Amyloid-β Oligomers Regulate ADAM10 Synaptic Localization Through Aberrant Plasticity Phenomena

A disintegrin and metalloproteinase 10 (ADAM10) is a synaptic enzyme that has been previously shown to limit amyloid-β1–42 (Aβ1–42) peptide formation in Alzheimer’s disease (AD). Furthermore, ADAM10 participates to spine shaping through the cleavage of adhesion molecules and its activity is under the control of synaptic plasticity events. In particular, long-term depression (LTD) promotes ADAM10 synaptic localization triggering its forward trafficking to the synapse, while long-term potentiation elicits ADAM10 internalization. Here, we show that a short-term in vitro exposure to Aβ1–42 oligomers, at a concentration capable of inducing synaptic depression and spine loss, triggers an increase in ADAM10 synaptic localization in hippocampal neuronal cultures. However, the Aβ1–42 oligomers-induced synaptic depression does not foster ADAM10 delivery to the synapse, as the physiological LTD, but impairs ADAM10 endocytosis. Moreover, Aβ1–42 oligomers-induced inhibition of ADAM10 internalization requires neuronal activity and the activation of the NMDA receptors. These data suggest that, at the synaptic level, Aβ1–42 oligomers trigger an aberrant plasticity mechanism according to which Aβ1–42 oligomers can downregulate Aβ generation through the modulation of ADAM10 synaptic availability. Moreover, the increased activity of ADAM10 towards its synaptic substrates could also affect the structural plasticity phenomena. Overall, these data shed new lights on the strict and complex relationship existing between synaptic activity and the primary mechanisms of AD pathogenesis. Electronic supplementary material The online version of this article (10.1007/s12035-019-1583-5) contains supplementary material, which is available to authorized users.


Background
Disentangling the initial steps of Alzheimer's disease (AD) pathogenesis from full-blown pathology at a molecular and cellular level remains a key step to fully understand disease onset and progression. In this frame, it has been shown that synapse dysfunction and spine loss represent an early event of the disease rather than just a consequence of cell death [1]. Further, the synapses have been shown to be the main target of the amyloid-β 1-42 (Aβ 1-42 ) peptide, whose deposition is one of the main hallmarks of AD [2,3]. The Aβ 1-42 peptide derives from a transmembrane protein, named amyloid-β precursor protein (APP), that is mainly localized in the pre-synaptic active zone and in the post-synaptic compartment in the hippocampus and in the cortex [4].
The Aβ 1-42 homeostasis is regulated by synapse activation: increased activity enhances secretion of Aβ 1-42 , while reduced activity inhibits it [8,9]. Coherently, also ADAM10 synaptic levels and activity towards APP are under the control of activitydependent synaptic plasticity [10]. Long-term depression (LTD) boosts ADAM10 membrane insertion by fostering its SAP97mediated forward trafficking to post-synaptic membrane, whereas long-term potentiation (LTP) reduces the enzyme membrane levels by inducing AP2-dependent endocytosis [10].
On the other hand, Aβ 1-42 can be considered a regulator of neuronal activity [8] since once released, it affects in turn synaptic transmission and plasticity. In particular, pathological Aβ 1-42 levels giving rise to the formation of Aβ 1-42 oligomers (oAβ  ) may indirectly cause a partial block of N-methyl-D-aspartate (NMDA)-type glutamate receptors and shift the activation of NMDA receptors-dependent signaling cascades towards pathways involved in the induction of LTD and synaptic loss [11][12][13].
Here, we hypothesize that the oAβ 1-42 -induced plasticity pathways have a feedback effect on ADAM10 synaptic localization. We show that short-term exposure to oAβ 1-42 reduces ADAM10 endocytosis, thus leading to an increase in ADAM10 synaptic localization. This effect is mediated by activation of synaptic NMDA receptors containing the GluN2A subunit.

Characterization of oAβ 1-42 Effect on the Synapse
In order to set up a reliable in vitro system to analyze the effect of oAβ 1-42 on ADAM10 synaptic localization, we performed a complete characterization of our experimental conditions. First, oAβ 1-42 preparation was monitored and controlled by different means. As a negative control, we used a peptide with the reverse sequence of Aβ (Aβ 42-1 ). Coomassie staining and Western Blot analysis with an antibody detecting the N-terminus of Aβ (6E10 antibody) indicated that our oAβ  preparation resulted in a spectrum of oligomeric Aβ species, from 4 up to 16 kDa (Fig. 1a). Transmission electron microscopy (TEM) analysis of oAβ  confirmed the presence of globular, oligomeric structures, while no fibrillar or protofibrillar species were observed (Fig. 1b). Aβ 42-1 preparation is mainly constituted of monomers and of aggregated species. Western Blot analysis performed with 6E10 antibody showed no signal in Aβ 42-1 samples, as expected (Fig. 1a).
Consistent with previous studies [14,15], we found that oAβ 1-42 species at a concentration of 500 nM applied to hippocampal neurons for 24 hours (h) resulted in a significant reduction in spine density as compared to non-treated cultures (CTRL) or to cells exposed to Aβ 42-1 (Fig. 1c). Moreover, such concentration of oAβ 1-42 did not increase the mortality of the cells after 24 h of treatment, different from higher concentrations as 1 μM, 5 μM, and 10 μM (Fig. 1d).
oAβ  species have been shown to facilitate synaptic depression of neurons in acute slices [13,16]. Accordingly, hippocampal cultures exposed to oAβ 1-42 for 30 min displayed a global synaptic depression as indicated by electrophysiological recordings of excitatory post-synaptic current in miniature (mEPSCs) (Fig. 1e) as well as upon the delivery of the classical LTD protocol (20 μM NMDA and 20 μM glycine for 3 min, chemical LTD; Fig. 1e; [17,18]). Moreover, we used a biochemical approach and we purified the Triton-insoluble fraction (TIF) that is enriched in postsynaptic proteins. As shown in Fig. 1f, the treatment with oAβ 1-42 (500 nM, 30 min) induced a significant dephosphorylation of serine 845 of the GluA1 subunit of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, without any significant change in GluA1 synaptic levels, confirming that this concentration elicits synaptic depression [19]. No changes in the total and synaptic levels of NMDA receptor subunits and in PSD-95 were observed ( Fig. 1f; Suppl. Fig. 1A, B).
These results suggest that our oAβ 1-42 preparation triggers a synaptic depression affecting specifically AMPA mediated conductance.
To determine the cellular mechanism underlying oAβ 1-42 -induced increase in ADAM10 synaptic localization, we analyzed the association of ADAM10 to SAP97 and AP2 complex. As shown in Fig. 2c, the oAβ 1-42 treatment does not alter the interaction with SAP97, while significantly decreases the interaction with β2adaptin, one of the subunits of AP2 complex. These results demonstrate that acute exposure to oAβ 1-42 affects ADAM10 synaptic localization because of a decrease of endocytosis rather than to an increase of forward trafficking. To identify which of the cellular pathways triggered by oAβ 1-42 is responsible for the increase in ADAM10 synaptic localization, we analyzed the effect of the blockade of different events. First, we took advantage of the action potential blocker tetrodotoxin (TTX) (500 nM), and we observed that TTX pre-incubation prevents the oAβ 1-42 -induced increase of ADAM10 in the TIF (Fig. 3a), indicating the involvement of neuronal synaptic activity.
In the adult forebrain, synaptic NMDA receptors are predominantly di-heteromeric GluN1/GluN2A and tri-heteromeric GluN1/GluN2A/GluN2B receptors [20,21]. In light of this observation and considering the key role of GluN2A-containing NMDA receptors in plasticity phenomena [22,23], we examined the effect of NVP-AAM077, a GluN2A-preferring antagonist [24]. The analysis of ADAM10 levels in the TIF shows that the presence of NVP-AAM077 fully prevents the increased localization of ADAM10 in the synapses induced by oAβ   (Fig. 3b). Notably, the presence of ifenprodil, an antagonist of GluN2Bcontaining NMDA receptors, does not affect the oAβ 1-42 -triggered increase in ADAM10 synaptic levels, thus indicating the specific involvement of GluN2A-containing NMDA receptors (Suppl. Fig. 3).

Discussion
In this study, we provide evidence for aberrant plasticity phenomena by which oAβ 1-42 control synaptic function and the in the TIF of primary hippocampal neurons. The quantitative analysis shows that oAβ 1-42 increase ADAM10 synaptic localization (CTRL 100 ± 16.55%, Aβ 42-1 102.7 ± 18.9%, oAβ 1-42 222 ± 32.46%; *p < 0.05 one-way ANOVA, n = 3). c Representative images of coimmunoprecipitation between ADAM10 and either SAP97 or β2adaptin. The quantitative analysis shows that oAβ 1-42 treatment reduces the interaction with β2-adaptin without affecting the interaction with SAP97 (SAP97 96.95 ± 5.26%, β2-adaptin 57.39 ± 17.81%, *p < 0.05 t test oAβ 1-42 vs Aβ 42-1 , n = 3) generation of Aβ 1-42 itself. Primary hippocampal cultures were treated with a preparation of oAβ 1-42 able to induce synaptic depression in 30 min and spine loss in 24 h, in the absence of cell death. In these experimental conditions, oAβ 1-42 short-term exposure results in an increase in ADAM10 synaptic availability. Although it has been shown that LTD fosters SAP97-mediated ADAM10 delivery to the postsynaptic compartment [10], oAβ 1-42 treatment leads to a decrease in the association between ADAM10 and AP2 complex, which is responsible for the endocytosis of the enzyme, without affecting the binding to SAP97 (Fig. 3c). Therefore, the increase in ADAM10 synaptic localization is due to the impairment of its endocytosis rather than to a stimulation of its forward trafficking, suggesting that the molecular pathways underlying physiological LTD and regulating ADAM10 are profoundly different from those responsible for oAβ Considering that several studies demonstrated that oAβ 1-42 inhibit the maintenance of hippocampal LTP [8,25,26] and that ADAM10 endocytosis is regulated by LTP, we can also hypothesize that the oAβ 1-42 -induced impairment in ADAM10 endocytosis is in line with a disorder in mechanism of LTP. Thus, overall our data indicate that oAβ 1-42 triggers aberrant plasticity phenomena.
We provide also a mechanistic inside for this aberrant plasticity of oAβ 1-42 action. oAβ 1-42 engage a pathway that requires neuronal activity and the activation of the GluN2A-containing NMDA receptors. Even if several studies reported the role of the extra-synaptic GluN2B-containing NMDA receptors [27][28][29], our data highlight that acute exposure to oAβ 1-42 triggers a synaptic event that involves GluN2A-containing NMDA receptors. Indeed, the presence of an inhibitor of GluN2B-containing NMDA receptors does not prevent the oAβ 1-42 -induced increase in ADAM10 synaptic localization.
Taken together, these data suggest that a short-term exposure to oAβ 1-42 engages a negative feedback mechanism according to which oAβ 1-42 can downregulate Aβ generation through the modulation of ADAM10 synaptic availability. Moreover, the increased activity of ADAM10 towards its synaptic substrates could tune synaptic transmission and structural plasticity. It has been shown that the sAPPα, which is released after ADAM10 cleavage of APP, is able to acutely modulate synaptic strength when applied in vitro [30]. Furthermore, ADAM10-mediated shedding of N-cadherin controls spine shaping and AMPA receptors function [31].
Overall, here, we show that oAβ 1-42 can trigger aberrant plasticity pathways and, thereby, affect synaptic plasticity. Since the synapses are considered to be an early site of pathology in AD [32] and loss of synapses is the best pathologic correlate of cognitive impairment in AD patients [33], understanding the molecular underpinnings leading to synaptic dysfunction will aid in the development of tailored synapsetargeted therapies for AD.

Neuronal Cultures Preparation, Transfection, and Treatments
Primary hippocampal neurons cultures were prepared from embryonic day 18-19 rat hippocampi as previously described [35]. The Institutional Animal Care and Use Committee of University of Milan and the Italian Ministry of Health (#326/2015) approved all the experiments involving primary neuronal cultures preparation.

Synaptic Fraction Purification, Western Blot, and Co-immunoprecipitation Analysis
After treatment, samples were processed for the purification of the Triton-insoluble fraction (TIF), a fraction enriched in postsynaptic density proteins [35]. After quantification, total homogenate and TIF proteins were resolved with SDS-PAGE method; co-immunoprecipitation experiments were performed as described in [10,35].

Immunocytochemistry and Confocal Microscope Acquisition
For ADAM10/PSD-95 co-localization and spine morphology studies, treated hippocampal neurons were fixed 7 min in 4% paraformaldehyde plus 4% sucrose in phosphate buffered saline (PBS) at room temperature. Then, cells were extensively washed with PBS supplemented with CaCl 2 and MgCl 2 , permeabilized with 0.2% Triton-X100 and incubated for 2 h at room temperature with 5% BSA in PBS. Primary and secondary antibodies were applied in 5% bovine serum albumin (BSA) in PBS. Cells were labeled with primary antibodies overnight at 4°C. Cells were washed and then incubated with secondary antibodies for 1 h at room temperature. Cells were then washed in PBS and mounted on glass slides with Fluoromount mounting medium (Sigma-Aldrich). Fluorescence images were acquired by using Zeiss Confocal LSM510 system (Zeiss, Jena, Germany) with a sequential acquisition setting at 1024 × 1024 pixels resolution; for each image, two up to four 0.5-μm sections were acquired and a z projection was obtained [31]. HEPES-NaOH, pH 7.4 in which also TTX (0.5 μM), bicuculline (20 μM, Tocris, Bristol, UK), and strychnine (1 μM, Sigma-Aldrich, Milan, Italy) were included. Then, to induce chemical LTD, we applied NMDA (20 μM) and glycine (20 μM, Sigma-Aldrich, Milan, Italy) for 3 min at room temperature in Mg 2+ -free KRH containing TTX (0.5 μM), bicuculline (20 μM), and strychnine (1 μM). Thirty minutes after this treatment, mEPSCs have been collected again in the starting KRH solution. Synaptic depression has been also induced in cultures by 30 min of oAβ 1-42 (500 nM) in Mg 2+free KRH containing only TTX (0.5 μM), bicuculline (20 μM) and strychnine; at the end of this treatment, mEPSCs have been recorded in normal KRH and analyzed. Off-line analysis of miniature events was performed by the use of Clampfit-pClamp-10 software.

Data Quantification and Statistical Analysis
Quantification of Western Blot analysis was performed by means of computer-assisted imaging (Image Lab, Biorad). The levels of the proteins were expressed as relative optical density (OD) measurements and normalized on tubulin. Values are expressed as mean ± S.E.M. of at least three independent experiments.
Co-localization analysis was performed using Zeiss AIM 4.2 software and spines analysis was performed with ImageJ software (National Institute of Health, Bethesda, MD, USA). For co-localization, and morphological analysis, cells were chosen randomly for quantification from 4 different coverslips (2-3 independent experiments), images were acquired using the same settings/exposure times, and at least 10 cells for each condition were analyzed. Statistical evaluations were performed by using 2-tailed Student's t test (a p value less than 0.05 was considered significant) or, when appropriate, by using one-way ANOVA followed by Bonferroni's post hoc test or Kruskal-Wallis analysis of variance followed by Dunn's post hoc test.