Blockade of Type 2A Protein Phosphatase Signaling Attenuates Complement C1q-Mediated Microglial Phagocytosis of Glutamatergic Synapses Induced by Amyloid Fibrils

Abstract

We previously reported the critical involvement of metabotropic GluR1 (mGluR1) signaling in complement C1q-dependent microglial phagocytosis of glutamatergic synapses in a rat model of Alzheimer’s disease (AD) injected with amyloid fibrils. Here, we explored the role of type 2A protein phosphatase (type 2A PPase), a key enzyme downstream of mGluR1 signaling, in the pathogenesis of AD in rats. Significant local upregulation of PP2A expression was observed in the hippocampal CA1 after bilateral microinjection of amyloid-beta (Aβ_1–40) fibrils. Amyloid fibrils induced remarkable dephosphorylation of pFMRP (fragile X mental retardation protein) and C1q upregulation in hippocampal glutamatergic synapses, which was ameliorated by microinjection of type 2A PPase inhibitor okadaic acid (OA). Microinjection of OA further attenuated the microglial phagocytosis of glutamatergic synapses, recovered the hippocampal glutamatergic transmission, and improved the performance in Morris water maze test. These findings demonstrated that dysfunction of type 2A PPase signaling contributed to complement C1q-dependent microglial phagocytosis of glutamatergic synapses and the cognitive impairments in the rat model of AD.

Introduction

Accumulating evidences demonstrated the pivotal role of complement C1q-dependent microglial phagocytosis of central glutamatergic synapse in the pathogenesis of various neurological diseases, such as Alzheimer’s disease (AD) [1–4]. Therefore, it is of substantial importance to explore the underlying mechanisms, which is crucial to further advance our understanding on the pathophysiology of AD. We recently reported that activation of metabotropic glutamate receptor 1 (mGluRl) signaling-enhanced dephosphorylation of fragile X mental retardation protein (FMRP), which facilitated the local translation of C1q mRNA in glutamatergic synapses and subsequent C1q-mediated microglial phagocytosis of these synapses. These contributed to synaptic and cognitive dysfunction in rodent models of AD [4].

The family members of type 2A group of phosphatases (including PP2A, PP4, and PP6) are closely related in sequence and can be functionally inhibited by small-molecule toxin okadaic acid (OA) [5]. Following mGluRs activation, the upregulation of protein phosphatase 2A (PP2A) activity critically mediated mGluRs-induced long-term synaptic depression and brain dysfunction in various neurological settings [6, 7]. PP2A is a typical type of Ser/Thr phosphatase and contributes to numerous physiological and pathological functions, including neural development, proliferation, cell cycle, and apoptosis [8]. Upon activation of mGluR, PP2A-induced dephosphorylation of pFMRP at Ser⁴⁹⁹, leading to rapid upregulation of dendritic Arc protein and synaptic depression in rat and mouse hippocampal neurons [9]. The dysfunction of PP2A signaling is involved in the pathogenesis of synaptic and cognitive impairment in AD models [10]. However, the dynamic activity of PP2A and its involvement in the pathogenesis of AD remains controversial [11–13].

Based on our recent reports [4, 14], we hypothesized that the upregulation of type 2A PPase (including PP2A) activity induced the dephosphorylation of pFMRP and increased complement C1q expression in hippocampal CA1, which subsequently contributed to microglial phagocytosis of glutamatergic synapses and cognitive impairment induced by amyloid fibrils.

Materials and Methods

Animals

All animal procedures were approved by the Institutional Committee of Animal Care and Use in Cleveland Clinic. Adult Sprague–Dawley rats (male, 250–300 g, Charles River) were randomly assigned to individual groups with different treatments. All experiments were performed by another party blinded to the grouping during the light cycle. The sample sizes were determined based on previous reports [15].

Microinjection in the Hippocampal CA1 Area

Bilateral hippocampal microinjection of Aβ_1–40 fibrils has been widely used for studying AD in the rodents [4, 15–18]. After anesthetized with sodium pentobarbital (45 mg/kg i.p.), the rats were restrained in a stereotaxic apparatus [19]. Aβ_1–40 fibrils [18] (10 μg/3 μl) or 3 μl of artificial cerebrospinal fluid (aCSF) were delivered into the bilateral hippocampus CA1 area (anteroposterior, − 3.5 mm; mediolateral, ± 2.0 mm; dorsoventral, − 3.0 mm) [20] using a 10-μl Hamilton syringe with a 27G injector at a rate of 0.5 μl /min.

For in vivo treatment by microinjection, a cannula (26-gauge) was implanted into the brain, 1 mm above the hippocampal CA1 area (anteroposterior. − 3.5 mm; mediolateral, ± 2.0 mm; dorsoventral, − 2.0 mm) [15], which was cemented in place to the skull and securely capped. The rat was allowed to recover for at least 5 days prior to subsequent experiment. Type 2A PPase inhibitor OA (25 ng/side), or artificial CSF, was delivered daily into the bilateral hippocampal CA1 through a 33-gauge injector at a rate of 0.5 μl/min at 14 days after microinjection of aCSF or amyloid fibrils [15, 21]. Behavioral, cellular, and molecular analysis were performed 1 day after the final injection of specific agents or vehicle. The injection sites for the hippocampal CA1 were histologically verified afterward [21].

Morris Water Maze Test

The Morris water maze test, as described previously [18, 21], was performed in a circular tank (diameter 1.8 m) filled with opaque water. A platform (15 cm in diameter) was submerged below the water’s surface in the center of the target quadrant. The swimming trace of an individual rat was digitally recorded by a video camera and analyzed by EthoVision XT software (Noldus Information Technology). Each rat underwent four trials per day for 5 days with a 10-min inter-trial interval. For each training session, a rat was placed into the maze from four different points of the tank and was allowed to search for the platform in target quadrant. In case the animal cannot locate the platform within 120 s, it was then gently guided to it. The rat was allowed to remain on the platform for 20 s. The latency for each trial was recorded. During the probe trial, the platform was removed from the tank, and rats were allowed to swim in the maze for 60 s.

Preparation of Hippocampal Synaptosome

Hippocampal synaptosome was prepared as previously described [15]. Hippocampal CA1 tissues obtained from the rats were gently homogenized in sucrose buffer (0 °C, 0.32 M) with proteinase and phosphatase inhibitors at pH 7.4 and then centrifuged for 10 min at 1000 g (4 °C). The supernatant was collected. After centrifugation for 20 min at 10,000 g (4 °C), the synaptosomal pellet was then obtained for further molecular analysis.

Protein Extraction and Immunoblotting

Protein extraction and immunoblotting were generally based on the protocols in previous reports with minor modification [22]. The hippocampal CA1 tissues or synaptosomal preparations were lysed in lysis buffer (0 °C) containing 50 mM Tris–Cl, 150 mM NaCl, 0.02 mM NaN₂, 100 μg/ml phenylmethyl sulfonyl fluoride, 1 μg/ml aprotinin, 1% Triton X-100, and proteinase and phosphatase inhibitor cocktail. The proteins were extracted and subjected to 7.5% SDS–PAGE followed by immunoblotting. The blots were incubated overnight at 4 °C with the following primary antibodies: monoclonal anti-PP2A antibody (1:1000; Cell Signaling Technology, #2259), anti-FMRP antibody (1:1000; Cell Signaling, 7104), monoclonal anti-pFMRP antibody (1:1000; Thermo Scientific, PA5-35,389), and monoclonal anti-β-actin antibody (1:2000; Santa Cruz Biotechnology, sc-81178). The blots were extensively washed and incubated with horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit IgG antibody (1:10,000; Jackson ImmunoResearch Laboratories Inc., West Grove, PA). The immunoreactivity was detected using enhanced chemiluminescence (ECL Advance Kit; Amersham Biosciences). The intensity of the bands was digitally captured and quantitatively analyzed with ImageJ software.

RNA Immunoprecipitation (RNA-IP) and Quantitative RT-PCR for mRNA Analysis

RNA-IP was performed in the hippocampal preparation as previously described [23]. Briefly, the hippocampal CA1 preparation was subjected to the lysis buffer. Protein/mRNA complex was pulled down by the monoclonal anti-pFMRP (1:100, Thermo Scientific, PA5-35,389) antibody and was recovered with protein A/G beads. The precipitated and input mRNA was extracted using a single-step RNA isolation protocol and quantified as previously described [15]. GAPDH was used internal control [24, 25]. Reverse transcription and real-time RT-PCR were performed in triplicate with C1q primers (5′-ACA AGG TCC TCA CCA ACC AG-3′, 5′-CGTT GCAATT GA AGCA CA GT-3′) and GAPDH primers (5′-AGA CAG CCG CAT CTT CTT GT-3′, 5′-CTT GCC GTG GGT AGA GTC AT-3′). IP/input ratio was calculated. To analyze C1q mRNA in synaptosomal preparation, fold change was calculated using the ΔΔCT method. The entire protocol was repeated in triplicate.

Immunostaining and 3-D Confocal Imaging

Immunostaining was performed on the serial Sects. (30 μm, 15–20/rat, n = 5 rats per group) containing hippocampal CA1 area as previously described [21]. Mouse monoclonal antibody against the microglial marker Iba1 (1:500, Abcam, ab5076), glutamatergic synapse marker PSD95 (1:200, Abcam, ab104898), lysosomal marker CD68 (1:200, Abcam, ab955), and monoclonal anti-C1q antibody (1:1000; Abcam, ab71940) were used. FITC- and Cy3-conjugated secondary antibody (1:500, Jackson ImmunoResearch) or Alex Fluor 633-conjugated secondary antibody (1:500, Invitrogen) was also used. Negative controls involving omission of primary antibodies and/or inclusion of an irrelevant IgG isotype were applied in parallel, which yielded no immunosignal. Leica TCS-SP8-AOBS-inverted confocal microscope (Leica Microsystems, GmbH, Wetzlar, Germany) was used to examine all sections. Immunofluorescence was rendered and analyzed with Image-Pro Plus (Media Cybernetics, Inc., Rockville, MD) and Velocity (PerkinElmer, Waltham, MA). The internalization of synaptic marker PSD95 in microglial lysosomes (as indicated by marker CD68) was defined by rotating the 3D micrographs to assure the colocalization of synaptic markers with CD68 immunosignal within the microglia (Iba1 immunosignal) in hippocampal CA1, as presented in our previous report [4, 14, 26]. The representative 3D images were demonstrated with resolution 1024 × 1024 dpi and z-step size 0.3 μm. Usually, a 200 × 200 × 20 μm neuropil (containing about 9–12 microglia) in each section and 4 sections in each animal were randomly sampled and analyzed in individual group. The volume of PSD95 within lysosome (CD68 +) and in a neuropil (200 μm × 200 μm × 20 μm) was measured by Velocity (PerkinElmer, Waltham, MA), and the ratio was calculated to measure microglial phagocytosis of synapses in hippocampal sections in all groups. Similarly, the colocalization of C1q with PSD95 was also defined by rotating the 3D micrographs as previously reported [4]. The ratio between the volume of PSD95 colocalized with C1q immunosignal and that of all PSD95 was calculated to measure the C1q expression in glutamatergic synapses in all groups.

Hippocampal Slice Preparation and Whole-Cell Recordings

Whole-cell recording on hippocampal slice was performed at room temperature as previously described [15]. Coronal slices (300 μm thick) containing the hippocampus were prepared. Whole-cell recordings from the CA1 neurons were performed using an Axopatch 200B amplifier (Molecular Devices) with 2–4 MΩ glass electrodes containing the internal solution (mM): K-gluconate, 125; NaCl, 5; M gCl₂, 1; EGTA, 0.5; Mg-ATP, 2; Na₃GTP, 0.1; HEPES, 10; pH, 7.3; and 290–300 mOsmol. A seal resistance of ≥ 2 GΩ and an access resistance of 15–20 MΩ were considered acceptable. The series resistance was optimally compensated by ≥ 70% and constantly monitored throughout the experiments. The miniature EPSCs (mEPSCs) were filtered at 2 kHz, digitized at 10 kHz, and acquired and analyzed using Axograph X software. Miniature EPSCs were recorded in the presence of tetrodotoxin (TTX, 1 μM) and bicuculline (10 μM) at a holding potential of − 70 mV.

Compounds

Aβ peptide consisting of residues 1–40 of human wild-type sequence (Aβ_1–40) was purchased from Bachem (Torrance, CA). Bicuculline, TTX, and all other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) or Tocris (Ellisville, MO).

Statistical Analyses

Two-way ANOVA with post hoc analysis was used for behavioral analysis. Histological and immunoblotting data were analyzed using one-way ANOVA test, Student’s t test, or Kruskal–Wallis statistical test with BMDP statistical software (Statistical Solutions, Saugus, MA). All data were presented as mean ± SEM. For all tests, a two-tailed P < 0.05 was considered statistically significant.

Results

Amyloid Fibrils Upregulate the Expression of PP2A in Hippocampal CA1

Our previous study demonstrated the critical role of mGluRs in the complement C1q-mediated microglial phagocytosis of glutamatergic synapses in rodent models of AD [4]. Considering the critical role of PP2A, a key downstream player of mGluR signaling, in regulating hippocampal synaptic plasticity and cognitive function [27], we first studied the expression of PP2A in hippocampal CA1 in the rats with microinjection of amyloid fibrils. As shown in Fig. 1, compared to control group, the expression of PP2A was significantly increased in hippocampal CA1 in the modeled rats, suggesting a potential involvement of PP2A signaling in amyloid fibril-induced synaptic and cognitive deficiency.

Fig. 1.

PP2A expression is significantly upregulated in hippocampal CA1 in the rats injected with amyloid fibrils. A The representative immunoblotting bands showing the expression of PP2A in the control (left 3 lanes) and amyloid-injected rats (right 3 lanes). B Graph showing the statistical difference of PP2A in these 2 groups. N = 8 per group; **P < 0.01 with student’s t test. Data represent as mean ± SEM

Blockade of Type 2A PPase Activity by OA Suppresses Synaptic C1q Upregulation Induced by Amyloid Fibrils

We next tested the effect of inhibition of type 2A PPase activity on complement C1q production. Type 2A PPase activity was suppressed by local administration of OA (25 ng per side for 7 days) into the hippocampal CA1, in the rats with microinjection of amyloid fibrils or aCSF. Consistent with previous report [4], significant dephosphorylation of FMRP, but no change in total FMRP expression, was observed in the hippocampal synaptosomal preparation in the amyloid-injected rats. Administration of OA significantly attenuated dephosphorylation of FMRP (Fig. 2a, b). While decreased binding between pFMRP and C1q mRNA (Fig. 2c) was observed in in the hippocampal synaptosomal preparation in the amyloid-injected rats, it is also recovered by OA (Fig. 2c). We further found that microinjection of OA decreased the upregulation of C1q mRNA in the synaptosomal preparation of hippocampal CA1 in the modeled rats (Fig. 2d). Note that OA treatment did not significantly change C1q mRNA and its interaction with pFMRP in the control rats. Further immunostaining study found that the C1q immunoreactivity, primarily colocalized with synaptic marker PSD95, was significantly increased in the modeled rats, which was largely attenuated by OA (Fig. 2e, f). OA treatment did not change C1q expression in control rats (Fig. 2). Together, these results suggested that blockade of type 2A PPase activity by OA attenuated the upregulation of C1q at both RNA and protein levels in hippocampal synapses induced by amyloid fibrils.

Fig. 2.

Inhibition of type 2A PPase signaling attenuates the C1q upregulation induced by amyloid fibrils. Microinjection of okadaic acid (OA) (25 ng for 7 days) significantly recovered the dephosphorylation of FMRP in the hippocampal CA1 in the amyloid-injected rats (a, b n = 7 rats in each group, F_3,24 = 7.2, P = 0.0013, one-way ANOVA). OA significantly increased the amount of C1q mRNA pulled down by pFMRP antibody in the hippocampal CA1 in the rats injected with amyloid fibrils (c n = 6 rats in each group, F_3,20 = 3.97, P = 0.02, one-way ANOVA). OA significantly decreased C1q mRNA in the hippocampal synaptosomal preparation in the modeled rats (d n = 6 rats in each group, F_3,20 = 4.3, P = 0.02, one-way ANOVA). OA significantly decreased C1q immunosignals colocalized with the PSD95 in the hippocampal CA1 in the modeled rats (e, f n = 5 rats in each group, F_3,16 = 66.1, P < 0.0001, scale bar = 10 μm, one-way ANOVA). Each dot represents the mean value of 4 brain slices of one rat. Data represent as mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001

Blockade of Type 2A PPase Activity by OA Attenuates Microglial Phagocytosis of Glutamatergic Synapses and Cognitive Impairment Induced by Amyloid Fibrils

Next, we investigated the role of PP2A signaling in glutamatergic synapse phagocytosis and cognitive dysfunction. Consistent with our previous report [4], we found an increased internalization of synaptic marker PSD95 in microglial (CD68-positive) lysosomes, indicating phagocytosis of glutamatergic synapses in the hippocampal CA1 areas injected with amyloid fibrils (Fig. 3A, B). Microinjection of OA significantly attenuated the colocalization of PSD immunoreactivity in microglial (Iba1-positive) lysosomes (CD68-positive) (Fig. 3A, B), indicating a reduction of microglial phagocytosis of glutamatergic synapses in hippocampal CA1 in these modeled rats.

Fig. 3.

Inhibition of type 2A PPase attenuates the microglial phagocytosis of glutamatergic synapses induced by amyloid fibrils. Microinjection of type 2A PPase inhibitor, okadaic acid (OA), significantly decreased the colocalization of PSD95 with lysosome marker CD68 in microglia in the rats injected with amyloid fibrils (A, B n = 5 rats in each group, F_3,16 = 48.1, P < 0.0001 with one-way ANOVA, scale bar = 10 μm). Right micrographs were presented showing the same microglia in which only the lysosomes (blue) and PSD95 (red) were visualized. Data represent mean ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001

Furthermore, microinjection of amyloid fibrils induced significant decrease in the amplitude and increase in inter-events interval of excitatory glutamatergic transmission in the hippocampal CA1 neurons. These changes were remarkably recovered by OA administration (Fig. 4A, B). Together, these results demonstrated that the suppression of PP2A signaling by OA attenuated microglial phagocytosis of glutamatergic synapses and mitigated the synaptic and cognitive dysfunction in this rodent model of AD.

Fig. 4.

Inhibition of type 2A PPase recovers the glutamatergic transmission in the rats injected with amyloid fibrils. Administration of OA significantly recovered the amplitude (A, B n = 39 neurons in each group, Kruskal–Wallis statistic KW = 36.4, P < 0.0001) and inter-event interval (A, C n = 39 neurons in each group, Kruskal–Wallis statistic KW = 32.2, P < 0.0001) of mEPSCs in the hippocampal CA1 neurons (n = 45) in the rats injected with amyloid fibrils. Data represent mean ± SEM, ***P < 0.001

Consistent with previous reports [4, 15, 21, 28], microinjection of amyloid fibrils significantly extended the escape latency in the Morris water maze test and decreased the time spent in the target quadrant during the probe trails, which were significantly recovered by OA treatment (Fig. 5A, B). Administration of OA did not remarkably change the behavioral performance in the control rats. These results indicated that inhibition of type 2A PPase by OA attenuated cognitive dysfunction in the rats injected with amyloid fibrils.

Fig. 5.

Inhibition of type 2A PPase improves the performance in Morris water maze test in the rats injected with amyloid fibrils. OA significantly decreased the escape latency (A n = 10 rats in each group, effect of group [F_3,36 = 10.9, P < 0.0001], effect of time [F_4,36 = 233.5, P < 0.0001], interaction between group and time [P = 0.26], two-way ANOVA with repeated measurements) and increased the time in the target quadrant (B n = 10 rats in each group, F_1,18 = 14.4, P = 0.001, one-way ANOVA) in the modeled rats. B Representative path tracings in each quadrant during the probe trial on day 6 (T, target quadrant; R, right quadrant; O, opposite quadrant; L, left quadrant). *P < 0.05, **P < 0.01

Discussion

Our recent reports demonstrated that mGluR1 signaling is critically involved in the pathogenic process of complement C1q-mediated microglial phagocytosis of glutamatergic synapses and the subsequent synaptic and cognitive deficiency in rodent models of AD [4]. In the preset study, we found an increased local expression of PP2A, a key downstream of mGluR1 signaling, in hippocampal CA1 in rats injected with amyloid fibrils. We further showed that suppression of local type 2A PPase activity by microinjection of OA into hippocampal CA1 remarkably ameliorated the synaptic and cognitive deficiency in the rodent model of AD.

Increasing evidences suggested that disturbance of PP2A activity played a prominent role in the pathogenesis of AD [10]. PP2A may contribute to tau phosphorylation in rodent models of AD through direct dephosphorylation or indirect inactivation of upstream inhibitor of GSK3-β [12, 29]. For instance, activation of PP2A activity by synthetic tricyclic sulfonamide decreased tau phosphorylation and Aβ_40/42 overproduction and prevented neuronal loss and cognitive impairment in the rodent AD model induced by DL-homocysteine [12]. However, studies reported either increased PP2A expression in postsynaptic supernatants in the temporal cortex [11] or decreased PP2A activity in gray and white matters in AD patients [13]. While a recent study reported the reduced PP2A activity in AD human brains, APP/PS1 transgenic mice, and primary neurons with Aβ treatment [30], the present study found an upregulation of local PP2A expression induced by amyloid fibril infusion in hippocampal CA1 area.

PP2A signaling, as one of the primary cascades upon mGluR1 activation, critically modulated the dynamic dephosphorylation and function of pFMRP, a key regulator of the transportation and local translation of mRNAs in glutamatergic synapses [31–33]. Previous evidence have documented that phosphorylation of FMRP (Ser⁴⁹⁹) generally leads to the stalled polyribosomes, while dephosphorylated FMRP, releasing from the polyribosome, can facilitate the transportation and translation of local mRNAs [34]. Disturbance of FMRP phosphorylation may induce the dysregulated local mRNA translation at the synapse and impair synaptic plasticity [32, 35]. It was revealed that dysfunction of FMRP also contributed to synaptic and cognitive deficiency in rodent models of AD [36–38]. Consistent with our previous reports [4], we found an increased dephosphorylation of pFMRP in hippocampal CA1 in the rats injected with amyloid fibrils. We further found that administration of type 2A PPase inhibitor OA significantly attenuated the dephosphorylation of pFMRP, which potentially contributed to OA-induced suppression of C1q production in rodent model of AD. Notably, an obvious limitation exists in the present study due to the nonselective inhibition of OA on the activity of type 2A PPase family members (PP2A, PP4, and PP6) [5]. Further studies to exclude the potential involvement of PP4 and PP6 are essential to clarify the signaling pathway in this pathological process. While previous reports existed to show the reduction of PP2A expression and/or function by OA in different settings [39], we also realized that further study was needed to examine the effect of OA on PP2A expression and function in various models of AD, together with the related cellular and behavioral evidence, which would further confirm the involvement of PP2A signaling in the pathogenesis of AD.

Complement C1q-mediated microglial phagocytosis of glutamatergic synapses exhibits an essential role in the pathogenic process of various neurodegenerative diseases, including AD [2, 4, 14, 40, 41]. Complement C1q is normally produced in microglia and/or neurons in brain, and aggregation of amyloid fibrils remarkably enhanced C1q production in hippocampal glutamatergic synapses in rodent models of AD [4]. We previously demonstrated that activation of mGluR1 signaling, via regulating pFMRP dephosphorylation, induced remarkable upregulation of C1q production in hippocampal synapses in various rodent models of AD [4]. Here, we further found that the inhibition of type 2A PPase activity by OA, via increasing FMRP dephosphorylation, significantly alleviated the C1q upregulation and microglial phagocytosis of glutamatergic synapse in the modeled rats. These results suggested the pivotal role of type 2A PPase in mGluR1-mediated C1q-dependent microglial phagocytosis of glutamatergic synapses in rodent models of AD [1, 2, 4, 14].

On the basis of cellular and behavioral findings in the present study, we also recognized some substantial pitfalls to further validate and extend the functional significance of PP2A signaling in the pathogenesis of AD. For example, additional approaches including immunoblotting will be helpful to further confirm C1q expression in various settings. A series of studies with artificial regulation of PP2A activity (loss or gain of PP2A function confirmed by various methods) will be also helpful to further clarify the involvement of PP2A signaling in the development of synaptic and cognitive deficiency in AD.

In conclusion, we demonstrated that inoculation of amyloid fibrils significantly upregulated PP2A expression in hippocampal CA1 and inhibition of type 2A PPase suppressed C1q production in hippocampal glutamatergic synapse in the rats injected with amyloid fibrils. We further found that inhibition of type 2A PPase signaling attenuated microglial phagocytosis of glutamatergic synapses and recovered the synaptic and cognitive deficiency in the rat model of AD.

Data Availability

The original dataset for analysis in the present study is available upon reasonable request from the corresponding author.