Exosomal MicroRNAs modulate the cognitive function in fasudil treated APPswe/PSEN1dE9 transgenic (APP/PS1) mice model of Alzheimer's disease

Alzheimer’s disease (AD) is characterized by cognitive impairment caused by the accumulation of beta-amyloid (A β ) plaques and trans-synaptic spread of tau pathology. Exosome has emerged as key mediators for neuronal development, maintenance, and cellular communication. However, the molecular mechanism of exosomal miRNAs related to AD remains unknown. In the present study, APPswe/PSEN1dE9 transgenic (APP/PS1) mice (AD) were treated with vehicle (ADNS) and fasudil (ADF), whereas C57BL/6 (control) mice were treated with vehicle (WT). Cognitive function was assessed by Y-maze test and AD pathology was con�rmed by immunostaining of A β plaque and phosphorylated tau. Exosomal RNAs from each mouse of the group were extracted, sequenced and analyzed. Our results showed amelioration of cognitive function, decreased A β plaque load, and phosphorylated tau protein after fasudil treatment. Exosomal miRNA analysis showed 3 miRNAs (mmu-let-7i-5p, mmu-miR-19a-3p, mmu-miR-451a) in the intersection of ADNS vs ADF and WT vs ADNS. GO annotation and KEGG pathway showed the target gene of miR-19a-3p are Pten and Tnf; mmu-miR-451a are Nsmaf, Gnai3 and Akt3. In conclusions, fasudil treatment improved cognitive function by regulating exosomal MicroRNAs (mmu-miR-451a and mmu-miR-19a-3p). These MicroRNAs could be potential biomarker of AD and therapeutic target for novel treatment for AD.


Introduction
Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by cognitive impairment and marked accumulation of beta-amyloid (Aβ) plaques and hyperphosphorylated tau tangles.Amyloid Beta (Aβ) and hyperphosphorylated tau (p tau) are hallmark pathological biomarkers of AD 1 .Extracellular vesicles (EVs) have been explored in diagnosis and therapeutic guidance in AD.EVs can cross the blood-brain barrier and served as a potential source for AD biomarkers.Several studies suggest that EVs circulate misfolded proteins involved in AD [2][3][4] .EVs include exosomes, microvesicles and apoptotic bodies that are known to participate in intercellular communication 5 .Exosomes have been given attention due to its role as pathological and therapeutic cargo (Haiyang Yu et al., 2020).Exosome vesicles are approximately 30-100 nm in diameter with a lipid bilayer membrane structure, released upon the exocytosis of a multivesicular body.Distant intercellular communication occurs through RNA signals triggered by pathogenic materials of AD during development and progression 6,7 .Pathogenic materials such as proteins, micro-RNAs, and messenger RNAs, are invaginated in exosome and released 8 .MicroRNAs (miRNAs) are a family of small non-coding RNA molecules reported to be involved in pathogenesis of neurodegenerative diseases and CNS injuries 9 .Exosomes carry these MicroRNAs (miRNAs) for cell-cell communication in neurons 10 .Exosomes transport miRNA, mRNA, and transcription factors (TFs) to neighboring cells and in uence their cellular function 11 .Exosomes represents several common proteins of cytotypes such as CD63, CD81, CD9, TSG-101, ALIX, and HSP70 and these markers aid in the identi cation of cell type 12 .Exosome-derived MicroRNAs (miRNAs) and long non-coding RNAs (lncRNAs) are involved in the pathophysiology of cancers and neurodegenerative diseases 13 .Exosomal miRNAs (ex-miRNAs) are highly stable, resistant to degradation and served as reliable biomarkers for the early clinical diagnosis of AD 14 .
Exosomes in blood are a novel source of biomarkers for the diagnosis and prognosis of diseases 15 .Aβ deposition and in ammation pathway have been implicated in AD pathology and progression.
In ammatory cytokines are released from exosomes after being triggered by sensitive cells 16 .A recent study showed increased level of cytokines in the mice, which was immunized with pooled exosomes isolated from the sera of lung transplant recipients (LTxR) with bronchiolitis obliterans syndrome (BOS) 17 .In ammatory markers are non-speci c due to pooling of samples during analysis, however, analysis of exosomal miRNA offers high speci city and sensitivity to predict the biomarkers of AD 18 .Exosomes carrying proteins, RNA, and DNA are derived from their originating cells and protected from degradation in the circulation 19 .Mesenchymal stem cells (MSCs) -derived exosomes contain high levels of miR-133b and miRNAs in the miR-17 ~ 92 cluster and represses the expression of PTEN, CTGF and RhoA to remodel the neurites and improve the functional recovery in a mouse model of stroke 20 .Exosomes isolated from pooled serum samples of patients and healthy controls have been used to analyze the protein pro les 21 .
Exosomes have been shown to establish communication between cells and exchanges genetic materials 22 such as mRNA, miRNA, transfer RNAs and long non-coding RNAs (lncRNAs) 23 to modulate the physiological processes such as immune response, in ammatory response, angiogenesis, apoptosis, blood coagulation, and cell debris disposal 24 .Cell type speci city of exosomes with RNA and protein components has been proposed as potential early diagnostic markers for several diseases and offers a targeted carrier of drugs for treatment 25 .
In the present study, we aimed to show the distinct expression pro les of miRNAs in exosomes derived from serum of fasudil-treated APP/PS1 mice to demonstrate the novel therapeutic strategies of exosomal miRNA-based diagnosis and treatment of AD.

Y-Maze spontaneous alternation test showed fasudil protects memory ability of AD mice
Y maze test measures the dynamic spatial working memory and willingness of rodents to explore new environments.The animals tend to explore a new arm of the maze rather than returning to one that was previously visited.The total number of entries, time and distance in new arm indicate the exploratory behavior.Alternation rate (%) re ects the memory retention.ADNS showed signi cantly lower alteration rate as compared to WT, and ADF restored to ADNS level (p < 0.01), suggesting that ADF treatment signi cantly improved the memory de cit (Fig. 1A).The ratio of number of novel arm entry to all arms entries did not show signi cant difference between all the three groups (Fig. 1B).The time spent in novel arm to all arms was signi cantly lower in ADNS group as compared to WT group (p < 0.05), and increase in ADF, but the difference was not signi cant (Fig. 1C).Overall, ADNS group had signi cantly less exploration in the novel environment as compared to WT (p < 0.05), and ADF improved their exploratory behavior.The ratio of distance traveled in the novel arm to all arms in ADNS group was lower as compared to WT group and increased in ADF as compared to ADNS but did not show statistically signi cant difference (Fig. 1D).

Fasudil treatment altered the expression of Aβ and p-Tau in the hippocampus of AD mice
Expression of Aβ and p-Tau in hippocampus of the three groups (WT, ADNS, ADF) was performed by immunostaining.Aβ 1−42 plaques in the hippocampus were signi cantly increased in ADNS group (776.49) and signi cantly decreased in ADF (459.39)(p < 0.01), suggesting reduction of Aβ burden after fasudil treatment (Fig. 2A).Tau phosphorylation is a key factor for AD pathogenesis 26 .Immuno uorescent intensity quanti cation of phospho-Tau (S404) in hippocampus was signi cantly increased in ADNS as compared to WT (4104.15)and decreased after fasudil treatment (3109.68)(p < 0.01), suggesting a protective effect of fasudil by decreasing Tau formation in AD (Fig. 2B).

Exsomes extraction and identi cation in serum of mice
Exosomes from serum of each mouse from the group were isolated using exosome Isolation and Puri cation Kit (from plasma or serum) (Umibio, China) following manufacturer's recommendations.NTA (Nanoparticle Tracking Analysis, NTA) measurements were recorded and analyzed at 11 positions.The ZetaView system was calibrated using 110 nm polystyrene particles and the temperature was maintained between 23°C and 30°C.
Transmission electron microscopy represent close-up images of single exosome.Nanoparticle tracking analysis provides an overview of size distribution and concentration of isolated exosomes.We observed a population of nanovesicles with a predominant size of 97 nm (Fig. 3A) which is commonly predicted size of the exosome.A cup-shaped morphology of exosomes was also observed (Fig. 3B).Western blotting analysis con rmed that the isolated particles expressed characteristics of exosome markers of CD63 and TSG101 (Fig. 3C).

MicroRNA Pro ling establishment
Exosomes of 3 mice were pooled together for sequencing as WT1 and another three mice were pooled as WT2.Similarly, we followed for ADNS1, ADNS2; ADF1 and ADF2.An unbiased high-throughput sequencing of serum exosome miRNAs was performed to capture the complete pro le of miRNAs.RNA extraction from each sample yielded the typical RNA pro le for exosomes, without ribosomal RNA and enriched small RNA species (< 200 nt).The expression pro les of exosomal miRNAs were determined by deep sequencing and analyzed.
A total of 1924 miRNAs were screened in each sample.Each sample achieved miRNA read counts > 10,000,000 and the error rate was around 0.09%-0.14%.Read counts were normalized to adjust the RNA and sample-level biases.Poor-quality read was detected if the number of bases ( < = 20nt) in reads exceeds 50% of the read length and the content of N in reading exceeds 10% of the length of the read.Reads with abnormal nal length were removed (supplementary table 1).
Small RNA (sRNA) in the 19-22 nt length range were screened from clean reads of each sample for comparison and analyzed using reference genome.These small RNA were obtained according to the priority order such as miRNA > rRNA > tRNA > snRNA > snoRNA > repeat > novel miRNA > exon > intron.
Priority order of detection results were annotated for small RNAs (Fig. 4, supplementary table 2)

Serum exosome MicroRNA target genes and pathway analysis
We used MiRanda[6] software and qTar software to predict and analyze target genes for these known miRNA and novel miRNA targeting mmu-let-7i-5p, mmu-miR-19a-3p and mmu-miR-451a (supplementary table 4).

Serum exosome MicroRNA target genes (GO)
Biological functions of differentially expressed miRNAs target genes in WT vs ADNS or ADNS vs ADF were analyzed by gene ontology (GO) enrichment analysis.GO terms showed important biological processes such as cellular process, single-organism process, biological regulation, regulation of biological process; the most important cellular components involved the cell, cell part, organelle, membrane; and the molecular functions includes binding, catalytic activity, nucleic acid binding transcription factor activity, signal and molecular transducer activity (Fig. 5D and 5E, supplementary table 5).

Serum exosome MicroRNA target genes (KEGG)
Differentially expressed miRNAs target genes between WT vs ADNS involved 181 pathways.Signal transduction includes AMPK, Jak-STAT, mTOR, PI3K-Akt, FoxO, MAPK, Ras, TNF and cGMP-PKG signaling pathway.Aging includes longevity regulating pathways in the nervous system such as neurotrophin signaling pathway and Cholinergic synapse.The endocrine system includes estrogen and prolactin signaling pathway.Other pathways are autophagy, axon guidance, cytokine-cytokine receptor interaction and signaling pathways regulating pluripotency of stem cells (Fig. 6, supplementary table 6).
Differentially expressed miRNAs target genes between ADNS vs ADF showed 195 pathways.Signal transduction includes sphingolipid, cGMP -PKG, FoxO, AMPK, TNF and MAPK signaling pathway.The immune system includes Toll-like receptor, NOD-like receptor and RIG-I-like receptor signaling pathway.Developmental includes osteoclast differentiation and axon guidance.The nervous system includes dopaminergic, glutamatergic and cholinergic synapse.Other pathways were neuroactive ligand-receptor interaction, adrenergic signaling in cardiomyocytes, signaling pathways regulating pluripotency of stem cells, fatty acid metabolism, longevity regulating pathway and apoptosis (Fig. 6, supplementary table 6).

Exosomal MicroRNA target genes (KEGG) in the intersection of ADNS vs ADF and WT vs ADNS
mmu-miR-19a-3p and mmu-miR-451a were in the intersection of ADNS vs ADF and WT vs ADNS and were signi cantly higher in ADNS as compared to WT and ADF.
The phosphatase and tensin homolog (Pten, ENSMUSG00000013663) is one of the target genes of mmu-miR-19a-3p which is associated with the sphingolipid and FoxO signaling pathway.Another target gene of mmu-miR-19a-3p is tumor necrosis factor (Tnf, ENSMUSG00000024401) which is involved in signal transduction pathways such as sphingolipid, TNF, immune system and Toll-like receptor signaling pathways.

Discussion
Our previous study showed that fasudil ameliorates cognitive function in the Morris water maze (MWM) behavior test 27 .The present study also showed amelioration of cognitive function in Y maze test after fasudil treatment in APP/PS1 mice.We believe that fasudil induced cognitive improvement could be due to the manipulation of serum exosomal miRNAs, and mmu-miR-19a-3p and mmu-miR-451a, which could be potential target biomarkers of fasudil treatment in APP/PS1mice.In our previous and recent studies 27,28 , we discovered that AD progression is associated with the alteration of gut microbiota and metabolites and impacts the cognitive function.There are several cellular communications in AD.One of them is exosomes that carry proteins, RNAs and DNAs and transport them from cell to cell.Our previous studies have shown the relationship between intestinal microbiota and AD through the microbiota -gut-brain axis.Intestinal microbiota and exosome miRNA could serve as new targets for AD therapeutic intervention.Exosomes are practical for long-term storage and easily pass through the blood-brain barrier (BBB) while protecting their molecules wrapped in their bilayer lipid structure 29 .We believe that brain-gut axis results in the transportation of biochemical molecules between the brain and gut via exosomes 30 .Therefore, intestinal ora and exosomes carrying biochemical products could trigger the neurodegenerative diseases through the brain-gut axis, which may be a potential mechanism for AD onset and a target for drug therapy 31 .Based on the metagenome and exosome results, we will continue our study of brain-gut axis pathway and fasudil treatment in AD to reveal the novel pathophysiology of AD and systematically elaborate the mechanism of fasudil treated AD (Fig. 7).
3.1 mmu-miR-19a-3p miRNA alterations in CNS and PNS in uence microglial activity and neuroin ammation by orchestrating the expression of proteins, which controls activation or inhibition of microglial activity.A recent clinical study reports that hsa-miR-19a-3p and hsa-miR-19b-3p expression distinctly discriminative the pain severity in spinal cord injury-induced neuropathic pain 32 .Increased expression of miR-19a-3p in the exosomes of α-synuclein gene transgenic SH-SY5Y cells has been reported, while enhanced expression of miR-19a-3p in exosomes suppresses autophagy in recipient microglia by targeting the phosphatase and tensin homolog/AKT/mTOR signaling pathway 33 .In multiple systems atrophy (MSA), miRNA expression and corresponding gene target suggests that miR-19a-3p acts as key regulators of neuroin ammation 34 .A similar trend was observed in our study, where mmu-miR-19a-3p expression was signi cantly decreased after fasudil treatment.Therefore, we speculate a potential link between AD and MSA pathologies that involves miRNAs and deregulation of BACE1.
Pten is the target gene of mmu-miR-19a-3p in the sphingolipid and FoxO signaling pathways.Pten is a tumor suppressor gene regulating axonal growth in the adult central nervous system 35 .Pten deletion enhances neurite outgrowth during neural stem cell differentiation 36 .Phosphatidylinositol-4,5bisphosphate 3-kinase (PI3K)/protein kinase B (PKB or Akt) pathway is deregulated in response to phosphatase and tensin homolog (Pten) overexpression in AD by orchestrating in ammation and oxidative stress.AD patient's serum has shown an increased level of amyloid β (Aβ) 42, p-tau as well as PTEN 37 .
Tnf was also observed as target gene of mmu-miR-19a-3p in signal transduction (sphingolipid and TNF signaling pathway, and immune system pathways (Toll-like receptor signaling pathway).Elevated tumor necrosis factor alpha (TNF-α), interferon gamma (IFN-y), interleukin 6 (IL-6) and IL-10 in AD patient has been compared with healthy controls 38 .Recent studies suggest that inhibition of neuroin ammation is a potential therapeutic or preventive strategy 39,40 .

mmu-miR-451a
miRNA451a expression in AD patients has been reported to increase signi cantly as compared to healthy control and positively correlates with Aβ-42/Aβ-40 ratio and tau protein 41,42 .We also observed signi cantly increased expression of miRNA451a in AD mice and restoration to normal level after fasudil treatment.
Sphingolipid metabolism has been associated with amyloid-beta production and AD neuropathology 43 .Nsmaf is a target gene of mmu-miR-451a in the sphingolipid signaling pathway and is reported to regulate AD patient's blood 44 .
Gnai3 is a target gene of mmu-miR-451a in the sphingolipid signaling pathway, glutamatergic synapse, dopaminergic synapse, cholinergic synapse and cGMP -PKG signaling pathway.Gnai3 assist in guanyl nucleotide binding, metal ion binding, nucleotide binding, protein binding and protein domain speci c binding.Gnai3 is associated with lipid metabolism in major depression 45 and AD pathogenesis in the mouse model 46 .cGMP-PKG signaling pathway in the hippocampus of AD mice has been shown to dysregulates circRNAs 47,48 .mmu-miR-298-3p/Smoc2 signaling axis may regulate the pathophysiology of AD by affecting the cGMP-PKG signaling pathway 49 .
Akt3 is also the target gene of mmu-miR-451a in the sphingolipid, Toll-like receptor, FoxO and TNF signaling pathway along with dopaminergic synapse, signaling pathways regulating pluripotency of stem cells.Akt3 is the predominant isoform of Akt expressed in the hippocampus and is primarily affected during AD progression.Mitochondrial dysfunction and AD like pathology have been reported in Akt3-null mice 50 .Akt3 modulates angiogenesis and orchestrate mitochondrial dynamics in the vascular endothelium by controlling autophagy and biogenesis through subcellular localization of the master regulator of nuclear mitochondrial gene expression, PGC-1α 51 .Akt3 is also required for the nuclear export receptor, CRM-1 52 .MiR-485-3p serves as a biomarker and therapeutic target of AD via regulating neuronal cell viability and neuroin ammation by targeting Akt3 32 .
Myc is the target gene of mmu-miR-451a in the signaling pathways regulating pluripotency of stem cells.Myc repair brain cells in neurodegenerative disease or CNS trauma, including stroke and traumatic brain and spinal cord injury and promote axonal growth and regeneration.It also regulates the cell cycle, metabolism, cell enhance the synaptic structure to restore cognitive function 52 .
Pten and Akt3 shares a common FoxO signaling pathway.FoxO transcription factors controls the proin ammatory pathways, affecting nervous system amyloid (Aβ) production and toxicity, leading to mitochondrial dysfunction, fostering neuronal apoptotic cell death, and accelerating the progression of degenerative disease 53 .However, FoxOs also offers a protective effect on the nervous system by promoting autophagy, reducing toxic intracellular protein accumulations and potentially limiting Aβ toxicity 54 .FoxO proteins may provide a potential molecular target for the treatment of AD 55 .

Conclusion
Fasudil treatment improved cognitive function by regulating exosomal MicroRNAs (mmu-miR-451a and mmu-miR-19a-3p).These MicroRNAs could be potential biomarker of AD and a potential target for novel treatment for AD.Our study provides a new strategy to elucidate the pathology of AD by targeting the signaling pathways such as signal transduction, immune system, signaling molecules and interaction, cellular community and nervous system pathways and pathways regulating pluripotency of stem cells.Currently, a major challenge in AD drug development is lack of precise mechanisms in AD pathogenesis.The new strategy of AD drug design employs a multi-targets target approach and focuses on preventive strategies at the early stages of disease development.Therefore, our novel strategies to target exosomal MicroRNAs have great potential in diagnosis, treatment and new AD drug development.

Animals and treatment
Animal Ethics Committee of Shanxi Datong University approved all procedures.All the experiments were performed in compliance with the guidelines and regulations of the Administration O ce of the International Council for Laboratory Animal Science and ARRIVE guidelines.Male APP /PS1 mice (8 months old) expressing human amyloid precursor protein (HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9) were purchased from Beijing Huafukang Bioscience CO., LTD (HFK, Beijing, China).All animals were housed in the pathogen-free facilities at the Institute of Brain Science, Shanxi Datong University and maintained at constant room temperature (25 ± 2 0 C) and humidity (50 ± 5%) in a 12-h light/12-h dark cycle.APP/PS1 mice were pre-screened based on the normal physiological behavior and randomly divided into two treatment groups: 1) vehicle-treated mice (ADNS); received normal saline (volume was adjusted similar to fasudil treatment); 2) fasudil-treated (ADF) mice; received a daily injection of fasudil (Tianjin Chase Sun Pharmaceutical Co., Ltd.), 25 mg/kg/day, i.p., 16 weeks).Aged and gender-matched C57BL/6 (WT) mice were used as normal controls, which received saline in the same volume (n = 6 per group).Animals had ad libitum access to food and water.All the experiments were performed in the compliance with the guidelines and regulations of the Administration O ce of the International Council for Laboratory Animal Science.The experimental protocols were approved by the Animal Ethics Committee of Shanxi Datong University, Datong, China.

Y Maze alternating behavior test
Y maze test measures the spatial working memory of animals.It consists of two stages and performed at 1h intervals.The rst stage is training period and the mouse was given limited food and water in one arm of Y maze, while other arm served as an experimental with reward.The new arm (left arm) was blocked by a barrier.Mice was kept at the starting point of the arm and allowed to move freely for 10 minutes.After the training, the mice were put back into the home cage.The second stage was the acquisition stage, where food or water was removed from experimental arm.The new arm barrier (left arm) was removed, and the mice were placed at the starting point and allowed to move freely for 5 min.Total time spent and the number of entries in each arm were recorded by a video tracking system (SMART V3.0 system, Panlab, Barcelona, Spain)).The apparatus was cleaned with ethanol at the end of each training or test to prevent the interference due to residual 56 .

Immuno uorescence staining with confocal microscopy for Aβ plaques and phospho-Tau
Mice were sacri ced at the end of treatment.The hippocampus tissues were collected from all groups and were xed in 4% cold paraformaldehyde for 30 min at room temperature.The sections were washed, permeabilized with 0.1% Triton X-100 for 15 min, rinsed with PBS and blocked with 1% bovine serum albumin (BSA) for 30 min at room temperature.Sections were subsequently incubated with anti-Tau (phospho S404) (dilution of 1:200, abcam, ab92676, USA) with mouse monoclonal primary antibodies, anti-Aβ1-42 (dilution of 1:500, Servicebio, GB111197, China) with rabbit monoclonal primary antibodies at 4ºC overnight.Slides were washed 3 times with PBS and incubated with FITC or Cy3-conjugated secondary antibodies (Invitrogen, USA) for 1hr.at RT, and then thoroughly washed with PBS for 3 times.
The negative control sections were treated using the same protocols without primary antibodies.The average values of six sections from each slice were analyzed for the nal analysis.Three samples were analyzed in each group.The modi ed coverslips were mounted onto glass slides and observed with confocal microscopy (Olympus FV1000, Japan).Images were taken at 10x (scale = 200 µm).The area (polygon) of positive cells were quantitatively analyzed by Image-Pro Plus software.All images taken for quanti cation were blinded during analysis.

Methods for extraction and puri cation of exosomes (serum)
Serum was diluted with precooled 1x PBS.Blood PureExo Solution (BPS) was added and kept it for 2 h at 2-8℃.The supernatant was removed after centrifugation at 10000 g centrifugal, 60 min at 4 ℃.Exsomes rich supernatant was precipitated by adding PBS (200µL, 1×x).Beat centrifugal sedimentation was performed at 12000 g for 2 min at 4 ℃ and supernatant rich in exosome particles were separated.
Harvested exosome particles were transferred into the upper chamber of the exosome puri cation lter (EPF column), and centrifuged at 3000 g, 4℃ for 10 min.The pellet at the bottom of the EPF column tube was collected as puri ed exosome particles.Puri ed exosomes were aliquoted (50-100µL) and stored at -80℃ for future experiments 57 .

Imaging of the exosomes by Transmission Electron Microscope(TEM)
Exosomes were resuspended in 50 µl of 2% paraformaldehyde.5 µl of exosome suspension was placed on Formvar-carbon grids (two or three grids for each exosome were prepared).100 µl of PBS was put on a para lm sheet for washing the grids (kept the grids wet on the side of membrane but dried on the opposite side).Grids were transferred to a 50 µl drop of 1% glutaraldehyde for 5 min and then transferred to 100 µl of ddH 2 O for 2 min.These steps were repeated seven times.
The grids were transferred from the above step to 50 µl of uranyl-oxalate solution (pH = 7) and kept for 5 min.Subsequently, the grids were transferred to a 50 µl drop of methyl cellulose-UA for 10 min on ice.The grids were removed to stainless steel loops and blotted excess uid by lter paper with drying 5 to 10 min in the air.After drying, the grids were stored in a storage box and images were captured under the electron microscope at 80 kV 58 (Frasergen Bioinformatics Co., Ltd (Wuhan, China).

Protein quanti cation by bicinchoninic acid(BCA)
BCA1-1KT kit (Sigma, USA) was used for protein quanti cation assay.Lysate (Umibio, no.UR33101) was added to puri ed extracted exosomes and kept on ice for 30 min.The volume ratio of reagent A/reagent B equaled to 50 was prepared.The total volume was calculated as 200 µl for each well multiplied by the number of total samples including standard and test samples.The standard dilution method was adopted by adding deionized water (180 µl in tube 1).Twenty-ve microliters of standard and test samples were added into 96-well plates with 200 µl BCA solution and mixed moderately.Plate was covered with aluminum foil and incubated at 37℃ for 30 minutes.The sample was analyzed by spectrophotometer and protein concentration was calculated from the standard curve 59 .

Western blot
Exosome protein concentration was mixed with 1/4 volume of 5x loading buffer and heated for 10 minutes.Fifteen microliters of each sample were loaded on 8% SDS-PAGE gels and run for 90 min.Protein was transferred onto PVDF membranes and probed with CD63 (1:1000 dilutions; Abcam company, no.ab193349) and TSG101 (1:1000 dilutions; Abcam company, no.ab83) primary antibodies and developed by secondary antibody, HRP reagent (Millipore, USA) 60 .

RNA isolation small library preparation and sequencing
Total RNA from the hippocampus tissue of each group were extracted.Trizol (Invitrogen, no.15596-026) method was used for RNA extraction and separated by 1.2% agarose gel electrophoresis.Quality was tested by using the NanoPhotometer® spectrophotometer (IMPLEN, CA, USA).Nanodrop detected the RNA purity (OD 260/280), concentration, and nucleic acid absorption peak was normal.1µg of total RNA was used as input for the Truseq Small RNA sample prep kit (Illumina) by following the manufacturer's recommendations and instructions 61 .The library was puri ed using 6% Novex TBE PAGE gel and quanti ed using Picogreen on the TBS-380 Fluorometer (Turner Biosystems).miRTarBase 62 software was used to predict and analyze target genes for known and novel miRNA 63 .Differentially expressed miRNAs target genes between WT vs ADNS and ADNS vs ADF were analyzed and performed gene ontology (GO) and KEGG enrichment were analyzed 64 .

Statistical Analysis
The SPSS software (International Business Machines Corporation, IBM, USA) was used for statistical analysis.All data was expressed as means ± SEM.Differences among multiple groups were analyzed by one-way analysis of variance (ANOVA) while differences between two groups were analyzed using Dunnett tests.A value of p < 0.05 was considered statistically signi cant.

Declarations
China (2020L0493 to HLY, 2019L0765 to YG), Leading Team of Medical Science and Technology, Shanxi Province (No.2020TD05 to G.C.M), Datong Municipal Science and Technology Bureau (2020061 to HLY).
We are grateful to thank Wuhan Frasergen Bioinformatics Co., Ltd for assisting in sequencing and/or bioinformatics analysis.