ABSTRACT
The development of novel treatments that restore brain function and improve patient outcomes for Alzheimer's disease (AD) is necessary, given the complications and lack of improvement in recently approved amyloid beta (Aβ)‐targeting drugs. Cell‐derived extracellular vesicles (EVs) have been found to improve cognitive function through reduced inflammation, oxidative stress, and apoptosis, restoring neuronal and blood‐brain barrier function, and inhibiting Aβ and phosphorylated tau build‐up in the brain. Given the recent emergence of EVs into clinical trials, it is essential to provide the field with an update on proposed mechanisms of action, gaps in knowledge for further study, and recommendations for producing EVs with high therapeutic efficacy to ensure success in subsequent clinical trials. This systematic review summarizes original research to date that reports effects of mammalian cell‐derived EVs for the treatment of AD. Evidence of therapeutic benefits and reported mechanisms of action are discussed. Further, methods for engineering EVs to increase their therapeutic efficacy and produce high‐quality EVs relevant to the AD field are outlined. The quality of evidence is discussed in terms of reporting guidelines from the Minimal Information for Studies of Extracellular Vesicles (MISEV). The review further discusses current preclinical AD models and provides direction to improve the quality of AD models for testing novel therapeutics.
Keywords: Alzheimer's disease, extracellular vesicles, neurodegeneration, therapeutic, treatment
Cell‐derived extracellular vesicles offer promise as a treatment for Alzheimer's disease and other dementias. Extracellular vesicles have been produced from various cell sources, with and without additional modifications to improve their therapeutic efficacy. In models of AD, they have found to improve neurogenesis, blood‐brain barrier integrity and cognitive function while reducing toxic protein deposits, inflammation, and oxidative stress. Created in BioRender. S. Willerth (2025) https://BioRender.com/s58m464
1. Background
Alzheimer's disease (AD) is the most prevalent form of dementia, accounting for 50%–70% of all dementia cases (Zhang et al. 2021). It is estimated that 1 in 5 women and 1 in 10 men over 45 will develop AD during their lifetime (Chêne et al. 2015), with around 50 million people worldwide currently suffering from dementia, and an expectation that this number will triple by 2050 (Zhang et al. 2021). Costs for health care, long‐term care, and hospice for those living with dementia are estimated at $345 billion in 2023, while unpaid dementia caregiving is valued at $449.5 billion in 2022 (2023 Alzheimer's Disease Facts and Figures 2023). Treatments to improve outcomes for AD patients are strongly needed to improve both patient and caregiver quality of life.
Defined pathological features of AD include raised levels of amyloid‐β (Aβ) forming senile plaques and hyperphosphorylated tau (p‐tau) forming neurofibrillary tangles. Aβ is a small protein released by the sequential cleavage of amyloid precursor protein (APP) by beta secretase (β‐amyloid cleaving enzyme, BACE) and gamma secretase (γ‐secretase). This process produces amyloidogenic fragments that can aggregate into soluble oligomers (e.g., Aβ40 or Aβ42) or deposit as plaques. Aβ plaques are formed when the protein is misfolded and begins to accumulate in the brain. Neurofibrillary tangles are formed by hyperphosphorylation of tau protein, which causes it to aggregate in an insoluble form. Elevated levels of Aβ oligomers and p‐tau often coincide with abnormal neuronal processes, activated astrocytes and microglia, and damage to the blood‐brain barrier (BBB). There are only two Food and Drug Administration (FDA)‐approved disease‐modifying treatments currently available: lecanemab and donanemab, which target removal of Aβ from the brain. Though some modest improvements are seen (Budd Haeberlein et al. 2022; Van Dyck et al. 2023), the drugs can have severe side effects, especially for those with cerebrovascular complications, leading to amyloid‐related imaging abnormalities (ARIA), which occur in ∼40% of treated cases and result in edema, hemorrhage, and superficial siderosis (Jäkel et al. 2022). Due to the complexity and heterogeneity of AD pathology, multimodal treatments that not only break up and reduce toxic protein deposits but also reduce neuroinflammation and restore BBB permeability and normal neural cell functions are necessary.
Cell‐derived extracellular vesicles (EVs) have shown promise for the treatment of AD and other dementias, and depending on the cell type they are isolated from, have been found to improve cognitive function through inhibitory effects on Aβ and p‐tau accumulation, improving neuronal and BBB function, restoring dendritic length and spine density, and reducing neuroinflammation, oxidative stress, and apoptosis in AD models (Deng et al. 2024). There is also evidence that they can penetrate the BBB, migrate to the brain, and incorporate into neurons, microglia, and astrocytes in animal models (Cone et al. 2021; Dou et al. 2021; Poltavtseva et al. 2021; Markoutsa et al. 2022, Xu et al. 2022; Attaluri et al. 2023; Ebrahim et al. 2024; Lin, Hsu, et al. 2024). EVs are lipid bilayer membrane‐bound particles released from cells that contain bioactive signalling molecules, or ‘cargo,’ that can modify the gene expression and pathological state of recipient cells (Phelps et al. 2018). Their lipid bilayer membrane is representative of their parent cells and can protect the cargo from degradation, while expressed receptors on the EV surface may play a role in targeting specific cell types (Kumar et al. 2024). EVs offer fewer translational hurdles compared to cells due to their non‐living nature and small size. Further, while allogeneic cells can result in an immune response and potential rejection by the recipient, it is speculated that EVs from allogeneic sources may not produce the same adverse effects (Xie et al. 2023; Johnson et al. 2023). One study found that EVs from induced pluripotent stem cells (iPSCs), cells that are not typically immune‐privileged, presented minimal immune response in preclinical animal studies (Gu et al. 2022), while a meta‐analysis of clinical trials assessing the safety of EV‐based therapies found a significantly higher number of adverse events for EVs from autologous sources (34.0%, n = 5) as opposed to those from allogeneic sources (0.9%, n = 14), though it should be noted that significant heterogeneity in methodology exists between studies (Van Delen et al. 2024). Allogeneic therapies are of benefit as cells can be cultured to produce stocks of EVs for off‐the‐shelf use. In addition, cells derived from aged and/or diseased patients may have reduced or impaired functionality (Pérez et al. 2018); therefore, patients would benefit from EVs isolated from the cells of healthy individuals.
All cells produce EVs, and they can be collected and isolated from cultured cell conditioned medium (CM). The therapeutic potential of EVs was first demonstrated in 2010 when it was shown that mesenchymal stem cell (MSC)‐derived EVs could induce the repair of cardiac tissue in a murine model of myocardial ischaemia/reperfusion injury with similar efficacy as their parent cells (Lai et al. 2010). Shortly thereafter, in 2011, it was demonstrated that EVs could be functionalized and loaded, engineering dendritic EVs with the rabies virus glycoprotein (RVG) peptide and loading them with GAPDH silencing RNA (siR) to knockdown beta‐secretase 1 (BACE1), an important target in AD pathogenesis (Alvarez‐Erviti et al. 2011). MSCs are widely recognized for their therapeutic capabilities, and therefore the field has naturally focused on MSC‐derived EVs, but increasingly, EVs from other cell types, including neural stem cells (NSCs), iPSCs, microglia, cerebral microvascular endothelial cells (CMECs), astrocytes, macrophages, and engineered HeLa, HEK293T, and dendritic cells, are being used therapeutically. Recent research has focused on ways to engineer EVs to improve their efficacy in specific applications, including using them as drug carriers and as delivery systems to enable precise gene editing (Yin et al. 2023). Engineering methods either indirectly modify the EVs by targeting their host cells (e.g., via pre‐conditioning or transfection) or directly modify EVs via the encapsulation or adhesion of specific biomolecules. Early‐phase clinical trials have demonstrated the safety of MSC‐EVs in patients with AD, with no adverse events reported (Xie et al. 2023; Morita et al. 2024). Therefore, it is necessary to provide the field with an update on proposed mechanisms of action and therapeutic targets, gaps in knowledge for further study, and recommendations for producing EVs with high therapeutic efficacy to ensure success in subsequent clinical trials.
The aims of this systematic review are (i) to summarize the evidence of therapeutic benefits and the proposed mechanisms of action of EVs to treat AD, (ii) to report on the methods for EV production that have been used in the AD field to date, including methods to engineer and improve the efficacy of EVs, (iii) to report on the cell and animal models that have been used, and the range and repeatability of EV dosing, and (iv) to better understand the quality of evidence based on the EV collection, isolation, and characterization methods used, following reporting guidelines from the 2023 Minimal Information for Studies of Extracellular Vesicles (MISEV2023) (Welsh et al. 2024). Included studies are original research publications to date that have reported effects of mammalian cell‐derived EVs for the treatment of AD. The review will further discuss challenges in the EV field and new technologies that could be applied to improve EV therapeutic efficacy, as well as current and potential preclinical AD models to improve the screening of novel EV‐based treatments.
2. Materials and Methods
2.1. Literature Search
The protocol for this systematic review was conducted according to the Preferred Reporting Items for Systematic reviews and Meta‐Analyses (PRISMA) guidelines (Page et al. 2021). The review protocol was defined a priori, including the search terms, inclusion and exclusion criteria, and data extraction process. The protocol was not registered; however, here we document the protocol used in detail. Two databases were searched: MEDLINE/PubMed and Web of Science on September 20, 2024, updated on December 2, 2024, and again on May 12, 2025. For the updated searches, date ranges were selected to prevent overlap. To ensure the inclusion of EVs, the search terms included extracellular vesicles and their commonly reported related terms: exosomes, microvesicles, and ectosomes. Since a majority of EV‐based treatments have focused on EVs from stem cells, treatment terms included stem cells and treatment. And lastly, to ensure the inclusion of Alzheimer's disease and other dementias, the search terms dementia and Alzheimer's disease were included. For PubMed, the search strategy was (“extracellular vesicle*” OR “exosome*” OR “microvesicle*” OR “ectosome*”) AND (“stem cell*” OR “treatment”) AND (“dementia” OR “Alzheimer's disease”). For Web of Science, the search strategy was: ALL = (“extracellular vesicle*” OR “exosome*” OR “microvesicle*” OR “ectosome”) AND ALL = (“stem cell*” OR “treatment”) AND ALL = (“dementia” OR “Alzheimer's disease”).
2.2. Inclusion and Exclusion Criteria
Inclusion criteria for the search included EV‐based treatments derived from mammalian cells for AD and other cognitive impairments such as those from metabolic disorders such as diabetes, vascular conditions, and Down syndrome, which are risk factors and often co‐occur with AD pathology. Exclusion criteria for the search included:
Review papers, non‐peer‐reviewed publications, and conference abstracts,
Pathological studies that did not describe EVs for the intended use as a therapeutic,
Studies that only evaluated internalization and/or biodistribution of EVs,
EVs derived from non‐mammalian cells (e.g., plant cells, plasma),
Treatment of cognitive impairment not strictly related to AD (e.g., Parkinson's disease),
Cell models that were not specific to AD or other cognitive impairments (e.g., tested on healthy cells),
EVs loaded with drugs and tested without an EV‐only control,
Co‐culture cell models that did not strictly isolate EVs for treatment (e.g., benefits attributed to EVs released during co‐culture).
2.3. Eligibility Screening
Two authors independently reviewed and screened titles and abstracts using the web‐based Covidence platform. A full text review of the selected abstracts was subsequently conducted independently by both authors. All conflicts were resolved through consensus review.
2.4. Data Extraction
Data extraction was completed by an individual author and checked by the second author. The following information was extracted from the included papers:
Title, first author, journal, and year of publication,
Cell type (including source and passage) used for producing the EVs,
Culture conditions during EV production (when reported), including cell medium, cell confluence at the start of EV collection, and length of collection period,
EV term that was used in the publication (e.g., EV, small EV, exosome, microvesicle),
Whether the cells or EVs were modified, and in what way,
EV delivery method (only applicable for in vivo studies),
EV quantification method (e.g., protein or particle concentration),
Models and assays used for testing, reported effects of EVs, and dosage,
Evidence of mechanism of action,
EV isolation method,
EV characterization techniques used,
EV labelling method (if applicable).
2.5. Assessment of Methodological and Reporting Quality
Methodological quality is discussed throughout the results, and individual studies were assessed by their adherence to the relevant and mandatory items on the MISEV2018 Checklist (Théry et al. 2018), with guidance from the MISEV2023 update (Welsh et al. 2024) and the ISEV cell culture‐conditioned medium‐derived EVs task force perspective on considerations for reporting cell culturing parameters (Shekari et al. 2023). Adherence to 20 criteria was assessed with a score given as a percentage of criteria adhered to as described in Table 1. Assessments were conducted by an individual author and checked by a second author.
TABLE 1.
Assessment of study adherence to MISEV guidelines.
| # | Category | Yes – 1 | Partial – 0.5 | No – 0 |
|---|---|---|---|---|
| 1 | EV nomenclature | Appropriate nomenclature used (e.g., generic term or specific term with demonstration of subcellular origin) | Two terms are used and only one term is appropriate | Specific term used without evidence of origin |
| 2 | Collection and pre‐processing: general cell characterization | Identity (cell type and source) and passage defined | Identity or passage not defined | Identity and passage not defined |
| 3 | Collection and pre‐processing: medium used | Medium defined before and during collection | Medium defined before or during collection | Medium not defined |
| 4 | Collection and pre‐processing: nature and size of culture vessels | Nature and size of culture vessels, and volume of medium during conditioning provided | Nature of culture vessels or volume provided | Nature and size of culture vessels, and volume of medium during conditioning not provided |
| 5 | Collection and pre‐processing: measure of cell confluence and viability | Cell confluence and viability reported at time of collection (or at time of seeding with estimation) | Confluence or viability provided | Cell confluence and viability not reported |
| 6 | Collection and pre‐processing: frequency and interval of CM harvest | Interval and frequency (if applicable) of CM harvest provided | Length of collection period provided but not clear on start time or frequency | Interval and frequency of CM harvest not provided |
| 7 | Storage and recovery | Storage conditions both before EV isolation and after isolation provided | Storage conditions before or after provided | Storage conditions not provided |
| 8 | EV separation and concentration | Experimental details provided for specified method | Protocol defined but lacking details (e.g., tubes, rotor, volume) | Experimental details not provided |
| 9 | EV characterization: volume of medium and/or cell number | A measure of quantification of EV source is provided | Not applicable | Quantification of EV source is not provided |
| 10 | EV characterization: global quantification by at least 2 methods | EVs are quantified using a minimum of two methods | EVs are quantified using one method | EVs are not quantified |
| 11 | EV characterization: ratio of 2 quantification figures | Provided | Not applicable | Not provided |
| 12 | Global characterization: transmembrane or GPI‐anchored protein (category 1) | Provided | Not applicable | Not provided |
| 13 | Global characterization: cytosolic protein (category 2) | Provided | Not applicable | Not provided |
| 14 | Global characterization: assessment of expected contaminants—purity (category 3) | Provided | Not applicable | Not provided |
| 15 | Global characterization: assessment of expected contaminants—negative marker (category 4) | Provided | Not applicable | Not provided |
| 16 | Single EV characterization: wide‐field and close‐up images of single EVs | Wide‐field and close‐up images provided | Images of wide‐field or close‐up images provided | Images not provided |
| 17 | Single EV characterization: non‐image‐based assessment (quantification) of large numbers of single EVs | Provided | Not applicable | Not provided |
| 18 | Functional studies: dose‐response | Provided | Not applicable | Not provided |
| 19 | Functional studies: negative control (nonconditioned medium) | Provided | Not applicable | Not provided |
| 20 | Functional studies: comparison to CM or EV‐depleted CM | Provided | Not applicable | Not provided |
Note: The scoring system for this review was developed based on the relevant and mandatory items on the MISEV2018 Checklist (Théry et al. 2018), with guidance from the MISEV2023 update (Welsh et al. 2024) and the ISEV cell culture‐conditioned medium‐derived EVs task force perspective on considerations for reporting cell culturing parameters (Shekari et al. 2023).
3. Results
3.1. Literature Search
A total of 608 papers were found in the search after removal of 291 duplicates. After complete screening, a total of 89 papers were included, with the selection process outlined in Figure 1 and conducted according to the PRISMA guidelines. Thirty‐eight publications were excluded after full‐text review for the following reasons: not peer reviewed (n = 4), EVs from plant or other non‐mammalian cells (n = 5), EVs not strictly tested as a therapeutic (n = 4), model not specific to AD or relevant neurodegeneration (n = 8), drug‐loaded EVs tested without an EV‐only control (n = 8), drug‐loaded EVs from biofluids (n = 5), only internalization or migration tested (n = 3), and one proposal publication that did not present research results (n = 1). A complete data extraction table detailing all studies can be found in the supplemental Excel file. The file can be used as a resource and sorted by extracted metrics as listed in Section 2.4 or by adherence to MISEV as outlined in Table 1.
FIGURE 1.
PRISMA flow diagram of the systematic literature search.
Figure 2 outlines the number of publications included in the review per year. A large portion of publications have focused on MSC‐derived EVs (pink overlay), but there is increasing interest in producing EVs from other cells, including iPSC‐derived progenitor cells, mature brain‐derived cells, and engineered cell lines.
FIGURE 2.
Number of publications per year. Total publications are seen in blue, with the number of MSC publications overlaid in pink, and the number of other cell types overlaid in yellow.
3.2. EV Production
Figure 3 depicts the major cell types used for producing EVs in the included studies, and the proportion that have modified cells or EVs to increase their efficacy. A majority of the included studies (64%, n = 57) evaluated EVs isolated from primary MSCs, of which 47% (n = 27) focused on bone marrow‐derived MSCs (BMSCs), 18% (n = 10) on adipose‐derived MSCs (ADSCs), 18% (n = 10) on umbilical cord‐derived MSCs (UMSCs), 7% (n = 4) on Wharton jelly‐derived MSCs (WJMSCs), 4% (n = 2) on amniotic fluid‐derived MSCs (AFMSCs) (Gatti et al. 2020; Zavatti et al. 2022), 2% (n = 1) on olfactory mucosa‐derived MSCs (OMMSCs) (Hu et al. 2025), and 5% (n = 3) did not specify source (Markoutsa et al. 2022; Reza‐Zaldivar et al. 2019; Amer et al. 2024). More recent publications have focused on EVs derived from other sources, including NSCs (10%, n = 9), iPSCs and derived progenitors (iNSCs, iMSCs, and iPSCs) (8%, n = 7), embryonic stem cells (ESCs) (1%, n = 1) (Hu et al. 2020), mature neural cells including anti‐inflammatory microglia (4%, n = 4) (Li et al. 2022; Chen et al. 2023; Wang, Yang, et al. 2023; Zhao et al. 2023), CMECs (3%, n = 3) (Pan et al. 2020; Zhang et al. 2022; Ding et al. 2022), astrocytes (2%, n = 2) (Deng et al. 2021; Chen, Deng, et al. 2024), and neurons (1%, n = 1) (Dou et al. 2021), and other cell types including HEK293T cells (1%, n = 1) (Cai et al. 2024), SH‐SY5Y cells (1%, n = 1) (Évora et al. 2025), and engineered macrophages (1%, n = 1) (Hao et al. 2022), HeLa cells (1%, n = 1) (Yu et al. 2021), and dendritic cells (DCs) (1%, n = 1) (Lin, Hsu, et al. 2024).
FIGURE 3.
EV production metrics. (A) Percentage of EV‐producing cell types, with (B) MSCs broken down into their tissue source. (C) Percentage of EVs engineered by modifying the cells, EVs, or both, with (D) types of modifications done on cells (including studies that modified both cells and EVs), and (E) types of modifications done on EVs (including studies that modified both cells and EVs). (F) Percentage of EVs produced in various types of culture media.
The therapeutic efficacy of EVs can be modified by altering the culture conditions of the parent cells, incubating the cells with different biological compounds, modifying the cells through transfection of RNAs, or directly loading or engineering the EVs following their isolation. Of the studies included, 63% (n = 56) evaluated the efficacy of EVs from baseline cell culture, where neither the cells nor EVs were modified or pre‐treated. Twenty‐seven percent of studies (n = 24) engineered the EVs through modification of only the cells or cell culture conditions to improve their efficacy. Of these, 46% (n = 11) modified the cells through transfection or overexpression of target genes, including myocardial infraction associated transcript (MIAT) (Qi et al. 2021), SHP2 (Xu et al. 2022), micro‐RNA (miR)‐132‐3pp (Ma et al. 2022), miR‐29b (Jahangard et al. 2020), miR‐22 (Zhai et al. 2021), miR‐211‐5p inhibitor (Chen, Huang, et al. 2024), RVG and CD10 (Yu et al. 2021), RVG, Gap43 and Snap25 mRNAs (Cai et al. 2024), miR‐29b‐2 and CD47 (Lin, Hsu, et al. 2024), miR‐214‐3p (Zhang 2025), and miR‐132‐3p (Wang et al. 2025). The remaining 54% (n = 13) altered culture conditions to modify EV content via their parent cells by culturing the cells as 3D spheroids (Cone et al. 2021; Pourhadi et al. 2024) or within 3D scaffolds (Yang et al. 2020), exposing them to pro‐inflammatory conditions (Markoutsa et al. 2022; Losurdo et al. 2020), heat shock (Huber et al. 2022), hypoxia (Liu, Jin, et al. 2022), ultrasound (Deng et al. 2021), specific wavelengths of light (Chen et al. 2023), or stimulating the cells with different compounds (Dou et al. 2021; Chen, Deng, et al. 2024; Sheykhhasan et al. 2022; Bashirrohelleh et al. 2025). Seven percent of the included studies (n = 6) modified EVs directly by tagging them with RVG (Cui et al. 2019), loading them with active molecules including AM1241 CB2 receptor by passive diffusion (Zhu et al. 2023), neprilysin (NEP) by freeze‐thaw cycling (Izadpanah et al. 2020), berberine and palmatine by sonication (Zhao et al. 2023), or berberine and BACE1 siR by electroporation (Sun et al. 2025), or encapsulating them into a hydrogel for enhanced release kinetics (Huang et al. 2024). Three percent (n = 3) modified both cells and EVs. One study treated macrophages with gemfibrozil and subsequently modified produced EVs by mixing them with a DSPE‐PEG‐mannose chain to specifically target microglia via mannose receptors (Hao et al. 2022). Another exposed ADSCs to turbulent vortex flow for 4 h to induce EV production and then loaded the EVs with donepezil by passive diffusion (Silva et al. 2025). A third study compared EVs loaded with miR‐124‐3p either by transfection of SH‐SY5Y cells or by direct loading into the isolated EVs using an Exo‐Fect Exosome Transfection Kit (Évora et al. 2025). Notably, EVs with miR‐124‐3p loaded using Exo‐Fect demonstrated a higher delivery efficiency and higher potency in reducing microglial activation.
Other factors within cell culture, including the medium that the cells are exposed to, their passage number, and their confluence, also affect EV content. Culture medium often contains serum which contains its own EVs, and therefore an alternative medium is needed during the EV collection phase. This is most commonly done by depleting the serum of EVs by overnight ultracentrifugation or by utilizing a defined serum‐free medium (Figure 3F). Thirty‐one percent of studies (n = 28) used EV‐depleted fetal bovine serum (FBS), another 6% (n = 5) defined their medium as depleted of EVs, and 22% (n = 20) used defined serum‐free medium. Twenty‐one percent (n = 19) exposed their cells to serum starved conditions during EV collection, meaning they simply removed the serum from culture medium prior to the EV collection phase to prevent co‐isolation of serum‐EVs. Three of these studies supplemented the serum‐removed medium with albumin. Ten percent of studies (n = 9) collected EVs from a medium containing FBS, and another 9% (n = 8) did not specify the culture medium used for EV collection. With regard to passage number, one study evaluated its effect on EV efficacy and found that BMSCs at a higher passage number (passage 8) produced EVs with reduced neuroprotective potential (Venugopal et al. 2018). Forty‐four percent of studies (n = 39) reported passage number at the time of EV collection, ranging from passage 1 to 17, and most commonly passage 3 (10/39 of reporting studies, 26%). Twenty‐seven percent of studies (n = 24) reported cell confluence at the beginning of EV collection, with a range of 60%–90%. Of the included studies that reported collection period (67%, n = 60), the collection period ranged from 3 h up to 6 days, with most studies collecting EVs within 24–72 h (50/60 of reporting studies, 83%).
3.3. EV Isolation and Characterization Methods
Appropriate isolation and characterization methods are needed to attribute effects to EVs. A pure population of EVs is difficult to isolate due to the high amount of protein aggregates and lipoproteins that co‐isolate with EV populations due to their similar size and density, and/or interactions at the EV surface. No single isolation method is perfect but rather sits somewhere on the spectrum of being highly specific with low yield, or high yield with low specificity. The choice of isolation method therefore depends on the downstream application. The isolation methods used in the included studies are visualized in Figure 4A. For EV treatments in AD, 61% of studies (n = 54) utilized differential ultracentrifugation (UC) to isolate the EV fraction, followed by polyethylene glycol (PEG) precipitation (19%, n = 17). A smaller number of studies used density gradient ultracentrifugation (DGUC) (4%, n = 4) (Huber et al. 2022; Zhu et al. 2023; Nakano et al. 2016; Zhao et al. 2019), ultrafiltration (UF) (3%, n = 3) (Poltavtseva et al. 2021; Zhdanova et al. 2021; Li, Ye, et al. 2024), or size exclusion chromatography (SEC) (2%, n = 2) (Zhao et al. 2023; Chen et al. 2021). Other studies used combinations of methods including UC and PEG precipitation (4%, n = 4) (Cone et al. 2021; Markoutsa et al. 2022; Xie et al. 2023; Khan et al. 2023), anion exchange chromatography (AEC), dialysis, and UF (1%, n = 1) (Lin, Huang, et al. 2024), AEC and SEC (2%, n = 2) (Madhu et al. 2024; Rao et al. 2025), tangential flow filtration (TFF) and UC (1%, n = 1) (Silva et al. 2025), or UF and UC (1%, n = 1) (Hu et al. 2025).
FIGURE 4.
EV isolation and quantification methods. (A) Percentage of studies utilizing specified isolation techniques. (B) Percentage of studies reporting measures of EV quantity by different means.
In terms of EV characterization, MISEV guidelines recommend at minimum reporting: (1) a quantitative measure of EV source (e.g., the number of secreting cells), (2) an approximate abundance of EVs (e.g., total particles, protein), (3) the abundance of proteins specific to EVs, including at least one membrane‐associated protein (e.g., CD63, CD81, CD9) and one cytosolic protein (e.g., ALIX, flotillin‐1/2, syntenin, TSG‐101), and (4) the abundance of co‐isolated non‐EV associated components (e.g., albumin, lipoproteins) (Welsh et al. 2024). The type of measurements used for approximate abundance in the included studies is seen in Figure 4B. Fifty‐eight percent of studies (n = 52) utilized protein measurements for dosing, followed by total particle measurements (28%, n = 25), and another 8% (n = 7) included both or a mix of protein and particle measures for dosing. Only 4% (n = 4) did not report an approximate abundance for the dose of EVs. Twenty‐two percent (n = 20) reported a quantitative measure of EV source based on the number of cells at the time of EV collection (or at time of seeding with an estimation at time of collection). Though some studies reported the volume of CM or the seeding density of the cell culture, this information does not give a quantitative measure of EV source, since this is dependent on cell density and the rate of proliferation.
With regard to EV characterization techniques, a combination of western blot (WB), transmission electron microscopy (TEM), and nanoparticle tracking analysis (NTA) was the most common (47%, n = 42). Overall, 72% of studies (n = 64) utilized WB for protein abundance measurements, 79% (n = 70) utilized TEM for size and morphology, and 58% (n = 52) utilized NTA. Eighteen percent (n = 16) used flow cytometry (FC), often in place of WB for the measurement of EV‐specific proteins. Dynamic light scattering (DLS) was used in 17% of studies (n = 15), and was at times confused with NTA, where authors reported using NTA despite the use of a DLS instrument (e.g., Malvern's Zetasizer). Where NTA was reported but a DLS instrument was used with corresponding size intensity results, the study was reported as using DLS. Some studies utilized methods such as scanning electron microscopy (SEM, 6%, n = 5) and atomic force microscopy (AFM, 2%, n = 2) (Pan et al. 2020; Khan et al. 2023) in place of TEM. Eleven percent (n = 10) did not provide any EV‐specific protein characterization, and 3% (n = 3) did not include any EV characterization at all (Wang et al. 2018; Liu et al. 2020; Campbell et al. 2023). Regarding protein characterization, 31% of studies (n = 28) included at least one membrane‐associated protein, one cytosolic protein, and one negative control associated with intracellular components other than EVs (e.g., calnexin). Only 3% (n = 3) included a purity control to measure co‐isolated non‐EV‐associated components from the culture medium (Wang, Yang, et al. 2023; Cai et al. 2024; Gao et al. 2023). Overall, membrane‐associated proteins CD9, CD63 and CD81 were the most measured, with 89% of studies (n = 79) measuring at least one. Seventy‐four percent (n = 66) measured CD63, 52% (n = 46) measured CD9, and 34% (n = 30) measured CD81. Fifty‐eight percent (n = 52) measured cytosolic proteins, with the most common being TSG101 (33%, n = 29), ALIX (25%, n = 22), and HSP70 (16%, n = 14). Calnexin was the most frequently measured negative control (25%, n = 22), with a total of 36% (n = 32) measuring a negative control. A summary of collection, isolation, and characterization methods for each individual study is summarized in Table 2, with reported adherence to collection and pre‐processing guidelines (items 2–6, Coll. Rep.) and characterization guidelines (items 9–17, Char. Rep.) as per Table 1.
TABLE 2.
EV collection, isolation, and characterization methods with percent adherence to MISEV collection and pre‐processing reporting (Coll. Rep.) and characterization reporting (Char. Rep.).
| EV source | Collection medium (interval) | Isolation | Char. methods | Protein | Coll. Rep. | Char. Rep. | Ref. |
|---|---|---|---|---|---|---|---|
| BMSCs | α‐MEM with 15% EV‐depleted FBS, 1% penicillin‐streptomycin (48 h starting at 60%–80% confluence) | DGUC | TEM, WB | 1, 2 | 70% | 33% | (Nakano et al. 2016) |
| DMEM‐KO with 10% EV‐depleted FBS, 100 U/mL penicillin‐streptomycin, 1% L‐glutamine | PEG precipitation | WB | 1, 2 | 50% | 28% | (Venugopal et al. 2018) | |
| α‐MEM with 15% EV‐depleted FBS, 1% penicillin‐streptomycin (48 h) | PEG precipitation | None | None | 50% | 6% | (Wang et al. 2018) | |
| DMEM/F‐12 with 100 U/mL penicillin, 100 µg/mL streptomycin (24 h) | UC | TEM, NTA, FC, dot blotting | 1 | 50% | 56% | (de Godoy et al. 2018) | |
| α‐MEM (3 h) | UC | TEM, NTA, WB, FC | 1, 2, 4 | 40% | 78% | (Elia et al. 2019) | |
| DMEM with 15% EV‐depleted FBS | UC | NTA | None | 30% | 17% | (Cui et al. 2019) | |
| Conventional medium with EV‐depleted FBS (48 h starting at 60–80% confluence) | DGUC | TEM, WB | 1, 2 | 40% | 33% | (Zhao et al. 2019) | |
| DMEM with or without 20 ng/mL TNF‐α, 25 ng/mL IFN‐γ (24–48 h) | UC | NTA, WB | 1, 2 | 50% | 56% | (Losurdo et al. 2020) | |
| RPMI with 0.5% BSA | UC | FACs | 1 | 40% | 22% | (Hassan et al. 2020) | |
| DMEM with 10% EV‐depleted FBS (24–48 h) | UC | SEM, DLS, WB | 1, 2 | 80% | 56% | (Izadpanah et al. 2020) | |
| DMEM with 10% EV‐depleted FBS, 1% penicillin‐streptomycin (48‐72 h) | UC | SEM, DLS | None | 40% | 17% | (Jahangard et al. 2020) | |
| L‐DMEM with 10% EV‐depleted FBS (48 h starting at 80% confluence) | UC | TEM, NTA, WB | 1, 4 | 70% | 50% | (Sha et al. 2021) | |
| Serum‐free medium (24 h starting at 80% confluence) | PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 80% | 44% | (Xiong et al. 2021) | |
| α‐MEM with 10% EV‐depleted FBS, 1% penicillin‐streptomycin (48 h) | UC and PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 80% | 72% | (Cone et al. 2021) | |
| DMEM | PEG precipitation | TEM, WB | 1 | 40% | 39% | (Wang and Yang 2021) | |
| RPMI‐1640 (overnight) | UC | TEM | None | 40% | 6% | (Aboulhoda et al. 2021) | |
| EV‐depleted DMEM with 10% FBS, 100 U/mL penicillin, 100 mg/mL streptomycin (48 h) | UC | TEM, DLS, WB | 1 | 50% | 33% | (Xu et al. 2022) | |
| DMEM with 15% FBS, 1% penicillin‐streptomycin, 2 ng/mL bFGF (72 h) | UC | NTA, FC, scanning helium‐ion microscopy | 1 | 60% | 33% | (Kaniowska et al. 2022) | |
| DMEM with 10% FBS (24 h) | UC | TEM, NTA, WB | 1 | 50% | 33% | (Ma et al. 2022) | |
| α‐MEM with EV‐depleted 10% FBS, 1% penicillin‐streptomycin, 20 ng/mL EGF, 20 ng/mL bFGF (48 h) | PEG precipitation | TEM, WB | 1, 2 | 40% | 39% | (Liu, Fan, et al. 2022) | |
| α‐MEM with 10% EV‐depleted FBS (24 h starting at 60–80% confluence) | UC | None | None | 70% | 11% | (Campbell et al. 2023) | |
| DMEM/F‐12 (12 h) | DGUC | TEM, DLS, WB | 1, 2 | 33% | 40% | (Zhu et al. 2023) | |
| DMEM with 0.5% HSA | UC | TEM, FC | 1 | 50% | 22% | (Ebrahim et al. 2024) | |
| Exosome‐free medium | UC | TEM | None | 20% | 11% | (Hamed et al. 2025) | |
| EV‐depleted medium with 10% FBS (72 h) | UC | TEM, WB | 1, 2, 4 | 80% | 56% | (Zhang 2025) | |
| DMEM with 1% penicillin‐streptomycin (48 h starting at 80–90% confluence) | PEG precipitation | FC, DLS, SEM | 1 | 40% | 22% | (Bashirrohelleh et al. 2025) | |
| DMEM with 10% EV‐depleted FBS (48 h) | UC | TEM, FC, DLS | 1 | 50% | 22% | (Sadeghi et al. 2025) | |
| ADSCs | EV‐depleted EGM‐2‐MV | PEG precipitation | WB | 1, 2 | 40% | 28% | (Lee et al. 2018) |
| Complete medium with 10% EV‐depleted FBS (72 h starting at 90% confluence) | UC | TEM, NTA, WB | 1, 2 | 40% | 50% | (Ma et al. 2020) | |
| DMEM with 10% EV‐depleted FBS (72 h) | UC | TEM, NTA, WB | 1, 2 | 50% | 50% | (Zhai et al. 2021) | |
| Serum‐free Prime‐XV (24 h starting at 80–90% confluence) | PEG precipitation | WB | 1 | 70% | 17% | (Garcia‐Contreras and Thakor 2021) | |
| FBS‐free EGM‐2‐MV with 1x serum replacement (48 h starting at 80–90% confluence) | UC | TEM, NTA, WB | 1 | 60% | 44% | (Liu, Jin, et al. 2022) | |
| FBS‐free medium (72 h) | PEG precipitation | TEM, FC, DLS | 1 | 40% | 17% | (Sheykhhasan et al. 2022) | |
| Low‐glucose DMEM with 10% EV‐depleted FBS, 20 ng/mL bFGF, 20 ng/mL EGF (72 h) | UC | TEM, NTA, WB, DLS | 1, 2 | 40% | 50% | (Huang et al. 2024) | |
| EV‐depleted α‐MEM with HPL (48 h starting at 90% confluence) | UC and PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 70% | 67% | (Xie et al. 2023) | |
| DMEM with 100 U/mL penicillin and streptomycin (4 h) | TFF and UC | TEM, NTA, WB, NanoFCM | 1, 2, 4 | 50% | 61% | (Silva et al. 2025) | |
| Serum free medium (36 h starting at 70–80% confluence) | UF | TEM, NTA, WB | 1, 2, 4 | 60% | 56% | (Li, Ye, et al. 2024) | |
| UMSCs | DMEM with 100 mg/mL penicillin (48 h) | PEG precipitation | TEM, WB | 1 | 50% | 28% | (Ding et al. 2018) |
| DMEM/F‐12 with EV‐depleted FBS, 1% penicillin‐streptomycin (48 h) | UC | TEM, NTA, WB | 1, 2 | 60% | 50% | (Yang et al. 2020) | |
| DMEM/F‐12 with 10% FBS | UC | TEM, NTA, WB | 1 | 20% | 44% | (Feng et al. 2020) | |
| α‐MEM with 10% EV‐depleted FBS, 100 U/mL penicillin, 100 µg/mL streptomycin (48 h) | UC | TEM, NTA, WB | 1, 2, 4 | 50% | 67% | (Wang, Liu, et al. 2021) | |
| α‐MEM with 10% FBS, 1% penicillin‐streptomycin | UC | TEM, NTA, WB | 1, 2, 4 | 40% | 61% | (Hou et al. 2023) | |
| α‐MEM (48 h starting at 70–80% confluence) | PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 70% | 61% | (Wang, Yi, et al. 2023) | |
| DMEM/F‐12 with 10% EV‐depleted FBS (48 h starting at 80% confluence) | UC | TEM, NTA, WB | 1 | 80% | 22% | (Chen, Huang, et al. 2024) | |
| DMEM/F‐12 plus 2 mM glutamine, 100 mg/mL streptomycin, 100 U/mL penicillin (48 h) | PEG precipitation | TEM, FC, SEM, DLS | 1 | 60% | 22% | (Pourhadi et al. 2024) | |
| 5% EliteGro‐Adv with 95% MEM (48 h) | UC | TEM, DLS, WB, FC | 1, 4 | 70% | 33% | (Li, Zhang, et al. 2024) | |
| Unknown | UC | TEM, WB, DLS | 1, 2 | 10% | 33% | (Sun et al. 2025) | |
| WJMSCs | DMEM/F‐12 (24 h starting at 90% confluence) | UC | TEM, NTA, FC/FACS | 1 | 70% | 44% | (Bodart‐Santos et al. 2019) |
| α‐MEM with 10% EV‐depleted FBS, 1% penicillin‐streptomycin, 1% L‐glutamine (48 h starting at 60–70% confluence) | SEC | TEM, NTA, WB | 1, 2, 4 | 60% | 61% | (Chen et al. 2021) | |
| DMEM/F‐12 with 200 U/mL penicillin and streptomycin (48 h starting at 80% confluence) | UF | FC | 1 | 80% | 11% | (Zhdanova et al. 2021) | |
| DMEM/F‐12 with 2 mM L‐glutamine, 100 U/mL penicillin, and 100 µg/mL streptomycin (48 h starting at 80% confluence) | UF | FC | 1 | 80% | 11% | (Poltavtseva et al. 2021) | |
| AFMSCs | αMEM with 2 mM L‐glutamine, 100 U/mL penicillin and streptomycin (4 days starting at 1E6 cells in T75) | PEG precipitation | TEM, WB | 1, 2 | 80% | 33% | (Gatti et al. 2020) |
| α‐MEM with 2 mM L‐glutamine, 100 U/mL penicillin, 100 µg/mL streptomycin (4 days) | UC | TEM, NTA, ELISA | 1 | 80% | 44% | (Zavatti et al. 2022) | |
| OMMSCs | DMEM/F‐12 with 10% EV‐depleted FBS (48 h starting at 90% confluence) | UF and UC | TEM, NTA, WB | 1, 4 | 70% | 44% | (Hu et al. 2025) |
| MSCs (source unspecified) | Exosome production medium (3 days) | UC | WB | 1 | 10% | 28% | (Reza‐Zaldivar et al. 2019) |
| α‐MEM with 16.5% EV‐depleted FBS | UC and PEG precipitation | TEM, NTA, dot blotting | 1, 2, 4 | 30% | 61% | (Markoutsa et al. 2022) | |
| RPMI with 0.5% BSA | UC | FC | 1 | 40% | 17% | (Amer et al. 2024) | |
| NSCs | EmbryoMax DMEM/F‐12 with L‐Glutamine, B27, retinoic acid, GlutaMAX, 20 ng/mL bFGF, antibiotics | UC | TEM, NTA, WB | 1, 2, 4 | 30% | 72% | (Micci et al. 2019) |
| Neurobasal medium with 2% B27, 20 ng/mL EGF, 20 ng/mL bFGF, 5 µg/mL heparin (1 day) | UC | TEM, NTA, WB | 1, 2 | 60% | 50% | (Li et al. 2020) | |
| DMEM/F‐12 with 2% B27, 20 µg/L EGF, 20 µg/L bFGF (24 h) | UC | TEM | None | 40% | 11% | (Qi et al. 2021) | |
| ENStem‐A neural expansion medium | UC | TEM, DLS | None | 30% | 22% | (Apodaca et al. 2021) | |
| DMEM/F‐12 with N2 medium, 20 ng/mL bFGF (24 h starting at 80% confluence) | DGUC | TEM, NTA, WB | 2 | 70% | 39% | (Huber et al. 2022) | |
| Full medium containing 1% exosome‐free FBS (48 h) | UC and PEG precipitation | NTA, AFM, WB, EXOCET | 1 | 50% | 39% | (Khan et al. 2023) | |
| Neurobasal medium with 2% B27, 20 ng/mL EGF, 20 ng/mL bFGF, 5 µg/mL heparin | UC | TEM, NTA, WB | 1, 2, 4 | 30% | 61% | (Li, Chen, et al. 2024) | |
| DMEM/F‐12 with L‐glutamine, B27, retinoic acid, GlutaMAX, 20 ng/mL bFGF, antibiotics | UC | TEM, NTA, dot blotting | 1, 2, 4 | 30% | 72% | (Krishnan et al. 2025) | |
| Neurobasal medium with 2% B27, 0.5 mM GlutaMAX, 10 ng/mL bFGF, PDGFbb and EGF | UC | TEM, NTA | 1, 2 | 20% | 44% | (Wang et al. 2025) | |
| iNSCs | Neural progenitor cell medium (48 h) | UC | None | None | 40% | 0% | (Liu et al. 2020) |
| NeuroCult NSC Proliferation Medium with 20 ng/mL bFGF and EGF, 2 ug/mL heparin, 2 mM L‐glutamine, 100 U/mL penicillin and streptomycin | UC | TEM, NTA, WB, SEM | 1, 2, 3 | 30% | 56% | (Gao et al. 2023) | |
| 1:1 Neurobasal media and DMEM/F‐12 with 1× neural induction supplement | AEC, SEC | TEM, NTA, WB | 1, 2, 4 | 40% | 61% | (Madhu et al. 2024) | |
| Unknown (48 h starting at 80% confluence) | UC | TEM, WB, DLS | 1, 2 | 50% | 33% | (Chen, Lan, et al. 2024) | |
| 1:1 Neurobasal media and DMEM/F‐12 with 1% neural induction supplement and penicillin‐streptomycin | AEC, SEC | TEM, NTA, WB | 1, 2, 4 | 90% | 72% | (Rao et al. 2025) | |
| iMSCs | High‐glucose DMEM with 10% FBS, 5 ng/mL bFGF, 10 ng/mL EGF | AEC, dialysis, UF | AEC, dialysis, UF | 1, 2, 4 | 20% | 56% | (Lin, Huang, et al. 2024) |
| iPSCs | mTeSR with 10 µM Y27632 (48 h) | PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 80% | 72% | (Marzano et al. 2019) |
| ESCs | ncEpic hPSC medium | UC | TEM, NanoFCM, WB | 1, 2, 4 | 30% | 72% | (Hu et al. 2020) |
| CMECs | DMEM with 10% EV‐depleted FBS | UC | AFM, WB, DLS | 1, 2, 4 | 40% | 50% | (Pan et al. 2020) |
| Endothelial growth medium | UC | TEM, NTA, WB | 1, 2, 4 | 20% | 56% | (Zhang et al. 2022) | |
| Endothelial cell culture medium | UC | TEM, NTA, WB | 1, 2 | 40% | 44% | (Ding et al. 2022) | |
| Microglia | DMEM with 10% EV‐depleted FBS, 20 ng/mL IL‐4, 1% penicillin‐streptomycin (48 h) | UC | TEM, NTA, WB | 1, 2 | 50% | 44% | (Li et al. 2022) |
| DMEM/F‐12 with 10% FBS, 1% penicillin‐streptomycin (24 h) | PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 40% | 67% | (Chen et al. 2023) | |
| MEM with 6% EV‐depleted FBS (24 h) | UC | TEM, NTA, WB | 1, 2, 3, 4 | 50% | 78% | (Wang, Yang, et al. 2023) | |
| DMEM with 10% EV‐depleted FBS (24 h starting at 60–70% confluence) | SEC | TEM, NTA, WB | 1, 2 | 50% | 44% | (Zhao et al. 2023) | |
| Astrocytes | DMEM with 10% EV‐depleted FBS (6‐day collection starting at 80% confluence) | PEG precipitation | TEM, NTA, WB | 1, 2, 4 | 60% | 61% | (Deng et al. 2021) |
| DMEM with 10% FBS, 1% penicillin‐streptomycin | UC | TEM, NTA, WB, NanoFCM | 1 | 28% | 30% | (Chen, Deng, et al. 2024) | |
| Neurons | Neurobasal medium with 1x B27, 1x GlutaMAX | UC | TEM, NTA, WB | 1, 2, 4 | 50% | 72% | (Dou et al. 2021) |
| HeLa cells | Serum‐free medium (36–48 h) | UC | TEM, NTA, WB | 1, 4 | 50% | 50% | (Yu et al. 2021) |
| Dendritic cells | Serum‐free medium (72 h starting at 80% confluence) | PEG precipitation | TEM, NTA, WB | 1, 2 | 50% | 50% | (Lin, Hsu, et al. 2024) |
| Macrophages | DMEM with 10% FBS, 1% penicillin‐streptomycin | UC | TEM, WB, DLS | 1, 2, 4 | 30% | 39% | (Hao et al. 2022) |
| HEK293T cells | EV‐depleted DMEM with 10% FBS | UC | TEM, NTA, WB | 1, 2, 3, 4 | 50% | 61% | (Cai et al. 2024) |
| SH‐SY5Y cells | EV‐depleted medium containing 1% FBS | UC | TEM, NTA, WB | 1, 2 | 30% | 56% | (Évora et al. 2025) |
Note: Protein categories are defined as in MISEV2018 – (1) transmembrane or GPI‐anchored proteins associated with plasma membrane and/or endosomes (e.g., CD63); (2) cytosolic proteins recovered in EVs (e.g., ALIX); (3) major components of non‐EV co‐isolated structures (e.g., APOA1, albumin); and (4) transmembrane, lipid‐bound and soluble proteins associated with other intracellular compartments than plasma membrane/endosome (e.g., calnexin).
Abbreviations: bFGF, basic fibroblast growth factor; BSA, bovine serum albumin; Char., characterization; Char. Rep., reporting adherence to MISEV characterization; Coll. Rep., reporting adherence to MISEV collection and pre‐processing; DMEM, Dulbecco's modified eagle medium; EGF, epidermal growth factor; FACs, fluorescence‐activated cell sorting; HPL, human platelet lysate; hPSC, human pluripotent stem cell; MEM, minimum essential medium; PDGFbb, platelet‐derived growth factor subunit B‐B; Ref., reference.
3.4. In Vitro Models for Testing EV Functionality in AD
Of the papers included, 63% (n = 56) tested the functionality of EVs in AD cell models in vitro. Many of these (68%, n = 38) performed in vitro testing to identify specific cellular mechanisms before or after preliminary testing in vivo, and the remaining 20% (n = 18) assessed EV treatment only in vitro (Figure 5). For in vitro models, induction‐based approaches were most common to induce AD phenotypes, including the use of Aβ (43%, n = 24), lipopolysaccharides (LPS) (7%, n = 4), a combination of LPS and Aβ (7%, n = 4), a combination of LPS and nigericin (2%, n = 1) (Lin, Huang, et al. 2024), a combination of okadaic acid and Aβ (2%, n = 1) (Lin, Hsu, et al. 2024), hydrogen peroxide (H2O2) (2%, n = 1) (Évora et al. 2025), kainic acid (2%, n = 1) (Venugopal et al. 2018), L‐glutamate (2%, n = 1) (Pourhadi et al. 2024), tau (2%, n = 1) (Krishnan et al. 2025), or oxygen‐glucose deprivation (OGD) (4%, n = 2) (Ma et al. 2022; Wang et al. 2025). One study investigated the effect of EVs on both SH‐SY5Y cells expressing the APP Swedish mutation (APPswe) and primary hippocampal neurons treated with Aβ42 (Wang, Liu, et al. 2021). Another study investigated the ability of EVs to reduce neuronal damage that was induced by either L‐glutamate or Aβ42 oligomers (Ma et al. 2020). In another study, human CMECs were pretreated with both Aβ42 and verapamil to inhibit P‐gp in normal human CMECs to assess BBB integrity (Pan et al. 2020). One study used both Aβ oligomers and LPS to induce AD phenotypes in a co‐culture model, where SH‐SY5Y cells were induced with Aβ42 fibrils, and BV‐2 microglia were induced with LPS before combining in co‐culture (Zavatti et al. 2022). Similarly, one study first primed BV‐2 microglia with LPS followed by incubation with Aβ oligomers (Kaniowska et al. 2022). In others, EVs were tested in both LPS‐ and Aβ‐induced microglia (Markoutsa et al. 2022; Zhao et al. 2023). LPS was used to induce HM3C microglia (Silva et al. 2025; Khan et al. 2023; Garcia‐Contreras and Thakor 2021) and iNSC‐derived astrocytes (Chen, Lan, et al. 2024). Okadaic acid, alongside the incubation of SH‐SY5Y cells with Aβ42, was used to trigger AD tauopathy in addition to Aβ deposition (Lin, Hsu, et al. 2024), and OGD models were generated with primary neurons through incubation in a hypoxic chamber (Ma et al. 2022; Wang et al. 2025). L‐glutamate was used to induce neurotoxicity in cultured C6 cells (Pourhadi et al. 2024), while kainic acid was used to generate a hippocampal neurodegeneration model (Venugopal et al. 2018). One study isolated CMECs and NSCs from streptozotocin (STZ)‐induced rats to perform ex vivo testing on BBB permeability and neurogenesis, respectively (Zhang et al. 2022).
FIGURE 5.
Summary of AD models used in EV‐based treatment research. (A) The number of studies that were presented in vitro only (pink), in vivo only (animal models, turquoise), in vitro and in vivo (animal models, blue), or clinical trials in humans (green). (B) The percentage of methods used to model AD in vitro. (C) The percentage of methods used to model AD in vivo in animals.
A total of 31 models used Aβ either on its own or in combination to induce AD phenotypes in vitro. Varying lengths of Aβ were used, where 68% (n = 20) used Aβ42, 10% (n = 3) used Aβ40, 13% (n = 4) used Aβ25–35, and the remaining 13% (n = 4) generally described using Aβ without specifying the peptide. Fifty‐two percent of studies (n = 16) using Aβ‐induced AD models in vitro used solubilized peptides, while 45% (n = 14) described creating and using Aβ oligomers or aggregates, and 3% (n = 1) described Aβ fibrils (Zavatti et al. 2022). The range of Aβ dose in treatments varied from 250 nM of oligomerized Aβ42 (Huber et al. 2022) to 25 µM of soluble Aβ monomers (Wang, Yang, et al. 2023; Chen, Deng, et al. 2024; Zhai et al. 2021). Of the 27 studies reporting molar units, the median dose of Aβ was 10 µM, with an interquartile range of 2.25–20 µM. Aβ induction with oligomers was at a lower dose compared to induction with peptides (mean of 6.1 ± 7.8 µM compared to 13.7 ± 8.4 µM, respectively, p = 0.04). The incubation time of the cells with Aβ ranged from 2 to72 h.
Several studies tested EVs in cell models with genetic mutations that predispose them to AD, either by transfection of cells or by isolating primary cells from AD mouse models. Cells with APP mutations were used in 9% of in vitro models (n = 5), while the APP/PS1 mutation was investigated in one study that cultured cortical neurons from the cerebral tissues of APP/PS1 mice (Wang et al. 2018). For APP mutations, SH‐SY5Y cells were either transfected with the Swedish KM670/671NL double mutation (APPswe) (Yang et al. 2020; Wang, Liu, et al. 2021) or with both the Swedish K670N/M671L and London V717I FAD mutations (Chen et al. 2021). Another study overexpressed APP in SH‐SY5Y cells, but the mutation was not specified (Sun et al. 2025). One study used N2a cells with a Swedish mutant gene, APP695, as an AD cell model (Chen et al. 2023). Two percent of studies (n = 2) used cell cultures derived from 5xFAD mice to test EV efficacy, isolating primary CMECs (Liu et al. 2020) or neurons (Gatti et al. 2020). Another study tested the effects of EVs on primary neurospheres extracted from TG2576 mice (Lee et al. 2018). One study evaluated EVs in Down syndrome (Campbell et al. 2023), which is a significant risk factor for developing AD‐like dementia. The Down syndrome model was generated from both male and female iPSC lines to generate spheroids, which were used to test EV treatments (Campbell et al. 2023). Seven percent of in vitro models (n = 4) included neurodegeneration not specific to AD, which was accompanied by AD‐relevant animal studies (Hu et al. 2020; Yu et al. 2021; Losurdo et al. 2020; Khan et al. 2023). For example, D‐galactose was used to produce an ageing model in conjunction with a vascular dementia mouse model (Hu et al. 2020), and inflammatory stimuli (TNF‐α and IFN‐γ) were used to induce neurodegeneration (Losurdo et al. 2020). Another study evaluated NSC‐EVs overexpressing MIAT in NSCs derived from MIAT knockdown rats, with the loss of MIAT being a key factor in neurovascular diseases that can lead to microvascular diseases, neurodegeneration, and AD (Qi et al. 2021). One study evaluated N2a cells treated with two differently transfected EV sources to compare the effect on Aβ40 clearance (Yu et al. 2021).
For delivery of EVs in vitro, EV doses varied in units with reported concentrations ranging from 0.05 to 5000 µg protein/mL, 3 × 107 to 2 × 1010 particles/mL, or total amounts of 1 to 90 µg protein, 1 × 106 to 1 × 1010 particles, or 5 to 15 µL per well or cell replicate (see Table S1). Most commonly, doses were reported as protein concentration in µg/mL of medium (22 studies, median 40 µg/mL, interquartile range 22–100 µg/mL). Where a dose was reported, it was not always clear what the dose was added to (e.g., number of cells, volume of culture medium), making repeatability and comparison across studies difficult. Some studies did not specify EV concentration, stating that cells were treated with Aβ25–35 and NSC‐derived EVs (Li, Chen, et al. 2024), Aβ42 with or without EVs (Xu et al. 2022), pre‐incubated with EVs (Pan et al. 2020), or that the models were tested in the presence of EVs (Markoutsa et al. 2022; Ma et al. 2022). EV exposure ranged from 6 h to 14 days, with most studies incubating the EVs and cells for 24 h prior to analysis (21/30 of studies that reported a single time point, 70%).
Various assays were used to assess the efficacy of EV treatments in vitro, including Aβ or p‐tau deposition, apoptosis, neuroinflammation, neurogenesis, BBB functionality, oxidative stress, or mitochondrial function. Overall, Aβ deposition was most often assessed (38%, n = 21), which was detected through enzyme‐linked immunosorbent assay (ELISA), WB, and/or immunocytochemistry (ICC). P‐tau concentration was measured in 13% of in vitro studies (n = 7), similarly through ELISA, WB, or ICC. To test the protective effects of EVs, assays such as Live/Dead (4%, n = 2), MTT (13%, n = 7), or the Cell Counting Kit‐8 (CCK‐8) (20%, n = 11) were used after treatment with Aβ or other neurotoxins. Cell apoptosis was detected through staining with Annexin V–FITC and propidium iodide (23%, n = 13) or TUNEL (4%, n = 2) (Dou et al. 2021; Hu et al. 2025), lactate dehydrogenase (LDH) (4%, n = 2) (Zhai et al. 2021; Venugopal et al. 2018), or measuring apoptosis‐related proteins (e.g., caspases, BAX, BCL2, Akt) through WB or ICC (32%, n = 18). Neuroinflammation was assessed in 34% of in vitro studies (n = 19), most often through ELISA for inflammatory and anti‐inflammatory factors in the culture medium (e.g., TNF‐α, IL‐1β, IL‐6, IL‐10, IL‐4), or through WB, RT‐qPCR, or ICC to measure expression of inflammatory factors or glial activation markers (e.g., IBA1). Neurogenesis was investigated in vitro by measuring neurite length and branch points (Évora et al. 2025; Wang et al. 2025; Gao et al. 2023; Ma et al. 2020), axonal length (Marzano et al. 2019), gene and protein expression related to synaptic plasticity (Évora et al. 2025; Ma et al. 2022; Wang et al. 2025; Chen et al. 2021), or intracellular calcium oscillations using calcium indicator Fluo‐8 AM (Wang, Liu, et al. 2021). BBB permeability was measured in one study through a Transwell permeability assay measuring sodium fluorescein leakage through a CMEC monolayer (Liu et al. 2020). A similar assay was used to evaluate Aβ clearance through the CMEC monolayer (Pan et al. 2020). Reactive oxygen species (ROS) are a group of highly reactive, oxygen‐containing molecules that can lead to oxidative stress and cell damage and are observed in neurodegenerative diseases such as AD (Marzano et al. 2019), with abundance estimated using dichloro‐dihydro‐fluorescein diacetate (DCFH‐DA) assays (23%, n = 13). Other measures of mitochondrial function were used, including measuring mitochondrial enzyme activities of superoxide dismutase (SOD) (Venugopal et al. 2018), malonaldehyde (MDA) (Rao et al. 2025), or catalase (de Godoy et al. 2018; Bodart‐Santos et al. 2019), or mitochondrial membrane potential through the JC‐1 assay (Xu et al. 2022; Li et al. 2022; Rao et al. 2025; Li, Chen, et al. 2024).
3.5. In Vivo Animal and Human Clinical Models for Testing EV Functionality in AD
In vivo animal models were used in 79% of included papers (n = 70), with 43% (n = 38) being investigated alongside in vitro models. Varying types of animal models were used, including transgenic mice with mutations in APP/PS1 (30%, n = 21), 5xFAD (14%, n = 10), APP (3%, n = 2) (Izadpanah et al. 2020; Feng et al. 2020), 3xTg‐AD (3%, n = 2) (Lin, Hsu, et al. 2024; Losurdo et al. 2020), or J20 (1%, n = 1) (Chen et al. 2021) (Figure 5). J20 mice obtain the Swedish and Indiana (V717F) mutations of the APP protein (Chen et al. 2021), while 3xTg‐AD mice express three mutant human transgenes (PS1M146V, APPSwe, and tauP301L) (Losurdo et al. 2020). Induction‐based animal models were also explored, such as induction with STZ (13%, n = 9), Aβ (13%, n = 9), aluminium chloride (AlCl3) (4%, n = 3), LPS (4%, n = 3), tau (1%, n = 1) (Krishnan et al. 2025), or sodium azide (NaN3) (1%, n = 1) (Amer et al. 2024). Three percent (n = 2) induced neuroinflammation in transgenic 5xFAD and APP models with LPS (Markoutsa et al. 2022; Feng et al. 2020). Induction‐based models often were administered with multiple doses over several days/weeks. For example, in one study, AlCl3 treatment was given once daily for 8 weeks to generate the AD model (Ebrahim et al. 2024), while in another NaN3 was added at a concentration of 12.5 mg/kg for the first 5 days and a lower dose of 10 mg/kg for the following 9 days (Amer et al. 2024). For models of vascular dementia, bilateral common carotid artery occlusion (BCCAO) was investigated in 7% of studies (n = 5). Further, 3% (n = 2) used animal models that had undergone olfactory bulbectomy (OBE) as a model of sporadic AD (Poltavtseva et al. 2021; Zhdanova et al. 2021). Finally, an aged model was analysed using 20‐month‐old mice (1%, n = 1) (Yu et al. 2021).
One clinical study was included, which was published in 2023. Human participants were treated with allogeneic ADSC‐EVs who were enrolled in the Memory Clinic of Neurology Department (Xie et al. 2023). A total of nine participants received a low, medium, or high dose of EVs at quantities of 2 × 108, 4 × 108, or 8 × 108 particles in 1 mL of saline intranasally twice per week for 12 weeks. The participants were certified for participation through informed consent, being 50 years or older, being diagnosed with mild or moderate AD, and not being treated with cognitive‐enhancing or therapeutic drugs. The safety of the EVs was assessed through vital signs, laboratory testing, and adverse or severe symptoms over 16 weeks. The study evaluated EV efficacy through measuring cognitive function and daily activities of the participants. The study reported increases in cognitive scores and reduced hippocampal volume loss for those who received the medium EV dose, with no significant differences in Aβ or p‐tau deposition across groups.
EV delivery was predominantly given intravenously (51% of in vivo studies, n = 36) in both human clinical testing and preclinical animal models. EVs were delivered by intranasal injection in 20% (n = 14) and through stereotaxic intracerebroventricular injection in 13% (n = 9). Other studies specified intraperitoneal (6%, n = 4) (Ebrahim et al. 2024; Reza‐Zaldivar et al. 2019; Hao et al. 2022; Bashirrohelleh et al. 2025), intrahippocampal (3%, n = 2) (Jahangard et al. 2020; Yang et al. 2020), intracerebral (3%, n = 2) (Silva et al. 2025; Elia et al. 2019), intracranial (1%, n = 1) (Zhao et al. 2019), retro‐orbital (1%, n = 1) (Apodaca et al. 2021), or intracisternal (1%, n = 1) (Lin, Huang, et al. 2024) delivery. Only 1% (n = 1) did not clarify the route of administration (Liu, Jin, et al. 2022). Similar to in vitro dosing, EV doses varied in concentration and reporting units for animal models (see Table S1). Doses measured by protein quantity ranged from 0.07 to 2000 µg (median 50 µg, interquartile range of 21 to 100 µg) at a concentration of 7 to 150,000 µg/mL (median 500 µg/mL, inter‐quartile range of 250 to 3750 µg/mL). Doses measured by particle concentration ranged from 2.25 × 107 to 5 × 1011 particles (median 1 × 1010 particles, inter‐quartile range of 1.8 × 109 to 2.1 × 1010). Some studies provided only a volume measurement, with total volumes injected ranging from 2 to 200 µL in mice, and 100 to 500 µL in rats. Doses were also reported based on the size of animals, from 100 ng/kg to 75 mg/kg.
The efficacy of EV treatment in animal models was primarily evaluated through behavioural tests such as the Morris water maze (70%, n = 49), novel object recognition (21%, n = 15), Y‐maze (13%, n = 9), open field (9%, n = 6), and passive avoidance (7%, n = 5). Other behavioural tests included fear conditioning (Zhu et al. 2023; Gao et al. 2023; Hou et al. 2023), Barnes maze (Cone et al. 2021; Jahangard et al. 2020; Huang et al. 2024), elevated pulse maze (Liu, Fan, et al. 2022; Apodaca et al. 2021), odour recognition (Zhang et al. 2022; Ding et al. 2022), rotarod (Li, Zhang, et al. 2024), voluntary activity (Li, Zhang, et al. 2024), nest building (Wang, Yang, et al. 2023), tail suspension (Liu, Fan, et al. 2022), and social interaction (Zhang et al. 2022). One study also investigated metrics related to neurological functional scores to quantify movement behaviour (Wang, Yi, et al. 2023). Another study used magnetic resonance imaging (MRI) to assess cerebral blood flow (CBF), BBB permeability, and relaxation time in STZ‐ and nicotinamide (NTM)‐induced Wistar rats (Ding et al. 2022). Label‐based methods were used to detect target cells and molecules following brain sectioning, such as immunohistochemistry (IHC) or immunofluorescence (IF) for Aβ, p‐tau, synaptic proteins, and other markers of cell state (81%, n = 57); Nissl staining to identify Aβ plaques and neurodegeneration (19%, n = 13); Thioflavin S to identify Aβ plaques (19%, n = 13); haematoxylin and eosin (H&E) staining (17%, n = 12); and Golgi‐Cox staining for neuron morphology (11%, n = 8). WB (40%, n = 28), RT‐qPCR (23%, n = 16), and ELISA (41%, n = 29) were used to detect Aβ, functional proteins, apoptotic and oxidative stress‐related proteins, and inflammatory markers in tissue lysates. Electrophysiological signals in brain slices were measured using whole‐cell patch clamps in three studies (Wang, Liu, et al. 2021; Krishnan et al. 2025; Chen, Lan, et al. 2024). Further, one study identified molecular energy indicators within cortical tissues using high‐performance liquid chromatography (Amer et al. 2024).
EV uptake within animal models was visualized through labelling techniques to track their localization and BBB penetration in vivo. Various fluorescent dyes were used including lipophilic membrane dyes: PKH26 (24%, n = 17), Dil (9%, n = 6), DiR (9%, n = 6), DiO (4%, n = 3), PKH67 (Silva et al. 2025), DiD (Huang et al. 2024), and CellTracker CM‐DiI (Yu et al. 2021); iodine radioisotopes: 125I (Ma et al. 2020); and other fluorophores tagged to EV‐specific components or through parent cell labelling: CD63‐mCherry (Lin, Huang, et al. 2024), Cy5.5 (Xu et al. 2022), Cy7 (Sun et al. 2025), C5 Maleimide‐Alexa 594 (Wang, Liu, et al. 2021), or CD63‐GFP (Zhang et al. 2022). Iron oxide was used to label EVs in one study, providing another method to identify EVs in tissues (Ebrahim et al. 2024). Several of the included studies provide evidence of EVs localizing to the brain and penetrating through the BBB, with EVs compared to dye‐only controls (Cone et al. 2021; Dou et al. 2021; Poltavtseva et al. 2021; Markoutsa et al. 2022; Xu et al. 2022; Lin, Hsu, et al. 2024).
3.6. Therapeutic Effects
EVs isolated from various cell cultures have demonstrated a broad range of therapeutic effects in AD models, summarized in Table 3. To date, 62 publications have reported that EV treatment improves cognitive deficits, measured by improvements in spatial memory and learning. These improvements have been attributed to a reduction in Aβ plaque formation and/or soluble Aβ (n = 45), reduced p‐tau (n = 13), improved cell viability and reduced apoptosis (n = 35), reduced neuroinflammation measured by a reduction in pro‐inflammatory cytokines, genes, or microglial and/or astrocyte activation (n = 41), an improvement in BBB integrity measured by BBB permeability or expression of tight junction proteins (n = 3), increased neurogenesis measured by an increase in dendritic spine density or length, synaptic density, synaptic plasticity‐associated genes or neurite formation (n = 31), and enhanced mitochondrial function measured by a reduction in ROS or other markers of oxidative stress (n = 20), or through mitochondrial membrane potential, calcium transients, increased ATP production, and altered mitochondrial function‐related gene expression (n = 7).
TABLE 3.
Cell types with evidence of producing EVs with therapeutic effects for AD, including any modifications done to engineer the EVs, the AD model with dose and effects reported, and the overall adherence to the MISEV guidelines as quantified by Table 1 (Rep.).
| EV source | Modification | Model (dose) | Effect | Rep. | Ref. |
|---|---|---|---|---|---|
| BMSCs | STZ‐induced rats (0.5 µg in 2 µL daily for 5 days) | Improved learning and memory, inhibited oxidative stress, increased synaptic density | 40% | (Nakano et al. 2016) | |
| Kainic acid‐induced H3 hippocampal cells (0.05, 0.1 0.5, or 1 µg/mL) | Reduced apoptosis, necrosis and oxidative stress | 38% | (Venugopal et al. 2018) | ||
| APP/PS1 primary cortical neurons (100 µg/mL), APP/PS1 mice (100 µg in 5 µL every 2 days for 2 weeks) | Reduced Aβ‐induced iNOS expression and impairment of synaptic transmission, improved cognitive function | 20% | (Wang et al. 2018) | ||
| Aβ‐induced hippocampal neurons (8 × 107 or 2.4 × 108 particles) | Protected neurons from Aβ‐induced oxidative stress and synapse damage | 55% | (de Godoy et al. 2018) | ||
| APP/PS1 mice (4 µL at 5.6 µg/µL, ∼1 × 109 particles) | Reduced Aβ and dystrophic neurites | 53% | (Elia et al. 2019) | ||
| EVs loaded with RVG | APP/PS1 mice (5 × 1011 particles in 100 µL monthly for 4 months) | Reduced Aβ, decreased astrocyte activation, improved cognitive function, normalized levels of inflammatory mediators | 18% | (Cui et al. 2019) | |
| STZ‐induced C57BL/6 mice (0.5 µg in 2 µL daily for 5 days) | Restored cognition | 33% | (Zhao et al. 2019) | ||
| Inflammatory primed | Pro‐inflammatory microglia (4.5 µg/mL), 3xTg‐AD mice (100 µL total—5 µL per nostril at 300 µg/mL, 15 × 109 particles twice) | Polarized microglia to M2 phenotype, decreased microglia activation, increased dendritic spine density | 45% | (Losurdo et al. 2020) | |
| AlCl3‐induced ovariectomized albino rats (100 µg in 1 mL) | Restored taste bud features: improved synaptophysin immunoreactivity and nerve fibre density | 23% | (Hassan et al. 2020) | ||
| EVs loaded with NEP | APP‐transfected Wistar rats (25 µL for 2 weeks) | Reduced expression of inflammatory and apoptotic genes. NEP‐loaded EVs improved cognition, reduced apoptosis and Aβ | 58% | (Izadpanah et al. 2020) | |
| Transfected with miR‐29b | Aβ‐induced Wistar rats (10 µg) | Improved spatial learning and memory | 25% | (Jahangard et al. 2020) | |
| Aβ42‐induced primary hippocampal neurons (30 µg in 100 µL), Aβ42‐induced SD rats (30 µg in 100 µL monthly for 2 months) | Reduced Aβ, apoptosis and inflammation, increased neuron viability, improved cognitive function | 58% | (Sha et al. 2021) | ||
| Aβ42‐induced SH‐SY5Y cells (30 µg/mL) | Increased viability, suppressed apoptosis and inflammation | 48% | (Xiong et al. 2021) | ||
| 3D spheroids | 5xFAD mice (20 × 108 particles in 5 µL per nostril every 4 days for 2–4 months) | Improved cognition, reduced Aβ and GFAP expression | 65% | (Cone et al. 2021) | |
| APP/PS1 mice (50 µg in 80 µL biweekly for 16 weeks) | Improved spatial learning and memory, reduced Aβ, enhanced expression of NeuN and NEP, reduced expression of BACE1 and PS1 via SphK/S1P | 35% | (Wang and Yang 2021) | ||
| LPS‐induced AD rats (0.2 mg/kg) | Attenuated inflammation and oxidative damage, reduced apoptosis, accelerated antioxidant capacity, improved cognitive function, reduced Aβ and p‐tau | 15% | (Aboulhoda et al. 2021) | ||
| Transfected with SHP2 | Aβ42‐induced SH‐SY5Y (u/k), Aβ‐induced C57BL/6 mice (100 µg every 2 days for 2 weeks) | Induced mitophagy, alleviated neuronal damage, reduced ROS generation, inflammatory response and Aβ, ameliorated cognitive decline, improved synaptic plasticity, relieved lipid peroxidation | 35% | (Xu et al. 2022) | |
| LPS‐ and Aβ‐induced BV‐2 microglia (8 particles per BV‐2 cell) | Murine BMSC‐EVs, but not human ADSC‐EVs prevented upregulation of inflammatory factors in response to Aβ, regulated CD36 and CD206 expression | 43% | (Kaniowska et al. 2022) | ||
| Transfected with miR‐132‐3pp | OGD‐induced neurons (u/k), BCCAO‐induced rats (1 × 1010 particles in 100 µL weekly for 21 days) | Improved cognitive function, neuron number, synaptic plasticity and dendritic spine density, reduced Aβ and p‐tau, reduced apoptosis, increased neurite elongation and branching | 30% | (Ma et al. 2022) | |
| STZ‐induced C57BL/6 mice (0.5 µg in 2 µL daily for 5 days) | Mitigated AD‐like behaviour, reduced expression of inflammatory factors, inhibited microglial activation, reduced Aβ and p‐tau, promoted neuron regeneration | 40% | (Liu, Fan, et al. 2022) | ||
| Down syndrome iPSC cortical spheroids (u/k weekly for 4 weeks) | Rescued neuron production, decreased levels of Aβ and p‐tau, reduced cell death, reduced GFAP expression | 35% | (Campbell et al. 2023) | ||
| EVs loaded with AM1241 | APP/PS1 mice (6 mg/kg AM1241 every 2 days for 2 weeks) | Restored cognitive ability, enhanced neuronal regeneration, reduced Aβ, reduced apoptosis | 33% | (Zhu et al. 2023) | |
| AlCl3‐induced albino rats (0.5 mL at 100 µg/mL for 4 weeks) | Improved memory, reduced Aβ and p‐tau, promoted neurogenesis, enhanced synaptic function, mitigated astrogliosis, inhibited neuroinflammation | 30% | (Ebrahim et al. 2024) | ||
| LPS‐induced male albino rats (100 µg in 1 mL) | Improved cognitive and locomotor deficits, suppressed serum MDA, hippocampal Aβ, BAX and TNF‐α, restored hippocampal pyramidal cells to control levels (only seen for EVs and not MSCs) | 13% | (Hamed et al. 2025) | ||
| Transfected with miR‐214‐3p | Aβ42‐induced SD rats (100 µL for 5 days) | Reduced neuronal apoptosis and oxidative stress (reduced levels of MDA and ROS), improved behaviour, increased activity of antioxidants GSH‐pX and SOD | 48% | (Zhang, 2025) | |
| Incubated with curcumin | STZ‐induced Wistar rats (30 µg/body every week for 30 days) | Improved memory and learning, reduced neuronal degeneration, inflammatory response and APP expression | 25% | (Bashirrohelleh et al. 2025) | |
| AlCl3‐induced Wistar rats (2–5 injections at 150 µg/µL) | Prevented loss of brain weight, improved spatial memory and passive avoidance learning, reduced expression of APP, protected against Aβ plaque deposition | 35% | (Sadeghi et al. 2025) | ||
| ADSCs | NSCs from TG2576 AD mice (200 µg/mL) | Reduced Aβ, ameliorated Aβ‐induced neuronal death, promoted neurite growth | 28% | (Lee et al. 2018) | |
| Aβ42‐induced primary neurons (0.05–5 µg/mL), APP/PS1 mice (10 µL at 1 mg/kg every 2 days for 2 weeks) | Alleviated Aβ or glutamate‐induced neuronal damage, ameliorated neurologic damage, increased neurogenesis, reduced Aβ and microglial activation, rescued memory deficits | 50% | (Ma et al. 2020) | ||
| Transfected with miR‐22 | Aβ42‐induced PC12 cells (15 µg/mL), APP/PS1 mice (50 µL at 100 µg/mL weekly for 30 days) | Inhibited neuron damage, improved nerve function and motor ability, suppressed neuroinflammation | 35% | (Zhai et al. 2021) | |
| LPS‐induced HMC3 microglia (50, 20, or 10 µg/mL) | Prevented upregulation of iNOS, inhibited secretion of proinflammatory factors, promoted expression of anti‐inflammatory mediators | 45% | (Garcia‐Contreras and Thakor 2021) | ||
| Hypoxia | APP/PS1 mice (EVs from 1 × 109 ADSCs in 100 µL monthly for 2 months) | Improved cognitive function, decreased neuronal damage, shifted hippocampal microglia from M1 to M2, reduced expression of inflammatory factors, suppressed apoptosis | 38% | (Liu, Jin, et al. 2022) | |
| Incubated with coenzyme Q10 | STZ‐induced Wistar rats (u/k) | Improved memory impairment, increased BDNF and SOX2, increased neuron density | 20% | (Sheykhhasan et al. 2022) | |
| Hydrogel delivery | 5xFAD mice (10 µL with 1 mg/kg weekly for 5 weeks) | Alleviated neuronal damage, promoted neurogenesis, rescued memory deficits | 45% | (Huang et al. 2024) | |
| Phase I/II clinical trial in patients with mild to moderate AD (2 × 108 to 8 × 108 particles in 1 mL twice/week for 12 weeks) | Improved cognitive scores, reduced hippocampal atrophy, no adverse events, not effective for those with moderate‐severe AD | 55% | (Xie et al. 2023) | ||
| Turbulent vortex flow, EVs loaded with DNZ | LPS‐induced HMC3 microglia (5 × 108 to 5 × 109 particles/cm2), LPS‐induced zebrafish (10 µM DNZ) | Reduced production of inflammatory factors, DNZ‐loaded EVs reduced microglia migration and AChE activity | 50% | (Silva et al. 2025) | |
| LPS‐induced BV‐2 microglia (20 µg) | Prevented M1 polarization | 48% | (Li, Ye, et al. 2024) | ||
| UMSCs | Aβ25‐35‐induced BV‐2 microglia (30 µg/mL), APP/PS1 mice (30 µg in 100 µL every 2 weeks for 4 doses) | Reduced microglia activation, improved behavioural performance, reduced Aβ | 33% | (Ding et al. 2018) | |
| 3D culture | APPswe‐transfected SH‐SY5Y cells (2 µg per 2 × 105 cells), APP/PS1 mice (2 mg/mL at 0.25 µL/h for 14 days) | Reduced Aβ, upregulated ADAM10, reduced BACE1 expression, improved spatial learning and memory, inhibited inflammation and oxidative stress | 40% | (Yang et al. 2020) | |
| LPS‐induced APP/PS1 mice (2 × 109 particles) | Mitigated inflammatory response, alleviated trained‐immunity‐induced increased Aβ load | 38% | (Feng et al. 2020) | ||
| Aβ42‐induced primary hippocampal neurons (10 µg/mL), SH‐SY5Y‐APP cells (10 µg/mL), APP/PS1 mice (50 µg in 150 µL) | Improved cognitive deficits, reduced Aβ and neuronal loss, repaired neuronal morphology, restored neuronal excitability and mitochondrial changes | 53% | (Wang, Liu, et al. 2021) | ||
| Aβ40‐induced SH‐SY5Y cells (2 µg per 2 × 105 cells), 5xFAD mice (1 × 109 in 100 µL monthly for 3 months) | Enhanced proliferation and migration, improved cognitive function, reversed hippocampal neuron apoptosis, reduced Aβ. Effects were mitigated when mice could obtain AD gut microbiota through faeces, while antibiotic modulation of gut microbiota enhanced therapeutic effects of EVs | 40% | (Hou et al. 2023) | ||
| BCCAO‐induced rats (100 ng/kg for 7 days) | Alleviated neurological damage and impairment, inhibited M1 polarization, inflammation and oxidative stress | 53% | (Wang, Yi, et al. 2023) | ||
| Transfected with miR‐211‐5p inhibitor | Aβ40‐induced SH‐SY5Y cells (20 µg) | Reduced apoptosis, restored migration | 38% | (Chen, Huang, et al. 2024) | |
| 3D spheroids | L‐glutamate‐induced C6 cells in 2D and 3D (0.7 or 7 µg/mL), STZ‐induced Wistar rats (10 µL at 7 or 70 µg/mL per nostril 5× over 2 weeks) | Increased cell viability, reduced Aβ, improved learning and memory, increased expression of neuroplasticity proteins | 33% | (Pourhadi et al. 2024) | |
| APP/PS1 mice (40 µL at 2.5 µg/µL every 5 days for total of 6 injections) | Improved spatial learning and memory, normalized levels of AP2A1 and AP2B1 | 40% | (Li, Zhang, et al. 2024) | ||
| EVs loaded with BACE1 siR and berberine | APP‐overexpressing SH‐SY5Y cells (u/k), BV‐2 microglia (u/k), 5xFAD mice (20 µL at 1 mg/mL twice weekly for 4 weeks) | Reduced Aβ, inflammatory cytokine production, and ROS, increased expression of NeuN and Nestin; promoted polarization of microglia from M1 to M2; improved spatial learning and memory | 20% | (Sun et al. 2025) | |
| WJMSCs | Aβ‐induced hippocampal cells (6.1 × 107 or 1.8 × 108 particles) | Protected neurons from oxidative stress and synapse damage | 55% | (Bodart‐Santos et al. 2019) | |
| APP‐overexpressed differentiated SH‐SY5Y cells (50 µg twice per week), J20 AD mice (50 µg weekly for 4 weeks) | Reduced Aβ, restored whole‐brain regional glucose metabolism, improved cognitive function, inhibited astrocyte activation, upregulated neuronal memory and synapse‐related genes | 50% | (Chen et al. 2021) | ||
| OBE‐induced mice (10 µL twice per week for 3 weeks) | Improved spatial memory | 28% | (Zhdanova et al. 2021) | ||
| OBE‐induced mice (108 particles in 10 µL twice per week for 3 weeks) | Prevented deterioration of spatial memory | 28% | (Poltavtseva et al. 2021) | ||
| AFMSCs | 5xFAD primary neurons (10 µg per 1 × 106 cells), Aβ42‐induced SH‐SY5Y cells (10 µg per 1 × 106 cells) | Reduced ROS, increased GSH, improved viability, reduced Aβ and p‐tau, increased number of neurites | 43% | (Gatti et al. 2020) | |
| LPS‐induced BV‐2 microglia, Aβ42‐induced SH‐SY6Y and BV‐2 co‐culture (0.5 × 109 particles/106 cells) | Prevented upregulation of pro‐inflammatory markers, inhibited oxidative stress and apoptosis | 43% | (Zavatti et al. 2022) | ||
| OMMSCs | Aβ42‐induced SH‐SY6Y cells (40 µg/mL), Aβ42‐induced C57BL/6 mice (1 mg/mL in 100 µL twice weekly for 4 weeks) | Reduced apoptosis, Aβ deposition, glial activation, inflammatory cytokine release, and expression of ER stress‐related proteins, enhanced object recognition, spatial learning and memory | 45% | (Hu et al. 2025) | |
| MSCs (source unspecified) | Aβ42‐induced C57BL/6 mice (10 µg in 2 µL) | Improved cognitive performance, enhanced neurogenesis | 23% | (Reza‐Zaldivar et al. 2019) | |
| Inflammatory primed | LPS‐ or Aβ‐induced HCMS microglia (2 × 109 particles/mL), LPS‐induced C57BL/6 mice (20 × 109 particles in 50 µL), 5xFAD mice (20 × 109 particles weekly for 10 weeks) | Reduced Aβ‐induced ROS, expression of inflammatory cytokines, astrocyte and glial activation Aβ deposition and demyelination, improved memory and learning | 48% | (Markoutsa et al. 2022) | |
| NaN3‐induced rats (100 µL) | Improved learning and memory, reduced Aβ and p‐tau, improved antioxidant and calcium levels, increased ATP production and COX‐IV expression | 20% | (Amer et al. 2024) | ||
| NSCs | Aβ42‐induced hippocampal neurons (u/k), Aβ‐induced C57Bl/6 mice (3 µL, at 109 particles/µL) | Abolished Aβ‐induced suppression of LTP and memory deficits, decreased Aβ binding to synapses and to neuronal processes | 43% | (Micci et al. 2019) | |
| APP/PS1 mice (200 µg in 10 µL twice per week for 4 weeks) | Rescued cognitive deficits, enhanced mitochondrial function and synaptic activity, reduced inflammatory response and oxidative stress, accelerated activation of SIRT1, did not alter Aβ | 45% | (Li et al. 2020) | ||
| Transfected with MIAT | NSCs with MIAT knockdown (u/k), BCCAO‐induced rats (100 µg at 3 × 108 particles/µL) | Increased cell viability, reduced apoptosis, oxidative stress, inflammation and Aβ, increased SOD levels, improved learning and memory | 18% | (Qi et al. 2021) | |
| 5xFAD mice (2.25 × 107 particles in 50 µL, one or two injections) | Improved memory, reduced anxiety, Aβ, microglial activation and inflammatory cytokine production, protected against synaptic loss | 23% | (Apodaca et al. 2021) | ||
| Heat shock | H2O2‐ or Aβ42‐induced neurons (2 µg/mL for H2O2‐induced, 4 mg/mL for Aβ‐induced) | Protected against oxidative stress and Aβ‐induced neurotoxicity | 48% | (Huber et al. 2022) | |
| SH‐SY5Y cells (5–15 µL per 0.3 × 106 cells), LPS‐induced HMC3 microglia (5–15 µL) | Reduced Aβ and p‐tau, increased viability, suppressed AD‐related mRNA expression, reduced pro‐inflammatory cytokine production | 43% | (Khan et al. 2023) | ||
| Aβ25‐35‐induced HT‐22 neurons (u/k), APP/PS1 GFAP‐Cre+ Green mice, SIRT1 knockout mice (200 µg in 10 µL twice per week for 4 weeks) | Preserved mitochondrial function, improved spatial learning and memory, did not alter Aβ, suppressed astrocyte activation, activated SIRT1 | 38% | (Li, Chen, et al. 2024) | ||
| Tau‐induced hippocampal neurons (1 × 106 particles), tau‐induced C57BL/6 mice (1 × 109 particles) | Prevented tau‐induced memory impairment and LTP suppression, reduced tau accumulation and internalization in synaptosomes | 43% | (Krishnan et al. 2025) | ||
| Transfected with miR‐132‐3p | OGD‐induced neurons (2 × 1010 particles/mL), BCCAO‐induced C57BL/6 mice (1 × 1010 in 12 µL weekly for three weeks) | Reduced tau phosphorylation, improved spatial learning and memory, increased dendritic and synaptic spine density, reduced neuronal loss and ROS overproduction | 28% | (Wang et al. 2025) | |
| iNSCs | CMECs from 5xFAD mice (30 µg/mL) | Reversed BBB leakage | 18% | (Liu et al. 2020) | |
| 5xFAD mice (200 µL at 0.5 µg/µL every 3 days for 1 month) | Improved cognitive function, reduced Aβ and p‐tau, increased dendritic length and spine density, reduced pro‐inflammatory cytokines, suppressed astrocyte activation | 45% | (Gao et al. 2023) | ||
| Aβ42‐induced iPSC‐derived microglia (6 × 109 particles), 5xFAD mice (30 × 109 particles in 100 µL weekly for 2 weeks) | Reduced microglial activation, Aβ and p‐tau, improved cognitive and mood function, reduced astrocyte hypertrophy, lowered expression of DAM genes | 53% | (Madhu et al. 2024) | ||
| LPS‐induced astrocytes from iNSCs (100 µg/mL), PBMNCs stimulated by PMA/PI (100 µg/mL), APP/PS1 mice (75 mg/kg every other day for 15 doses) | Inhibited secretion of proinflammatory cytokines, suppressed T cell activation and proliferation, improved cognitive dysfunction, suppressed activated astrocytes, restored neuron morphology | 30% | (Chen, Lan, et al. 2024) | ||
| Aβ42‐induced iPSC‐derived neurons (1.5 × 109, 3 × 109, or 6 × 109 particles) | Protected neurons from Aβ‐induced neurodegeneration, suppressed oxidative stress, improved mitochondrial membrane potential, prevented mitochondria loss, improved autophagy, reduced tau phosphorylation | 73% | (Rao et al. 2025) | ||
| iMSCs | LPS/nigericin‐induced BV‐2 microglia (200 µg/mL), STZ‐induced C57BL/6J mice (30 µg biweekly for 6 weeks) | Abolished inflammasome activation, alleviated neuroinflammation, decreased Aβ and neuronal apoptosis, mitigated cognitive dysfunction | 43% | (Lin, Huang, et al. 2024) | |
| iPSCs | Aβ42‐induced iNSC spheroids (50 µg/mL) | Enhanced proliferation, reduced apoptosis and oxidative stress. EVs from differentiated cortical neurons increased axonal length | 60% | (Marzano et al. 2019) | |
| ESCs | D‐gal‐induced senescent hippocampal NSCs (1 × 1010 particles/mL for 3 passages), BCCAO‐induced rats (1 × 1010 particles in 200 µL every 2 days for 1 week, then once per week) | Reversed cognitive deficits, ameliorated NSC senescence | 53% | (Hu et al. 2020) | |
| CMECs | Aβ42‐ and verapamil‐induced CMECs (u/k), Aβ42‐induced C57BL/6 mice (20 mg/mL, 100 µL/day for 6 days) | Facilitated clearance of Aβ | 35% | (Pan et al. 2020) | |
| NTM/STZ‐induced Wistar rats (1 × 1011 particles twice/week for 4 weeks), NSCs from STZ‐induced rats (3 × 107 particles/mL for 7 days) | Ameliorated cognitive deficits, reduced cerebral vascular damage, enhanced neurogenesis | 38% | (Zhang et al. 2022) | ||
| NTM/STZ‐induced rats (1 × 1011 particles twice/week for 4 weeks) | Increased CBF, reduced BBB permeability and cognitive deficits | 33% | (Ding et al. 2022) | ||
| Microglia | Aβ42‐induced HT‐22 neurons (90 µg per 30 × 104 cells), APP/PS1 mice (100 µg in 100 µL weekly for 2 months) | Increased cell viability, restored mitochondrial membrane potential, reduced ROS, decreased Aβ plaque deposition | 35% | (Li et al. 2022) | |
| Modulation by 1070 nm light | N2a/APP695swe cells (1 µg per 5 × 104 cells), 5xFAD mice (2.8 × 109 particles in 20 µL every 2 days for 2 months) | Protected against inflammation and Aβ accumulation, promoted neurite growth, improved spatial learning and memory, reduced Aβ, lowered inflammation, protected dendritic spine integrity and synaptic plasticity, alleviated ER stress | 43% | (Chen et al. 2023) | |
| Aβ40‐induced SH‐SY5Y cells (50 or 2.5 µg/mL), APP/PS1 mice (200 µg/kg daily for 4 weeks) | Microvesicles (not small EVs) rescued cell viability, prevented Aβ fibrillation, reduced Aβ, ameliorated cognitive impairment | 53% | (Wang, Yang, et al. 2023) | ||
| EVs loaded with berberine and palmatine | LPS‐ or Aβ25‐35‐induced microglia (u/k), APP/PS1 mice (10 mg/kg every 2 days for 21 days) | Reduced nitric oxide and inflammatory factors, microglial activation, number of plaques and neuroinflammation, improved cognitive function | 35% | (Zhao et al. 2023) | |
| Astrocytes | Ultrasound stimulated | Aβ42‐induced SH‐SY5Y cells (50 µg), APP/PS1 mice (60 µg in 200 µL) | Improved viability, reduced Aβ | 50% | (Deng et al. 2021) |
| Incubation with catalpol and tetramethylpyrazine | Aβ42‐induced HT‐22 neurons (100 µg/mL), APP/PS1 mice (50 µL at 100 µg/mL twice/week for 4 weeks) | Restored cell viability and expression of CDK5, improved cognitive function, alleviated neural damage | 28% | (Chen, Deng, et al. 2024) | |
| Neurons | GABA‐treated | Aβ42‐induced primary neurons (10–80 µg/mL), APP/PS1 mice (1 × 1010 particles/g every 2 days for 40 days) | Alleviated Aβ‐induced apoptosis, improved spatial memory deficits | 58% | (Dou et al. 2021) |
| HeLa cells | Transfected with RVG and CD10 | N2a cells (1 × 108–1 × 1010 particles), aged mice (150 µg in 150 µL) | Decreased intracellular and secreted Aβ40, reduced proinflammatory genes | 43% | (Yu et al. 2021) |
| Dendritic cells | Transfected with miR‐29b‐2 and CD47 | Aβ42‐ and okadaic acid‐induced SH‐SY5Y cells (3 × 108 particles/mL), 3xTg‐AD mice (2.5 × 1010 particles) | Inhibited PSEN1, reduced Aβ | 43% | (Lin, Hsu, et al. 2024) |
| Macrophages | Incubated with gemfibrozil | Aβ42‐induced BV‐2 microglia (5 mg/mL), Aβ42‐induced C57BL/6 mice (1 mL Exo‐gemfibrozil, 100 µg/mL gemfibrozil for 7 days) | Improved learning and memory, enhanced lysosomal activity of microglia, accelerated lysosome‐mediated clearance of Aβ, reduced neuronal damage | 28% | (Hao et al. 2022) |
| HEK293T cells | EVs loaded with RVG, Gap43 and Snap25 mRNA | 5xFAD mice (1 × 1010 particles in 0.2 mL every 3 days for 12 days) | Upregulated synaptic proteins and dendritic density, ameliorated cognitive impairment, no significant differences in Aβ | 48% | (Cai et al. 2024) |
| SH‐SY5Y cells | Transfected cells or EVs with miR‐124‐3p | H2O2‐induced triculture of APP‐SH‐SY5Y, HMC3 microglia, IM‐HA astrocytes (2.16 ± 1.41 × 107 particles/µg) | Reduced neuronal apoptosis and neurite deficits, caspase‐12 activation, and glial activation | 35% | (Évora et al. 2025) |
Abbreviations: BDNF, brain derived neural factor; ER, endoplasmic reticulum; GABA, gamma‐aminobutyric acid; NTM, nicotinamide; PMA/PI, phorbol 12‐myristate 13‐acetate/ionomycin; Rep, reporting adherence to MISEV; S1P, sphingosine‐1‐phosphate; SD, Sprague Dawley; u/k, unknown dose; VD, vascular dementia.
The positive effects of EVs have been attributed largely to encapsulated or associated miRs due to their ability to elicit changes in the gene expression of recipient cells, but have also been associated with specific lipids, proteins, and enzymes contained within or associated with the EVs. The breadth of proposed factors in the included studies is listed in Table 4, highlighting the multimodal function of EVs. To improve the efficacy of EVs as therapeutics for AD, several studies have transfected or overexpressed specific miRs into the parent cells, which have been effectively encapsulated into their secreted EVs (Évora et al. 2025; Qi et al. 2021; Ma et al. 2022; Jahangard et al. 2020; Chen, Huang, et al. 2024; Wang et al. 2025). Others have exposed the parent cells to stimuli in culture, such as low oxygen (Liu, Jin, et al. 2022), inflammation (Markoutsa et al. 2022), or temperatures exceeding physiological temperatures (i.e., heat shock) (Huber et al. 2022), which subsequently modified the EV cargo.
TABLE 4.
Proposed bioactive molecules encapsulated or associated with EVs.
| Proposed factor | Function | EV source | Model | Ref. |
|---|---|---|---|---|
| miR‐223‐3p | Targets NLRP3 to alleviate neuroinflammation | iMSCs | STZ‐induced sporadic AD mice | (Lin, Huang, et al. 2024) |
| miR‐17‐5p, miR‐18a‐5p, miR‐21‐5p, miR‐29a‐3p, and let‐7a‐5p | Inhibit mTORC1 activation, promote transcription factor EB nuclear translocation and lysosome resumption to reverse NSC senescence and cognitive decline | ESCs | BCCAO‐induced VD mice | (Hu et al. 2020) |
| miR‐125a, miR‐125b, and miR‐124 | Protect against synaptic loss, reduce microglial activation, regulate anxiety | NSCs | 5xFAD mice | (Apodaca et al. 2021) |
| Anti‐apoptotic and DNA repair proteins | Protect against oxidative stress and Aβ‐induced neurotoxicity | Heat‐shocked NSCs | H2O2‐ or Aβ42‐induced murine neuronal cells | (Huber et al. 2022) |
| MIAT | Inhibits miR‐34b‐5p and upregulates CALB1 to reduce hippocampal pathological features and improve cognitive function | NSCs | BCCAO‐induced VD rats | (Qi et al. 2021) |
| miR‐485, miR‐17 and miR‐322 | Reduce Aβ oligomer binding to the synapse | NSCs | Aβ‐induced C57Bl/6 mice | (Micci et al. 2019) |
| GDF‐15 | Upregulates NEP and IDE via activation of AKT/GSK‐3β/β‐catenin pathway to degrade Aβ42 | BMSCs | Aβ42‐induced SH‐SY5Y cells | (Xiong et al. 2021) |
| miR‐132‐3p (transfected) | Activates Ras/Akt/GSK‐3β pathway by downregulation of RASA1 to improve neuronal and synaptic dysfunction | BMSCs | BCCAO‐induced VD mice, OGD‐induced neurons | (Ma et al. 2022) |
| miR‐29b (transfected) | Restores Aβ‐induced cognitive deficits | BMSCs | Aβ‐induced Wistar rats | (Jahangard et al. 2020) |
| miR‐29c‐3p | Targets BACE1 and activates Wnt/β‐catenin pathway to restore Aβ‐induced cognitive deficits, inflammation, neuronal damage, and Aβ deposition | BMSCs | Aβ42‐induced SD rats and neurons | (Sha et al. 2021) |
| miR‐21, miR‐155, and miR‐126–3p, S1P | miRs influence PI3K/AKT/mTOR pathway via S1P signalling, to alleviate pathogenic inflammation and apoptosis, clear Aβ. S1P binds to Aβ to remove deposits and lower synaptic toxicity | BMSCs | AlCl3‐induced AD albino rats | (Ebrahim et al. 2024) |
| Circ‐Epc1 | Targets TREM2 and miR‐770‐3p to reduce neuronal damage and microglial activation | Hypoxic ADSCs | APP/PS1 mice | (Liu, Jin, et al. 2022) |
| miR‐211‐5p inhibitor (transfected) | Increases NEP expression to reduce apoptosis and inflammation | UMSCs | Aβ40‐induced SH‐SY5Y cells | (Chen, Huang, et al. 2024) |
| miR‐219a‐2‐3, miR‐10527, miR‐329, and miR‐578 | Target key genes on the TLR4 signalling pathway | Inflammatory primed MSCs | 5xFAD mice, LPS‐induced C57BL/6 mice, LPS‐ or Aβ‐induced HCMS microglia | (Markoutsa et al. 2022) |
| Catalase | Protects hippocampal neurons from Aβ‐induced damage | WJMSCs | Aβ‐induced hippocampal cells | (Bodart‐Santos et al. 2019) |
| miR‐7670‐3p | Targets ATF6 to ameliorate ER stress, attenuate inflammatory response, and protect dendritic spine integrity | BV‐2s | 5xFAD mice, N2a/APP695swe cells | (Chen et al. 2023) |
| miR‐132 | Exerts protective effect against Aβ toxicity | GABA‐neurons | APP/PS1 mice, Aβ42‐induced primary neurons | (Dou et al. 2021) |
| P‐gp | Facilitates cerebral clearance of Aβ | CMECs | Aβ42‐induced C57BL/6 mice, Aβ42‐ and verapamil‐induced CMECs | (Pan et al. 2020) |
| IL‐10 | Alleviates trained immunity | UMSCs | LPS‐induced APP/PS1 mice | (Feng et al. 2020) |
| miR‐214‐3p | Targets CD151 to enhance antioxidant capacity | BMSCs | Aβ42‐induced SD rats | (Zhang 2025) |
| miR‐124‐3p (transfected) | Regulates neuron–microglia paracrine signalling | SH‐SY5Ys | H2O2‐induced triculture of APP‐SH‐SY5Ys, HMC3 microglia, immortalized human astrocytes | (Évora et al. 2025) |
| miR‐132‐3p (transfected) | Targets RASA1 to promote neurite outgrowth | NSCs | OGD‐induced neurons, BCCAO‐induced C57BL/6 mice | (Wang et al. 2025) |
Abbreviations: ATF6, activating transcription factor 6; CALB1, calbindin 1; ER, endoplasmic reticulum; GABA, gamma‐aminobutyric acid; GDF‐15, growth differentiation factor‐15; IDE, insulinase; IL, interleukin; P‐gp, P‐glycoprotein; S1P, sphingosine‐1‐phosphate; SD, Sprague Dawley; TLR4, toll‐like receptor‐4; TREM2, triggering receptor expressed on myeloid cells 2; VD, vascular dementia.
3.7. Attribution of Effects to EVs
Appropriate isolation and characterization methods should be paired with proper controls to be able to attribute effects to EVs as opposed to components within the cell medium and/or other secreted factors. This is particularly important given the large number of studies using low‐ to medium‐specificity isolation methods and that EVs can be identified within CM prior to their isolation. In addition, several studies appear to have isolated EVs directly from serum‐containing medium, which is well known to contain its own EVs. Of the 89 studies in this review, only two studies (2%) compared EVs to a control of CM and/or the supernatant remaining after EV isolation (Markoutsa et al. 2022; Sha et al. 2021). One study compared the effects of CM, EVs, and the supernatant remaining after EV isolation (termed soluble fraction) on LPS‐induced microglial activation, and found both CM and EVs had a substantial anti‐inflammatory effect, while the soluble fraction did not significantly alter microglial activation (Markoutsa et al. 2022). The second study included a CM condition compared to BMSC‐EVs in Aβ‐induced hippocampal neurons and a rat model and found that only BMSC‐EVs could reduce Aβ and neuroinflammation, increase neuron viability and neurogenesis, and improve cognitive function (Sha et al. 2021). None of the included studies evaluated the contribution of components within fresh or nonconditioned medium in functional studies.
3.8. Methodological and Reporting Quality
Individual study adherence to reporting guidelines was assessed using 20 relevant mandatory items on the MISEV2018 checklist with guidance from the MISEV2023 update and the ISEV cell culture‐CM‐derived EVs task force perspective on considerations for reporting cell culturing parameters. Reporting adherence for each study can be found in Table 3, under heading ‘Rep.’ with the number of studies per category outlined in Table S2 and details for each individual study in the supplemental Excel file. Overall adherence to MISEV reporting guidelines was low (average 39%). Adherence was lowest for functional assays, with only 2% (n = 2) comparing EVs to CM as mentioned above, and 13% (n = 12) evaluating dose‐response, likely due to the difficulty in scaling EV production and running large experiments to test different fractions and doses. Thirty‐nine percent of studies (n = 35) defined isolated fractions as EVs, with the remaining studies largely using the term exosomes despite non‐specific isolation approaches. With regard to reporting for collection and pre‐processing (items 2–6 in Table 1, average 50%), most studies (97%, n = 86) reported on cell type except for 3% (n = 3) that did not report MSC source. Fifty‐six percent of studies (n = 50) did not report passage number. Culture medium before and during EV collection was generally well reported, with 74% of studies (n = 66) reporting both, as was the collection interval, which was reported by 67% of studies (n = 60). Thirty‐seven percent (n = 33) reported both interval and frequency. Cell viability during EV collection was only reported by 2% of studies (n = 2). The nature, size, and volume of culture vessels were rarely reported (7%, n = 6), while 31% (n = 28) reported on the type of culture vessel without details on volume. EV isolation protocols largely lacked the level of detail laid out by MISEV guidelines, with only 3% (n = 3) reporting on details such as the volume of CM and the type and size of tubes used. Metrics for EV characterization are largely reported in Section 3.3 (items 9–17, average 42%). Thirty‐seven percent of studies (n = 33) quantified EVs using two different methods (e.g., particle concentration and protein), while 51% (n = 45) reported quantification using only a single method, and 12% (n = 11) did not report a measure of EV quantification. Three percent (n = 3) provided a ratio of two quantification metrics. Of 84% of studies (n = 75) providing images of single EVs, 15% (n = 13) provided both wide‐field and close‐up images.
4. Discussion
The therapeutic effects of cell‐derived EV fractions in reducing symptoms of AD are well reported. Studies have demonstrated the localization of EVs to the brain, and subsequent alleviation of cognitive deficits and neurodegeneration in AD cell and animal models. Therapeutic EVs have been derived from MSCs, NSCs, iPSC‐derived progenitor cells, mature neural cells, and other cell types engineered to carry active cargo with effects that reduce AD pathological features. A wide range of proposed mechanisms of action have been reported, highlighting the multimodal effects that EVs may have on AD pathology and associated neurodegeneration. Most commonly, studies have focused on measuring EV efficacy by evaluating their effects on improving learning and memory in transgenic mice, reducing Aβ deposition/production, reducing pro‐inflammatory cytokine production, and improving cell viability after exposure to Aβ or LPS. Despite promising results, the overall quality of reporting is low, leading to reduced clarity on the contributions that EVs have over other components that co‐isolate during the isolation process, and reduced ability to be able to repeat and compare studies that use different production and collection protocols. Below we further discuss sources of EVs, methods to engineer them, models of AD to accurately test functionality, and challenges in their clinical translation.
4.1. Sources of EVs
Much of the literature to date has focused on MSC‐derived EVs, including early clinical trials that have demonstrated the safety of ADSC‐EVs as a treatment for AD. MSCs have been more traditionally studied due to the potential tumorigenicity of other stem cells like ESCs and iPSCs, and their relative ease of isolation and culture (Thanaskody et al. 2022). However, the production of EVs from iPSC‐derived cell lines, mature neural cells, and other engineered cell types has been gaining more traction in recent years. NSCs may be more relevant due to their role in homeostasis in the brain but are considered less clinically relevant due to the difficulty and invasive nature of isolating them from brain tissue. New methods of programming iPSCs have created iNSCs with comparable benefits to NSCs, and iPSC‐derived EVs have shown minimal immune response in preclinical studies (Gu et al. 2022). Studies comparing iMSC‐ to MSC‐derived EVs and iNSC‐ to NSC‐derived EVs have displayed comparable therapeutic effects on cognitive function, Aβ deposition, neuroinflammation, neurogenesis, and immunomodulation (Gao et al. 2023; Buitrago et al. 2024), demonstrating high potential for iPSC‐derived progenitors for EV production.
There is, however, a lack of comparative studies to provide evidence of what cell type produces the most therapeutic EVs for the treatment of AD using the same functional assays, isolation methods, and dosage. In a murine thromboembolic stroke model, a study compared EVs derived from iNSCs and iMSCs differentiated from the same iPSC cell line, and found iNSCs were more effective in improving cellular, tissue, and functional outcomes (Webb et al. 2018). In 5xFAD mice, it was found iNSC‐EVs were more effective than MSC‐EVs in improving cognitive outcomes, inhibiting Aβ42 accumulation and p‐tau propagation, and restoring dendritic length and dendritic spine density (Gao et al. 2023). However, this study is limited by the use of cells from different individuals, which may not be comparable due to their heterogenous nature. Other studies have compared stem cell‐derived EVs to those produced from other cell types. NSC‐EVs and not EVs from mature neurons were able to restore Aβ oligomer‐induced suppression of hippocampal neuron long‐term potentiation and subsequent memory deficits in an AD mouse model, highlighting the advantages of stem cells as source cells for EV therapies (Micci et al. 2019). Another study briefly reported that MSC‐EVs induced greater permeability than microglia‐EVs in the extracellular matrix, with a proteomic comparison revealing specific membrane proteins on MSC‐EVs exhibiting matrix degradative activity, which may enhance their abilities to cross biological barriers (Huang et al. 2024). While some studies have found functional benefits from EVs derived from mature neural cells including gamma‐aminobutyric acid (GABA)‐treated neurons, CMECs, astrocytes, and microglia, mature cell types do not exhibit the same capacity for proliferation as compared to stem cells and thus may be limited in the number of EVs that can be feasibly produced.
Recent studies excluded during the formal review process have also explored novel sources of EVs that may have functional relevance in AD, including Lactobacillus paracasei (Kwon et al. 2023), Tomafran (a bioengineered tomato) (Etxebeste‐Mitxeltorena et al. 2024), citrus lemon (Dolma et al. 2024), coffee extract (Esmekaya and Ertekin 2024), and Lycium ruthenicum Murray (Zhang et al. 2025). Lactobacillus paracasei‐derived EVs were tested in transgenic APP/PS1 mice and were found to counteract Aβ‐induced downregulation of neurotrophic factors, nuclear factors methyl‐CpG binding protein 2 (MECP2) and sirtuin 1 (SIRT1), and Aβ‐degrading MMP‐2, MMP‐9, and NEP, alleviating neuroinflammation, Aβ accumulation, and cognitive decline (Kwon et al. 2023). Tomafran and citrus lemon‐derived EVs were tested in SH‐SY5Y cells, showing neuroprotective effects, and antioxidant activity, respectively (Etxebeste‐Mitxeltorena et al. 2024, Dolma et al. 2024). Coffee‐derived EVs were tested in Aβ‐induced HT‐22 neurons and protected them against Aβ‐induced toxicity (Esmekaya and Ertekin 2024). Lycium ruthenicum Murray‐derived EVs were tested in transgenic AD Caenorhabditis elegans, extending the worm lifespan and chemotaxis index, reducing Aβ, inhibiting acetylcholinesterase activity to mitigate cholinergic dysfunction, restoring mitochondrial membrane potential and ATP production to ameliorate mitochondrial dysfunction, and reducing oxidative stress and inflammation (Zhang et al. 2025).
4.2. Engineering EVs
The therapeutic efficacy for EVs can be further improved through engineering either parent cells by endogenous methods, or through direct modification of EVs by exogenous methods. EVs may also be considered as drug carriers to improve drug biodistribution and facilitate movement across the BBB, and have been used to successfully deliver proteins, small molecules, RNAs, and gene editing systems such as CRISPR‐Cas9 to enable precise, non‐viral gene editing (Zhu et al. 2023; Han et al. 2024; Liang et al. 2025; Li et al. 2023; Wang et al. 2024).
Of the included studies, therapeutic improvements from endogenous modifications were seen from cells grown in 3D conditions, exposed to hypoxia, inflammation, or heat shock, or genetically modified to overexpress specific genes or miRs. Genetic modification of parent cells enables site‐specific insertion, deletion or modifications within the genome to indirectly improve EV function, preserving EV integrity and functionality as compared to exogenous methods. However, genetic modification is complex and has low reproducibility and transfection efficiency (Huang et al. 2023). Alternatively, EV content can be modified through various cell pre‐conditioning strategies (Yin et al. 2023; Chen et al. 2022), though work is needed to evaluate the optimal growth conditions of the cells towards specific applications. These strategies include optimizing the oxygen levels cells are exposed to, the culture medium prior to and during EV collection (including effects of serum deprivation), the mode of culture (i.e., 2D vs. 3D), cell confluency, and exposure of the cells to specific stimuli including shear, electromagnetic forces, ultrasound, biochemical factors such as inflammatory factors, and changes in temperature or pH, all of which have been shown to alter EV content (Patel et al. 2017; Haraszti et al. 2019; Losurdo et al. 2020; Deng Z et al. 2021; Nakase et al. 2021; Huber CC et al. 2022; Liu H et al. 2022; Markoutsa et al. 2022; Yang et al. 2020; Phelps et al. 2023, 2024; Rayamajhi et al. 2023; Wang, Worrell and Wang 2023; Pourhadi M et al. 2024). The scaled production of EVs should also be considered. Large‐scale production of cell lines is typically performed in bioreactors, where cells are exposed to shear forces and higher levels of oxygen and nutrient transfer. EVs have been produced in stirred suspension (Phelps et al. 2024, 2022; Costa et al. 2023), vertical wheel (De Almeida Fuzeta et al. 2020; Jeske et al. 2023), perfusion (Kronstadt et al. 2023), flat plate (Kang et al. 2022), and hollow fibre (Cao et al. 2020; Yan and Wu 2020; Gobin et al. 2021) bioreactors, all of which have reportedly higher EV production rates per cell when compared to static culture, and have demonstrated similar or enhanced functionality in different types of models.
EVs can also be modified exogenously after production through encapsulation or adhesion of specific miRs, proteins, or other categories of drugs, enabling higher control over their cargo. Of the studies included in the current review, EVs have been exogenously modified to target their delivery to the brain or specific cell types, and by loading them with active molecules shown to be effective in improving AD‐related neurodegeneration and reducing pathological gene expression. For enhancing targeting mechanisms, the RVG peptide interacts specifically with acetylcholine receptors to enable viral entry into neuronal cells (Cui et al. 2019). Conjugation of mannose, an epimer of glucose, enables specific targeting of EVs to microglia, which are rich in mannose receptors (CD206) (Hao et al. 2022). Regarding drug delivery, AM1241 is a CB2R‐specific agonist, where activation of CB2R is hypothesized to be a promising target for neurodegenerative diseases (Zhu et al. 2023). NEP is an active enzyme with the function of clearing Aβ in the brain (Izadpanah et al. 2020). CoQ10 is a supplement shown to reduce inflammation and oxidative stress associated with AD (Sheykhhasan et al. 2022). Gemfibrozil is an FDA‐approved drug for hyperlipidaemia, which has also demonstrated a protective effect on dopaminergic neurons, reduced neuroinflammation, and promotion of lysosomal activity for clearing Aβ, but, like many other drugs, has difficulty crossing the BBB (Hao et al. 2022). Encapsulation of such drugs into EVs can enhance their efficacy by enabling targeted delivery across the BBB and improving their stability. Unfortunately, these methods can be limited by low encapsulation efficiencies or by subjecting the EVs to harsh chemical and physical environments. Drugs can be loaded by incubation of the EVs with the drugs for chemical loading; however, the loading efficiency of this method is low, and it is limited to hydrophobic drugs (Wang, Chen, et al. 2021). The use of transfection agents or permeabilizers may improve drug entry into EVs. For example, the permeabilizer saponin removes membrane cholesterol, inducing pore formation without destroying the lipid bilayer structure (Shany et al. 1974); however, such methods should be used with caution as they could also have detrimental effects on specific cell types. For example, in the case of saponin, there are concerns that it could induce a haemolytic effect on blood cells (Wang, Chen, et al. 2021). Physical methods such as acoustofluidics, electroporation, hypotonic dialysis, sonication, and others, may improve loading without the need for chemical modification (Yin et al. 2023). While these methods exhibit higher loading efficiencies, they can result in deformation and aggregation of EVs and/or their cargo. Newer physical methods such as acoustofluidics may improve EV integrity but may be limited by the availability of equipment and method complexity (Wang et al. 2022).
Though excluded from formal analysis in the current review, EVs from bodily fluids have also been considered as drug carriers, with studies showing the loading of plasma‐derived EVs with curcumin (Ding et al. 2025), quercetin (Qi et al. 2020), and donepezil (Oliveira Silva et al. 2024); urine‐derived EVs with resveratrol and platinum nanoparticles and modified with RVG (Bai et al. 2023); and one study evaluating the effect of plasma‐derived EVs isolated following acute exercise (Fuller et al. 2025). Interestingly, donepezil‐loaded EVs were compared to PLA‐PEG nanoparticles with the same amount of drug and were found to improve the drug's activity and biodistribution (Oliveira Silva et al. 2024). An additional eight studies were excluded from the current analysis, as they did not compare the modified EVs to non‐modified EVs. Resveratrol and olesoxime were loaded into BMSC‐EVs, which cleave Aβ42 peptides into smaller fragments and promote the survival of motor neurons (Wang et al. 2024). Berberine and palmatine were loaded into transferrin‐decorated EVs from microglia, a follow‐up to the group's previous study (Zhao et al. 2023), with the additional use of transferrin to facilitate transport across the BBB (Zhou et al. 2025). One study evaluated EVs from curcumin‐treated macrophages and demonstrated their ability to restore brain‐derived neural factor (BDNF) levels, reduce inflammation and Aβ aggregation, and improve cognitive function in AD (Wang et al. 2019). Another study loaded macrophage‐derived EVs with curcumin and methylene blue, demonstrating inhibition of tau phosphorylation and reversal of cognitive decline in okadaic acid‐induced mice (Yang et al. 2025). Other studies have focused on more extensive EV engineering applications. Superparamagnetic iron oxide nanoparticles were conjugated to IL‐8‐stimulated neutrophil‐derived EVs and loaded with curcumin to restore mitochondrial function in APP/PS1 mice (Zhang et al. 2024). RVG‐conjugated MSC‐derived EVs loaded with siRs targeting BACE1 and caspase 3, contained within a shell of ROS‐responsive polymer, reduced astrocyte reactivity and restored memory deficits in 3xTg‐AD mice (Li et al. 2023). Exosome‐liposome hybrid nanovesicles were developed to co‐deliver BACE1 siR and TREM2 plasmid to reprogram microglia to an anti‐inflammatory phenotype and restore Aβ phagocytosis activity (Jiang et al. 2024). Finally, CMEC and macrophage EV membranes were hybridized and combined with polydopamine nanoparticles, resveratrol, and Aβ‐targeting aptamers to improve BBB penetration and targeting to the brain (Du et al. 2025). Engineering EVs as drug carriers offers an exciting avenue for the EV field and could enable large‐scale production from lower maintenance cell lines such as HeLa cells or CHO cells, with functional EVs engineered through exogenous and/or endogenous methods.
As more research emerges regarding pathological mechanisms of AD, targeted engineering approaches may be more desired, as demonstrated in recent studies (Cai et al. 2024; Huang et al. 2024; Li et al. 2023; Jiang et al. 2024). Loading of siRs within EVs could specifically target and downregulate key pathological genes such as BACE1, APP, PSEN2, PSEN1, MAPT, APOE, TREM2, glutamate receptors, calcium channels, and inflammatory factors such as TNF‐α, IL‐1β, and COX2 (Jain et al. 2024), and the addition of synthetic RNAs could interact with different signalling pathways to improve functional deficits (Rezayof et al. 2024). Recent research has further explored the use of EVs for direct genome and epigenome editing. A major challenge in gene therapy is the intracellular delivery of proteins and RNA, with issues such as endosomal entrapment, toxicity, and detection by neutralizing antibodies reported in common delivery systems such as cell‐penetrating peptides, liposomes, and adeno‐associated virus (AAV) vectors (Liang et al. 2025; György et al. 2014). EVs have the natural ability to evade neutralizing antibodies, thereby reducing immune response. For example, EV‐associated AAV vectors were found to be 136‐fold more resistant to neutralizing antibodies compared to standard AAV vectors (György et al. 2014). Recent engineering technologies have been developed to enrich specific proteins into EVs so they can escape endosomes once taken up by cells. Liang et al. combined an engineered mini‐intein protein with self‐cleavage activity for active cargo loading and release with a fusogenic VSV‐G protein for endosomal escape to enable highly efficient recombination and genome editing, achieving up to 40% gene editing efficiency by HEK293T‐EVs delivering Cre in the hippocampus (Liang et al. 2025). In a similar approach, Han et al. engineered EVs with a photoinducible cargo protein release system termed ‘mMaple3 mediated protein loading into and release from exosome’ (MAPLEX), and successfully delivered MAPLEX carrying a Cas9 ribonucleoprotein complex to edit the BACE1 promoter, successfully reducing Aβ and improving memory and cognition in 5xFAD and 3xTg‐AD mice (Han et al. 2024).
4.3. Models of AD
Therapeutic research on EVs has primarily focused on the Aβ cascade hypothesis, a controversial theory that presumes the buildup and lack of clearance of Aβ leads to other pathological markers of AD, including neurofibrillary tangles, neuroinflammation, and other neurodegenerative features (Solis et al. 2020; Kurkinen et al. 2023). This is seen through the high amount of cell studies focused on Aβ stimulation as a means of modelling AD, and transgenic animal models that are driven by mutations in APP and PSEN1 genes. However, many of these animal models do not replicate all aspects of AD pathology, including neuronal loss and neurofibrillary tangle development, as thoroughly reviewed by Drummond et al. (Drummond and Wisniewski 2017). Similarly, Aβ‐induced models replicate AD‐like behavioural abnormalities and exhibit Aβ pathology, but do not replicate p‐tau pathology (Kim et al. 2016). A recent meta‐analysis by Deng et al. included 16 studies that tested the therapeutic efficacy of stem cell‐derived EVs in animal models of AD, with pooled analyses confirming significant function in increased learning as measured by Morris water maze and Y‐maze tests, reduced Aβ plaques in the hippocampus, and reduced proinflammatory cytokines TNF‐α and IL‐1β (Deng et al. 2024). Included animal models were variations of APP/PS1 transgenic mice and rats, 5xFAD mice, C57BL/6 mice with Aβ42 injection, and Sprague Dawley (SD) rats with Aβ42 injection. Focusing on a broader definition of AD, the current study found a significantly higher number of studies (n = 70) that included AD‐related animal models and included additional studies reporting on AD‐related cell models only (n = 18), and a clinical trial (n = 1). Of the included studies, rodent animal models were used in 69/70 studies, which is limiting due to their inability to naturally develop the neuropathological features seen in AD (Drummond and Wisniewski 2017). APP/PS1 and 5xFAD transgenic mice were the most frequent for testing EV‐based treatments for AD. However, not only do these lack the ability to fully replicate AD pathology, but familial AD due to genetic mutations (PSEN1, PSEN2, APP) accounts for < 5% of cases (Drummond and Wisniewski 2017; Bali et al. 2012). Sporadic AD, caused by a complex combination of genes, environment, and lifestyle, accounts for > 95% of AD cases (Bali et al. 2012). Both sporadic and familial AD exhibit similar pathological features, but the underlying mechanisms leading to sporadic AD are not yet fully elucidated. Besides genetic risk factors, including APOE4 and TREM2, vascular risk factors (e.g., high blood pressure, diabetes, inflammation, hypoxia and stroke) play a major role in the development of sporadic AD, with some hypothesizing that sporadic AD is in fact a vascular disease (De La Torre 2002).
Studies testing EVs have induced sporadic AD‐like pathologies using LPS, STZ, okadaic acid, kainic acid, AlCl3, NaN3, OBE, or hypoxia through BCCAO. LPS is a potent inducer of inflammation. STZ is toxic to pancreatic beta cells and induces insulin resistance, which can lead to neuroinflammation and oxidative stress. Experimentally, low sub‐diabetogenic STZ injection leads to the production of Aβ fragments and tau proteins, activation of astrocytes and microglia, neuron apoptosis, and cognitive impairment mimicking AD pathology (Lin, Huang, et al. 2024). Okadaic acid is a polyether toxin that inhibits serine/threonine phosphatases PP1 and PP2A, which have been implicated in AD, and induces tau hyperphosphorylation, Aβ deposition, oxidative stress, neuroinflammation, synaptic loss, neurotoxicity, and memory impairments (Kamat and Nath 2015). Kainic acid binds to specific excitatory amino acid receptors in the central nervous system and leads to microglial and astrocyte activation, aggregation of Aβ, and cognitive impairments (Ruan et al. 2019). AlCl3 is a neurotoxin, with aluminium acting as a crosslinker of Aβ to result in aggregation, inducing oxidative damage, cholinergic dysfunction, activation of the NF‐κβ pathway, and learning and memory impairments (Shunan et al. 2021). NaN3 causes acute neurotoxicity that inhibits mitochondrial cytochrome c oxidase activity (COX‐IV) and increases ROS production, leading to AD‐like learning and memory impairments. OBE animals exhibit symptoms of depression, impaired spatial memory, neurodegeneration, and increased deposition of Aβ (Aleksandrova et al. 2004). BCCAO can recapitulate the cognitive symptoms and pathology of AD by disrupting the neurovascular unit and increasing astrocyte reactivity (Solis et al. 2020). EVs from various cell types have been demonstrated to provide benefits in each of these models. This includes the alleviation of LPS‐induced inflammation by EVs from iNSCs (Chen, Lan, et al. 2024), NSCs (Khan et al. 2023), BMSCs (Aboulhoda et al. 2021; Kaniowska et al. 2022; Hamed et al. 2024), ADSCs (Silva et al. 2025; Garcia‐Contreras and Thakor 2021), UMSCs (Feng et al. 2020), AFMSCs (Zavatti et al. 2022), inflammatory‐primed MSCs (Markoutsa et al. 2022), and STZ‐induced symptoms from EVs of iMSCs (Lin, Huang, et al. 2024), 3D spheroid UMSCs (Pourhadi et al. 2024), and BMSCs (Liu, Fan, et al. 2022). EVs from dendritic cells transfected with miR‐29b‐2 and CD47 inhibited PSEN1 and Aβ oligomer deposition in SH‐SY5Y cells induced with okadaic acid (Lin, Hsu, et al. 2024). BMSC‐EVs were found to reduce apoptosis, necrosis, and oxidative stress in hippocampal cells treated with kainic acid (Venugopal et al. 2018). NaN3‐induced toxicity has been mitigated through MSC‐EV injection (Amer et al. 2024), as has AlCl3‐induced neurodegeneration (Ebrahim et al. 2024). WJMSC‐EVs have demonstrated the benefits of improving spatial memory in OBE mice (Poltavtseva et al. 2021; Zhdanova et al. 2021). Finally, EVs from UMSCs and NSCs have been found to reduce neuronal damage, oxidative stress, inflammation, and improve learning and memory in BCCAO‐induced rats (Hu et al. 2020; Qi et al. 2021; Wang, Yi, et al. 2023).
Zebrafish are emerging as real‐time models to study AD and drug discovery, as they have a high degree of neurological and behavioural similarities to humans, and similar gene orthologs to those mutated in human familial AD (Saleem and Kannan 2018). However, one major disadvantage to zebrafish models is their ability to regenerate neurons throughout life, unlike humans and other mammals. Injection of Aβ42 peptides in zebrafish leads to neurogenesis, as opposed to the toxic effect of Aβ seen in other models (Bhattarai et al. 2017). Despite this disadvantage, zebrafish models present advantages in their size and scalability and their ability to real‐time image physiological processes, including neurodegeneration. Novel screening methods such as human iPSC‐derived brain organoids (generated through spheroid culture or with higher control using 3D bioprinting) may provide more insight and relevance into EV mechanisms for treating AD and other diseases (Sharma et al. 2020). The organoids can be generated using a variety of different AD donors with varying pathology to better interpret and predict EV behaviour in different phenotypes. Despite the lack of a single robust AD model, EVs have been successful in reducing neurodegeneration after exposure to differing pathologies and neurotoxins.
4.4. Challenges in the Translation of EVs
Several challenges exist that limit the translational potential of EVs. This includes a lack of standardized and reliable methods to isolate and characterize EVs at high purity and efficiency, limitations in visualizing EVs in vivo and targeting EVs to specific areas of the brain, and generally low circulation times and retention in the target area. The translation of EVs to the clinic is further hampered by the complexity and heterogeneity of EVs, which limits the ability to identify their mechanism of action and creates uncertainty as to their long‐term effects, especially given that they can play a role in disease pathogenesis. Further research to validate the proposed factors and mechanisms and better understand the basic biological properties of EVs is needed to address these limitations; however, appropriate uses of controls and novel technologies can mitigate some of the associated challenges.
A major limitation is the absence of standardized and reliable methods for isolating and characterizing EV populations. Current methods of isolating EVs are limited in both purity and yield and contain varying amounts of protein aggregates and lipoproteins that may come from the cells or from the culture medium they are grown in and may or may not be associated with EVs as part of their corona (Welsh et al. 2024). Protein aggregates and lipoproteins appear in the same size range as EVs and impact commonly used EV quantification methods (e.g., protein content, NTA). These methods also isolate an array of heterogenous types of EVs that may have different biological functions. The type of isolation method used can therefore impact characterization and functionality of EVs, which has implications on dosing and determining their mechanism of action. Further technologies that improve and ease EV isolation are beginning to emerge. For example, novel label‐free methods that take advantage of the dielectric properties of EVs are able to obtain both higher yield and purity compared to conventional techniques (Sharma et al. 2023; Back et al. 2024; Hou et al. 2024). Regarding being able to isolate specific subtypes of EVs, the field is still lacking the essential biological properties that can identify subtypes to enable isolation technologies to be developed. Overall, there is a significant need for studies to compare the functionality of EVs isolated using different methods to optimize production protocols for the highest therapeutic efficacy in AD and other applications.
In creating a therapeutic product, high purity may not be needed if the co‐isolating components are not harmful. In fact, they may provide additional therapeutic benefits. However, a thorough understanding of their contribution or impact is required for clinical studies. Few studies within the AD field have evaluated the contributions of culture medium or the non‐EV fraction on varying AD models and assays, which is a major limitation in understanding the contribution of EVs in improving cognitive function and reducing AD pathology. Markoutsa et al. found the CM to have similar effects on reducing microglia activation to the isolated EVs but did not see the same effects in the supernatant following isolation (Markoutsa et al. 2022). On the other hand, Sha et al. found the CM to have no effects on hippocampal neuron cultures, while EVs protected them from Aβ toxicity (Sha et al. 2021). EV‐depleted FBS, which was used in both of these studies, and 34% of studies found in this review, has been found to still contain a population of vesicles that varies by batch (Pham et al. 2021; Karttunen et al. 2022). Similarly, studies have found serum‐free media to have components that co‐purify with EVs and may elicit false positive results in therapeutic studies (Auber et al. 2019). It is important to consider the culture medium that has undergone the same isolation process as a control within such studies and further consider the potential of the culture medium to alter therapeutic efficacy.
Further, despite excitement over the stability of EVs in vivo in comparison to liposomal drug carriers, studies have found EVs to have a half‐life as short as 2 min in circulation, and distribution predominantly to the liver, spleen, lungs, and kidneys (Takahashi et al. 2013; Lai et al. 2014). EVs may be engineered to improve their targeting capabilities, stability, and retention. Targeting of EVs could be done through engineering EVs: as discussed earlier, RVG and mannose enable specific targeting to neuronal and microglia populations (Hao et al. 2022; Cui et al. 2019). Technologies like focused ultrasound have been shown to transiently and non‐invasively increase the permeability of the BBB, enabling drug delivery to specific regions of the brain (Burgess et al. 2015), and may be an effective strategy to improve EV delivery to regions that are most affected (Deng et al. 2021; Haroon et al. 2023). Specific molecules may also be used to enhance circulation time. For example, in an application of targeted cancer drug delivery, Belhadj et al. showed that CD47 ('don't eat me' signal)‐enriched EVs had prolonged circulation time in vivo by helping them to evade phagocytosis by the mononuclear phagocyte system, while at the same time functionalizing them with a novel homing peptide to promote cancer cell uptake (Belhadj et al. 2020). Another strategy that has been explored to improve EV circulation time is associating albumin binding domains on the EV surface either by fusion to EV‐associated proteins or through lipid anchorage. This has been found to increase circulation time by 10–19‐fold compared to unmodified EVs (Liang et al. 2022; Zheng et al. 2023). The encapsulation of EVs in hydrogels may be another method to limit early degradation and improve their retention in a target area. For example, Huang et al. embedded EVs in a hydrogel designed to be susceptible to degradation, as its design was based on the membrane enzymes present in MSC‐EVs (Huang et al. 2024). The hydrogel was used for intranasal injection to retain the EVs in the nasal cavity. Another limitation is the lack of robust methods to visualize EVs in vivo to enable live studies that could help to appropriately identify circulation times and biodistribution. Lipophilic membrane dyes like PKH and Dil are the most common methods for labelling and visualizing EVs in cell culture and animal models but are constrained by the need for animal sacrifice to visualize the EVs in brain sections. A recent study also found that PKH dyes form micelles from the dye itself that can create false positive staining, thereby warranting caution when using them as a labelling method to verify biodistribution (Wan et al. 2024). To enable in vivo imaging of EVs, both iron oxide and gold nanoparticles have been loaded into EVs, enabling EVs to be tracked using live imaging methods such as MRI and x‐ray computed tomography (Perets et al. 2019; Liu, Helsper, et al. 2022).
5. Conclusion
There is increasing interest in the use of EVs from stem cells and other engineered cell types for use as therapeutics in applications of regenerative medicine, including AD and other dementias. There is preliminary evidence that supports the use of EVs for the treatment of AD, since they host protective and regenerative cargo that can reduce inflammation, oxidative stress, and apoptosis, restore neuronal function, and contribute to Aβ clearance. Further engineering of EVs has demonstrated improved function using pre‐conditioning strategies or transfection of cells during culture, as well as direct loading following their isolation. These strategies enable the production and modification of EVs from cell types other than traditionally used stem cells. However, there is a lack of direct evidence denoting therapeutic effects purely to EVs, with many studies using low purity isolation methods that may contain other functional biological molecules such as proteins and RNAs. In addition, study adherence to the MISEV guidelines was low (average 39%) for reporting on EV collection and pre‐processing, characterization, and providing dose‐response and appropriate controls for functional assays. Adherence to well‐set out standards, as well as the development of new isolation and characterization technologies, is needed to better understand the contribution of EVs. Finally, the use of robust AD models is necessary to better compare different culture methods, cell sources, and isolation methods for producing EVs, and to understand the mechanisms by which EVs exert their benefits.
Author Contributions
Jolene Phelps: conceptualization (lead), data curation (lead), formal analysis (lead), funding acquisition (equal), methodology (lead), visualization (lead), writing – original draft (lead), writing – review and editing (lead). Amanda Orr: data curation (supporting), formal analysis (supporting), visualization (supporting), writing – original draft (supporting). Katherine S. Elvira: supervision (supporting), writing – review and editing (supporting). Stephanie M. Willerth: funding acquisition (equal), supervision (lead), writing – review and editing (supporting).
Conflicts of Interest
Stephanie M. Willerth is the CEO of Axolotl Biosciences. The remaining authors declare no conflicts of interest.
Supporting information
Supplementary Excel file: jex270077‐sup‐0001‐Tables.xlsx
Supplementary Table: jex270077‐sup‐0002‐SuppMat.docx
Acknowledgements
This research was supported by a Post‐Doctoral award from the Alzheimer Society Research Program to J.P., a Research Trainee Award from Michael Smith Health Research B.C./Canadians for Leading Edge Alzheimer Research to A.O., the Natural Sciences and Engineering Research Council of Canada (Discovery grant: RGPIN‐2017‐04044), and the Canadian Institutes of Health Research (Project grant: 460840).
Phelps, J. , Orr A., Elvira K. S., and Willerth S. M.. 2025. “Extracellular Vesicles for the Treatment of Alzheimer's Disease: A Systematic Review.” Journal of Extracellular Biology 4, no. 8: 4, e70077. 10.1002/jex2.70077
Funding: This research was supported by a Post‐Doctoral award from the Alzheimer Society Research Program to J.P. a Research Trainee Award from Michael Smith Health Research B.C/Canadians for Leading Edge Alzheimer Research to A.O., the Natural Sciences and Engineering Research Council of Canada (Discovery grant: RGPIN‐2017‐04044), and the Canadian Institutes of Health Research (Project grant: 460840).
Data Availability Statement
The data that supports the findings of this study are available in the supplementary material of this article.
References
- 2023. “2023 Alzheimer's Disease Facts and Figures.” Alzheimers Dementia 19, no. 4: 1598–1695. 10.1002/alz.13016. [DOI] [PubMed] [Google Scholar]
- Aboulhoda, B. E. , Rashed L. A., Ahmed H., et al. 2021. “Hydrogen Sulfide and Mesenchymal Stem Cells‐Extracted Microvesicles Attenuate LPS‐Induced Alzheimer's Disease.” Journal of Cellular Physiology 236: 5994–6010. 10.1002/jcp.30283. [DOI] [PubMed] [Google Scholar]
- Aleksandrova, I. Y. , Kuvichkin V. V., Kashparov I. A. et al. 2004. “Increased Level of β‐Amyloid in the Brain of Bulbectomized Mice.” Biochemistry (Moscow) 69: 176–180. 10.1023/B:BIRY.0000018948.04559.ab. [DOI] [PubMed] [Google Scholar]
- Alvarez‐Erviti, L. , Seow Y., Yin H., Betts C., Lakhal S., and Wood M. J. A.. 2011. “Delivery of siRNA to the Mouse Brain by Systemic Injection of Targeted Exosomes.” Nature Biotechnology 29: 341–345. 10.1038/nbt.1807. [DOI] [PubMed] [Google Scholar]
- Amer, A. S. , Ali E., Zahra M. M., and Sabry H. A.. 2024. “Mesenchymal Stem Cell‐Derived Exosomes Modulate the COX‐IV Pathway via Inhibition of Amyloidogenesis and Mitoprotection in Sodium Azide‐ Alzheimer Model in Rats.” Scientific African 25: e02274. 10.1016/j.sciaf.2024.e02274. [DOI] [Google Scholar]
- Apodaca, L. A. , Baddour A. A. D., Garcia C. Jr, et al. 2021. “Human Neural Stem Cell‐Derived Extracellular Vesicles Mitigate Hallmarks of Alzheimer's Disease.” Alzheimer's Research & Therapy 13: 57. 10.1186/s13195-021-00791-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Attaluri, S. , Jaimes Gonzalez J., Kirmani M., et al. 2023. “Intranasally Administered Extracellular Vesicles From Human Induced Pluripotent Stem Cell‐Derived Neural Stem Cells Quickly Incorporate Into Neurons and Microglia in 5xFAD Mice.” Frontiers in Aging Neuroscience 15: 1200445. 10.3389/fnagi.2023.1200445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Auber, M. , Fröhlich D., Drechsel O., Karaulanov E., and Krämer‐Albers E.‐M.. 2019. “Serum‐Free Media Supplements Carry miRNAs That Co‐Purify With Extracellular Vesicles.” Journal of Extracellular Vesicles 8: 1656042. 10.1080/20013078.2019.1656042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Back, W. , Bang M., Jung J.‐H., et al. 2024. “Charge‐Based Isolation of Extracellular Vesicles From Human Plasma.” ACS Omega 9: 17832–17838. 10.1021/acsomega.3c07427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bai, Z. , Ge K., Fu J., et al. 2023. “Engineered Urinary‐Derived Extracellular Vesicles Loaded Nanoenzymes as Trojan Horses to Regulate the Inflammatory Microenvironment for Treatment of Alzheimer's Disease.” Chemical Engineering Journal 465: 142955. 10.1016/j.cej.2023.142955. [DOI] [Google Scholar]
- Bali, J. , Gheinani A. H., Zurbriggen S., and Rajendran L.. 2012. “Role of Genes Linked to Sporadic Alzheimer's Disease Risk in the Production of β‐amyloid Peptides.” Proceedings of the National Academy of Sciences of the United States of America 109: 15307–15311. 10.1073/pnas.1201632109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bashirrohelleh, M.‐A. , Bavarsad K., Khodadadi A., Shohan M., and Asadirad A.. 2025. “Curcumin‐Enhanced Stem Cell Exosomes: A Novel Approach to Modulating Neuroinflammation and Improving Cognitive Function in a Rat Model of Alzheimer's Disease.” European Journal of Pharmacology 999: 177695. 10.1016/j.ejphar.2025.177695. [DOI] [PubMed] [Google Scholar]
- Belhadj, Z. , He B., Deng H. et al. 2020. “A Combined 'Eat Me/Don't Eat Me' Strategy Based on Extracellular Vesicles for Anticancer Nanomedicine.” Journal of Extracellular Vesicles 9: 1806444. 10.1080/20013078.2020.1806444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bhattarai, P. , Thomas A. K., Zhang Y., and Kizil C.. 2017. “The Effects of Aging on Amyloid‐β42‐Induced Neurodegeneration and Regeneration in Adult Zebrafish Brain.” Neurogenesis 4: e1322666. 10.0180/23262133.2017.1322666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bodart‐Santos, V. , de Carvalho L. R. P., de Godoy M. A., et al. 2019. “Extracellular Vesicles Derived From Human Wharton's Jelly Mesenchymal Stem Cells Protect Hippocampal Neurons From Oxidative Stress and Synapse Damage Induced by Amyloid‐β Oligomers.” Stem Cell Research & Therapy 10: 332. 10.1186/s13287-019-1432-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Budd Haeberlein, S. , Aisen P. S., Barkhof F., et al. 2022. “Two Randomized Phase 3 Studies of Aducanumab in Early Alzheimer's Disease.” The Journal of Prevention of Alzheimer's Disease 9: 197–210. 10.14283/jpad.2022.30. [DOI] [PubMed] [Google Scholar]
- Buitrago, J. C. , Morris S. L., Backhaus A., et al. 2024. “Unveiling the Immunomodulatory and Regenerative Potential of iPSC‐Derived Mesenchymal Stromal Cells and Their Extracellular Vesicles.” Scientific Reports 14: 24098. 10.1038/s41598-024-75956-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Burgess, A. , Shah K., Hough O., and Hynynen K.. 2015. “Focused Ultrasound‐Mediated Drug Delivery Through the Blood–Brain Barrier.” Expert Review of Neurotherapeutics 15: 477–491. 10.1586/14737175.2015.1028369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cai, H. , Pang Y., Ren Z., Fu X., and Jia L.. 2024. “Delivering Synaptic Protein mRNAs via Extracellular Vesicles Ameliorates Cognitive Impairment in a Mouse Model of Alzheimer's Disease.” BMC Medicine 22: 138. 10.1186/s12916-024-03359-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Campbell, N. B. , Patel Y., Moore T. L., Medalla M., and Zeldich E.. 2023. “Extracellular Vesicle Treatment Alleviates Neurodevelopmental and Neurodegenerative Pathology in Cortical Spheroid Model of Down Syndrome.” International Journal of Molecular Sciences 24: 3477. 10.3390/ijms24043477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cao, J. , Wang B., Tang T., et al. 2020. “Three‐Dimensional Culture of MSCs Produces Exosomes With Improved Yield and Enhanced Therapeutic Efficacy for Cisplatin‐Induced Acute Kidney Injury.” Stem Cell Research & Therapy 11: 206. 10.1186/s13287-020-01719-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen, C. , Bao Y., Xing L., et al. 2023. “Exosomes Derived From M2 Microglial Cells Modulated by 1070‐nm Light Improve Cognition in an Alzheimer's Disease Mouse Model.” Advanced Science 10: 2304025. 10.1002/advs.202304025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen, C. , Lan Z., Tang X., et al. 2024. “Human‐Derived Induced GABAergic Progenitor Cells Improve Cognitive Function in Mice and Inhibit Astrocyte Activation With Anti‐Inflammatory Exosomes.” Annals of Neurology 96: 488–507. 10.1002/ana.27001. [DOI] [PubMed] [Google Scholar]
- Chen, H. , Deng C., Meng Z., et al. 2024. “Combined Catalpol and Tetramethylpyrazine Promote Axonal Plasticity in Alzheimer's Disease by Inducing Astrocytes to Secrete Exosomes Carrying CDK5 mRNA and Regulating STAT3 Phosphorylation.” Molecular Neurobiology 61, no. 12: 10770–10791. 10.1007/s12035-024-04251-z. [DOI] [PubMed] [Google Scholar]
- Chen, H. , Huang Z., Lei A., Yu X., Shen M., and Wu D.. 2024. “miRNA‐211‐5p Inhibition Enhances the Protective Effect of hucMSC‐Derived Exosome in Aβ1‐40‐Induced SH‐SY5Y Cells by Increasing NEP Expression.” Journal of Biochemical and Molecular Toxicology 38: e23624. 10.1002/jbt.23624. [DOI] [PubMed] [Google Scholar]
- Chen, S. , Sun F., Qian H., Xu W., and Jiang J.. 2022. “Preconditioning and Engineering Strategies for Improving the Efficacy of Mesenchymal Stem Cell‐Derived Exosomes in Cell‐Free Therapy.” Stem Cells International 2022: 1779346. 10.1155/2022/1779346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen, Y.‐A. , Lu C.‐H., Ke C.‐C., et al. 2021. “Mesenchymal Stem Cell‐Derived Exosomes Ameliorate Alzheimer's Disease Pathology and Improve Cognitive Deficits.” Biomedicines 9: 594. 10.3390/biomedicines9060594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chêne, G. , Beiser A., Au R., et al. 2015. “Gender and Incidence of Dementia in the Framingham Heart Study From Mid‐Adult Life.” Alzheimer's & Dementia 11: 310–320. 10.1016/j.jalz2013.10.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cone, A. S. , Yuan X., Sun L., et al. 2021. “Mesenchymal Stem Cell‐Derived Extracellular Vesicles Ameliorate Alzheimer's Disease‐Like Phenotypes in a Preclinical Mouse Model.” Theranostics 11: 8129–8142. 10.7150/thno.62069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Costa, M. H. G. , Costa M. S., Painho B., et al. 2023. “Enhanced Bioprocess Control to Advance the Manufacture of Mesenchymal Stromal Cell‐derived Extracellular Vesicles in Stirred‐tank Bioreactors.” Biotechnology and Bioengineering 120: 2725–2741. 10.1002/bit.28378. [DOI] [PubMed] [Google Scholar]
- Cui, G. H. , Guo H.‐D., Li H., et al. 2019. “RVG‐Modified Exosomes Derived From Mesenchymal Stem Cells Rescue Memory Deficits by Regulating Inflammatory Responses in a Mouse Model of Alzheimer's Disease.” Immunity & Ageing 16: 10. 10.1186/s12979-019-0150-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- De Almeida Fuzeta, M. , Bernardes N., Oliveira F. D., et al. 2020. “Scalable Production of Human Mesenchymal Stromal Cell‐Derived Extracellular Vesicles Under Serum‐/Xeno‐Free Conditions in a Microcarrier‐Based Bioreactor Culture System.” Frontiers in Cell and Developmental Biology 8: 553444. 10.3389/fcell.2020.553444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- de Godoy, M. A. , Saraiva L. M., de Carvalho L. R. P., et al. 2018. “Mesenchymal Stem Cells and Cell‐Derived Extracellular Vesicles Protect Hippocampal Neurons From Oxidative Stress and Synapse Damage Induced by Amyloid‐β Oligomers.” Journal of Biological Chemistry 293: 1957–1975. 10.1074/jbc.M117.807180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- De La Torre, J. C 2002. “Alzheimer Disease as a Vascular Disorder: Nosological Evidence.” Stroke 33: 1152–1162. 10.1161/01.str.0000014421.15948.67. [DOI] [PubMed] [Google Scholar]
- Deng, H. , Zhao J., Li J., et al. 2024. “Therapeutic Efficacy of Extracellular Vesicles Derived From Stem Cell for Alzheimer's Disease: A Meta‐Analysis Study.” Frontiers in Bioscience‐Landmark 29: 340. https://doi/org/10.31083/j.fbl2909340. [DOI] [PubMed] [Google Scholar]
- Deng, Z. , Wang J., Xiao Y., et al. 2021. “Ultrasound‐Mediated Augmented Exosome Release From Astrocytes Alleviates Amyloid‐β‐Induced Neurotoxicity.” Theranostics 11: 4351–4362. 10.7150/thno.52436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ding, G. , Li L., Zhang L., et al. 2022. “MRI Metrics of Cerebral Endothelial Cell‐Derived Exosomes for the Treatment of Cognitive Dysfunction Induced in Aging Rats Subjected to Type 2 Diabetes.” Diabetes 71: 873–880. 10.2337/db21-0754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ding, M. , Shen Y., Wang P., et al. 2018. “Exosomes Isolated From Human Umbilical Cord Mesenchymal Stem Cells Alleviate Neuroinflammation and Reduce Amyloid‐Beta Deposition by Modulating Microglial Activation in Alzheimer's Disease.” Neurochemical Research 43: 2165–2177. 10.1007/s11064-018-2641-5. [DOI] [PubMed] [Google Scholar]
- Ding, Y.‐N. , Li M.‐Q., Song J.‐Y., Guan P.‐P., and Wang P.. 2025. “Engineered Exosomes Improve the Effects of Curcumin on Protecting Mitochondria of Neurons in Alzheimer's Disease.” Materials Today Bio 32: 101738. 10.1016/j.mtbio.2025.101738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dolma, L. , Damodaran A., Panonnummal R., and Nair S. C.. 2024. “Exosomes Isolated From Citrus Lemon: a Promising Candidate for the Treatment of Alzheimer's Disease.” Therapeutic Delivery 15: 507–519. 10.1080/20415990.2024.2354119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dou, Y. , Xie J., Tan Y., Zhang M., Zhao Y., and Liu X.. 2021. “Neurotransmitter‐Stimulated Neuron‐Derived sEVs Have Opposite Effects on Amyloid β‐Induced Neuronal Damage.” Journal of Nanobiotechnology 19: 324. 10.1186/s12951-021-01070-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Drummond, E. , and Wisniewski T.. 2017. “Alzheimer's Disease: Experimental Models and Reality.” Acta Neuropathologica 133: 155–175. 10.1007/s00401-016-1662-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Du, B. , Zou Q., Wang X., et al. 2025. “Multi‐Targeted Engineered Hybrid Exosomes as Aβ Nanoscavengers and Inflammatory Modulators for Multi‐pathway Intervention in Alzheimer's Disease.” Biomaterials 322: 123403. 10.1016/j.biomaterials.2025.123403. [DOI] [PubMed] [Google Scholar]
- Ebrahim, N. , Al Saihati H. A., Alali Z., et al. 2024. “Exploring the Molecular Mechanisms of MSC‐Derived Exosomes in Alzheimer's Disease: Autophagy, Insulin and the PI3K/Akt/mTOR Signaling Pathway.” Biomedicine & Pharmacotherapy 176: 116836. 10.1016/j.biopha.2024.116836. [DOI] [PubMed] [Google Scholar]
- Elia, C. A. , Tamborini M., Rasile M., et al. 2019. “Intracerebral Injection of Extracellular Vesicles From Mesenchymal Stem Cells Exerts Reduced Aβ Plaque Burden in Early Stages of a Preclinical Model of Alzheimer's Disease.” Cells 8: 1059. 10.3390/cells8091059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Esmekaya, M. A. , and Ertekin B.. 2024. “Neuroprotective Effects of Coffee‐Derived Exosome‐Like Nanoparticles Against Aβ‐Induced Neurotoxicity.” General Physiology and Biophysics 43: 535–543. 10.4149/gpb_2024025. [DOI] [PubMed] [Google Scholar]
- Etxebeste‐Mitxeltorena, M. , Niza E., Fajardo C. M., et al. 2024. “Neuroprotective Properties of Exosomes and Chitosan Nanoparticles of Tomafran, a Bioengineered Tomato Enriched in Crocins.” Natural Products and Bioprospecting 14: 9. 10.1007/s13659-023-00424-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Évora, A. , Garcia G., Rubi A., et al. 2025. “Exosomes Enriched With miR‐124‐3p Show Therapeutic Potential in a New Microfluidic Triculture Model That Recapitulates Neuron–Glia Crosstalk in Alzheimer's Disease.” Frontiers in Pharmacology 16: 1474012. 10.3389/fphar.2025.1474012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Feng, Y. , Guo M., Zhao H., Han S., Dong Q., and Cui M.. 2020. “Mesenchymal‐Stem‐Cell‐Derived Extracellular Vesicles Mitigate Trained Immunity in the Brain.” Frontiers in Bioengineering and Biotechnology 8: 599058. 10.3389/fbioe.2020.599058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fuller, O. K. , McLennan E. D., Egan C. L., et al. 2025. “Extracellular Vesicles Contribute to the Beneficial Effects of Exercise Training in APP/PS1 Mice.” iScience 28: 111752. 10.1016/j.isci.2025.111752. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gao, G. , Li C., Ma Y., et al. 2023. “Neural Stem Cell‐Derived Extracellular Vesicles Mitigate Alzheimer's Disease‐Like Phenotypes in a Preclinical Mouse Model.” Signal Transduction and Targeted Therapy 8: 228. 10.1038/s41392-023-01436-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Garcia‐Contreras, M. , and Thakor A. S.. 2021. “Human Adipose Tissue‐Derived Mesenchymal Stem Cells and Their Extracellular Vesicles Modulate Lipopolysaccharide Activated human Microglia.” Cell Death Discovery 7: 98. 10.1038/s41420-021-00471-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gatti, M. , Zavatti M., Beretti F., et al. 2020. “Oxidative Stress in Alzheimer's Disease: In Vitro Therapeutic Effect of Amniotic Fluid Stem Cells Extracellular Vesicles.” Oxidative Medicine and Cellular Longevity 2020: 2785343. 10.1155/2020/2785354. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gobin, J. , Muradia G., Mehic J., et al. 2021. “Hollow‐Fiber Bioreactor Production of Extracellular Vesicles From human Bone Marrow Mesenchymal Stromal Cells Yields Nanovesicles That Mirrors the Immuno‐Modulatory Antigenic Signature of the Producer Cell.” Stem Cell Research & Therapy 12: 127. 10.1186/s13287-021-02190-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gu, Z. , Yin Z., Song P., et al. 2022. “Safety and Biodistribution of Exosomes Derived From Human Induced Pluripotent Stem Cells.” Frontiers in Bioengineering and Biotechnology 10: 949724. 10.3389/fbioe.2022.949724. [DOI] [PMC free article] [PubMed] [Google Scholar]
- György, B. , Fitzpatrick Z., Crommentuijn M. H. W., Mu D., and Maguire C. A.. 2014. “Naturally Enveloped AAV Vectors for Shielding Neutralizing Antibodies and Robust Gene Delivery In Vivo.” Biomaterials 35: 7598–7609. 10.1016/j.biomaterials.2014.05.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hamed, G. M. , Morsy W. E., Elnahrawy E. A. B., and Abd‐el‐Hamid M. S.. 2025. “Comparative Study of the Rat‐Labeled PKH26‐MSCs and Its Derived Exosomes as a Neurotherapeutic Approach on LPS‐Induced Alzheimer's Disease.” Regenerative Engineering and Translational Medicine 11: 416–433. 10.1007/s40883-024-00362-0. [DOI] [Google Scholar]
- Han, J. , Sul J. H., Lee J., et al. 2024. “Engineered Exosomes With a Photoinducible Protein Delivery System Enable CRISPR‐Cas–Based Epigenome Editing in Alzheimer's Disease.” Science Translational Medicine 16: eadi4830. 10.1126/scitranslmed.adi4830. [DOI] [PubMed] [Google Scholar]
- Hao, Y. , Su C., Liu X., Sui H., Shi Y., and Zhao L.. 2022. “Bioengineered Microglia‐targeted Exosomes Facilitate Aβ Clearance via Enhancing Activity of Microglial Lysosome for Promoting Cognitive Recovery in Alzheimer's Disease.” Biomaterials Advances 136: 212770. 10.1016/j.bioadv.2022.212770. [DOI] [PubMed] [Google Scholar]
- Haraszti, R. A. , Miller R., Dubuke M. L., et al. 2019. “Serum Deprivation of Mesenchymal Stem Cells Improves Exosome Activity and Alters Lipid and Protein Composition.” iScience 16: 230–241. 10.1016/j.isci.2019.05.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Haroon, J. , Aboody K., Flores L., et al. 2023. “Use of Transcranial Low‐Intensity Focused Ultrasound for Targeted Delivery of Stem Cell‐Derived Exosomes to the Brain.” Scientific Reports 13: 17707. 10.1038/s41598-023-44785-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hassan, R. , Rabea A. A., Ragae A., and Sabry D.. 2020. “The Prospective Role of Mesenchymal Stem Cells Exosomes on Circumvallate Taste Buds in Induced Alzheimer's Disease of Ovariectomized Albino Rats: (Light and Transmission Electron Microscopic Study).” Archives of Oral Biology 110: 104596. 10.1016/j.archoralbio.2019.104596. [DOI] [PubMed] [Google Scholar]
- Hou, G. , Li Y., Cui X., et al. 2024. “Electric Field Assisted Tangential Flow Filtration Device for Highly Effective Isolation of Bioactive Small Extracellular Vesicles From Cell Culture Medium.” Analytical Chemistry 96: 13345–13351. 10.1021/acs.analchem.4c02807. [DOI] [PubMed] [Google Scholar]
- Hou, X. , Jiang H., Liu T., et al. 2023. “Depletion of Gut Microbiota Resistance in 5×FAD Mice Enhances the Therapeutic Effect of Mesenchymal Stem Cell‐derived Exosomes.” Biomedicine & Pharmacotherapy 161: 114455. 10.1016/j.biopha.2023.114455. [DOI] [PubMed] [Google Scholar]
- Hu, G. , Xia Y., Zhang J., et al. 2020. “ESC‐sEVs Rejuvenate Senescent Hippocampal NSCs by Activating Lysosomes to Improve Cognitive Dysfunction in Vascular Dementia.” Advanced Science 7: 1903330. 10.1002/advs.201903330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hu, X. G. , Ma Y. N., Peng J., Wang Z. J., Liang Y. C., and Xia Y.. 2025. “Exosomes Derived From Olfactory Mucosa Mesenchymal Stem Cells Attenuate Cognitive Impairment in a Mouse Model of Alzheimer's Disease.” BioScience Trends 19: 189–201. 10.5582/bst.2025.01065. [DOI] [PubMed] [Google Scholar]
- Huang, L. , Wu E., Liao J., Wei Z., Wang J., and Chen Z.. 2023. “Research Advances of Engineered Exosomes as Drug Delivery Carrier.” ACS Omega 8: 43374–43387. 10.1021/acsomega.3c04479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huang, M. , Zheng M., Song Q., et al. 2024. “Comparative Proteomics Inspired Self‐Stimulated Release Hydrogel Reinforces the Therapeutic Effects of MSC‐EVs on Alzheimer's Disease.” Advanced Materials 36: e2311420. 10.1002/adma.202311420. [DOI] [PubMed] [Google Scholar]
- Huber, C. C. , Callegari E. A., Paez M. D., Romanova S., and Wang H.. 2022. “Heat Shock‐Induced Extracellular Vesicles Derived From Neural Stem Cells Confer Marked Neuroprotection against Oxidative Stress and Amyloid‐β‐Caused Neurotoxicity.” Molecular Neurobiology 59: 7404–7412. 10.1007/s12035-022-03055-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Izadpanah, M. , Dargahi L., Ai J., et al. 2020. “Extracellular Vesicles as a Neprilysin Delivery System Memory Improvement in Alzheimer's Disease.” Iranian Journal of Pharmaceutical Research 19: 45–60. 10.22037/ijpr.2020.112062.13508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jahangard, Y. , Monfared H., Moradi A., Zare M., Mirnajafi‐Zadeh J., and Mowla S. J.. 2020. “Therapeutic Effects of Transplanted Exosomes Containing miR‐29b to a Rat Model of Alzheimer's Disease.” Frontiers in Neuroscience 14: 564. 10.3389/fnins.2020.00564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jain, S. , Bhushan B., Mishra A. K., and Singh R.. 2024. “Unlocking Therapeutic Potential of siRNA‐Based Drug Delivery System for Treatment of Alzheimer's Disease.” Journal of Drug Delivery Science and Technology 102: 106413. 10.1016/j.ddst.2024.106413. [DOI] [Google Scholar]
- Jäkel, L. , De Kort A. M., Klijn C. J. M., Schreuder F. H. B. M., and Verbeek M. M.. 2022. “Prevalence of Cerebral Amyloid Angiopathy: A Systematic Review and Meta‐analysis.” Alzheimer's & Dementia 18: 10–28. 10.1002/alz.12366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jeske, R. , Liu C., Duke L., et al. 2023. “Upscaling Human Mesenchymal Stromal Cell Production in a Novel Vertical‐Wheel Bioreactor Enhances Extracellular Vesicle Secretion and Cargo Profile.” Bioactive Materials 25: 732–747. 10.1016/j.bioactmat.2022.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jiang, S. , Cai G., Yang Z., et al. 2024. “Biomimetic Nanovesicles as a Dual Gene Delivery System for the Synergistic Gene Therapy of Alzheimer's Disease.” ACS Nano 18: 11753–11768. 10.1021/acsnano.3c13150. [DOI] [PubMed] [Google Scholar]
- Johnson, J. , Law S. Q. K., Shojaee M., et al. 2023. “First‐In‐Human Clinical Trial of Allogeneic, Platelet‐Derived Extracellular Vesicles as a Potential Therapeutic for Delayed Wound Healing.” Journal of Extracellular Vesicles 12: 12332. 10.1002/jev2.12332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kamat, P. K. , and Nath C.. 2015. “Okadaic Acid: A Tool to Study Regulatory Mechanisms for Neurodegeneration and Regeneration in Alzheimer′s Disease.” Neural Regeneration Research 10: 365–367. 10.4103/1673-5374.153679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kang, H. , Bae Y.‐H., Kwon Y., Kim S., and Park J.. 2022. “Extracellular Vesicles Generated Using Bioreactors and Their Therapeutic Effect on the Acute Kidney Injury Model.” Advanced Healthcare Materials 11: 2101606. 10.1002/adhm.202101606. [DOI] [PubMed] [Google Scholar]
- Kaniowska, D. , Wenk K., Rademacher P., et al. 2022. “Extracellular Vesicles of Mesenchymal Stromal Cells Can be Taken Up by Microglial Cells and Partially Prevent the Stimulation Induced by β‐Amyloid.” Stem Cell Reviews and Reports 18: 1113–1126. 10.1007/s12015-021-10261-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karttunen, J. , Heiskanen M., Joki T., et al. 2022. “Effect of Cell Culture Media on Extracellular Vesicle Secretion From Mesenchymal Stromal Cells and Neurons.” European Journal of Cell Biology 101: 151270. 10.1016/j.ejcb.2022.151270. [DOI] [PubMed] [Google Scholar]
- Khan, M. I. , Jeong E. S., Khan M. Z., Shin J. H., and Kim J. D.. 2023. “Stem Cells‐Derived Exosomes Alleviate Neurodegeneration and Alzheimer's Pathogenesis by Ameliorating Neuroinflamation, and Regulating the Associated Molecular Pathways.” Scientific Reports 13: 15731. 10.1038/s41598-023-42485-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim, H. Y. , Lee D. K., Chung B.‐R., Kim H. V., and Kim Y.. 2016. “Intracerebroventricular Injection of Amyloid‐β Peptides in Normal Mice to Acutely Induce Alzheimer‐Like Cognitive Deficits.” Journal of Visualized Experiments 109: e53308. 10.3791/53308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Krishnan, B. , Marcatti M., Fracassi A., et al. 2025. “Hippocampal Neural Stem Cell Exosomes Promote Brain Resilience Against the Impact of Tau Oligomers.” Journal of Neuroscience 45: e1664242025. 10.1523/JNEUROSCI.1664-24.2025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kronstadt, S. M. , Patel D. B., Born L. J., et al. 2023. “Mesenchymal Stem Cell Culture Within Perfusion Bioreactors Incorporating 3D‐Printed Scaffolds Enables Improved Extracellular Vesicle Yield With Preserved Bioactivity.” Advanced Healthcare Materials 12: 2300584. 10.1002/adhm.202300584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kumar, M. A. , Baba S. K., Sadida H. Q., et al. 2024. “Extracellular Vesicles as Tools and Targets in Therapy for Diseases.” Signal Transduction and Targeted Therapy 9: 27. 10.1038/s41392-024-01735-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kurkinen, M. , Fułek M., Fułek K., Beszłej J. A., Kurpas D., and Leszek J.. 2023. “The Amyloid Cascade Hypothesis in Alzheimer's Disease: Should We Change Our Thinking?” Biomolecules 13: 453. 10.3390/biom13030453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kwon, H. , Lee E.‐H., Park S.‐Y., et al. 2023. “Lactobacillus‐Derived Extracellular Vesicles Counteract Aβ42‐Induced Abnormal Transcriptional Changes Through the Upregulation of MeCP2 and Sirt1 and Improve Aβ Pathology in Tg‐APP/PS1 Mice.” Experimental & Molecular Medicine 55: 2067–2082. 10.1038/s12276-023-01084-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lai, C. P. , Mardini O., Ericsson M., et al. 2014. “Dynamic Biodistribution of Extracellular Vesicles In Vivo Using a Multimodal Imaging Reporter.” ACS Nano 8: 483–494. 10.1021/nn404945r. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lai, R. C. , Arslan F., Lee M. M., et al. 2010. “Exosome Secreted by MSC Reduces Myocardial Ischemia/Reperfusion Injury.” Stem Cell Research 4: 214–222. 10.1016/j.scr.2009.12.003. [DOI] [PubMed] [Google Scholar]
- Lee, M. , Ban J. J., Yang S., Im W., and Kim M.. 2018. “The Exosome of Adipose‐Derived Stem Cells Reduces β‐Amyloid Pathology and Apoptosis of Neuronal Cells Derived From the Transgenic Mouse Model of Alzheimer's Disease.” Brain Research 1691: 87–93. 10.1016/j.brainres.2018.03.034. [DOI] [PubMed] [Google Scholar]
- Li, B. , Chen Y., Zhou Y., et al. 2024. “Neural Stem Cell‐Derived Exosomes Promote Mitochondrial Biogenesis and Restore Abnormal Protein Distribution in a Mouse Model of Alzheimer's Disease.” Neural Regeneration Research 19: 1593–1601. 10.4103/1673-5374.385839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li, B. , Liu J., Gu G., Han X., Zhang Q., and Zhang W.. 2020. “Impact of Neural Stem Cell‐Derived Extracellular Vesicles on Mitochondrial Dysfunction, Sirtuin 1 Level, and Synaptic Deficits in Alzheimer's Disease.” Journal of Neurochemistry 154: 502–518. 10.1111/jnc.15001. [DOI] [PubMed] [Google Scholar]
- Li, J. , Peng H., Zhang W., et al. 2023. “Enhanced Nose‐to‐Brain Delivery of Combined Small Interfering RNAs Using Lesion‐Recognizing Nanoparticles for the Synergistic Therapy of Alzheimer's Disease.” ACS Applied Materials & Interfaces 15: 53177–53188. 10.1021/acsami.3c08756. [DOI] [PubMed] [Google Scholar]
- Li, N. , Shu J., Yang X., Wei W., and Yan A.. 2022. “Exosomes Derived From M2 Microglia Cells Attenuates Neuronal Impairment and Mitochondrial Dysfunction in Alzheimer's Disease Through the PINK1/Parkin Pathway.” Frontiers in Cellular Neuroscience 16: 874102. 10.3389/fncel.2022.874102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li, P. , Ye L., Sun S., et al. 2024. “Molecular Intersections of Traumatic Brain Injury and Alzheimer's Disease: The Role of ADMSC‐Derived Exosomes and Hub Genes in Microglial Polarization.” Metabolic Brain Disease 40: 77. 10.1007/s11011-024-01503-8. [DOI] [PubMed] [Google Scholar]
- Li, S. , Zhang J., Liu X., et al. 2024. “Proteomic Characterization of hUC‐MSC Extracellular Vesicles and Evaluation of Its Therapeutic Potential to Treat Alzheimer's Disease.” Scientific Reports 14: 5959. 10.1038/s41598-024-56549-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liang, X. , Gupta D., Xie J., et al. 2025. “Engineering of Extracellular Vesicles for Efficient Intracellular Delivery of Multimodal Therapeutics Including Genome Editors.” Nature Communications 16: 4028. 10.1038/s41467-025-59377-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liang, X. , Niu Z., Galli V., et al. 2022. “Extracellular Vesicles Engineered to Bind Albumin Demonstrate Extended Circulation Time and Lymph Node Accumulation in Mouse Models.” Journal of Extracellular Vesicles 11: e12248. 10.1002/jev2.12248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lin, E.‐Y. , Hsu S.‐X., Wu B.‐H., et al. 2024. “Engineered Exosomes Containing microRNA‐29b‐2 and Targeting the Somatostatin Receptor Reduce Presenilin 1 Expression and Decrease the β‐Amyloid Accumulation in the Brains of Mice With Alzheimer's Disease.” International Journal of Nanomedicine 19: 4977–4994. 10.2147/IJN.S442876. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lin, L. , Huang L., Huang S., et al. 2024. “MSC‐Derived Extracellular Vesicles Alleviate NLRP3/GSDMD‐Mediated Neuroinflammation in Mouse Model of Sporadic Alzheimer's Disease.” Molecular Neurobiology 61: 5494–5509. 10.1007/s12035-024-03914-1. [DOI] [PubMed] [Google Scholar]
- Liu, C. , Helsper S., Marzano M., et al. 2022. “Human Forebrain Organoid‐Derived Extracellular Vesicle Labeling With Iron Oxides for In Vitro Magnetic Resonance Imaging.” Biomedicines 10: 3060. 10.3390/biomedicines10123060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu, H. , Jin M., Ji M., Zhang W., Liu A., and Wang T.. 2022. “Hypoxic Pretreatment of Adipose‐Derived Stem Cell Exosomes Improved Cognition by Delivery of Circ‐Epc1 and Shifting Microglial M1/M2 Polarization in an Alzheimer's Disease Mice Model.” Aging 14: 3070–3083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu, S. , Fan M., Xu J. X., et al. 2022. “Exosomes Derived From Bone‐Marrow Mesenchymal Stem Cells Alleviate Cognitive Decline in AD‐Like Mice by Improving BDNF‐related Neuropathology.” Journal of Neuroinflammation 19: 35. 10.1186/s12974-022-02393-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu, Y. , Huber C. C., and Wang H.. 2020. “Disrupted Blood‐Brain Barrier in 5×FAD Mouse Model of Alzheimer's Disease Can be Mimicked and Repaired In Vitro With Neural Stem Cell‐Derived Exosomes.” Biochemical and Biophysical Research Communications 525: 192–196. 10.1016/j.bbrc.2020.02.074. [DOI] [PubMed] [Google Scholar]
- Losurdo, M. , Pedrazzoli M., D'Agostino C., et al. 2020. “Intranasal Delivery of Mesenchymal Stem Cell‐derived Extracellular Vesicles Exerts Immunomodulatory and Neuroprotective Effects in a 3xTg Model of Alzheimer's Disease.” Stem Cells Translational Medicine 9: 1068–1084. 10.1002/sctm.19-0327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ma, X. , Huang M., Zheng M., et al. 2020. “ADSCs‐Derived Extracellular Vesicles Alleviate Neuronal Damage, Promote Neurogenesis and Rescue Memory Loss in Mice With Alzheimer's Disease.” Journal of Control Release 327: 688–702. 10.1016/j.jconrel.2020.09.019. [DOI] [PubMed] [Google Scholar]
- Ma, X. , Wang Y., Shi Y., et al. 2022. “Exosomal miR‐132‐3p From Mesenchymal Stromal Cells Improves Synaptic Dysfunction and Cognitive Decline in Vascular Dementia.” Stem Cell Research & Therapy 13: 315. 10.1186/s13287-022-02995-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Madhu, L. N. , Kodali M., Upadhya R., et al. 2024. “Extracellular Vesicles From Human‐Induced Pluripotent Stem Cell‐Derived Neural Stem Cells Alleviate Proinflammatory Cascades Within Disease‐Associated Microglia in Alzheimer's Disease.” Journal of Extracellular Vesicles 13: e12519. 10.1002/jev2.12519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Markoutsa, E. , Mayilsamy K., Gulick D., Mohapatra S. S., and Mohapatra S.. 2022. “Extracellular Vesicles Derived From Inflammatory‐Educated Stem Cells Reverse Brain Inflammation‐Implication of miRNAs.” Molecular Therapy 30: 816–830. 10.1016/j.ymthe.2021.08.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Marzano, M. , Bejoy J., Cheerathodi M. R., et al. 2019. “Differential Effects of Extracellular Vesicles of Lineage‐Specific Human Pluripotent Stem Cells on the Cellular Behaviors of Isogenic Cortical Spheroids.” Cells 8: 993. 10.3390/cells8090993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Micci, M. A. , Krishnan B., Bishop E., et al. 2019. “Hippocampal Stem Cells Promotes Synaptic Resistance to the Dysfunctional Impact of Amyloid Beta Oligomers via Secreted Exosomes.” Molecular Neurodegeneration 14: 25. 10.1186/s13024-019-0322-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morita, Y. , Izawa H., Ohga H., et al. 2024. “Safety and Clinical Efficacy on Intranasal Administration of Mesenchymal Stem Cell‐Derived Secretome in Patients With Alzheimer's Disease and Its Future Prospect.” Glycative Stress Research 11, no. 3: 103–110. 10.24659/gsr.11.3_103. [DOI] [Google Scholar]
- Nakano, M. , Nagaishi K., Konari N., et al. 2016. “Bone Marrow‐Derived Mesenchymal Stem Cells Improve Diabetes‐induced Cognitive Impairment by Exosome Transfer Into Damaged Neurons and Astrocytes.” Scientific Reports 6: 24805. 10.1038/srep24805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nakase, I. , Ueno N., Matsuzawa M., et al. 2021. “Environmental pH Stress Influences Cellular Secretion and Uptake of Extracellular Vesicles.” FEBS Open Bio 11: 753–767. 10.1002/2211-5463.13107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Oliveira Silva, R. , Counil H., Rabanel J.‐M., et al. 2024. “Donepezil‐Loaded Nanocarriers for the Treatment of Alzheimer's Disease: Superior Efficacy of Extracellular Vesicles Over Polymeric Nanoparticles.” International Journal of Nanomedicine 19: 1077–1096. 10.2147/IJN.S449227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Page, M. J. , McKenzie J. E., Bossuyt P. M., et al. 2021. “The PRISMA 2020 Statement: an Updated Guideline for Reporting Systematic Reviews.” BMJ 372: n71. 10.1136/bmj.n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pan, J. , He R., Huo Q., Shi Y., and Zhao L.. 2020. “Brain Microvascular Endothelial Cell Derived Exosomes Potently Ameliorate Cognitive Dysfunction by Enhancing the Clearance of Aβ Through Up‐Regulation of P‐gp in Mouse Model of AD.” Neurochemical Research 45: 2161–2172. 10.1007/s11064-020-03076-1. [DOI] [PubMed] [Google Scholar]
- Patel, D. B. , Gray K. M., Santharam Y., Lamichhane T. N., Stroka K. M., and Jay S. M.. 2017. “Impact of Cell Culture Parameters on Production and Vascularization Bioactivity of Mesenchymal Stem Cell‐Derived Extracellular Vesicles.” Bioengineering & Translational Medicine 2: 170–179. 10.1002/btm2.10065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perets, N. , Betzer O., Shapira R., et al. 2019. “Golden Exosomes Selectively Target Brain Pathologies in Neurodegenerative and Neurodevelopmental Disorders.” Nano Letters 19: 3422–3431. 10.1021/acs.nanolett.8b04148. [DOI] [PubMed] [Google Scholar]
- Pérez, L. M. , de Lucas B., and Gálvez B. G.. 2018. “Unhealthy Stem Cells: When Health Conditions Upset Stem Cell Properties.” Cellular Physiology and Biochemistry 46: 1999–2016. 10.1159/000489440. [DOI] [PubMed] [Google Scholar]
- Pham, C. V. , Midge S., Barua H., et al. 2021. “Bovine Extracellular Vesicles Contaminate Human Extracellular Vesicles Produced in Cell Culture Conditioned Medium When ‘Exosome‐Depleted Serum’ Is Utilised.” Archives of Biochemistry and Biophysics 708: 108963. 10.1016/j.abb.2021.108963. [DOI] [PubMed] [Google Scholar]
- Phelps, J. , Hart D. A., Mitha A. P., Duncan N. A., and Sen A.. 2023. “Physiological Oxygen Conditions Enhance the Angiogenic Properties of Extracellular Vesicles From Human Mesenchymal Stem Cells.” Stem Cell Research & Therapy 14: 218. 10.1186/s13287-023-03439-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Phelps, J. , Hart D. A., Mitha A. P., Duncan N. A., and Sen A.. 2024. “Extracellular Vesicles Generated by Mesenchymal Stem Cells in Stirred Suspension Bioreactors Promote Angiogenesis in Human‐Brain‐Derived Endothelial Cells.” International Journal of Molecular Sciences 25: 5219. 10.3390/ijms25105219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Phelps, J. , Leonard C., Shah S., et al. 2022. “Production of Mesenchymal Progenitor Cell‐Derived Extracellular Vesicles in Suspension Bioreactors for Use in Articular Cartilage Repair.” Stem Cells Translational Medicine 11: 73–87. 10.1093/stcltm/szab008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Phelps, J. , Sanati‐Nezhad A., Ungrin M., Duncan N. A., and Sen A.. 2018. “Bioprocessing of Mesenchymal Stem Cells and Their Derivatives: Toward Cell‐Free Therapeutics.” Stem Cells International 2018: 9415367. 10.1155/2018/9415367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Poltavtseva, R. A. , Bobkova N. V., Zhdanova D. Y., Svirshchevskaya E. V., and Sukhikh G. T.. 2021. “Alzheimer's Type Neurodegeneration. Possible Correction of Memory Impairment With Intravenous Administration of Exosomes.” Biochemistry (Moscow) Supplement Series A ‐Membrane and Cell Biology 15: 306–318. 10.1134/S1990747821050068. [DOI] [Google Scholar]
- Pourhadi, M. , Zali H., Ghasemi R., Faizi M., Mojab F., and Zomorrod M. S.. 2024. “Restoring Synaptic Function: How Intranasal Delivery of 3D‐Cultured hUSSC Exosomes Improve Learning and Memory Deficits in Alzheimer's Disease.” Molecular Neurobiology 61: 3724–3741. 10.1007/s12035-023-03733-w. [DOI] [PubMed] [Google Scholar]
- Qi, D. , Hou X., Jin C., et al. 2021. “HNSC Exosome‐Derived MIAT Improves Cognitive Disorders in Rats With Vascular Dementia via the miR‐34b‐5p/CALB1 Axis.” American Journal of Translational Research 13: 10075–10093. [PMC free article] [PubMed] [Google Scholar]
- Qi, Y. , Guo L., Jiang Y., Shi Y., Sui H., and Zhao L.. 2020. “Brain Delivery of Quercetin‐Loaded Exosomes Improved Cognitive Function in AD Mice by Inhibiting Phosphorylated Tau‐Mediated Neurofibrillary Tangles.” Drug Delivery 27: 745–755. 10.1080/10717544.2020.1762262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rao, S. , Madhu L. N., Babu R. S., et al. 2025. “Extracellular Vesicles From hiPSC‐derived NSCs Protect Human Neurons Against Aβ‐42 Oligomers Induced Neurodegeneration, Mitochondrial Dysfunction and Tau Phosphorylation.” Stem Cell Research & Therapy 16: 191. 10.1186/s13287-025-04324-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rayamajhi, S. , Sulthana S., Ferrel C., Shrestha T. B., and Aryal S.. 2023. “Extracellular Vesicles Production and Proteomic Cargo Varies With Incubation Time and Temperature.” Experimental Cell Research 422: 113454. 10.1016/j.yexcr.2022.113454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rezayof, A. , Ghasemzadeh Z., and Kouhkan F.. 2024. “microRNAs Mediate Signaling Pathways in Alzheimer's Disease: Biomarkers and Therapeutic Targets.” In The Textbook of Nanoneuroscience and Nanoneurosurgery, edited by Kateb B., Heiss J. D., and Yu J. S., 657–686. Springer Nature Switzerland. 10.1007/978-3-030-80662-0_39. [DOI] [Google Scholar]
- Reza‐Zaldivar, E. E. , Hernández‐Sapiéns M. A., Gutiérrez‐Mercado Y. K., et al. 2019. “Mesenchymal Stem Cell‐derived Exosomes Promote Neurogenesis and Cognitive Function Recovery in a Mouse Model of Alzheimer's Disease.” Neural Regeneration Research 14: 1626–1634. 10.4103/1673-5374.255978. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ruan, Y. , Qiu X., Lv Y.‐D., et al. 2019. “Kainic Acid Induces Production and Aggregation of Amyloid β‐protein and Memory Deficits by Activating Inflammasomes in NLRP3‐ and NF‐κB‐Stimulated Pathways.” Aging 11: 3795–3810. 10.18632/aging.102017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sadeghi, A. , Noorbakhshnia M., and Khodashenas S.. 2025. “Protective Potential of BM‐MSC Extracted Exosomes in a Rat Model of Alzheimer's Disease.” PLoS ONE 20: e0320883. 10.1371/journal.pone.0320883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Saleem, S. , and Kannan R. R.. 2018. “Zebrafish: an Emerging Real‐Time Model System to Study Alzheimer's Disease and Neurospecific Drug Discovery.” Cell Death Discovery 4: 45. 10.1038/s41420-018-0109-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sha, S. , Shen X., Cao Y., and Qu L.. 2021. “Mesenchymal Stem Cells‐Derived Extracellular Vesicles Ameliorate Alzheimer's Disease in Rat Models via the microRNA‐29c‐3p/BACE1 Axis and the Wnt/β‐catenin Pathway.” Aging 13: 15285–15306. 10.18632/aging.203088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shany, S. , Bernheimer A. W., Grushoff P. S., and Kim K. S.. 1974. “Evidence for Membrane Cholesterol as the Common Binding Site for Cereolysin, Streptolysin O and Saponin.” Molecular and Cellular Biochemistry 3: 179–186. 10.1007/BF01686643. [DOI] [PubMed] [Google Scholar]
- Sharma, A. , Sances S., Workman M. J., and Svendsen C. N.. 2020. “Multi‐Lineage Human iPSC‐Derived Platforms for Disease Modeling and Drug Discovery.” Cell Stem Cell 26: 309–329. 10.1016/j.stem.2020.02.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sharma, M. , Sheth M., Poling H. M., Kuhnell D., Langevin S. M., and Esfandiari L.. 2023. “Rapid Purification and Multiparametric Characterization of Circulating Small Extracellular Vesicles Utilizing a Label‐Free Lab‐on‐a‐Chip Device.” Scientific Reports 13: 18293. 10.1038/s41598-023-45409-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shekari, F. , Alibhai F. J., Baharvand H., et al. 2023. “Cell Culture‐Derived Extracellular Vesicles: Considerations for Reporting Cell Culturing Parameters.” Journal of Extracellular Biology 2: e115. 10.1002/jex2.115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sheykhhasan, M. , Amini R., Soleimani Asl S., Saidijam M., Hashemi S. M., and Najafi R.. 2022. “Neuroprotective Effects of Coenzyme Q10‐Loaded Exosomes Obtained From Adipose‐Derived Stem Cells in a Rat Model of Alzheimer's Disease.” Biomedicine & Pharmacotherapy 152: 113224. 10.1016/j.biopha.2022.113224. [DOI] [PubMed] [Google Scholar]
- Shunan, D. , Yu M., Guan H., and Zhou Y.. 2021. “Neuroprotective Effect of Betalain Against AlCl3‐Induced Alzheimer's Disease in Sprague Dawley Rats via Putative Modulation of Oxidative Stress and Nuclear Factor Kappa B (NF‐κB) Signaling Pathway.” Biomedicine & Pharmacotherapy 137: 111369. 10.1016/j.biopha.2021.111369. [DOI] [PubMed] [Google Scholar]
- Silva, R. O. , Haddad M., Counil H., et al. 2025. “Exploring the Potential of Plasma and Adipose Mesenchymal Stem Cell‐derived Extracellular Vesicles as Novel Platforms for Neuroinflammation Therapy.” Journal of Controlled Release 377: 880–898. 10.1016/j.jconrel.2024.11.060. [DOI] [PubMed] [Google Scholar]
- Solis Jr, E. , Hascup K. N., and Hascup E. R.. 2020. “Alzheimer's Disease: The Link Between Amyloid‐β and Neurovascular Dysfunction.” Journal of Alzheimer's Disease 76: 1179–1198. 10.3233/JAD-200473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sun, C. , Sha S., Shan Y., et al. 2025. “Intranasal Delivery of BACE1 siRNA and Berberine via Engineered Stem Cell Exosomes for the Treatment of Alzheimer's Disease.” International Journal of Nanomedicine 20: 5873–5891. 10.2147/IJN.S506793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Takahashi, Y. , Nishikawa M., Shinotsuka H., et al. 2013. “Visualization and In Vivo Tracking of the Exosomes of Murine Melanoma B16‐BL6 Cells in Mice After Intravenous Injection.” Journal of Biotechnology 165: 77–84. 10.1016/j.jbiotec.2013.03.013. [DOI] [PubMed] [Google Scholar]
- Thanaskody, K. , Jusop A. S., Tye G. J., Zaman W. S. W. K., Dass S. A., and Nordin F.. 2022. “MSCs vs. iPSCs: Potential in Therapeutic Applications.” Frontiers in Cell and Developmental Biology 10: 1005926. 10.3389/fcell.2022.1005926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Théry, C. , Witwer K. W., Aikawa E., et al. 2018. “Minimal Information for Studies of Extracellular Vesicles 2018 (MISEV2018): A Position Statement of the International Society for Extracellular Vesicles and Update of the MISEV2014 Guidelines.” Journal of Extracellular Vesicles 7: 1535750. 10.1080/20013078.2018.1535750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Van Delen, M. , Derdelinckx J., Wouters K., Nelissen I., and Cools N.. 2024. “A Systematic Review and Meta‐analysis of Clinical Trials Assessing Safety and Efficacy of human Extracellular Vesicle‐based Therapy.” Journal of Extracellular Vesicles 13: e12458. 10.1002/jev2.12458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Van Dyck, C. H. , Swanson C. J., Aisen P., et al. 2023. “Lecanemab in Early Alzheimer's Disease.” New England Journal of Medicine 388: 9–21. 10.1056/NEJMoa2212948. [DOI] [PubMed] [Google Scholar]
- Venugopal, C. , Shamir C., Senthilkumar S., et al. 2018. “Dosage and Passage Dependent Neuroprotective Effects of Exosomes Derived From Rat Bone Marrow Mesenchymal Stem Cells: An In Vitro Analysis.” Current Gene Therapy 18. 10.2174/1566523218666180125091952. [DOI] [PubMed] [Google Scholar]
- Wan, Z. , Liu T., Xu N., et al. 2024. “PKH Dyes Should Be Avoided in the EVs Biodistribution Study of the Brain: a Call for Caution.” International Journal of Nanomedicine 19: 10885–10898. 10.2147/IJN.S475060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang, C. , Yang Y., Zhang X., et al. 2023. “Secreted Endogenous Macrosomes Reduce Aβ Burden and Ameliorate Alzheimer's Disease.” Science Advances 9: eade0293. 10.1126/sciadv.ade0293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang, H. , Liu Y., Li J., et al. 2021. “Tail‐vein Injection of MSC‐Derived Small Extracellular Vesicles Facilitates the Restoration of Hippocampal Neuronal Morphology and Function in APP /PS1 Mice.” Cell Death Discovery 7: 230. 10.1038/s41420-021-00620-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang, H. , Sui H., Zheng Y., et al. 2019. “Curcumin‐Primed Exosomes Potently Ameliorate Cognitive Function in AD Mice by Inhibiting Hyperphosphorylation of the Tau Protein Through the AKT/GSK‐3β Pathway.” Nanoscale 11: 7481–7496. 10.1039/c9nr01255a. [DOI] [PubMed] [Google Scholar]
- Wang, J. , Chen D., and Ho E. A.. 2021. “Challenges in the Development and Establishment of Exosome‐Based Drug Delivery Systems.” Journal of Controlled Release 329: 894–906. 10.1016/j.jconrel.2020.10.020. [DOI] [PubMed] [Google Scholar]
- Wang, P. , Yi T., Mao S., and Li M.. 2023. “Neuroprotective Mechanism of Human Umbilical Cord Mesenchymal Stem Cell‐Derived Extracellular Vesicles Improving the Phenotype Polarization of Microglia via the PI3K/AKT/Nrf2 Pathway in Vascular Dementia.” Synapse 77: e22268. 10.1002/syn.22268. [DOI] [PubMed] [Google Scholar]
- Wang, S. S. , Jia J., and Wang Z.. 2018. “Mesenchymal Stem Cell‐Derived Extracellular Vesicles Suppresses iNOS Expression and Ameliorates Neural Impairment in Alzheimer's Disease Mice.” Journal of Alzheimer's Disease 61: 1005–1013. 10.3233/JAD-170848. [DOI] [PubMed] [Google Scholar]
- Wang, X. , and Yang G.. 2021. “Bone Marrow Mesenchymal Stem Cells‐Derived Exosomes Reduce Aβ Deposition and Improve Cognitive Function Recovery in Mice With Alzheimer's Disease by Activating Sphingosine Kinase/Sphingosine‐1‐Phosphate Signaling Pathway.” Cell Biology International 45: 775–784. 10.1002/cbin.11522. [DOI] [PubMed] [Google Scholar]
- Wang, Y. , Worrell G. A., and Wang H.‐L.. 2023. “It Is the Frequency That Matters — Effects of Electromagnetic Fields on the Release and Content of Extracellular Vesicles.” Preprint, bioRxiv, August 11. 10.1101/2023.08.08.552505. [DOI]
- Wang, Y. , Zhao J., Guo Q. et al. 2025. “Neural Stem Cell‐Derived Exosomes Improve Neurite Outgrowth and Cognitive Function Through Transferring miR‐132‐3p.” Experimental Neurology 388: 115224. 10.1016/j.expneurol.2025.115224. [DOI] [PubMed] [Google Scholar]
- Wang, Z. , Gao C., Zhang L., and Sui R.. 2024. “Novel Combination of Olesoxime/Resveratrol‐Encapsulated Exosomes to Improve Cognitive Function by Targeting Amyloid β‐Induced Alzheimer's Disease: Investigation on In Vitro and In Vivo Model.” Inflammopharmacology 32: 2613–2628. 10.1007/s10787-024-01476-1. [DOI] [PubMed] [Google Scholar]
- Wang, Z. , Rich J., Hao N., et al. 2022. “Acoustofluidics for Simultaneous Nanoparticle‐Based Drug Loading and Exosome Encapsulation.” Microsystems & Nanoengineering 8: 45. 10.1038/s41378-022-00374-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Webb, R. L. , Kaiser E. E., Scoville S. L., et al. 2018. “Human Neural Stem Cell Extracellular Vesicles Improve Tissue and Functional Recovery in the Murine Thromboembolic Stroke Model.” Translational Stroke Research 9: 530–539. 10.1007/s12975-017-0599-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Welsh, J. A. , Goberdhan D. C. I., O'Driscoll L., et al. 2024. “Minimal Information for Studies of Extracellular Vesicles (MISEV2023): From Basic to Advanced Approaches.” Journal of Extracellular Vesicles 13: e12404. 10.1002/jev2.12404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xie, X. , Song Q., Dai C., et al. 2023. “Clinical Safety and Efficacy of Allogenic Human Adipose Mesenchymal Stromal Cells‐Derived Exosomes in Patients With Mild to Moderate Alzheimer's Disease: A Phase I/II Clinical Trial.” General Psychiatry 36: e101143. 10.1136/gpsych-2023-101143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xiong, W. P. , Yao W.‐Q., Wang B., and Liu K.. 2021. “BMSCs‐Exosomes Containing GDF‐15 Alleviated SH‐SY5Y Cell Injury Model of Alzheimer's Disease via AKT/GSK‐3β/β‐Catenin.” Brain Research Bulletin 177: 92–102. 10.1016/j.brainresbull.2021.09.008. [DOI] [PubMed] [Google Scholar]
- Xu, F. , Wu Y., Yang Q., et al. 2022. “Engineered Extracellular Vesicles With SHP2 High Expression Promote Mitophagy for Alzheimer's Disease Treatment.” Advanced Materials 34: e2207107. 10.1002/adma.202207107. [DOI] [PubMed] [Google Scholar]
- Yan, L. , and Wu X.. 2020. “Exosomes Produced From 3D Cultures of Umbilical Cord Mesenchymal Stem Cells in a Hollow‐Fiber Bioreactor Show Improved Osteochondral Regeneration Activity.” Cell Biology and Toxicology 36: 165–178. 10.1007/s10565-019-09504-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yang, D. , Deng Z., Zhou H., Zhang Q., Zhang X., and Gong J.. 2025. “Exosome‐Mediated Dual Drug Delivery of Curcumin and Methylene Blue for Enhanced Cognitive Function and Mechanistic Elucidation in Alzheimer's Disease Therapy.” Frontiers in Cell and Developmental Biology 13: 1562565. 10.3389/fcell.2025.1562565. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yang, L. , Zhai Y., Hao Y., Zhu Z., and Cheng G.. 2020. “The Regulatory Functionality of Exosomes Derived From hUMSCs in 3D Culture for Alzheimer's Disease Therapy.” Small 16: e1906273. 10.1002/smll.201906273. [DOI] [PubMed] [Google Scholar]
- Yin, T. , Liu Y., Ji W., et al. 2023. “Engineered Mesenchymal Stem Cell‐Derived Extracellular Vesicles: A State‐of‐the‐Art Multifunctional Weapon Against Alzheimer's Disease.” Theranostics 13: 1264–1285. 10.7150/thno.81860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yu, Y. , Li W., Mao L., et al. 2021. “Genetically Engineered Exosomes Display RVG Peptide and Selectively Enrich a Neprilysin Variant: a Potential Formulation for the Treatment of Alzheimer's Disease.” Journal of Drug Targeting 29: 1128–1138. 10.1080/1061186X.2021.1929257. [DOI] [PubMed] [Google Scholar]
- Zavatti, M. , Gatti M., Beretti F., Palumbo C., and Maraldi T.. 2022. “Exosomes Derived From Human Amniotic Fluid Mesenchymal Stem Cells Preserve Microglia and Neuron Cells From Aβ.” International Journal of Molecular Sciences 23: 4967. 10.3390/ijms23094967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhai, L. , Shen H., Sheng Y., and Guan Q.. 2021. “ADMSC Exo‐MicroRNA‐22 Improve Neurological Function and Neuroinflammation in Mice With Alzheimer's Disease.” Journal of Cellular and Molecular Medicine 25: 7513–7523. 10.1111/jcmm.16787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhang, L. 2025. “MicroRNA‐214‐3p Delivered by Bone Marrow Mesenchymal Stem Cells‐Secreted Exosomes Affects Oxidative Stress in Alzheimer's Disease Rats by Targeting CD151.” Organogenesis 21: 2489673. 10.1080/15476278.2025.2489673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhang, L. , Li C., Huang R., et al. 2022. “Cerebral Endothelial Cell Derived Small Extracellular Vesicles Improve Cognitive Function in Aged Diabetic Rats.” Frontiers in Aging Neuroscience 14: 926485. 10.3389/fnagi.2022.926485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhang, L. , Lin J., Xiang K., Shi T., and Guo B.. 2024. “Omnidirectional Improvement of Mitochondrial Health in Alzheimer's Disease by Multi‐Targeting Engineered Activated Neutrophil Exosomes.” Journal of Controlled Release 376: 470–487. 10.1016/j.jconrel.2024.10.033. [DOI] [PubMed] [Google Scholar]
- Zhang, X.‐X. , Tian Y., Wang Z.‐T., Ma Y.‐H., Tan L., and Yu J.‐T.. 2021. “The Epidemiology of Alzheimer's Disease Modifiable Risk Factors and Prevention.” Journal of Prevention of Alzheimer's Disease 8, no. 3: 313–321. 10.14283/jpad.2021.15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhang, Y. , Zhang X., Zhou J., Li Y., Kai T., and Zhang L.. 2025. “Lycium Ruthenicum Murray Exosome‐Like Nanovesicles Alleviated Alzheimer's Disease–Like Symptoms Induced by Aβ Protein in Transgenic Caenorhabditis Elegans Through the DAF‐16 Pathway.” International Journal of Biological Macromolecules 304: 140758. 10.1016/j.ijbiomac.2025.140758. [DOI] [PubMed] [Google Scholar]
- Zhao, W. , Zhang H., Yan J., and Ma X.. 2019. “An Experimental Study on the Treatment of Diabetes‐Induced Cognitive Disorder Mice Model With Exosomes Deriving From Mesenchymal Stem Cells (MSCs).” Pakistan Journal of Pharmaceutical Sciences 32: 1965–1970. [PubMed] [Google Scholar]
- Zhao, X. , Ge P., Lei S., et al. 2023. “An Exosome‐Based Therapeutic Strategy Targeting Neuroinflammation in Alzheimer's Disease With Berberine and Palmatine.” Drug Design, Development and Therapy 17: 2401–2420. 10.2147/DDDT.S417465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhdanova, D. Y. , Poltavtseva R. A., Svirshchevskaya E. V., and Bobkova N. V.. 2021. “Effect of Intranasal Administration of Multipotent Mesenchymal Stromal Cell Exosomes on Memory of Mice in Alzheimer's Disease Model.” Bulletin of Experimental Biology and Medicine 170: 575–582. 10.1007/s10517-021-05109-3. [DOI] [PubMed] [Google Scholar]
- Zheng, W. , Schürz M., Wiklander R. J., et al. 2023. “Surface Display of Functional Moieties on Extracellular Vesicles Using Lipid Anchors.” Journal of Controlled Release 357: 630–640. 10.1016/j.jconrel.2023.04.033. [DOI] [PubMed] [Google Scholar]
- Zhou, P. , Chao Q., Li C., et al. 2025. “Microglia‐Targeting Nanosystems That Cooperatively Deliver Chinese Herbal Ingredients Alleviate Behavioral and Cognitive Deficits in Alzheimer's Disease Model Mice.” Journal of Nanobiotechnology 23: 313. 10.1186/s12951-025-03385-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhu, Y. , Huang R., Wang D., et al. 2023. “EVs‐Mediated Delivery of CB2 Receptor Agonist for Alzheimer's Disease Therapy.” Asian Journal of Pharmaceutical Sciences 18: 100835. 10.1016/j.ajps.2023.100835. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Excel file: jex270077‐sup‐0001‐Tables.xlsx
Supplementary Table: jex270077‐sup‐0002‐SuppMat.docx
Data Availability Statement
The data that supports the findings of this study are available in the supplementary material of this article.