Amyloid β-Mediated Zn2+ Influx into Dentate Granule Cells Transiently Induces a Short-Term Cognitive Deficit

We examined an idea that short-term cognition is transiently affected by a state of confusion in Zn2+ transport system due to a local increase in amyloid-β (Aβ) concentration. A single injection of Aβ (25 pmol) into the dentate gyrus affected dentate gyrus long-term potentiation (LTP) 1 h after the injection, but not 4 h after the injection. Simultaneously, 1-h memory of object recognition was affected when the training was performed 1 h after the injection, but not 4 h after the injection. Aβ-mediated impairments of LTP and memory were rescued in the presence of zinc chelators, suggesting that Zn2+ is involved in Aβ action. When Aβ was injected into the dentate gyrus, intracellular Zn2+ levels were increased only in the injected area in the dentate gyrus, suggesting that Aβ induces the influx of Zn2+ into cells in the injected area. When Aβ was added to hippocampal slices, Aβ did not increase intracellular Zn2+ levels in the dentate granule cell layer in ACSF without Zn2+, but in ACSF containing Zn2+. The increase in intracellular Zn2+ levels was inhibited in the presence of CaEDTA, an extracellular zinc chelator, but not in the presence of CNQX, an AMPA receptor antagonist. The present study indicates that Aβ-mediated Zn2+ influx into dentate granule cells, which may occur without AMPA receptor activation, transiently induces a short-term cognitive deficit. Extracellular Zn2+ may play a key role for transiently Aβ-induced cognition deficits.


Introduction
Memory function normally declines along with aging, and is believed to deteriorate initially because of changes in synaptic function rather than loss of neurons [1]. Some individuals then go on to develop Alzheimer's disease with progressive neurodegeneration [2][3][4]. Neuronal loss is not observed in animal models overexpressing amyloid-b (Ab) or infused with Ab, which develop early synaptic alterations but lack the extensive cell death [5,6]. One of the pathophysiological mechanisms of Ab is to alter synaptic functions such as longterm potentiation (LTP) and long-term depression (LTD), which are believed to be cellular mechanisms of learning and memory [7,8]. It has been reported that soluble Ab oligomers inhibit LTP [7,9] and facilitate LTD [10]. Thus, the cognitive deficit and memory loss occurring before any prominent neuronal loss, which is observed in patients with mild cognitive impairment (MCI) or earlyphase Alzheimer's disease, may be associated with Ab-induced synapse dysfunctions such as alterations in LTP [11].
Approximately 20-40% of cognitively normal elderly people have evidence of significant brain Ab-plaque deposition, either from amyloid imaging or CSF Ab42 concentrations [12][13][14]. Ab peptides derive from cleavage of the Ab precursor protein (APP) [15]. APP is initially cleaved by aor b-secretases, generating large, soluble, secreted fragments (sAPPa and sAPPb) and membrane-associated carboxy-terminal fragments (CTFs). Ab peptides of various lengths (e.g., Ab , 142 and Ab , 140) are produced after b-secretase cleavage, followed by c-secretase cleavage. Ab is normally produced in the brain, where the in vivo concentration in the rodent is estimated to be in the picomolar range [16]. Because Ab levels in the extracellular fluid are linked to cognitive activity [17,18], the mechanism of Abmediated modulation in cognition is important to understand cognitive function under both pathological and physiological conditions and also to pursue a strategy to prevent cognitive deficits in the pre-dementia stage of Alzheimer's disease. On the other hand, postmortem studies suggest that the hippocampus and entorhinal cortex are the first brain regions to be affected in Alzheimer's disease [19][20][21].
There is evidence that Zn 2+ in the extracellular fluid induces amyloid deposition [22,23], and clinical trials indicate that zinc chelation inhibits Ab-plaque deposition and improves executive function on a Neuro-psychological Test Battery in patients with Alzheimer's disease [24,25]. It is likely that amyloid pathology arises in a milieu of constitutively high Zn 2+ flux. On the other hand, homeostasis of extracellular Zn 2+ is critical for cognitive activity [26]. Both LTP and object recognition memory are impaired by excess intracellular Zn 2+ signaling, which is linked to dishomeostasis of extracellular Zn 2+ induced by excess excitation [27]. It is possible that extracellular Zn 2+ dynamics is changed in the presence of Ab and involved in cognitive decline in the pre-dementia stage of Alzheimer disease.
To examine an idea that short-term cognition is transiently affected by a state of confusion in Zn 2+ transport system due to a local increase in Ab concentration, a single injection of Ab , 142 into the dentate gyrus was performed in normal rats, focused on Ab-mediated Zn 2+ dynamics.

Ab-induced attenuation of LTP
SDS-PAGE showed that Ab in saline prepared in the present study was mainly monomers. It was estimated that dimers and a very small portion of trimers were also contained in the prepared solution (Fig. 1). LTP was induced 1 h after injection of Ab into the dentate gyrus via an injection cannula attached to a recording electrode. LTP was significantly attenuated with 25 pmol Ab ( Fig. 2A-2C). The attenuation of LTP was also observed when 25 pmol Ab was injected into the dentate gyrus 5 min before tetanic stimulation (saline: 265¡12% (n55), Ab 194¡22% (n55), p,0.05 (Dunnett's test) vs. saline). Ab-induced attenuation of LTP was enhanced by co-injection of 50 pmol ZnCl 2 , which had no effect on LTP ( Fig. 2D-2F). On the other hand, Ab-induced attenuation of LTP was rescued by co-injection of CaEDTA, a membrane-impermeable zinc chelator, and ZnAF-2DA, a membrane-permeable zinc chelator (Fig. 3).
The data that the action of Ab against LTP was rescued with zinc chelators suggest that Zn 2+ is involved in the action of Ab. Furthermore, the effect of neuronal depolarization on the action of Ab was evaluated by using KCl. The recording region where the perforant pathway-dentate granule cell synapses exist was pre-stimulated by co-injection of KCl. However, Ab-induced attenuation of LTP was not enhanced by co-injection of KCl (Fig. 4), unlike the case of coinjection of ZnCl 2 ( Fig. 2D-2F).

Ab-induced memory deficit
To examine whether the attenuation of LTP is linked to an object recognition memory deficit, training of the object recognition test was performed 1 h after injection of Ab into the dentate gyrus. The exploring time during training was not significantly influenced by injection of Ab or by co-injection of Ab and zinc chelators (Dunnett's test). The time exploring objects was 72¡4 s (vehicle), 67¡3 s (Ab), 77¡9 s (Ab + CaEDTA), 79¡8 s (CaEDTA), and 76¡6 s (Ab + ZnAF-2DA). However, a recognition memory deficit in the test was induced by injection of Ab, but not by co-injection of CaEDTA or ZnAF-2DA ( Fig. 5A and 5B).
ZnAF-2DA has the merits to visualize the area of Zn 2+ bound to intracellular ZnAF-2, which is produced in cells after ZnAF-2DA injection into the dentate gyrus. To assess the rescue effect of ZnAF-2DA on Ab-induced impairment of memory, the area containing Zn 2+ chelated with intracellular ZnAF-2 was checked in brain slices prepared after injection of ZnAF-2DA into the dentate gyrus. Intracellular ZnAF-2 fluorescence was observed in the injected area in the dentate gyrus, but not in the CA3 and CA1 (Fig. 5C). When the brain slices were stimulated with high K + , which increases Zn 2+ influx [27], intracellular ZnAF-2 fluorescence intensity was increased in the injected area ( Fig. 5D) and no change in fluorescence intensity was observed in other hippocampal regions, suggesting that intracellular ZnAF-2 in the injected area has a capacity to chelate Zn 2+ increased by Ab injection 1 h after ZnAF-2DA injection into the dentate gyrus. The injection of Ab may induce the influx of extracellular Zn 2+ in the dentate gyrus, while the co-injection of ZnAF-2DA rescued Ab-induced impairment of memory (Fig. 5B), probably via blocking Ab-mediated increase in intracellular Zn 2+ with intracellular ZnAF-2.
Ab-induced attenuation of LTP was not observed when LTP was induced 4 h after injection of 25 pmol Ab into the dentate gyrus ( Fig. 5E-5G). When training of the object recognition test was performed 4 h after injection of Ab into the dentate gyrus, the recognition index was almost the same as the control rats (Fig. 5H). Involvement of Ab-induced Zn 2+ influx in memory deficit Furthermore, Ab-mediated Zn 2+ dynamics was checked by co-injection of Ab and ZnAF-2DA into the dentate gyrus. Ab increased intracellular Zn 2+ levels, which were detected with intracellular ZnAF-2, in the injected area 15 min after injection ( Fig. 6A and 6B), suggesting that Ab increases the influx of Zn 2+ into cells in the injected area. When Ab was added to hippocampal slices in ACSF, intracellular Zn 2+ levels in the dentate granule cell layer were not changed (vehicle, 100.16¡0.85% (n59); Ab, 99.53¡0.91% (n59), Dunnett's test). The brain extracellular fluid contains zinc. In hippocampal slices in ACSF containing 1.2 mM ZnCl 2 at the final concentration, intracellular Zn 2+ levels were increased in the presence of Ab (final concentration, 10 mM). This increase was inhibited in the presence of CaEDTA, but not in the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), an a-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptor antagonist ( Fig. 6C and 6D). On the other hand, Ab did not influence intracellular Ca 2+ levels in the dentate granule cell layer.

Discussion
The aggregation of Ab-peptide is widely considered to be the critical step in the pathology of Alzheimer's disease. Small, soluble Ab oligomers have been shown to be more neurotoxic than large, insoluble aggregates and fibrils [5,6,28,29]. Ab1-42 is the major component of amyloid deposits in senile plaques [30]. Ab1-42 promotes amyloid deposition but Ab1-40 inhibits the process [31]. The soluble forms of Ab1-42 are the most toxic species that cause neuronal damage in the brains of Alzheimer's disease patients [32]. Zinc plays an important role in Ab aggregation process [33][34][35]. It is possible that the interaction between Ab and Zn 2+ in the extracellular compartment is involved in cognitive deficits in the presymptomatic phase of Alzheimer's disease.
In the dentate gyrus of the hippocampus, each granule cell receives two distinct inputs from the entorhinal cortex known as the medial and lateral perforant pathways (MPP and LPP, respectively). MPP fibers originate from the medial entorhinal area and terminate in the middle one-third of the molecular layer of the hippocampus where minimally stained by Timm's sulfide-silver method to detect zinc in the presynaptic vesicles [36][37][38]. Electrical stimulation of MPP fibers  The object recognition test was performed 1 h after injection of saline (control) (n516), Ab (25 pmol) (n513), Ab + CaEDTA (500 pmol) (n510), and CaEDTA (n59) in saline (1 ml). * , p,0.05, vs. saline, # , p,0.05 (Tukey's test). (B) The object recognition test was performed 1 h after injection of saline (control) (n516), Ab (25 pmol) (n513), and Ab + ZnAF2-DA (100 pmol) (n512) in saline (1 ml). ** , p,0.01, vs. saline, # , p,0.05 (Tukey's test). (C) Brain slices were prepared 1 h after bilateral injection of ZnAF-2DA (100 pmol, 100 mM/1 ml). Intracellular ZnAF-2 fluorescence was measured as the basal level for 30 s and then measured under stimulation with 50 mM KCl for 170 s. Note that intracellular ZnAF-2 fluorescence was observed only in a few slices including the injected area in the dentate gyrus. ML, molecular layer; GCL, granule cell layer. Bars, 50 mm. (D) The data represent the rate (%) of fluorescence intensity in the dentate granule cell layer 170 s after stimulation to that before the stimulation, which was represented as 100% (n55). ***, p,0.001 (paired t-test). vs. base. (E and F) LTP was induced 4 h after injection of produces a characteristic waveform of evoked field excitatory postsynaptic potential [fEPSP] superimposed by a population spike (PS) [39][40][41]. Judging from the PS amplitude recorded in the present study, it is estimated that LTP was induced by tetanic stimulation to the MPP [41], which is non-zincergic [42]. Thus it is estimated that extracellular Zn 2+ concentration is hardly increased during MPP LTP (dentate gyrus LTP) induction. In the present study, the action of Ab in LTP and cognitive activity was examined under the basal (static) condition. LTP was attenuated by local injection of Ab into the recording region 5 min or 1 h before tetanic stimulation. Ab-induced attenuation of LTP was enhanced by coinjection of ZnCl 2 . It is possible that the basal levels of extracellular Zn 2+ are involved in the action of Ab.
The basal concentration of extracellular Zn 2+ is estimated to be around 10 nM [43] and higher than that (,1 nM) of intracellular (cytosolic) Zn 2+ [44,45]. An idea that Ab-mediated Zn 2+ influx into dentate granule cells affects LTP was examined by co-injection of zinc chelators. The attenuation of LTP was rescued with CaEDTA to block the influx of extracellular Zn 2+ [46] and, the rescue effect was also observed by co-injection of ZnAF-2DA, which blocks intracellular Zn 2+ signaling [47], supporting the idea that Ab-mediated Zn 2+ influx affects LTP.
If LTP is involved in learning of object recognition, it is estimated that Ab simultaneously affects cognitive activity during the learning. Ab transiently induced the impairment of memory in agreement with the attenuation of LTP. Furthermore, Ab-induced impairment of memory was rescued with both CaEDTA and ZnAF-2DA. The present study indicates that cognitive activity is transiently affected via Ab-induced attenuation of LTP and that a state of confusion in Zn 2+ transport system in the dentate gyrus may be involved in the Ab-induced cognition deficit.
To examine whether neuronal depolarization aggravates Ab action, Ab was coinjected with KCl. KCl shortly induces depolarization of all neurons in the injected area prior to LTP induction. KCl had no effect on Ab-induced attenuation of LTP. It is likely that extracellular Zn 2+ is minimally increased in the recording area after injection of KCl into the dentate gyrus because of no release of Zn 2+ from the MPP, resulting in no enhancement of Ab-induced attenuation of LTP. In contrast, it is possible that Ab-mediated attenuation of LTP is enhanced after neurons are depolarized in the presence of Ab at zincergic synapses because of the increase in extracellular Zn 2+ .
doi:10.1371/journal.pone.0115923.g005 Fig. 6. Increase in Ab-mediated Zn 2+ influx. (A) Brain slices were prepared 15 min after injection of saline containing ZnAF-2DA (100 pmol) and Ab (25 pmol) in saline containing ZnAF-2DA (100 pmol) (1 ml) via an injection cannula and stained with calcium orange AM. Note that intracellular ZnAF-2 fluorescence was observed only in a few slices including the injected area in the dentate gyrus and not observed in the CA3 and CA1. G, dentate granule cell layer. H, hilus. Bars, 100 mm. (B) The data represent the ratios (%) of the fluorescence intensity after Ab injection to that after saline injection that was expressed as 100% (n54). * , p,0.05 (t-test). (C) Hippocampal slices doubly stained with ZnAF-2DA and Calcium Orange AM were immersed in 400 ml ACSF containing 1500 nM ZnCl 2 . After measuring each baseline fluorescence for 10 sec, 100 ml of saline (n513), 50 mM Ab in saline (n515), 50 mM Ab +200 mM CNQX in saline (n58), and 50 mM Ab +5 mM CaEDTA in saline (n58) was added to hippocampal slices in ACSF and both fluorescence was measured for 160 sec. The images of the dentate gyrus were obtained 160 s after addition. Bars; 50 mm. (D) The data represent the ratios (%) of the fluorescence after the addition (150-160 sec) to the baseline fluorescence (0-10 sec) that was expressed as 100%. * , p,0.05, vs. saline, # , p,0.05 (Tukey's test). No significant difference in calcium orange intensity, vs. saline (Dunnett's test). deficits after a single injection of Ab into the brain; Lesné et al. [5] show that a 56 kDa form of Ab purified from Tg2576 mouse brain disrupts maze performance when injected into the lateral ventricle of rats. Soluble Ab dimers, isolated from cerebral cortex of AD subjects, inhibit LTP, reduce dendritic spine density in rodent hippocampus, and disrupt memory in normal rats [6]. While multiple forms of Ab impair cognitive activity, there are significant differences in effective concentration and potency [48].
On the other hand, Ab has been shown to disturb intracellular calcium homeostasis that may alter calcium-related signal transduction pathways [49]. Intracellular Ca 2+ levels are increased in astrocytes exposed to Ab [50]. Ab increases intracellular Ca 2+ concentration by Ca 2+ influx through calciumpermeable channels formed by Ab and those channels are Zn 2+ -sensitive at high doses (50,1000 mM) [51,52]. Zn 2+ passes through calcium-permeable channels such as calcium-permeable AMPA receptors and N-methyl-D-aspartate (NMDA) receptors and high doses of Zn 2+ block NMDA receptors and voltage-dependent calcium channels [46,53,54]. Multiple forms of Ab are produced in the extracellular compartment and changed by the surrounding environment such as the presence of heavy metals. The changes may be markedly dynamic in the brain. Ab-mediated Zn 2+ dynamics was assessed after Ab injection into the dentate gyrus of rats. Intracellular Zn 2+ levels were increased only in the injected area. Furthermore, in vitro hippocampal slice experiments indicated that Ab also increased intracellular Zn 2+ levels in the dentate granule cell layer in ACSF containing 1.2 mM ZnC1 2 but not in ACSF without Zn 2+ . Ab-mediated Zn 2+ influx, which did not influence intracellular Ca 2+ levels, was inhibited in the presence of CaEDTA, but not in the presence of CNQX, which blocks glutamatergic neuron activation. These data suggest that extracellular Zn 2+ is critical for Ab-mediated Zn 2+ influx into dentate granule cells. Ab-mediated Zn 2+ influx may occur under the basal condition that is independent on AMPA receptor activation. It is possible that Zn 2+ influx into dentate granule cells occurs through zinc-permeable channels (ionophores) formed by Ab and the channels interact with multiple receptors such as NMDA receptors and prion protein [55]. Thus Zn 2+ potentially flows into the perforant pathway including MPP terminals in the same manner. Cognitive activity is transiently affected by unknown mechanism after intracellular Zn 2+ levels are excessively increased by Ab-mediated Zn 2+ influx. Ab-induced attenuation of MPP LTP, which is induced NMDA receptor-dependently [22,41], may be involved in cognitive deficits. This type of LTP is sensitive to excess Zn 2+ [56]. It is necessary to identify target molecules of excess Zn 2+ to understand the mechanism of Ab-induced attenuation of MPP LTP.
In conclusion, the present study indicates that Ab-mediated Zn 2+ influx into dentate granule cells, which may occur without AMPA receptor activation, transiently induces a short-term cognitive deficit. Extracellular Zn 2+ may play a key role for transiently Ab-induced cognition deficits. In subjects in the predementia stage of Alzheimer's disease, their cognitive deficits are limited to memory alone and/or their everyday abilities are preserved. A novel strategy focusing on a state of confusion in Zn 2+ transport system due to Ab may contribute to not only understanding transient cognition deficits in the predementia stage of Alzheimer's disease but also preventing pathogenesis of Alzheimer's disease.

Animals and chemicals
Male Wistar rats (6-8 weeks of age, Japan SLC (Hamamatsu, Japan) were used for the experiments. They were housed under the standard laboratory conditions (23¡1˚C, 55¡5% humidity) and had access to tap water and food ad libitum. All experiments were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals of University of Shizuoka that refer to American Association for Laboratory Animals Science and the guidelines laid down by the NIH (NIH Guide for the Care and Use of Laboratory Animals) in the USA. The Animal Experiment Committee of University of Shizuoka approved all protocols for animal experiments (#136043).
Synthetic Ab1-42 for human was purchased from ChinaPeptides (Shanghai, China). Ab was dissolved in saline and immediately used when the experiments were performed. ZnAF-2DA (K d 52.7610 29 M for zinc), a membrane-permeable zinc indicator, was kindly supplied from Sekisui Medical Co., LTD (Tokai, Japan). ZnAF-2DA is taken up into the cells through the cell membrane and is hydrolyzed by esterase in the cytosol to yield ZnAF-2, which cannot permeate the cell membrane [57,58]. Calcium orange AM, a membrane-permeable calcium indicator, was purchased from Molecular Probes, Inc. (Eugene, OR). These indicators were dissolved in dimethyl sulfoxide (DMSO) and then diluted to artificial cerebrospinal fluid (ACSF) containing 119 mM NaCl, 2.5 mM KCl, 1.3 mM MgSO 4 , 1.0 mM NaH 2 PO 4 , 2.5 mM CaCl 2 , 26.2 mM NaHCO 3 , and 11 mM D-glucose (pH 7.3).

Slice preparation and imaging
Rats were anesthetized with ether and decapitated. The brain was quickly removed and immersed in ice-cold choline-ACSF containing 124 mM choline chloride, 2.5 mM KCl, 2.5 mM MgCl 2 , 1.25 mM NaH 2 PO 4 , 0.5 mM CaCl 2 , 26 mM NaHCO 3 , and 10 mM glucose (pH 7.3) to suppress excessive neuronal excitation. Transverse hippocampal slices (400 mm) or coronal brain slices (400 mm) were prepared using a vibratome ZERO-1 (Dosaka, Kyoto, Japan) in an ice-cold choline-ACSF. Slices were then maintained in ACSF at 25˚C (for at least 1 h). All solutions used in the experiments were continuously bubbled with 95% O 2 and 5% CO 2 .
For intracellular zinc and calcium imaging, the hippocampal slices were loaded with 10 mM ZnAF-2DA and 10 mM calcium orange AM, respectively, for 30 min and then transferred a chamber filled with ACSF to wash out extracellular ZnAF-2DA and calcium orange AM for at least 30 min. The hippocampal slices were transferred to a recording chamber filled with ACSF. The fluorescence of ZnAF-2 (excitation, 488 nm; monitoring, 505-530 nm) and calcium orange (excitation, 543 nm; monitoring, above 560 nm) was measured in the granule cell layer of the dentate gyrus with a confocal laser-scanning microscopic system LSM 510 (Carl Zeiss), equipped with the inverted microscope (Axiovert 200 M, Carl Zeiss), at the rate of 1 Hz through a 106 objective.
To identify the area containing Zn 2+ chelated with intracellular ZnAF-2, coronal brain slices were prepared 15 min or 1 h after injection of 1 ml ZnAF-2DA (100 mM) or 1 ml ZnAF-2DA (100 mM) containing Ab (25 mM) into the dentate granule cell layer of unanesthetized rats at the rate of 0.25 ml/min for 4 min through injection cannulae. The fluorescence of ZnAF-2 was measured in the dentate granule cell layer in the same manner. In another experiment, brain slices were loaded with 10 mM calcium orange AM for 30 min and the fluorescence of ZnAF-2 and calcium orange was measured in the dentate granule cell layer in the same manner.

In vivo Dentate Gyrus LTP
We used the surgical procedure described previously with a slight modification [41]. Male rats were anesthetized with chloral hydrate (400 mg/kg) and placed in a stereotaxic apparatus. A bipolar stimulating electrode and a monopolar recording electrode made of tungsten wire attached to an injection cannula (internal diameter, 0.15 mm; outer diameter, 0.35 mm) were positioned stereotaxically so as to selectively stimulate the perforant pathway while recording in the dentate gyrus. The electrode stimulating the perforant pathway was positioned 8.0 mm posterior to the bregma, 4.5 mm lateral, 3.0-3.5 mm inferior to the dura. A recording electrode was implanted ipsilaterally 4.0 mm posterior to the bregma, 2.3-2.5 mm lateral and 3.0-3.5 mm inferior to the dura. All the stimuli were biphasic square wave pulses (200 ms width) and their intensities were set at the current that evoked 40% of the maximum PS amplitude. Test stimuli (0.05 Hz) were delivered at 20 s intervals to monitor PS amplitude.
At the beginning of the experiments, input/output curves were generated by systematic variation of the stimulus current (0.1-1.0 mA) to evaluate synaptic potency. After stable baseline recording for at least 30 min, agents in 1 ml saline were locally injected into the dentate gyrus of anesthetized rats at the rate of 0.25 ml/min for 4 min via an injection cannula attached to a recording electrode. LTP was induced by delivery of high-frequency stimulation (HFS; 10 trains of 20 pulses at 200 Hz separated by 1 s) 1 h or 4 h after injection and recorded for 60 min. PS amplitudes (test frequency: 0.05 Hz) were averaged over 120-second intervals and expressed as percentages of the mean PS amplitude measured during the 15-min baseline period perfused with ACSF prior to the injection, which was expressed as 100%. PS amplitudes for the last 10 min were also averaged and represented as the magnitude of LTP. In all experiments, we confirmed the location of the recording/injection sites after completing the experiments. The stimulation artifact was removed at the gaps in the representative traces.

Object recognition memory
LTP recording and the object recognition test were separately done. Rats were placed for 10 min into an open field, which was a 70660 cm arena surrounded by 70 cm high walls, made of a black-colored plastic. Twenty-four hours after open field exploration, agents in saline (1 ml) were bilaterally injected via injection cannulae into the dentate gyrus of unanesthetized rats at the rate of 0.25 ml/min for 4 min. One hour or four hours after injection, rats were trained and tested in a novel object recognition task. Training in the object recognition task took place in the same area used for the open field exploration. The open field exploration was thus used as a context habituation trial for the recognition memory task. The object recognition test requires that the rats recall which of two earthenware objects they had been previously familiarized with. Training was conducted by placing individual rats into the field, in which two identical objects (objects A1 and A2; sake bottle) were positioned in two adjacent corners, 15 cm from the walls. Rats were left to explore the objects for 5 min. Rats were not used for the test when the total of the object exploration time was less than 20 s. In the test given 1 h after training, the rats explored the open field for 3 min in the presence of one familiar (A) and one novel (B; cup) object. Behavior of rats was recorded with a video camera during the training and the test, and then two persons independently measured exploratory time and the averaged time was used. All objects presented similar textures, colors and sizes, but distinctive shapes. A recognition index calculated for each rat was expressed by the ratio T B /(T A +T B ) [T A 5 time spent to explore the familiar object A; T B 5 time spent to explore the novel object B]. Between trials the objects were washed with 70% ethanol solution. Exploration was defined as sniffing or touching the object with the nose and/or forepaws. We confirmed that there was no preference for the objects used.

Data analysis
For statistical analysis, Student's paired t-test was used for comparison of the means of paired data. For multiple comparisons, differences between treatments were assessed by one-way ANOVA followed by post hoc testing using the Dunnett's test or the Tukey's test (the statistical software, GraphPad Prism 5). A value of p,0.05 was considered significant. The Dunnett's test was used to compare between the control and treatments. The Tukey's test was used to compare between treatments in addition to the comparison between the control and treatments. Data were expressed as means ¡ standard error. The results of statistical analysis are described in each figure legend.