Abstract
Brain Aβ1–42 accumulation is considered an upstream event in pathogenesis of Alzheimer's disease. However, accumulating evidence indicates that other neurochemical changes potentiate the toxicity of this constitutively generated peptide. Here we report that the interaction of Aβ1–42 with extracellular Zn2+ is essential for in vivo rapid uptake of Aβ1–42 and Zn2+ into dentate granule cells in the normal rat hippocampus. The uptake of both Aβ1–42 and Zn2+ was blocked by CaEDTA, an extracellular Zn2+ chelator, and by Cd2+, a metal that displaces Zn2+ for Aβ1–42 binding. In vivo perforant pathway LTP was unaffected by perfusion with 1000 nm Aβ1–42 in ACSF without Zn2+. However, LTP was attenuated under preperfusion with 5 nm Aβ1–42 in ACSF containing 10 nm Zn2+, recapitulating the concentration of extracellular Zn2+, but not with 5 nm Aβ1–40 in ACSF containing 10 nm Zn2+. Aβ1–40 and Zn2+ were not taken up into dentate granule cells under these conditions, consistent with lower affinity of Aβ1–40 for Zn2+ than Aβ1–42. Aβ1–42-induced attenuation of LTP was rescued by both CaEDTA and CdCl2, and was observed even with 500 pm Aβ1–42. Aβ1–42 injected into the dentate granule cell layer of rats induced a rapid memory disturbance that was also rescued by coinjection of CdCl2. The present study supports blocking the formation of Zn-Aβ1–42 in the extracellular compartment as an effective preventive strategy for Alzheimer's disease.
SIGNIFICANCE STATEMENT Short-term memory loss occurs in normal elderly and increases in the predementia stage of Alzheimer's disease (AD). Amyloid-β1–42 (Aβ1–42), a possible causing peptide in AD, is bound to Zn2+ in the extracellular compartment in the hippocampus induced short-term memory loss in the normal rat brain, suggesting that extracellular Zn2+ is essential for Aβ1–42-induced short-term memory loss. The evidence is important to find an effective preventive strategy for AD, which is blocking the formation of Zn-Aβ1–42 in the extracellular compartment.
Introduction
Cognitive function normally declines along with aging and is thought to be initially due to changes in synaptic function rather than loss of neurons (Morrison and Hof, 1997). Alzheimer's disease (AD) is the most common cause of dementia and has a preclinical phase of 20–30 years before clinical onset (Nestor et al., 2004; Querfurth and LaFerla, 2010). Amyloid-β (Aβ) accumulation in the neocortex, the hallmark pathology of AD, is thought to play an upstream role in disease pathogenesis. Through mechanisms that are uncertain, Aβ oligomers can induce synapse dysfunction that contributes to cognitive decline in the predementia stage of AD (Perrin et al., 2009; Kepp, 2016).
Aβ is normally produced in the brain, where the concentration has been estimated to be in the picomolar range in rodents (Cirrito et al., 2003). The peptide is prone to self-assembly into oligomers, protofibrils, and fibrils (Gu et al., 2016). Aβ1–40 and Aβ1–42 are the two most abundant isoforms. Aβ1–40 is ≈10 times as abundant as Aβ1–42 in biological fluids (Schoonenboom et al., 2005). Importantly, Aβ1–42 far more readily forms aggregates and is more neurotoxic than Aβ1–40 (Mucke et al., 2000).
Aβ levels in the brain extracellular fluid are linked to cognitive activity (Cirrito et al., 2005; Puzzo et al., 2011). Synaptic vesicle release may be likely to be the primary mediator of dynamic changes in extracellular Aβ levels, which in turn may modify synaptic activity and are independent of changes in amyloid-β precursor protein (APP) processing (Cirrito et al., 2005). Studies of normal young animals report that endogenous Aβ is involved in learning and memory (Morley et al., 2010) ant that endogenous Aβ1–42 supports LTP expression (Puzzo et al., 2011). Together, it is possible that Aβ1–42 supports LTP and memory at picomolar concentrations under physiological conditions, whereas it impairs them at pathological nanomolar concentrations (Rammes et al., 2011; Puzzo et al., 2012).
Zn2+ has been implicated in the pathogenesis of AD by inducing Aβ oligomerization (Bush et al., 1994; Ayton et al., 2013; Bush, 2013). Here, we determine that Aβ1–42 takes Zn2+ as a cargo into the dentate granule neurons in the normal brain causing LTP and memory impairment.
Materials and Methods
Animals and chemicals.
Male Wistar rats (7–9 weeks of age) were purchased from Japan SLC. Rats were housed under the standard laboratory conditions (23 ± 1°C, 55 ± 5% humidity) and had access to tap water and food ad libitum. All the experiments were performed in accordance with the Guidelines for the care and use of laboratory animals of the University of Shizuoka that refer to the American Association for Laboratory Animals Science and the guidelines laid down by the National Institutes of Health Guide for the care and use of laboratory animals. The Ethics Committee for Experimental Animals in the University of Shizuoka has approved this work.
Synthetic human Aβ1–42 and Aβ1–40 were purchased from ChinaPeptides. Aβ was dissolved in saline and used immediately when the experiments were performed. SDS-PAGE showed that Aβ1–42 prepared in saline was mainly monomers with a small fraction of low-order oligomers (Takeda et al., 2014a). ZnAF-2DA (Kd = 2.7 × 10−9 m for zinc), a membrane-permeable zinc indicator was kindly supplied by Sekisui Medical. 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 (Hirano et al., 2002; Ueno et al., 2002). Calcium Orange AM, a membrane-permeable calcium indicator, was purchased from Invitrogen. These fluorescence indicators were dissolved in DMSO and then diluted to ACSF containing 119 mm NaCl, 2.5 mm KCl, 1.3 mm MgSO4, 1.0 mm NaH2PO4, 2.5 mm CaCl2, 26.2 mm NaHCO3, and 11 mm d-glucose, pH 7.3.
Hippocampal slice preparation.
Wistar rats were anesthetized with ether and decapitated in accordance with the Japanese Pharmacological Society Guide for the care and use of laboratory animals. The brain was quickly removed and immersed in ice-cold choline-ACSF containing 124 mm choline chloride, 2.5 mm KCl, 2.5 mm MgCl2, 1.25 mm NaH2PO4, 0.5 mm CaCl2, 26 mm NaHCO3, and 10 mm glucose, pH 7.3, to suppress neuronal excitation. Horizontal hippocampal slices (400 μm) were prepared in ice-cold choline-ACSF using a vibratome ZERO-1 (Dosaka) in an ice-cold choline-ACSF. Slices were then maintained in ACSF at 25°C for at least 30 min. All solutions used in the experiments were continuously bubbled with 95% O2 and 5% CO2.
In vitro immunostaining.
Hippocampal slices were incubated with 50 μm Aβ1–42 in ACSF, or with 50 μm Aβ1–42 and either 50 μm metals or 500 μm CaEDTA in ACSF for 15 min. Slices were then washed twice with ACSF for 5 min to remove extracellular agents, and fixed with PFA (4% in 0.01 m PBS) for 15 min. Slices were rinsed in 0.01 m PBS three times. Tissues were then blocked in 10% normal goat serum for 30 min, rinsed in 0.01 m PBS three times, incubated with 70% formic acid for 5 min, rinsed with 0.01 m PBS three times, and incubated at 4°C with 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, incubated with AlexaFluor-633 goat antimouse IgG secondary antibody (1: 200 dilution in 0.01 m PBS) for 1 h, rinsed in 0.01 m PBS three times, incubated with DAPI for 10 min, and then rinsed again with 0.01 m PBS three times before mounting on glass slides. Images for immunostaining were captured using a confocal laser-scanning microscopic system LSM 510 META (Carl Zeiss), equipped with an inverted microscope (Axiovert 200 m, Carl Zeiss) through a 10× and 40× objective. Florescence intensity was analyzed using National Institutes of Health ImageJ. The region of interest was set in the dentate granule cell layer in the dentate gyrus or in the pyramidal cell layer in the CA1 and CA3 subfields of the hippocampus.
In vivo immunostaining.
Male rats were anesthetized with chloral hydrate (400 mg/kg) and placed in a stereotaxic apparatus. The skull was exposed, two burr holes were drilled, and 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). Thirty minutes later, following recovery from the insertion damage, 50 μm Aβ1–42 in ACSF or 50 μm Aβ1–42 with either 50 μm metals or 500 μm CaEDTA in ACSF was bilaterally injected via the injection cannulae into the dentate granule cell layer of unanesthetized rats at the rate of 0.25 μl/min for 8 min. Five minutes later, the brain was quickly removed as described above and immunostaining using hippocampal slices was performed in the same manner except for exchanging the 10% goat serum with 5% goat serum.
In vivo intracellular Zn2+ imaging.
According to the procedure described above, ACSF containing 100 μm ZnAF-2DA was bilaterally injected via injection cannulae into the dentate granule cell layer of unanesthetized rats at the rate of 0.25 μl/min for 8 min. One hour later, 50 μm Aβ1–42 in ACSF or 50 μm Aβ1–42 + 50 μm metals in ACSF was bilaterally injected in the same manner. Five minutes later, the hippocampal slices were also prepared as noted above. In another experiment, 50 μm Aβ1–42 in ACSF containing 100 μm ZnAF-2DA or 50 μm Aβ1–42 + 500 μm CaEDTA in ACSF containing 100 μm ZnAF-2DA was bilaterally injected via injection cannulae into the dentate granule cell layer of unanesthetized rats at the rate of 0.25 μl/min for 8 min. The hippocampal slices were prepared in the same manner as above, transferred to a chamber filled with ACSF, loaded with 2 μm Calcium Orange AM in ACSF for 30 min to identify hippocampal regions, and then rinsed in ACSF for 10 min. The hippocampal slices were transferred to a recording chamber filled with ACSF. The fluorescence of ZnAF-2 (laser, 488 nm; emission, 505–530 nm) and Calcium Orange (laser, 543 nm; emission, >560 nm) was measured with a confocal laser-scanning microscopic system LSM 510.
In vivo LTP.
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 a microdialysis probe (AtmosLM, 1000 kDa cutoff, outer diameter 0.44 mm, Eicom) were positioned stereotactically so as to selectively stimulate the perforant pathway while recording under local perfusion with agents in ACSF (127 mm NaCl, 2.5 mm KCl, 0.9 mm MgCl2, 1.0 mm NaH2PO4, 1.3 mm CaCl2, 21 mm NaHCO3, and 3.4 mm d-glucose, pH 7.3) at the rate of 1.0 μl/min 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. The 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 μs width), and their intensities were set at the current that evoked 40% of the maximum population spike (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 were added to the ACSF perfusate either before or after LTP induction. LTP was induced by delivery of high-frequency stimulation (10 trains of 20 pulses at 200 Hz separated by 1 s) and recorded for 60 min. PS amplitudes (test frequency: 0.05 Hz) were averaged over 120 s intervals and expressed as percentages of the mean PS amplitude measured during the 30 min baseline recordings, which was expressed as 100%. PS amplitudes for the last 10 min were also averaged and represented as the magnitude of LTP.
In another set of experiments, 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 stereotactically so as to selectively stimulate the perforant pathway while recording in the dentate gyrus. After stable baseline recording for at least 30 min, agents in 1 μl saline were locally injected into the dentate granule cell layer of anesthetized rats at the rate of 0.25 μl/min for 4 min via an injection cannula attached to a recording electrode. LTP was induced in the same manner.
Object recognition memory.
The object recognition tests were performed in a separate cohort of animals. Briefly, rats were allowed to explore an open field (70 × 60 cm arena surrounded by 70 cm high walls, made of a black-colored plastic) for 10 min. Twenty-four hours later, agents in saline (1 μl) were bilaterally injected via injection cannulae into the dentate granule cell layer of unanesthetized rats at the rate of 0.25 μl/min for 4 min. One hour 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. Thus, the open field exploration was 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 in 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 <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. All objects were of similar texture, color, and size but were a distinctive shape (we confirmed that there was no preference for the objects used). All objects were washed with 70% ethanol between trials. The behavior of the rats was recorded with a video camera during the training and the test phases of the experiment, and then two people independently measured exploratory time and the averaged time was used. Exploration was defined as sniffing or touching the object with the nose and/or forepaws. A recognition index calculated for each rat was expressed by the ratio TB/(TA +TB), with TA as time spent to explore the familiar object A and TB as time spent to explore the novel object B.
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 are expressed as mean ± SE. The results of statistical analysis are described in each figure legend.
Results
Rapid hippocampal uptake of Aβ1–42 is mediated by extracellular Zn2+
Rat hippocampal slices were incubated for 15 min with Aβ1–42, and peptide retention was determined by Aβ immunohistochemistry (monoclonal antibody 4G8). Aβ was observed to attach mainly in the dentate granule cell layer. The staining was markedly enhanced by coincubation with ZnCl2, although it was not influenced by the presence of CuCl2 and FeCl3 (Fig. 1A,B). We then tested whether endogenous Zn2+ released from the hippocampal slices could be promoting Aβ1–42 retention in the absence of additional extracellular Zn2+ (Takeda et al., 2017). Indeed, Aβ1–42 retention was completely blocked in the presence of CaEDTA, an extracellular Zn2+ chelator (Fig. 1A,C). We confirmed that Zn2+ interaction with Aβ1–42 was responsible for peptide retention in the absence of exogenous Zn2+ by displacing the endogenous Zn2+ released by the slices with Cd2+, a nonphysiological metal ion that competes with Zn2+ for the histidine residues of Aβ (Syme and Viles, 2006). CdCl2 also abolished Aβ retention on the slices (Fig. 1A,C).
In vitro Aβ1–42 uptake in the dentate gyrus. Hippocampal slices were incubated with 50 μm Aβ1–42 in ACSF (n = 29), 50 μm Aβ1–42 + 50 μm ZnCl2 (n = 27), 50 μm Aβ1–42 + 50 μm CuCl2 (n = 10), 50 μm Aβ1–42 + 50 μm FeCl3 (n = 9), 50 μm Aβ1–42 + 50 μm CdCl2 (n = 8), or 50 μm Aβ1–42 + 500 μm CaEDTA in ACSF (n = 13). A, Aβ immunostaining in the dentate gyrus 15 min after incubation. GCL, Dentate granule cell layer. Scale bar, 50 μm. B, C, Aβ uptake in the dentate granule cell layer determined with Alexa-633 intensity, which is represented by the ratio to the control (n = 32) without 50 μm Aβ1–42 in ACSF expressed as 100%. B, *p < 0.05 versus control. ***p < 0.001 versus control. ###p < 0.001 versus Aβ. C, ***p < 0.001 versus control. ###p < 0.001 versus Aβ.
We studied the hippocampal retention of Aβ1–42 in vivo by performing Aβ immunohistochemistry of the region captured 5 min after local injection of Aβ1–42 into the dentate granule cell layer. Aβ1–42 staining, which was also observed in the dentate granule cell layer, was observed around the nuclei of dentate granule cells (Fig. 2A,C,D), consistent with intracellular uptake. Aβ1–42 staining was not observed in the CA1 and CA3 (Fig. 2B). Aβ1–42 staining was completely blocked upon coinjection of CaEDTA and CdCl2 (Fig. 2A,D). In contrast to Aβ1–42, when Aβ1–40 was injected, its retention in the hippocampus was not detectable in this time period (Fig. 2E,F).
In vivo differential uptake of Aβ1–42 and Aβ1–40 in the dentate gyrus and involvement of Zn2+. Fifty micromolar Aβ1–42 in ACSF (n = 29), 50 μm Aβ1–42 + 50 μm CdCl2 in ACSF (n = 10), and 50 μm Aβ1–42 + 500 μm CaEDTA in ACSF (n = 11) were bilaterally injected via injection cannulae into the dentate granule cell layer of unanesthetized rats. A, Aβ1–42 immunostaining in the dentate gyrus determined 5 min after injections. GCL, Dentate granule cell layer. Scale bar, 50 μm. B, Aβ1–42 immunostaining in the CA1 and CA3 subfields 5 min after injection of 50 μm Aβ1–42 injection. PCL, Pyramidal cell layer. C, Magnified image of Aβ1–42 in A. D, Aβ1–42 uptake in the dentate granule cell layer determined with Alexa-633 intensity, which is represented by the ratio to the control (n = 28) without 50 μm Aβ1–42 in ACSF expressed as 100%. ***p < 0.001 versus control. ##p < 0.01 versus Aβ. ###p < 0.001 versus Aβ. E, Aβ immunostaining images were determined in the dentate gyrus 5 min after bilateral injection of Aβ1–42 (n = 20) and Aβ1–40 (n = 11) (50 μm) in ACSF. F, Aβ uptake in the dentate granule cell layer determined with Alexa-633 intensity, which is represented by the ratio to the control (n = 28) without 50 μm Aβ in ACSF expressed as 100%. ***p < 0.001 versus control. ###p < 0.001 versus Aβ1–42. G, Intracellular Zn2+ images were determined in the dentate gyrus 5 min after bilateral injection of Aβ1–42 (n = 15) and Aβ1–40 (n = 19) (50 μm) in ACSF containing 100 μm ZnAF-2DA. H, Intracellular Zn2+ levels in the dentate granule cell layer determined with intracellular ZnAF-2, which is represented by the ratio to the control (n = 6) without 50 μm Aβ in ACSF expressed as 100%. *p < 0.05 versus control. #p < 0.05 versus Aβ1–42.
Rapid hippocampal uptake of extracellular Zn2+ is mediated by Aβ1–42
The in vivo status of intracellular Zn2+, which was measured with ZnAF-2, was determined 5 min after local injection of Aβ1–42 into the dentate granule cell layer. Intracellular ZnAF-2 fluorescence was increased after injection of Aβ1–42, and the increase was enhanced after coinjection of ZnCl2 (Fig. 3A,C). Intracellular ZnAF-2 fluorescence was not observed in the CA1 and CA3 (Fig. 3B), probably because there was no diffusion of ZnAF-2DA or Aβ1–42 from the injection zone. Aβ1–42-mediated increase in intracellular ZnAF-2 florescence was not influenced by coinjection of CuCl2 or FeCl3 (Fig. 3A,C) but blocked by coinjection of CaEDTA or CdCl2 (Fig. 3A,D,E,F). Because CaEDTA does not enter the cell, we conclude that extracellular Zn2+ entering into the cell is the source of the increased ZnAF-2 fluorescence induced by Aβ1–42.
In vivo Aβ1–42-mediated Zn2+ uptake in the dentate gyrus. A, ACSF containing 100 μm ZnAF-2DA was bilaterally preinjected in the dentate gyrus to obtain intracellular ZnAF-2 images. One hour later, intracellular Zn2+ images were determined in the dentate gyrus 5 min after bilateral injection of 50 μm Aβ1–42 in ACSF (n = 13), 50 μm Aβ1–42 + 50 μm ZnCl2 (n = 11), 50 μm Aβ1–42 + 50 μm CuCl2 (n = 12), 50 μm Aβ1–42 + 50 μm FeCl3 (n = 11), and 50 μm Aβ1–42 + 50 μm CdCl2 (n = 11) in ACSF. B, Intracellular Zn2+ images were also determined in the CA1 and CA3 5 min after bilateral injection. C, D, Quantitation of intracellular Zn2+ levels in the dentate granule cell layer determined with intracellular ZnAF-2, which is represented by the ratio to the control (n = 8) without 50 μm Aβ1–42 in ACSF expressed as 100%. **p < 0.01 versus control. ***p < 0.001 versus control. ##p < 0.01 versus Aβ1–42. ###p < 0.001 versus control. E, Intracellular Zn2+ images were determined in the dentate gyrus 5 min after bilateral injection of 50 μm Aβ1–42 (n = 12) and 50 μm Aβ1–42 + 500 μm CaEDTA (n = 8) in ACSF containing 100 μm ZnAF-2DA. F, Intracellular Zn2+ levels in the dentate granule cell layer determined in the same manner. **p < 0.01 versus control (n = 9). #p < 0.05 versus Aβ1–42.
In contrast to the effect of Aβ1–42, local injection of Aβ1–40 did not increase intracellular ZnAF-2 fluorescence (Fig. 2G,H).
Aβ1–42-induced attenuation of LTP requires extracellular Zn2+
We examined the impact of Aβ1–42 drawing extracellular Zn2+ into the cell upon in vivo LTP in perforant pathway-dentate granule cell synapses, where the recording area was locally perfused with Aβ in ACSF. LTP was not attenuated upon perfusion with 5–1000 nm Aβ1–42 (Fig. 4A). LTP was not attenuated upon perfusion with ACSF containing 10 nm ZnCl2 (representing the concentration of extracellular Zn2+) (Frederickson et al., 2006) but was significantly attenuated upon perfusion with 5 nm Aβ1–42 in ACSF containing 10 nm ZnCl2 (Fig. 4B). LTP was also attenuated by the preperfusion with 5 nm Aβ1–42 in ACSF containing 10 nm ZnCl2 before tetanic stimulation, but not by the perfusion during and after tetanic stimulation (Fig. 4C). In the absence of Aβ1–42, LTP was not attenuated by preperfusion with 100 nm ZnCl2, but continuous perfusion with 100 nm ZnCl2 did attenuate LTP (Fig. 5A). The preperfusion of 5 nm Aβ1–42 with 10 nm CuCl2, FeCl3, or CdCl2 instead of ZnCl2 did not attenuate LTP (Fig. 5B). On the other hand, the attenuation of LTP seen with 5 nm Aβ1–42 in ACSF containing 10 nm ZnCl2 was rescued when 1 mm CaEDTA or 10 nm CdCl2 was added to the perfusate (Fig. 5C). LTP was not attenuated under preperfusion with 5 nm Aβ1–40 in ACSF containing 10 nm ZnCl2 (Fig. 4C).
Involvement of extracellular Zn2+ in Aβ1–42-induced attenuation of LTP. A, Recording region was perfused with either ACSF (n = 9) or 5 (n = 5), 200 (n = 6), or 1000 nm Aβ1–42 (n = 7) in ACSF for 60 min. High-frequency stimulation (10 trains of 20 pulses at 200 Hz separated by 10 s) was delivered at time 0 min and perfused for 60 min under the same condition. Left, PS amplitude over time. Bar represents the perfusion period with Aβ. Middle panels, Average PS amplitude (mean ± SEM) during the last 10 min of recording. Right, Representative fEPSP traces with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red). B, Left, LTP was induced in the same manner under perfusion with 5 nm Aβ1–42 (n = 5), 10 nm ZnCl2 (n = 7), and 5 nm Aβ1–42 + 10 nm ZnCl2 (n = 11). Middle panels, Average PS amplitude (mean ± SEM) during the last 10 min of recording. Right, Representative fEPSP traces with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red). **p < 0.001 versus control (0/0) (n = 9). #p < 0.05 versus 5 nm Aβ1–42. C, LTP was induced under differential perfusion period (−60 to −5 min, n = 8; −10 to 5 min, n = 5; 5 to 60 min, n = 5) with 5 nm Aβ1–42 + 10 nm ZnCl2 and also induced under preperfusion (−60 to −5 min) with 5 nm Aβ1–40 + 10 nm ZnCl2 (n = 8). Left, Bars represent the perfusion period. Middle, Each bar (mean ± SEM) represents the averaged PS amplitude of the last 10 min of recording. Right, LTP was attenuated under preperfusion with 5 nm Aβ1–42 + 10 nm ZnCl2, but not with 5 nm Aβ1–40 + 10 nm ZnCl2. Bottom, Representative fEPSP with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red). *p < 0.05 versus control (n = 11).
Rescue of Aβ1–42- and Zn2+-induced attenuation of LTP in the presence of CaEDTA and CdCl2. A, LTP was induced under differential perfusion period (−60 to 60 min, n = 9; −60 to −5 min, n = 5; −10 to 5 min, n = 7) with 100 nm ZnCl2 in ACSF. Left, Bars represent the perfusion period. Middle, Each bar (mean ± SEM) represents the averaged PS amplitude of the last 10 min. LTP was not significantly attenuated under preperfusion with 100 nm ZnCl2. Right, Representative fEPSP with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red). ***p < 0.001 versus control (n = 11). B, LTP was induced under preperfusion with 5 nm Aβ1–42 + 10 nm CuCl2 (n = 5), 5 nm Aβ1–42 + 10 nm FeCl3 (n = 4), and 5 nm Aβ1–42 + 10 nm CdCl2 (n = 5) in ACSF. Left, Bar represents the perfusion period. Middle, Each bar and line (mean ± SEM) indicate the averaged PS amplitude of the last 10 min. Right, Representative fEPSP with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red). C, LTP was induced under preperfusion with 5 nm Aβ1–42 + 10 nm ZnCl2 (n = 8), 5 nm Aβ1–42 + 10 nm ZnCl2 + 1 mm CaEDTA (n = 5), and 5 nm Aβ1–42 + 10 nm ZnCl2 + 10 nm CdCl2 (n = 7) in ACSF. Left, Bar represents the perfusion period. Middle, Each bar and line (mean ± SEM) indicate the averaged PS amplitude of the last 10 min. Right, Representative fEPSP with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red). *p < 0.05 versus control (n = 11). #p < 0.05 versus 5 nm Aβ1–42 + 10 nm ZnCl2.
Aβ1–42-induced short-term memory decline is rescued by Cd2+
LTP was attenuated 1 h after local injection of Aβ1–42 (25 μm, 1 μl) into the dentate granule cell layer (control, 230 ± 7%; Aβ1–42, 174 ± 9%, p < 0.001 vs control) and the attenuation was rescued by coinjection of CdCl2 (50 μm, 226 ± 9%, p < 0.05 vs Aβ1–42) (Fig. 6A), consistent with Cd2+ displacing Zn2+ from Aβ in the slice experiments (Fig. 3D). The novel object recognition test was then performed 1 h after local injection of Aβ1–42 in the same manner (exploring time, control, 44.2 ± 13.8 s; Aβ1–42, 51.8 ± 2.8 s; Aβ1–42/Cd, 50.1 ± 5.4 s; Cd, 44.6 ± 4.6 s). One hour later, object recognition memory was impaired in the animals injected with Aβ1–42, whereas the impairment was rescued by coinjection of CdCl2 (Fig. 6B). LTP and object recognition memory were not impaired by injection of CdCl2 alone.
Rescue of Aβ1–42-induced impairments of LTP and memory in the presence of CdCl2 and picomolar Aβ1–42-induced attenuation of LTP. A, LTP was induced 1 h after injection of 25 μm Aβ1–42 (n = 18), 25 μm Aβ1–42 + 50 μm CdCl2 (n = 5), and 50 μm CdCl2 (n = 6) in ACSF (1 μl) via an injection cannula into the dentate granule cell layer (time 0 min). LTP was significantly (p < 0.001 vs control; n = 24) attenuated by coinjection of Aβ1–42, and the attenuation was significantly (p < 0.05 vs Aβ1–42) rescued by coinjection of CdCl2. B, Training of object recognition test was performed 1 h after bilateral injection of 25 μm Aβ1–42 (n = 9), 25 μm Aβ1–42 + 50 μm CdCl2 (n = 9), and 50 μm CdCl2 (n = 8) (1 μl) in ACSF into the dentate granule cell layer. One hour later, the test was performed. ***p < 0.001 versus control (n = 10). ##p < 0.01 versus Aβ1–42. ###p < 0.001 versus Aβ1–42. Recognition indices in the test were significantly elevated in the control (p < 0.001, t test), Aβ1–42/Cd (p < 0.001, t test), and Cd (p < 0.01, t test) groups. C, LTP was induced under preperfusion with 100 pm Aβ1–42 + 10 nm ZnCl2 (n = 5) and 500 pm Aβ1–42 + 10 nm ZnCl2 (n = 11) in ACSF. Bar represents the perfusion period (left). Scale bar and line (mean ± SEM) indicate the averaged PS amplitude of the last 10 min (middle). Representative fEPSP with PS recordings at the time −70 min (gray), −30 min (black), and 50–60 min (red) (right). *p < 0.05 versus control (n = 11).
Finally, the action of Aβ1–42 in LTP induction was assessed under local perfusion of the animal at the concentrations of <1 nm. LTP was significantly attenuated under preperfusion with 500 pm Aβ1–42 in ACSF containing 10 nm ZnCl2, but not with 100 pm Aβ1–42 in ACSF containing 10 nm ZnCl2 (Fig. 6C).
Discussion
It is reported that the concentrations of zinc, copper, and iron in the human CSF are 0.38, 0.34, and 0.54 μm, respectively (Gellein et al., 2008; Michalke and Nischwitz, 2010). Although the concentrations of these metals are unknown in the brain extracellular fluid, a fraction is exchangeable. It is reported that Aβ is bound to Zn2+ via histidine residues and that the Kd values of Zn2+ to Aβ1–40 are in the range of 0.1–60 μm (Tõugu et al., 2008). However, the Kd value of Zn2+ to Aβ1–42 is unreported. It is likely that, as is the case with Cu2+ binding to Aβ (Atwood et al., 2000), the apparent Kd of metal binding to Aβ1–42 is higher in affinity than to Aβ1–40 probably due to the perturbed equilibrium caused by the increased self-assembly of Aβ1–42 oligomers. In vivo LTP at medial perforant pathway-dentate granule cell synapses, which is closely linked to object recognition memory (Takeda et al., 2014a, b; Suzuki et al., 2015), was not affected by perfusion with 1000 nm Aβ1–42 in ACSF without Zn2+, but attenuated under preperfusion with 5 nm Aβ1–42 in ACSF containing 10 nm Zn2+, as estimated Zn2+ concentration in the brain extracellular compartment under the basal (static) conditions (Frederickson et al., 2006). The attenuation was rescued by extracellular Zn2+ chelation with CaEDTA, whereas the attenuation was not observed under preperfusion with 5 nm Aβ1–40 in ACSF containing 10 nm Zn2+. These data indicate that low nanomolar Aβ1–42, unlike Aβ1–40, rapidly binds extracellular Zn2+ and subsequently attenuates LTP, consistent with subsequent object recognition memory decline. Thus, extracellular Zn2+ may impact on Aβ1–42-induced cognitive decline via attenuated LTP in the normal brain.
When Aβ1–42 was added to hippocampal slices, its uptake into dentate granule cells was increased in ACSF-containing Zn2+, whereas it was blocked in ACSF containing CaEDTA. In vivo rapid Aβ1–42 uptake into dentate granule cells was also completely blocked by coinjection of Aβ1–42 with CaEDTA into the dentate gyrus. Simultaneously, in vivo rapid Zn2+ uptake into dentate granule cells, which was induced by injection of Aβ1–42, was completely blocked by coinjection of CaEDTA. Together, our findings indicate that the interaction of Aβ1–42 with extracellular Zn2+ promotes uptake of Aβ1–42 into dentate granule cells in the normal brain, causing an increase in intracellular Zn2+, and leading rapidly to cognitive impairment (Fig. 7). Preclinical and clinical data show the potential for metal chelation-based drug therapy for AD: Clioquinol reduces zinc accumulation in neuritic plaques and inhibits the amyloidogenic pathway in the AβPP/PS1 transgenic mouse brain (Wang et al., 2012). Clioquinol also promotes the degradation of metal-dependent Aβ oligomers to restore endocytosis and ameliorates Aβ toxicity (Matlack et al., 2014). Furthermore, PBT2, a copper/zinc ionophore and second-generation 8-hydroxyquinoline analog, significantly lowers Aβ levels in the CSF and improves cognitive performance over baseline in several key executive function tests (Lannfelt et al., 2008; Faux et al., 2010). Zn2+- or Cu2+-induced aggregates of Aβ (i.e., soluble oligomers) have been implicated as the neurotoxic form of the peptides against synapse function and structure.
A model for neuronal intoxication by Zn-Aβ1–42 complexes. Dentate granule cells will only take up Aβ1–42 when it is bound to Zn2+, whereupon Zn-Aβ1–42 complexes enter neurons and either the complexes or the liberated Zn2+ induces cognitive decline via attenuated LTP. Cd2+, but not Cu2+ or Fe3+, can displace Zn2+ from Aβ1–42 and prevent it from being taken up into granule cells, neutralizing its toxicity. Zn2+ is withdrawn into peptide complexes more readily by oligomers; hence, Aβ1–42, which more rapidly forms oligomers, fosters neuronal uptake more readily than Aβ1–40.
The medial perforant pathway is nonzincergic and does not release Zn2+ (Sindreu et al., 2003). Extracellular Zn2+ concentration at medial perforant pathway-dentate granule cell synapses may be relatively static in the hippocampus and may be maintained at ∼10 nm (Frederickson et al., 2006). Such rapid uptake into dentate granule cells was not observed after injection of Aβ1–40, probably due to less interaction of low nanomolar Aβ1–40 with Zn2+ in the extracellular compartment. Hence, Aβ1–40 perfusion had little effect on LTP induction. Increasing evidence has suggested that formation and propagation of misfolded aggregates of Aβ1–42, rather than of Aβ1–40, contribute to AD pathogenesis. However, structural details of misfolded Aβ1–42 remain to be clarified (Ahmed et al., 2010). Masuda et al. (2009) report that C-terminal carboxylate anion of Aβ1–42 forms the C-terminal hydrophobic core that accelerates neurotoxic oligomerization. Xiao et al. (2015) report that C-terminal Ala42, absent in Aβ1–40, forms a salt bridge with Lys28 to create a self-recognition molecular switch that is the Aβ1–42-selective self-replicating amyloid-propagation machinery. The aggregation property of Aβ1–42 is promoted with Zn2+, resulting in higher affinity of Aβ1–42 to Zn2+ than Aβ1–40 that leads to synaptic dysfunction via neuronal Zn-Aβ1–42 uptake.
Aβ can bind up to 3.5 equivalents of Zn2+ and Cu2+ simultaneously (Atwood et al., 2000) and can bind several other transition metals. However, at pH 7.4, only Zn2+, but not Cu2+, causes significant Aβ aggregation. The reversible oligomerization of Aβ induced by Zn2+ forming salt-bridges between peptide subunits (Huang et al., 1997) may be the physical basis for our observations that Cu2+ and Fe3+ do not compete for Zn2+ uptake into cells. Aβ1–40 possesses selective affinity Cu2+ binding sites, and the binding affinity of Cu2+ is greater than for Zn2+. If Cu2+ is preferentially bound to Aβ1–42 in the extracellular compartment, it blocks Aβ1–42-mediated Zn2+ accumulation. CuCl2 (50 μm) did not modify in vitro Aβ1–42 uptake into dentate granule cells, which might be mediated by endogenous Zn2+ released from the hippocampal slices (Takeda et al., 2017). Furthermore, in vivo increase in Aβ1–42-mediated Zn2+ uptake into dentate granule cells was not modified by coinjection of Aβ1–42 and CuCl2 (50 μm). These data suggest that extracellular Zn2+ is bound to Aβ1–42 even in the presence of micromolar Cu2+ and Fe3+ at neutral pH, resulting in the increase in intracellular Zn2+. Because ZnAF-2 (Kd, 2.7 nm) must have a higher affinity than Aβ1–42 for Zn2+, it is estimated that in vivo Kd value of Zn2+ to Aβ1–42 is in the range of ∼3–30 nm. The free intracellular Zn2+ concentration is estimated to be <1 nm (Sensi et al., 1997; Colvin et al., 2008). Therefore, once ferried by Aβ1–42 into dentate granule cells, the Zn2+ cargo is released in dentate granule cells. This may be critical for Aβ1–42-induced cognitive decline in the normal rats. Because Zn2+ has many potential targets in dentate granule cells, the mechanism of the Zn2+ neurotoxicity is complex. Zhang et al. (2008) report abundant expression of Zn2+ transporters in the amyloid plaques of AD brain. Altered protein levels of the membrane Zn2+ transporters ZnT1, ZnT4, and ZnT6 have been reported in AD postmortem brain tissue (Beyer et al., 2012). The evidence suggests that Zn2+ transporters are involved in the pathological processes that lead to plaque formation. In a short period when the Zn2+ neurotoxicity is induced in the present study, however, the involvement of Zn2+ transporters might be modest.
In contrast, both Aβ1–42 uptake and intracellular Zn2+ increase were blocked by Cd2+. Together, these data are consistent with the possibility that Aβ1–42 preferentially enters the cell when it forms a complex with Zn2+, and that Cu2+ and Fe3+ are unable to compete off the Zn2+ entirely, whereas Cd2+ can. Thus, the neuronal dysfunction that we observed would be mediated by the intrusion of Aβ, Zn2+, or both together (Fig. 7).
It is reported that the mean concentration of Aβ1–42 in the CSF is significantly reduced in subjects with Alzheimer's disease compared with age-matched controls and is ∼500 pm in age-matched controls (Motter et al., 1995). On the other hand, 200 pm human Aβ1–42 improves LTP and memory in normal mice (Puzzo et al., 2008) and in mice challenged with antirodent Aβ monoclonal antibody and siRNA against murine APP (Puzzo et al., 2011). In the hippocampus of young mice, extracellular Aβ concentration measured by microdialysis is ∼160 pm and extracellular Aβ1–42 is ∼20 pm, and not significantly changed in aged mice (Cirrito et al., 2003). Synaptic activity increases both extracellular Aβ1–42 (Cirrito et al., 2005; Kim et al., 2010) and extracellular Zn2+ at zincergic synapses (Takeda and Tamano, 2016). So we hypothesize that a decrease in clearance mechanisms may lead to the inappropriate combination of Zn2+ with Aβ1–42, leading to a toxic aggregate that enters neurons.
In conclusion, Zn2+ induces Aβ1–42 uptake in the normal dentate gyrus, and memory dysfunction, when extracellular Aβ1–42 reaches high picomolar concentrations. Blocking the formation of Zn-Aβ1–42 in the extracellular compartment may be an effective strategy for preventing Aβ1–42-mediated cognitive decline.
Footnotes
The authors declare no competing financial interests.
- Correspondence should be addressed to Dr. Atsushi Takeda, Department of Neurophysiology, School of Pharmaceutical Sciences, University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka 422-8526, Japan. takedaa{at}u-shizuoka-ken.ac.jp
