Blocking DP1 Receptors Reduces Brain Immune Cell Aging and Memory Loss Linked to PTGDS Protein in Tiny Particles from Older Brains

To investigate the impact of brain‐derived exosomes from aged mice on the cognitive function of normal young mice, we collected exosomes sourced from the brains of aged mice and employed transmission electron microscopy (TEM) and Nanoparticle Tracking Analysis (NTA) to measure their morphology and size (Figure 1A). Subsequently, these brain‐derived exosomes from aged mice were injected into the hippocampus of normal young mice, and we assessed the Y‐maze performance of mice at 36 days and evaluated the effect of A‐exo on recognition memory by calculating alternation performance. Compared to the Sham group, the alternation performance of the Sham+A‐exo group significantly decreased, indicating impaired memory. In comparison to the Sham group, there was no decline in cognitive function in the Sham+Y‐exo group (Figure 1B).

Between days 37 and 42 after injection of exosomes in the hippocampus, the water maze was performed. Mice in the Sham+A‐exo group exhibited impaired spatial learning and memory compared to the Sham group. In comparison to the Sham group, there was no decline in cognitive function in the Sham+Y‐exo group, as evidenced by their swimming paths and learning curves (Figure 1C,D). During the training phase, mice of the Sham+Y‐exo group showed a decrease in escape latency, while mice in the Sham+A‐exo group exhibited significantly longer escape latencies compared to sham‐treated mice (Figure 1C). In the probe test, the number of crossings over the target quadrant platform significantly decreased in the Sham+A‐exo group (Figure 1F), while the latency to first platform arrival increased (Figure 1G), indicating impaired spatial learning and memory function. Compared to the Sham+A‐exo group, mice in the Sham+Y‐exo group spent significantly more time in the target quadrant (Figure 1H) and had reduced latency to first platform arrival (Figure 1G). Additionally, there were no significant differences in average swimming speed among the three groups of mice (Figure 1E). Immunofluorescence analysis demonstrated significant upregulation of Iba1 and GFAP expression in the hippocampal regions of Sham+A‐exo group brains compared with Sham+Vehicle controls (Figure 1I–L). Additionally, Western blot (WB) results showed that compared with the control group, A‐exo reduced the expression of the synaptic‐associated protein PSD95 in the brain (Figure S1A,B). In conclusion, mice treated with A‐exo showed impaired spatial learning and memory function compared to the Sham group. Conversely, mice treated with Y‐exo demonstrated no decline in spatial learning and memory function.

To investigate the causes of cognitive decline in normal young mice induced by brain‐derived exosomes from aged mice, we conducted proteomic sequencing of aged mouse brain‐derived exosomes (A‐exo) and young mouse brain‐derived exosomes (Y‐exo). Proteomic sequencing results revealed that, based on the selection criteria of |log2(FC)| ≥ 1.00 and p value ≤ 0.05 during the differential expression protein analysis, there were 1066 differentially expressed proteins between A‐exo and Y‐exo. Compared with Y‐exo, A‐exo had 726 upregulated proteins and 340 downregulated proteins (Figure 2A). To further elucidate the function of the significantly differentially expressed protein PTGDS, bioinformatics enrichment analysis was performed, revealing that PTGDS is primarily involved in immune system processes and immune responses (Figure 2B). Figure 2C displays the differentially expressed proteins related to immune system processes and immune responses. GSEA enrichment analysis of differential proteins found that the top five KEGG enrichment analyses were ALZHEIMERS_DISEASE, CALCIUM_SIGNALING_PATHWAY, PARKINSONS_DISEASE, PEROXISOME, and PROTEASOME (Figure 2D), which are significantly related to cognitive function. The Deeply Integrated human Single‐Cell Omics data (DISCO) database indicated that the main cells expressing PTGDS in the brain include astrocytes, oligodendrocytes, endothelial cells, and neurons (Figure 2E). In patients with Alzheimer's disease, PTGDS is primarily expressed in oligodendrocytes (Figure 2F). The AgeAnno single‐cell database showed that in the elderly, PTGDS is mainly expressed in astrocytes, oligodendrocytes, and endothelial cells (Figure 2G). To ascertain whether A‐exo leads to an increase in PGD2, the ligand for DP1, in the brain and periphery of mice, we conducted Elisa assays for PGD2 in mouse brain tissue and peripheral blood after exosome intervention. The results indicated that the A‐exo intervention group significantly increased PGD2 levels in both mouse brain tissue and peripheral blood (Figure 2H).

Data from HU_0127_Cerebral‐Cortex_GSE140231 indicated that the DP1 receptor is mainly expressed in microglia in the cerebral cortex (Figure 2I). To determine the effect of aged mouse brain‐derived exosomes (A‐exo) on the expression of DP1 receptors in microglia, we stimulated BV2 microglia cultured in vitro with A‐exo. Flow cytometry results showed that the expression of DP1 receptors was significantly increased in the Con+A‐exo group compared with the Con+Y‐exo group (Figure 2J).

To investigate whether A‐exo directly acts on the DP1 receptor in microglia, we concurrently treated microglia with A‐exo and the DP1 receptor inhibitor asapiprant (abbreviated as asap in subsequent text and figures) to observe whether asap could reverse microglia activation, senescence‐associated secretory phenotype (SASP) secretion, and cellular senescence induced by A‐exo.

We injected A‐exo into the hippocampal region of the mouse brain and performed immunofluorescence staining on the hippocampal region of the mouse brain. Immunofluorescence analysis revealed a significant upregulation of DP1 receptor expression in microglia of the Sham+A‐exo group compared to the Sham+Vehicle group (Figure 3A–D). This indicates that A‐exo robustly activates the DP1 receptor on microglia.

We conducted interventions on BV2 microglia in vitro, grouping them as Con+DMSO, Con+A‐exo, A‐exo+asap, and Con+asap. Flow cytometry analysis revealed that A‐exo led to an increase in the senescence markers P16 and P53 in microglia. Compared to the A‐exo group, the expression of P16 and P53 in microglia was significantly reduced in the A‐exo+asap group. Additionally, asap had no effect on microglia (Figure 3E–H). Further immunofluorescence staining similarly found that A‐exo reduces the expression of CD206 in microglia. In comparison to the A‐exo group, the A‐exo+asap group promoted microglia to express more CD206, a marker of the anti‐inflammatory phenotype (Figure 3I). Similarly, flow cytometry analysis also showed that microglia treated with both A‐exo and asap exhibited increased expression of the anti‐inflammatory cytokine CD206, reduced lipid droplet accumulation, and decreased secretion of SASP factors such as TNF‐α and IL‐6 compared to microglia treated solely with A‐exo (Figure 3J–M). In summary, A‐exo leads to microglial DP1 activation and cellular senescence, and DP1 receptor inhibitors significantly ameliorate A‐exo‐induced microglial senescence and neuroinflammation.

To assess the influence of A‐exo on leukocyte infiltration into the brain. Using flow cytometry, we measured the counts of immune cell subsets in the brains of mice administered with A‐exo (Figure 4A). We found that, compared to normal mice, mice receiving A‐exo exhibited significantly increased numbers of microglia (CD45^int CD11b⁺), neutrophils (CD45^highCD11b⁺Ly6G⁺), and monocytes (CD45^highCD11b⁺Ly6C^high) in their brains (Figure 4A–C). Conversely, the counts of other immune cell subsets, such as CD4⁺ T cells (CD45^highCD3⁺CD4⁺), CD8⁺ T cells (CD45^highCD3⁺CD8⁺), NK cells (CD45^highCD3⁻NK1.1⁺), and B cells (CD45^highCD3⁻CD19⁺), remained unchanged in the brains of A‐exo‐administered mice (Figure 4B). Notably, A‐exo increased the number of microglia expressing CD86, lipid droplets (BODIPY), IL‐1β and TNF‐α (Figure 4E–G), and decreased microglial expression of CD206 (Figure 4D). Brain tissue inflammatory cytokines were measured by enzyme‐linked immunosorbent assay (ELISA), and the results are presented in Figure S1C–E. The ELISA results from brain tissues were consistent with flow cytometry findings, demonstrating that A‐exo increased the expression of TNF‐α, IL‐6, and IL‐1β in the brain, thereby exacerbating neuroinflammation (Figure S1C–E). These results indicate that A‐exo promotes leukocyte infiltration and microglial inflammatory activity in the brains of normal young mice.

To investigate whether A‐exo activates microglia by acting on the DP1 receptor, we simultaneously intervened in mice with a DP1 receptor inhibitor (asap) and A‐exo. The results showed that, compared to the A‐exo+Vehicle group, the A‐exo+asap group significantly reduced the number of immune cells such as CD45^highCD11b⁺Ly6G⁺ and CD45^highCD11b⁺Ly6C⁺ in the brain (Figure 4B). Notably, the DP1 receptor inhibitor (asap) decreased the number of microglia expressing CD86, lipid droplets (BODIPY), IL‐1β and TNF‐α (Figure 4E–G), and increased microglial expression of CD206 (Figure 4D). These findings suggest that A‐exo activates microglia through the DP1 receptor, which in turn regulates the infiltration and activation of immune cells in the brain.

These results indicate that A‐exo induces the infiltration of immune cells such as CD45^highCD11b⁺Ly6G⁺ and CD45^highCD11b⁺Ly6C⁺ into the brain by activating the microglial DP1 receptor. A‐exo also stimulates microglia to express increased levels of CD86, IL‐1β, and TNF‐α through activation of the microglial DP1 receptor. DP1 receptor blockage reverses A‐exo‐mediated hyperactivation of microglia and leukocyte infiltration.

To further investigate whether A‐exo impairs cognitive function and induces cellular senescence in mice by acting on microglia, we first depleted microglia in the brain using the small‐molecule CSF1R inhibitor PLX3397 (Elmore et al. 2014). Microglia are the only immune cells in the central nervous system that express CSF‐1R under physiological conditions, and their development and survival depend on the CSF‐1R signaling pathway (Langmann et al. 2015; Waisman et al. 2015). Administration of a CSF‐1R inhibitor effectively eliminates microglia without adverse effects on other brain tissues. Mice were fed either a regular diet containing PLX3397 or a standard diet for 28 days, which was continued until the end of the experiment, to prevent subsequent microglial proliferation in the brain. The efficacy of PLX3397 in depleting microglia was confirmed by flow cytometry of brain tissue. Compared with mice fed the standard diet, the number of CD45^intCD11b⁺ microglia in the brains of mice fed the PLX3397‐containing diet was significantly reduced (Figure 5B). Immunofluorescence analysis further confirmed robust microglial depletion following 28‐day PLX3397 intervention, as evidenced by quantitative assessment of Iba1⁺ cells (Figure S1F,G).

Y‐maze was performed at 36 days after A‐exo injection into the hippocampus to detect the cognitive function of mice. Compared to the A‐exo+PLX3397 group, the alternation performance of the A‐exo+Vehicle group significantly decreased, indicating impaired memory. In comparison to the A‐exo+Vehicle group, the A‐exo+PLX3397 group exhibited higher alternation performance. The above results showed that microglial depletion alleviates cognitive decline due to A‐exo (Figure 5C).

Between days 37 and 42 after A‐exo injection into the hippocampus, the water maze was performed. Mice in the A‐exo+Vehicle group exhibited impaired spatial learning and memory compared to the A‐exo+PLX3397 group. Treatment with PLX3397 significantly alleviates cognitive decline due to A‐exo, as evidenced by their swimming paths and learning curves (Figure 5D,E). During the training phase, mice in the A‐exo+Vehicle group exhibited significantly longer escape latencies compared to the A‐exo+PLX3397 mice (Figure 5D). In the probe test, the number of crossings over the target quadrant platform significantly decreased in the A‐exo+Vehicle group (Figure 5G), while the latency to first platform arrival increased (Figure 5H), indicating impaired spatial learning and memory function. Conversely, compared to the A‐exo group, mice in the A‐exo+PLX3397 group spent significantly more time in the target quadrant (Figure 5I) and had reduced latency to first platform arrival (Figure 5H), indicating recovery of spatial learning and memory function. Additionally, there were no significant differences in average swimming speed among the four groups of mice (Figure 5F). In conclusion, microglial depletion improves the effects of A‐exo on cognitive impairment.

Subsequently, we examined the effects of A‐exo treatment on cellular senescence in mice fed either the PLX3397‐containing diet or the standard diet. Compared with A‐exo treatment alone, the A‐exo+PLX3397 group has less expression of the DNA damage marker H2A.X (Figure 5J). This suggests that the A‐exo+PLX3397 treatment significantly ameliorated neuronal senescence. Flow cytometry results also indicated that, compared with the A‐exo+Vehicle group, the A‐exo+PLX3397 group exhibited a reduction in cellular lipid droplets, accompanied by a significant decrease in the expression of P16 and P53 (Figure 5K,L).

We next evaluated the impact of microglial depletion on A‐exo‐induced leukocyte infiltration in the brain. Using flow cytometry, we quantified immune cell subsets in the brains of mice subjected to hippocampal A‐exo injection with or without microglial depletion (Figure 6A). Compared to controls fed a standard diet, microglial depletion significantly reduced the counts of microglia (CD11b⁺CD45^int), neutrophils (CD45^highCD11b⁺Ly6G⁺), and monocytes (CD45^highCD11b⁺Ly6C^high) in A‐exo‐injected mice (Figure 6B,C). In contrast, other immune cell subsets—including CD4⁺ T cells (CD45^highCD3⁺CD4⁺), CD8⁺ T cells (CD45^highCD3⁺CD8⁺), NK cells (CD45^highCD3⁻NK1.1⁺), and B cells (CD45^highCD3⁻CD19⁺) remained unchanged following microglial depletion (Figure 6B).

We further investigated the impact of microglial depletion on splenic immune subsets in mice subjected to hippocampal A‐exo injection. Flow cytometry analysis revealed that the counts of splenic immune cells, including neutrophils (CD11b⁺Ly6G⁺), monocytes (CD11b⁺Ly6C^high), NK cells (CD3⁻NK1.1⁺), CD4⁺ T cells (CD3⁺CD4⁺), CD8⁺ T cells (CD3⁺CD8⁺), and B cells (CD3⁻CD19⁺) remained comparable between the microglia‐depleted and control groups (Figure 6D,E). These results indicate that microglial depletion does not significantly alter splenic immune cell composition in A‐exo‐injected mice.

To identify whether the cellular origins of A‐exo and Y‐exo are the same, we analyzed the exosomes using flow cytometry. We found that A‐exo exhibited higher expression of GFAP and CD31 compared with Y‐exo, while there were no significant differences in the expression of L1CAM and CD11b between A‐exo and Y‐exo (Figure 7A,B). These results suggest that A‐exo may primarily originate from astrocytes and endothelial cells in the aged brain.

Compared with Y‐exo, A‐exo exhibited significantly higher expression of PTGDS, which is predominantly expressed in astrocytes. Based on this observation, we further investigated the effects of brain‐derived astrocytic exosomes from aged mice on microglia. We isolated astrocytic exosomes with high GFAP expression (A‐exo‐GFAP) from brain‐derived exosomes of aged mice and used them to intervene in BV2 microglia cultured in vitro. TEM revealed that A‐exo‐GFAP displayed intact exosomal morphology (Figure 7C). Western blot (WB) analysis showed that A‐exo‐GFAP had significantly higher GFAP expression compared to A‐exo (Figure 7D,E). Flow cytometry further demonstrated that A‐exo‐GFAP increased the expression of P16, P53, TNF‐α, IL‐6, and BODIPY in microglia (Figure 7F,G). Collectively, these findings indicate that astrocytic exosomes (A‐exo‐GFAP) from the brains of aged mice exacerbate neuroinflammation and induce microglial senescence.

To investigate whether A‐exo induces microglial senescence in aged mice, we injected A‐exo into the hippocampus of aged mice (18 months old) and examined microglial senescence in aged mice. The results, presented in Figure , show that compared with young mice, microglia in the brains of aged mice exhibited a senescent phenotype, characterized by increased expression of P16, P53, and lipid droplets. A‐exo further enhanced microglial senescence in the brains of aged mice (Figure ). S2 S2A,B

To investigate the impact of the DP1 receptor inhibitor (asap) on cognitive function in aged mice, we administered the DP1 receptor inhibitor (asap) via gavage to 18‐month‐old mice. We assessed the Y‐maze performance of mice at 36 days and evaluated the effect of asap on recognition memory in aged mice by calculating alternation performance. Compared to the Young group, the alternation performance of the Aged+Vehicle group significantly decreased, indicating impaired memory. In comparison to the Aged+Vehicle group, the Aged+asap group exhibited higher alternation performance. This suggests that asap treatment improves cognitive decline in aged mice (Figure 8B).

Between days 37 and 42 after orally administering DP1 receptor inhibitor Asapiprant, a water maze was performed. Mice in the Aged+Vehicle group exhibited impaired spatial learning and memory compared to the Young group. Treatment with asap significantly improved the spatial learning ability of the aged mice, as evidenced by their swimming paths and learning curves (Figure 8C,D). During the training phase, the Aged+ Vehicle mice of the group showed a decrease in escape latency, while mice in the Aged+Vehicle group exhibited significantly longer escape latencies compared to the Young mice. In contrast, mice treated with asap showed shorter escape latencies compared to the Aged+Vehicle counterparts (Figure 8D). In the probe test, the number of crossings over the target quadrant platform significantly decreased in the Aged+Vehicle group (Figure 8F), while the latency to first platform arrival increased (Figure 8G), indicating impaired spatial learning and memory function. Conversely, compared to the vehicle‐treated counterparts, mice in the Aged+asap group spent significantly more time in the target quadrant (Figure 8H) and had reduced latency to first platform arrival (Figure 8G), indicating recovery of spatial learning and memory function. Additionally, there were no significant differences in average swimming speed among the four groups of mice (Figure 8E). In conclusion, the aged mice showed impaired spatial learning and memory function compared to the Young mice, while treatment with asap significantly alleviated cognitive decline in aged mice.

To our amazement, we also found that the DP1 receptor inhibitor significantly improved hair loss in aged mice (Figure 8I). Flow cytometry results further indicated that the DP1 receptor inhibitor markedly reduced the expression of P16 and P53 in microglia within the brains of aged mice (Figure 8K,L). Immunofluorescence staining was conducted, and the results showed that compared to aged mice, asap treatment significantly decreased the expression of the neuronal DNA damage marker (H2A.X) (Figure 8J), improving cellular senescence in aged mice. Additionally, flow cytometry results demonstrated that the DP1 receptor inhibitor significantly decreased lipid droplet accumulation (BODIPY) in microglia and reduced the expression of senescence‐associated secretory phenotype (SASP) cytokines, including IL‐1β, TNF‐α, and IL‐6, in microglia within the brains of aged mice (Figure 8K,L), thereby ameliorating neuroinflammation. These findings suggest that DP1 receptor blockage ameliorates cellular senescence and cognitive decline in aged mice.

In this research, specific‐pathogen‐free (SPF) elderly male C57BL/6J mice (aged 18–20 months, weighing 30–35 g) and young male C57BL/6J mice (6–8 weeks old, weighing 20–22 g) were obtained from Beijing Vitalstar Biotechnology Co. Ltd. All the mice were kept in the Animal Center of Tianjin Medical University under SPF conditions. The environment featured a 12‐h light/dark cycle, a temperature of 22°C ± 2°C, and a humidity of 45%. They had free access to standard rodent feed and sterile water. Experimental animal operations were carried out following the protocols approved by the Animal Care and Use Committee of Tianjin Medical University (Tianjin, China), in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals. All procedures were approved by the Animal Care and Use Committee of Tianjin Medical University General Hospital (Approval No. IRB2022‐DWFL‐074) and strictly conformed to the ARRIVE guidelines.

Brain tissues from either 18‐ to 20‐month‐old or 6‐ to 8‐week‐old wild‐type mice, which were finely minced using a sterile scalpel and placed into a 15 mL conical tube containing 3 mL of Hibernate A solution (A1247501, Thermo Fisher Scientific) with 20 units/mL of papain (Worthington) for enzymatic digestion at 37°C for 20 min (Bodart‐Santos et al. 2023). The enzymatic digestion process was halted by adding 7 mL of ice‐cold Hibernate A solution, which was supplemented with protease and phosphatase inhibitors. The tissue suspension was gently pipetted approximately 30 times using a 10 mL plastic pipette to ensure dissociation and homogeneity. The suspension was then subjected to sequential centrifugation steps at 300 × g for 10 min, 2000 × g for 10 min, and 10,000 × g for 30 min at 4°C to remove cells and debris of varying membrane sizes. The resulting supernatant was filtered through a 0.2 μm syringe filter and centrifuged at 100,000 × g for 70 min at 4°C using a SW41Ti Beckman rotor. The pellet containing exosomes was resuspended in 0.5 mL of PBS (Bodart‐Santos et al. 2023).

The particle size distribution of brain‐derived exosomes was assessed using NanoSight NS300 (Malvern Instruments, Worcestershire, UK) and ZetaView PMX 110 V3.0 (Particle Metrix, Meerbusch, Germany). For both instruments, exosomes were diluted in PBS. NanoSight results were averaged from three 1‐min readings per sample (Lei et al. 2021). ZetaView analysis involved two cycles, examining 11 cell positions with pre‐acquisition settings of sensitivity 85, frame rate 30 fps, and shutter/laser pulse at 55. Post‐acquisition settings were set to a minimum brightness of 20, with particle sizes ranging from 5 to 500 pixels (Bodart‐Santos et al. 2023).

Brain‐derived exosome suspensions were fixed with 4% paraformaldehyde (diluted 1:1 to 2% final concentration) at 4°C overnight and stored at −20°C. After thawing, they were vortexed and centrifuged for 30 s. Exosomes were adsorbed onto Formvar‐coated Ni mesh grids by placing 5‐μL drops on the grids for 20 min in a dry environment. Negative staining was done using 100 μL of 1% aurothioglucose or 2% uranyl acetate, followed by excess removal with lens paper. The grids were then examined under Tecnai Spirit or Talos L120C transmission electron microscopes at 120 kV. Electron micrographs were captured at magnifications ranging from 8500× to 17,500× (Bodart‐Santos et al. 2023).

The cerebral tissue was initially fixed in phosphate‐buffered saline (PBS) containing 4% paraformaldehyde (PFA) for a duration of 12 h. This was followed by immersion in PBS with 15% sucrose at 4°C for 24 h, and subsequently in PBS with 30% sucrose at 4°C for 48 h. Once the tissue had settled, it was embedded in OCT compound and sliced into 10 μm sections using a cryostat. The sections were then exposed to 0.3% Triton X‐100 for 30 min at ambient temperature (Ge et al. 2020; Santos‐Garcia et al. 2023). Following this, the sections were blocked with PBS containing 10% goat serum for 1 h at room temperature and incubated overnight at 4°C with primary antibodies targeting CD206 (1:200, Santa Cruz Biotechnology), H2A.X (1:500, CST), NeuN (1:500, Abcam), and Iba1 (1:500, Abcam). After thorough washing with PBS, the sections were treated with secondary antibodies for 1 h (Alegre‐Zurano et al. 2023). The cell nuclei were counterstained with DAPI (S2110, Solarbio) (Q. Yang et al. 2020). Co‐localization was quantified using ImageJ Coloc2 and Pearson correlation coefficient (r) values. Imaging was performed using a Leica confocal microscope TCS SP8 or Nikon C2, and analysis was conducted using ImageJ software. ImageJ Coloc2 was used for quantifying co‐localization, and Pearson's correlation coefficient (r) values were calculated to assess the correlation between co‐localized voxels in different color channels.

ECHO Biotech conducted proteomic analysis on exosomes isolated from the brain tissues of young and aged rats. Protein concentrations of each exosome were measured using a Qubit Fluorometer (Invitrogen), and each sample (10 μg) was processed via 10% SDS‐PAGE. The gel bands were excised and processed for in‐gel trypsin digestion utilizing a ProGest robotic system (DigiLab). Comparative analysis of proteomic profiles between Y‐exo and A‐exo was conducted using nano‐liquid chromatography coupled with tandem mass spectrometry (nanoLC–MS/MS). This was performed with a Waters Nano Acquity HPLC system connected to a Thermo Fisher quadrupole‐orbitrap high‐resolution mass spectrometer. The acquired LC–MS/MS raw data were analyzed using Mascot (Matrix Science) and queried against the UniProt Rat database. The Mascot Dial‐A‐Truck (DAT) files were subsequently imported into Scaffold (Proteome Software) for validation, filtering, and compilation of a non‐redundant protein list for each sample. The data were refined based on a 1% false discovery rate (FDR) for both proteins and peptides, with a requirement of at least 2 unique peptides per protein. Protein identifications were considered valid if they were identified by a minimum of 2 peptides or a single unique peptide with an FDR below 1%.

Functional annotation of the identified proteins was performed using Blastp (Blast2GO version 4) and cross‐referenced with the Gene Ontology database (accessible at http://geneontology.org↗). To categorize proteins based on their similarity, hierarchical clustering was applied, utilizing the Pheatmap package (available at https://CRAN.R‐project.org/package=pheatmap↗, Estonia). This method organizes data into clusters according to their similarity matrix. For the visualization of significant changes across large datasets, a volcano plot was generated using the Ggplot2 package (https://ggplot2.tidyverse.org↗, New Zealand), which helps in identifying variations in data points among replicates. The analysis used mouse proteins as a reference for determining enrichment values. The functional enrichment of differentially expressed exosomal proteins was statistically assessed using Fisher's exact test within Blast2GO. Differentially expressed proteins were considered significant with an absolute fold change ≥ 1.5 and a corrected p‐value < 0.05. PPI networks were obtained with the STRING database and visualized using Cytoscape 3.9.1 (Y. Zhang et al. 2023). Gene Set Enrichment Analysis (GSEA) (Subramanian et al. 2005) was performed using R package "clusterProfiler" (Yu et al. 2012) and "GSEABase", The DISCO databasethe, ScAgeAnno database, and HUSCH database help analyze the functionality of PTGDS.

Male wild‐type (WT) mice aged 6 to 8 weeks received intracranial infusions of brain‐derived exosomes from aged mice (18–20 months). The mice were anesthetized with 5% isoflurane (Cristalia) for 3 to 5 min and maintained at 2.5% isoflurane using a vaporizer system during the injection process. The exosomes (2.3 μg in 3 μL) were bilaterally injected into the hippocampus at coordinates (−2.0 AP, ±1.3 ML, −2.2 DV) using stereotaxic surgery. The mice were euthanized 42 days post‐infusion.

Asapiprant (asap) is a potent and selective DP1 receptor antagonist. Asapiprant (in 50% DMSO) at a dosage of 5 mg/kg was administered starting 2 h post‐exosome injection into the hippocampus of mice, once every 2 days for a total of 28 days. The vehicle control group received an equivalent volume orally (Sang et al. 2023).

Aged male C57BL/6J mice (19 months old, weighing 30–35 g) were treated with asapiprant at a dose of 5 mg/kg (in 50% DMSO) every 2 days for a total of 28 days. The vehicle control group received an equal amount orally.

Serum and brain PGD2 concentrations were determined using a mouse PGD2 quantitative enzyme‐linked immunoassay kit (MLBIO, YJ911182) (Iwasa et al. 2021). Obtain serum samples and brain tissue samples and follow the instructions of the ELISA kit.

The Morris water maze (MWM) test was utilized to evaluate the memory and spatial learning capabilities of murine subjects (Baeske et al. 2024). All mice underwent a training regimen aimed at locating a submerged platform within an opaque aquatic environment. For the initial 5 days of training, the platform (with a diameter of 10 cm) was submerged 1 cm beneath the water surface. Each mouse participated in four trials daily, each lasting 90 s, with a consistent starting quadrant and trial sequence for individual trials but varied among mice to prevent bias. The primary outcome measure was escape latency, defined as the time taken to discover the submerged platform, serving as an indicator of memory and spatial learning proficiency (capped at 90 s). Mice that did not locate the platform within this period were gently guided to it (Katz et al. 2015). Once the platform was found, the mice were allowed to stay on it for 15 s. On the sixth day, a spatial probe test (SPT) was performed by removing the platform and allowing each mouse to swim freely for 90 s. The Stoelting video tracking system was then used to measure various parameters, such as the number of times the mice crossed the platform's previous location, the time taken to first reach that location, and the swimming speed (Xiao‐Hang et al. 2024).

The Y‐maze apparatus is employed as a tool to evaluate short‐term learning and memory functions in murine models. Constructed from durable blue opaque polyethylene plastic, the Y‐maze comprises three arms designated as A, B, and C. During the procedure, mice are positioned at the central nexus of the three arms and granted unrestricted access to explore the maze for a duration of 8 min (Z. J. Yang et al. 2024). Spontaneous alternation behavior is recorded when an animal sequentially enters three distinct arms (e.g., ABC, CAB, ACB, BCA, CBA, BAC). The percentage of spontaneous alternations (%) is calculated using the formula: (number of spontaneous alternations/[total number of arm entries minus 2]) × 100. This metric serves as a quantitative indicator of short‐term memory and learning efficiency in the mice (Lee et al. 2024).

To achieve in vivo microglial depletion, mice were administered a standard diet enriched with PLX3397 (Catalog No. HY‐16749; MedChemExpress Inc.), a CSF1R (colony‐stimulating factor 1 receptor) inhibitor (X. Li, Li, Jin, et al. 2023). This dietary regimen was initiated at least 28 days before the intrahippocampal administration of exosomes to ensure optimal microglial elimination, which was sustained throughout the duration of the study (Chauquet et al. 2024).

BV2 microglia were sourced from the Chinese Academy of Sciences Cell Resource Center for Basic Research (Beijing, China) and cultured in DMEM/F12 medium (Thermo Fisher Scientific) enriched with 10% ex‐fetal bovine serum (FBS), 100 U/mL penicillin, and 100 mg/mL streptomycin. The cells were incubated in a humidified environment with 5% CO₂ at 37°C (Yin et al. 2020). Asapiprant (asap), a selective and potent DP1 receptor antagonist, was used at a concentration of 10 μM (D. Sang et al. 2023). BV2 cells were pretreated with 10 μM asap for 1 h, and then were intervened with exosomes at a concentration of 80 μg/mL for 24 h before detecting relevant indicators.

Flow cytometry was employed to measure immune cell populations and evaluate cytokine production in the brain, following established protocols (Cheng et al. 2024). Between 42 and 45 days after the injection of exosomes into the mouse hippocampus, the brains were perfused with PBS, harvested, and treated with collagenase IV to create a single‐cell suspension. Myelin was removed using a 30% Percoll solution, and the resulting cell pellet was resuspended in a 1% BSA solution for staining. All antibodies were purchased from Biolegend (San Diego, CA, USA), unless otherwise specified. The antibodies used in this study included: CD45 (30‐F11), CD11b (M1/70), Ly6G (1A8), F4/80 (BM8), Ly6C (HK1.4), CD3 (17A2), CD4 (GK1.5), CD8 (53–6.7), NK1.1 (PK136), CD19 (6D5), tumor necrosis factor alpha (TNF‐α) (MP6‐XT22), IL‐1β (NJTEN3), IL‐6 (48‐7061‐82) (Cheng et al. 2024). P16(sc‐1661) and P53(sc‐126) were sourced from Santa. DP1(56773) was sourced from SAB. BODIPY (GC42959) was sourced from Glpbio (USA) (Guo et al. 2023). Flow cytometry was performed using a FACS Aria flow cytometer, and the data were analyzed with FlowJo 10.6.2 software.

Western blotting was performed as previously described (Li et al. 2025). For protein extraction, brain tissues from each experimental group were lysed using a lysis buffer containing protease and phosphatase inhibitors. Brain tissues were homogenized to extract proteins into the supernatant fraction, followed by protein quantification using the bicinchoninic acid (BCA) assay. Specific primary antibodies were applied to polyvinylidene fluoride (PVDF) membranes and incubated overnight at 4°C. All antibodies were diluted at a 1:1000 ratio, except GAPDH, which was prepared at a 1:10,000 dilution. The next day, membranes underwent three rounds of washing to remove unbound antibodies. Subsequently, membranes were incubated with secondary antibodies (Zhongshan Golden Bridge, Beijing, China) diluted at 1:5000 for 1 h at room temperature. Immunoblots were developed using an enhanced chemiluminescence detection system and visualized with a ChemiDoc Touch Imaging System (Bio‐Rad, Hercules, CA, USA). Densitometric analysis of protein band intensity was performed using ImageJ software (National Institutes of Health, Version 1.53), providing a quantitative assessment of protein expression levels.

To determine the effects of brain‐derived astrocytic exosomes from aged mice on microglia, we collected GFAP‐positive A‐exo (A‐exo‐GFAP) and performed in vitro interventions on microglia. GFAP⁺ exosomes were isolated from total exosomes using a co‐immunoprecipitation assay (L. Li, Li, Bai, et al. 2023). Briefly, GFAP antibody (Cat. No. 60190‐1‐lg, Proteintech) was incubated with Protein G magnetic beads (Cat. No. 10007D; Thermo Fisher Scientific) at room temperature for 1 h. The isolated total exosomes were then incubated with antibody‐coated magnetic beads overnight at 4°C (Ge et al. 2020). Immunoprecipitated GFAP⁺ exosomes were washed with washing buffer and released from the beads using elution buffer.

GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA) and IBM SPSS Statistics version 26.0 (IBM Corp., Armonk, NY, USA) (Schroer et al. 2023) were used for carrying out data analysis. To compare the data from two independent groups, a two‐tailed unpaired Student's t‐test was utilized. For the comparison of data from three or more groups with a single factor, one‐way Analysis of Variance (ANOVA) followed by a Tukey post‐hoc test was applied. Two‐way ANOVA followed by a Tukey post‐hoc test was implemented for comparing two or more factors among multiple groups. Significance was considered when the p‐value was less than 0.05. Data analysis was conducted with GraphPad Prism 8.0. All data were displayed as the mean ± standard deviation (SD) (Liang et al. 2022). In the figure legends, the number of samples (n) and p‐values are all indicated.