|Home | About | Journals | Submit | Contact Us | Français|
Amyloid-β (Aβ) peptide-induced impairment of hippocampal synaptic plasticity is considered an underlying mechanism for memory loss in the early stages of Alzheimer’s disease and its animal models. We previously reported inhibition of long-term potentiation (LTP) and miniature excitatory postsynaptic currents by oligomeric Aβ1–42 at hippocampal synapses. While multiple cellular mechanisms could be involved in Aβ-induced synaptic dysfunction, blockade of activity-dependent autophosphorylation of Ca2+ and calmodulin-dependent protein kinase II (CaMKII) appeared to be a major component of Aβ action in our studies. The present study further tested this hypothesis and examined the therapeutic potential of trkB receptor-acting neurotrophins in rescuing Aβ-induced synaptic and signaling impairments. As expected, treatment of rat hippocampal slices with Aβ1–42 significantly reduced LTP in the Schaffer collateral-CA1 pathway and dentate medial perforant path. LTP-associated CaMKII activation and AMPA receptor phosphorylation were blocked by Aβ1–42 at the same concentration that inhibited LTP. Aβ-induced LTP impairment, however, was prevented when slices were co-treated with neurotrophin 4 (NT4). Western blotting and immunohistochemical analyses confirmed that treatment with NT4 or brain-derived neurotrophic factor, another trkB-acting neurotrophin, could oppose Aβ action, enhancing autophosphorylation of CaMKII, and AMPA receptor phosphorylation at a CaMKII-dependent site. These findings support the view that CaMKII is a key synaptic target of Aβ toxicity as well as a potential therapeutic site of neurotrophins for Alzheimer’s disease.
Accumulation of amyloid-β (Aβ) peptides in the brain is a crucial step in the pathogenesis of Alzheimer’s disease (AD). Earlier studies linked neurotoxic action of Aβ with its insoluble, fibrillar forms deposited in senile plaques in AD brain. Such a link was controversial because of lack of strong correlations between fibrillar Aβ load and the extent of memory loss in AD and transgenic mouse models. Studies in recent years have suggested that soluble oligomeric forms of Aβ may be a major player in causing synaptic dysfunction and memory loss during early AD. Accumulation of various Aβ assemblies, most notably a soluble dodecamer Aβ*56, reportedly disrupts cognitive function in AβPP transgenic mice . Both synthetic and naturally existing soluble Aβ oligomers can bind to synaptic sites in rat hippocampal and cortical cultures . Application of Aβ oligomers leads to rapid inhibition of hippocampal long-term potentiation (LTP) [3–5], facilitation of long-term depression , and suppression of spontaneous synaptic activity in hippocampal and cortical cultures [7–9]. These findings support the view that Aβ oligomer-induced synaptic dysfunction in brain regions crucial for memory formation and storage, such as hippocampus and cortex, has a causative role in memory loss of early AD [10,11]. While this view has been increasingly accepted in recent years, the exact synaptic targets and key signaling events underlying Aβ-induced synaptic depression remain to be further elucidated.
We previously reported that acute application of Aβ1–42 strongly inhibited LTP induction and maintenance in the hippocampal Schaffer collateral-CA1 pathway and dentate medial perforant path . Further analyses showed that Aβ-induced LTP deficits were linked with inhibition of N-methyl-D-aspartate (NMDA) receptor channel currents and activation of the calcineurin-protein phosphatase 1 pathway [13, 14]. Aβ1–42 also blocked LTP-triggered activation of Ca2+ and calmodulin-dependent protein kinase II (CaMKII) and subsequent phosphorylation of the α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor at Ser831 of the GluR1 subunit, a CaMKII-dependent site . These findings suggest that multiple Ca2+-regulated signaling pathways are involved in mediating Aβ action at hippocampal synapses. The essential role of CaMKII in long-term synaptic plasticity and cognitive function is well documented. Synaptic activity-triggered Ca2+ influx through NM-DA receptor channels can activate CaMKII and promote its autophosphorylation at Thr286, which results in a persistently active form of the kinase that is required for LTP . Activation of a calcineurin-dependent phosphatase pathway, however, can dephosphorylate CaMKII and reduce its activity . It is thus highly plausible that CaMKII is a key synaptic target for Aβ. By integrating several cellular effects of Aβ, changes in CaMKII activity could lead to synaptic weakening and impaired synaptic plasticity . Drugs that enhance CaMKII signaling, therefore, should be able to oppose Aβ action and rescue synaptic and cognitive function. The present study tested these hypotheses and demonstrated that application of trkB-acting neurotrophins, which are known to enhance synaptic plasticity in adult hippocampus , could stimulate CaMKII activity and effectively rescue Aβ-induced LTP deficits at hippocampal synapses.
Hippocampal slices were prepared from 30- to 45-day-old male Sprague-Dawly rats as previously described . Rats were sacrificed under isoflurane anesthesia following a protocol in accordance with PHS Guidelines, reviewed and approved by the Chancellor’s Animal Research Committee at University of California, Los Angeles. The brain was removed and sectioned into 500-μm-thick slices with a vibroslicer. Individual slices were trimmed to remove the surrounding areas of the hippocampal section. After recovering for 1–3 h in a holding chamber containing oxygenated artificial cerebrospinal fluid (aCSF), the hippocampal slices were treated with Aβ1–42 and/or neurotrophins for 30 min and used for electrophysiological recordings, Western blotting, or immunohistochemical analysis.
Hippocampal slices were placed in a submerged recording chamber and perfused with 29–30°C oxygenated aCSF at 2–3 ml/min. Field potentials were evoked and recorded in the CA1 region and dentate gyrus using standard extracellular recording techniques as described . In the CA1 region, Schaffer collateral-evoked field excitatory postsynaptic potentials (fEPSPs) were recorded in the stratum radiatum. In the dentate gyrus, the medial perforant path was stimulated in the middle third of the molecular layer, and evoked population spike and fEPSPs were recorded from the granule cell body and molecular layers. Baseline responses were collected using 0.008 or 0.016 Hz test pulses that yielded 40–50% of the maximal evoked responses in aCSF. The early LTP was induced using a single high-frequency stimulation (HFS) of 100 Hz for 1 s at the same stimulus intensity as used for the baseline. The normal test pulses were resumed after the HFS. LTP was measured 30 or 45 min after HFS in the dentate or CA1 pathways and expressed as percent changes of evoked responses from the average baseline values.
Drug- and LTP-induced changes in the level of total CaMKII, phospho-Thr286-CaMKII (p-CaMKII), total GluR1, and phospho-Ser831-GluR1 (p-GluR1) were measured in hippocampal slice tissues following the protocols previously described with some modifications [15,20]. Individual hippocampal slices were quickly frozen on dry ice after treatment and homogenized in ice-cold buffer containing 50 mM Tris/HCl (pH 7.0), 50 mM NaF, 10 mM EGTA, 10 mM EDTA, 80 mM sodium molybdate, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 0.01% Triton X-100 and 4 mM para-nitrophenylphosphate. To stop the activity of proteases and protein phosphatases, the homogenization buffer also contained Protease Inhibitors Complete (Roche Molecular Biochemicals) and Protein Phosphatase Inhibitor Cocktail I and II (Sigma-Aldrich). Aliquots of homogenates were taken to determine protein concentrations using a Bio-Red Protein Assay Kit. A cold denaturing loading buffer was added to the remaining homogenates, and sample aliquots containing 20 μg of protein each were electrophoresed on 12% SDS/PAGE gels. The proteins were transferred onto nitrocellulose membranes that were probed overnight at 4°C with various primary antibodies and incubated for 2 h with horseradish peroxidase-conjugated secondary antibodies. The immunoreactive bands were visualized with enhanced chemiluminescence (ECL-plus, GE Healthcare), and signals were quantified with densitometry using ImageQuant TL software (GE Healthcare). For loading controls, all blots were reprobed with a β-actin antibody (Sigma-Aldrich) and the signal value for each band of interest was normalized to that of β-actin. As multiple slices from the same rats were used for different drug treatments, percent changes in the protein of interest due to treatments were calculated relative to aCSF-treated controls run in the same experiments using slices from the same animals.
After drug treatment, hippocampal slices were fixed using 4% paraformaldehyde overnight, cryoprotected in phosphate-buffered saline (PBS) containing 30% sucrose, and cut into 40 μm cryostat sections. The sections were washed with PBS and incubated at 24°C for 30 min in a blocking solution containing 20% goat serum, 0.2% Triton X-100, and 0.1% bovine serum albumin in PBS. The brain sections were then probed with the primary antibodies against p-CaMKII and p-GluR1 at 24°C for 3 h, followed by 4°C overnight. After washing, brain sections were further incubated at 24°C for 2 h with Cy3- (1:2000; Jackson Laboratory) and FITC-conjugated (1:1000; Abcam) secondary antibodies. Confocal fluorescent images of double labeling were obtained using a Zeiss LSM 5 Pascal laser scanning microscope with a 5× water immersion objective. To control for nonspecific labeling, immunostaining was performed in control slices without adding the primary or secondary antibodies.
Aβ1–42 was purchased from Bachem (Torrance, CA) and prepared as previously described to yield oligomeric Aβ [9,15]. Brain-derived neurotrophic factor (BD-NF) and neurotrophin 4 (NT4) were obtained from Regeneron and Sigma-Aldrich. Drugs were diluted with aCSF to the desired final concentrations immediately before application. A monoclonal phosphorylation-site specific antibody recognizing phospho-Thr286-αCaMKII was obtained from Upstate (Lake Placid, NY) and used for Western blot (1:1000) or immunohistochemical analysis (1:100). A polyclonal phospho-specific antibody against phospho-Ser831-GluR1 was obtained from Abcam (Cambridge, UK) and used for Western blot (1:1000) or immunohistochemical analysis (1:200). The antibodies for total CaMKII (Upstate) and total GluR1 (Abcam) were used for Western blot analysis at 1:2000.
Data are presented as group means ± SE. One-way ANOVA was applied to test for overall statistical significance across multiple group means, followed by Bonferroni post hoc test for pairwise mean comparisons. Student’s t-test was used in LTP experiments for two-group comparisons. Statistical significance was defined as p < 0.05.
We first examined whether exogenously applied neurotrophin could affect synaptic plasticity and rescue impaired LTP in Aβ-treated hippocampal slices. As shown in Fig. 1, a single HFS induced early LTP in the Schaffer collateral-CA1 pathway of control slices, increasing the EPSP slope to 176 ± 8% of the baseline at 45 min after HFS (n = 23). The CA1 LTP was significantly reduced in slices pretreated with 1 μM Aβ1–42 for 30 min (137 ± 4%, n = 27, p < 0.01 compared to control LTP). NT4 pretreatment alone (100 ng/ml, 30 min) caused no evident changes in either baseline synaptic responses or early LTP (167 ± 9%, n = 9). However, when slices were co-treated with Aβ and NT4, the subsequent LTP (165 ± 9%, n = 11) was significantly higher than that in slices treated with Aβ alone (p < 0.01) and indistinguishable from control LTP (Fig. 1A, 1B). Similarly, in the dentate medial perforant path (Fig. 1C), early LTP induced in Aβ-treated slices (113 ± 5%, n = 15) was significantly smaller than that in control slices (132 ± 5%, n = 12, p < 0.05). NT4 treatment did not alter dentate LTP by itself (130 ± 5%, n = 17) but prevented LTP inhibition by Aβ when co-applied. The synaptic responses recorded 30 min after HFS in slices treated with both NT4 and Aβ (134 ± 6%, n = 19) was significantly higher than that in slices treated with Aβ alone and indistinguishable from the control LTP. Thus, when it did not affect LTP under control conditions, NT4 could rescue Aβ-induced LTP deficits in two major hippocampal pathways.
Next, we determined whether neurotrophin treatment could enhance CaMKII autophosphorylation and function in hippocampal neurons. Hippocampal slices were incubated in oxygenated aCSF containing 50 or 100 ng/ml BDNF or NT4 for 30 min and immediately processed for Western blot analysis. As shown in Fig. 2A, while neither affected the level of total CaMKII, both BDNF and NT4 enhanced CaMKII autophosphorylation in a dose-dependent manner. The p-CaMKII level was increased by 68–130% after NT4 treatment and by 97–143% after BDNF treatment, compared to the level in control slices treated with drug-free aCSF (n = 15 for all groups, p < 0.05 or 0.01). Consistent with upregulation of CaMKII activity, CaMKII-dependent phosphorylation of AMPA receptors at Ser831-GluR1 was dose-dependently increased in BDNF- or NT4-treated slices (Fig. 2B). The p-GluR1 level was increased by 80–131% and 89–172%, respectively, after NT4 or BDNF treatments (n = 15, p < 0.05 or 0.01 compared to the controls). No significant changes in the total GluR1 level were detected. Immunohistochemical analysis using phosphospecific antibodies showed increased fluorescent labeling for p-CaMKII and p-GluR1 in the CA1 pyramidal cell layer, dentate granule cell layer and respective dendritic layers (Fig. 3).
Last, we examined whether neurotrophins can oppose Aβ action, promoting LTP-associated CaMKII and GluR1 phophorylation. Hippocampal slices were pretreated with neurotrophin alone or combined with Aβ for 30 min, received a single HFS to induce dentate LTP, and harvested at 15 min after HFS. The dentate area of the slices was quickly dissected out, frozen on dry ice, and processed for Western blot analyses. To determine drug-induced changes in basal phosphorylation without induction of LTP, additional slices from the same animals were used as paired controls that received matching drug treatment but not HFS. As shown in Fig. 4, Aβ1–42 treatment (1 μM) did not affect the basal level of p-CaMKII and p-GluR1, but blocked HFS-induced increases in CaMKII and GluR1 phosphorylation. Neurotrophin treatment (100 ng/ml) increased p-CaMKII and p-GluR1 levels and prevented their inhibition by Aβ during LTP. In comparison, NT4 appeared to induce less basal phosphorylation than BD-NF, which allowed further increases in phosphorylation by HFS relative to paired, unstimulated slices (Fig. 4A). We therefore further focused on the effect of NT4 on LTP-associated CaMKII and GluR1 phosphorylation in the presence or absence of Aβ (Fig. 4B, 4C). Following the HFS, the p-CaMKII level in slices pretreated with both NT4 and Aβ increased to 163 ± 12% of the control level, which was significantly higher than the post-HFS level in slices treated with Aβ alone (104 ± 13%, p < 0.01) and comparable to the post-HFS level in slices treated with aCSF or NT4 alone (146 ± 10% or 169 ± 11%, p > 0.05 for both comparisons). In slices not subject to HFS, NT4 treatment significantly elevated the basal p-CaMKII level (131 ± 9% of the controls, p < 0.05) and this effect was not blocked by co-treatment with Aβ (Fig. 4B). Similar patterns of changes in p-GluR1 levels were induced by NT4 and/or Aβ under basal condition and during LTP (Fig. 4C).
TrkB-acting neurotrophins play a crucial role in hippocampal synaptic plasticity in the adult brain. Targeted deletion of the gene for BDNF [21,22] or trkB receptors  in mice results in severe impairments in early and late LTP at CA1 synapses and behavioral deficits in hippocampus-dependent tasks. Exogenously applied BDNF, on the other hand, restores hippocampal LTP in brain slices derived from BDNF knockout mice  and other mutant mice lacking gene products required for activity-dependent synaptic plasticity [25,26]. Although not as well-known as BDNF, NT4 has also been implicated in maintaining hippocampal synaptic function. The NT4 knockout mice display normal basal synaptic transmission and short-term synaptic plasticity but impaired late LTP at hippocampal CA1 synapses and show long-term memory deficits during contextual fear conditioning . When replacing the BDNF coding sequence with the NT4 sequence in a knock-in mouse, expression of NT4 completely rescues BD-NF deletion-induced morphological and biochecmical phenotypes and enhances formation of hippocampal synapses . In the present study, application of exogenous NT4 did not affect hippocampal basal synaptic transmission and early LTP, but it was able to prevent LTP deficits induced by Aβ when co-applied. This LTP-rescuing action of NT4 was reminiscent to that previously reported for exogenously applied BDNF. These findings support the notion that trkB-acting neurotrophins not only are involved in maintaining normal synaptic function, but can also prevent or alleviate neurodegenerative disease-related synaptic dysfunction. Furthermore, BDNF and NT4 strongly enhanced CaMKII autophosphorylation, a signaling event stimulated by LTP but suppressed by Aβ. These results are consistent with the proposed role of CaMKII as a key regulatory target for both Aβ toxicity and neurotrophin rescue.
Previous studies have indicated that LTP enhancement by BDNF can be mediated at presynaptic sites by increasing glutamate release . Evidence also exists for postsynaptic action of BDNF which evokes depolarization and Ca2+ transients in dendrites and spines of the postsynaptic neuron . More recently, it has been shown that application of BDNF simultaneously stimulates pre- and postsynaptic NMDA receptors in hippocampal slices . As membrane depolarization and subsequent Ca2+ influx through NM-DA receptor channels is the major driving force for CaMKII autophosphorylation, these actions of neurotrophins could readily lead to enhanced CaMKII autophosphorylation as seen in this study, and the resulting CaMKII activation is likely to occur at both pre-and postsynaptic sites. Autophosphorylation at Thr286 is known to render persistent, Ca2+- and calmodulin-independent CaMKII activity and promote translocation of the kinase to the postsynaptic density . This process enables CaMKII to modulate multiple synaptic proteins in a highly efficient manner and hence enhances LTP. At the presynaptic site, CaMKII-mediated phosphorylation of synapsin 1 promotes its dissociation from synaptic vesicles, causing increased neurotransmitter release . At the postsynaptic site, the best known substrate of CaMKII is the AMPA type of glutamate receptors. CaMKII-induced phosphorylation of AMPA receptors at Ser831-GluR1 increases channel conductance of the receptor , which can directly contribute to LTP expression. Consistent with this scenario, our results showed significant increases in GluR1 phosphorylation after neurotrophin treatments, in parallel with enhanced CaMKII autophosphrylation and LTP rescue. Activation of postsynaptic CaMKII is also known to trigger rapid delivery of AM-PA receptors from non-synaptic sites into the synapses  and promote activity-dependent growth of dendritic spines and formation of new synapses . In contrast, Aβ reportedly reduces the number of synaptic AMPA receptors [37,38], blocks activity-dependent AMPA receptor phosphorylation , and may cause loss of dendritic spines in the hippocampus . Taken together, CaMKII activity and Aβ have opposing effects in regulating multiple aspects of synaptic function. The substantial enhancement of CaMKII activity by neurotrophins could therefore well explain their LTP-rescuing effect in the presence of Aβ. It should be mentioned, however, our study does not exclude the possibility that neurotrophins can act through other signaling pathways to rescue synaptic function. Through trkB receptors, neurotrophins can activate multiple signaling cascades, including phosphoinositide 3-kinases, phospholipase C, and mitogen-activated protein kinases (MAPK) . Future studies are needed to determine the involvement of these pathways in the protective effect of neurotrophins.
The reversal of Aβ-induced synaptic and signaling impairments by BDNF or NT4 indicates therapeutic potentials of trkB-acting neurotrophins for AD. There is ample evidence showing rapid regulation of neurotrophin synthesis by neuronal activity and release of endogenous neurotrophins in an activity-dependent manner from neuronal dendrites . The increased neurotrophin level at synapses provides strong support to synaptic plasticity  and protects neurons from excitotoxic stress and oxidative insults . These properties of neurtrophins, combined with the finding that BDNF expression is significantly reduced in the hippocampus and frontal cortex of AD brain [43,44], make neurotrophin-based strategies attractive candidates for AD treatment. Previous studies have investigated the protective role of neurotrophins in AD, focusing to large extent on their ability to improve neuronal cell survival against Aβ toxicity . Further studies, however, are in need to determine whether and how neurotrophins can reverse Aβ-induced synaptic dysfunction via regulation of key signaling pathways involved in synaptic modulation, such as CaMKII and MAPK. Elucidation of such mechanisms is particularly relevant for treatment of early AD in which synaptic and cognitive impairments may be developed prior to synaptic loss and neuronal cell death. Promising new approaches that upregulate neurotrophin signaling, such as development of functional mimetics of neurotrophins capable of penetrating blood-brain-barrier  and utilization of neural stem cell-derived BDNF , have provided valuable tools for further investigation in this area.
This study is supported by National Institute of Health Grant R01AG-17542 to C.W. Xie.
We wish to thank Dr. Joseph B. Watson, Hongmei Ruan, and Kevah Navab for their input and assistance to this project.
Authors’ disclosures available online (http://www.j-alz.com/disclosures/view.php?id=419).