Amyloid Aβ25-35 Aggregates Say ‘NO’ to Long-Term Potentiation in the Hippocampus through Activation of Stress-Induced Phosphatase 1 and Mitochondrial Na+/Ca2+ Exchanger

The search for strategies for strengthening the synaptic efficiency in Aβ25-35-treated slices is a challenge for the compensation of amyloidosis-related pathologies. Here, we used the recording of field excitatory postsynaptic potentials (fEPSPs), nitric oxide (NO) imaging, measurements of serine/threonine protein phosphatase (STPP) activity, and the detection of the functional mitochondrial parameters in suspension of brain mitochondria to study the Aβ25-35-associated signaling in the hippocampus. Aβ25-35 aggregates shifted the kinase–phosphatase balance during the long-term potentiation (LTP) induction in the enhancement of STPP activity. The PP1/PP2A inhibitor, okadaic acid, but not the PP2B blocker, cyclosporin A, prevented Aβ25-35-dependent LTP suppression for both simultaneous and delayed enzyme blockade protocols. STPP activity in the Aβ25-35-treated slices was upregulated, which is reverted relative to the control values in the presence of PP1/PP2A but not in the presence of the PP2B blocker. A selective inhibitor of stress-induced PP1α, sephin1, but not of the PP2A blocker, cantharidin, is crucial for Aβ25-35-mediated LTP suppression prevention. A mitochondrial Na+/Ca2+ exchanger (mNCX) blocker, CGP37157, also attenuated the Aβ25-35-induced LTP decline. Aβ25-35 aggregates did not change the mitochondrial transmembrane potential or reactive oxygen species (ROS) production but affected the ion transport and Ca2+-dependent swelling of organelles. The staining of hippocampal slices with NO-sensitive fluorescence dye, DAF-FM, showed stimulation of the NO production in the Aβ25-35-pretreated slices at the dendrite-containing regions of CA1 and CA3, in the dentate gyrus (DG), and in the CA1/DG somata. NO scavenger, PTIO, or nNOS blockade by selective inhibitor 3Br-7NI partly restored the Aβ25-35-induced LTP decline. Thus, hippocampal NO production could be another marker for the impairment of synaptic plasticity in amyloidosis-related states, and kinase–phosphatase balance management could be a promising strategy for the compensation of Aβ25-35-driven deteriorations.


Introduction
Alzheimer's disease (AD) is actually a severe challenge for basic and translational neuroscience, annually affecting millions of new patients and creating billions of dollars burden around the world [1,2]. Mechanistically, the progression of AD is correlated with the accumulation of misfolded structural proteins and their fragments in the brain, leading to neuroinflammation, neuronal loss, and, ultimately, a memory decline accompanied by cognitive dysfunctions [1,3]. The forming of amyloid fibrils, plaques, and agglomerates of misfolded tau protein in different models of AD pathology is well-documented, including post-mortem analysis of brains from people with AD [4][5][6]. Functionally, amyloid aggregates strongly suppress synaptic plasticity, leading to memory impairment in both 2 of 25 invertebrate and vertebrate animal species, which apparently indicates there are conservative mechanisms mediating synaptic deteriorations [7][8][9]. One of the key amyloid peptides established in the respect of neuronal excitotoxicity is Aβ [25][26][27][28][29][30][31][32][33][34][35] , which has oligomers that disrupt the long-term potentiation (LTP) in crucial brain structures [10][11][12]. Unambiguous targets for Aβ [25][26][27][28][29][30][31][32][33][34][35] -driven signaling are still unknown, which is due to the complexity and prolongation of AD-developed events. Different studies indicate that amyloidosis could be accompanied by the downregulation of muscarinic [13,14], nicotinic cholinoreceptors' sensitivity [2,15], AMPA receptor trafficking dysregulation [16,17], changes in the NMDAreceptor-mediated fluxes [18,19], upregulation of voltage-gated Ca 2+ channels [20,21], or outward K + current inhibition in neurons, or by its activation in microglia [22,23], the activation of ryanodine and inositol triphosphate receptors [24,25], or global mitochondrial bioenergetic disturbances [26,27]. Several studies showed the involvement of phosphatases in the development of a number of neurodegenerative models, including AD and parkinsonism. Thus, in mutant mice overexpressing the amyloid precursor protein (APP), an increase in the activity of PTP1B tyrosine phosphatase was found, which correlates with the development of neuroinflammation and the progression of cognitive deficit. Systemic blockade of PTP1B phosphatase partially prevented the deterioration of synaptic plasticity [28]. Another tyrosine phosphatase and tensin homolog (PTEN), which negatively regulates the prosurvival pathway PI3K-Akt-mTOR, increased the expression 48 h after Aβ [25][26][27][28][29][30][31][32][33][34][35] application to organotypic cultures, which correlates with cell death [29]. A decrease in the activity of neuronal-specific striatal-enriched tyrosine phosphatase, STEP, prevented the internalization of the GluA1/GluA2 subunits of glutamate receptors, which is responsible for the partial suppression of the synaptic transmission induced by the action of amyloid aggregates [30]. Sparse and somewhat controversial data also exist for serine/threonine phosphatases. Thus, crosstalk between impaired phosphatase 2A (PP2A) activity and the cancerous inhibitor protein of PP2A (CIP2A) expression enhancement is responsible for the chronic increase in hyperphosphorylated tau protein [31]. At the same time, incubation of amyloid oligomers caused the induction of PP2A activity, since the decrease in LTP in amyloid-treated slices was prevented by the specific PP2A inhibitor foestriscin, as well as the application of pyruvate [32]. The inconsistency of data regarding the activity of serine/threonine phosphatases during amyloidosis progression indicates the complexity of amyloid-driven neurochemical events in the hippocampus, and, probably, at least in part, can be explained by the existence of critical 'time windows' in which the kinase-phosphatase balance can be bidirectionally switched.
In addition, the mechanisms underlying the pathogenesis of neurodegenerative states are closely associated with mitochondrial dysfunction and oxidative stress, an imbalance between the production and utilization of reactive oxygen species (ROS). According to the hypothesis of the 'mitochondrial cascade' [33,34], the decrease in the synthesis of adenosine triphosphate (ATP) and oxidative stress leads to an excessive synthesis of APP, which can directly have toxic effects on mitochondria, aggravating neurodegenerative processes [33,35]. Mitochondrial dysfunction results in suppressing the bioenergetics processes in nerve cells, the damage of membrane structures by free radicals, neuroinflammation, violation of the synaptic transmission, increase in the release of glutamate from the presynaptic terminals, decrease in the plasticity of synaptic contacts, and, ultimately, the death of neurons [35][36][37][38]. The ability of amyloid oligomers to initiate multiple intracellular signal pathways exacerbates the neurodegenerative processes. In the brains of patients with Aβ-amyloidosis, amyloids are accumulated in mitochondria and impair the activity of the glycolysis and tricarboxylic acid cycle enzymes, stimulating ROS production [39][40][41]. The direct action of oligomers and amyloid fibrils on the outer mitochondrial membrane (OMM) is a disruption of energy metabolism and electron chain transport [34,42]. Currently, the relationships between mitochondrial dysfunction and the kinase-phosphatase balance during amyloid-associated synaptic deteriorations is also studied quite fragmentarily. In this work, we used field excitatory postsynaptic potential (fEPSPs) recording, fluorescent NO-sensitive imaging in slices, direct measurements of serine/threonine phosphatase activity, and recordings of functional mitochondrial parameters (transmembrane mitochondrial potential and the ROS production and swelling of organelles in the suspension) to describe the Aβ 25-35 -driven neurochemical events in the hippocampus. We observed that Aβ [25][26][27][28][29][30][31][32][33][34][35] -dependent suppression of LTP is strictly related to the induction of PP1 activity and enhancement of NO production in CA1, CA3, and the dentate gyrus. Curiously, a delayed PP1/PP2A blockade 1 h after the Ab 25-35 treatment of slices was quite effective in preventing fEPSP decline after LTP induction. The direct action of the Aβ 25-35 aggregates on mitochondria revealed a moderate effect in the speed of mitochondrial swelling, an effect that disappeared during the increase in external Ca 2+ in the chamber solution, which was well-correlated to the mitochondrial Na + /Ca 2+ exchanger (mNCX) operations. Thus, the mNCX blockade strongly rescued the late LTP phase in Aβ 25-35 -treated slices. The data obtained clarified immediate influences of the Aβ 25-35 oligomers on synaptic plasticity in the hippocampus and can be useful as a potential strategy for the compensation of some neurodegenerative states.

Aβ25-35 Aggregates Impaired Synaptic Plasticity through PP1/PP2A-Sensitive Mechanism
To prepare Aβ25-35 aggregate solution, we used once-frozen 1 mM Aβ25-35 aliquots dissolved in mQ water. The final Aβ25-35 concentration in the ACSF was adjusted to 50 nM, after which the solution was kept for 24 h at +4 °C for monomer aggregation ( Figure 1A).  (B) Averaged curves for control slices (blue circles, n = 13), okadaic acid (OA)-treated slices (100 nM, gray circles, n = 9), Aβ 25-35 -treated slices (dark red circles, n = 9), and OA+ Aβ 25-35 -treated slices (red circles, n = 9). Arrow at 0 min indicates start of tetanic stimulation, and dotted line 100% represents normalized fEPSPs corresponding to the pre-tetanic level. Right panels display typical fEPSP responses of studied groups for pre-tetanic levels (baseline), at the start of post-tetanic recordings (tetanus) and 180 min after LTP induction (L-LTP). (C) Summarized statistics for (B) curves during early phase (E-LTP, 0-3 min) and late phase (L-LTP, 178-180 min) of long-term potentiation. * denotes p < 0.05 for comparing group with control; # indicates p < 0.05 for comparison of groups between themselves. The number of independent experiments is indicated in parentheses. (D) Paired-pulse facilitation (PPF) curves at the 50 ms ISI, corresponding to the (B) recordings. PPF ratios reflect short-term plasticity (STP), relating to the residual Ca 2+ in the presynaptic endings, which increases fEPSPs in response to the second stimulus of the same strength. (E)-PPF ratios for different interstimulus intervals (30-400 ms) before tetanic stimulation (left panel, pre-tetanus) and for L-LTP, 3 h after tetanus (right panel, post-tetanus).

Aβ 25-35 Aggregates 'Say' NO to the LTP in CA3-CA1 Synapses
Previously, we showed that LTP suppression in several neuropathological conditions such as protein synthesis inhibitors' application was accompanied by an increase in the NO production, suggesting the potential role of NO in labeling some neurophysiological stress reactions [43,44]. In the next series of experiments, a direct estimation of NO production in the hippocampus was completed. In non-stimulated hippocampal slices, the treatment by 50 nM Aβ 25-35 significantly increased the NO-sensitive fluorescence of the DAF-FM probe in key hippocampal layers including CA1 somas (Stratum pyramidale), CA1 dendrites (Stratum radiatum, SR; and Stratum lacunosum-moleculare, SLM), CA3 dendrites, dentate gyrus (DG) dendrites, and somas ( Figure 6). which is closely related to the global changes in mitochondrial ion transport. Mitochondrial swelling was detected as a decrease in the optical density absorption spectra at 540 nm in different conditions: with the addition of 50 μM Ca 2+ to the organelle suspension ( Figure 5H right panels) or without it ( Figure 5H, left panels). An Aβ25-35 (10-200 nM) dose independently protected mitochondria from the swelling in the Ca 2+ -free medium and had no protection effect in the presence of 50 μM Ca 2+ . So, the optical densities in 10, 50, 100, 150, and 200 nM Aβ25-35-treated mitochondria without external Ca 2+ were 0.718 ± 0.014, 0.720 ± 0.011, 0.707 ± 0.015, 0.723 ± 0.019, and 0.722 ± 0.010 a.u., respectively (vs. 0.563 ± 0.010 a.u in control probes, for all n = 3, p ˂ 0.05, Figure 5H, left panels). Dose-independent protection against mitochondrial swelling disappeared in the presence of 50 µ M Ca 2+ (for example, optical density in the 10 nM Aβ25-35 was 0.371 ± 0.024 a.u. vs. 0.346 ± 0.026 a.u. in control, n = 3, p = 0.7, Figure 5H, right panels). These observations are well consistent with the mNCX mode operations. CGP37157-sensitive mNCX operation is responsible for the inverted ion antiport, increasing the Ca 2+ leakage from mitochondria into the cytosol. The mNCX-dependent Ca 2+ output somewhat increases mitochondrial Ca 2+ capacity, attenuating the swelling of mitochondria. In the conditions of increased external Ca 2+ , the mNCX drives the forward mode, loading Ca 2+ into mitochondria together with Na + outflow to the cytosol. Thus, inversion of the ion antiport for Ca 2+ and Na + attenuated the mNCX-dependent protection effect against mitochondrial swelling.

Discussion
Amyloid peptide aggregates noticeably change ion channel transport, leading to adaptive neurochemical stress responses in the neuronal and glial cells [45][46][47]. At the synaptic level, Aβ oligomers suppress post-tetanic and long-term potentiation, decreasing the number of released vesicles [48][49][50]. In accordance with these data, we revealed the decreased PPF ratio at 30-100 ms interstimulus intervals in the Aβ 25-35 -treated hippocampal slices ( Figure 1E). Furthermore, 1 h incubation of slices in the presence of 50 nM Aβ 25-35 disrupted the late LTP phase, 3 h after the tetanic stimulation ( Figure 1A,B), which was previously shown for the different stimulations' protocols of invertebrates and mammal species [7,[51][52][53]. Curiously, the Aβ 25-35 monomers in the freshly prepared amyloid solution (50 nM) did not change the LTP kinetics, after tetanic stimulation of the Aβ 25-35 -pretreated hippocampal slices (Supplementary Figure S5). Impairment of both short-term and long-term plasticity denotes the multiple influences of the Aβ 25-35 aggregates at pre-and postsynaptic endings. In this paper, we observed at least three important players: the PPPR15A subunit of the PP1 phosphatase complex (PP1α), the mitochondrial Na + /Ca 2+ exchanger (mNCX), and the neuronal NO synthase (nNOS) involved in the Aβ 25-35 -associated neurochemical events in the rat hippocampus.
The regulatory subunit of the protein phosphatase 1 complex (PPPR15A), well known as GADD34 (growth arrest and DNA damage-inducible gene), is a cell stress protein that recruits PP1α activity under conditions of oxidative, genotoxic, and other forms of stress [54,55]. Immunocytochemistry analysis revealed GADD34 localization both in the cytosol and membrane organelles, such as the endoplasmic reticulum, Golgi apparatus, and mitochondria [56]. GADD34 trafficking between cytosol and the intracellular compartments drives its proteasomal degradation, changing the dephosphorylation status of PP1-targeted proteins [56,57]. In our experiments, the specific blocker of GADD34, sephin1, strongly prevented Aβ 25-35 -dependent LTP suppression in the protocols of both the transient and chronic blockade of the enzyme (Figure 4). Curiously, sephin1 regulates integrated stress response (ISR), preventing the accumulation of misfolded proteins and promising therapeutic potency against neuronal excitotoxicity, multiple sclerosis, and neurodegeneration [58,59]. Moreover, GADD34-expression enhancement is a hallmark of some neuropathological states, including traumatic brain injury, cerebral ischemia/reperfusion injury, and glioblastoma [60][61][62]. Furthermore, increased GADD34 activity was observed in glial cells, such as oligodendrocytes during AD pathology development [63], and overexpression was observed in AD transgenic J20 mice [64].
Direct measurements of serine/threonine phosphatase (STPP) activity revealed that the incubation of hippocampal slices in the presence of Aβ 25-35 amyloid aggregates significantly elevates it ( Figure 2E). An Aβ 25-35 -mediated STPP activity increase was prevented by the PP1/PP2A blockade by okadaic acid but not by the PP2B blocker cyclosporin A ( Figure 2E). Considering the fact that a selective PP2A inhibitor, cantharidin, was ineffective in preventing Aβ 25-35 -dependent LTP suppression ( Figure 4F,H), the STTP increase is most likely related with the induction of PP1 activity rather than that of PP2A. It should be noted that the application of Aβ [25][26][27][28][29][30][31][32][33][34][35] aggregates switches the kinase-phosphatase balance to a prevalence of phosphatase activity ( Figure 2E), whereas during the E-LTP phase in control slices STPP activity was strongly downregulated ( Figure 2E). Probably, STPP-activity enhancement is partly involved in the impairment of synaptic transmission during LTP development in the Aβ 25-35 -pretreated slices. During the E-LTP phase, the stimulation of multiple protein kinases, including serine/threonine A, B, C, and G and Ca 2+ -calmodulindependent II occurs, which is crucial for the enhancement of synaptic transmission in tetanized slices [65][66][67]. The induction of phosphatase activity during the E-LTP phase can interfere the kinase-mediated signaling, leading to the disappearance of fEPSPs and the impairment of synaptic effectivity.
Earlier, we observed that the increase in PP2B activity during LTP suppression by protein synthesis inhibitors could be associated with the increase in nitric oxide (NO) production by neuronal NO synthase [68]. Thus, NO is a neuromediator, which supports synaptic plasticity and is involved in the reconsolidation and memory storage processes [69][70][71][72], and it can be a hallmark of synaptic deteriorations. NO-sensitive imaging of hippocampal slices reveals that Aβ [25][26][27][28][29][30][31][32][33][34][35] aggregates strongly increase the NO levels in both the somas and dendrites of the CA1 and DG layers, as well as in the CA3 inputs including Shaffer's collaterals (Figures 6 and 7). The PP2B inhibitor cyclosporin A was ineffective in preventing Aβ 25-35 -dependent NO synthesis (Figure 7). In contrast, a PP1/PP2A blocker, okadaic acid, partly prevented the Aβ 25-35 -mediated NO increase in the CA1 and DG somas, as well as in the CA1, CA3, and DG dendrites (Figure 7). The enhancement of neuronal NOS activity can be associated with the PP1-dependent changes in the phosphorylation status of NOS. Thus, the nNOS dephosphorylation, when 'inhibiting' Ser-847 (Ser-852 in mice) or Thr-1296 residues by the PP1 complex, increases the NO production in different cell models [73][74][75][76]. Furthermore, PP1 could indirectly lead to the enhancement of nNOS phosphorylation when 'stimulating' Ser-1417 (Ser-1412 in mice) [77][78][79], which is also well-correlated with the increase in NO production in the Aβ 25-35 -pretreated hippocampal slices.
The involvement of mitochondria in the development of neurodegenerative states, including AD, is well-known as the concept 'mitochondrial cascade' [33,34]. The interaction between oxidative stress and mitochondrial dysfunction and the impairment of Ca 2+ homeostasis induced by misfolded tau and β-amyloid play important roles in the progressive neuronal loss occurring in specific areas of the brain [80,81]. In our experiments, Aβ 25-35 aggregate mediated LTP suppression, which was partly prevented in the presence of CGP37187, a blocker of the reverse mode of the mitochondrial Na + /Ca 2+ exchanger (mNCX) (Figure 5A,C). The mNCX represents an ion transporter in the inner mitochondrial membrane, which carries three Na + outwards in exchange for the one Ca 2+ that is uploaded into the mitochondrion [82,83]. The reverse mode of the mNCX is responsible for the inverted ion antiport, increasing Ca 2+ leakage from mitochondria into the cytosol [84,85]. The heightened [Ca 2+ ] in could lead to the additional elevation of NO synthase activity, through Ca 2+ -calmodulin protein-protein interaction [86], and it can be involved in the regulation of synaptic transmission, through exocytosis of the neuromediator vesicles [87,88]. Apparently, incubation of hippocampal slices in the presence of Aβ 25-35 amyloid aggregates depletes the intracellular neuromediator sources, which is well-correlated with the significant decrease in the PPF ratio in the CA3-CA1 synapses ( Figure 1E). Although CGP37187 did not alter the short-term plasticity and had no influence on the LTP kinetics ( Figure 5A,C), the data obtained strongly indicate mNCX's participation in the Aβ [25][26][27][28][29][30][31][32][33][34][35] amyloid response in the rat hippocampus. At the same time, neuronal deletion of the mitochondrial Na + /Ca 2+ exchanger (NCLX, Slc8b1 gene) accelerated memory decline and increased amyloidosis, tau pathology, and, ultimately, memory loss in 3XTg-AD mice, [89]. Inhibition of the mNCX protects striatal neurons from α-synuclein cytotoxicity in models of Parkinson's disease [90]. In addition, NCX1, NCX2, and NCX3 isoforms were upregulated in the synaptic terminals, accumulating amyloid-beta (Aβ), the neurotoxic peptide responsible for AD neurodegeneration [91,92]. More recently, the hyperfunction of a specific NCX subtype, NCX3, has been shown to delay endoplasmic reticulum stress and apoptotic neuronal death in hippocampal neurons exposed to Aβ insult [93]. Direct measurements of the mitochondrial parameters indicated that Aβ 25-35 could directly regulate some mitochondrial functions, such as swelling and Ca 2+ uptake ( Figure 5H). Aβ [25][26][27][28][29][30][31][32][33][34][35] , in nanomolar concentrations, did not change the transmembrane mitochondrial potential ( Figure 5F) or ROS production ( Figure 5G). Curiously, the Aβ 25-35 effects, in respect to the swelling of brain mitochondria, disappeared with an increase in the external Ca 2+ in the chamber solution ( Figure 5H, right panels). This fact is in good agreement with the operation of the mNCX modes. At heightened external Ca 2+ concentrations, Ca 2+ transport from mitochondria through the mNCX is not working [85,94]. The insufficient influence of mNCX-mediated Ca 2+ fluxes, in respect to mitochondrial transmembrane potential, could be explained by the involvement in ∆Ψ m maintenance of major players, such as the mitochondrial Ca 2+ uniporter (MCU), which operates a large Ca 2+ influx to mitochondria, successfully compensating for the outwards mNCX-dependent Ca 2+ leak.
At a minimum, the Aβ 25-35 -driven PP1α-and mNCX-related events may occur independently, because purified mitochondrial suspension, apparently, contains low quantities of GADD34 protein. However, there are indirect data concerning the mNCX's modulation by mitochondrial serine/threonine PTEN-induced kinase 1 (PINK1) [95]. In this case, it is logical to suggest the participation of PP1α as another signaling player that provides negative feedback for the kinase-phosphatase switching for the mNCX's transport regulation. On the other hand, changes in the intracellular [Ca 2+ ] in , as in the case of the mNCX's reverse mode stimulation, could underlie the enhancement of PP1α activity. Thus, endoplasmic reticulum store depletion (ER stress) is sufficient to induce GADD34 mRNA expression, which can be the mechanism, for example, of the calcium homeostasis disturbances during transient forebrain ischemia in rats [96,97]. Thus, a positive correlation between PP1α and the mNCX blockade was observed, in respect to the prevention of the Aβ 25-35 aggregate-dependent LTP suppression (Supplementary Figure S6). Figure 9 presents a speculative scheme of Aβ 25-35 -driven signaling in the hippocampal neurons. Aβ [25][26][27][28][29][30][31][32][33][34][35] aggregates (in)directly induced GADD34 and PP1α phosphatase complex activation. PP1α can dephosphorylate nNOS, leading to the disinhibition of nNOS activity followed by NO production. NO acts both pre-and postsynaptically, reducing the number of mediator vesicles and nitrosylating the protein targets, such as ion channels, enzymes, etc. Furthermore, Aβ [25][26][27][28][29][30][31][32][33][34][35] aggregates stimulate the mNCX's transport, elevating [Ca 2+ ] in in the cytosol. An increase in [Ca 2+ ] in can, additionally, facilitate nNOS activity through the Ca 2+ -calmodulin binding site at the nNOS. Moreover, the elevation of [Ca 2+ ] in can provide additional GADD34 induction and the enhancement of PP1α phosphatase activity. PP1α in the postsynaptic cells interferes with kinase-mediated signals, preventing fEPSP maintenance after LTP induction. PP1α in the presynapse may limit vesicular exocytosis of the neuromediator, impairing short-term plasticity. mNCX's reverse mode stimulation, could underlie the enhancement of PP1α activity. Thus, endoplasmic reticulum store depletion (ER stress) is sufficient to induce GADD34 mRNA expression, which can be the mechanism, for example, of the calcium homeostasis disturbances during transient forebrain ischemia in rats [96,97]. Thus, a positive correlation between PP1α and the mNCX blockade was observed, in respect to the prevention of the Aβ25-35 aggregate-dependent LTP suppression (Supplementary Figure S6). Figure 9 presents a speculative scheme of Aβ25-35-driven signaling in the hippocampal neurons. Aβ25-35 aggregates (in)directly induced GADD34 and PP1α phosphatase complex activation. PP1α can dephosphorylate nNOS, leading to the disinhibition of nNOS activity followed by NO production. NO acts both pre-and postsynaptically, reducing the number of mediator vesicles and nitrosylating the protein targets, such as ion channels, enzymes, etc. Furthermore, Aβ25-35 aggregates stimulate the mNCX's transport, elevating [Ca 2+ ]in in the cytosol. An increase in [Ca 2+ ]in can, additionally, facilitate nNOS activity through the Ca 2+ -calmodulin binding site at the nNOS. Moreover, the elevation of [Ca 2+ ]in can provide additional GADD34 induction and the enhancement of PP1α phosphatase activity. PP1α in the postsynaptic cells interferes with kinase-mediated signals, preventing fEPSP maintenance after LTP induction. PP1α in the presynapse may limit vesicular exocytosis of the neuromediator, impairing short-term plasticity.

Hippocampal Slice Preparation
All experimental procedures were conducted in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC), on the protection of animals used for scientific purposes. The study protocol was approved by the Ethics Committee of the Institute of Higher Nervous Activity and Neurophysiology of RAS. Male Wistar rats (6-10 weeks old) were anesthetized by sevoflurane and decapitated. Brains were quickly submerged in ice-cold dissection solution (concentrations in mM: 124 NaCl, 3 KCl, 1.25 NaH 2 PO 4 , 26 NaHCO 3 , 0.5 CaCl 2 , 7 MgCl 2 , and 10 D-glucose, pH equilibrated with 95% O 2 -5% CO 2 ). Parasagittal hippocampal slices (400 µm) were prepared using a vibratome (Leica VT1000S, Wetzlar, Germany) and immediately transferred to a recording solution (ACSF, composition as above, except the CaCl 2 and MgCl 2 concentrations were adjusted to 2.5 and 1.3 mM, respectively). Slices were heated to 34 • C in a water bath for 40 min and then kept at room temperature. Amyloid Aβ 25-35 (H1192, Bachem, Bubendorf, Switzerland) was prepared 24 h before the slice incubation from 1 mM stock solution in mQ water, adjusting to 50 nM of final concentration in the ACSF. Hippocampal slices were incubated for 1 h with or without specific blocker, depending on the experimental groups.

Electrophysiology
During the experiments, slices were perfused by a continuous ACSF flow (appr. 4 mL/min) at 32-33 • C. Electrophysiological recordings were carried out using a SliceMaster system (Scientifica, Uckfield, UK). Field excitatory postsynaptic potentials (fEPSPs) were recorded from Stratum radiatum in the CA1 area using glass microelectrodes (1-2 MΩ) filled with the chamber solution. Baseline synaptic responses were evoked by paired-pulse stimulation of the Schaffer collaterals with 50 ms interval at 0.033 Hz with a bipolar electrode. Test stimulation intensity was adjusted, to evoke fEPSPs with amplitude 50% of maximal, and was kept constant throughout the experiment. Long-term potentiation (LTP) was induced with four 100 Hz trains spaced 5 min apart, as in [98]. The data were recorded and analyzed by Spike2 and SigmaPlot 11.0 (Systate Software Inc., Chicago, IL, USA). For statistical analysis, the first 3 min after tetanization (0-3 min, early LTP) and 3 min at 178-180 min after LTP induction (late LTP) were used. For baseline responses, the amplitudes and appropriate fEPSP slopes during test stimulation were evaluated. Paired-pulse facilitation (PPF) ratio was calculated as PPF = (S 2 EPSP/S 1 EPSP), where S 1 EPSP and S 2 EPSP are the slopes of fEPSP in responses to the first and the second stimuli, respectively. PPF measures were carried out just before (pre-tetanus) and after LTP recordings (post-tetanus).

NO Imaging of Hippocampal Slices
Parasagittal slices (400 µm), after 1 h incubation in the 95% O 2 /5% CO 2 perfused ACSF at 34 • C, were transferred into the light-shielded camera containing continuously aerated ACSF with 5 µM of the NO-sensitive dye-DAF-FM (4-Amino-5-methylamino-2 ,7 -Difluorofluorescein) diacetate. After 1 h incubation with the dye, the slices were washed in the ACSF. Further, slices were placed to experimental chamber and exposed to the drug treatment and/or tetanus. Control slices were in parallel placed in aerated ACSF without any influences. At the end of the experiment, slices were fixed in 4% formaldehyde in PBS solution for 20 min. Fixation and storage of slices occurred in light-shielded plates. Next, the slices were investigated for NO levels using a fluorescence microscope, Keyence BZ-9000 (Keyence, Osaka, Japan), equipped with mercury lamp, objective lenses (×2, ×10, ×20), excitation filter (450-490), dichroic mirror, and emission filter (520-540). Optical images were recorded and further analyzed by BZ-Analyzer (Keyence, Japan) and ImageJ (NIH, NY, USA). For semi-quantitative estimation of the NO production, a ratio of digitized dye signals was used in the region of interest (ROI) to the DAF fluorescence in the Stratum oriens, a hippocampal layer having weak NO synthesis activity.

Phosphatase Assay
Serine/threonine phosphatase activity was determined by using a nonradioactive molybdate dye-based phosphatase assay kit (Promega, Madison, WI, USA), in accordance with the instructions of the manufacturer, as reported in [99]. Following 20 min after tetanus of control or drug-treated slices (total time of the tested drug exposure to hippocampal slices in aerated ACSF is 40 min), slices were harvested and homogenized, 2 slices per 0.5 mL of buffer containing (in mM): sucrose, 250; β-mercaptoethanol, 15; EDTA, 0.1; phenylmethylsulfonyl fluoride, 0.1; TRIS-HCl, 50 (pH 7.4). Free phosphate was removed from the lysate supernatants using a Sephadex G-25 resin spin column. Phosphatase reactions were assessed in 50 µL samples at 34 • C, in buffer containing 50 mM imidazole, 0.2 mM EGTA, 0.02% β-mercaptoethanol, and 0.1 mg/mL bovine serum albumin (pH 7.2), based on the dephosphorylating rate of the synthetic 754 Da phosphopeptide RRA(pT)VA, a substrate for serine/threonine phosphatases. Specificity of the phosphatase reaction was tested by okadaic acid (100 nM) or cyclosporin A (5 µM), using PP1/PP2A and PP2B inhibitors, respectively. The reaction was stopped by addition of the dye-additive mixture (50 µL), and samples were incubated to color development for 20 min at room temperature. Absorbance was measured at 640 nm using an Infinite 200 Pro plate reader (Tecan, Salzburg, Austria). For normalization of phosphatase activity, we also measured total protein amounts in hippocampal lysates using BCA Protein Assay Kit (Thermofisher, Waltham, MA, USA), in accordance with the recommendations of the manufacturer. Protein-dependent reduction of Cu 2+ to Cu + in an alkaline medium, followed by the bicinchoninic acid sensitive detection of Cu + ions, was determined in hippocampal samples, using bovine serum albumin (Amresco, Solon, OH, USA) as a calibration standard (125-2000 µg/mL). Then, 50 µL samples were incubated in the well of microplates at 37 • C for 30 min to color development and were measured by Infinite 200 Pro plate reader (Tecan, Austria) at 562 nm.

Isolation and Purification of Mitochondria
All manipulations with animals, before the isolation of the organs, were performed in accordance with the Helsinki Declaration of 1975 (revised in 1983), national requirements for the care and use of laboratory animals, and protocol 9/2020 of 17 February 2020, approved by the Commission on Biological Safety and Bioethics at the ITEB RAS. Male Wistar rats (2-3 months) were sacrificed after anesthesia with CO 2 . Rat brain mitochondria (RBM) were isolated as described in [100], with minor modifications. The homogenization medium contained 320 mM sucrose, 10 mM Tris (pH adjusted to 7.4 with Trizma Base), 0.5 mM EGTA, and 0.5 mM EDTA. The brain was homogenized in a Potter homogenizer (800 rpm, 20 passages). The homogenate was centrifuged twice for 4 min at 2000 g, and each time the precipitate was discarded. A "crude" mitochondrial pellet containing synaptosomes, myelin, and non-synaptic RBM was obtained by sedimentation from supernatant at 12,500 g for 11 min. Then, RBM was purified by the sedimentation (17,500 g for 11 min) in a discontinuous Percoll gradient (3%, 10%, 15%, and 24% Percoll in the isolation medium). The fraction of nonsynaptic mitochondria was collected and washed once in an isolation medium, free of EGTA (11,500 g, 11 min). The concentration of isolated RBM was approximately 20 mg protein/mL. Mitochondrial protein was assayed by the Biuret method, using BSA as a standard [101]. All measurements were performed at 30 • C in the standard KCl-based medium (KCl-BM: 120 mM KCl, 20 mM sucrose, 2 mM KH 2 PO 4 , and 2 mM MgCl 2 ; 10 mM HEPES (pH 7.3), substrates: 5 mM pyruvate + 5 mM malate or 5 mM succinate + 1 µM rotenone, respectively). Other experimental details are given in the figures and figure legends.

Ca 2+ -Dependent Swelling of Mitochondria
The opening of mitochondrial permeability transition pore (mPTP) in isolated mitochondria was registered as the initiation of EGTA-and CsA-sensitive high-amplitude swelling. Mitochondrial swelling was determined by measuring a decrease in optical density at the 540 nm in mitochondrial suspension using a plate reader (Infinite 200 Tecan, Salzburg, Austria) and 96-well plates. Other details are given in the figures and figure legends.

Statistical Analysis
Results are presented as mean ± standard error (S.E.M.) of n number of slices, from at least three different animals (n = 3-8, depending on the experimental series). All statistical tests were performed using SigmaPlot 11.0 (Systat Software Inc, San Jose, CA, USA) or using SPSS Statistics software (version 21, IBM Corp., Armonk, NY, USA). Significance of changes in the fEPSP characteristics, the DAF-FM fluorescence, and the phosphatase activity was tested by ANOVA with the Bonferroni test. Phosphatase reaction values are expressed as percentage (%) from control measurements (taken as 100%) and normalized to the total protein amounts in the sample. Mitochondrial measurements were made in three repeats from different animals. All tests used were two-sided; p < 0.05 was predetermined as defining statistically significant differences and in figures denoted by * or # for multiple comparisons.

Conclusions
The combination of the electrophysiological recordings of fEPSPs together with the direct measurements of STPP activity and NO production, as well as the monitoring of the physiological parameters in the suspension of the brain mitochondria, allows for deeper understanding of the Aβ 25-35 -mediated LTP suppression in the hippocampus. Aβ [25][26][27][28][29][30][31][32][33][34][35] aggregates lead to activation of stress-induced PP1α phosphatase, shifting the kinasephosphatase balance during the LTP induction protocol, toward the enhancement of STPP. An increase in STPP activity was well-correlated with the fEPSP suppression as well as with stimulation of the nNOS-derived NO production in the CA1, CA3, and DG layers. Furthermore, the Aβ 25-35 -dependent influence involved stimulation of the Ca 2+ released from mitochondria through the mNCX. The results obtained in our experiments suggest the involvement of PP1α, nNOS, and mNCX in the realization of Aβ 25-35 -driven signaling in the hippocampus. These data provide at least three potential strategies for the compensation of Aβ 25-35 -mediated synaptic insufficiency, involving the blockade of the stress-induced GADD34/PP1α complex, and/or inhibition of mNCX, and targeting NO production via the nNOS blockade. Additionally, the management of the kinase/phosphatase balance through a reduction in STPP activity and/or a strengthening of serine/threonine kinase activity can be useful in preventing the Aβ 25-35 -associated impairment of hippocampal neural network functionality.

Institutional Review Board Statement:
The study was conducted according to the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes 1986, 86/609/EEC, and institutional requirements for the care and use of laboratory animals (the Institute of Higher Nervous Activity and Neurophysiology, Russian Academy of Sciences, Russia, and the Institute of Theoretical and Experimental Biophysics, Russian Academy of Sciences, Russia). All efforts were made to minimize the quantity of animals.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.