Physical activity-associated extracellular vesicles inhibit inflammagen-induced microglial activation

Abstract

Excessive microglial activation and neuroinflammation are pivotal contributors to the progression of neurodegenerative diseases. Physical activity has emerged as a complementary therapy recognized for its anti-inflammatory effects. Long-term treadmill-induced mandatory physical activity (TMPA), a rodent paradigm modeling exercise with controllable intensity, has been shown to reduce age-related microglial activation, dopaminergic neuron degeneration in the substantia nigra, and motor impairments in mice. However, the mechanisms underlying these effects remain unclear. Given the role of physical activity-induced exerkines in modulating anti-inflammatory responses, this study investigated the effects of plasma-derived extracellular vesicles (EVs) from TMPA-treated animals on inflammagen-induced microglial activation in both in vitro and in vivo models. We showed that plasma from TMPA-treated rats suppressed lipopolysaccharide (LPS)-induced inflammatory signaling pathways in BV2 microglia cell line. Within 2 h, over 60% of BV2 microglia internalized EVs, and EVs from rats significantly inhibited LPS-induced microglial activation. In vivo, DiI-labeled EVs from TMPA-treated rats, delivered via tail vein injection, were detected in the brain parenchyma of recipient mice within one day. LPS-treated mice receiving EVs from TMPA-treated rats exhibited significantly lower microglial activation in the substantia nigra compared to those receiving EVs from sedentary rats. These findings demonstrate that physical activity-derived EVs effectively suppress inflammagen-induced microglial activation, providing new insights into their anti-neuroinflammatory properties and potential as therapeutic agents for neuroinflammation-related diseases.

Highlights

• •Mandatory physical activity does not alter basal quantity of EVs in plasma.
• •Exogenous EVs can be internalized by microglia in vitro and in vivo.
• •Physical activity-derived EVs inhibit LPS-induced NF-κB activation in microglia.
• •Physical activity-derived EVs suppress LPS-induced microglial activation in vivo.

1. Introduction

Neuroinflammation plays a pivotal role in the development and progression of neurodegenerative diseases, such as Alzheimer's and Parkinson's diseases (Wang et al., 2022a, Wang et al., 2022b; Kiani Shabestari et al., 2022; Heppner et al., 2015). A central player in this process is microglia, the brain's resident macrophage-like innate immune cells, which tightly regulate neuroinflammatory responses. In physiological conditions, microglia adopt a “surveillance” state characterized by a ramified morphology. In this state, microglial processes continuously scan distinct regions of the brain parenchyma, actively maintaining the homeostasis of the central nervous system (Nimmerjahn et al., 2005; Hanisch and Kettenmann, 2007). When exposed to stimuli, microglia undergo a dynamic process of activation that involves changes in morphology, gene expression, and functional behavior (Hanisch and Kettenmann, 2007). Microglia activation varies in extent depending on the nature, intensity, and duration of the stimulus. While microglial activation can have protective effects, such as clearing debris and promoting tissue repair, it can also lead to harmful outcomes (Hanisch and Kettenmann, 2007). Overactive or dysregulated microglia release pro-inflammatory cytokines and neurotoxins, contributing to neuronal injury and death (Hanisch and Kettenmann, 2007). Thus, the modulation of microglial activation is a critical therapeutic target for controlling neuroinflammation and mitigating its impact on neurodegenerative diseases.

Regular physical activity is widely recognized for its extensive health benefits, which extend beyond enhanced physical fitness to include improvements in brain function (Cotman and Engesser-Cesar, 2002; Lin and Kuo, 2013). Notably, physical activity has gained attention as a complementary therapeutic approach for neurodegenerative diseases, primarily due to its anti-inflammatory properties (Gleeson et al., 2011; Scheffer and Latini, 2020). For instance, we previously demonstrated that long-term treadmill-induced mandatory physical activity (TMPA), a rodent paradigm modeling exercise with controllable intensity, reduces age-related dopaminergic neuron loss in the substantia nigra and alleviates motor deficits associated with Parkinson's disease by inhibiting local microglial activation (Wang et al., 2022a, Wang et al., 2022b). The anti-inflammatory effects of physical activity are thought to involve several mechanisms, including reduced circulating levels of pro-inflammatory factors, decreased body fat, shifts in immune cell populations, and increased circulating levels of anti-inflammatory exerkines, the signaling moieties released in response to acute and/or chronic physical activity (Lancaster and Febbraio, 2014; Pedersen, 2006; Kawanishi et al., 2010, 2013; Yu et al., 2017). However, the exact mechanisms through which physical activity exerts its widespread influence on diverse cell types and organs, ultimately modulating microglial activity in the central nervous system through these peripheral pathways, remain incompletely understood.

Recent studies suggest that extracellular vesicles (EVs), released in response to physical activity and carrying exerkines, may play an important role in mediating interorgan communication and enhancing the physiological benefits of physical activity (Whitham et al., 2018a; Trovato et al., 2019; Safdar et al., 2016a, 2016b; Nederveen et al., 2020; Siqueira et al., 2021). EVs are heterogeneous cell-derived particles that are present in various biological fluids such as blood, urine, saliva, and cerebrospinal fluid. They play critical roles in both physiological and pathological processes (Valadi et al., 2007). One of the key functions of EVs is facilitating intercellular communication by enabling the exchange of proteins, lipids, and nucleic acids between cells (Kumar et al., 2024). Furthermore, it has been demonstrated that EVs can regulate microglial activation (Duan et al., 2020; Zagrean et al., 2018; Li et al., 2019; Zheng et al., 2019). These findings suggest that physical activity-associated EVs may play a role in the systemic effects of physical activity, including its potential to modulate neuroinflammatory processes in the central nervous system.

To address this issue, we aimed to evaluate the effects of EVs derived from TMPA-treated animals on lipopolysaccharide (LPS)-induced microglial activation both in vitro and in vivo. Plasma samples were collected from rats subjected to a 4-week TMPA regimen and from sedentary (SED) controls. Initially, the effects of these plasma on LPS-induced inflammation were examined in BV2 microglial cells. Subsequently, EVs were isolated and their effects on LPS-induced microglial activation were examined both in vitro and in vivo.

2. Methods

2.1. BV2 microglial cell cultures and treatments

Immortalized murine microglial BV2 cells (Cat#: ABC-TC212S, AcceGen Biotechnology, Fairfield, NJ, USA; RRID: CVCL_0182) were cultured in Dulbecco's modified Eagle's medium/nutrient mixture F-12 (DMEM/F12, Cat#: 12500062, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS) with extremely low endotoxin levels (<0.05 EU/ml; Lot#: VP2002200, Cat#: TMS‐013‐BKR, Merck‐Millipore, Burlington, MA, USA) and penicillin-streptomycin (Cat#: 15140122, Thermo Fisher Scientific). The cultures were maintained in a humidified incubator at 37 °C with 5% CO₂ and 95% air. Subculturing was conducted every two days when cell confluency reached 80%. BV2 cells used in the experiments were between passage numbers four and nine. Cells were seeded onto culture plates at a density of 3 × 10⁴ cells/cm² and underwent treatments 16 h after seeding.

To determine the optimal dose of LPS (from Escherichia coli O55:B5, Cat#: L2880, Sigma-Aldrich, St. Louis, MO, USA; stock concentration: 1 mg/mL, dissolved in the culture medium), cells were exposed to LPS concentrations of 1, 2, 5, 10, 25, 50, 100, 200, or 500 ng/mL at 16 h post-seeding. Protein lysates were prepared from all cultures 1 h after LPS treatment.

To examine the effects of plasma from TMPA rats on LPS-induced inflammatory responses, the 10% FBS in the culture medium was replaced with 10% SED or TMPA plasma, 15 h after seeding. One hour later, these cultures were extra treated with 100 ng/mL LPS, and cell lysates were collected 1 h after treatment.

To assess EV uptake in BV2 cells, the cells were treated with fluorescently labeled EVs at a concentration of 10 μg protein content/mL, 15 h after seeding. The cultures were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 8 mM Na₂HPO₄, and 2 mM KH₂PO₄) at 1, 2, 4, and 8 h post-treatment, prior to immunocytofluorescent stain. Additionally, to examine the effects of plasma EVs from TMPA rats on LPS-induced inflammatory responses, BV2 cells were exposed to 80 μg protein content/mL of EVs isolated from the plasma of SED or TMPA rats, 15 h after seeding. One hour later, these cells were extra treated with 100 ng/mL LPS, and cell lysates were collected 1 h post-treatment. Sample sizes for each assay are specified in the corresponding figure legends.

2.2. Cell counting Kit-8 assay

The Cell Counting Kit-8 (CCK-8) assay was used to assess cell viability in experiments evaluating the effects of plasma and EVs derived from SED or TMPA rats on LPS-induced inflammatory responses in BV2 cells. BV2 cells were seeded in 96-well plates at a density of 3 × 10⁴ cells/cm² in 200 μL of medium per well, with nine replicates per condition. After 16 h, cells were treated with LPS and plasma/EVs as described in the experimental design. One hour before the end of treatment, 22 μL of CCK-8 reagent (Cat#: C0005, TargetMol, Boston, MA, USA) was added to each well and incubated for 1 h at 37 °C. Absorbance at 450 nm was then measured to evaluate cell viability.

2.3. Western blotting

Western blotting was performed as described (Huang et al., 2024). Briefly, the BV2 cells were homogenized using chilled RIPA buffer (Cat#: 89900, Thermo Fisher Scientific) with protease and phosphatase inhibitors (Cat#: 04693116001 and 04906837001, Roche, Basel, Switzerland). The homogenates were centrifuged at 15,000×g for 20 min at 4 °C to obtain the cell protein lysates. Protein concentrations of cell lysates and isolated EV samples were measured using a BCA kit (Cat#: 23225, Thermo Fisher Scientific). Protein samples (20 μg) were mixed with sample buffer (Cat#: S3401, Sigma‐Aldrich) containing 2% β-mercaptoethanol (Cat#:19‐1335, Sigma‐Aldrich), denatured by boiling, and resolved on polyacrylamide gels. Proteins were transferred to PVDF membranes (Cat#: IPVH00010, Merck‐Millipore), blocked with 5% skim milk, and incubated with primary antibodies at 4 °C overnight. Membranes were probed with HRP-conjugated secondary antibodies (goat anti‐mouse IgG: Cat#: 115‐035‐166; goat anti‐rabbit IgG: Cat#: 111‐035‐144, Jackson ImmunoResearch, West Grove, PA, USA). Bound antibodies were detected using an ECL kit (Cat#: WBKLS0500, Merck‐Millipore) and developed on medical X-ray film (Cat#: Super RX, Fujifilm, Minato City, Tokyo, Japan). In certain cases, membranes were stripped using a buffer containing 2% SDS, 62.5 mM Tris, and 0.8% β-mercaptoethanol at 55 °C before re-probing. Band densities were analyzed with ImageJ software (v2.0.0‐rc‐69/1.52p, U.S. National Institutes of Health, Bethesda, MD, USA). Relative protein expression was analyzed by normalizing with levels of respective total proteins. Antibody dilution ratios and exposure times were optimized for linear luminescence detection. Primary antibody details are listed in Supplementary Table 1. The obtained data values were normalized to the average value of the vehicle/control group, which was set to 1. This normalization strategy facilitated easier comparison of relative changes across treatment conditions.

2.4. Animals

Animal experiments were approved by the Institutional Animal Care and Use Committee of National Cheng Kung University, Tainan, Taiwan (approval code: 111302) and conducted in compliance with local and national guidelines. In adherence to the 3R principles, we used Sprague-Dawley (SD) rats as plasma donors, while C57BL/6N mice served as recipients to evaluate the in vivo anti-inflammatory effects of plasma and plasma-derived EVs.

The 7-week-old male SD rats were procured from BioLASCO (Taipei, Taiwan), and the 7-week-old male C57BL/6N mice were obtained from the National Cheng Kung University Laboratory Animal Center (NCKULAC), an AAALAC-accredited facility. All animals were group-housed, with two rats per cage and four to five mice per cage, in a specific-pathogen-free breeding unit at NCKULAC. Housing conditions were maintained at 55 ± 10% humidity and 24 ± 1 °C temperature, with a 13-h light/11-h dark cycle (lights on at 7 a.m.). Food and water were provided ad libitum. Animal health and the housing environment were monitored by qualified personnel and veterinarians. Animals were randomly assigned to groups using a computer-based randomization method (random number generator in Microsoft Excel, Microsoft Corporation, Redmond, WA, USA). Details on the treatment protocols and the number of animals used in each experiment are provided below and in Supplementary Table 2.

2.5. Treadmill-induced mandatory physical activity

The 7-week-old SD rats underwent a 1-week pre-experiment familiarization period to minimize stress from handling and environmental stimuli. Following this, the rats were randomly assigned to either the SED or TMPA groups. The TMPA group participated in a chronic physical activity protocol, beginning with a 1-week familiarization phase. During this phase, they walked on a leveled, motor-driven treadmill (Model #T510E; Diagnostic & Research Instruments Co., Taoyuan, Taiwan) at 10 m/min for 10 min per day, five days per week. This was followed by a formal 4-week training program. In the first week, the TMPA rats ran at 15 m/min for 20-60 min per day, with the duration increasing by 10 min daily, five days per week. From the second to fourth weeks, the rats ran for 60 min daily, five days a week, with the treadmill speed increasing by 1 m/min per week (16-18 m/min). The maximum speed reached during the TMPA protocol was 18 m/min (approximately 1 km/h), which corresponds to a walking pace for rats rather than a high-speed running condition. Accordingly, this protocol represents a controlled, mild-intensity mandatory locomotor stimulus. The front of the treadmill was covered with a dark cloth to encourage the rats to run forward, and a gentle touch on the tail was enough to keep them running. The entire training process was conducted without the use of tail shocks. The SED rats were placed on a stationary treadmill during the same time periods to ensure exposure to the same environment.

2.6. Plasma collection and EV isolation

Two days after completing the 4-week TMPA, both SED and TMPA rats were anesthetized with Zoletil (30 mg/kg, i.p.; Virbac, Carros, France) and xylazine (5 mg/kg, i.p.; Bayer AG, Leverkusen, Germany). Fifteen milliliters of blood were collected from the right cardiac ventricle using EDTA (Cat#: E4884, Sigma-Aldrich; 0.5M, 150 μL)-containing syringes and centrifuged at 3000×g for 10 min at 4 °C to separate the plasma. One batch of plasma (from five rats per group) was stored at −80 °C for assessing the anti-inflammatory effects of TMPA rat plasma in BV2 cell cultures. Another batch of five rats was processed for EV isolation.

EVs were isolated from plasma (8 mL/rat) using size exclusion chromatography columns (Model: qEV10/35 nm, Izon Science, Christchurch, New Zealand). Plasma samples were first centrifuged at 3000×g for 10 min to remove any large particles, followed by centrifugation of the supernatant at 10,000×g for 10 min. The resulting supernatant was loaded onto the column, and PBS filtered through a 0.22 μm membrane was used to elute the EVs at room temperature. Eight fractions (5 mL/fraction) were collected, and the second, third, and fourth fractions were used as the EV samples for subsequent experiments. A portion (1 mL/sample) of EVs was subjected to nanoparticle tracking analysis (NTA), while the remaining portions (14 mL/sample) of EVs were pooled and concentrated 12.5-fold using a TFF-Easy EV concentrator (Cat#: HBM-TFF/1, HansaBioMed, Tallinn, Republic of Estonia). For protein quantification, a small aliquot (2 μL) of the concentrated EV sample was lysed with an equal volume of RIPA buffer (Cat#: 89900, Thermo Fisher Scientific) containing protease and phosphatase inhibitors (Cat#: 04693116001 and 04906837001, Roche, Basel, Switzerland). The EV lysates were then diluted 40-fold with ultrapure water and subjected to BCA assay (Cat#: 23225, Thermo Fisher Scientific) to determine protein concentration.

2.7. Nanoparticle Tracking Analysis (NTA)

The collected EV specimens were diluted 10,000-fold with filtered PBS and analyzed using the NanoSight LM10HS equipped with a 405-nm laser (Malvern Panalytical, Malvern, Worcestershire, UK). EVs were visualized using an sCMOS camera at room temperature, and their mean particle size and concentration were measured. The detection range was maintained between 20 and 100 particles per frame. To prevent contamination, the PBS used for dilution was analyzed prior to each experiment, ensuring a particle count of fewer than 10 particles per frame.

2.8. Cryo-Transmission Electron Microscopy (cryo-TEM)

The ultrastructure of EVs was examined using cryo-TEM. Isolated EVs were applied onto charged carbon-coated grids and stained with 1% uranyl acetate (Cat#: 102092-284, Avantor-VWR, Radnor, PA, USA). Then the images of the stained EVs were captured by a field emission electron microscope (Model: JEM-2100F, JEOL, Tokyo, Japan) operated at an acceleration voltage of 200 kV.

2.9. DiI labeling for EVs

The isolated SED and TMPA EVs were labeled with the lipophilic fluorescent dye DiI (Cat#: D282, Thermo Fisher Scientific). DiI was prepared in absolute ethanol (Cat#: E7023, Sigma-Aldrich) at a concentration of 100 mg/16.7 mL and diluted 50-fold with PBS to create a 2% working solution. The 5 × 10¹² isolated EVs were incubated with 1 mL of DiI working solution for 30 min at room temperature. Subsequently, the labeled EVs were centrifuged at 100,000×g for 30 min at 4 °C, and the pellet was resuspended in PBS. This process was repeated three times to remove residual DiI.

2.10. In vitro EV uptake assay and immunocytofluorescence staining

BV2 cells were seeded on chamber slides (Cat#: 154534, Thermo Fisher Scientific) and incubated with DiI-labeled EVs (10 μg protein content/mL) for 1, 2, 4, and 8 h. After incubation, the cells were fixed with 4% paraformaldehyde in PBS for 10 min to prepare for immunocytofluorescence analysis. The cell membrane structure was visualized using Na⁺/K⁺ pump staining. Fixed cells were permeabilized with 0.1% Triton-X100 (Cat#: X100, Sigma-Aldrich) in PBS for 30 min, followed by overnight incubation at 4 °C with primary antibodies against the Na⁺/K⁺-ATPase α (Cat#: sc-28800, Santa Cruz Biotechnology, Dallas, TX, USA; Supplementary Table 1) in PBS containing 1% normal goat serum (Cat#: S26-M, Sigma-Aldrich). After washing, the cells were incubated for 1 h at room temperature with Alexa Fluor™ 488-conjugated goat anti-mouse secondary antibodies (Cat#: A-11001, Thermo Fisher Scientific). Slides were mounted using DAPI-containing mounting media (Cat#: ab104139, Abcam, Cambridge, UK) to counterstain cellular nuclei. Fluorescence images were captured using an optical fluorescence microscope (Model: Axio Imager A1, Carl Zeiss, Oberkochen, Germany) equipped with a digital camera (Model: Axiocam 305 Color, Carl Zeiss). The EV uptake index was determined by calculating the percentage of DiI-positive cells relative to the total number of DAPI-positive cells.

2.11. In vivo tracking of DiI-labeled EVs

To minimize animal use in accordance with welfare regulations, mice were chosen as recipients for rat-derived EVs in this study. In this part of experiment, two 8-week-old male C57BL/6N mice were used. One mouse received a tail vein injection of DiI-labeled EVs containing 400 μg of protein, while the other mouse was injected with an equal amount of unlabeled EVs via the same route to serve as a control. Two hours post-injection, the mice were anesthetized with Zoletil (50 mg/kg, i.p.; Virbac) and xylazine (2.5 mg/kg, i.p.; Bayer AG), and blood was collected from their heart to prepare blood smears. Then, they were transcardially perfused with PBS, and their brains were collected, fixed in 4% paraformaldehyde in PBS for two days, and dehydrated daily with gradually increasing concentrations (10, 20, 30, and 35%, twice per concentration) of sucrose (Cat# S9378, Sigma-Aldrich) solutions prepared in PBS. The dehydrated brains were embedded in a cutting compound (Cat# 3801480, Leica Biosystems, Wetzlar, Hessen, Germany) and sectioned into 40-μm coronal slices using a cryostat (Model: CM1950, Leica Biosystems). To adhere to stereological principle, sections were systematically sampled at a ratio of one out of every six. Fluorescent and bright-field images of the blood smear and four selected brain regions, including, the corpus callosum, hippocampus, substantia nigra, and cerebellum, were acquired using the previously described imaging system. ImageJ software (v2.0.0-rc-69/1.52p, U.S. National Institutes of Health) was utilized to stack the fluorescent and bright-field images.

2.12. LPS and EVs treatments in mice

Twelve 8-week-old male C57BL/6N mice, following a 1-week acclimatization period in our holding unit, were utilized for the experiment. The mice were randomly divided into four groups: PBS, LPS, LPS + SED-EVs, and LPS + TMPA-EVs. Mice in the LPS + SED-EVs and LPS + TMPA-EVs groups received tail vein injections of EVs containing 400 μg of protein, derived from SED and TMPA donor rats, respectively. One hour later, mice in the PBS and LPS groups were administered an intraperitoneal injection of either LPS (150 μg/kg in 10 mL PBS; Cat#: L2880, Sigma-Aldrich) or an equal volume of PBS as a control. All mice were sacrificed 24 h post-injection, and 40-μm coronal brain sections were prepared as described above. These brain sections were subjected to immunohistochemistry for a microglial marker to assess microglial reactivity in the targeted brain region of these mice.

2.13. Immunohistochemistry and image analysis

Immunohistochemistry for ionized calcium-binding adapter molecule-1 (Iba-1) was performed to evaluate the interactive effects of administered EVs and LPS on microglial activation in the substantia nigra of mice as described (Wang et al., 2022a, Wang et al., 2022b; Wu et al., 2020). Briefly, brain sections were rinsed in PBS containing 0.3% Triton X-100 (PBST, Cat#: X100, Sigma-Aldrich) to remove residual cutting compound, followed by treatment with 3% H₂O₂ (Cat#: H1009, Sigma-Aldrich) in PBST for 20 min at room temperature. The sections were blocked with PBST containing 3% normal goat serum (Cat#: S26-M, Sigma-Aldrich) for 1 h and incubated with primary antibodies against Iba-1 (Cat#: 019-19741, FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan; Supplementary Table 1) overnight at room temperature. Unbound primary antibodies were removed by washing with PBST, and the sections were incubated with HRP-conjugated goat anti-rabbit secondary antibodies (Cat#: 111-035-144, Jackson ImmunoResearch, West Grove, PA, USA) for 2 h at room temperature. Chromogen staining was performed using 3,3′-Diaminobenzidine (Cat#: D12384, Sigma-Aldrich). The stained sections were dehydrated and mounted using a xylene-based mounting medium (Cat#: 3801730, Leica Biosystems). To control for nonspecific binding, primary antibodies were omitted in some samples. Additionally, all sections from the same experimental set were processed simultaneously to minimize batch-to-batch variation.

In C57BL/6N mice, the substantia nigra spans approximately −2.5 mm to −3.9 mm relative to the bregma, covering a total rostrocaudal distance of 1.4 mm. This region was sectioned at 40 μm thickness, yielding around 35 sections in total, which were then divided into six sets. Each set consisted of approximately 5-6 evenly spaced sections, with one set selected for analysis per mouse. The five sections, spaced 240 μm apart, collectively covered the entire extent of the substantia nigra, ensuring systematic and representative sampling. The density of Iba-1-immunoreactive cells and the area fraction of Iba-1-positive signals in the entire substantia nigra were quantified for each section. A uniform background signal threshold was applied to all images, and signals exceeding this threshold were classified as positive. The data from the five sections were averaged to yield a single value per mouse.

2.14. Statistical analyses

Statistical analyses and graph generation were performed using Prism software (version 10.4.0, GraphPad Software Inc., San Diego, CA, USA), with significance defined as p < 0.05. The D'Agostino-Pearson normality test was applied to assess the assumption of normality. For normally distributed data, values are presented as mean ± standard deviation and analyzed using parametric tests. Data that did not meet the normality assumption or had relatively small sample sizes (n < 8) are presented as the median and interquartile range and analyzed using non-parametric methods. Parametric datasets with two independent variables (LPS vs. vehicle × TMPA vs. SED) were analyzed using ordinary two-way ANOVA followed by Tukey's multiple comparisons test. Mixed-model two-way ANOVA followed by Tukey's multiple comparisons test was used to analyze the size distributions of isolated EVs. For non-parametric datasets, comparisons between two groups were conducted using the Mann-Whitney U test, while experiments with more than two independent groups were analyzed using the Kruskal-Wallis test followed by Dunn's multiple comparisons test.

3. Results

3.1. Establishment of an LPS-induced inflammation model in BV2 microglial cells

To identify the optimal LPS treatment dosage, BV2 microglial cells were treated with a range of LPS concentrations (0, 1, 2, 5, 10, 25, 50, 100, 200, and 500 ng/mL) 16 h after seeding. One hour after treatment, the activation of NF-κB and MAP kinases, the key signaling pathways associated with LPS-induced inflammatory responses (Lu et al., 2008; Kawasaki and Kawai, 2014), was assessed (Fig. 1A). The results showed that LPS dose-dependently increased phosphorylation of p65 (p-p65), a subunit of NF-κB (p = 0.002), and JNK (p-JNK) (p 0.001, vs. 0 ng/mL of LPS) at this time point (Fig. 1B–D). In contrast, the phosphorylation levels of p38 and ERK1/2 were not significantly affected under these conditions (Fig. 1B). Therefore, we chose 100 ng/mL of LPS, which effectively and consistently activated LPS signaling pathways (Fig. 1B–D) to stimulate BV2 microglia in subsequent experiments.

Fig. 1.

Plasma derived from TMPA rats inhibits the LPS-induced activation of NF-κB in BV2 microglial cells. (A) Timeline for experiments identifying the optimal dosage of LPS treatment for BV2 cell. (B) Representative Western blot images showing phosphorylated levels of molecules involved in MAPK and NF-κB signaling in BV2 cells treated with various LPS doses at indicated time points. (C) Quantitative results of phosphorylated level of p65 in BV2 cells treated with various LPS doses at indicated time points. (D) Quantitative results of phosphorylated level of JNK in BV2 cells treated with various LPS doses at indicated time points. (E) Timeline for experiments evaluating the effects of plasma from SED and TMPA rats on p65 and JNK phosphorylation in BV2 cells treated with 100 ng/mL of LPS. (F) Quantitative results of cell viability using CCK-8 assay. (G) Representative Western blot images showing phosphorylated levels of p65 and JNK in BV2 cells treated with plasma from SED or TMPA rats followed by 100 ng/mL of LPS at indicated time points. (H) Quantitative results of phosphorylated level of p65 in BV2 cells treated with plasma from SED or TMPA rats followed by 100 ng/mL of LPS at indicated time points. (I) Quantitative results of phosphorylated level of JNK in BV2 cells treated with plasma from SED or TMPA rats followed by 100 ng/mL of LPS at indicated time points. Data in panels (C) and (D) were expressed as the median with interquartile range and analyzed with Kruskal-Wallis test followed by Dunn's multiple comparisons test. Data in panels (F), (G) and (H) were expressed as mean ± standard deviation and analyzed with ordinary two-way ANOVA followed by Tukey's multiple comparisons test. Statistical significances from post hoc multiple comparisons were indicated by asterisks: *p < 0.05, **p < 0.01, ***p < 0.001. Asterisks in panels (C) and (D) indicate significant differences compared to the 0 ng/mL group. n = 4 cultures/group in assays shown in panels (B–D) and n = 9 cultures/group in assays shown in panels (F–I).

3.2. Plasma derived from TMPA-treated rats attenuates LPS-induced NF-κB activation in BV2 microglial cells

We then investigated whether physical activity-associated peripheral factors could attenuate LPS-induced activation in BV2 microglial cells. Plasma samples from TMPA rats (collected two days after the final running session) and SED rats were added to BV2 cells 1 h before the LPS challenge. The levels of p-p65 and p-JNK were evaluated 1 h post-LPS treatment (Fig. 1E). CCK-8 assay showed that neither LPS (p > 0.5) nor plasma treatment (p = 0.405) affected BV2 cell viability (Fig. 1F). LPS treatment (100 ng/mL) significantly increased phosphorylation of both p65 and JNK in BV2 cells pre-incubated with either SED plasma (p-p65/p65: p < 0.001; p-JNK/JNK: p < 0.001) or TMPA plasma (p-p65/p65: p < 0.001; p-JNK/JNK: p < 0.001) (Fig. 1G–I). Notably, BV2 cells pre-treated with TMPA plasma exhibited lower levels of p-p65 following LPS stimulation than those pre-treated with SED plasma (p = 0.038) (Fig. 1G and H). However, no significant difference was observed in LPS-induced JNK phosphorylation between the BV2 cells pre-treated with TMPA and SED plasma (p > 0.5) (Fig. 1G and I). These findings suggested that plasma derived from TMPA-treated rats could mitigate LPS-induced NF-κB activation in BV2 microglial cells.

3.3. EVs isolated from SED and TMPA rats exhibit comparable quantity, size, and purity

Given that plasma components suppressed LPS-induced microglial activation, we investigated whether circulating EVs were responsible for this effect. EVs were isolated from the plasma of SED and TMPA rats (termed SED-EVs and TMPA-EVs, respectively) and their quantity and quality were evaluated. NTA results revealed that SED-EVs and TMPA-EVs exhibited similar particle numbers (p > 0.5) (Fig. 2A–D), size distributions (p = 0.054) (Fig. 2A–C), and mean sizes (p > 0.5) (Fig. 2E). Cryo-TEM confirmed that the majority of the EVs had diameters ranging from 130 to 150 nm (Fig. 2F), consistent with the NTA findings and within the typical size range of exosomes (Gurung et al., 2021). Western blot analysis showed no detectable albumin signals in the EV samples from either group (Fig. 2G), indicating the effective removal of plasma contaminants during the EV isolation process. Additionally, SED-EVs and TMPA-EVs expressed common EV protein markers, including CD9, CD81, TSG101, HSP90, and flotillin-1 (Deng and Miller, 2019). These results confirmed the comparability of SED-EVs and TMPA-EVs in terms of quantity, size, and purity.

Fig. 2.

Quantitative and qualitative characterization of EVs isolated from the plasma of SED and TMPA rats. (A) NTA results illustrating particle concentrations in EV samples derived from SED rats, with each trace representing the analysis of an individual sample from a single rat. (B) NTA results illustrating particle concentrations in EV samples derived from TMPA rats, with each trace representing the analysis of an individual sample from a single rat. (C) NTA results showing the average particle concentrations in EV samples from SED and TMPA rats, with each trace representing the mean values from EV samples isolated from five rats. (D) Quantitative results of total particle concentration of EV samples from SED and TMPA rats. (E) Quantitative results of particle mean size of EV samples from SED and TMPA rats. (F) Representative cryo-TEM images for circulating EVs derived from SED and TMPA rats. Scale bar: 100 nm. (G) Representative micrograph of Western blotting for albumin. (H) Representative micrographs of Western blotting for EV markers. Data in panel (C) are expressed as mean with standard deviation and were analyzed using a mixed-model two-way ANOVA followed by Tukey's multiple comparisons test. Data in panels (D) and (E) were expressed as the median with interquartile range and analyzed with Mann-Whitney U test. n = 5 rats/group in assays shown in panels (A–E).

3.4. TMPA-EVs attenuate LPS-induced NF-κB activation in BV2 microglial cells

We then investigated the effects of SED-EVs and TMPA-EVs on LPS-induced microglial activation. To track EVs, we labeled the isolated EVs with the hydrophobic fluorescent dye DiI. Our results showed that EVs were efficiently internalized by BV2 cells, with approximately 30% of the cells showing DiI signals 1 h after exposure to DiI-labeled EVs (Fig. 3A and B). The proportion of cells exhibiting DiI signals gradually increased, reaching a plateau of about 70% within 2 to 4 h post-treatment (Fig. 3A and B). Based on this EV incorporation dynamic, we treated BV2 microglia with SED-EVs and TMPA-EVs 1 h before LPS treatment and collected cell lysates 1 h after LPS exposure (Fig. 3C). The results of CCK-8 assay showed that neither LPS (p = 0.459) nor EVs treatment (p = 0.414) affected BV2 cell viability (Fig. 3D). Notably, BV2 cells pre-treated with TMPA-EVs exhibited lower levels of p-p65 compared to those pre-treated with SED-EVs (p < 0.001) (Fig. 3E and F). No significant differences were observed in LPS-induced JNK phosphorylation between BV2 cells pre-treated with SED-EVs and TMPA-EVs (p > 0.5) (Fig. 3E and G). These findings indicated that EVs derived from TMPA-treated animals attenuate LPS-induced NF-κB activation in BV2 microglial cells.

Fig. 3.

Circulating EVs derived from TMPA rats inhibits the LPS-induced activation of NF-κB in BV2 microglial cells. (A) Representative fluorescence images from the EV uptake assay in BV2 cells. The images in the rightmost column are magnified views of the areas outlined by red frames in the color-merged images. Scale bar: 50 μm. (B) Quantitative results of EV uptake assay in BV2 cells. (C) Timeline for experiments evaluating the effects of circulating EVs from SED and TMPA rats on p65 and JNK phosphorylation in BV2 cells treated with 100 ng/mL of LPS. (D) Quantitative results of cell viability using CCK-8 assay. (E) Representative Western blot images showing phosphorylated levels of p65 and JNK in BV2 cells treated with circulating EVs from SED or TMPA rats followed by 100 ng/mL of LPS at indicated time points. (F) Quantitative results of phosphorylated level of p65 in BV2 cells treated with circulating EVs from SED or TMPA rats followed by 100 ng/mL of LPS at indicated time points. (G) Quantitative results of phosphorylated level of JNK in BV2 cells treated with circulating EVs from SED or TMPA rats followed by 100 ng/mL of LPS at indicated time points. Data in panel (B) was expressed as the median with 95% confidence interval. Data in panels (D), (F) and (G) were expressed as mean ± standard deviation and analyzed with ordinary two-way ANOVA followed by Tukey's multiple comparisons test. Statistical significances from post hoc multiple comparisons were indicated by asterisks: ***p < 0.001. n = 3 cultures/group in assays shown in panels (A) and (B) and n = 9 cultures/group in assays shown in panels (D–G). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.5. TMPA-EVs attenuate LPS-induced microglial activation in the substantia nigra of mice

To monitor EV distribution in vivo, DiI-labeled EVs were injected into C57BL/6N mice via the tail vein, and blood and brain samples were collected 2 h later (Fig. 4A). Blood smear imaging showed detectable DiI signals in the circulating blood cells, confirming the successful injection (Fig. 4B). Furthermore, after thorough perfusion, DiI signals were detected in the corpus callosum, hippocampus, substantia nigra, and cerebellum (Fig. 4C). No DiI signals were detected in these brain regions of mice injected with unlabeled EVs (Fig. 4C), excluding the possibility of autofluorescence or other potential sources of false signals. These findings suggest that the administered exogenous EVs are capable of infiltrating the brain parenchyma of mice within relatively short period (e.g., 2 h).

Fig. 4.

Rat-derived EVs administered peripherally are able to penetrate into brain parenchyma of recipient mice. (A) Experimental timeline. (B) Representative images showing DiI fluorescent signals in prepared blood smears. Scale bar: 50 μm. (C) Representative images depicting DiI fluorescent signals in the brain parenchyma of mice following intravenous administration of DiI-labeled or unlabeled rat EVs. The rectangular frames in the leftmost column indicate the imaging areas for the subsequent columns. Scale bar: 100 μm.

To evaluate the effects of TMPA-EVs on LPS-induced microglial activation, mice were intravenously injected with either SED-EVs or TMPA-EVs, each containing an equal protein amount (400 μg), via the tail vein 1 h prior to intraperitoneal administration of LPS at a dose of 150 μg/kg (Fig. 5A). Mice were sacrificed 24 h later (Fig. 5A) to examine the degree of microglial activation. Microglia were identified by immunostaining Iba1, a marker for resting and activated microglia. We defined the degrees of microglial activation by the density of Iba-1-immunoreactive (Iba-1⁺) cells and the extent of Iba-1⁺ signal areas in the sampled sections as previously described (Wang et al., 2022a, Wang et al., 2022b; Wu et al., 2020; Yang et al., 2013). The expression pattern of Iba-1 was specifically analyzed in the substantia nigra, as this region exhibited significant infiltration of exogenous EVs (Fig. 4C), a high density of microglia, and increased susceptibility to LPS-induced microglial activation (Yang et al., 2013). Our findings revealed that intraperitoneal injection of LPS at 150 μg/kg did not significantly increase the density of Iba-1⁺ microglia in the substantia nigra (p > 0.5) (Fig. 5B and C). Similarly, the administration of SED-EVs and TMPA-EVs had no effect on the density of Iba-1⁺ microglia (p > 0.5) (Fig. 5B and C). However, LPS significantly increased the Iba-1⁺ signal areas in the substantia nigra (p = 0.011) (Fig. 5B and D). This increase was reversed by treatment with TMPA-EVs (p = 0.032) (Fig. 5B and D). Given that neither LPS nor EV treatments altered the density of Iba-1⁺ cells, the increased Iba-1 signal likely reflects functional activation rather than microglial proliferation. These findings suggest that physical activity-associated EVs suppressed LPS-induced microglial activation in the substantia nigra.

Fig. 5.

Circulating EVs derived from TMPA rats inhibits the LPS-induced microglia activation in the substantia nigra in recipient mice. (A) Timeline for experiments evaluating the effects of circulating EVs from SED and TMPA rats on nigra microglial activation in mice intraperitoneally injected with 150 μg/kg of LPS. (B) Representative immunohistochemistry images for Iba-1 staining. Rectangular frames in the upper row mark the imaging areas shown in the lower row. Scale bar: 50 μm. (C) Quantitative analysis of Iba-1-immunoreactive cell density in the substantia nigra of recipient mice. (D) Quantitative analysis of the area fraction of Iba-1-immunoreactive signals in the substantia nigra of recipient mice. Data in panels (C) and (D) were expressed as the median with interquartile range and analyzed with Kruskal-Wallis test followed by Dunn's multiple comparisons test. Statistical significances from post hoc multiple comparisons were indicated by asterisks: *p < 0.05. n = 3 mice/group in assays shown in panels (B–D).

4. Discussion

Excessive activation of microglia, resulting in neuroinflammation, is a significant contributor to the onset of various neurodegenerative diseases (Muzio et al., 2021; Gao et al., 2023). Excessive microglial activation in the substantia nigra is strongly associated with the pathogenesis of Parkinson's disease, leading to dopaminergic neuron loss and motor impairments. While prolonged TMPA has been shown to mitigate these effects by reducing local microglial activation (Wang et al., 2022a, Wang et al., 2022b; Wu et al., 2020), the mechanisms underlying these anti-inflammatory effects remain unclear. This study aimed to explore the role of circulating physical activity-related EVs in this process. Using LPS-induced in vitro and in vivo models of microglial activation, we initially found that plasma from TMPA rats inhibits LPS-induced NF-κB activation in BV2 cells, emphasizing the anti-inflammatory potential of physical activity-associated circulating factors. Furthermore, we demonstrated that exogenous EVs can be absorbed by BV2 microglia and penetrate the brain parenchyma in recipient mice. Treating with circulating EVs derived from TMPA rats replicated the inhibitory effects on LPS-induced microglial activation in both in vitro and in vivo models. These findings not only deepen our understanding of the anti-neuroinflammatory properties of physical activity-derived EVs but also open avenues for future research into their bioactive components as potential therapeutic agents for neuroinflammation-related diseases.

Our results demonstrated that both plasma and EVs from TMPA rats attenuated LPS-induced activation of the NF-κB pathway in BV2 cells, while LPS-induced changes in the MAPK pathway were unaffected. This selective modulation aligns with previous reports (Schlegel et al., 2012; Yuan et al., 2013) indicating that the NF-κB and MAPK pathways can be independently regulated. These findings suggest that specific, yet unidentified, factors present in the circulation or EVs of TMPA-treated animals may target the NF-κB pathway selectively, rather than broadly inhibiting LPS-induced signaling. The NF-κB family plays a central role in mediating immune and inflammatory responses by regulating the expression of genes involved in cytokine and chemokine production, apoptosis, and cell proliferation. The NF-κB family functions as key transcription factors in immune-driven inflammatory responses, regulating the expression of genes involved in diverse processes such as cytokine and chemokine modulation, apoptosis, and cell proliferation (Oeckinghaus and Ghosh, 2009; Pahl, 1999). As the resident macrophage-like innate immune cells in the central nervous system, microglia are tightly regulated by NF-κB during inflammatory responses in various pathological conditions. For example, activated NF-κB signaling promotes pro-inflammatory M1 polarization in an intracerebral hemorrhage model (Chen et al., 2024). In Alzheimer's disease, amyloid plaques stimulate microglial NF-κB pathways, resulting in increased pro-inflammatory gene expression and contributing to a neurodegenerative environment (Ho et al., 2005; Wang et al., 2023). Additionally, NF-κB signaling activated by tau drives microglial-mediated tau propagation and toxicity in a tauopathy mouse model (Wang et al., 2022a, Wang et al., 2022b). In a mouse model of amyotrophic lateral sclerosis (ALS) with mutated superoxide dismutase 1, inhibition of microglial NF-κB rescued motor neurons and improved survival rates (Frakes et al., 2014). Moreover, mutant TAR DNA-binding protein 43 (TDP-43), another factor implicated in ALS pathogenesis, has been linked to NF-κB activation. Spinal cord extracts from patients with sporadic ALS showed colocalization of TDP-43 with the p65 NF-κB subunit in microglia, suggesting that the canonical NF-κB pathway may contribute to the deregulation of TDP-43 characteristic of ALS (Swarup et al., 2011). These findings collectively underscore the importance of targeting NF-κB signaling in microglia to mitigate inflammation-related neurodegeneration and highlight the therapeutic potential of physical activity-associated EVs, which may serve as a natural intervention to counteract neurodegenerative diseases by leveraging the innate mechanisms of body. On the other hand, it is worth noting that phosphorylation of other MAPKs, such as p38 and ERK1/2, was not significantly induced following LPS treatment at various dosages. This may be attributed to the selected time point for assessment (i.e., 1 h after LPS treatment). In a previous study, we reported significant increases in phosphorylated p38 and ERK1/2 levels 2 h after LPS treatment at a dosage of 500 ng/mL in BV2 cells (Huang et al., 2024). This temporal variation should be considered when interpreting our results.

Aerobic physical activity is well-documented to transiently increase the release of EVs into circulation (Fruhbeis et al., 2015; Vanderboom et al., 2021; Whitham et al., 2018b). In the plasma of healthy individuals who performed an incremental cycling or running protocol to exhaustion, EVs measuring 100-130 nm and expressing exosome protein markers significantly increased immediately after cycling, then declined after a 90-min rest. In treadmill running, the elevation of these small EVs was more moderate but sustained, gradually declining between 90 and 360 min post-physical activity (Fruhbeis et al., 2015). The kinetics of circulating EV concentration revealed that EV release begins in the early phase of physical activity, prior to reaching the anaerobic threshold marked by lactate elevation. Moreover, this acute physical activity-induced EV release was not evident in individuals undergoing resistance physical activity training (Vanderboom et al., 2021; Conkright et al., 2022; Gu et al., 2022; Just et al., 2020). Collectively, these findings indicate that a single bout of physical activity transiently, rather than persistently, elevates circulating EV levels and that both the type and intensity of physical activity influence the acute release patterns of EVs. Our results of NTA showed that the number of circulating EVs was similar between TMPA and SED rats two days after completing a 4-week TMPA training regimen. This result suggested that long-term physical activity training did not alter the basal levels of circulating EVs. In agreement with our findings, studies in humans have also reported no significant changes in baseline circulating EV levels before and after chronic aerobic physical activity training (Maggio et al., 2023; Kargl et al., 2024). Notably, a diminished physical activity-induced release of EVs has been observed in trained individuals (Maggio et al., 2023), indicating that chronic physical activity training induces adaptations that regulate the magnitude of EV release in response to physical activity. These findings imply that repeated exposure to aerobic physical activity may fine-tune the EV release mechanism, potentially reflecting an improved physiological efficiency or a reduced need for acute stress signaling. Further research is needed to elucidate the underlying mechanisms of this adaptation and to determine whether it confers specific benefits related to physical activity performance or recovery.

Although long-term physical activity training does not affect the basal levels of circulating EVs, some studies have reported its effect on EV compositional content (Kargl et al., 2024; Hou et al., 2019). In humans, a 12-week combined resistance and endurance physical activity program increased EV microRNA (miRNA) content by 29% (Kargl et al., 2024). While the most abundantly expressed miRNAs showed similarities between men and women, notable sex-dependent differences were observed in miRNA expression profiles following physical activity training. In men, resting EV miRNA content underwent substantial changes, with 31 miRNAs downregulated and 8 upregulated after 12 weeks of training compared to baseline. Conversely, women exhibited a smaller effect, with only 7 miRNAs downregulated and 5 upregulated during the same period (Kargl et al., 2024). These altered EV miRNA components were predicted to be involved in various biological pathways, including muscle function and hypertrophy, myokine signaling, bone health, and metabolism (Kargl et al., 2024). Similarly, in rats, a 4-week swimming training regimen significantly elevated the expression levels of 11 miRNAs in circulating EVs (Hou et al., 2019). Among these, miR-342-5p emerged as a critical component contributing to the cardioprotective effects of physical activity-associated EVs (Hou et al., 2019). While few studies have investigated the effects of long-term physical activity training on EV cargo composition, numerous studies have demonstrated that acute physical activity significantly alters EV content. At the protein level, Kobayashi and colleagues identified 558 distinct proteins within circulating EVs isolated from participants. They further demonstrated that a high-intensity interval training regimen significantly altered 20 of these proteins, including alpha-2-antiplasmin, von Willebrand factor, multimerin-1, fibrinogen, and carbonic anhydrase-1, among others. These proteins are involved in various pathways, such as coagulation cascade activation, cellular oxidant detoxification, and acid-base balance regulation (Kobayashi et al., 2021). Additionally, acute physical activity was shown to elevate the expression of exosome protein markers such as flotillin-1, Alix, CD81, and exosome-carried heat shock proteins (Estebanez et al., 2021). At the miRNA level, acute cycling exercise has been shown to increase the expression of miR‐1 in circulating EVs, with levels returning to baseline within 4 h of recovery (D'Souza et al., 2018). Moreover, exosome-carried miRNAs, including miR‐1, miR‐133a, miR‐133b, miR‐206, miR‐208a, and miR‐499, were elevated immediately after physical activity and reverted to baseline levels 48 h into recovery (Yin et al., 2019). Importantly, Sullivan et al. recently demonstrated that miRNA composition changes in skeletal muscle-derived EVs following one week of concurrent aerobic and resistance physical activity training were linked to the regulation of inflammatory processes, including the Toll-like receptor and NF-κB pathways (Sullivan et al., 2022). These findings, combined with our results, highlight the critical importance of identifying the bioactive components in physical activity-derived EVs. Such efforts are essential for understanding whether, and through what mechanisms, physical activity-associated EVs could serve as therapeutic agents for neuroinflammation-related diseases. Our previous work demonstrated that long-term TMPA suppresses nigral microglial activation via a brain-derived neurotrophic factor-dependent pathway (Wang et al., 2022a, Wang et al., 2022b). Therefore, investigating EV cargos responsive to physical activity and involved in the brain-derived neurotrophic factor pathway warrants initial exploration and further investigation.

We observed that DiI-labeled exogenous EVs administered peripherally were detectable in the brains of recipient mice. This aligns with established findings that EVs can cross the blood-brain barrier bidirectionally, facilitating communication between the bloodstream and brain parenchyma (Ramos-Zaldivar et al., 2022). Additionally, some studies suggest that exosomes can traverse the blood-cerebrospinal fluid barrier, further highlighting their role in blood-brain communication (Balusu et al., 2016). These findings underscore the importance of EVs in mediating interactions between the peripheral circulation and the central nervous system. However, a limitation of lipophilic dyes (e.g., DiI) is that detergents such as Triton-X 100 or alcohol-based fixatives can strip away the lipid layer and dye during cell permeabilization and fixation. Consequently, it becomes challenging to determine which cell types are the primary recipients of injected exosomes using immunofluorescence staining. In this context, a novel CD63-GFP transgenic rat model may provide a potentially more effective method for tracing these EVs (Yoshimura et al., 2016). Regarding the source of EVs induced by physical activity, Alexandra Brahmer et al. introduced the term “ExerVs” for EVs released during physical activity and proposed that platelets, endothelial cells, and leukocytes significantly contribute to their release (Brahmer et al., 2019). They also suggested other cell types as potential sources, but the absence of highly specific markers leaves this question unresolved (Brahmer et al., 2019). The Rab protein family, known to regulate endocytic and secretory pathways, may offer valuable insights. Certain Rab proteins, such as Rab11, Rab27, and Rab35, play crucial roles in EV biogenesis and secretion (Blanc and Vidal, 2018). Exploring the functions of these proteins using loss- and gain-of-function approaches, both in vitro and in vivo, could enhance our understanding of the cellular origins of physical activity-associated EVs.

In addition to the proposed future directions, several limitations of the present study should be acknowledged. First, only male animals were included in our experiments. As estrogen has been shown to modulate microglial reactivity (Wu et al., 2016), potential sex differences in response to the intervention remain unexplored. Future studies should consider including both sexes to fully elucidate the role of sex-specific factors. Second, this study employed a TMPA paradigm, a controlled and non-volitional form of physical activity characterized by low-speed, treadmill-based walking (maximum 1 km/h) that may not fully reflect the physiological, behavioral, or motivational features of voluntary human exercise and may engage stress-related responses influencing neuroimmune signaling. The intensity of this paradigm corresponds to a mild physical activity rather than a high-intensity exercise condition. These factors should be considered when interpreting the findings and their translational relevance. Third, our in vivo experiments did not include LPS-untreated control mice that received either SED-EVs or TMPA-EVs. Moreover, the sample sizes in several experiments were relatively small, as these studies were intentionally designed as proof-of-concept investigations to support our in vitro findings while adhering to ethical guidelines for the use of laboratory animals. Despite limited numbers, the experiments were performed using rigorous and appropriate statistical analyses, yielding biologically meaningful and internally consistent results. In particular, the in vivo extracellular vesicle tracking experiment was designed as an exploratory, qualitative assessment to evaluate feasibility and anatomical distribution rather than to draw quantitative or inferential conclusions. These observations are not intended for statistical comparison, and no quantitative claims are derived from them. Additionally, the cross-species transfer of EVs from rats to mice was used as an exploratory approach to minimize animal use and assess feasibility; species-specific differences may exist, and these findings should be interpreted cautiously. Future studies using same-species EVs and larger cohorts with additional control groups will be essential to validate and extend these observations. Fourth, our study was designed to evaluate the prophylactic effects of physical activity-associated EVs by administering them prior to LPS stimulation, emphasizing the preventive benefits linked to regular physical activity. This approach does not capture the therapeutic potential of EVs, and future studies should investigate EV administration following LPS exposure to better assess clinical and translational applicability. Fifth, a notable limitation is the absence of co-staining with cell-type-specific markers to identify which brain cell types internalized the DiI-labeled EVs. Determining the recipient cell populations is crucial for understanding the cellular mechanisms underlying EV-mediated neuroimmune modulation. Finally, while our findings indicate that exogenous EVs can reach the brain parenchyma within 2 h, fluorescent signals alone do not confirm blood-brain barrier transit. Further investigations using time-course imaging, in vitro models, and cell-specific EV cargo detection are needed to clarify the transport mechanism. These limitations, including sample size, exploratory design, TMPA characteristics, and cross-species EV transfer, should be considered when interpreting the results.

In conclusion, our study highlighted the pivotal role of physical activity-associated EVs in mitigating neuroinflammation through the suppression of microglial activation. By demonstrating that EVs derived from TMPA rats inhibited LPS-induced NF-κB activation and microglial activation in both in vitro and in vivo models, we provide compelling evidence of their anti-inflammatory potential. These findings suggested that physical activity-associated EVs may serve as crucial mediators of the neuroprotective effects of physical activity, offering a novel perspective on the interplay between peripheral circulation and central nervous system health. Notably, our results indicate that even low levels of enforced locomotion, such as slow treadmill walking, may be sufficient to elicit these EV-mediated anti-inflammatory effects. Additionally, the finding that physical activity-associated EVs from rats exerted anti-inflammatory effects in mouse recipients provides initial evidence supporting the potential of cross-species EV-based interventions, highlighting their relevance for future translational applications. Furthermore, our results underscore the need for future investigations into the bioactive components of these EVs and their molecular mechanisms, paving the way for innovative therapeutic strategies targeting neuroinflammation and related neurodegenerative diseases.

CRediT authorship contribution statement

Chien-Yu Su: Writing – review & editing, Methodology, Investigation, Funding acquisition. Heng-Juei Hsu: Writing – review & editing, Methodology, Investigation, Funding acquisition. Tzu-Feng Wang: Visualization, Methodology, Investigation, Formal analysis, Data curation. Hong-Lin Chou: Visualization, Methodology, Investigation, Formal analysis, Data curation. Sheng-Feng Tsai: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation. Chu-Wan Lee: Methodology. Ching-Li Cheng: Methodology. Yang-Kao Wang: Writing – review & editing, Resources, Conceptualization. Yu-Min Kuo: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT and Perplexity tool/service to improve language and readability. After using these tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Funding

This study was supported by grants from Ditmanson Medical Foundation Chia-Yi Christian Hospital Research Program, Taiwan (Grant #: CNGMP11301, CYS and YMK) and National Science and Technology Council, Taiwan (Grant #: NSTC 113-2320-B-006-010-, YMK).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is/are the supplementary data to this article:

Data availability

Data will be made available on request.