuronal-conditioned medium or microglial culture medium treated with drugs into the lower chamber of the Transwell plate. Membrane filters were then put in vacant wells, and 200 μl of microglia-containing medium (1.0 × 10
5 cells/well) was carefully added on top of the filter membrane to avoid bubbles. These plates were incubated for 24 h. Cells that migrated into the lower wells were counted by fluorescence-activated cell sorting (FACS). Chemokine-treated T cells (combination of FKN and CCL21 (100 nM each)) were used as positive controls for this method, as previously described [
32].
RT-PCR
Total RNA was extracted from astrocytes, microglia, and neurons using an RNeasy Mini Kit (Qiagen, Tokyo, Japan). A first-strand cDNA library was obtained using SuperScript II (Invitrogen, Carlsbad, CA) and oligo (dT) 12-18 (Invitrogen) as the first-strand primer. Negative control reactions were performed using the same system after heat denaturation of reverse transcriptase. RT-PCR was used to amplify transcripts encoding mouse FGF-2, each receptor subtypes and glyceraldehydes-3-phosphate dehydrogenase (GAPDH), using 0.1 μg of first-strand cDNA, Blend Taq polymerase (Toyobo Co., Osaka, Japan), and oligonucleotide primers (Table
1; except for previously described primers for GAPDH [
25]).
Table 1
Oligonucleotide primers used in RT-PCR
FGF-2 sense antisense | 5′-AGCGGCTCTACTGCAAGAAC | 371 |
| 5′-AGCAGACATTGGAAGAAACAGT | |
FGFR1 sense antisense | 5′-GTTGGGCTCTGTCATCATCTAT | 522 |
| 5′-GCGTACTCCACAATGACATAAA | |
FGFR2 (IIIb, IIIc) sense antisense | 5′-CTCATCCTGCTGGGTCTGAG | 748 |
| 5′-AGGAGTAGCAGCTGATGTGAC | |
FGFR3 sense antisense | 5′-CCTGTGTAGTTGAGAACAAGTTT | 625 |
| 5′-GTGTTGGAGTTCATAGAGGAGT | |
FGFR4 sense antisense | 5′-GAGGTCTTGTATCTGAGGAACG | 651 |
| 5′-GTTCTTGTGTCTTCCGATTAGC | |
FGFR5 sense antisense | 5′-ATGATATTAGTCCAGGGAAGG | 366 |
| 5′-GGATTACATCCACTTTGTAGGT | |
Statistical analysis
Statistically significant differences between experimental groups were determined by one-way analysis of variance (ANOVA) followed by Dunnett’s or Tukey’s tests for multiple comparisons. Statistical analysis was performed using the software program Prism 4 for Windows (GraphPad Software, San Diego, CA, USA). P values less than 0.05 were considered significant.
Discussion
Our results indicate that FGF-2 is released from degenerating neurons and induces microglial migration and neuroprotection, which are mediated through the FGFR3-Wnt-ERK signaling pathway. Neurons were fine responders of glutamate and oAβ, and then allowed the release of FGF-2 in relatively short times. FGF receptors are expressed in neurons and glial cells. FGFR3, in particular, is activated by FGF-2 via the ERK MAPK-dependent signaling pathway in microglia. The other FGF, FGF-19, is reported to negatively regulate NFκB via FGFR4 [
38]. In the developmental morphogenic stages and angiogenesis, coordinated action of Wnt/β-catenin and FGF signaling has been reported [
12,
23,
24,
39]. Recently, expression of Wnt receptors Frizzled and LDL receptor-related protein 5/6 has been reported in mouse primary microglia [
37]. In this study, we revealed that FGF-2 directly controlled the Wnt signaling pathway in mouse primary microglia, and that Wnt signaling could also directly regulate microglial migration induced by FGF-2. FGF-2 and the extracellular matrix protein Anosmin-1 have dynamic roles in cellular proliferation and migration from the subventricular zone in CNS development [
40]. FGF-2 enhances the proliferation and differentiation of neuronal stem cells. Anosmin-1 and FGF-2 could possibly be diagnostic markers in multiple sclerosis (MS), because their expression level varies between different types of MS [
16]. In experimental autoimmune encephalomyelitis, the animal model of MS, FGF-2 may act as a remyelinating and nerve fiber preserving agent [
41]. Therefore, FGF-2/Wnt signaling has a potential to regulate cellular proliferation and migration to maintain adult CNS function.
Localized delivery of FGF-2 and brain-derived neurotrophic factor (BDNF) to the lesioned hippocampus increases neurogenesis and reduces epileptogenesis in a rat model of epilepsy [
42]. The overexpression of FGF-2/BDNF also attenuates neuroinflammation through suppression of IL-1β [
43]. Moreover, FGF-2 gene delivery restores hippocampal functions in an Alzheimer’s disease mouse model [
44]. FGF-2 has a deep connection with tumorigenicity. CD44-mediated migration of human inflammatory macrophages into the extravascular compartment depends on binding of FGF-2 to the CD44 receptor [
45]. Therefore, it is possible that FGF-2 has functional association with a new counterpart other than FGFRs.
The brain concentration of FGF2 is reported to be around 30 to 120 ng/mg [
46]; however, some reports show that the concentration is around 50 pg/ml [
47,
48]. In a future study, we will attempt to clarify the effect of 100 ng/ml FGF2
in vivo. Taken together, the present study shows that FGF-2 from damaged neurons functions as help-me and eat-me signals. Targeting the FGF-2/FGFR3 pathway may give us clues for future therapeutic strategy against neurodegenerative diseases.
Acknowledgements
This study was supported by JSPS KAKENHI Grant Number 24659430, a grant from the Advanced Research for Medical Products Mining Programme of the National Institute of Biomedical Innovation (NIBIO), and grants from the Ministry of Health, Labour and Welfare of Japan.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MN conducted the ELISAs, microglial phagocytosis assay, FACS analysis, and statistical analysis, and drafted the manuscript. KT performed the RT-PCR experiments and helped draft the manuscript. BP and JK performed the cell culture and were involved in the conception of the study. YS and HT were also involved in the conception of the study. TM carried out the immunocytochemistry and statistical analysis. He was also involved in the conception and design of the study, and helped draft the manuscript. AS was also involved in the conception and design of the study, as well as in preparing the manuscript. All authors read and approved the final manuscript.