In addition to their neurotoxic role in Alzheimer's disease (AD), β-amyloid peptides (Aβs) are also known to play physiological roles. Here, we show that recombinant Aβ40 significantly increased the outward current of the GABAA receptor containing (GABAAα6) in rat cerebellar granule neurons (CGNs). The Aβ40-mediated increase in GABAAα6 current was mediated by an increase in GABAAα6 protein expression at the translational rather than the transcriptional level. The exposure of CGNs to Aβ40 markedly induced the phosphorylation of ERK (pERK) and mammalian target of rapamycin (pmTOR). The increase in GABAAα6 current and expression was attenuated by specific inhibitors of ERK or mTOR, suggesting that the ERK and mTOR signaling pathways are required for the effect of Aβ40 on GABAAα6 current and expression in CGNs. A pharmacological blockade of the p75 neurotrophin receptor (p75NTR), but not the insulin or α7-nAChR receptors, abrogated the effect of Aβ40 on GABAAα6 protein expression and current. Furthermore, the expression of GABAAα6 was lower in CGNs from APP−/− mice than in CGNs from wild-type mice. Moreover, the internal granule layer (IGL) in APP−/− mice was thinner than the IGL in wild-type mice. The injection of Aβ40 into the cerebellum reversed this effect, and the application of p75NTR blocking antibody abolished the effects of Aβ40 on cerebellum morphology in APP−/− mice. Our results suggest that low concentrations of Aβ40 play a role in regulating CGN maturation through p75NTR.
In addition to its neurotoxic role in Alzheimer's disease, Aβ is known to play important physiological roles. Whether Aβ improves neuronal development and maturation remains elusive. Our results demonstrate that low concentrations of Aβ40 significantly increase the GABAA receptor α6 subunit expression and associated current in cerebellar granule neurons (CGNs) via the p75NTR and MEK/ERK pathway. Aβ also increases the thickness of the internal granule layer in APP−/− mice cerebellum. Our data provide new evidence for the role of Aβ40 in regulating the maturation of CGNs.
After β-amyloid peptide (Aβ) was first identified in the brain of Alzheimer's disease (AD) patients (Glenner and Wong 1984), its neurotoxic effects were well established. However, soluble Aβ has been detected in the cerebrospinal fluid and serum of healthy individuals (Shoji 2002; Cleary et al. 2005; Walsh et al. 2005), and the inhibition of Aβ by antibodies or inhibitors significantly impairs memory and learning in young mice (Morley et al. 2010). In addition, neurogenesis is enhanced in the hippocampus of patients with AD and in a transgenic mouse model (Jin et al. 2004). Therefore, it is important to re-examine the physiological role of Aβ (Pearson and Peers 2006).
Several lines of evidence indicate that Aβ may function similarly to neurotrophins with respect to its ability to protect neurons. The synthetic Aβ monomer has been reported to aid the survival of neurons under conditions of trophic deprivation and to protect mature neurons from excitotoxic death (Giuffrida et al. 2009). Inhibiting the activity of secretases that cleave β-amyloid precursor protein (APP) and release Aβ induces death in neuronal cells (Plant et al. 2003). Aβ was reported to stimulate neurogenesis in the subventricular zone (SVZ) through p75NTR (Sotthibundhu et al. 2009). In hippocampal neurons, low concentrations of Aβ mimic the activity of nerve growth factor (NGF) at the levels of gene expression, NF-κB activation, neuronal morphology, and connectivity (Arevalo et al. 2009). Moreover, endogenously released Aβ peptides positively regulate the release probability of synapses and control synaptic activity (Abramov et al. 2009). In addition, Aβ has been shown to exert effects on A-type K+ currents in primary cultures of central neurons (Ramsden et al. 2001). Although the physiological role of Aβ in maintaining neuronal survival has been suggested, the effects of Aβ on the maturation of central neurons remain unknown.
Cerebellar granule neurons (CGNs) constitute the largest homogeneous neuronal population in the mammalian brain. Because of their post-natal generation and well-defined developmental pathway in cell culture, CGNs have been established as a model for studying neuronal development and maturation (D'Angelo et al. 2009). In addition to the changes in the types and current amplitudes of K+ channels during CGN development and maturation (Shibata et al. 1999; Mei et al. 2004), there is a predominant switch in the expression of the gamma-amino butyric acid type A receptor (GABAA) subunit from α2 and α3 to α1 and α6. This switch is likely the marker of terminal differentiation and maturation in neurons (Mellor et al. 1998; Nakanishi and Okazawa 2006). Previous studies have indicated that in CGNs, brain-derived neurotrophic factor regulates GABAAα6 expression via the MEK/ERK pathway (Bulleit and Hsieh 2000), and pre-treatment of CGNs with unaggregated Aβ40 for 24 h results in a 60% increase in the A-type K+ current (Ramsden et al. 2001). These findings led us to investigate whether Aβ40 regulates neuronal maturation.
In this study, we examined the effects of Aβ40 on the regulation of neuronal maturation by measuring GABAA α6 expression in rat CGNs. We reveal for the first time that Aβ40 increases the expression of the GABAA α6 subunit via p75NTR and the MEK/ERK pathway and that low levels of Aβ40 play a role in the regulation of CGN maturation.
Materials and methods
All experimental procedures were carried out in accordance with the European guidelines for the care and use of laboratory animals (Council Directive 86/609/EEC). The protocol was approved by the Committee on the Ethics of Animal Experiments of Fudan University (Permit Number: 20090614-001). All surgeries were performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.
Primary cultures of rat cerebellar granule neurons
Sprague–Dawley female rats were purchased from the Laboratory Animal Center of Shanghai at the Chinese Academy of Sciences (Shanghai, China). CGNs were derived from the cerebella of 7-day-old pups as described previously (Zhou et al. 2012). Isolated cells were plated onto 35 mm Petri dishes coated with poly-l-lysine (1 μg/mL) at a density of 106 cells/mL. Cultured cells were incubated at 37°C under 5% CO2 in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, insulin (5 μg/mL), KCl (25 mM) and a 1% antibiotic–antimycotic solution. Experiments were performed using primary CGNs after 3–8 days in culture (DIC).
GABA current recordings
Whole-cell currents from granule neurons were recorded using a conventional patch-clamp technique. Prior to GABA current recording, the culture medium was replaced with a bath solution containing the following (in mM):NaCl 125, KCl 2.5, HEPES 10, MgCl2 1 and glucose 10 (pH adjusted to 7.4 with NaOH). Soft-glass recording pipettes were filled with an internal solution containing the following (in mM): KCl 145, HEPES 10, CaCl2 1, MgCl2 1, EGTA 10, and ATP 1 (pH adjusted to 7.2 with KOH). The recordings were performed at 23–25°C. While recording, 100 μM GABA was added to the cells via a rapid perfusion system, which ensured constant perfusion with the bath solution. The GABA receptor current was recorded for 4 s every 40 s. GABAAα6 currents were recorded in cells treated with or without Aβ40. To correctly evaluate the effects of the experimental drug (with and without Aβ), each experiment had a corresponding control, and the results are displayed using histograms.
The cell surface proteins were biotinylated according to the manufacturer's protocol as previously described (Yao et al. 2009). Briefly, the neurons were incubated with 0.25 mg/mL sulfo-NHS-SS-biotin (Thermo Scientific, Rockford, IL, USA) for 45 min at 4°C and subsequently blocked with 50 mM Tris (pH 8.0) for 20 min at 4°C. The cells were lysed in HEPES-NP40 lysis buffer. Biotinylated proteins were pulled down using streptavidin-agarose beads (Thermo Scientific) overnight at 4°C and subsequently washed four times with lysis buffer. The bound proteins were eluted with the sample buffer and used for western blotting.
CGNs were incubated with Aβ40 for 30 min and harvested as described in the western blotting section. The supernatants were incubated overnight with or without a mouse monoclonal antibody against Aβ (1 : 1000). The immune complexes were precipitated with protein A/G PLUS-agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 45 min at 4°C, and 3–4 sequential washes were then performed in HEPES-NP40 lysis buffer. After the final wash, the pellet was resuspended in sample buffer and boiled for 5 min at 95°C. After centrifugation, western blot analysis was performed on the supernatants. The p75NTR protein was labeled with a rabbit monoclonal antibody (clone D4B3, 1 : 1000; Cell Signaling Technology, Beverly, MA, USA).
Protein samples were prepared with HEPES-NP40 lysis buffer (20 mM HEPES, 150 mM NaCl, 0.5% NP-40, 10% glycerol, 2 mM EDTA, 100 μM Na3VO4, 50 mM NaF and 1% protease inhibitor cocktail at pH 7.5) as previously described (Zhou et al. 2012). The proteins were probed with the following antibodies: rabbit polyclonal against GABAAα6 (1 : 2000; Millipore, Billerica, MA, USA); rabbit monoclonal against phosphorylated ERK1/2 or total ERK1/2 (1 : 1000; Cell Signaling Technology); or rabbit polyclonal against phosphorylated mTOR (1 : 1000; Cell Signaling Technology). Mouse monoclonal antibodies against GAPDH (1 : 10 000; KangChen Bio-Tech, Shanghai, China) and the transferrin receptor (1 : 1000; Invitrogen, Carlsbad, CA, USA) were used as loading controls. After washing in TBST, the membrane was incubated with horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG (1 : 10 000) (KangChen Bio-Tech) for 2 h at 23–25°C. Chemiluminescent signals were generated using a Super Signal West Pico Trial Kit (Pierce Biotechnology, Rockford, IL, USA) and detected via exposure to X-ray film or by using the ChemiDoc XRS System (Bio-Rad Laboratories Inc., Hercules, CA, USA).
Quantitative real-time PCR
Quantitative real-time PCR (qPCR) was performed to measure GABAA α6 mRNA levels. Total RNA was harvested from primary cultured rat cerebellar granule cells and purified with an RNeasy Mini Kit (Qiagen, Valencia, CA, USA). Complementary DNA (cDNA) was synthesized from 1 μg of total RNA using M-MLV reverse transcriptase (Promega, Madison, WI, USA) following the manufacturer's instructions. qPCR was performed with the following primer pairs: GABAA α6 subunit forward (5′-TGTTGCTTCTCCCCTGG-3′) and reverse (5′-ACGGTTGTCATAGCCCTCC-3′) (NM_021841.1) and cyclophilin D forward (5′-AAGGGCACTGGACCGACAA-3′) and reverse (5′-GCCATGCTCAGCAAACC-3′)(NM_001004279.1). The 20-μL reactions contained 2.0 μL of the diluted cDNA, 0.2 μM each primer pair and 1XPower SYBR Green PCR Master Mix (Toyobo, Osaka, Japan). Each reaction was performed in triplicate. The annealing temperature was set at 61°C for both GABAAα6 and cyclophilin D, and 45 amplification cycles were performed. A standard curve was constructed to verify that the efficiency of the primers was between 100 ± 5%. The relative expression levels of the genes of interest were normalized against cyclophilin D housekeeping genes using the comparative cycle threshold (Ct) method. The difference between the mean Ct values of the gene of interest and the housekeeping gene is labeled ΔCt, and the difference in ΔCt between the calibrator and the test sample is labeled ΔΔCt. The log2 (ΔΔCt) is the relative gene expression value.
Animal injection procedures and cerebellar slice preparation
APP+/− mice (c57BL/6J) were purchased from the Model Animal Research Center of Nanjing University (Nanjing, Jiangsu, China). Homozygous APP−/− mice were obtained from APP+/− mice and genotyped by reverse transcription PCR (RT-PCR) using the Jackson Lab protocol. Five-day postnatal mice were injected with either Aβ40 solvent, 100 nM Aβ40 or p75NTR blocking antibody as previously described (Buffo et al. 2000). Acetic acid or 100 nM Aβ40 (0.2 μL) was injected into the dorsal vermisat a depth of 0.2–0.5 mm at 1–2 mm left and right of the cerebellar midline. The frequency and duration of the pressure pulses were adjusted such that 1 μL of the solution was injected over a period of 10 min. The pipette was left in situ for an additional 3 min to avoid excessive leakage of the injected solution. The injected animals were kept for 24 h and then killed. Frozen sections (14 μm thickness) were cut after embedding in OCT compound. The cerebellum was dissected sagittally at 300–500 μm away from the center joint, and each slice was cut at an interval of 45 μm. The slices were stained with thionin solution (Nissl staining) for 3–5 min. Images were acquired with a Carl Zeiss Axiovert 200 microscope equipped with an AxioCamMRc 5 camera and analyzed with AxioVision Rel. 4.8 software (Carl Zeiss, Oberkochen, Germany).
Recombinant rat Aβ40 and Aβ42 were purchased from Millipore Merck (Darmstadt, Germany) and resolved in 5% acetic acid according to the manufacturer's instructions. The proteins were stored at −20°C in 10–15-μL aliquots at a concentration of 100 μM. The aggregation state of the Aβ40 and Aβ42 stock solutions was assessed by studying protein migration patterns after tricine sodium dodecyl sulfate polyacrylamide gel electrophoresis (tricine-SDS-PAGE) and staining with Coomassie brilliant blue.
RT-PCR and semi-quantitative PCR
Total RNA was isolated from primary cultures of CGNs using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. First-strand synthesis was performed using SuperScript II reverse transcriptase (Invitrogen), and semi-quantitative PCR was performed using the Applied Biosystems Venti 96-well thermo cycler (Life Technologies, Grand Island, NY, USA). The samples were analyzed as duplicates of four independent runs. Five percent of each first-strand reaction was assayed by PCR using 0.5 μM oligonucleotides and 200 μM dNTPs. Amplification was performed using the following primer sets: p75NTRS (5′-AGGGCACATACTCAGACGAAG-3′) and R (5′-CAAGATGGAGCAATAGACAGGA-3′); GAPDHS (5′-ATCTTCTTGTGCAGTGCCAGCC-3′) and R (5′-GGTCATGAGCCCTTCCACGATG-3′).
Semi-quantitative PCR was performed in 20 μL reactions containing10 μL of 2x Taq Plus PCR Master Mix (DBI Bioscience, Shanghai, China), 0.2 μL each primer (10 μM), 7.6 μL of ddH2O and 2 μL of the RT product. The PCR conditions were as follows: 95°C for 1 min followed by 35 cycles of 95°C for 15 s, 61°C for 30 s and 72°C for 30 s with a final extension step of 72°C for 7 min. The products were separated on 1% agarose gels using Tris acetate buffer and visualized with ethidium bromide. Images were obtained and analyzed using a ChemiDoc XRS System and Image Lab software (Bio-Rad Laboratories).
The p75NTR blocking antibody 9651/9650 was kindly provided by Dr. Moses Chao (Skirball Institute, New York University, New York, NY, USA). The β-amyloid peptide antibody (6E10; Millipore) used for the Co-IP experiments was kindly provided by Dr. Huang (Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China). The brain-derived neurotrophic factor blocking antibody (2 mg/mL) was obtained from Millipore. Hydroxy-2-naphthalenylmethylphosphonic acid (HNMPA) was obtained from Santa Cruz Biotechnology. Insulin, furosemide, U0126, rapamycin, cycloheximide, α-bungarotoxin, k252a, and poly-l-lysine were purchased from Sigma (St. Louis, MO, USA). Fetal calf serum, DMEM, and the antibiotic–antimycotic solution were purchased from Gibco Life Technologies (Grand Island, NY, USA).The low-range unstained protein ladder 26632 was purchased from Thermo Scientific (Pierce Biotechnology). Nissl buffer was purchased from Beyotime (Shanghai, China).
Statistical analysis was carried out using two-sample Student's t-test for two samples or one-way analysis of variance (anova) followed by Tukey test and Fisher's least significant difference test for more than two samples with Originpro software (OriginLab Corporation, Northampton, MA, USA). Statistical significance was determined at *p < 0.05. The results are expressed as the mean ± standard error of the mean (SEM).
Aβ40, rather than Aβ42, increased the membrane expression of the GABAA receptor α6 subunit on CGNs
To address the effects of Aβ40 on the GABAA receptor α6 subunit, we first performed whole-cell patch clamping and current analysis. GABA-induced currents were recorded while the membrane potential remained at −70 mV. To estimate the contribution of the α6-containing receptors, we inhibited GABA-induced currents with 100 μM furosemide, which is a specific antagonist of GABAA receptors containing the α6 subunit (Engblom et al. 2003; Yamashita et al. 2006). We measured the decrease in current using the equation (I0−I1)/I0, where I0 and I1 represent the stable GABA receptor current before and after perfusion with furosemide, respectively. The furosemide-induced decrease in GABAA current is referred to as the ‘α6 current’. Pre-treatment of CGNs in culture for 5 days (5 DIC) with 100 nM Aβ40 for 24 h significantly increased the GABAAα6 current (Fig. 1a) from 22.1 ± 3.7% (n = 19) to 34.1 ± 4.1% (n = 19). We also tested the effects of 10 nM Aβ40, which increased the GABAAα6 current to 30.8 ± 3.7% (n = 12) without affecting the controls (22.1 ± 3.7%, n = 19). Therefore, we chose to use 100 nM Aβ40 in the subsequent experiments. However, we noted that 100 nM Aβ42 did not affect the GABAAα6 current, which remained at 20.2 ± 1.5% (n = 16) after treatment (Fig. 1b). Previous studies by Hepler (Hepler et al. 2006) indicated that the state of Aβ might result in differential effects; thus, we measured the state of aggregation of Aβ40 and Aβ42 by Coomassie blue staining after Tricine-SDS-PAGE. The results shown in Fig. 1c indicate a significant difference between the Aβ40 and Aβ42 states and reveal that Aβ42 is more vulnerable to aggregation because it stains at a higher molecular weight. This result suggests that the Aβ40 monomer may be pivotal in the Aβ40-induced up-regulation of the GABAAα6 current.
The Aβ40-mediated increase in GABAAα6 current was dependent on the number of days that the CGNs were cultured (Fig. 1d). After incubating 3 DIC CGNs with Aβ40 for 24 h, the GABAAα6 currents were 20.8 ± 2.4% (n = 16) and 24.0 ± 2.5% (n = 9) in the presence and absence of Aβ40, respectively, indicating that there was no significant difference between the two groups. When 5 DIC and 7 DIC CGNs were used, Aβ40 significantly increased the GABAAα6 currents from 22.1 ± 3.7% (n = 19) to 34.1 ± 4.1% (n = 19) and 21.6 ± 3.1% (n = 20) to 38.6 ± 4.8% (n = 19), respectively, indicating that there were significant differences between the control and Aβ40-treated groups.
We therefore investigated whether the Aβ40-mediated increase in GABAAα6 currents might be because of up-regulated channel expression. Western blotting data obtained from four independent experiments indicated that there was a significant increase in the protein levels of the GABAA α6 subunit following treatment with Aβ40 for 24 h (Fig. 1e). Furthermore, the changes in GABAA α6 protein expression occurred in parallel with the increase in GABAA α6 currents in 5 and 7 DIC CGNs. The levels of total GABAAα6 protein in 5 and 7 DIC CGNs significantly increased by 54.3 ± 12.3% (n = 5) and 92.5 ± 23.3% (n = 5), respectively. Meanwhile, we investigated whether the membrane-bound GABAAα6 proteins in the neurons treated with Aβ40 were affected. The surface proteins in the vehicle- or Aβ40-treated neurons were labeled using a biotinylation assay, and biotinylated GABAAα6 proteins were then detected by immunoblotting with an anti-GABAAα6 antibody. As shown in Fig. 1f, the GABAAα6 proteins in the membrane fractions were greatly increased by 70.7 ± 15.3% (n = 3) after Aβ40 treatment for 24 h. These findings suggest that Aβ40 up-regulates total cellular GABAAα6 levels and their associated currents.
Aβ40 induces the translation of GABAAα6
Gene expression changes in response to neurotrophins are generally thought to take place at the transcriptional or translational level; therefore, primers targeting GABAA α6 were used to measure its mRNA expression levels by Quantitative real-time PCR (qPCR) after incubation with and without Aβ40 for 24 h. qPCR revealed that the mRNA levels of GABAAα6 in 3, 5, and 7 DIC CGNs were not altered upon treatment with Aβ40 (Fig. 2a), suggesting that Aβ40 may increase GABAAα6 protein expression at the translational level.
To identify a potential increase in GABAAα6 translation, the effects of cycloheximide (CHX), which is a broad translation inhibitor, on the Aβ40-mediated increase in GABAAα6 current and protein expression were measured. Co-incubation of CGNs with Aβ40 plus CHX (10 μM) significantly reduced the Aβ40-mediated increase in GABAAα6 current from 29.6 ± 2.0% (n = 10) to 22.0 ± 1.6% (n = 16) (Fig. 2b). Western blotting data revealed that CHX (10 μM) significantly decreased the Aβ40-mediated increase in GABAAα6 protein expression from 156.2 ± 13.7% (n = 5) to 90.3 ± 9.1% (n = 5) (Fig. 2c). Thus, it is likely that Aβ40 induces GABAAα6 translation.
The ERK/mTOR pathway is involved in the Aβ40-mediated up-regulation of the GABAAα6 subunit
Previous studies revealed that the ERK pathway is involved in the regulation of GABAAα6 expression (Bulleit and Hsieh 2000). We further examined whether the ERK signaling pathway was activated by Aβ40 by measuring the levels of phosphorylated ERK (pERK) (Ji et al. 1999; Karim et al. 2001). Western blotting revealed that the phosphorylation of ERK1 (44 kDa) and ERK2 (42 kDa) was significantly increased upon treatment with Aβ40 (Fig. 3a); the levels of pERK1 and pERK2 relative to controls were 138.7 ± 14.5% (n = 4) and 132.7 ± 11.7% (n = 5), respectively. The effects of Aβ40 on ERK activation were confirmed with the specific MEK inhibitor U0126; 20 μM U0126 significantly reduced the levels of pERK1 and pERK2 to 11.5 ± 1.4% (n = 4) and 41.3 ± 7.8% (n = 5) of the controls. However, upon co-treatment with Aβ40 and U0126, there were no significant differences in the levels of pERK1 or pERK2 (9.0 ± 2.2% and 36.1 ± 9.4%, respectively) (Fig. 3b). Moreover, treatment with U0126 decreased the Aβ40-mediated induction of the GABAAα6 current from 34.1 ± 4.1% (n = 19) to 25.3 ± 3.0% (n = 6) and GABAAα6 subunit protein expression from 146.1 ± 14.4% to 92.6 ± 14.1% (n = 4) (Fig. 3c and d). These data suggest that the Aβ40-mediated increase in GABAAα6 current and subunit expression involves the activation of the ERK signaling pathway.
The mTOR pathway has been suggested to regulate mRNA translation downstream of the ERK pathway (Laplante and Sabatini 2012). We therefore measured whether the mTOR pathway was required for the Aβ40-mediated increase in GABAAα6 current and subunit protein expression. The phosphorylation of mTOR (pmTOR) was used as an indicator of mTOR activation. As shown in Fig. 4a and b, upon Aβ40 treatment, there was a significant increase in the relative pmTOR levels (193.7 ± 52.2% of the control, n = 4). This increase in pmTOR correlates with changes in GABAAα6 expression upon treatment with Aβ40. Treatment with rapamycin (50 nM), which is a selective inhibitor of the mTOR1 complex, decreased the expression levels of pmTOR to 76.5 ± 12.0% of the controls. In the presence of rapamycin, the Aβ40-mediated activation of mTOR remained at 68.2 ± 23.1% (n = 4). Moreover, blocking mTOR activity with rapamycin prevented the Aβ40-mediated increase in both GABAAα6 current and subunit expression (Fig. 4b and c). Rapamycin (50 nM) alone did not affect the GABAAα6 current (19.6 ± 1.7%, n = 14). In the presence of rapamycin, the Aβ40-mediated increase in GABAAα6 current decreased from 31.4 ± 4.0% (n = 9) to 21.8 ± 1.7% (n = 13). Similarly, the Aβ40-mediated increase in GABAAα6 protein expression decreased from 196.7 ± 34.0% (n = 6) to 82.0 ± 9.2% (n = 6) of the controls. These data indicate that the mTOR pathway is required for the Aβ40-mediated up-regulation of GABAAα6 subunit expression.
To address whether the activation of mTOR occurs downstream of the ERK pathway, ERK was inhibited with U0126, and the effects on the pmTOR levels were determined. The data in Fig. 4d indicate that in the presence of 20 μM U0126, the Aβ40-mediated increase in relative pmTOR levels decreased from 193.7 ± 52.2% to 83.5 ± 16.0% (n = 5), suggesting that activating the ERK pathway is necessary for pmTOR activation by Aβ40.
Aβ40 activates the ERK/mTOR pathway via p75NTR
Previous studies have reported that several receptors mediate Aβ-induced cytotoxic or neurotrophic effects (Verdier and Penke 2004). We chose to study two candidate receptors, the insulin receptor (IR) and the α7-nicotine receptor, both of which activate the ERK pathway. Our data indicate that blocking IR activity with HNMPA, which is a selective antagonist of IR, did not affect the Aβ40-mediated increase in GABAAα6 current and/or subunit protein levels (Fig. 5a and b). In the presence of HNMPA, the Aβ40-mediated increase in GABAAα6 current remained at 33.3 ± 2.7% (n = 12), and the relative GABAAα6 protein levels increased to 201.4 ± 24.6% (n = 6). These results are similar to the results obtained from CGNs treated with Aβ40 alone. Similarly, incubating CGNs with α-bungarotoxin (α-BGT, 100 nM), which is an antagonist of α7-nAChR, did not alter the Aβ40-induced up-regulation of the GABAAα6 subunit (Fig. 5a and c). In the presence of α-bungarotoxin, the Aβ40-mediated increase in GABAAα6 current remained at 34.1 ± 3.6% (n = 8), and the relative GABAAα6 protein levels increased to 186.1 ± 29.8% (n = 5) of the controls. These results indicate that there was no significant difference from the results obtained with CGNs treated with Aβ40 alone.
p75NTR has been reported to be related to neurodevelopment, neural death and survival (Friedman 2000; Lin et al. 2007; Geetha et al. 2012), and activation of p75NTR may mediate the effects of Aβ on neurons (Knowles et al. 2009; Sotthibundhu et al. 2009). We thus tested the effects of the 9651 antibody, which was reported to block the binding between Aβ and p75NTR, on Aβ-mediated induction of GABAAα6 subunit expression (Huber and Chao 1995; Knowles et al. 2009). CGNs were pre-treated with the 9651 antibody for 2 h and subsequently treated with Aβ40 or normal rabbit IgG for 24 h. Blocking p75NTR activity significantly reduced the Aβ40-mediated increase in both GABAAα6 current and protein expression from 31.2 ± 2.3% (n = 11) to 24.2 ± 2.0% (n = 16) and from 198.1 ± 15.7% to 144.0 ± 12.1% (n = 6), respectively (Fig. 6a and b). Similarly, the Aβ40-mediated increase in ERK1 and ERK2 phosphorylation was significantly decreased by the 9651 antibody to 61.6 ± 8.9% (n = 6) and 72.9 ± 17.6% (n = 5) of the controls, respectively (Fig. 6c). The pmTOR levels in the CGNs co-treated with Aβ40 and the 9651 antibody were significantly reduced from 143.0 ± 12.1% to 97.7 ± 8.8% (n = 6) (Fig. 6d). Taken together, these results suggest that Aβ40 may interact with p75NTR to mediate downstream signaling.
We performed co-IP experiments to further examine the interactions between Aβ40and p75NTR. CGNs were treated with Aβ40 or acetic acid (the solvent for Aβ40). The results in Fig. 6e reveal the interaction between p75NTR and Aβ40, as indicated by a clear band recognized by the p75NTR antibody at approximately 75 kD. We also measured the mRNA expression levels of p75NTR by semi-quantitative PCR. The data shown in Fig. 6f indicate that the mRNA levels of p75NTR in CGNs that were cultured from 5 to 7 DIC were significantly enhanced in parallel with the increase in GABAAα6 current and GABAAα6 protein expression after treatment with Aβ40.
Aβ40 improves CGN maturation in vivo
In vivo, GABAAα6 is not expressed until the cerebellar granule cells have matured in the internal granule layer. The increased expression of the α6 subunit of the GABAA receptor is generally the marker of CGN maturation (Zheng et al. 1993; Mellor et al. 1998). The migration of granule cells from the external granular layer (EGL) to the internal granular layer (IGL) is a morphological characteristic of a mature cerebellum. To assess the effects of Aβ40 on CGN maturation, APP−/− and wild-type mice (c57BL/6J) were used to compare the expression of GABAAα6 and the morphology of the cerebellum using Nissl's staining. Cerebella were obtained from APP−/− mice, APP+/+ mice and APP−/− mice that received cerebellar injections of Aβ40 (100 nM). An equal amount of 0.5% acetic acid (the solvent for Aβ40) was injected into mice as a control. The data in Fig. 7a show that the protein level of GABAAα6 was decreased in APP−/− mice (70.4 ± 5.7% of the control, n = 3). After Aβ40 injections, the expression of GABAAα6 was elevated to a normal level (106.4 ± 6.7%, n = 3) in APP−/− mice and to 127.8 ± 8.2% (n = 3, p < 0.05) in wild-type mice. The data in Fig. 7b reveal that the morphology of the cerebellum was also altered. APP−/− mice exhibited a thin and obscure IGL, whereas the APP+/+ mice showed a significantly thicker IGL. The injection of Aβ40 abolished this effect (Fig. 7b, n = 3). The morphology changes observed from APP−/− and wild-type mice treated with or without Aß were consistent in the same cerebellum. We also injected the p75NTR blocking antibody 9650 into the cerebella of APP−/− mice prior to applying Aβ40 to address whether p75NTR mediated the effect of Aβ40 on CGN maturation. The 9650 antibody was produced in the same way as 9651 but originated from another rabbit. After inhibiting p75NTR with 9650, Aβ40 no longer exerted effects on the thickness of the IGL in APP−/− mice (Fig. 7b, n = 3). Taken together, these results indicate that low concentrations of Aβ40 play an important role in the maturation of CGNs.
Here, we have shown for the first time that Aβ40 up-regulates the expression of the GABAA α6 subunit by inducing translation. This induction occurs through the activation of p75NTR and the ERK/mTOR signaling pathway. APP knockout mice express less GABAA α6 protein and have a thinner IGL in their cerebella compared with wild-type mice. Aβ40 increases the thickness of the IGL in APP−/− mice, and blocking p75NTR inhibits this effect. Thus, it is likely that the physiological effects of the Aβ40 peptide may be similar to the activity of neurotrophins, and Aβ40 may participate in neuronal development.
Aβ40 and Aβ42 are the two main Aβ isoforms found in brain tissue and have different physicochemical properties and functions (Walsh et al. 2005; Morley et al. 2010). In total, 90% of the Aβ peptides detected in human cerebrospinal fluid (CSF) are the Aβ40 isoform, whereas 10% are the Aβ42 isoform. Although the relative proportions of Aβ40 and Aβ42 are 50% each in Alzheimer's disease patients (Mehta et al. 2001), the effects of exogenous Aβ40 and Aβ42 on neuronal function are different. Notably, previous studies have indicated that exogenous Aβ42 has a cytotoxic effect, although synthetic Aβ42 monomers aid in the survival of neurons under conditions of trophic deprivation (Giuffrida et al. 2009). In contrast, exogenous Aβ40 has been shown to have greater neuroprotective effects compared with Aβ42, especially on CGNs (Ramsden et al. 2001; Plant et al. 2003; Arevalo et al. 2009). Furthermore, Aβ40 has been shown to be neuroprotective against Aβ42-induced neurotoxicity in vitro and in vivo (Zou et al. 2003). Our study indicates that Aβ40, rather than Aβ42, significantly increases GABAA α6 protein expression and current. This result is consistent with previous studies indicating that Aβ40 functions as a neurotrophin.
CGNs have been used as a model to study Aβ-induced neuronal death and protection. A study performed by Plant et al. reported that concentrations of Aβ40 as low as 10 pM alleviated the toxicity of secretase inhibitors in CGNs (Plant et al. 2003). A report by Ramsden et al. indicated that Aβ increased the A-type K+ current in CGNs by 60%, whereas no effects on any other K+ currents were observed in cortical neurons (Ramsden et al. 2001). In our study, 100 nM Aβ40 led to the maturation of 5–7 DIC CGNs by up-regulating the expression of the GABAAα6 subunit. These results reveal an unanticipated and previously unrecognized role for Aβ in CGNs and suggest that Aβ40 might have different functions during the development of CGNs.
Aβ has been reported to induce cytotoxic or neurotrophic effects by binding multiple membrane receptors such as the insulin receptor, a7-nAChR, NMDA-R and p75NTR (Verdier and Penke 2004). The results of our study indicate that p75NTR activation is involved in the Aβ40-mediated up-regulation of GABAAα6 in cultured CGNs. p75NTR has been reported to bind all neurotrophins and mediate neural death, survival and neurodevelopment (Friedman 2000; Lin et al. 2007; Geetha et al. 2012). For example, it is known that p75NTR mediates Aβ-induced cell death in neurons (Hashimoto et al. 2004; Longo et al. 2007; Knowles et al. 2009) and that this receptor contributes to neuronal survival during development and cerebellar developmental patterning (Ferri et al. 1998; Wiese et al. 1999; Carter et al. 2003; Howard et al. 2013). These differing functions may result from differences in intracellular mediators and co-receptors between developing and adult cells (Ibanez and Simi 2012). We used RT-PCR to examine p75NTR mRNA expression, and our data were similar to those of a reporter-based study in which Aβ induced Ras and ERK activation in cerebellar neurons expressing p75NTR (Susen and Blochl 2005). Interestingly, p75NTRmRNA levels were lower in 3DIC CGNs than in 5–9 DIC CGNs. These results suggest that the expression of p75NTR in CGNs changes during development and explains why Aβ40 did not affect the GABAAα6 current or protein expression in 3DIC CGNs.
The results of this study indicate that ERK and mTOR are involved in the Aβ40-induced up-regulation of GABAAα6. ERK and mTOR signaling are both involved in neuronal development (Lhuillier and Dryer 2000; Kelleher et al. 2004; Swiech et al. 2008) and are modulated in response to various stimuli that lead to the transcriptional and translational control of gene expression. Previous studies have indicated that p75NTR activates different intracellular signaling events and pathways, including NFκB translocation, Jun kinase phosphorylation, PI3K phosphorylation and the MEK/ERK pathway, in a cell-type- and cell-state-dependent manner (Carter et al. 1996; Yoon et al. 1998; Meldolesi et al. 2000; Susen and Blochl 2005). Interestingly, the Aβ40-mediated induction of GABAAα6 in this study appeared to be stimulated by p75NTR rather than by the insulin receptor, which is generally known to be the upstream receptor of mTOR. It is possible that Aβ40 binds p75NTR to activate the MEK/ERK pathway, which then induces the phosphorylation of mTOR. Because inhibiting ERK with U0126 reduced the Aβ40-mediated induction of pmTOR phosphorylation, our results suggest that activating the ERK pathway is necessary for the Aβ40-mediated phosphorylation of mTOR.
The cerebellum is one of the few regions of the rat brain that undergoes profound morphological changes after birth. In particular, during the first 2–3 weeks of post-natal life, the immature CGNs generated in the EGL migrate through the molecular layer (ML) to reach their final destination within the IGL, where they complete their maturation (Komuro and Rakic 1998) and express GABAAα6 (Zheng et al. 1993; Mellor et al. 1998). The regulation of GABAAα6 by Aβ40 has been confirmed in vivo. We thus investigated the effects of Aβ40 on morphological changes in the cerebellum layers in APP−/− mice. In our experiments, the IGL layer in the APP−/− mice was thinner than that in wild-type mice. Notably, the morphological changes induced by APP depletion were reversed by the injection of Aβ40 into the cerebellum, suggesting that Aβ40 is important for CGN maturation and the process of IGL formation. However, APP and other APP products, such as the APP intracellular domain (AICD), have been reported to regulate neuronal function (Leissring et al. 2002) and therefore cannot be ruled out as potential causes of the thinner IGL in the APP knockout mice. The results obtained with cultured rat cerebellar slices treated with Aβ40 were in accordance with the IGL morphology in the cerebellum, which further supports our hypothesis.
In conclusion, our study suggests for the first time that low concentrations of Aβ40 significantly induce GABAAα6 subunit expression and likely have physiological effects on CGN development. Although it is clear that GABAAα6 is one of the markers of CGN differentiation and maturation, the effects of Aβ40 on other biological markers, such as the NMDA receptor subunit switch (Hogberg et al. 2009) and the transcription factor Neuro D2 (Olson et al. 2001), warrant further investigation.
This work was supported by the National Natural Science Foundation of China (NSFC 31070745) and the Shanghai Leading Academic Discipline Project (B111). The P75 receptor blocking antibody 9651/9650 was kindly provided by Dr. Moses Chao (New York University). The β-amyloid peptide antibody used in the co-IP experiments was a gift from Dr. Fude Huang (Shanghai Institutes for Biological Sciences, China). The authors also thank Dr. Huang for valuable comments on the manuscript.
The authors declare that they have no conflict of interest.