In vivo synaptic activity-independent co-uptakes of amyloid β1–42 and Zn2+ into dentate granule cells in the normal brain

Neuronal amyloid β1–42 (Aβ1–42) accumulation is considered an upstream event in Alzheimer’s disease pathogenesis. Here we report the mechanism on synaptic activity-independent Aβ1–42 uptake in vivo. When Aβ1–42 uptake was compared in hippocampal slices after incubating with Aβ1–42, In vitro Aβ1–42 uptake was preferentially high in the dentate granule cell layer in the hippocampus. Because the rapid uptake of Aβ1–42 with extracellular Zn2+ is essential for Aβ1–42-induced cognitive decline in vivo, the uptake mechanism was tested in dentate granule cells in association with synaptic activity. In vivo rapid uptake of Aβ1–42 was not modified in the dentate granule cell layer after co-injection of Aβ1–42 and tetrodotoxin, a Na+ channel blocker, into the dentate gyrus. Both the rapid uptake of Aβ1–42 and Zn2+ into the dentate granule cell layer was not modified after co-injection of CNQX, an AMPA receptor antagonist, which blocks extracellular Zn2+ influx, Both the rapid uptake of Aβ1–42 and Zn2+ into the dentate granule cell layer was not also modified after either co-injection of chlorpromazine or genistein, an endocytic repressor. The present study suggests that Aβ1–42 and Zn2+ are synaptic activity-independently co-taken up into dentate granule cells in the normal brain and the co-uptake is preferential in dentate granule cells in the hippocampus. We propose a hypothesis that Zn-Aβ1–42 oligomers formed in the extracellular compartment are directly incorporated into neuronal plasma membranes and form Zn2+-permeable ion channels.

It had been reported that Aβ oligomers form Ca 2+ -permeable plasma membrane pores that form Aβ channels, leading to a disruption of neuronal Ca 2+ homeostasis 17,18 , which may be linked with synaptic dysfunction and neurodegeneration 19 . Aβ-mediated Ca 2+ channels interact with Zn 2+ and Zn 2+ blocks extracellular Ca 2+ influx in the range of high micromolar concentrations (>250 μM) 20 , although the gating kinetics and Ca 2+ permeability of Aβ pores are not well understood 21 .
In the present study, we postulate that Zn-Aβ 1-42 oligomers formed in the extracellular compartment form Zn 2+ -permeable plasma membrane pores, based on the evidence of permeation of Zn 2+ through Ca 2+ channels 22,23 . Here we report the mechanism on synaptic activity-independent Aβ 1-42 uptake in vivo.

Results
When Aβ 1-42 is bound to extracellular Zn 2+ in vivo, Aβ 1-42 is rapidly taken up into dentate granule cells and affects memory via attenuated LTP [14][15][16] . Both uptake of Aβ  and Zn 2+ are observed 5 min after Aβ 1-42 injection into the dentate gyrus, followed by Aβ toxicity and are blocked by coinjection of CaEDTA, an extracellular Zn 2+ chelator, followed by blockade of Aβ toxicity. Thus, we need to clarify the mechanism of the rapid uptake of Aβ  in vivo, which is linked with Aβ toxicity. On the other hand, endogenous Zn 2+ released from the hippocampal slices promotes Aβ 1-42 uptake (retention), which is determined by in vitro Aβ immunohistochemistry (monoclonal antibody 4G8), in the absence of additional extracellular Zn 2+ 15 .
In the present study, Aβ 1-42 uptake was determined in rat hippocampal slices 15 min after incubation with Aβ 1-42 , Aβ 1-42 uptake was preferentially high in the dentate granule cell layer in the hippocampus, compared with the CA3 and CA1 pyramidal cell layer (Fig. 1).
To examine the involvement of neuronal activity in Aβ 1-42 uptake, Aβ 1-42 and tetrodotoxin, a Na + channel blocker, was co-injected into the dentate gyrus of rats. The present dose (15 μM) of tetrodotoxin was determined based on the in vivo action as a Na + channel blocker (10 μM) 24 . Aβ 1-42 uptake was determined by ex vivo Aβ immunohistochemistry 5 min after the co-injection. The rapid uptake of Aβ  was not modified in the dentate granule cell layer even by co-injection of tetrodotoxin (Fig. 2).
AMPA receptor activation induces extracellular Zn 2+ influx in vivo, which is blocked in the presence of 2 mM CNQX, and the excess influx leads to cognitive decline via attenuated LTP 25,26 . The present dose of CNQX (2 mM) was used according to the previous papers. Both the rapid uptake of Aβ  and Zn 2+ into the dentate granule cell layer was not modified by co-injection of CNQX (Fig. 3), suggesting Aβ 1-42 -mediated extracellular Zn 2+ influx.

Discussion
In vivo LTP at medial perforant pathway-dentate granule cell synapses, which is closely linked to object recognition memory 29 , is not affected even under perfusion with 1,000 nM Aβ 1-42 in ACSF without Zn 2+ , but attenuated under pre-perfusion with 500 pM Aβ 1-42 in ACSF containing 10 nM Zn 2+ . The attenuation is rescued by extracellular Zn 2+ -chelation with CaEDTA. These data indicate that high picomolar Aβ 1-42 captures extracellular Zn 2+ and subsequently attenuates LTP 15 . The evidence is consistent with Aβ 1-42 -induced object recognition memory deficit, which is also rescued by CaEDTA and ZnAF-2DA, an intracellular Zn 2+ chelator 13 . Thus, extracellular Zn 2+ is essential for Aβ 1-42 -induced cognitive decline via attenuated LTP in the normal brain, consistent with in vivo complete blocking rapid uptake of Aβ 1-42 and Zn 2+ into dentate granule cells after co-injection of Aβ 1-42 and CaEDTA into the dentate gyrus 15 . In the present study, we tested the uptake mechanism of Aβ 1-42 oligomers in vitro and in vivo.
When Aβ 1-42 uptake was determined in hippocampal slices after incubation with Aβ 1-42 , Aβ 1-42 uptake was preferentially high in the dentate granule cell layer in the hippocampus. Aβ 1-42 uptake was also considerably high in the CA3 pyramidal cell layer, while Aβ 1-42 uptake was not observed in the CA1 pyramidal cell layer. Extracellular Zn 2+ interaction with Aβ 1-42 is essential for Aβ 1-42 uptake into dentate granule cells in vivo and its essentiality is the same even under in vitro hippocampal slice condition 15 . Endogenous Zn 2+ release from the hippocampal slices is required for the Aβ 1-42 uptake. In hippocampal slices bathed in ACSF without Zn 2+ , extracellular Zn 2+ level determined with ZnAF-2 is the highest in the hilus close to the dentate granule cell layer and the second highest in the stratum lucidum close to the CA3 pyramidal cell layer 30 . Endogenous Zn 2+ , which is released from mossy fibers, may be closely linked with the Aβ 1-42 uptake. Although the Schaffer collaterals also release Zn 2+ , it is estimated that its release is not enough for the Aβ 1-42 uptake into CA1 pyramidal cells in vitro. In contrast, it is estimated that Aβ 1-42 captures extracellular Zn 2+ and is taken up into CA1 pyramidal cell in vivo. In vivo CA1 LTP is impaired after intracerebroventricular injections of Aβ peptide fragments 31 , implying that extracellular Zn 2+ is potentially involved in the impairment. Intracellular infusion of oligomerised Aβ 1-42 via passive diffusion from the patch pipette induces the rapid synaptic insertion of Ca 2+ -permeable AMPA receptors in CA1 pyramidal cells 32 . When oligomeric Aβ 1-42 captures extracellular Zn 2+ in vivo, it is also possible that intracellular oligomeric Aβ 1-42 affect neuronal functions. Postmortem studies suggest that the hippocampus and entorhinal cortex are the first brain regions to be affected in Alzheimer's disease 1,33,34 . The perforant pathway-dentate granule cell synapses are vulnerable to Aβ synapse toxicity 35 , which may be linked with extracellular Zn 2+ -mediated Aβ 1-42 uptake into dentate granule cells. This uptake may be associated with the high level of extracellular Zn 2+ in the hilus that may be due to Zn 2+ release form mossy fibers.
The mechanism of Aβ 1-42 uptake was tested in the dentate granule cell layer in vivo. When Aβ 1-42 was co-injected with tetrodotoxin, a Na + channel blocker, into the dentate gyrus, the rapid uptake of Aβ 1-42 was not modified in the dentate granule cell layer. AMPA receptor activation induces extracellular Zn 2+ influx and the www.nature.com/scientificreports www.nature.com/scientificreports/ excess influx leads to cognitive decline via attenuated LTP 25,26 . In contrast, both the rapid uptake of Aβ 1-42 and Zn 2+ into the dentate granule cell layer was not modified after co-injection of CNQX, an AMPA receptor antagonist, suggesting that Zn-Aβ 1-42 oligomers are ionotropic glutamate receptor activation-independently taken up into dentate granule cells. Both the rapid uptake of Aβ 1-42 and Zn 2+ into the dentate granule cell layer was not also modified after either co-injection of chlorpromazine or genistein, an endocytic repressor. It is likely that Zn-Aβ 1-42 oligomers are rapidly taken up into dentate granule cells without interaction with plasma membrane receptor proteins.
In human SH-SY5Y neuroblastoma, monomeric Aβ 1-42 is selectively internalized via clathrin-and dynamin-independent endocytosis compared to monomeric Aβ 1-40 36 . Genistein, a major phytoestrogen in soybean, may reduce the Aβ 1-42 -induced cell toxicity by suppressing the formation of toxic, cell membrane penetrating Aβ 1-42 oligomers in human SH-SY5Y neuroblastoma 37 . Zn 2+ interaction with Aβ 1-42 has not been taken into account in these in vitro cell culture systems, suggesting that Zn-Aβ 1-42 oligomers formed in vivo are taken up into dentate granule cells via a novel mechanism and causes cytotoxicity. On the other hand, Aβ production occurs in acidic vesicular organelles and contributes both to Aβ secretion and to the direct accumulation of Aβ within neurons 37 . Intracellular Aβ produced by the direct accumulation can exist as a monomeric form that further aggregates into oligomers and it may mediate pathological events 38 , although it is unknown whether Zn 2+ is involved in the pathological effects.
In conclusion, the present study suggests that amyloid β 1-42 and Zn 2+ are synaptic activity-independently co-taken up into dentate granule cells in the normal brain and the co-uptake is preferential in dentate granule cells. We propose a hypothesis that Zn-Aβ oligomers formed in the extracellular compartment are directly incorporated into neuronal membranes and form Zn 2+ -permeable ion channels. Because extracellular Zn 2+ is age-relatedly increased in the rat hippocampus 40 , Zn-Aβ 1-42 oligomers are more readily produced in the extracellular compartment of the aged hippocampus, followed by vulnerability to Zn-Aβ 1-42 oligomers in aging 16,41 . Therefore, controlling intracellular Zn 2+ dysregulation may be an effective strategy for overcoming AD pathogenesis.
www.nature.com/scientificreports www.nature.com/scientificreports/ to water and food ad libitum. Rats were used for experiments approximately 1 week after housing. All the experiments were performed according to the Guidelines for the Care and Use of Laboratory Animals in the University of Shizuoka, which refer to the 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 present study has been approved by the Ethics Committee for Experimental Animals in the University of Shizuoka.
In vitro immunostaining. Hippocampal slices were bathed in 50 μM Aβ 1-42 in ACSF for 15 min. Slices were then rinsed twice with ACSF for 5 min to remove extracellular agents, and fixed with paraformaldehyde (4% in 0.01 M PBS) for 15 min. Slices were rinsed in 0.01 M PBS three times. Slice tissues were blocked in 10% normal goat serum for 30 min, followed by rinse in 0.01 M PBS three times, incubated with 70% formic acid for 5 min, rinsed with 0.01 M PBS three times, and bathed at 4 °C in Aβ monoclonal antibody, 4G8 (COVANCE, 1:500 dilution in 0.01 M PBS) for 48 h. Slices were then rinsed with 0.01 M PBS three times, bathed in Alexa Fluor 633 goat anti-mouse IgG secondary antibody (1: 200 dilution in 0.01 M PBS) for 1 h, rinsed with 0.01 M PBS three times, bathed in 4′,6-diamidino-2-phenylindole (DAPI) for 10 min, rinsed again with 0.01 M PBS three times, and mounted on glass slides. Immunostaining images were obtained by using a confocal laser-scanning microscopic system (Nikon A1 confocal microscopes, Nikon Corp.) through a 10× objective. Florescence intensity was analyzed by the NIH Image J.
Ex vivo immunostaining. Rats anesthetized with chloral hydrate (400 mg/kg) were placed in a stereotaxic apparatus. The skull was exposed and two burr holes were drilled. Injection cannulae (internal diameter, 0.15 mm; outer diameter, 0.35 mm) were bilaterally inserted into the dentate granule cell layer (4.0 mm posterior to the bregma, 2.3 mm lateral, 2.9 mm inferior to the dura) of the anesthetized rats. Thirty minutes later following recovery from the insertion damage, 50 μM Aβ 1-42 in ACSF, or 50 μM Aβ 1-42 with either 15 μM tetrodotoxin, a Na 2+ channel blocker, 2 mM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), an α-amino-3-hydroxy-5-methyl -4-isoxazolepropionate (AMPA) receptor antagonist, 500 μM chlorpromazine, a clathrin/caveolae-mediated endocytosis blocker, or 250 μM genistein, a caveolae/raft-mediated endocytosis blocker, in ACSF were bilaterally injected into the dentate granule cell layer at the rate of 0.25 μl/min for 8 min via the injection cannulae. Five minutes later, the brain was quickly removed from the rats and immunostaining images in hippocampal slices were obtained in the same manner except for changing the 10% goat serum with 5% goat serum. The positions of the injection cannulae were confirmed in the slice preparation.
In vivo intracellular Zn 2+ imaging. Fifty μM Aβ 1-42 in ACSF containing 100 μM ZnAF-2DA, or 50 μM Aβ 1-42 with either 15 μM tetrodotoxin, 2 mM CNQX, 500 μM chlorpromazine, or 250 μM genistein in ACSF containing 100 μM ZnAF-2DA was bilaterally injected via injection cannulae into the dentate granule cell layer of anesthetized rats at the rate of 0.25 μl/min for 8 min. Five minutes later, the hippocampal slices were prepared in the same manner. Slices were transferred to a chamber filled with ACSF, loaded with 2 μM Calcium Orange AM in ACSF for 30 min, and then rinsed in ACSF for 30 min. The hippocampal slices were transferred to a recording chamber filled with ACSF. The fluorescence of ZnAF-2 (laser, 488.4 nm; emission, 500-550 nm), and calcium orange (laser, 561.4 nm; emission, 570-620 nm) was measured with a confocal laser-scanning microscopic system (Nikon A1 confocal microscopes, Nikon Corp.). Calcium Orange AM was used to identify hippocampal regions in slices. The positions of the injection cannulae were confirmed in the slice preparation.
Data analysis. For multiple comparisons, differences between treated groups were analyzed by one-way ANOVA followed by post hoc testing using the Tukey's test. A value of p < 0.05 was considered significant (the statistical software, GraphPad Prism 5). Data were expressed as means ± standard error. The control group with vehicle was compared with the treated groups in all figures for the statistical analysis. In Fig. 1, Aβ group in the DG was also compared with those in the CA3 and the CA1. ethics statement. All experiments were done according to the Guidelines for the Care and Use of Laboratory Animals of the University of Shizuoka, which 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. All experimental protocols were approved by the ethics committee of the University of Shizuoka