PSD-93 Attenuates Amyloid-β-Mediated Cognitive Dysfunction by Promoting the Catabolism of Amyloid-β

Abstract

Amyloid-β (Aβ) is a key neuropathological hallmark of Alzheimer’s disease (AD). Postsynaptic density protein 93 (PSD-93) is a key scaffolding protein enriched at postsynaptic sites. The aim of the present study was to examine whether PSD-93 overexpression could alleviate Aβ-induced cognitive dysfunction in APPswe/PS1dE9 (APP/PS1) mice by reducing Aβ levels in the brain. The level of PSD-93 was significantly decreased in the hippocampus of 6-month-old APP/PS1 mice compared with that in wild-type mice. Following lentivirus-mediated PSD-93 overexpression, cognitive function, synaptic function, and amyloid burden were investigated. The open field test, Morris water maze test, and fear condition test revealed that PSD-93 overexpression ameliorated spatial memory deficits in APP/PS1 mice. The facilitation of long-term potentiation induction was observed in APP/PS1 mice after PSD-93 overexpression. The expression of somatostatin receptor 4 (SSTR4) and neprilysin was increased, while the amyloid plaque load and Aβ levels were decreased in the brains of APP/PS1 mice. Moreover, PSD-93 interacted with SSTR4 and affected the level of SSTR4 on cell membrane, which was associated with the ubiquitination. Together, these findings suggest that PSD-93 attenuates spatial memory deficits and decreases amyloid levels in APP/PS1 mice, which might be associated with Aβ catabolism, and overexpression of PSD-93 might be a potential therapy for AD.

Keywords

Alzheimer’s disease amyloid-β postsynaptic density protein 93 somatostatin receptor 4

INTRODUCTION

Alzheimer’s disease (AD), as the major cause of dementia, is an irreversible neurodegenerative disorder with progressive cognitive dysfunction, memory impairment, and behavioral damage. Amyloid-β (Aβ) is the major constituent of the tissue deposits detected in parenchymal plaques and leads to AD pathology during middle age [1]. Aβ is produced from amyloid-β protein precursor (AβPP) through sequential β- and γ-secretase cleavage, and could be cleared through degradation by proteases, cerebrospinal fluid (CSF) absorption, perivascular, and blood-brain barrier systems, etc. [2]. It has been suggested that the altered balance between production and clearance is responsible for the accumulation of brain Aβ [3]. Impaired clearance of Aβ from the brain of late onset AD plays a critical role in the formation of amyloid plaques and the pathogenesis of the disease [4]. Thus, modulating the transport or metabolic balance of Aβ could regulate the brain levels of Aβ, which might affect the progression of AD. Regrettably, alterations in Aβ transport mechanisms through the blood-brain barrier (BBB) remain difficult due to the poor capability of most drugs to cross the BBB [5]. Therefore, increasing Aβ catabolism within the brain might provide an unprecedented choice for therapeutic development. Recent studies have also suggested that stimulating the activity of amyloid-degrading enzymes could enhance the clearance of Aβ and reduce its deposition in the brain [6]. The proteolytic enzyme neprilysin (NEP) is now considered as the most important Aβ-degrading peptidase, and the enhancement of NEP caused significant reductions in Aβ levels and increased synaptic density [7].

Excitatory synapses of the brain are characterized by dense thickenings of protein referred to as postsynaptic densities (PSDs), which contain glutamate receptors, cell adhesion proteins, scaffold proteins, and signaling molecules [8, 9]. PSD-93/chapsyn-110, PSD-95/SAP-90, SAP-97, and SAP-102 are the primary members of a family of scaffold proteins called membrane-associated guanylate kinases (MAGUKs) which are abundantly expressed in the PSD. Each of these proteins contains three PDZ domains, an SH3 domain, and a catalytically inactive guanylate kinase domain, which together mediate protein-protein interactions that are important for channel clustering and the recruitment of signaling complexes for synaptic function. Recent evidence clearly supports Aβ as a key factor in synaptic impairment, which is closely correlated with indices of cognitive decline [10]. The synaptic damage is detected at the early stages of AD, which is shown as the loss of pre-synaptic proteins and post-synaptic markers, such as PSD-95 and Shank1 [11, 12]. PSD-93 is similar to PSD-95 in terms of amino-acid sequence, domain organization, developmental expression and function; however, it is unclear whether PSD-93 is able to attenuate Aβ-induced synapse injury in AD.

There is compelling evidence of a close relationship between somatostatin (SST) and Aβ fibril accumulation in AD [13]. In addition, the activity of NEP is upregulated by SST [14]. In the hippocampus of APP/PS1 mice, the expression of SST was reduced, and a linear correlation was observed between Aβ content and SST loss [15]. Somatostatin receptor 4 (SSTR4), a member of the SSTRs family, is expressed in cortical and hippocampal tissues, and binds to the PDZ domains of MAGUKs in rat neurons [16]. Thus, in the present study, we hypothesized that PSD-93 could attenuate memory impairment and decrease Aβ-induced synaptic function damage in AD mice, which might be associated with the interaction of SSTR4.

MATERIALS AND METHODS

Animals and treatment

Male APPswe/PS1dE9 (APP/PS1) double-transgenic mice expressing a human/mouse APP construct with the Swedish double mutation and the exon-9-deleted variant of PSEN1 were obtained from the Model Animal Research Center of Nanjing University, and male age-matched wild-type (WT) littermates were used as controls. At the age of 6 months, APP/PS1 mice showed amyloid plaques deposition and cognitive impairment [17, 18]. The lentivirus which mediated overexpression (Lv-PSD-93) or inhibition of PSD-93 (Lv-shPSD-93) was provided by Shanghai Genechem. As previously described [19, 20], the PSD-93 overexpressing lentivirus (lenti-PSD-93, 5×10⁸ TU/ml) or control lentivirus (lenti-con, 5×10⁸ TU/ml) was slowly injected into the bilateral hippocampus of 6-month-old APP/PS1 mice using a stereotaxic apparatus. Behavior tests were conducted 4 weeks after the lentivirus injection, and then the mice were sacrificed for the other experiments. All animal experiments were approved through the Animal Care Committee of Nanjing University (reference number: 20150105), and all efforts were made to reduce animal suffering.

Cognitive and behavioral tests

To exclude the influence of locomotor activity and anxiety on cognitive performance, an open field test (OFT) was conducted. Briefly, every mouse was placed in the center of an open chamber (40×40×50 cm) with white walls. A camera was fixed above the chamber to track the exploratory behavior of the mice for 10 min (ANY-maze, Stoelting, USA); subsequently, the video was analyzed using ANY-maze software (Stoelting, USA). Prior to placing the next mice, the device was cleaned with 75% ethanol.

As previously described [19], Morris water maze (MWM) test was performed to evaluate the spatial memory of the mice after the OFT. Briefly, during the acquisition trial (days 1–5), the mice were trained to find a submerged platform within 60 s and allowed to stay on the platform for an additional 5 s. However, when the mice could not find the platform within 60 s, they were allowed to remain on the platform for 30 s to become familiar with their surroundings, and 60 s was the latency. For the probe trial (day 6), the platform was removed, and the mice were allowed to swim freely for 60 s, and the number of platform crossing, time in the target quadrant, and latency to the target quadrant were recorded. The data for all mice were analyzed using the ANY-maze system (Stoelting, USA).

To further evaluate hippocampus-dependent and hippocampus-independent memory function, a fear conditioning (FC) test was conducted, followed by the MWM. The mouse was placed in a conditioning chamber (XR-XC404, Shanghai Softmaze Information Technology Co. Ltd.) for 3 min and subsequently one tone-foot-shock pairing (tone, 30 s, 65 dB, 1 kHz; foot-shock, 2 s, 0.75 mA) was conducted. To examine postshock freezing, every mouse was allowed to remain in the chamber for another 30 s after shock. The context-dependent test was measured at 24 h after training, which allowed the mice to explore in the same chamber for 5 min. The mouse was placed in a new chamber with a different shape, color, and smell for 3 min after 2 h. The training tone was delivered to examine tone fear conditioning. The chamber was cleaned with 75% ethanol to avoid the olfactory cues observed after each test, and the freezing time was scored using analysis software.

Electrophysiological recordings

Electrophysiological recordings of hippocampal slices were performed as previously described [18 , 22]. One month later, mice exhibiting the lentivirus-medicated overexpression of PSD-93 and their littermates were deeply anesthetized using sevoflurane. The brains were cut into transverse slices (350 μm) using a vibrating microtome (Leica, Germany) in an ice bath of oxygenated cutting solution containing 75 mM sucrose, 87 mM NaCl, 2.5 mM KCl, 7.0 mM MgCl₂, 21.4 mM NaHCO₃, 0.5 mM CaCl₂, 1.25 mM NaH₂PO₄, and 20 mM D-glucose. After incubation with oxygenated artificial CSF (ACSF) (119 mM NaCl, 2.5 mM CaCl₂, 2.5 mM KCl, 1.3 mM MgSO₄, 26.2 mM NaHCO₃, 1.0 mM NaH₂PO4, and 11 mM D-glucose) for 60 min at 32°C, and the hippocampal slices were continuously recovered at room temperature for another 60 min. The cutting solution and ACSF were continually perfused with a gas mixture of 95% O₂/5% CO₂. Bipolar stimulating electrodes were placed in the stratum radiatum to elicit the action potentials of the Schaffer collateral/commissural pathway. Excitatory postsynaptic potential (EPSP) was recorded at a resistance of 0.5–3 MΩ using an ACSF-filled glass microelectrode placed in the stratum radiatum region of CA1. The data were gathered using a MultiClamp 700B amplifier (Axon Instruments, USA) filtered at 2 kHz and digitized at 10 kHz. The amplitude of peak EPSP was at least 2 mV, and 40% of the maximal response was set as the stimulus intensity. After recording baseline responses with a low frequency (2 per min) for 20 min, high-frequency stimulations (HFS, 100 Hz, 1 s) were delivered to evoke CA1 synapses. Immediately, long-term potentiation (LTP) was recorded at the same intensity as pre-HFS for 45 min. The slope of EPSP was analyzed using the pCLAMP system (Axon Instrument, CA).

Aβ extraction and measurements

Soluble and insoluble forms of Aβ were extracted as previously reported with some modifications [23]. The frozen brain was homogenized with 15 volumes (w/v) of TBS buffer, containing phosphatase and protease inhibitor cocktails (Sigma, USA), followed by centrifugation at 100,000 g for 1 h at 4°C. The soluble fraction was collected as the TBS-soluble fraction, and the pellet was resuspended with 15 volumes of 1% Triton X-100/TBS (TBSX). After incubation for 30 min on ice, the tube was centrifuged at 100,000 g for 1 h at 4°C. The supernatant was removed as the TBSX-soluble fraction, and 70% formic acid (FA) was added to the pellet, followed by centrifugation at 100,000 g at 4°C for 1 h. The FA fraction was neutralized with 20 volumes 1 M Tris-base (pH 11), and subsequently aliquoted and stored at –80°C. The total protein level of TBS-, TBSX-, and FA- was determined using the BCA protein assay kit (Bioworld, USA) and the Bradford protein assay kit (Bioworld, USA). The amount of Aβ₄₀ and Aβ₄₂ was quantified using Quantikine ELISA Human Amyloid β aa1-40/aa1-42 Immunoassay kits (R&D System, USA) according to the manufacturer’s instructions.

Primary cortical neuron culture

Primary cortical neuron culture was prepared from E16 C57/BL6J mouse embryos as previously described [24]. The cells were maintained in Neurobasal/B27 for 8 to 10 days for subsequent use in this experiment. The purity of the neuron was more than 90% as determined by microtubule-associated protein-2 (MAP-2) staining. Lenti-shPSD-93 or Lv-PSD-93 was added to the medium (MOI = 5) at day-in-vitro (DIV) 3-4 and treated for 72 h.

At DIV 9, the neurons were washed twice and subsequently fixed with 4% paraformaldehyde (PFA). After blocking with 2% BSA for 1.5 h, the neurons were incubated with primary antibodies against PSD-93 (1:500, Abcam, USA) and SSTR4 (1:500, Thermo, USA) at 4°C overnight, and then incubated with the secondary antibodies for 2 h at room temperature, and DAPI reagent (1:500, Bioworld, USA) was used for nuclear staining. The images were captured using a microscope (Olympus, Japan).

Real-time PCR

Total RNA was isolated from neurons using Trizol (Invitrogen, USA) as previously described [25] and subsequently reversed transcribed to cDNA (Takara, Dalian, China). Quantitative PCR was performed using the StepOne system (Applied Biosystems, USA) and a SYBR green kit (Takara, Dalian, China) according to the manufacturer’s instructions. The value of every sample was normalized to GAPDH. The following primers (Invitrogen, USA) were used: SSTR4: F:

AACGGAGGCGCTCAGAGAAGAAGA, R: AGGCGAGGTGAGGGAGGGTAAAAT;

GCCAAGGCTGTGGGCAAGGT, R:TCTCCAGGCGGCACGTCAGA.

Immunofluorescence assay

The levels of PSD-93 and Aβ were further confirmed using immunofluorescence assay (IFA). Following anesthesia, the mice were perfused with pre-chilled 0.9% saline and 4% PFA. The brain was immediately removed, and cryosectioned at 18 μm thickness. Similarly, the neurons were treated with 4% PFA after washing 3 times. The brain sections were blocked with 2% BSA for 1.5 h, followed by incubating with primary antibodies against PSD-93 (1:500, Invitrogen, USA), Aβ (1:500, Biolegend, USA), SSTR4 (1:500, Thermo Scientific, USA), and PSD-93 (1:500, Abcam, USA) at 4°C overnight. The secondary antibodies were applied for 2 h at room temperature, followed by nuclear staining using DAPI reagent (1:500, Bioworld, USA). The images were captured using a fluorescence microscope (Olympus, Japan). All images were quantitatively analyzed using Image J software, and the analyses were performed in a randomized and blinded manner.

Membrane protein extraction

Membrane protein extraction was performed using a Mem-PER^TM Plus Membrane Protein Extraction Kit (Thermo Scientific, USA). The cells were scraped and washed, and subsequently Permeabilization Buffer was added to the cell pellet. The homogeneous cell suspension was incubated 10 min at 4°C, followed by centrifugation for 15 min at 16,000 g at 4°C. The supernatant, containing cytosolic proteins, was removed, and the pellet was resuspended with Solubilization Buffer. After incubation at 4°C for 30 min, the tubes were centrifuged at 16,000 g for 15 min at 4°C. The membrane and membrane-associated proteins were transferred and quantified using a BCA protein assay kit (Bioworld, USA).

Immunoprecipitation and western blotting

Immunoprecipitation (IP) was performed as previously described to analyze the interaction between PSD-93 and SSTR4 and characterize the ubiquitination of SSTR4 [26]. Briefly, high-density neurons were incubated with 20 μM MG132 (carbobenzoxy-L-leucyl-L-leucylL-leucine, Selleck Chemicals, USA) for 12 h, and proteins were extracted using RIPA Lysis Buffer (Bioworld, USA). The protein content was determined using a BCA protein assay kit (Bioworld, USA). The lysates were incubated with anti-SSTR4 (1:500, Thermo, USA) or anti-PSD-93 (1:500, Invitrogen, USA) antibody at 4°C overnight. The immunoprecipitates were incubated with protein G agarose beads (Millipore, USA) for 2 h at 4°C and subsequently washed and boiled in loading buffer.

For western blotting, equal quantities of protein extracts were subjected to SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, USA). After blocking in 5% non-fat milk for 1 h at room temperature, the membranes were incubated overnight at 4°C with the following primary antibodies: anti-PSD-95 (1:1000, Cell Signaling, USA), anti-SAP102 (1:1000, Cell Signaling, USA), anti-SAP97 (1:500, Abcam, USA), anti-SST (1:1000, Bioworld, USA), anti-SSTR1 (1:1000, Bioworld, USA), anti-SSTR2 (1:500, Abcam, USA), anti-SSTR4 (1:500, Thermo, USA), anti-APP (1:500, Sigma, USA), anti-NEP (1:500, Millipore, USA), anti-IDE (1:1000, Abcam, USA), anti-Aβ (1:1000, Biolegend, USA), anti-LRP1 (1:1000, Abcam, USA), anti-ubiquitin (1:500, Abcam, USA), and anti-GAPDH (1:5000, Bioworld, USA). After washing with TBST, the membranes were incubated with HRP-conjugated secondary antibodies for 2 h at room temperature. The signals were visualized using an ECL kit (Bioworld, USA), and the band intensities were detected using Image J software.

Statistical analysis

All data were expressed as the means±standard error of mean (SEM) and analyzed using SPSS 16.0 statistical analytical software (SPSS, USA). Differences in the escape latency of MWM were analyzed using two-way analysis of variance (ANOVA) with repeated measures, followed by Bonferroni’s multiple comparison test, with day and treatment as the sources of variation. Student’s t test was performed to analyze data between groups. A value of p < 0.05 was considered statistically significant.

RESULTS

PSD-93 is decreased in the hippocampus of 6-month-old APP/PS1 mice

Prior to measuring the PSD-93 level in 6-month-old APP/PS1 mice, the OFT, MWM, and FC behavior tests were performed to determine the cognitive function. In the OFT test, there was no significant difference between the two groups in the distance, mean speed, and time spent in the corner and center zone (Supplementary Figure 1), which indicated that locomotor activity and anxiety had no effect on cognitive performance. Next, the MWM test was used to determine the spatial memory, which was compromised in AD patients and animal models [27, 28], of 6-month-old APP/PS1 mice. During the training phase, the mean escape latency was significantly increased in APP/PS1 mice (Supplementary Figure 2A, groups: F (1, 64) = 5.916, p = 0.029; days: F (4, 64) = 13.47, p < 0.001; group×day: F (4, 64) = 2.103, p = 0.093). During the probe trial, the platform crossing times and the time spent in the target quadrant were significantly decreased in APP/PS1 mice (Supplementary Figure 2B, 2C, p < 0.01 and p < 0.05, respectively). In addition, the latency to the target quadrant of the APP/PS1 group was significantly higher than that of the WT group (Supplementary Figure 2D, p < 0.05). In the FC test, increased freezing percent of APP/PS1 mice was shown in the context test (Supplementary Figure 3A, p < 0.01). However, there was no significant difference in the condition of cued test between the two groups (Supplementary Figure 3B, p > 0.05). These results revealed spatial memory deficits of 6-month-old APP/PS1 mice.

Subsequently, we evaluated the expression of the MAGUK family in the cortex and hippocampus of APP/PS1 mice. As shown in Fig. 1A, the protein levels of SAP-97, SAP-102, and PSD-95 were not significantly altered in the cortex and hippocampus of APP/PS1 mice (p > 0.05). The level of PSD-93 remained unchanged in the cortex. However, the protein level of PSD-93 was significantly decreased in the hippocampus of APP/PS1 mice compared with that in WT mice as demonstrated by western blot and IFA (Fig. 1A, C-E, p < 0.05 and p < 0.01, respectively). The expression of PSD-95 was decreased in the hippocampus of AD mice, but this difference was not significant (Fig. 1A, C).

Fig.1

The expression of PSD-93 was decreased in the hippocampus of 6-month-old APP/PS1 mice. A) Western blotting was used to determine the levels of MAGUKs in the hippocampal and cortical lysates. The expression of PSD-93 was reduced in the hippocampus, and no significant change was shown in the cortex (n = 4 per group). B, C) Quantification of the intensities normalized to GAPDH as a loading control. D) Immunofluorescence assay was used to detect the level of PSD-93 in the hippocampus. The expression of PSD-93 (green) was decreased in 6-month-old AD mice (n = 4 per group), as analyzed in (E). DAPI: blue (as a location indicator). Scale bars, 50 μm. *p < 0.05; **p < 0.01. Error bars show SEM.

PSD-93 overexpression ameliorates the cognitive deficits of APP/PS1 mice

To determine the role of PSD-93 in the cognitive deficits of AD, lenti-PSD-93 (Lv-PSD-93) or lenti-con (Lv-con) was injected into the hippocampus of APP/PS1 mice (6-month-old). As shown in Fig. 2A, GFP-tagged lentivirus was primarily detected in the hippocampus, and the effects of PSD-93 overexpression were also confirmed by western blotting (Fig. 2B). In the OFT test, there was no significant difference in the total distance travelled in the 10 min trial between the Lv-PSD-93 group and Lv-con group (p > 0.05) suggesting that PSD-93 overexpression has no effect on motor performance in AD model mice. Compared with that of the Lv-con group, the time in the center and corner zone of Lv-PSD-93 group was not significantly different (Fig. 2C, 2D, 2E), which indicated that PSD-93 overexpression did not affect the exploration behavior in APP/PS1 mice.

Fig.2

Lentivirus-medicated overexpression of PSD-93 in the hippocampus did not affect the basal activity of AD mice. The 6-month-old APP/PS1 mice were injected with lenti-PSD-93 (n = 12) or lenti-con (n = 13), and subsequently an open-field test was performed. A) Representative image of the lenti-PSD-93-injected hippocampus of APP/PS1 mice. Scale bars, 100 μm. B) The expression of PSD-93 was determined by western blotting. (n = 4 per group) There were no significant differences in the corner time (C), center time (D), and distance (E) between Lv-con (n = 13) and Lv-PSD-93 (n = 12) group. Error bars show SEM.

Spatial memory was assessed using a 6-day MWM test after the OFT test. The analysis of the mean latency revealed that Lv-con mice spent more time searching the hidden platform compared to the Lv-PSD-93 group (Fig. 3A, groups: F (1, 92) = 21.529, p < 0.001; days: F (4, 92) = 36.805, p < 0.001; group×day: F (4, 92) = 1.881, p = 0.123). On 6th test day, when the hidden platform was removed, the time of target platform crossings and the time in the target quadrant of Lv-PSD-93 group were significantly increased compared with that of the Lv-con group (Fig. 3B, 3C, p < 0.05 and p < 0.01, respectively). Moreover, the latency to the target quadrant of the Lv-PSD-93 group was significantly lower than that of the control group (Fig. 3D, p < 0.05), indicating that PSD-93 overexpression improved the spatial memory in AD mice.

Fig.3

Overexpression of PSD-93 in the hippocampus attenuates spatial memory deficits in AD mice. Morris water maze and fear condition tests were conducted to determine the effect of PSD-93 on spatial memory of AD mice (Lv-con, n = 11; Lv-PSD-93, n = 9). A) In the acquisition trial, the escape latency in the Lv-PSD-93 group was shorter than that in control mice. In the probe phase, Lv-PSD-93 mice showed significantly more crossing platform times (B) and time in the target quadrant (C). D) Latency to the target quadrant of Lv-PSD-93 mice performed better than WT mice. E) Representative image of path tracings in the probe test on day 6. To confirm the memory enhancement of AD mice, fear condition test was performed. F) No significant changes were shown in the percent of freezing of cued phase. G) The percentage of freezing time in Lv-PSD-93 mice was increased than control mice. *p < 0.05; **p < 0.01. Error bars show SEM.

To confirm the enhancement of cognitive function, the FC test was also conducted. In the context test, we observed an increased freezing percent in the Lv-PSD-93 group compared with that in the control group (Fig. 3G, p < 0.05). However, if the tone components of the cued were performed, then there was no difference between the two groups (Fig. 3F), which indicated that PSD-93 overexpression could rescue hippocampus-dependent memory in AD mice.

PSD-93 overexpression facilitates long-term potentiation of synapses in the hippocampus of APP/PS1 mice

Following the overexpression of PSD-93 using lentivirus, we detected the expression of the other members of MAGUK, such as PSD-95, SAP-97, and SAP-102, in the cortex and hippocampus of AD mice. As shown in Fig. 4A, C, and D, the levels of postsynaptic proteins were not significantly changed in the brains of Lv-PSD-93 mice. In addition, synaptophysin, a presynaptic vesicle protein, was also no changed (Fig. 4A, C, D). To examine the effect of PSD-93 overexpression on synaptic function of APP/PS1 mice, LTP was examined. The slope of hippocampus slices from Lv-PSD-93 mice was significantly increased compared with control mice (Fig. 4B, p < 0.01). The data showed that PSD-93 overexpression could ameliorate the LTP in the hippocampus of APP/PS1 mice.

Fig.4

PSD-93 overexpression facilitated the LTP of synapses in APP/PS1 mice. A) The levels of MAGUK family proteins and synaptophysin were examined in the hippocampus and cortex of Lv-PSD-93-injected AD mice (n = 4-5 mice per group). B) LTP of hippocampus slices was performed in the CA1 region of Lv-PSD-93 and their control APP/PS1 mice. The slope of the regression line in Lv-PSD-93 mice was markedly increased. C, D) Quantification of the signal intensities normalized to GAPDH as a loading control. *p < 0.05; **p < 0.01. Error bars show SEM.

PSD-93 overexpression increases NEP level and decreases Aβ levels in the brains of APP/PS1 mice

To further confirm the effects of PSD-93 overexpression on Aβ content, the levels of Aβ₄₀ and Aβ₄₂ in the hemisphere of the brains of APP/PS1 mice after lentivirus injection were examined using ELISA. The results showed that PSD-93 overexpression significantly decreased the levels of both TBS-soluble Aβ₄₀ and Aβ₄₂ (Fig. 5A, B, p < 0.05 and p < 0.01, respectively). For the TBSX-soluble forms, the levels of Aβ₄₀ were significantly decreased in the Lv-PSD-93 group (Fig. 5A, p < 0.01), whereas the corresponding change was not detected in the levels of Aβ₄₂ (Fig. 5B, p > 0.05). In addition, the levels of Aβ_40, but not Aβ₄₂, were decreased in the FA fractions in PSD-93-treated mice (Fig. 5A, B, p < 0.01 and p > 0.05 respectively). We also examined the Aβ plaque load in the brains of APP/PS1 mice using IFA, revealing that the Aβ plaque load was significantly decreased after PSD-93 injecting (Fig. 5C, D, p < 0.05).

Fig.5

PSD-93 overexpression increased NEP level and decreased Aβ levels in the brains of APP/PS1 mice. A) The levels of Aβ₄₀ in the brain were determined using ELISA (n = 7-8 mice per group). The fraction of TBS-, TBS-X-, and FA-soluble were decreased in the Lv-PSD-93 group. B) The Aβ₄₂ level of the TBS- fraction in the brains of Lv-PSD-93 group was decreased. No significant changes were observed in TBS-X and FA-soluble forms. C) Representative images of Aβ plaques in the hippocampus of lenti-PSD-93 injected APP/PS1 mice (n = 4 mice per group). Scale bars, 200 μm. D) The counts of Aβ plaque load in the hippocampus. E) The levels of Aβ metabolism associated enzymes and AβPP were determined using western blotting (n = 4-5 mice per group). The expression of NEP was increased in the hippocampus of Lv-PSD-93 mice. F, G) Quantification of the signal intensities normalized to GAPDH as a loading control. *p < 0.05; **p < 0.01. Error bars show SEM.

To explore the possible mechanisms of Aβ reduction after PSD-93 overexpression, the levels of amyloid-β protein precursor (APP), insulin degrading enzyme (IDE), NEP, and low density lipoprotein receptor-related protein 1 (LRP1) were analyzed. As shown in Fig. 5E, the expression of APP, IDE, and LRP1 was not affected in the brains of PSD-93 overexpressed AD mice. However, NEP was significantly increased in the hippocampus of Lv-PSD-93 group (Fig. 5E, p < 0.05). These data demonstrated that PSD-93 overexpression reduced the levels of Aβ₄₀ and Aβ₄₂ and increased the NEP expression in the APP/PS1 mice.

PSD-93 overexpression enhances the expression of SSTR4 in the hippocampus of APP/PS1 mice

The expression of SST and SSTR1 was not changed in the cortex and hippocampus of PSD-93 overexpressed APP/PS1 mice (Fig. 6A-C). However, SSTR4 was significantly increased in the hippocampus of APP/PS1 mice after PSD-93 overexpression (Fig. 6A-C, p < 0.01). The level of SSTR2 was moderately increased in the hippocampus, but the differences were not significant (Fig. 6A, C, p > 0.05).

Fig.6

The expression of SSTR4 was increased in the hippocampus of Lv-PSD-93 APP/PS1 mice. A) The level of SSTR4 was increased in the hippocampus of Lv-PSD-93-injected AD mice but not in the cortex. The levels of SSTR1, SSTR2, and SST were not significantly changed in the hippocampus and cortex in the Lv-PSD-93 group. B, C) Quantification of the signal intensities normalized to GAPDH as a loading control. **p < 0.01. Error bars show SEM.

PSD-93 interacts with SSTR4 and increases the level of SSTR4 in the cytomembrane

Given that PSD-95 was able to bind to SSTR4 in primary rat neurons, we next explored whether PSD-93 could interact with SSTR4 using IP and IFA. As shown in Fig. 7A, PSD-93 could be precipitated with SSTR4 in the cortex and hippocampus of APP/PS1 mice and vice versa. In addition, PSD-93 was found to merge with SSTR4 using IFA, indicating that PSD-93 interacted with SSTR4 in primary cortical neurons (Fig. 7B).

Fig.7

PSD-93 could interact with SSTR4 in the brains of APP/PS1 mice. A, B) The combination between PSD-93 and SSTR4 were examined using immunoprecipitation. C) Representative images of PSD-93 (green) and SSTR4 (red) in neurons. Scale bars, 20 μm. Error bars show SEM.

To determine the effect of PSD-93 on SSTR4 levels, we knocked down and overexpressed PSD-93 in neurons using lentivirus. The efficiency of lenti-shPSD-93 was determined through fluorescence microscopy and western blot analysis (Supplementary Figure 4, p < 0.01). The levels of PSD-93 and SSTR4 in membrane and cytosolic fractions were detected 72 h after lentivirus infection. The expression of SSTR4 was decreased in the membrane of PSD-93 knocked down neurons (Fig. 8A), while the SSTR4 level in the cytosolic fraction was not significantly changed (Fig. 8A). In addition, the increased levels of SSTR4 in PSD-93 overexpression neurons were primarily observed in the membrane fraction but not in the cytoslic fraction (Fig. 8A).

Fig.8

Ubiquitination of SSTR4 was changed by PSD-93 in the membrane fraction. A) The levels of SSTR4 in the membrane and cytoplasmic fraction affected by PSD-93 knockdown and overexpression were determined using western blotting. B) The effect of MG 132 on the level of SSTR4 after PSD-93 knockdown was determined by western blotting (n = 4). C) Quantification of the signal intensities normalized to GAPDH as a loading control. D) Relative SSTR4 mRNA levels in the neurons of PSD-93 knockdown and overexpression was determined by real-time PCR. E) Representative immunoblot of overexpression and inhibition of PSD-93 in neurons affected ubiquitination of SSTR4 pretreated with MG132. *p < 0.05. Error bars show SEM.

Subsequently, we explored whether the decreased SSTR4 was resulted from ubiquitination. In primary cortical neurons, the level of SSTR4 was decreased after PSD-93 inhibition, which was reversed by MG 132 (Fig. 8B, C, p < 0.05). As shown in Fig. 8D, overexpression or inhibition of PSD-93 did not change the mRNA level of SSTR4 in MG132-pretreated neurons. PSD-93 inhibition increased in the accumulation of high molecular-weight bands which were reactive to anti-ubiquitin antibody, while PSD-93 overexpression decreased the levels of SSTR4 ubiquitination (Fig. 8E). The results indicated that PSD-93 regulated the cytomembrane levels of SSTR4, which might be associated with the ubiquitination pathway.

DISCUSSION

In the present study, we have found that the expression of PSD-93 was decreased in the hippocampus of 6-month-old APP/PS1 mice, and PSD-93 overexpression could attenuate cognitive impairment, decrease the amyloid levels and ameliorate Aβ-induced LTP injury in APP/PS1 mice. Moreover, PSD-93 could interact with SSTR4, and regulate the membrane level of SSTR4, which was associated with the ubiquitination pathway. Increasing evidence indicates that Aβ-induced LTP impairment in the hippocampus is correlated with the clinical symptoms of AD [29], and plays a crucial role in learning and memory dysfunction. The MAGUK family plays an important role in synaptic function and memory formation [30, 31]. The dendrite ramification and spine generation are regulated by the expression of post synaptic proteins [32]. Studies using PSD-93 mutant mice exhibited deficits in hippocampal learning and memory-related LTP, but long-term depression remained normal [33]. In addition, PSD-93 knockout mice showed significant impairments in visual discrimination, cognitive flexibility, visuo-spatial learning and memory in attention [31]. In the brain of cold water stress-induced spatial memory impaired animals, the expression of PSD-93 decreased [34]. In the present study, we investigated the expression of MAGUKs in the hippocampus and cortex, and observed that the decline of PSD-93 in the hippocampus of APP/PS1 mice was found in 6 months old, which was not significantlychanged in cortex and other members of MAGUKs family. The results indicated that the level of PSD-93 might serve as an important biomarker to assess the early synaptic injury reduced by Aβ.

This study also showed that following PSD-93 overexpression, the memory impairment was attenuated, and the LTP was facilitated in APP/PS1 mice. Studies in 3X Tg-AD mice showed that after spatial training in the MWM, memory acquisition was significantly improved and maintained. The training enhanced synaptic plasticity and remarkably activated PSD-93 and PSD-95 in the hippocampus [35], which was consistent with our data. Notably, it was found that the level of the other MAGUKs and synaptophysin, a presynaptic vesicle protein, did not change. This finding suggests that short-term PSD-93 overexpression did not lead to detectable changes in the synaptic structure, but the functional declines could be ameliorated. Moreover, PSD-93 overexpression could decrease the Aβ₄₀ and Aβ₄₂ levels and Aβ plaques in APP/PS1 mice. The steady-state levels of Aβ in the brain are maintained via the balance of production and clearance. The catabolism of Aβ is controlled through several amyloid-degrading enzymes. In the brain, NEP has showed substantial affinity for Aβ, and the enhanced activity of NEP corresponds with significant decreases of Aβ in conjunction with cognitive function improvement [36, 37]. In our study, the expression of NEP in the hippocampus was increased after PSD-93 overexpression. Until recently, little is known about the relationship between PSD-93 and SSTR4 in APP/PS1 mice. It was observed that the protein level of SSTR4 in the hippocampus of PSD-93 overexpression APP/PS1 mice was also increased. The expression of PSD-93 presents an early-up, late-down pattern over the Braak stages in the hippocampus CA1 of AD patients [38]. Moreover, it has been reported that human mutations in PSD-93 was identified in several psychiatric disorders, such as major depressive disorder, and schizophrenia [39 –42]. Humans carrying PSD-93 mutations have significantly deficits in visual discrimination acquisition and cognitive flexibility and visuo-spatial learning and memory [31], which indicates that the mutation of PSD-93 might be a predictor of several diseases. It is a potential strategy to access the stages of AD by the level of PSD-93, and PSD-93 agonist might be used totreat AD.

A previous study showed that the decreased SST levels observed with AD progression were directly responsible for the loss of NEP activity and subsequent increases of Aβ₄₂ within the brain [14]. In the cortex or CSF samples of AD patients, the decline of SST is correlated with the extent of cognitive impairment. SSTR, which mediates the neuromodulatory signals of SST, is also markedly reduced in the AD brain [43]. The SSTRs can be divided into five subtypes: SSTR1-5. Among the five subtypes, SSTR2 and SSTR4 have the highest brain expression, but only SSTR4 is well expressed in cortical and hippocampal tissues [16]. Studies have shown that treated with the agonist of SSTR4 NNC 26-9100 improved the learning memory in AD mice and reduced the Aβ₄₂ oligomers within extracellular, intracellular, and membrane fractions [44]. Previous studies identified a combination between SSTR4 and SSTR1 and PDZ domains 1 and 2 of PSD-95 and PSD-93 using peptide affinity purification in rat neurons and this association might serve to localize the receptor to postsynaptic sites, especially in dendrites [45]. In the present study, we observed the interaction between PSD-93 and SSTR4 using immunoprecipitation and IFA. The levels of membrane SSTR4 could be regulated by PSD-93 in the cytomembrane. Additionally, the results of this study indicated that PSD-93 inhibition promoted the ubiquitination of SSTR4. Further studies are needed to better understand the mechanisms underlying these results.

Conclusions

In conclusion, the current study demonstrates that PSD-93 overexpression improves synaptic dysfunction and attenuates the memory deficits in AD mice. Furthermore, the mechanism might be associated with interactions between PSD-93 and SSTR4 and regulation the level of NEP, which plays an important role in the catabolism of Aβ.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the National Nature Science Foundation of China (81471102, 81230026, 81630028 and 81671055), the Natural Science Foundation (BE2016610) and Jiangsu Province Commision of Health and Family Planing (LJ201101), and the Six Talent Peak Project of Jiangsu Province (2015-WSN-083), the Key Project of Nanjing Medical Science and Technology Development Project (ZKX13020). Jiangsu Provinical Key Medical Discipline (Laboratory) (ZDXKA2016020).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Bettens

, Sleegers

, Van Broeckhoven

(2013) Genetic insights in Alzheimer’s disease. Lancet Neurol 12, 92–104.

Tarasoff-Conway

, Carare

, Osorio

, Glodzik

, Butler

, Fieremans

, Axel

, Rusinek

, Nicholson

, Zlokovic

, Frangione

, Blennow

, Menard

, Zetterberg

, Wisniewski

, de Leon

(2015) Clearance systems in the brain-implications for Alzheimer disease. Nat Rev Neurol 11, 457–470.

Zolezzi

, Bastias-Candia

, Santos

, Inestrosa

(2014) Alzheimer’s disease: Relevant molecular and physiopathological events affecting amyloid-beta brain balance and the putative role of PPARs. Front Aging Neurosci 6, 176.

Mawuenyega

, Sigurdson

, Ovod

, Munsell

, Kasten

, Morris

, Yarasheski

, Bateman

(2010) Decreased clearance of CNS beta-amyloid in Alzheimer’s disease. Science 330, 1774.

Jia

, Deng

, Qing

(2014) Potential therapeutic strategies for Alzheimer’s disease targeting or beyond beta-amyloid: Insights from clinical trials. Biomed Res Int 2014, 837157.

Kuruppu

, Rajapakse

, Spicer

, Parkington

, Smith

(2016) Stimulating the activity of amyloid-beta degrading enzymes: A novel approach for the therapeutic manipulation of amyloid-beta levels. J Alzheimers Dis 54, 891–895.

Blurton-Jones

, Spencer

, Michael

, Castello

, Agazaryan

, Davis

, Muller

, Loring

, Masliah

, LaFerla

(2014) Neural stem cells genetically-modified to express neprilysin reduce pathology in Alzheimer transgenic models. Stem Cell Res Ther 5, 46.

Nagura

, Ishikawa

, Kobayashi

, Takao

, Tanaka

, Nishikawa

, Tamura

, Shiosaka

, Suzuki

, Miyakawa

, Fujiyoshi

, Doi

(2012) Impaired synaptic clustering of postsynaptic density proteins and altered signal transmission in hippocampal neurons, and disrupted learning behavior in PDZ1 and PDZ2 ligand binding-deficient PSD-95 knockin mice. Mol Brain 5, 43.

Kim

, Sheng

(2004) PDZ domain proteins of synapses. Nat Rev Neurosci 5, 771–781.

10.

Vickers

, Mitew

, Woodhouse

, Fernandez-Martos

, Kirkcaldie

, Canty

, McCormack

, King

(2016) Defining the earliest pathological changes of Alzheimer’s disease. Curr Alzheimer Res 13, 281–287.

11.

Overk

, Masliah

(2014) Pathogenesis of synaptic degeneration in Alzheimer’s disease and Lewy body disease. Biochem Pharmacol 88, 508–516.

12.

Pham

, Crews

, Ubhi

, Hansen

, Adame

, Cartier

, Salmon

, Galasko

, Michael

, Savas

, Yates

, Glabe

, Masliah

(2010) Progressive accumulation of amyloid-beta oligomers in Alzheimer’s disease and in amyloid precursor protein transgenic mice is accompanied by selective alterations in synaptic scaffold proteins. FEBS J 277, 3051–3067.

13.

Grosser

, Neumann

, Horsthemke

, Zeschnigk

, van de Nes

(2014) Methylation analysis of SST and SSTR4 promoters in the neocortex of Alzheimer’s disease patients. Neurosci Lett 566, 241–246.

14.

Saito

, Iwata

, Tsubuki

, Takaki

, Takano

, Huang

, Suemoto

, Higuchi

, Saido

(2005) Somatostatin regulates brain amyloid beta peptide Abeta42 through modulation of proteolytic degradation. Nat Med 11, 434–439.

15.

Ramos

, Baglietto-Vargas

, del Rio

, Moreno-Gonzalez

, Santa-Maria

, Jimenez

, Caballero

, Lopez-Tellez

, Khan

, Ruano

, Gutierrez

, Vitorica

(2006) Early neuropathology of somatostatin/NPY GABAergic cells in the hippocampus of a PS1xAPP transgenic model of Alzheimer’s disease. Neurobiol Aging 27, 1658–1672.

16.

Moller

, Stidsen

, Hartmann

, Holst

(2003) Somatostatin receptors. Biochim Biophys Acta 1616, 1–84.

17.

Lachen-Montes

, Gonzalez-Morales

, de Morentin

, Perez-Valderrama

, Ausin

, Zelaya

, Serna

, Aso

, Ferrer

, Fernandez-Irigoyen

, Santamaria

(2016) An early dysregulation of FAK and MEK/ERK signaling pathways precedes the beta-amyloid deposition in the olfactory bulb of APP/PS1 mouse model of Alzheimer’s disease. J Proteomics 148, 149–158.

18.

Jin

, Liou

, Shi

, Yin

, Chen

, Li

, Zhu

, Qian

, Yang

, Chen

, Xu

(2015) CART treatment improves memory and synaptic structure in APP/PS1 mice. Sci Rep 5, 10224.

19.

, Wang

, Chen

, Yang

, Li

, Xu

, Zhu

(2015) Orientin alleviates cognitive deficits and oxidative stress in Abeta1-42-induced mouse model of Alzheimer’s disease. Life Sci 121, 104–109.

20.

Zhu

, Wang

, Yu

, Yang

, Tan

, Yin

, Jin

, Zhao

, Guan

, Xu

(2014) TL-2 attenuates beta-amyloid induced neuronal apoptosis through the AKT/GSK-3beta/beta-catenin pathway. Int J Neuropsychopharmacol 17, 1511–1519.

21.

Guan

, Haggarty

, Giacometti

, Dannenberg

, Joseph

, Gao

, Nieland

, Zhou

, Wang

, Mazitschek

, Bradner

, DePinho

, Jaenisch

, Tsai

(2009) HDAC2 negatively regulates memory formation and synaptic plasticity. Nature 459, 55–60.

22.

Gao

, Wang

, Mao

, Graff

, Guan

, Pan

, Mak

, Kim

, Su

, Tsai

(2010) A novel pathway regulates memory and plasticity via SIRT1 and miR-134. Nature 466, 1105–1109.

23.

Durairajan

, Huang

, Yuen

, Chen

, Kwok

, Liu

, Song

, Han

, Xue

, Chung

, Huang

, Baum

, Senapati

, Li

(2014) Effects of Huanglian-Jie-Du-Tang and its modified formula on the modulation of amyloid-beta precursor protein processing in Alzheimer’s disease models. PLoS One 9, e92954.

24.

Meng

, Li

, Chen

, Yu

, Zhang

, Lao

, Zhang

, Xu

(2016) Neuronal soluble Fas ligand drives M1-microglia polarization after cerebral ischemia. CNS Neurosci Ther 22, 771–781.

25.

Pan

, Jin

, Ge

, Yin

, Chen

, Han

, Chen

, Qian

, Li

, Xu

(2015) Malibatol A regulates microglia M1/M2 polarization in experimental stroke in a PPARgamma-dependent manner. J Neuroinflammation 12, 51.

26.

Zhang

, Li

, Chen

, Li

, Zhang

, Chen

, Zhang

, Shao

, Xu

(2014) PSD-93 deletion inhibits Fyn-mediated phosphorylation of NR2B and protects against focal cerebral ischemia. Neurobiol Dis 68, 104–111.

27.

Braak

, Braak

(1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82, 239–259.

28.

Roy

, Arons

, Mitchell

, Pignatelli

, Ryan

, Tonegawa

(2016) Memory retrieval by activating engram cells in mouse models of early Alzheimer’s disease. Nature 531, 508–512.

29.

Xiang

, Janc

, Grochowska

, Kreutz

, Reymann

(2016) Dopamine agonists rescue Abeta-induced LTP impairment by Src-family tyrosine kinases. Neurobiol Aging 40, 98–102.

30.

Levy

, Chen

, Reese

, Nicoll

(2015) Synaptic consolidation normalizes AMPAR quantal size following MAGUK loss. Neuron 87, 534–548.

31.

Nithianantharajah

, Komiyama

, McKechanie

, Johnstone

, Blackwood

, St Clair

, Emes

, van de Lagemaat

, Saksida

, Bussey

, Grant

(2013) Synaptic scaffold evolution generated components of vertebrate cognitive complexity. Nat Neurosci 16, 16–24.

32.

Engert

, Bonhoeffer

(1999) Dendritic spine changes associated with hippocampal long-term synaptic plasticity. Nature 399, 66–70.

33.

Carlisle

, Fink

, Grant

, O’Dell

(2008) Opposing effects of PSD-93 and PSD-95 on long-term potentiation and spike timing-dependent plasticity. J Physiol 586, 5885–5900.

34.

Jiang

, Jia

, Li

, Cheng

, Xie

, Ma

, Xu

, Liu

, Yao

, Du

, Zhou

(2013) Capsaicin ameliorates stress-induced Alzheimer’s disease-like pathological and cognitive impairments in rats. J Alzheimers Dis 35, 91–105.

35.

Jiang

, Chai

, Wang

, Hu

, Li

, Ma

, Wang

, Liu

(2015) Spatial training preserves associative memory capacity with augmentation of dendrite ramification and spine generation in Tg2576 mice. Sci Rep 5, 9488.

36.

Nalivaeva

, Beckett

, Belyaev

, Turner

(2012) Are amyloid-degrading enzymes viable therapeutic targets in Alzheimer’s disease? J Neurochem 120 Suppl 1, 167–185.

37.

Saido

, Leissring

(2012) Proteolytic degradation of amyloid beta-protein. Cold Spring Harb Perspect Med 2, a006379.

38.

Hondius

, van Nierop

, Li

, Hoozemans

, van der Schors

, van Haastert

, van der Vies

, Rozemuller

, Smit

(2016) Profiling the human hippocampal proteome at all pathologic stages of Alzheimer’s disease. Alzheimers Dement 12, 654–668.

39.

Walsh

, McClellan

, McCarthy

, Addington

, Pierce

, Cooper

, Nord

, Kusenda

, Malhotra

, Bhandari

, Stray

, Rippey

, Roccanova

, Makarov

, Lakshmi

, Findling

, Sikich

, Stromberg

, Merriman

, Gogtay

, Butler

, Eckstrand

, Noory

, Gochman

, Long

, Chen

, Davis

, Baker

, Eichler

, Meltzer

, Nelson

, Singleton

, Lee

, Rapoport

, King

, Sebat

(2008) Rare structural variants disrupt multiple genes in neurodevelopmental pathways in schizophrenia. Science 320, 539–543.

40.

Duric

, Banasr

, Stockmeier

, Simen

, Newton

, Overholser

, Jurjus

, Dieter

, Duman

(2013) Altered expression of synapse and glutamate related genes in post-mortem hippocampus of depressed subjects. Int J Neuropsychopharmacol 16, 69–82.

41.

Kristiansen

, Beneyto

, Haroutunian

, Meador-Woodruff

(2006) Changes in NMDA receptor subunits and interacting PSD proteins in dorsolateral prefrontal and anterior cingulate cortex indicate abnormal regional expression in schizophrenia. Mol Psychiatry 11, 737–747, 705.

42.

, Roos

, Levy

, van Rensburg

, Gogos

, Karayiorgou

(2008) Strong association of de novo copy number mutations with sporadic schizophrenia. Nat Genet 40, 880–885.

43.

Perez-Cruz

, Nolte

, van Gaalen

, Rustay

, Termont

, Tanghe

, Kirchhoff

, Ebert

(2011) Reduced spine density in specific regions of CA1 pyramidal neurons in two transgenic mouse models of Alzheimer’s disease. J Neurosci 31, 3926–3934.

44.

Sandoval

, Farr

, Banks

, Crider

, Morley

, Witt

(2013) Somatostatin receptor subtype-4 agonist NNC 26-9100 mitigates the effect of soluble Abeta(42) oligomers via a metalloproteinase-dependent mechanism. Brain Res 1520, 145–156.

45.

Christenn

, Kindler

, Schulz

, Buck

, Richter

, Kreienkamp

(2007) Interaction of brain somatostatin receptors with the PDZ domains of PSD-95. FEBS Lett 581, 5173–5177.