Materials and methods

Chemical and reagents

The RVG-29 was synthesized by Shenzhen Meluo Technology Co., Ltd. (Shenzhen, China). DSPE-PEG2000 was obtained from Shenzhen Meluo Technology Co., Ltd. (Shenzhen, China). Exosome Isolation and Purification Kit (from Plasma or Serum) Plus (UR52151) and the exosome-labeling dyes PKH67 (UR52303) and DiRm (UR21017) were obtained from Umibio (Shanghai) Co., Ltd. (Shanghai, China). Anti-CD63 antibody (A5271) was obtained from ABclonal Biotechnology Co., Ltd. (Wuhan, China). Anti-ALIX antibody (ab186429), Anti-P62 antibody (ab109012), Anti-phospho-Tau (Ser396) antibody (ab32057), Anti-neun antibody (AB104224), Anti-Synaptophysin antibody (AB32127), and Anti-MAP2 antibody (ab5392) were obtained from Abcam (Waltham, USA). Anti-Aβ antibody (15126), Anti-LAMP1 antibody (15665), and Anti-Raptor antibody (2280) were obtained from Cell Signaling Technology (Danvers, USA). LC3B-Specific Polyclonal antibody (18725-1-AP), Anti-GAPDH antibody (60004-1-lg), Anti-phospho-Tau (Ser202/Thr205) antibody (82568-1-RR), Anti-rabbit IgG (SA00001-2) antibody, Anti-mouse IgG (SA00001-1) antibody, and Anti-Beta Actin antibody (20536-1-AP) were obtained from Proteintech (Wuhan, China). Dulbecco's Modified Eagle Medium (DMEM) (PM150210, Procell, China), fetal bovine serum (FBS) (164210), and Minimum Essential Medium (PM150410) were obtained from Procell Life Science & Technology Co., Ltd. (Wuhan, China). Amyloid β peptide (1-42) human (P9001-5 mg, Beyotime, China), okadaic acid (S1786-10 μg), Calcein-AM/PI Cell Viability/Cytotoxicity Assay Kit (C1371S), RIPA Lysis Buffer (strong) (P0013B), and Western and IP Cell Lysis Buffer (P0013) were obtained from Beyotime Biotechnology (Shanghai, China). Lentiviral vectors for overexpressing RPTOR (human) and Rptor (mouse) were constructed by Shanghai GeneChem Co., Ltd. (Shanghai, China). 4% paraformaldehyde fixative (143174) was obtained from Biosharp Life Sciences (Beijing, China). Methanol was obtained from Tianjin Fuyu Fine Chemical Co., Ltd. (Tianjin, China). 20× Tris-EDTA Antigen Retrieval Buffer (pH 9.0) (G1203-250 ML), Antifade Mounting Medium with DAPI (G1407-25 ML), SWE Rapid High-Resolution Electrophoresis Buffer (powder) (G2081-1L), and Tris-Glycine Transfer Buffer (Powder) (G2017-1L) were obtained from Wuhan Servicebio Technology Co., Ltd. (Wuhan, China). Phosphatase Inhibitor Cocktail (100×) (GRF102) was obtained from Shanghai Yamoen Biomedical Technology Co., Ltd. (Shanghai, China). Triton X-100 (T8200), PMSF (P0100), and BCA Protein Assay Kit (PC0020) were obtained from Beijing Solarbio Science & Technology Co., Ltd. (Beijing, China). Mouse IL-6 ELISA Kit (EK0411), Mouse IL-1 beta/IL1B ELISA Kit (EK0394), and Mouse TNF Alpha/TNFA ELISA Kit (EK0527) were obtained from Boster Biological Technology Co., Ltd. (Pleasanton, USA)

Animals

Eight-month-old triple-transgenic AD (3×Tg-AD) mice [B6C3-Tg; (APPswe, Psen1M146V, tauP301L)/V, male] and their age- and sex-matched wild-type (WT) mice were purchased from Jinan Xingkang Biotech (Jinan, China) and housed in a temperature-controlled facility under a 12-h light/dark cycle with ad libitum access to sterile food and water. The animals were allowed to acclimate for 1 week prior to the initiation of experimental procedures, which were approved by the Animal Care and Use Committee of Shandong Provincial Hospital affiliated to Shandong First Medical University and were conducted following the institutional guidelines.

Isolation of exosomes from young human plasma

Healthy young donors were recruited through Shandong Provincial Hospital. All participants underwent a detailed health assessment prior to donation, including questionnaire surveys and medical report reviews. The inclusion criteria were healthy individuals aged between 18 and 25 years. The exclusion criteria included a diagnosis of cancer, HIV, COVID-19, any form of brain disease, or testing positive for hepatitis B or C surface antigens or antibodies. All donors provided written informed consent. Blood samples collected from donors were centrifuged at 1150 × g for 20 min to isolate plasma, which was then aliquoted into 500 μL portions and immediately stored at −80 °C for subsequent EXO extraction. EXOs were purified from 500 μL of plasma using the Exosome Isolation and Purification Kit (from Plasma or Serum) Plus (UR52151, Umibio Biotechnology, Shanghai, China), strictly following the manufacturer's instructions. The precipitation-based kit employed was inherently scalable and could be adapted for large-scale production by processing larger volumes of plasma or utilizing automated systems.

To minimize the impact of individual variability, plasma samples were collected from multiple healthy young donors (aged 18–25 years). EXOs were isolated separately from each donor's plasma, and equal amounts of EXOs from each donor were then pooled to obtain the final EXOs preparation. This pooling strategy has been widely adopted in EXO research to reduce batch effects and individual bias, thereby enhancing the reproducibility and generalizability of the experimental findings [57].

This study was approved by the Ethics Committee of the Provincial Hospital affiliated with Shandong First Medical University for Biomedical Research Involving Human Subjects (NSFC No. 2022–511). All procedures were conducted in accordance with the ethical principles of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrollment.

RVG-engineered exosome production

RVG-EXOs were constructed by Shenzhen Meluo Technology Co., Ltd. (Shenzhen, China). First, the RVG-29 peptide and the lipid anchor DSPE-PEG2000 were mixed. Then, the solution was dried under a nitrogen stream to form a lipid film, which was subsequently hydrated with PBS (pH 7.4) in a 55 °C water bath. Finally, ultrasonication was applied to prepare the DSPE-PEG2000-RVG29 nanomicelle solution.

The extraction of EXOs from young human plasma was performed as previously described (see the “Exosome Isolation” section). The purified EXOs were mixed with the prepared DSPE-PEG2000-RVG29 nanomicelle solution and co-incubated at 37 °C with gentle agitation. Following incubation, the mixture was washed twice via ultracentrifugation to remove unbound micelles and free peptides. The final pellet was resuspended in sterile PBS to obtain RVG-EXOs, which were aliquoted and stored at −80 °C for future use.

Characterization of EXOs and RVG-EXOs

For morphological observation of EXOs and RVG-EXOs, the samples resuspended in PBS filtered through a 0.02-μm membrane were applied onto a 200-mesh carbon-coated copper grid and left to adsorb at room temperature for 5 min. Subsequently, excess liquid was carefully removed with filter paper, and the grid was negatively stained with uranyl acetate solution for 10 s. After drying at room temperature for 30 min, the samples were finally imaged using TEM (H7700, Hitachi, Japan).

To determine the size distribution and concentration of EXOs and RVG-EXOs, samples were diluted to 500 μL with PBS filtered through a 0.02-μm membrane. NTA was then performed using a NanoSight NS300 system (Malvern Instruments, UK). During detection, the NTA software (version 3.4, Build 3.4.4) was employed with the camera level set to 11 and the detection threshold set to 5 for data acquisition and analysis.

To measure the zeta potential of EXOs and RVG-EXOs, purified EXOs from young human plasma and RVG-EXOs were diluted in 1× PBS to a final concentration of approximately 1 × 109 particles/mL. Measurements were performed at Shenzhen Meluo Technology Co., Ltd. (Shenzhen, China) using a ZetaPlus analyzer (Brookhaven Instruments Corporation, New York, USA). Each sample was loaded into a disposable clear zeta potential cuvette, equilibrated at 25 °C for 60 s, and then the absorbance was measured. Three independent measurements were conducted per sample, with each measurement consisting of 15 runs. The zeta potential values were calculated using the instrument's built-in Smoluchowski model. Data were presented as the mean ± standard deviation (SD) of the three independent experiments.

To quantitatively analyze the binding efficiency of RVG peptides on the EXO surface, nanoscale flow cytometry analysis was performed for EXO and RVG-EXO samples by Shenzhen Meluo Technology Co., Ltd. (Shenzhen, China). Briefly, samples were diluted in PBS to approximately 1 × 108 particles/mL and incubated with an FITC-conjugated RVG antibody at 4 °C in the dark for 30–60 min, with an isotype control set in parallel. Unbound antibodies were then removed using a size-exclusion column, and the eluate was resuspended and calibrated with standard fluorescent nanoparticles. Finally, more than 10,000 particle events were acquired on a nanoflow cytometer (e.g., NanoFCM). The EXO population was gated based on side scatter, and the FITC fluorescence signal was analyzed to calculate the percentage of positive particles and the median fluorescence intensity, thereby assessing the RVG modification efficiency.

Tracking of EXOs and RVG-EXOs in vivo and in vitro

To monitor the in vivo biodistribution and cellular uptake of EXOs and RVG-EXOs, the EXOs were labeled with the lipophilic fluorescent dyes DiR (UR21017) and PKH67 (UR52303) obtained from Umibio Biotechnology (Shanghai, China) for in vivo and in vitro tracking experiments, respectively.

To evaluate the long-term in vivo distribution of EXOs and RVG-EXOs, purified EXOs were labeled by co-incubation with DiR dye at 37 °C in the dark for 30 min. The labeling reaction was terminated by adding an equal volume of PBS containing 1% BSA. Unbound dye was removed through a secondary purification step (either ultracentrifugation or size-exclusion chromatography). Six experimental mice received intravenous injections of 200 μg DiR-labeled EXOs or RVG-EXOs via the tail vein, while control mice were injected with free DiD dye. For three of the mice, the heart, liver, spleen, lungs, kidneys, and brain were collected 6 h after injection to capture fluorescence signals from these major organs. For the other three mice, whole-body fluorescence signals were collected at 6, 24, and 48 h post-injection using the IVIS Lumina II in vivo imaging system from Keygen BioTECH (Jiangsu, China).

To investigate the brain-targeting capability and cellular internalization of EXOs and RVG-EXOs, the EXOs were labeled with the PKH67 dye. The suspension of EXOs or RVG-EXOs was thoroughly mixed with the PKH67 dye by vortexing for 1 min, followed by incubation at 37 °C in the dark for 10 min. The staining was terminated by adding 1% BSA/PBS, and free dye was removed through a purification step. Mice were intravenously injected via the tail vein with 200 μg of PKH67-labeled EXOs or RVG-EXOs. In 4 h and 24h, brain tissues were collected and processed to make frozen sections for observation of EXO distribution using fluorescence microscopy.

To directly observe the cellular uptake of EXOs and RVG-EXOs, cultured neuronal cells were co-incubated with PKH67-labeled EXOs or RVG-EXOs under standard incubation conditions. After incubation, the culture medium was removed, and the cells were washed with PBS. Subsequently, the cells were fixed and examined under a fluorescence microscope to assess the internalization of EXOs or RVG-EXOs.

Animal Treatments

Two animal treatment plans were conducted.

Plan A. This study utilized 8-month-old 3xTg AD model mice (male), with their WT littermates serving as controls. All mice were randomly assigned to four groups (n = 8 per group): the Tg-Vehicle group (AD control), the WT-Vehicle group (WT control), the Tg-EXO group (treated with young plasma-derived EXOs), and the Tg-RVG-EXO group (treated with RVG-engineered EXOs). The protein concentrations of the young plasma-derived EXOs and RVG-EXOs were determined using a BCA protein assay kit (PC0020, Solarbio, Beijing, China). The administration concentration was selected based on previous studies [9], which demonstrated that EXOs isolated from young plasma, diluted with PBS to a total protein concentration of 1.8 μg/μL and administered via tail vein injection to aged mice, effectively improved cognitive function without causing significant toxicity to other organs. To ensure consistency with the effective dose reported in the literature, the working concentration of EXOs and RVG-EXOs in this study was set at 1.8 μg/μL. During the treatment, mice in the Tg-EXO and Tg-RVG-EXO groups received intravenous injections via the tail vein of 100 μL of EXOs or RVG-EXOs at a concentration of 1.8 μg/μL, administered three times per week for 4 consecutive weeks. The Tg-Vehicle and WT-Vehicle control groups were injected with an equal volume of PBS following the same dosing regimen.

Plan B. Eight-month-old male 3xTg mice were randomly divided into four groups (n = 8 per group): the Tg-EXO group (treated with EXOs), the Tg-RVG-EXO group (treated with RVG-EXOs), the Tg-RVG-EXO-NC group (injected with empty lentivirus and RVG-EXOs), and the Tg-RVG-EXO-OE group (injected with Rptor-overexpressing lentivirus and RVG-EXOs). The lentiviruses containing mouse Rptor (LV-gcGFP) and the empty lentivirus were provided by Shanghai GeneChem Co., Ltd. (Shanghai, China). Lentivirus with a titer of 1.5 × 109 TU/mL were stereotactically injected into the hippocampus of mice at the following coordinates: AP, −2.0 mm; ML, ± 1.0 mm; DV, −2.0 mm [58], with a rate of 400 nL/min and a total volume of 1 μL unilaterally. The procedure for intravenous injection of EXOs or RVG-EXOs into mice was performed as described above in the treatment protocol of Plan A. Upon completion of behavioral testing, all mice underwent perfusion, and brain tissue samples were immediately collected for further analysis.

Morris Water Maze test

To assess spatial learning and memory abilities, the MWM test was performed on mice from each group, as previously reported [59]. Mice were first trained in a circular pool (diameter: 120 cm; water temperature: 22 ± 1 °C) that was divided into four quadrants with distinct visual cues. The experimental protocol consisted of an adaptation and visible platform session on day 1, followed by 5 consecutive days of hidden platform training (4 trials per day), and a probe test on day 7. On day 1 (visible platform session), the platform (10 cm in diameter), marked with a visible cue, was placed above the water surface for mice to identify. During days 2–6 (hidden platform navigation), the platform was fixed in the center of the target quadrant, submerged approximately 2 cm below the water surface. Each mouse underwent multiple trials daily. In each trial, the mouse was gently placed into the water and allowed to swim freely for a maximum of 60 s. If the mouse successfully located and mounted the platform, then it was allowed to remain there for 5 s before being removed. If the mouse failed to find the platform within 60 s, then it was gently guided to the platform and allowed to stay for 5 s. On day 7 (probe test), the platform was completely removed. Each mouse was introduced into the pool from the entry point in the quadrant opposite to the original platform location and allowed to swim freely for 60 s. An overhead tracking system was used to record escape latency and the number of platform crossings.

Y-maze test

The Y-maze test was employed to assess short-term spatial working memory in mice [60]. The Y-maze consisted of three identical arms (each of 40 cm in length, 15 cm in height, and 10 cm in width) arranged at 120° angles to each other, with interiors made of grey opaque material. Prior to testing, all mice (grouping consistent with the Plan A or Plan B of Animal Treatments) were placed in a quiet behavioral laboratory for a 30-min habituation period. During the formal test, a single mouse was placed in the central area and allowed to freely explore all three arms for a total of 5 min. The movement trajectory of the mouse was recorded using the ANY-maze video tracking system. The primary analytical metric was based on the Spontaneous Alternation Rate, defined as the percentage of actual alternations (consecutive entries into three different arms) relative to the maximum possible alternations (total arm entries minus 2). All experiments were conducted under consistent lighting, temperature, and humidity conditions at the same fixed time each day. The maze was thoroughly wiped with 75% ethanol after each mouse to eliminate odor cues.

Open Field Test

The OFT was conducted to evaluate autonomous exploratory behavior and the anxiety-like state in mice [61]. The open field apparatus contained a white square open chamber (length × width × height = 50 × 50 × 40 cm). The floor was divided by black grid lines into 16 squares of equal area, with the central region defined as a square area of 25 cm × 25 cm. The experiment was conducted in a quiet behavioral laboratory with uniform lighting. During testing, a single mouse was gently placed in the central square and allowed to explore freely for 10 min. A video tracking system was used to record the movement trajectory of the mouse throughout the session. Analyzed parameters included duration spent in the central area, number of entries into the central area, total distance traveled, and average movement speed. After each test, the interior walls and floor of the open field chamber were thoroughly cleaned with 75% ethanol to eliminate residual odors.

Novel object recognition test

The NOR test was employed to assess non-spatial recognition memory in mice [62]. The experimental apparatus contained a white square open field (length × width × height = 50 × 50 × 40 cm). Two identical objects that were difficult for mice to displace (e.g., building blocks or glass bottles, approximately 10 cm in height) were used for testing. The procedure consisted of three phases. (1) During the habituation phase (day 1), each mouse was placed individually into the empty apparatus and allowed to explore freely for 5 min (2) During the familiarization phase (24 h after the habituation phase), two identical objects were placed symmetrically in the apparatus, and the mouse was allowed to explore freely for 5 min (3) During the test phase (1 h after the familiarization phase), one of the two objects was replaced with a novel object (different in shape, color, or texture), and the mouse was again allowed to explore freely for 5 min. All sessions were recorded using a video tracking system, and active exploration behavior—defined as when the nose of the mouse was within 2 cm of an object—was analyzed with manual assistance. The primary outcome measure was the RI, calculated using the formula: [(time spent exploring the novel object)/(time spent exploring the novel object + time spent exploring the familiar object)] × 100%. After each test, the apparatus and objects were thoroughly cleaned with 75% ethanol to eliminate odor cues.

Nissl staining

To assess the survival of hippocampal neurons, Nissl staining was performed on brain tissue sections, as previously described [63]. After transcardial perfusion under anesthesia, the mouse brains were removed and fixed in 4% paraformaldehyde for 24 h. Following dehydration in a graded sucrose series, coronal sections of 30 μm thickness were prepared using a cryostat. Prior to staining, the sections were rinsed with 0.1 M PBS and then stained in 0.5% toluidine blue solution at 37 °C for 30 min in the dark. After staining, the sections were rapidly differentiated and dehydrated through a graded ethanol series (70%, 95%, and 100%), cleared in xylene, and mounted with neutral balsam. All sections were scanned in their entirety using a digital slide scanner. The number of neurons with clearly defined Nissl bodies was quantified using ImageJ software (National Institutes of Health, USA), with three non-consecutive sections randomly selected per mouse for counting.

Transmission electron microscopy analysis

TEM was employed to observe and quantitatively analyze the ultrastructure of autophagolysosomes in the hippocampal tissue of mice. Following cardiac perfusion, brain tissues were fixed in electron microscopy-grade fixative, post-fixed with 1% osmium tetroxide, dehydrated, and embedded in 812 resin. Ultrathin sections (60–80 nm thickness) were prepared, stained with uranyl acetate and lead citrate, and imaged using a Hitachi HT7800 TEM. Quantitative analysis of autophagolysosomes was performed using ImageJ software (National Institutes of Health, USA).

Golgi staining

To observe and analyze the morphology and density of neuronal dendritic spines, Golgi staining was performed on brain tissues, as previously reported [64]. Freshly obtained mouse brain tissues were immersed in Golgi staining solution (containing potassium dichromate, mercuric chloride, etc.) and impregnated in the dark for 14 d. The tissues were then transferred to a 30% sucrose solution until the samples sank to the bottom of the solution. Coronal sections of 150 μm thickness were prepared using a vibratome. Following dehydration through a graded ethanol series, the sections were cleared in xylene and mounted with neutral resin. All sections were observed under a light microscope to examine dendritic spine morphology. Dendritic spine density and morphology were quantitatively analyzed using ImageJ software (National Institutes of Health, USA).

Immunofluorescence staining of mouse brain tissues

The immunofluorescence staining was used to localize and detect specific antigens in brain tissue sections. The specific procedures were as follows: brain sections stored at −80 °C were retrieved and equilibrated to room temperature in a humidified chamber. The tissue area was circled with an immunohistochemistry pen, covered with PBS for 2 min, and then the PBS was removed. The sections were fixed with 4% paraformaldehyde at room temperature for 30 min and then washed three times (each of 5 min) with an adequate volume of PBS on a shaker. Antigen retrieval was then performed as follows: the sections were immersed in 1× EDTA retrieval buffer (pH 8.0) and heated in a microwave oven on medium power for 8 min, allowed to stand for 8 min, and then heated again on medium-low power for 7 min. After retrieval, the sections were cooled naturally to room temperature. Permeabilization was subsequently conducted by incubating with 0.5% Triton X-100 at room temperature for 10–15 min. The tissue area was re-circled with the pen, followed by three washes with PBS. The sections were blocked with 3% BSA blocking solution at room temperature for 1 h. After removing the blocking solution, the corresponding primary antibody working solution was applied, and the sections were incubated at 4 °C overnight. Then, the sections were washed three times with PBST. A species-matched fluorescent secondary antibody (diluted 1:500) was then applied, and the sections were incubated at room temperature for 1 h in the dark, followed by three washes with PBST. Finally, the sections were counterstained for nuclei by applying an antifade mounting medium containing DAPI, covered with a coverslip, protected from light, and observed under a confocal microscope for image acquisition. Fluorescence intensity and positive cell quantification were performed using ImageJ software (National Institutes of Health, USA). The primary antibodies used in this experiment included Anti-Aβ antibody (15126, Cell Signaling Technology, Danvers, USA), Anti-phospho-Tau (Ser202/Thr205) antibody (82568-1-RR, Proteintech, Wuhan, China), and Anti-P62 antibody (ab109012), Anti-neun antibody (AB104224), Alexa Fluor® 555 (ab150118), and Alexa Fluor® 488 (ab150081) obtained from Abcam (Waltham, USA).

Hematoxylin and eosin staining

Paraffin sections were dewaxed and rehydrated through a graded ethanol series, then the sections underwent staining with an hematoxylin and eosin (H&E) HD constant dye kit (G1076, Servicebio, China) and a Masson dye solution set (G1006, Servicebio, China) according to the manufacturer's instructions [65]. Stained sections were scanned using a light microscope (Nikon Eclipse 100), and histological features were evaluated by image analysis.

Cell culture

The human neuroblastoma SH-SY5Y cell line was used for all in vitro experiments. Cells were cultured in MEM medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin solution and maintained in a humidified incubator at 37 °C with 5% CO2. The cells grew adherently and were passaged every 2–3 d using 0.25% trypsin-EDTA solution when reaching 80–90% confluence [66]. All experiments were performed using cells in the logarithmic growth phase with a passage number of less than 30.

Cell treatments

Two treatment plans were applied to the cells.

Plan A. SH-SY5Y cells were pretreated with 20 μM Aβ1-42 for 48 h and subsequently co-cultured for 24 h with either PBS, EXOs, or RVG-EXOs. The group receiving PBS only was designated as the control. The dosage of EXOs and RVG-EXOs in cell experiments was determined based previous studied [9], which confirmed that incubating young plasma-derived EXOs with a total protein content of 50 μg with 1 × 106 NE-4C or C2C12 cells effectively improved cellular function without inducing cytotoxicity. Accordingly, in this study, young plasma-derived EXOs or RVG-EXOs with a total protein content of 50 μg were incubated with 1 × 106 SH-SY5Y cells for treatment.

Plan B. Untransfected SH-SY5Y cells were treated with 20 μM Aβ1-42 and then co-cultured with either EXOs or RVG-EXOs. SH-SY5Y cells transfected with either an empty lentivirus or an RPTOR-overexpressing lentivirus, after the same Aβ1-42 pretreatment as that in Plan A, were co-cultured with RVG-EXOs for 24 h.

Protein extraction and Western blot analysis

Protein samples were prepared by lysing cells or tissues with RIPA lysis buffer containing protease and phosphatase inhibitor cocktails, followed by centrifugation to collect the supernatant. Protein concentrations of the samples were determined using a BCA protein assay kit (PC0020, Solarbio, Beijing, China). Equal amounts of total protein were separated by 12% SDS-polyacrylamide gel electrophoresis and then transferred onto PVDF membranes (ISEQ00010, Merck Millipore, Shanghai, China) using a wet transfer method. The membranes were blocked with 5% skim milk in TBST at room temperature for 1 h, followed by incubation with corresponding primary antibodies at 4 °C overnight. GAPDH or β-actin was used as an internal reference protein for normalization. Horseradish peroxidase (HRP)-conjugated secondary antibodies, including anti-rabbit IgG (SA00001-2) and anti-mouse IgG (SA00001-1), were used for detection. The primary antibodies used in this experiment included Anti-CD63 antibody (A5271, Abclonal, Wuhan, China); Anti-Synaptophysin antibody (AB32127), Anti-ALIX antibody (ab186429), Anti-P62 antibody (ab109012), and Anti-phospho-Tau (Ser396) antibody (ab32057) obtained from Abcam (Waltham, USA); LC3B-Specific Polyclonal antibody (18725-1-AP), Anti-GAPDH antibody (60004-1-lg), and Anti-Beta Actin antibody (20536-1-AP) obtained from Proteintech (Wuhan, China); Anti-Raptor antibody (2280, Cell Signaling Technology, Danvers, USA).

Construction of RPTOR-overexpressing SH-SY5Y stable cell line

RPTOR-overexpressing SH-SY5Y stable cell lines were established by lentiviral transduction. Briefly, SH-SY5Y cells were seeded in 24-well plates at an appropriate density and cultured for 16–24 h to reach 20–40% confluence. For infection, 40 μL of 25× HiTransG P reagent and the calculated volume of lentivirus (corresponding to MOI = 20 for SH-SY5Y cells; virus volume = MOI × cell number/viral titer) were added to each well. After 16 h of incubation at 37 °C, the medium was replaced with fresh complete medium. At 72 h post-infection, when cell confluence reached 70–80%, puromycin selection was initiated at 3.5 μg/mL for 48 h, followed by maintenance in medium containing 1 μg/mL puromycin until stably transduced polyclonal populations were expanded. Parallel infections with empty vector lentivirus were performed as controls.

Immunofluorescence staining of cells

Immunofluorescence staining of cells was performed as follows: after discarding the culture medium, cells were washed three times with PBS (each of 5 min) and then fixed with 4% paraformaldehyde (1 mL per well for a 12-well plate format) at room temperature for 10 min, followed by three PBS washes. Subsequently, permeabilization was carried out using 0.1% Triton X-100 for 1 min, then cells were washed three times with PBS. Blocking was performed by applying 5% BSA blocking solution at room temperature for 0.5–2 h. After removing the blocking solution, cells were directly incubated with PBS-diluted primary antibody at 4 °C overnight. Then, the primary antibody was retrieved, and cells were washed three times with PBST. The corresponding fluorescent secondary antibody (diluted 1:500 in PBST) was then applied and incubated at room temperature for 1 h in the dark, followed by three PBST washes. Finally, cell nuclei were counterstained by applying an antifade mounting medium containing DAPI. A coverslip was placed over the sample and sealed with nail polish. Imaging was performed using a confocal microscope. Fluorescence intensity and positive cell quantification were performed using ImageJ software (National Institutes of Health, USA). The primary antibodies used in this experiment included Anti-P62 antibody (ab109012), Anti-neun antibody (AB104224), Anti-MAP2 antibody (ab5392), Alexa Fluor® 488 (ab150081), and Alexa Fluor® 555 (ab150118) obtained from Abcam (Waltham, USA); Anti-LAMP1 antibody (15665, Cell Signaling Technology, Danvers, USA); and LC3B-Specific Polyclonal antibody (18725-1-AP, Proteintech, Wuhan, China).

Cell viability assay by Calcein-AM/PI double staining

Cell viability was assessed using the Calcein-AM/PI double-staining assay [67]. Briefly, SH-SY5Y cells were pre-treated with 20 μM Aβ1-42 for 48 h and then co-cultured with different treatment reagents for 24 h. Then, the cells were collected by centrifugation at 1000 × g for 5 min in a 24-well plate. The detection working solution was prepared by mixing 1 μL of Calcein-AM and 1 μL of PI into 1 mL of assay buffer. Then, 250 μL of the working solution was added to each well, followed by incubation at 37 °C in the dark for 30 min. Then, cells were observed and photographed under a fluorescence microscope using the green and red channels, respectively. Cell survival rate was assessed by counting the green (live cells) and red (dead cells) fluorescent signals.

Assessment of autophagic flux by tandem fluorescent LC3b reporter

To monitor autophagic flux dynamics, the tandem fluorescent-tagged autophagy reporter plasmid pmCherry-EGFP-LC3B (Research Cloud Biology, Jinan, China) was transfected into SH-SY5Y cells using Lipofectamine 3000 transfection reagent according to the manufacturer's instructions. Following transfection, the punctate accumulation of yellow fluorescence (representing autophagosomes) and red fluorescence (representing autolysosomes) was observed under a fluorescence microscope to assess the progression of autophagic flux. The fluorescent puncta within cells were counted and quantitatively analyzed using ImageJ software (National Institutes of Health, USA).

Assessment of Aβ clearance in SH-SY5Y cells

To evaluate the Aβ clearance capacity of SH-SY5Y cells, fluorescein-labeled 647-Aβ1-42 was used for tracing analysis [28]. The specific steps were as follows: SH-SY5Y cells were co-incubated with 2.5 μM Aβ1-42 and EXOs from different treatment groups (EXOs or RVG-EXOs) for 24 h. Subsequently, the cells were washed three times with PBS, and the medium was replaced with fresh medium containing 2 μg/mL (approximately 0.4 μM) 647-Aβ1-42 for an additional 3-h incubation. Then, the medium containing the fluorescently labeled Aβ was removed and replaced with fresh DMEM, followed by a 24-h culture to simulate the intracellular degradation and clearance of Aβ. Finally, the medium was discarded, the cells were washed three times with PBS, and subsequent cell staining and fluorescence imaging analysis were performed. The fluorescence signal intensity was quantified using ImageJ software (National Institutes of Health, USA) to assess the level of residual intracellular Aβ.

Assessment of P-Tau clearance in SH-SY5Y cells

To evaluate the clearance effect of EXOs on P-Tau, cells were seeded in 6-well plates at a density of 2 × 105 cells per well and allowed to adhere overnight. To induce Tau protein hyperphosphorylation, cells were first treated with 40 nM OA for 24 h. Subsequently, the OA-containing medium was replaced with fresh medium containing either EXOs or RVG-EXOs for a 24-h co-culture. The untreated cells were used as a negative control, OA-treated cells alone as a positive control, and experimental groups contained OA-treated cells co-cultured with either EXOs or RVG-EXOs. After the co-culture period, the levels of the corresponding P-Tau (Ser396) were detected by Western blotting analysis.

Dual-luciferase reporter assay

The interaction of hsa-miR-23a-3p and hsa-miR-23b-3p with the 3′ UTR of RPTOR was predicted using the TargetScan database (https://www.targetscan.org/vert_80/). Based on the predicted binding site, WT and seed region-mutated (MUT) fragments of RPTOR were cloned into the pmirGLO vector (GenePharma, China). HEK-293T cells were seeded in 24-well plates and transfected with 500 ng of RPTOR-WT or RPTOR-MUT reporter plasmid, along with 50 nM hsa-miR-23a-3p mimic, hsa-miR-23b-3p mimic, or negative control mimic (GenePharma, China) using Lipofectamine 3000. After 48 h, luciferase activity was measured using the Dual-Luciferase® Reporter Assay System (11402ES60, YEASEN, China). Firefly luciferase activity was normalized to Renilla luciferase activity. All experiments were performed in triplicate, and results were presented as the mean ± standard deviation (SD).

RNA isolation and quantitative real-time PCR

Total RNA from cells and tissues was extracted using TRIzol reagent (AG21101, Accurate Biotechnology, China) and exosomal RNA was isolated with the ExoRNeasy Serum/Plasma Maxi Kit (Qiagen, Frankfurt, Germany). cDNA synthesis was performed using Evo M-MLV RT Premix (AG11706, Accurate Biotechnology, China) for mRNA and a miRNA First Strand cDNA Kit (Stem-loop) (AG11742, Accurate Biotechnology, China) for miRNA. Quantitative real-time PCR (qRT-PCR) was carried out using the SYBR® Green Pro Taq HS Premixed qPCR Kit (AG11701, Accurate Biotechnology, China) with gene-specific primers. GAPDH was used as internal control for mRNA normalization; U6 and cel-miR-39-3p were used as internal and external controls for miRNA normalization, respectively. Relative expression was calculated using the 2–ΔΔCT method. Primer sequences were listed in Table S1.

Cell transfection

For cell transfection experiments, 50 pmol of miRNA inhibitor (GenePharma, China) was transfected into SH-SY5Y cell line using Lipofectamine 3000 (L3000015, Thermo Fisher Scientific, USA) following the manufacturer's instructions [68]. Total RNA and protein were extracted 48 h after transfection. The sequence of has-miR-23a-3p inhibitor was 5′-GGAAAUCCCUGGCAAUGUGAU-3′, and the sequence of has-miR-23b-3p inhibitor was 5′-GUGGUAAUCCCUGGCAAUGUGAU-3’.

Enzyme-linked immunosorbent assay (ELISA)

The concentrations of interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), and immunoglobulin G (IgG) in mouse samples were measured using commercial ELISA kits according to the manufacturer's instructions. Plasma samples were collected and stored at −80 °C until analysis. For tissue samples, brain and liver tissues were homogenized in ice-cold PBS containing protease inhibitors, followed by centrifugation to collect the supernatants for protein quantification. The following kits from Boster Biological Technology Co., Ltd. (Pleasanton, USA) were utilized: Mouse IL-6 ELISA Kit (catalog number: EK0411), Mouse IL-1 beta/IL1B ELISA Kit (Cat. No.: EK0394), Mouse TNF Alpha/TNFA ELISA Kit (Cat. No.: EK0527), and Mouse IgG ELISA Kit (Cat. No.: EK0101). All assays were performed with three replicates. The optical density was measured at 450 nm using a microplate reader. The concentrations of the target proteins were calculated by comparing the sample optical densities with the standard curves.

Small RNA sequencing and analysis

Samples were obtained from six randomly selected elderly donors (aged ≥ 75 years), six donors with AD, and six young donors (aged ≤ 30 years). Total RNA was extracted from isolated exosomes, and sequencing libraries were prepared using the Multiplex Small RNA Library Prep Set for Illumina® (San Diego, CA, USA) following the manufacturer's protocol. Sequencing was conducted by Novogene Bioinformatics Technology Co., Ltd. (Beijing, China). Clean reads were filtered based on length distribution to enrich for small RNAs (sRNAs). Reads were aligned to the human reference genome (GRCh38) using Bowtie. Known miRNAs were annotated by alignment to miR Base v20.0, and novel miRNAs were predicted using a combination of miR Evo and miRDeep2. Differential expression analysis between the elderly and young groups was performed using the DESeq R package, and miRNAs with a log2 (fold change) > 1 and adjusted p–value (FDR) < 0.05 were considered significantly differentially expressed.

Clinical and demographic characterization of plasma donors

Samples were obtained from six randomly selected elderly donors (aged ≥ 75 years), six donors with AD, and six young donors (aged ≤ 30 years). Demographic details of the participants were provided in Table S2.

Peripheral blood was collected from each participant and placed into EDTA-coated anticoagulant tubes and processed within 4 h of collection. Plasma was separated by centrifugation at 1150 × g for 15 min, aliquoted into 500 μL portions, and stored at −80 °C until analysis. For Alzheimer's disease-related biomarker measurement, frozen plasma samples were thawed at 4 °C and then centrifuged at 10,000 × g for 10 min to remove residual cell debris. The resulting supernatant was then divided into 75 μL aliquots for quantification of Alzheimer's disease-related biomarkers using a single-molecule detection system (AXL-2000, Lychix, Suzhou, China).

This study was approved by the Ethics Committee of the Provincial Hospital affiliated with Shandong First Medical University for Biomedical Research Involving Human Subjects (NSFC No. 2022–511). All procedures were conducted in accordance with the ethical principles of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrollment.

Single-nucleus RNA sequencing (snRNA-seq)

To investigate the transcriptional landscape of brain tissues following RVG-EXO treatment, 8-month-old 3xTg mice were intravenously injected (tail vein) with either RVG-EXOs (1.8 μg/μL, 100 μL per injection, three times per week for 4 weeks) or an equal volume of PBS (vehicle control) according to the treatment protocol described in Plan A (n = 3 mice per group). At the end of the treatment period, mice were euthanized, and whole brains were rapidly dissected. The hippocampal and cortical regions were isolated, flash-frozen in liquid nitrogen, and stored at −80 °C until further processing.

From fresh frozen mouse brain tissue, approximately 500 mg of tissue was cut and weighed on dry ice using sterile disposable scalpels. The brain tissue was homogenized in ice-cold homogenization buffer (0.25 M sucrose, 5 mM CaCl2, 3 mM MgAc2, 10 mM Tris-HCl pH 8.0, 0.1 mM EDTA, 1×protease inhibitor, and 1 U/μL Ribolock RNase inhibitors) with a glass-on-glass Dounce homogenizer: 10 strokes with the A pestle, followed by 10 strokes with the B pestle. Homogenates were passed through a 70-μm cell strainer to collect the nuclear fraction. The nuclear fraction was mixed with an equal volume of 50% iodixanol and added on top of a 30–33% iodixanol solution, then centrifuged for 20 min at 10,000×g, 4 °C. After removal of the myelin layer from the top of the gradient, the nuclei were collected from the 30–33% iodixanol interface. The nuclei were resuspended in nuclei wash and resuspension buffer (0.04% bovine serum albumin, 0.2 U/μL Ribolock RNase inhibitors, 500 mM mannitol, and 0.1 mM PMSF protease inhibitor in phosphate-buffered saline) and pelleted for 5 min at 500×g, 4 °C. The nuclei were passed through a 40-μm cell strainer to remove cell debris and large clumps. Nucleus concentration was manually determined using trypan blue counterstaining and a hemocytometer. The concentration was adjusted to 700–1200 nuclei/μL, and the nuclei were immediately processed following the 10x Genomics® Single Cell Protocol.

Single-nucleus libraries were constructed using the DNBelab C Series High-throughput Single-cell RNA Library Preparation Set V3.0 (TaiM 4, MGI, Shenzhen, China) following the manufacturer's protocol. Briefly, the nuclei suspension was loaded onto a DNBelab C4 single-cell instrument for droplet generation and barcoding. After reverse transcription, cDNA was recovered and amplified by PCR. The amplified cDNA was then used to construct both a cDNA library and an oligo library. The libraries were sequenced on a DNBSEQ-T7RS platform (MGI, Shenzhen, China) in paired-end 150 bp (PE150) mode.

Raw sequencing data were processed using dnbc4tools (version 2.1.0) for quality control, alignment to the mouse reference genome (mm10), and gene expression quantification. The resulting cell-by-gene matrix was analyzed using the Seurat package (version 4.3.0). Cells with low-quality metrics (e.g., high mitochondrial gene percentage, abnormal UMI or gene counts) and potential doublets were filtered out. Data were normalized and scaled, followed by principal component analysis (PCA), batch effect correction with Harmony, and cell clustering using the Louvain algorithm. Clusters were visualized by t-SNE and annotated using the SingleR R package. The snRNA-seq experiment and data analysis were performed by Genedenovo Biotechnology Co., Ltd. (Guangzhou, China).

Detection of APP and CD74 protein in vitro and in vivo

In cell experiments, human SH-SY5Y neurons were co-cultured with HMC3 microglia, and the following four treatment groups were established: (1) Negative control group, treated with PBS only; (2) Cu2+ control group, treated with 1 μM Cu2+ for 3 days (a condition previously confirmed to significantly upregulate APP protein levels in SH-SY5Y cells [69]), followed by incubation with PBS for 24 h; (3) RVG-EXOs alone group, treated with Cu2+ for 3 days, then incubated with 50 μg RVG-EXOs (per 1 × 106 cells) for 24 h; (4) Combination treatment group, treated with Cu2+ for 3 days, then co-incubated with 50 μg RVG-EXOs and 5 mM 3-methyladenine (3-MA, a well-known autophagy inhibitor) for 24 h (this concentration and treatment duration of 3-MA have been shown to suppress cellular autophagy [70]). PBS was used as the solvent control in all cell experiments.

In the in vivo study, 8-month-old 3xTg mice were injected via the tail vein with 100 μL of RVG-EXOs at a concentration of 1.8 μg/μL, three times per week for two weeks. In addition, another group of mice received both the above tail-vein injections of RVG-EXOs and intraperitoneal injections of 3-MA at a dose of 30 mg/kg (a dose that has been shown to significantly suppress autophagy activity in the mouse brain [71]) at the same frequency. Control mice were injected with an equal volume of PBS.

After the respective treatments, cells and brain tissues were either fixed or lysed, and then processed for immunofluorescence staining and Western blot analysis following standard protocols. The reagents used included APP (E4H1U) Rabbit Monoclonal Antibody (76600, Cell Signaling Technology, Danvers, USA); CD74 Antibody (MU601117, Abmart, Shanghai, China); and Alexa Fluor® 488 anti-mouse CD74 (CLIP) Antibody (151005, BioLegend, San Diego, USA); Goat Anti-Rabbit IgG H&L (Alexa Fluor® 555) preadsorbed (ab150086, abcam, Waltham, USA); Goat Anti-Mouse IgG H&L (Alexa Fluor® 488) (ab150113, abcam, Waltham, USA); 3-Methyladenine (HY-19312, MedChemExpress LLC, New Jersey, USA).

Detection of excitatory neurons, inhibitory neurons, and DAM

For in vivo immunofluorescence assessment of neuronal and microglial subtypes, 8-month-old 3xTg mice received tail-vein injections of 100 μL RVG-EXOs (1.8 μg/μL) or an equivalent volume of PBS three times per week for two consecutive weeks. Subsequently, brain tissues were harvested and processed for immunofluorescence staining. Excitatory neurons were identified as NeuN+CaMK2a+, inhibitory neurons as NeuN+GAD1+, and DAM as IBA1+CLEC7A+. The primary antibodies used for this detection included CaMK2 alpha Polyclonal antibody (13730-1-AP, Proteintech, Wuhan, China), GAD1 Monoclonal antibody (67648-1-Ig, Proteintech, Wuhan, China), Dectin-1/Clec7a (E3P5W) Rabbit Monoclonal Antibody (30260, Cell Signaling Technology, Danvers, USA), Iba1 Mouse IgG1 mAb (OB-MMS039-02, Oasis biofarm, Zhejiang China), Goat Anti-Rabbit IgG H&L (Alexa Fluor® 555) preadsorbed (ab150086, abcam, Waltham, USA); Goat Anti-Mouse IgG H&L (Alexa Fluor® 488) (ab150113, abcam, Waltham, USA), Goat Anti-Rabbit IgG H&L (Alexa Fluor® 488) preadsorbed (ab150081, abcam, Waltham, USA), Goat Anti-Mouse IgG H&L (Alexa Fluor® 555) preadsorbed (ab150118, abcam, Waltham, USA), Anti-NeuN antibody (ab104224, abcam, Waltham, USA), Anti-NeuN antibody (ab177487, abcam, Waltham, USA). All stained sections were examined under a fluorescence microscope, and images were captured for quantitative analysis.

Statistical analyses

Bioinformatic analysis of single-cell RNA-sequencing data was performed using OmicsMaster, a dynamic real-time interactive online platform (https://report.omicsmaster.com). All other statistical analyses were performed using GraphPad Prism version 10.2 (RRID: SCR_002798). Data were presented as mean ± standard deviation (SD) unless otherwise specified. Comparative analyses between two groups were conducted to reveal the significant differences using unpaired two-tailed Student's t-tests. Comparative analyses among multiple groups were performed using one-way or two-way ANOVA followed by appropriate post hoc tests. Statistically significant difference was determined based on P < 0.05.