Granulocyte-Colony Stimulating Factor Attenuates Oligomeric Amyloid β Neurotoxicity by Activation of Neprilysin

Soluble oligomeric amyloid β (oAβ) causes synaptic dysfunction and neuronal cell death, which are involved in the pathogenesis of Alzheimer's disease (AD). The hematopoietic growth factor granulocyte-colony stimulating factor (G-CSF) is expressed in the central nervous system (CNS) and drives neurogenesis. Here we show that G-CSF attenuated oAβ neurotoxicity through the enhancement of the enzymatic activity of Aβ-degrading enzyme neprilysin (NEP) in neurons, while the NEP inhibitor thiorphan abolished the neuroprotection. Inhibition of MEK5/ERK5, a major downstream effector of G-CSF signaling, also ablated neuroprotective effect of G-CSF. Furthermore, intracerebroventricular administration of G-CSF enhanced NEP enzymatic activity and clearance of Aβ in APP/PS1 transgenic mice. Thus, we propose that G-CSF may be a possible therapeutic strategy against AD.


Introduction
Alzheimer's disease (AD) is a neurodegenerative disorder and the most common cause of dementia in the elderly. One of the pathological hallmarks of AD is senile plaque, whose major component is fibrillar amyloid b (fAb). While fAb-induces neuronal dystrophy and tau hyperphosphorylation [1,2], soluble oligomeric Ab (oAb) has been reported to exhibit higher neurotoxicity than fAb. oAb reportedly inhibits hippocampal long-term potentiation and disrupts synaptic plasticity [3,4].
Granulocyte-colony stimulating factor (G-CSF) is a major growth factor in the differentiation and proliferation of neutrophilic-granulocytic lineage cells that modulates the immune response by inhibiting the production of inflammatory cytokines [5,6]. Both G-CSF and its receptor G-CSFR are widely expressed in neurons in the central nervous systems (CNS), and their expression is induced by ischemia [7]. G-CSFR is also reportedly expressed in adult neural stem cells, and G-CSF can induce neuronal differentiation in vitro [7]. However, the exact functions of G-CSF await further elucidation.
Administration of G-CSF has been shown to improve cognitive performance in an AD model mouse carrying the Tg2576 transgene without reduction of Ab burden [8]. The mechanism is reported to be due to local neurogenesis surrounding Ab aggregates and the enhancement of acetylcholine levels. Another report shows that G-CSF ameliorates cognitive impairments with accompanying decreases of Ab burden in APP/PS1 transgenic (Tg) mouse model of AD [9]. The study reported that the effects of G-CSF are due to upregulation of neurogenesis by neuronal stem cells and Ab clearance by microglia. However, the precise functions of G-CSF on mature neurons are not fully understood. Increasing zinc-metalloprotease neprilysin (NEP) activity in AD mouse models reportedly improves cognitive impairments [10]. Indeed, NEP is one of the most prominent Ab degrading enzymes. In this study, we show that G-CSF attenuates oAb 1-42 toxicity via activation of NEP.

Materials and Methods
Preparation of oligomeric Ab  Soluble oligomeric amyloid b 1-42 (oAb 1-42 ) was prepared as described previously [11]. Briefly, synthetic human Ab 1-42 (Peptide Institute, Osaka, Japan) was dissolved in 100% 1,1,1,3,3,3-hexafluoro-2-propanol at a concentration of 1 mM. This solution was completely dried by the vacuum desiccator. The obtained film was resuspended in dimethyl sulfoxide to a concentration of 5 mM, and diluted with Dulbecco's Modified Eagle Medium/F12 (Invitrogen, Carlsbad, CA, USA) at a concentration of 100 mM. This solution was incubated at 4uC for 24 h to obtain oAb  . A final concentration of 5 mM oAb  was used in all experiments.

Animals
This study was carried out in strict accordance with the guideline for the care and use of laboratory animals of Nagoya University. All protocols for animal experiments were approved by the Animal Experiment Committee of Nagoya University. Transgenic mice expressing mutant variants of human amyloid precursor protein (APP) with K595N and M596L mutations and presenilin 1 (PS1) with A264E mutation were purchased from the Jackson Laboratory (B6C3-Tg(APP695)3Dbo Tg(PSEN1)5Dbo/J; #003378) and were backcrossed to C57BL/6J mice for more than 10 generations after purchase (here designated as APP/PS1 Tg mice).
G-CSF (30 ng/3 ml) or vehicle [phosphate-buffered saline (PBS)] was injected into the cerebral ventricular space of 12month-old APP/PS1 Tg mice as previously described [12,13]. Three days after injection, deep-anesthetized mice were transcardially perfused with ice-cold PBS, and the brains were collected. The left hemispheres were used for histological analysis, and the right hemispheres were used for assessments of neprilysin enzymatic activity and Ab concentration.

Neuronal culture
Primary mouse cortical neurons were prepared as previously described [11,14]. Briefly, cerebral cortices were isolated from C57BL/6J mouse embryos on the 17 th embryonic day, minced and treated with dissociation solution (Sumitomo Bakelite, Akita, Japan). Neurons were resuspended in Nerve Culture Medium (Sumitomo Bakelite), plated on polyethylenimine-coated glass coverslips (Asahi Techno Glass, Chiba, Japan) at a density of 5610 3 cells/well in 96-well multidishes, 5610 4 cells/well in 24well multidishes, or 6610 6 cells/well in 6-well multidishes, and incubated at 37uC in an atmosphere containing 5% CO 2 at 100% humidity. The purity of the cultures was greater than 95% based on NeuN-specific immunostaining. Neurons were used at 14 days in vitro for the following assessments.

Immunohistochemistry
Ten-micrometer-thick frozen sections of APP/PS1 Tg mouse brains were prepared using a previously described method [11]. Sections were permeabilized with 1% Triton X-100 after blocking with 10% normal goat serum for 30 min, and then were incubated with anti-Ab mouse monoclonal antibody (clone 4G8, 1:500, Chemicon) overnight at 4uC. After rinsing, they were incubated with Alexa488-conjugated secondary antibody (1:1,000, Invitrogen) and 1 mg/ml Hoechst33342 for 1 h at room temperature. After rinsing, they were mounted in Fluoromount-G (South-ernBiotech). Images were analyzed with a deconvolution fluorescence microscope system (Keyence).

RNA extraction and reverse transcription-PCR (RT-PCR)
The mRNA expression of neprilysin was detected by RT-PCR. Neurons were plated at a density of 5610 4 cells per well in 24-well multidishes, and stimulated with or without 100 ng/ml G-CSF (R&D Systems, Minneapolis, MN, USA) for 6 h. Total RNA was extracted from neurons using RNeasy Mini Kit (Qiagen, Valencia, CA, USA). cDNA synthesis was performed using SuperScript II (Invitrogen). PCR was carried out using the following primers.

Measurement of protein level and enzymatic activity of NEP
The cell membrane fractions from the mouse primary neurons or the APP/PS1 Tg mouse brains were harvested and assessed for NEP protein levels using specific ELISA (R&D Systems). NEP enzymatic activity was also examined as described previously [16]. The fluorescence of each samples were measured by a Wallac 1420 ARVO MX (PerkinElmer Japan, Yokohama, Japan). Human Ab ELISA To evaluate oAb  in neuronal culture, we used a human Ab oligomer specific ELISA kit (IBL, Gunma, Japan). Neurons were pre-treated with 10 mM thiorphan for 1 h and then treated with G-CSF for 3 h prior to the addition of 5 mM oAb 1-42 for 24 h. The neuronal culture supernatants were assessed with an ELISA kit. To evaluate the amount of human Ab 1-40 and Ab 1-42 in mouse brains, we used a human Ab 1-40 and Ab 1-42 specific ELISA kit (Wako Pure Chemical Industries, Osaka, Japan) as previously described [17]. Brains were homogenized with TNE lysis buffer [50 mM Tris-HCl at pH 7.6, 1% Nonidet P-40, 150 mM NaCl, 2 mM EDTA, and protease inhibitor mixture (Complete Mini EDTA-free, Roche, Germany)] and centrifuged at 10,000 g for 15 min at 4uC. The supernatants were analyzed by each Ab specific ELISA kit. The values obtained were corrected with the wet weight of each brain sample.

Statistical Analysis
Statistical significance was analyzed with a Student's t-test or one-way analysis of variance followed by Tukey's post-hoc test

G-CSF rescues oAb 1-42 -induced neuronal damage
We first examined the effects of G-CSF on oAb 1-42 -induced neurotoxicity using mouse primary neuronal culture (Figure 1). We found that treatment with 5 mM oAb 1-42 for 24 h resulted in severe neurotoxicity ( Figure 1B; Figure 1F and 1G, black columns). Three hours before the addition of 5 mM oAb 1-42 , we then applied 1-100 ng/ml G-CSF for 24 h. Treatment with G-CSF significantly suppressed oAb 1-42 -induced neuronal damage in a dose-dependent manner ( Figure 1C-E; Figure 1F and 1G, shaded columns).

G-CSF enhances oAb degradation through activation of NEP
Next, we assessed whether G-CSF treatment alters the amount of Ab applied in neuronal culture. We found that G-CSF significantly decreased Ab concentration in neuronal culture ( Figure 2D, black column). We then assessed the expression levels of Ab-degrading enzymes [NEP and insulin-degrading enzyme (IDE)] in G-CSF-treated neurons. RT-PCR data indicated that the addition of G-CSF upregulated the expression level of NEP in neurons, whereas IDE was not affected (Figure 2A and data not shown). Next, we assessed the protein levels and enzymatic activity of NEP. G-CSF treatment significantly enhanced NEP enzymatic activity, but not NEP protein level ( Figure 2B and 2C). Inhibition of NEP by thiorphan completely reversed the amount of Ab ( Figure 2D, dotted column), suggesting that Ab degradation by G-CSF stems from the activation of neuronal NEP. Treatment with thiorphan alone did not affect the amount of Ab.

NEP is critical for the neuroprotective effect of G-CSF
We assessed whether the neuroprotective effect of G-CSF results from NEP (Figure 3). We found that treatment with the NEP inhibitor thiorphan almost completely ablated the neuroprotective effects of G-CSF ( Figure 3D-F; Figure 3G and 3H, shaded columns). These findings imply that treatment with G-CSF enhanced neuronal NEP activity and protected against oAb 1-42induced neurotoxicity through Ab degradation.

The MEK5/ERK5 pathway contributes to G-CSF-mediated neuroprotection
The MEK5/ERK5 pathway is a major downstream effector of G-CSF signaling. We examined the role of the MEK5/ERK5 pathway in G-CSF-mediated neuroprotection. We found that inhibition of MEK5/ERK5 by BIX02189 almost completely suppressed the neuroprotective effects of G-CSF against oAbinduced neurotoxicity ( Figure 4D-F; Figure 4G and 4H, shaded columns). We confirmed BIX02189 decreased NEP activity in G-CSF-treated neurons. These results suggest that G-CSF-mediated neuroprotection depended on MEK5/ERK5 signaling.
In vivo G-CSF treatment enhances Ab 1-42 degradation by activation of NEP Finally, we examined whether G-CSF treatment enhances NEP activity and Ab degradation using APP/PS1 Tg mice, a mouse model of Alzheimer's disease. G-CSF was injected into the cerebral ventricular space of APP/PS1 mice. Three days after injection, mouse brains were assessed by histological and biochemical analysis. Histological analysis revealed that G-CSF treatment reduced the Ab burden in the hippocampus, whereas PBS-treated mice showed substantial amounts of Ab deposits ( Figure 5A-D). As expected, G-CSF treatment significantly enhanced NEP activity in the brains of APP/PS1 Tg mice, whereas NEP protein levels were not affected ( Figure 5E and 5F). Human Ab-specific ELISAs also revealed that G-CSF injection significantly reduced the amount of Ab 1-42 in APP/PS1 transgenic mice, whereas Ab 1-40 load was not affected ( Figure 5G and 5H).

Discussion
The present study revealed a novel neuroprotective function of G-CSF against oAb toxicity. We found that G-CSF significantly enhanced neuronal NEP activity and led to increased degradation of oAb. Furthermore, injection of G-CSF into the cerebral ventricular space of APP/PS1 mice also enhanced oAb degradation by activation of NEP.
NEP is the major Ab degrading peptidase. NEP deficiency results in elevated levels of endogenous Ab 1-40 and Ab 1-42 in the hippocampus, cortex, thalamus/striatum, and cerebellum [18].
NEP is also reported to degrade Ab 1-40 more rapidly than Ab  in vitro [19]. However, our in vivo data show that G-CSF reduced the amount of Ab 1-42 in APP/PS1 Tg mice, though Ab 1-40 was not affected. Clearance of Ab 1-40 may not depend on NEPcatalyzed proteolysis as that of Ab 1-42 . These results suggest that G-CSF has an effect on Ab 1-42 degradation via NEP activation in vivo. While NEP is capable of cleaving Ab monomers, its ability to degrade oAb is controversial [10]. However, a recent report shows that NEP gene transfer into an AD mouse model significantly reduces oAb [20]. In the present study, we have shown that G-CSF treatment reduced the amount of oAb in the supernatants of neuronal cultures via activation of NEP. Therefore, NEP is clearly able to degrade oAb.
Another Ab degrading enzyme, IDE, is reported to be reduced in the hippocampus of AD [21]. The enhanced IDE activity in IDE and APP double-transgenic mice decreases Ab levels and prevents the formation of AD pathology. However, G-CSF did not induce activation of IDE in neurons in that study. The reduced level of oAb was small. Other mechanism such as neurogenesis may be involved in neuroprotection.
G-CSF activates the Jak/Stat, MAPK (Erk1/2, JNK and p38), PI3-K, and Src family kinase cascades [22]. A recent study shows that the MEK5/ERK5 pathway is a major downstream effector of G-CSF signaling in the regulation of cell proliferation and survival. [23,24]. In the present study, inhibition of MEK5/ERK5 by BIX02189 almost completely suppressed the neuroprotective effects of G-CSF against oAb-induced neurotoxicity. The results suggest that G-CSF-induced NEP is activated by the MEK5/ ERK5 pathway. MEK5/ERK5 pathway is involved in cell proliferation, cell survival, and angiogenesis. However, the precise mechanism of NEP expression by MEK5/ERK5 remains unknown.
The G-CSF receptor is also expressed in microglia, and expression is increased after spinal cord injury [25]. G-CSF has been shown to promote the recruitment of microglia to the injury site, which regulates microglial activation by inhibiting the activity of NF-kB. [26]. In the previous study, G-CSF increased microglial burden, reduced Ab deposition, and ameliorated the cognitive impairments in APP/PS1 mice. This mechanism is considered to be microglial Ab clearance and neurogenesis in neural stem cells [9]. Therefore, microglial Ab clearance may also contribute to decreasing the amount of Ab 1-42 by G-CSF injection in APP/PS1 transgenic mice in the present study. Taken together, the present study shows that G-CSF significantly enhances the expression level and enzymatic activity of NEP in neurons, indicating that G-CSF could be a useful therapeutic strategy against oAb 1-42 neurotoxicity in AD.