Proteomic analysis involved with synaptic plasticity improvement by GABAA receptor blockade in hippocampus of a mouse model of Alzheimer’s disease

GABAergic system plays a part in synaptic plasticity in the hippocampus. We had reported a long-term potentiation (LTP)-like facilitation in vivo, known as synaptic plasticity, through GABAA receptor blockade by bicuculline and the expression of proteins involved with this synaptic plasticity in mouse hippocampus. In the present study, we aimed to show improvement of impaired synaptic plasticity through GABAA receptor blockade and to clarify the molecular mechanisms involved with this improvement in the hippocampus of mice overexpressing human amyloid precursor protein with the E693Δ mutation (APPOSK-Tg) as an Alzheimer's disease model showing impaired synaptic plasticity. Electrophysiological study showed that the LTP-like facilitation expressed with application of bicuculline in vivo was significantly greater than impaired tetanic LTP in APPOSK-Tg mice, which was improved by bicuculline. Proteomic analysis showed that the expression of 11 proteins in the hippocampus was significantly changed 8 h after bicuculline application to APPOSK-Tg mice. The identified proteins could be functionally classified as chaperome, cytoskeletal protein, energy metabolism, metabolism, neuronal development, and synaptic component. Additionally, western blotting validated the changes in four proteins. We therefore propose that the improvement of impaired synaptic plasticity through GABAA receptor blockade could be mediated by the changed expression of these proteins.


Introduction
Alzheimer's disease (AD) is a progressive, age-related dementia of unknown etiology. Diverse lines of evidence suggest that amyloid ␤ (A␤) plays important roles in AD pathogenesis (LaFerla and Oddo, 2005). Especially, soluble A␤ oligomers could be a cause of synaptic and cognitive dysfunction by inhibiting long-term potentiation (LTP) and accelerating neuronal cell death in AD (Lambert et al., 1998;Shankar et al., 2008). We identified the E693 mutation in amyloid precursor protein (APP) in patients with AD who displayed almost no signals of amyloid plaques in amyloid imag-central GABAergic processes on memory consolidation is suggested: GABA A receptor blockades enhance retention of aversively motivated tasks (Brioni and McGaugh, 1988), active avoidance task (Yonkov and Georgiev, 1985) and inhibitory avoidance task (Castellano and Pavone, 1988). LTP, an example of synaptic plasticity that may contribute to learning and memory functions, at excitatory glutamatergic synapses is paralleled by long-term depression at inhibitory GABAergic synapses in the hippocampus (Wingstrom and Gustafsson, 1983;Maroun and Richter-Levin, 2002). LTP induction in hippocampal slice preparations in vitro was facilitated by bicuculline (Wingstrom and Gustafsson, 1983), and local application of bicuculline to the rat dentate gyrus caused a persistent enhancement of the population spike in vivo (Maroun and Richter-Levin, 2002). We also showed that peripheral application of bicuculline expressed an LTP-like facilitation in vivo in the mouse dentate gyrus (Matsuyama et al., 2008). Furthermore, proteomic analysis during the expression of LTP-like facilitation in the hippocampus demonstrated the levels of 15 proteins significantly changed , in agreement with the observation that late-phase LTP induced with chemical stimulation is dependent on new gene expression (Kelleher et al., 2004). Taken together, a persistent activation of neural circuit through glutamatergic function stimulated by GABAergic depression in the hippocampus could contribute to the restoration from impaired synaptic plasticity of AD. Furthermore, to elucidate the precise mechanism of this restoration, proteomic analysis during restoration of synaptic plasticity will be required.
In this study, we examined the effects of GABA A receptor blockade, bicuculline, on the restoration from impaired synaptic plasticity in the dentate gyrus of APP OSK -Tg mice, and changes in the expression of hippocampal proteins to elucidate the molecular mechanisms involved in this restoration using two-dimensional gel electrophoresis followed by mass spectrometry.

Animal subjects
Transgenic mice overexpressing human APP 695 with the E693 mutation (APP OSK -Tg) under the mouse prion promoter (Tomiyama et al., 2010) were housed under standard illumination conditions (12-h light/dark cycle; light on at 7:00 A.M.) and given free access to food and water. Male heterozygous APP OSK -Tg mice at 13-months of age were used. All animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the experimental protocol was approved by Kobe University Graduate School of Medicine and Himeji Dokkyo University Animal Experiment Committee. All efforts were made to minimize animal discomfort and the number of animals used.

LTP recording in mouse dentate gyrus in vivo
Experiments were performed on mice in vivo prepared as described previously (Namgung et al., 1995). Briefly, mice were anesthetized by injecting urethane (1.2 g/kg, intraperitoneally followed by supplemental injections of 0.2−0.6 g/kg as needed) and placed in the stereotaxic apparatus. Body temperature was maintained at 37 • C using a heated mat (BRC, Nagoya, Japan). A glass recording electrode with 9−12 m tip diameter, back-filled with saline, was lowered to the cell body layer of dentate granule cells. Initial responses were obtained using a cathodal stimulation (6.0-8.0 V, 0.1 Hz, 0.1 ms duration) of the perforant path. After electrode insertion and population responses were obtained, the preparation was allowed to stabilize for 60 min prior to baseline recording. Voltage was reduced so that baseline spike amplitude was one-third the maximum asymptotic value. The LTP-inducing voltage used was the lowest voltage level that could evoke a maximum asymptotic spike amplitude. The parameters for tetanus (HFS, high frequency stimulation) to induce LTP were three trains with an intertrain interval of 10 s, and each train consisted of eight 0.4-ms 400-Hz pulses. The amplitude of the population spike was measured from initial positivity to peak negativity. At 5 min intervals, the population spikes induced by five successive stimulations were averaged and analyzed with a personal computer (PowerLab System; BRC). Bicuculline was first dissolved in a few drops of 0.1 N HCl, after which the final volume was made up with sterile saline. Bicuculline at a volume of 0.1 ml was injected intraperitoneally at the indicated dose and time.
Data are expressed as the mean ± S.E.M. of n values. Multiple group means were compared using the Bonferroni/Dunn test following one-factor or two-factor factorial analysis of variance (ANOVA). Differences with a P value of less 0.05 were considered statistically significant.

Sample preparation
Under pentobarbital sodium anesthesia administered by intraperitoneal injection, APP OSK -Tg mice treated with bicuculline or vehicle were killed at 8 h after application. Sample preparation was performed as previously described (Matsuura et al., 2016). In brief, isolated mouse hippocampal homogenate was prepared by sonication for 30 s in 250 l of lysis buffer containing 7 M urea, 2 M thiourea, 5 % CHAPS, 2 % IPG (immobilized pH gradient) buffer (GE Healthcare UK Ltd., Buckinghamshire, UK), 50 mM 2mercaptoethanol, 2.5 g/mL DNase I, and 2.5 g/mL RNase A. The supernatant was obtained by centrifugation at 15,000 × g for 30 min to remove cellular debris, and was recovered for use in 2-DE.

Two-dimensional electrophoresis
Two-DE was carried out by a method previously described (Matsuura et al., 2016). Briefly, after sample preparation, Immobiline Dry Strips (IPG strips) pH 4-7NL (7 cm) (GE Healthcare, Madison, WI) were used as a first-dimension isoelectric focusing gel. Approximately 1000 g of protein from each animal was separated at 50 V for 6 h, at 100 V for 6 h, and at 2000 V for 6 h. The IPG strips were equilibrated for 15 min in 50 mM Tris−HCl (pH 8.8), 6 M urea, 30 % (v/v) glycerol, 1 % SDS, and 1 % (w/v) DTT, and then for 15 min in the same buffer with 2.5 % (w/v) iodoacetamide instead of DTT. After equilibration, the gels were subjected to a seconddimensional electrophoresis on 12.5 % SDS-polyacrylamide gels at 5 mA/gel for 7 h.

SYPRO ruby staining
Proteins on SDS-polyacrylamide gels were detected using SYPRO Ruby Protein Gel Stain (Thermo Fisher Scientific, MA) as previously described (Matsuura et al., 2016). Gels after 2-DE were fixed in a solution containing 10 % acetic acid / 50 % methanol for 30 min, then 7 % acetic acid / 10 % methanol for 30 min. After fixing, the gels were incubated in the undiluted stock solution of SYPRO Ruby for 90 min, and destained with 7 % acetic acid / 10 % methanol for 30 min. After rinsing with H 2 O for 10 min, digital images were acquired using a FluoroPhoreStar 3000 image capture system (Anatech, Tokyo, Japan) with 470 nm excitation and 618 nm emission for SYPRO Ruby detection.

Image analysis
Image analysis was performed as previously described (Matsuura et al., 2016). Following image acquisition, 2-DE gel imaging and analysis software Prodigy SameSpots (Nonlinear Dynamics, Newcastle upon Tyne, UK) version 1.0 was used for gel-to-gel matching and identifying differences between bicuculline-treated and non-treated APP OSK -Tg mouse samples. Each of five sets of samples was represented by two independent biological replicates of 2-DE gels. The gel images were normalized in the Prodigy SameSpots software to normalize differences in staining intensities among gels. ANOVA was performed with 95 % significance level to determine which proteins were differentially expressed between bicuculline-treated and non-treated APP OSK -Tg mice. A minimum of 1.2-fold change was considered to identify the increased proteins and 0.9-fold change for decreased proteins.

In-gel digestion and peptide extraction
In-gel digestion was performed in accordance with the method of Yokoyama et al. (Yokoyama et al., 2004). Protein spots in the gels stained with SYPRO Ruby were cut out and subjected to trypsin digestion with porcine trypsin (Promega, Madison, WI). Briefly, gel pieces were washed with 200 l of 25 mM ammonium hydrogen carbonate with 5 % acetonitrile (v/v), and then dried under vacuum. Trypsin (5 l, 10 ng/l) was added and the digestion was incubated for 18 h at 37 • C. After separation of supernatant, gel pieces were washed again and then extracted with 50 % (v/v) acetonitrile / 0.3 % (v/v) trifluoroacetic acid for 10 min by sonication. The supernatant was once again collected, mixed with the two fractions, and evaporated under vacuum.

Mass spectrometry analysis and protein identification
Mass spectrometry (MS) analysis was performed according to a previously described (Matsuura et al., 2016). Mass spectra were recorded in positive reflection mode using a MALDI-TOF MS/MS analyzer (ABI PLUS 4800, Applied Biosystems, Foster City, CA), equipped with delayed ion technology. The samples were dissolved in 5 l of 50 % (v/v) acetonitrile / 0.3 % (v/v) trifluoroacetic acid. For the matrix, 1 g/l ␣-cyano-4-hydroxycinnamic acid (Wako Pure Chemical Industries) dissolved in the same mixture was used. Analyte and matrix were spotted consecutively in a 1:1 ratio on a stainless steel target and dried under ambient conditions. All spectra acquired by MALDI-TOF MS were externally calibrated with peptide calibration standard II (Bruker Daltonics, Bremen, Germany). An MS condition of 2500 shots per spectrum was used. Automatic monoisotopic precursor selection for MS/MS was performed using an interpretation method based on the 12 most intense peaks per spot, with an MS/MS mode condition of 4000 laser shots per spectrum. Minimum peak width was one fraction and mass tolerance was 80 ppm. Adduct tolerance was (m/z) ± 0.003. MS/MS was performed with a gas pressure of 1 × 10 −6 bar in the collision cell. Ambient air was used as collision gas. Data analyses were performed using Data Explorer version 4.9 software (Applied Biosystems), and proteins were identified through the search engine Mascot (http://www.matrixscience.com; Matrix Science, Boston, MA) (peptide mass tolerance: 60 ppm; MS/MS tolerance: 0.3 Da; maximum missed cleavages: 1) using the protein database NCBInr. Proteins identified by MALDI-TOF MS with a score of 79 or higher were considered significant (P < 0.05).

Results
3.1. Improvement of impaired tetanic LTP in vivo in the dentate gyrus of APP OSK -Tg mice through GABA A receptor blockade APP OSK -Tg mice showed an impaired tetanic LTP in the dentate gyrus in vivo at 8 months (Tomiyama et al., 2010). In these mice, the LTP-like facilitation expressed by bicuculline at a dose of 1 mg/kg was significantly greater than this impaired tetanic LTP, and the mean population spike of 5−120 min period was 207 % or 158 % of baseline spike amplitude in the LTP-like facilitation or tetanic LTP, respectively (P < 0.05, Bonferroni / Dunn test, n = 5 for each group, Fig. 1A). The application of bicuculline (1 mg/kg) 60 min after tetanic stimulation significantly enhanced the population spike of tetanic LTP, and the mean population spike of 5-60 min period or the mean population spike of 65-120 min period was 162 % or 216 % of baseline spike amplitude in the dentate gyrus of APP OSK -Tg mice, respectively (P < 0.05, Bonferroni / Dunn test, n = 5 for each group, Fig. 1B). On the other hand, the application of tetanic stimulation 60 min after the application of bicuculline did not change the population spike of the LTP-like facilitation by bicuculline (1 mg/kg), and the mean population spike of 5-60 min period or the mean population spike of 65-120 min period was 232 % or 231 % of baseline spike amplitude, respectively (P > 0.05, Bonferroni / Dunn test, n = 5 for each group, Fig. 1B).

Identification of altered proteins in the hippocampus of bicuculline-treated APP OSK -Tg mice
Proteomic analysis was performed to examine the change in levels of proteins in the hippocampus of bicuculline-treated and non-treated APP OSK -Tg mice. Proteins were quantified and identified from 2-DE gels using Prodigy Same Spot software and Fig. 1. Improvement of impaired tetanic LTP in vivo in the dentate gyrus of APPOSK-Tg mice through bicuculline. (A) Tetanic LTP, but not LTP-like facilitation by bicuculline, was impaired in the dentate gyrus of APPOSK-Tg mice. Open squares represent that tetanic stimulation (HFS) at 0 min was applied. Closed circles represent that bicuculline (1 mg/kg) at 0 min was injected intraperitoneally. (B) Bicuculline improved impaired tetanic LTP in the dentate gyrus of APPOSK-Tg mice. Open squares represent that tetanic stimulation (HFS) at 0 min was applied and bicuculline (1 mg/kg) at 60 min was injected intraperitoneally. Closed circles represent that bicuculline (1 mg/kg) at 0 min was injected intraperitoneally and tetanic stimulation (HFS) at 60 min was applied. The time course of the action of bicuculline and/or tetanic stimulation (HFS) on the population spike, n = 5 for each group, is shown. Each point represents the mean ± S.E.M. percentage of basal population spike amplitude at 0 min.
MALDI-TOF MS/MS. The levels of more than 400 spots significantly changed in representative 2-DE gels stained with SYPRO Ruby for the hippocampus of bicuculline-treated and non-treated APP OSK -Tg mice (Fig. 2). Finally, the expressed levels of 16 proteins signifi- cantly increased and the expressed levels of 2 proteins significantly decreased ( Table 1). The proteins with increased expression levels were identified as heat shock protein 8 (HSP8, also known as Hsc70), tubulin alpha-1C chain, Atp5b protein, aldehyde dehydrogenase, mitochondrial precursor (ALDH2), guanine deaminase (GDA), dihydropyrimidinase-related protein 2 (DPYSL2), V-ATPase subunit A and subunit B, and unnamed protein product (Table 2), and with decreased expression levels were identified as tubulin beta-5 chain and secernin-1. These proteins were subsequently categorized according to their cellular functions using MOTIF (http:// www.genome.jp/tools/motif/) database (Table 2).

Validation for the altered proteins in the hippocampus of bicuculline-treated APP OSK -Tg mice
We performed western blot analysis to validate the identity of HSP8, DPYSL2, V-ATPase subunit A, and secernin-1 as differentially expressed proteins (Fig. 3A). The protein levels of HSP8 and DPYSL2  Proteins of the mouse hippocampus were separated by 2-DE followed by in-gel digestion with trypsin and identified by MALDI-TOF MS/MS. The spots representing the identified proteins are indicated in Fig. 2 and are designated by their gene ID accession numbers (GI acc. no.) used in the NCBI non-redundant (NCBInr) protein database. Scores that relate to the probability assignment, molecular weight, pI, and sequence coverage (SC) are given. The score and sequence coverage were obtained by Mascot database searching (http://www.matrixscience.com). The spot densities were compared with control values. P values were obtained by ANOVA, P < 0.05. were significantly increased to 155 % and 145 %, respectively, and of secernin-1 was significantly decreased to 75 % in the hippocampus of bicuculline-treated mice as compared with non-treated mice (Fig. 3B). The protein level of V-ATPase subunit A was increased to 143 % as compared with non-treated mice (Fig. 3B).

Discussion
This study assessed the beneficial effects of GABA A receptor blockade by bicuculline on impaired synaptic plasticity in the hippocampus of APP OSK -Tg mice showing impaired tetanic LTP and enhanced A␤ oligomerization without fibrillization (Tomiyama et al., 2010). We showed that intraperitoneal application of bicuculline (1 mg/kg) recovered impaired tetanic LTP in the dentate gyrus of APP OSK -Tg mice (Fig. 1) and that the expression level of 11 proteins in the hippocampus of APP OSK -Tg mice significantly changed 8 h after bicuculline application (Fig. 2, Tables 1 and 2). Thus, these proteins might play important roles in the improvement of impaired learning and memory functions of AD.

Effects of bicuculline on impaired LTP in the dentate gyrus of APP OSK -Tg mice
We have shown that bicuculline induces the LTP-like facilitation in vivo in the mouse dentate gyrus in a dose-dependent manner (Matsuyama et al., 2008). As shown in Fig. 1, the LTP-like facilitation in vivo induced by bicuculline was significantly greater than impaired tetanic LTP, and bicuculline recovered tetanic LTP impaired in the dentate gyrus of APP OSK -Tg mice. Similarly, we have shown that bicuculline shows the same effects on impaired LTP in the dentate gyrus of transgenic mice overexpressing human tau protein with the N279 K mutation (Matsuyama et al., 2008) as a mouse model of AD (Taniguchi et al., 2005). Taken together, it is suggested that bicuculline could ameliorate impairment of synaptic plasticity in AD model mice, providing a clue for the clinical use of GABA A receptor blockades to improve cognitive disorders.

Identification of altered proteins in the hippocampus of bicuculline-treated APP OSK -Tg mice
Previous proteomics studies showed significantly altered levels of proteins, including those involved in synaptic plasticity, neurite outgrowth, microtubule dynamics, pH regulation, and antioxidant function in the cortex or hippocampus of AD model mice (Martin et al., 2008) and the hippocampus of AD patients (Sultana et al., 2007). We previously showed altered expression levels of cytoskeletal and their interacting proteins, chaperone and their interacting proteins, energy metabolism, synaptic component, vesicle transport, and recycling and signaling proteins in the hippocampus of APP OSK -Tg mice (Takano et al., 2012. Also, we were interested in the involvement of GABA A receptor blockade in synaptic plasticity and examined the expression of proteins changed during bicuculline-induced LTP-like facilitation in the mouse hippocampus. We showed altered expression levels of effectors of cellular functions including neuronal differentiation, cytoskeletal dynamics, folding of proteins, stress response, energy metabolism, synapse formation, and unknown function in the mouse hippocampus during enhanced synaptic plasticity in vivo induced by bicuculline . In the present study, we found that the expression of 11 proteins significantly changed during the process of recovery from impaired synaptic plasticity by GABA A receptor blockade bicuculline in the hippocampus of APP OSK -Tg mice (Table 1). These identified proteins were classified into several groups on a functional basis as shown in Table 2. Interestingly, HSP8, tubulin alpha-1C chain, Atp5b protein, and DPYSL2 were identified from the areas having more than two protein spots on 2-DE gels stained with SYPRO Ruby staining (Fig. 2,  Tables 1 and 2). The spot shift of these proteins may imply translational modifications. Together, DPYSL2 and V-ATPase subunit A were changed during synaptic plasticity induced by bicuculline in the hippocampus of both wild-type and APP OSK -Tg mice.

Chaperone
HSP8 acts as a molecular chaperone that induces protein folding through binding to nascent polypeptides (Stricher et al., 2013). HSP8 has been shown to inhibit ␣-synuclein fibril formation (Luk et al., 2008), protect against ␣-synuclein toxicity in vitro (Klucken et al., 2004), and to reduce the amount of ␣-synuclein aggregates in vivo (Klucken et al., 2004). Furthermore, the complex that is formed by HSP8 and phosphorylated tau is ubiquitinated by E3 Ub ligase carboxyl terminus of the Hsc70-interacting protein (CHIP), and then degraded through the proteasome system, leading to attenuation of phosphorylated tau-induced cell death (Shimura et al., 2004). Our findings showed that HSP8 was significantly increased in the hippocampus of bicuculline-treated APP OSK -Tg mice (Tables 1 and 2). Taken together, this increased expression of HSP8 could attenuate degraded changes of cells by reducing levels of aggregated ␣-synuclein or tau proteins.

Cytoskeletal proteins
Cytoskeletal proteins are responsible for cell motility, movement of organelles and vesicles through the cytoplasm, establishment of the intracellular organization of the cytoplasm, and many other functions that are essential for cellular homeostasis and survival. In the nervous system, cytoskeletal proteins such as microtubules, neurofilaments, and microfilaments are a vital, dynamic component, and regulate several types of functions. The expression levels of cytoskeletal proteins such as tubulin and actin were altered in mutated APP transgenic mice . Tubulin has been shown to be involved in the transport of membrane-bound organelles and is required for extension and maintenance of neuritis (Stiess et al., 2010). Tubulin composed of ␣ and ␤ chains is the major protein of microtubules, which play a pivotal role in cytoskeletal maintenance (Verhey and Gaertig, 2007) and regulate dendritic spine morphology and synaptic plasticity (Jaworski et al., 2009). Recent studies reported that total expression levels of ␣-tubulin and the number of ␣-tubulin-positive axonal processes were decreased in the brain of patients with AD (Zhang et al., 2015). In the present study, we showed that the expression level of tubulin alpha-1C and beta-5 chains were also significantly altered in the hippocampus of bicuculline-treated APP OSK -Tg mice (Tables 1 and 2). Thus, morphological dynamics of synapses in synaptic plasticity could depend on changes in the expression of tubulin.

Energy metabolism
The brain is a heavy user of energy metabolic systems such as glycolytic pathway. Energy metabolic insufficiency has been proposed to be involved in the pathological processes of AD (Hoyer, 2004), and the expression of some genes involved in energy metabolism such as Atp5b protein has been shown to be significantly downregulated in the hippocampus of AD patients (Brooks et al., 2007;Liang et al., 2008). Atp5b protein is a subunit of mitochondrial ATP synthase that produces ATP in the presence of an electrochemical gradient of protons across the inner membrane during oxidative phosphorylation. ATP acts as a synaptic transmitter, and is released from pre-synapse together with noradrenaline, acetylcholine, or other substances in the central nervous system (Zimmermann, 1994). Furthermore, ATP causes facilitation of excitatory post-synaptic responses for a prolonged period (Wieraszko and Seyfried, 1989;Fujii et al., 1995;O'Kane and Stone, 2000). In the present study, we showed that the expression level of Atp5b protein was significantly increased in the hippocampus of bicucullinetreated APP OSK -Tg mice (Tables 1 and 2). Taken together, the increased expression of Atp5b protein could improve the impaired synaptic plasticity of AD pathology by increasing ATP level.

Metabolism
ALDH2 is important for the detoxification of endogenous toxic aldehydes such as lipid peroxidation product 4-hydroxynonenal (HNE) that are known to accumulate in the brain of neurodegenerative disease patients (Singh et al., 2010). Several studies reported that loss of ALDH2 leads to HNE accumulation in the brain and memory loss of mice compared with age-matched littermates (Ohsawa et al., 2008). HNE-adducted proteins are increased in the brain of patients with AD (Sayre et al., 1997). Furthermore, gene expression level of ALDH2 is decreased in the hippocampus of AD model mice (Griñán-Ferré et al., 2016). In the present study, we showed that the expression level of ALDH2 was significantly increased in the hippocampus of bicuculline-treated APP OSK -Tg mice (Tables 1 and 2). Taken together, the increased expression of ALDH2 could improve the impaired synaptic plasticity through preventing neuronal cell death by reducing toxic aldehydes.

Neuronal development
GDA, also known as cypin, an enzyme that converts guanine to xanthine, dramatically increases dendritic branching in the central nervous system (Akum et al., 2004;Chen and Firestein, 2007;Fernández et al., 2009). Postsynaptic density protein 95 (PSD-95) and GDA regulate the patterning of dendrites in the hippocampal neurons (Tseng and Firestein, 2011). GDA increases dendrite branching by promoting microtubule polymerization and negatively regulating PSD-95 clustering (Akum et al., 2004;Firestein et al., 1999). Accordingly, GDA and PSD-95 may be involved in the machinery that maintains dendrite integrity and promotes spine retraction. Activity-dependent dynamic changes in spines of hippocampal pyramidal neurons are essential for learning and memory (Yuste and Bonhoeffer, 2001).
DPYSL2, also known as CRMP2, is the most abundant of five mammalian CRMP subunits in the developing nervous system (Charrier et al., 2003). CRMP2 is highly expressed in neuronal tissues of plasticity and/or neurogenesis areas such as pyramidal cells of the hippocampus and olfactory bulb during the middle embryonic to early postnatal period, but it is constitutively expressed at a reduced level throughout adulthood (Charrier et al., 2003;Wang and Strittmatter, 1996). Also, CRMP2 plays important roles in the structural and regulatory functions related to cytoskeletal dynamics, vesicle trafficking, and synaptic physiology in the developing brain and cultured hippocampal neurons (Arimura and Kaibuchi, 2007). Importantly, we previously showed that CRMP2 present in the hippocampus of adult mice is increased during synaptic plasticity induced by nicotine (Kadoyama et al., 2015). Furthermore, proteomic analysis during the expression of LTP-like facilitation induced by bicuculline showed that CRMP2 is increased in the hippocampus of wild-type mice . Taken together with our present findings, the increased expression of GDA and CRMP2 in the hippocampus could recover impaired synaptic plasticity of AD pathology.

Synaptic component
V-ATPase hydrolyzes ATP to pump protons across cell membranes, and acidification of the synaptic vesicles by V-ATPase is essential for loading with neurotransmitter (Di Giovanni et al., 2010). Also, a direct interaction between V-ATPase and vesicleassociated membrane protein 2 is involved in a late step of neurotransmitter release (El Far and Seagar, 2011). A recent review summarized that mutations of V-ATPase components underlie neurodegenerative disorders, including AD (Colacurcio and Nixon, 2016). Furthermore, disruption of ATP6AP2, which encodes for a critical V-ATPase-regulating protein, leads to neurodegeneration and cognitive impairment in both fly and mouse models (Dubos et al., 2015). Taken together, the increased expression of V-ATPase subunit A and subunit B in the hippocampus could improve the impaired synaptic plasticity by increasing the neurotransmitter release, resulting in long-lasting synaptic plasticity.

Conclusion
We found that GABA A receptor blockade by bicuculline improved impaired synaptic plasticity in the dentate gyrus of APP OSK -Tg mice as a model of A␤ oligomer-dependent AD, and that the expression levels of 11 proteins were significantly changed for this recovering process in the hippocampus. Thus, we propose that these proteins play important roles in the restoration from impaired synaptic plasticity, and that GABA A receptor blockade might contribute to AD cure.