RNA from mobile DNA in blood particles may cause brain inflammation and memory problems in aging through the cGAS-STING pathway

To investigate the relationship between plasma EV LINE‐1 mRNA levels and aging, the LINE‐1 mRNA levels in EVs isolated from plasma samples of 185 individuals aged 20–95 years were quantified (Figure 1A). Plasma EVs were characterized using NTA, TEM, and Western blotting. NTA confirmed the typical size distribution of EVs (~50–300 nm) (Figure 1B), while TEM revealed their characteristic morphological features (Figure 1C). Western blot analysis validated the presence of EV‐specific markers (ALIX and CD63) and the absence of the negative control markers (GM130 and Calnexin), confirming the purity of the isolated EVs (Figure 1D).

Quantitative PCR (qPCR) using three distinct LINE‐1‐specific primer sets (targeting ORF1, ORF2, and 5′UTR regions, respectively; Table S1) revealed a significant age‐dependent increase in plasma EV LINE‐1 mRNA levels (Figure 1E,G,H). LINE‐1 mRNA levels were the lowest in young individuals (20–45 years; ORF1: 0.6 ± 0.3; ORF2: 0.6 ± 0.3; 5′UTR: 0.4 ± 0.2) and markedly higher in older individuals (> 65 years; ORF1: 7.6 ± 5.0; ORF2: 7.9 ± 5.5; 5′UTR: 6.7 ± 4.4; p < 0.001 for all primer sets; Table 1). This increase was consistent across all primer sets, indicating a robust upregulation of LINE‐1 mRNA in plasma EVs with age.

Next, we assessed correlations between plasma EV LINE‐1 mRNA levels and brain aging biomarkers, including neurofilament light chain (NFL), amyloid‐beta 42 (Aβ42), amyloid‐beta 40 (Aβ40), and phosphorylated tau 181 (pTau181). Stratification of participants into tertiles based on LINE‐1 mRNA expression revealed significantly elevated levels of NFL, Aβ42, and Aβ40 in individuals in the highest tertile compared to those in the lowest tertile (Figure 1F,I,K). For example, NFL levels were 4.3 ± 3.2 pg/mL in younger individuals versus 49.3 ± 19.4 pg/mL in older individuals (p < 0.001; Table 1). Additionally, Spearman correlation analysis and multivariable regression analysis demonstrated strong positive associations between LINE‐1 mRNA levels and aging biomarkers, particularly NFL (Figure 1J; Table 2). These findings identified plasma EV LINE‐1 mRNA as a non‐invasive marker of aging and brain degeneration.

To investigate the characteristics and tissue origins of plasma EVs in aged mice, organs were collected from 21‐month‐old mice and 3‐month‐old mice (Figure 2A). To determine tissue‐specific contributions to circulating EVs, expression levels of EV‐specific markers (ALIX and CD63) were analyzed in lysates from the brain, heart, kidney, liver, lung, and spleen of aged and young mice (n = 3). Western blot results revealed significantly increased ALIX and CD63 levels in the brain and heart of aged mice, compared to controls (p < 0.05), whereas no significant differences were observed in the kidney or liver (Figure 2B). These findings suggested that age‐dependent upregulation of EV secretion occurred in specific tissues, particularly the brain and heart, potentially contributing to elevated circulating EV levels in aged mice.

Furthermore, EVs were isolated from plasma of 21‐month‐old WT mice and compared to those of 3‐month‐old controls (n = 6). NTA confirmed that the size distribution of aged mouse plasma EVs (~50–300 nm) was consistent with expectation, with a prominent peak at 149 nm (Figure 2C). TEM further revealed the typical morphology of EVs (Figure 2D), and Western blot analysis validated the presence of EV‐specific markers (ALIX and CD63) and the absence of cellular contaminants (GM130 and Calnexin) in aged mouse EVs (Figure 2E). Meanwhile, we also used qPCR to measure the levels of LINE‐1 RNA in plasma EVs from aged and young mice (n = 3). The results showed that LINE‐1 RNA expression in plasma EVs from aged mice was significantly elevated compared to young mice, which was consistent with the findings revealed in our human studies (Figure 2F).

Then, to determine tissue‐specific contributions to LINE‐1 RNA in circulating EVs, expression levels of LINE‐1 RNA were analyzed in EVs isolated from the brain, heart, kidney, liver, lung, and spleen of 21‐month‐old and 3‐month‐old mice (n = 6; Table S2). qPCR results revealed significantly increased LINE‐1 RNA levels in the brain and lung of aged mice, compared to controls (p < 0.05), whereas no significant differences were observed in the heart, kidney, liver, or spleen (Figure 2G). These findings suggested that age‐dependent upregulation of LINE‐1 activities occurred in specific tissues, particularly the brain and lung, potentially contributing to elevated LINE‐1 RNA levels in circulating EVs in aged mice.

To examine the functional effects of aged EVs on cognition, EVs from 21‐month‐old mice and 3TC treated 21‐month‐old mice were injected into 10‐month‐old recipient mice via tail vein (Figure 3A). Confocal microscopy of PKH67‐stained EVs confirmed successful entry of old EVs into the brain (Figure 3B), while in vivo imaging of Dir‐labeled EVs revealed their accumulation in the brain and major organs, including the liver, kidney, and heart, over a 4‐week period (Figure 3C,D).

Behavioral tests revealed significant cognitive and exploratory impairments in old EV‐treated mice. In the Y‐maze test, old EV‐treated mice displayed reduced spontaneous alternation percentages, indicative of impaired working memory (Figure 3J,L; p < 0.0001). Similarly, new object recognition (NOR) tests revealed reduced exploration ratios in old EV‐treated mice, demonstrating recognition memory deficits (Figure 3F; p < 0.0001). In the open field test, old EV‐treated mice exhibited diminished central zone exploration, reduced total distance traveled, and decreased entry frequency into the central zone (Figure 3G–I,K; p < 0.0001). Furthermore, prolonged immobility times were observed in the tail suspension test (Figure 3E; p < 0.0001), indicative of depressive‐like behavior. Notably, the behavioral performance of the recipient mice injected with EVs extracted from 21‐month‐old mice that were treated with the LINE‐1 inhibitor 3TC were close to near‐control levels across all tests (vs. old EV group; p < 0.01). These results demonstrated that old EVs impaired cognition through mechanisms linked to LINE‐1 activity.

Histological analysis of hippocampal sections revealed significant structural damage, reduced neuronal density, and disrupted cellular architecture in old EV‐treated mice, as observed by H&E staining (Figure 4A). Nissl staining further confirmed a reduction in the number of surviving neurons in old EV‐treated mice, which was significantly more in old EV (without LINE‐1)‐treated mice (p < 0.05) (Figure 4B,D). As shown by SA‐β‐gal staining, old EV‐treated mice exhibited a significant increase in SA‐β‐gal activity in the hippocampus sections (p < 0.001) (Figure 4C,E). Microglial activation, assessed by IBA1 immunofluorescence, was significantly increased in old EV‐treated mice (Figure 4H,K), showing morphological changes, for example, enlarged soma and decreased number of branches (Figure S1A,B). Compared to old EV‐treated mice, old EV (without LINE‐1)‐treated mice showed restored neuronal density and reduced microglial activation (p < 0.05).

LINE‐1 ORF1P and P21 expression was enriched in IBA1‐positive microglial cells in old EV‐treated mice, linking LINE‐1 RNA to microglial activation and senescence (Figure 4L,M). Old EV (without LINE‐1)–treated mice showed decreased expression of ORF1P and P21 and reduced microglial activation and senescence (Figure 4F,G,I,J). These findings suggested that old EVs promoted brain aging through LINE‐1 RNA‐mediated microglial activation, which could be mitigated by LINE‐1 inhibition.

Immunofluorescence and Western blot analysis revealed significant activation of microglia and the cGAS/STING pathway in old EV‐treated HMC3 (Figure 5B,C), as indicated by elevated expression of cGAS, STING, phosphorylated TBK1 (pTBK1), and phosphorylated IRF3 (pIRF3) (Figure 5I,P,Q–X; p < 0.001). Co‐localization studies confirmed the enrichment of ORF1P, cGAS, STING, and P21 in microglia, highlighting the linkage between cGAS/STING activation and cellular senescence (Figure 5A,D–H). Both 3TC and STING inhibitor H151 suppressed cGAS/STING signaling and reduced downstream inflammation. These findings demonstrated that LINE‐1 RNA in plasma EVs promoted neuroinflammation through activation of microglia and the cGAS/STING pathway, which could be targeted pharmacologically to mitigate aging‐associated brain pathology.

To test whether inhibiting LINE‐1 activity or the cGAS/STING pathway could rescue cognitive impairments induced by old EVs, mice injected with old EVs were treated with either the LINE‐1 reverse transcriptase inhibitor 3TC or the STING inhibitor H151. The experimental timeline was illustrated in Figure 6A. Behavioral assessments showed significant improvements in cognitive and exploratory behaviors in both treatment groups.

In the Y‐maze test, old EV‐treated mice exhibited a significant reduction in spontaneous alternation percentages (p < 0.001), indicating impaired working memory (Figure 6B,D). Both 3TC and H151 treatments restored alternation percentages to levels comparable to that of the old EV group (p < 0.01). Similarly, in the open field test, old EV‐treated mice showed reduced central zone (CZ) residence time, total distance traveled, and entry frequency, which were consistent with diminished exploratory activity (Figure 6C,E–G; p < 0.001). Both inhibitors significantly rescued these impairments (p < 0.05), bringing behavioral metrics close to control levels.

In the NOR test, old EV‐treated mice demonstrated reduced exploration of the novel object, as observed by a lower new‐to‐old object exploration ratio (p < 0.0001; Figure 6H). Treatment with 3TC or H151 restored recognition memory, with exploration ratios significantly increased compared to untreated old EV groups (p < 0.01). Furthermore, in TST, old EV‐treated mice displayed prolonged immobility times, indicative of depressive‐like behavior (Figure 6I; p < 0.001). Both inhibitors significantly reduced immobility times, restoring affective behavior to near‐control levels (p < 0.01).

These results demonstrated that inhibition of LINE‐1 activity with 3TC or suppression of the cGAS/STING pathway with H151 effectively mitigated cognitive, exploratory, and affective deficits induced by old EVs. These findings suggested the therapeutic potential of targeting the LINE‐1/cGAS/STING axis to preserve brain function during aging.

To further explore the mechanistic basis of the pathological effects of old EVs on the brain, hippocampal sections from mice treated with old EVs, with or without 3TC or H151, were analyzed. Immunofluorescence analysis revealed elevated LINE‐1 ORF1P and IBA1 expression in old EV‐treated mice, consistent with increased LINE‐1 activity and microglial activation (Figure 7A,F,G, S2B; p < 0.05). Both 3TC and H151 treatments significantly reduced these markers (p < 0.05), indicating successful suppression of LINE‐1 activity and microglial activation.

Histological analysis of hippocampal sections using H&E staining showed disrupted neuronal architecture, reduced neuronal density, and structural abnormalities in the old EV group (Figure 7B). Nissl staining confirmed a significant reduction in healthy neurons in old EV‐treated mice compared to controls (p < 0.001; Figure 7C). SA‐β‐gal staining also showed that old EV‐treated mice exhibited a significant increase in SA‐β‐gal activity compared to controls (p < 0.001) (Figure 7E and Figure S2A). Both 3TC and H151 treatments ameliorated cellular senescence, restored neuronal density, and preserved hippocampal structure.

Co‐localization studies of STING and IBA1 revealed increased STING expression in activated microglia of old EV‐treated mice (Figure 7D,G and Figure S2C; p < 0.01). Western blot analysis confirmed the activation of the cGAS/STING pathway, with elevated expression of cGAS, STING, phosphorylated STING (pSTING), TBK1, phosphorylated TBK1 (pTBK1), IRF3, phosphorylated IRF3 (pIRF3), and the pro‐inflammatory cytokine TNF‐α in the old EV group (p < 0.05; Figure 7H,K–T). Both 3TC and H151 significantly suppressed the activation of these signaling molecules, reducing neuroinflammation (p < 0.01). Microglial morphology analysis revealed increased soma size and reduced branching in old EV‐treated mice, indicative of a pro‐inflammatory state (p < 0.01; Figure 7I,J). Both treatments restored microglial morphology to quiescent states, further confirming their anti‐inflammatory effects (p < 0.05).

Building upon these findings, we further investigated whether similar alterations occurred in neuronal populations across the brain. Immunofluorescence analysis and co‐localization studies of LINE‐1 ORF1P, STING, P21, and NeuN revealed elevated expression of LINE‐1 ORF1P, STING, and P21 in activated neurons in the cortical sections of the old EV‐treated mice (Figures S3 and S4A,C,D,F). Both 3TC and H151 treatments significantly reduced these markers (p < 0.05), indicating successful suppression of LINE‐1 activity, cGAS/STING pathway activation, and neuron senescence, which is consistent with the microglial expression patterns observed in immunofluorescence and Western blot analyses. Nissl staining also confirmed a significant reduction in healthy neurons in the cortical sections of old EV‐treated mice (Figure S4B,E).

These findings established that LINE‐1 RNA drove pathological changes in the brain via the cGAS/STING pathway, promoting microglial activation and neuroinflammation. Inhibition of either LINE‐1 activity or cGAS/STING signaling ameliorated these pathological changes, highlighting the potential of these pathways as therapeutic targets.

The Ethics Committee for Biomedical Research Involving Human Beings of Provincial Hospital affiliated to Shandong First Medical University approved the human studies. Informed consent was obtained from all participants. A total of 185 individuals were chosen for this study, excluding those who were diagnosed with cancer, HIV, any kind of brain disease, or tested positive for the hepatitis C or B surface antigen antibodies, and divided into three age groups: young (20–45 years old), middle‐aged (46–65 years old), and old (66–95 years old). Smoking and alcohol use were identified as either current users or non‐users. The demographics of the participants were provided in Table 1. Blood sample from each participant was collected, kept in EDTA anticoagulant‐coated tubes, and processed within 4 h of hospitalization. Plasma was obtained by centrifuging blood samples at 3000 rpm for 15 min, and then divided into small portions (500 μL) and stored at −80°C for later use. For the measurement of brain aging markers, the frozen plasma was melted at 4°C and centrifuged at 10,000 × g for 10 min to remove cell debris. Then, the plasma sample was divided into small portions (75 μL) to measure the levels of brain aging biomarkers using Single‐Molecule Analyzer (AXL‐2000, Lychix, Suzhou, China).

The 10‐ and 21‐month‐old wild‐type (WT) C57BL/6J mice were purchased from Jinan Xingkang Bio‐tech (Jinan, China) and housed in a temperature‐controlled animal facility with a 12‐h/12‐h cycle of light/darkness and free access to sterile water and food. One week was given for the mice to adapt to the environment before any treatments. For donor mice, the 21‐month‐old mice were selected into two groups: one group received oral 3TC at a dose of 100 mg/kg per day 6 day a week for 4 weeks, and the other received an equivalent volume of PBS for the same duration. Then the mice were euthanized by inhalation of isoflurane. Blood was collected in EDTA anticoagulant‐coated tubes via cardiac puncture. Then, brain, kidney, liver, lung, spleen, and heart tissues of mice were collected. Plasma was obtained by centrifuging blood samples at 3000 rpm for 15 min and then stored at −80°C prior to use.

Mice or human plasma EVs were extracted from 500 μL of plasma using the Total Exosome Isolation Kit (UR52151, Umibio Biotechnology, Shanghai, China) according to the manufacturer's instructions. The EVs from various organ tissues of 21‐month‐old and 3‐month‐old mice were extracted from 0.1 g of the corresponding organ tissues using the Exosome Isolation and Purification Kit (UR52160, Umibio Biotechnology, Shanghai, China) according to the manufacturer's instructions. For the in vivo tracking of EV distribution, EVs were stained with Dir (UR21017, Umibio Biotechnology, Shanghai, China) and PKH67 (UR52303, Umibio Biotechnology, Shanghai, China), respectively. For Dir staining, EVs were incubated with Dir dye in the dark at 37°C for 30 min. For PKH67 staining, PKH67 dye was added into EVs, and the mixture was shaken for 1 min and then incubated in the dark at 37°C for 10 min. Both of the reactions were terminated by the addition of an equal volume of 1% BSA dissolved in PBS. The samples were then processed using the EV isolation method to remove excess dye, and 200 μg of Dir‐stained EVs and PKH67‐stained EVs were injected into the tail vein of recipient mice, respectively, with the same volume PBS injected into the tail veins of control mice. The mice that were injected with PKH67‐stained EVs were euthanized by inhalation of isoflurane 4 h after the injection, and the brain was collected. The Dir‐stained EVs were injected to the mice every other day for 4 weeks, and then the mice were eventually euthanized by inhalation of isoflurane. Brain, kidney, liver, and heart tissues were collected and imaged using an IVIS spectrum in vivo imaging system (IVIS Lumina II, PerkinElmer).

Freshly isolated human or mice plasma EVs were allowed to adsorb onto formvar/carbon coated grids, rinsed briefly in water, and negatively stained in 2% aqueous uranyl acetate. Then, the samples were observed with a transmission electron microscope (TEM, Hitachi H7700, Japan) at 100 kV.

Filtered PBS was used to dilute isolated EV samples from mouse plasma by a factor of 35. The mouse plasma EV size distribution and concentration were examined with nanoparticle tracking analysis (NTA) on a NanoSight NS300 (Malvern Instruments Ltd., Malvern, UK). Samples were recorded at camera level of 11 and detect threshold of 5 using NTA 3.4 Build 3.4.4 software. The human plasma EVs were observed on Particle Metrix (PMX, German) using ZetaView version 8.05.14 SP7.

The 3TC (HY‐BO250, USA) and H151 (HY‐112693, USA) were purchased from MCE (USA). On the day of each experiment, the drugs were freshly prepared. Old EVs extracted from the 21‐month‐old mice were quantified using the BCA Protein Analysis Kit (PC0020, Solarbio, Beijing, China). The recipient mice (44‐week‐old) were first randomly divided into three groups (n = 5 mice per group): a test group injected with 200 μg protein equivalent of plasma EVs extracted from 21‐month‐old mice into the tail vein every other day for 4 weeks, a drug treatment group injected with 200 μg protein equivalent of plasma EVs extracted from 21‐month‐old mice treated with 3TC into the tail vein every other day for 4 weeks, and a control group injected with an equal amount of PBS for the same duration.

After the findings that Old EVs could induce cognitive impairments, the mechanisms underlying this induction were further explored. The mice were randomly divided into five groups (n = 5 mice per group), four of which were injected with 200 μg protein equivalent plasma EVs extracted from 21‐month‐old mice into the tail vein every other day for 4 weeks, and one group was injected with an equal amount of PBS in the tail vein for the same duration as a control. Among the four groups injected with 21‐month‐old mice EVs, one group remained untreated, one group received oral 3TC at a dose of 100 mg/kg per day 6 day a week for 4 weeks, one group was injected intraperitoneally with 750 nmol H‐151 in 200 μL PBS of 5% Tween‐80 and 5% DMSO 6 day a week for 4 weeks, and one group was injected intraperitoneally with vehicle as control. All experimental procedures were authorized by Shandong Provincial Hospital's Laboratory Animal Management and Ethics Review Committee (NSFC: No. 2022–511) and complied with the National Institutes of Health's (NIH, USA) Guide for the Care and Use of Laboratory Animals.

Through the detection of spontaneous alternating behavior in mice, the Y‐maze is a classic experiment used to test short‐term memory ability. In the maze facility, each mouse was positioned at the end of an arm of the maze. The order in which the mice entered and exited each arm over a period of 5 min was recorded and analyzed. Three consecutive entries into three different arms were considered correct choices. The number of correct choices and the total number of arms entered were recorded for each mouse, and the percentage of alternation was calculated using the following formula: alternation percentage (%) = [(number of correct choices)/(total number of arms entered − 2)] × 100%. The maze facility was wiped with alcohol after each mouse finished its test.

The anxiety behaviors of mice were evaluated using the open field test, the classic experimental method for evaluating the anxiety behaviors of mice. The facility was a cubic cage with 50 cm high walls, enclosing a 56 cm × 56 cm wooden floor. Each mouse was placed in the corner of the open field at the beginning of each experiment and given 5 min to roam around. The time spent in the central zone (26 × 26 cm), the total traveled distance, and the number of entries in the central region were all recorded using video tracking software. The facility was cleaned with alcohol after each mouse finished its test.

Recollection memory of mice was evaluated using the new object recognition test. The open field test was carried out the day before the new object recognition test to serve as the adaptation phase of the new object recognition test (Day 1). On Day 2, mice were given 10 min to explore the same field with two identical objects. Then, the mice were put back into their cages for 2 h. Then, mice were placed in their designated locations for the last time, which now included one familiar object and one new object. They were given 5 min to explore at their leisure. The analyses included mice that investigated the study object for at least 10 s. The amount of time spent with each object was calculated by analyzing the full 5‐min video. The score was calculated using the following formula: new object recognition index (NOI) = (TN − TF)/(TN + TF), where TN was the amount of time spent on the new object and TF was the amount of time spent on investigating the familiar object.

The depression level in mice was evaluated using the traditional tail suspension test (TST). The mice were placed in a 40 cm high white open box and suspended for 6 min by adhesive tape about 1 cm from the tip of their tails, and their immobility time was monitored and captured by a video camera. The mice were deemed immobile only if they hung passively or remained completely still.

Mice were all euthanized by inhalation of isoflurane after behavior tests and their brains were collected. Hematoxylin and eosin (H&E) staining was used to analyze the brain morphology. A gradient of alcohol with varying concentrations was used to dewax and dehydrate the paraffin sections of hippocampus tissues. The H&E staining was carried out using an H&E staining kit (GP1031, Servicebio, Wuhan, China) according to the manufacturer's instructions. The sections were submerged in ethanol and xylene, mounted with resin, and observed and photographed under a light microscope (Olympus).

Sections of the hippocampus tissues were examined for neuronal damage using Nissl staining. After 30 min of Nissl stain incubation (GP2087, Servicebio, Wuhan, China), the sections were cleaned with 95% ethanol. The Olympus light microscope was used to count and image the labeled cells. On each slide of the hippocampus tissue, cells were counted in three randomly chosen non‐overlapping fields.

A commercial β‐galactosidase staining detection kit (G1580, Solarbio, Beijing, China) was used to perform SA‐β‐gal staining. The provided fixative solution was used to fix the sections of the hippocampus and cortical tissues for 15 min at room temperature. Following fixation, the sections were submerged in a brand‐new β‐gal staining solution in buffer and left to incubate for the entire night at 37°C. The Olympus light microscope was used to take microscopic pictures of the stained sections.

Human microglia cell line HMC3 was purchased from Pricella (CL‐0620, Wuhan, China) and cultured in MEM containing 10% fetal calf serum and 1% penicillin/streptomycin. Cells were seeded in 6‐well plates 1 day prior to experimental treatments. HMC3 cells were divided into 5 groups: the Old EV group treated with 200 μg protein equivalent EVs extracted from plasma of individuals older than 80 years for 24 h, the Old EV + DMSO group treated with old EVs and the same volume of DMSO as the prepared drugs, the Old EV + 3TC group treated with both old EVs and 40 μmol/L 3TC, the Old EV + H151 group treated with old EVs and 1 μmol/L H151, and the control group cultured with an equal amount of PBS as the old EVs.

Cells or frozen mouse brain slices were first cleaned with PBS and treated with 4% paraformaldehyde prior to immunofluorescence examination. After being rendered transparent with PBS containing 0.1% Triton X‐100, the samples were blocked for 1 h at room temperature using a 5% goat serum/PBS solution. Slides were coated with primary antibodies diluted in PBS and incubated overnight at 4°C. Then, the slides were exposed to DAPI for 7 min at room temperature and then secondary antibodies (Alexa Fluor555, Abcam, ab150078, and Alexa Fluor488, Abcam, ab150113) for 50 min. Lastly, a fluorescent inverted microscope (Leica Microsystems, Biberach, Germany) was used to monitor and image the sealed samples. The antibodies used in this study included anti‐IBA1 (OB‐MMS039, Oasis, Hangzhou, China), anti‐LINE1 ORF1P (ab230966, Abcam, USA), anti‐cGAS (26416‐1‐AP, Proteintech, Wuhan, China), anti‐STING (19851‐1‐AP, Proteintech, Wuhan, China), anti‐P21 (GT1032, Invitrogen, USA), and anti‐beta actin (20536‐1‐AP, Proteintech, Wuhan, China).

Trizol (AG21101, Accurate Biotechnology, China) was used to extract the total RNA of brain tissues of each mouse and EVs derived from various organ tissues and cells, which was then reverse‐transcribed into cDNA. The real‐time PCR experiment was performed using SYBR Green Pro Taq HS premixed qPCR Kit (AG11701, Accurate Biotechnology, China) with the primers for each gene tested. GAPDH was used as an internal reference (Tables S2 and S3). For the clinical study, the RNA was extracted from plasma EVs using Total Exosome RNA and Protein Isolation Kit (ThermoFisher, cat4478545, Rockford, USA) and reverse‐transcribed into cDNA using Taqman Reverse Transcription Kit (Applied Biosystems, Carlsbad, USA). The real‐time PCR experiment was performed using specific primers (Table S1), TaqMan probe, TaqMan Fast Advanced Master Mix, and sterilized water, and hSATA was used as the endogenous gene. At least three independent technical replicates were performed for each sample to ensure reliability and reproducibility, and the 2^−ΔΔCt method was used to calculate the gene expression levels.

Cells or tissues were lysed with RIPA Lysis Buffer with the addition of protease phosphatase inhibitor mixture (P0013B, Beyotime Biotechnology, China). Total protein was collected, measured, and standardized utilizing the BCA Protein Analysis Kit (PC0020, Solarbio, Beijing, China). Electrophoresis was performed on SDS‐PAGE gel, which was transferred to the PVDF membrane (ISEQ00010, Merck Millipore, China). The membranes were then blocked with 5% skimmed milk or 5% BSA diluted in TBST for 1 h at room temperature. After washing in TBST, the membranes were incubated with primary antibodies overnight at 4°C. Then, the membranes were washed in TBST and incubated with secondary antibodies for 1 h at room temperature. Finally, the protein bands were visualized using Tanon's imaging equipment (Shanghai, China) and the Immobilon Western HRP substrate Luminol reagent (Millipore, USA). ImageJ software (NIH, USA) was used to calculate the optical density of immunoblots. The primary antibodies used in this experiment included those for immunofluorescence, as well as anti‐ALIX (ab186429, Abcam, USA), anti‐GM130 (ab52649, Abcam, USA), anti‐CD63 (A5271, ABclonal, China), anti‐Calnexin (66903‐1‐lg, Proteintech, Wuhan, China), anti‐NeuN (66836‐1‐lg, Proteintech, Wuhan, China), anti‐TBK1 (28397‐1‐AP, Proteintech, Wuhan, China), anti‐pTBK1 (5483T, CST, USA), anti‐IRF3 (A2172, ABclonal, USA), anti‐pIRF3 (AP1412, ABclonal, USA), anti‐pSTING (50907T, CST, USA), anti‐TNFα (60291‐1‐lg, Proteintech, Wuhan, China), and anti‐GFAP (Abcam, ab33922, USA).

The statistical packages R (The R Foundation; http://www.r.project.org↗; version 3.4.3) and Empower (R) (www.empowerstates.com↗, X&Y solutions Inc., Boston, Massachusetts, USA) were used for the clinical study analysis. Continuous variables were reported as median, percent 25–percent 75, or mean ± standard deviation (SD). Group differences were examined using the Kruskal–Wallis H test and one‐way ANOVA. The correlation between the levels of EV LINE‐1 RNA and plasma brain aging biomarkers was examined using Spearman correlation and multivariable linear regression analysis. For the animal and cellular experiments, data were analyzed using GraphPad and Prism version 9.0 (RRID: SCR_002798). Comparisons between groups were made using unpaired t‐tests and one‐way or two‐way ANOVA. The significant difference was determined based on p < 0.05.