Activated Microglia-Derived Extracellular Vesicles Elicit a Pro-Inflammatory Astrocytic Response via Cargo-Dependent Mechanisms

Abstract

Neuroinflammation plays a dual role in brain health supporting defense and repair, but causes neurotoxicity when persistent. Microglia and astrocytes coordinate these responses through cytokine signaling and extracellular vesicles (EVs), though their vesicle-mediated communication remains unclear. This study investigated whether EVs from activated microglia (ABEVs) influence astrocyte polarization and inflammatory signaling. BV-2 microglial cells were activated with lipopolysaccharide (LPS), and microvesicle (ABMVs) and exosome (ABEXs) EVs were isolated via sequential ultracentrifugation. Primary mouse astrocytes were treated with LPS, ABMVs, or ABEXs, and expression of reactive astrocyte markers (C3, Serpina3n, S100a10, Sphk1) and inflammatory mediators (Lcn2, Il-1β, Ccl2, Ccl5, Cxcl10) was quantified, and EV protein cargo was analyzed by mass spectrometry and proteomics. LPS-treated astrocytes exhibited increased C3 and Serpina3n and decreased S100a10, consistent with reactive polarization. ABEXs mimicked this effect, significantly inducing C3, Serpina3n, and Sphk1, whereas ABMVs had a weaker influence. ABEXs also upregulated Lcn2 and Il-1β, partially reproducing microglial inflammatory effects. Proteomic profiling revealed marked cargo differences: ABEXs exhibited 16 upregulated proteins linked to NOD-like receptor signaling compared to non-activated BEXs, and 165 proteins associated with ribosome biogenesis and spliceosome pathways compared to ABMVs, indicating subtype-specific signaling potential. Collectively, our findings demonstrate that microglia-derived EVs modulate astrocytic polarization and cytokine profiles in a cargo-dependent manner, emphasizing their importance in interglial communication and revealing novel targets for neuroinflammatory modulation.

1. Introduction

Neuroinflammation is a crucial immune response in the central nervous system (CNS), defending against pathogens, restoring damaged glia and neurons, removing cell debris and aiding in tissue repair [1,2,3]. However, when excessive, it can hinder neuronal regeneration, leading to neurodegenerative diseases (NDDs) like multiple sclerosis (MS), Parkinson’s disease, or Alzheimer’s disease (AD) [2,4]. This inflammation, triggered by various stimuli, involves microglial activation, cytokine release, immune cell recruitment, and local tissue damage, orchestrated mainly by microglia and astrocytes [3,5,6,7].

Microglia, which comprise 10–15% of all glial cells, are pivotal in CNS immune surveillance, expressing various pattern recognition receptors to detect pathogens and damage-associated molecules [8,9,10]. Upon activation, they polarize into M1/pro-inflammatory or M2/anti-inflammatory phenotypes, releasing pro- and anti-inflammatory factors, respectively [4,10]. Dysregulated microglial activation has been implicated in the aforementioned NDDs as well as amyotrophic lateral sclerosis (ALS), perpetuating chronic inflammation and disease progression [11,12,13]. Toll-like receptors and cytokines mediate microglial inflammatory responses, with lipopolysaccharides (LPS) commonly inducing neuroinflammation and M1 state of microglial activation [4,11,14,15].

Astrocytes, the most abundant cell type in the CNS (40–60%), play crucial roles in maintaining homeostasis, supporting neuronal metabolism, forming synapses, and regulating neuronal activity [16,17,18,19,20]. In response to stimuli like inflammation or injury, astrocytes undergo transcriptional remodeling, leading to morphological changes and the secretion of various, like inflammatory or neuroprotective, factors, termed reactive astrogliosis [21]. Reactive astrocyte phenotypes are complex and context-dependent, and include different transcriptomic patterns, morphological variability, and functional diversity [22]. As one concept to describe different states of reactive astrocytes, a set of molecular markers were used to describe two distinct phenotypes referred to as A1/pro-inflammatory and A2/anti-inflammatory states, analogous to the polarization seen in microglia [3,23,24]. A1 astrocytes are associated with neuroinflammation and are described to exhibit detrimental properties, such as reduced phagocytic activity and neurotoxicity, whereas A2 astrocytes upregulate neurotrophic factors promoting neuronal survival and growth [3,24,25].

The interaction between microglia and astrocytes is crucial for maintaining normal CNS health and function. Recent research has highlighted extracellular vesicles (EVs) as essential mediators of physiological and pathological intercellular communication, providing insights into inflammatory diseases [14,26]. Once considered cellular debris, EVs now reveal a novel mode of cell-to-cell signaling, allowing for long-range regulation of gene expression and cellular differentiation [27,28,29]. EV production by microglia occurs in both resting and activated states, with different effects depending on microglial activation [28,30,31]. Accumulating evidence links EVs to neuroinflammation and NDDs, demonstrating that pro-inflammatory EVs transmit inflammatory cues and carry cytokines such as IL-1β, IL-6, TNF-α, and caspase-1 [32,33,34,35]. Moreover, EVs encapsulated anti-inflammatory drugs hold therapeutic potential, as demonstrated by studies showing protective effects against brain inflammation and tumor growth when administered intranasally in mice [36].

Mechanistically, EVs are generated either through inward budding of endosomal membranes or direct outward budding of the plasma membrane, giving rise to three main subtypes—apoptotic bodies, microvesicles (ectosomes), and exosomes—distinguished by their size and biogenetic pathways [37,38,39,40]. The classification of EVs remains challenging due to their overlapping physical and biochemical characteristics [41]. Apoptotic bodies (500–4000 nm) originate from cells undergoing programmed cell death, whereas microvesicles (MVs), ranging from 100–1000 nm in diameter, are generated through the outward budding of the plasma membrane [27,40,42,43]. In contrast, exosomes (EXs), typically 50–150 nm in size, are formed via the endosomal pathway during multivesicular body maturation and subsequent exocytosis [39,42]. Upon release, EXs migrate through the interstitial space and circulation to interact with recipient cells, mediating a wide range of biological functions such as antigen presentation, immune modulation, tumor progression, and regulation of synaptic activity within the central nervous system [44,45]. EVs, particularly EXs, are present in diverse body fluids—including plasma, saliva, urine, and cerebrospinal fluid—and carry bioactive cargos such as proteins and microRNAs that play key roles in physiological and pathological processes [44,46]. The role of EVs in modulating inflammatory pathways and their potential as biomarkers for early detection and therapeutic monitoring of NDDs is increasingly recognized, offering new avenues for diagnosis and disease management [7,33]. This research aimed to elucidate the modulatory effects of microglia-derived EVs on astrocyte functional states and cytokine expression, providing mechanistic insight into glia-to-glia communication during neuroinflammation.

2. Materials and Methods

2.1. Cell Culture

2.1.1. Microglial Culture and Primary Astrocyte Isolation

Cells of the microglial cell line BV-2 were cultured according to Dr. E. Blasi (RRID: CVCL_0182; Modena, Italy) [47]. This cell line was not authenticated before experiments and is not listed as a commonly misidentified cell line by the International Cell Line Authentication Committee. All animals used in the present study were acquired and cared for in accordance with the Federation of European Laboratory Animal Science Associations (FELASA) recommendations. The animals were housed in a 12 h light and dark cycle with controlled temperature and humidity (23 ± 2 °C; 55 ± 10%). Wild-type C57BL/6J mice were obtained from Janvier Labs (Le Genest-Saint-Isle, France). Primary astrocyte-enriched cultures were prepared from postnatal day 1–3 mice as previously described [48]. Briefly, the telencephalic cortex and adjacent white matter were isolated, meninges and plexus choroideus removed, and brain tissue incubated in ice-cold HEPES buffer. After mechanical homogenization, the cell suspension was strained with a 70 µm cell strainer and pelleted by centrifugation at 800× g. The pellet was resuspended in Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal calf serum (FCS), 50 U/mL penicillin, and 50 μg/mL streptomycin (0.5% Pen/Strep), transferred to poly-L-ornithine (PLO)-coated flasks, and incubated at 37 °C and 5% CO₂. All DMEM used in cell culture throughout the study contained 0.5% Pen/Strep. After 4 days, other glial cell types were removed by shaking and medium replacement. Subsequent medium changes were done every 2 to 3 days until the cell culture was confluent. For passaging the BV-2 cells and primary astrocytes, cells were detached using 0.1% trypsin in 2% EDTA/PBS and collected in DMEM with 10% FCS. After centrifugation at 300× g at room temperature for 5 min, the pellet was suspended in fresh DMEM with 10% FCS and passaged 1:20 into uncoated flasks for BV-2 cells and 1:3 into PLO-coated flasks for primary astrocytes. Cells were always incubated at 37 °C and 5% CO₂.

2.1.2. Treatment of Microglial Cells and Primary Astrocytes

BV-2 cells were seeded into poly-D-lysine hydrobromide (PDL)-coated (Sigma Aldrich, St. Louis, MO, USA) 10 cm dishes at a density of 86,000 cells/cm² in DMEM with 10% FCS. The next day, the cells were treated with 100 ng/mL lipopolysaccharides (LPS, Sigma-Aldrich, Merck, St. Louis, MO, USA) in DMEM without FCS for 24 h. Primary astrocytes were seeded into PLO-coated 6-well dishes at a density of 30,000 cells/cm² in DMEM with 10% FCS. The next day, the medium was replaced with DMEM with 0.5% FCS, and the day after, treatment with 100 ng/mL LPS in DMEM with 0.5% FCS or with 40 µL isolated EVs (4 × 10¹⁰ particles/mL) in DMEM with 0.5% FCS occurred for 24 h. See the Section 2.4 and Section 2.5 for further astrocyte analysis.

2.1.3. Extracellular Vesicle Isolation

EVs were isolated from the cell culture supernatants of BV-2 cells. After two successive centrifugation steps at 300× g for 10 min and at 2000× g for 20 min at 4 °C, the supernatants were transferred into ultracentrifugation tubes (Beckman Coulter, Brea, CA, USA). Supernatants were ultracentrifuged at 8000× g at 4 °C for 45 min using the rotor SW 32 Ti and the Optima LE-80K ultracentrifuge (Beckman Coulter) and filtered through a 0.20 µm filter (Corning, New York, NY, USA) into new ultracentrifugation tubes. Microvesicle-enriched pellets were dried and either resuspended in DPBS for further treatment of astrocytes or in EDTA-free protease inhibitor cocktail (RIPA+), consisting of a 1:7 dilution of cOmplete mini (Roche, Basel, Switzerland) and RIPA buffer for protein analysis. The supernatant was then ultracentrifuged at 110,000× g at 4 °C for 105 min, and the pellets were washed with ice-cold DPBS using another ultracentrifugation step at 110,000× g at 4 °C for 105 min. Exosome-enriched pellets were dried and resuspended in DPBS or RIPA+, respectively.

2.2. Nanoparticle Tracking Analysis (NTA)

The analysis was conducted with the NS300 and the Nano Sight 3.2 Dev Build 3.2.16 software using a 488 nm laser module. All settings were induced according to the manufacturer’s instructions. Briefly, the flow cell was cleaned by introducing water into the flow chamber using a syringe. Samples were diluted 1:100 with DPBS and introduced into the chamber with a syringe, and optimal image settings were made by adjusting the focus, camera level, and detection threshold. After reaching the default set temperature of 25 °C, particles were captured five times for 60 s as raw data for statistical analysis. After the measurement, different-sized cut-offs at 100 nm, 200 nm, and 299 nm were set to calculate the concentrations.

2.3. IncuCyte Tracking Analysis and Confocal Microscopy

Before IncuCyte Tracking analysis and confocal microscopy, BV-2-derived exosomes (BEXs) were stained with Vybrant DiI Cell-Labeling Solution (V22885, Invitrogen, Carlsbad, CA, USA) for 10 min on ice in the dark. The solution was diluted with DPBS and ultracentrifuged at 110,000× g at 4 °C for 105 min. The BEX-pellet was resuspended in DPBS for primary astrocyte treatment. For the IncuCyte tracking analysis, primary astrocytes were seeded into a PLO-coated 96-well plate with a density of 46,000 cells/cm² in DMEM with 10% FCS. The next day, cells were treated with 2 µL dyed BEXs. Immediately after treatment, nine images per hour were taken with the IncuCyte SX5 Live-Cell Analysis System from Sartorius for 24 h. The images were then analyzed using the IncuCyte Software IncuCyte 2021C. For the confocal microscopy, primary astrocytes were seeded on PLO-coated coverslips at a density of 30,000 cells/cm² in DMEM with 10% FCS. The next day, the medium was replaced with DMEM supplemented with 0.5% FCS, and the day after, treatment with dyed BEXs in DMEM with 0.5% FCS occurred for 3 h. Cells were washed with DPBS and fixed with 4% formaldehyde/PBS. Immunofluorescence staining was performed in a wet chamber. Cells were permeabilized with 0.2% Triton X-100 (#3051, Roth, Karlsruhe, Germany) in 1X PBS and blocked with IFF buffer (1% bovine serum albumin (BSA) and 2% FCS in PBS). All further steps were conducted in the dark. Incubation with the primary antibody Aldehyde Dehydrogenase 1 Family Member L1 (ALDH1L1, ab87117, Abcam, Cambridge, UK) diluted 1:600 in IFF buffer was conducted overnight at 4 °C. The next day, the cells were incubated with the secondary antibody Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody (Invitrogen, #A-11008) diluted 1:500 in IFF buffer. Cells were counterstained 1:10,000 with Hoechst 33342 solution (#H3570, Thermo Fisher Scientific, Waltham, MA, USA) and mounted in Immu-Mount (#9990402, Epredia, Kalamazoo, MI, USA) on coverslips. For confocal imaging, an LSM 710 confocal microscope with Airyscan and Zen 2.1 SP3 Software (ZEISS, Oberkochen, Germany) was used. The acquired images were further processed with the Fiji software based on ImageJ 1.54f [49]. 3D projections were generated out of 12 individual slices with a distance of 0.18 µm between them.

2.4. Gene Expression Analysis

RNA of BV-2 cells and primary astrocytes was isolated by using RNA-Solv^® reagent (#R6830-02, VWR, Darmstadt, Germany) with a standardized protocol as described previously [50] and reverse-transcribed using the M-MLV reverse transcription (RT) kit and random hexanucleotide primer (Invitrogen™, ThermoFisher Scientific). cDNA levels were analyzed by semi-quantitative real-time PCR using SensiMix SYBR^® & Fluorescein Kit (QT615-05, Bioline, Cincinnati, OH, USA). The expression levels were calculated relative to the reference genes coding for Hypoxanthine-guanine phosphoribosyltransferase (Hprt) or 18S, applying the ∆∆Ct method according to Pfaffl [51]. Primer sequences and annealing temperatures are given in

Table 1. Primer sequences and annealing temperatures.

Primer	Sense	Anti-Sense	Annealing Temperature [°C]
18S	CGGCTACCACATCCAAGGAA	GCTGGAATTACCGCGGCT	60 °C
Arg1	CTCCAAGCCAAAGTCCTTAGAG	AGGAGCTGTCATTAGGGACATC	61 °C
C3	TCCCAATGTCCTACGGCTG	ACGTACTTGTGCCCCTCCTT	60 °C
Ccl2	TTAAAAACCTGGATCGGAACCAA	GCATTAGCTTCAGATTTACGGGT	65 °C
Ccl5	GCTGCTTTGCCTACCTCTCC	TCGAGTGACAAACACGACTGC	65 °C
Cxcl10	CCAAGTGCTGCCGTCATTTTC	GGCTCGCAGGGATGATTTCAA	65 °C
Hprt	TCAGTCAACGGGGGACATAAA	GGGGCTGTACTGCTTAACCAG	65 °C
Iba-1	ATCAACAAGCAATTCCTCGATGA	CAGCATTCGCTTCAAGGACATA	64 °C
Il-1β	GCCCATCCTCTGTGACTCAT	AGGCCACAGGTATTTTGTCG	61 °C
Il-4	GGTCTCAACCCCCAGCTAGT	GCCGATGATCTCTCTCAAGTGAT	65 °C
iNos/Nos2	ACATCGACCCGTCCACAGTAT	CAGAGGGGTAGGCTTGTCTC	64 °C
Lcn2	GCAGGTGGTACGTTGTGGG	CTCTTGTAGCTCATAGATGGTGC	65 °C
Mrc1	GTGGTCCTCCTGATTGTGATAG	CACTTGTTCCTGGACTCAGATTA	65 °C
S100a10	ACCACTTGACAAAGGAGGACC	AAAGCTCTGGAAGCCCACTTT	60 °C
Serpina3n	AACCAGAGACCCTGAGGAAGT	AGTTTCGCAGACATTGGGACAA	60 °C
Sphk1	TATGCTGGGTACGAGCAGGT	CCCACTGTGAAACGAATCTCC	65 °C
Tnf-α	GCCATAGAACTGATGAGAGGGAG	GGTGCCTATGTCTCAGCCTCTT	62 °C

. Experiments were performed with at least 3 biological and 2 technical replicates if not stated otherwise.

2.5. Western Blot Analysis

BV-2 cells and primary astrocytes were detached using cold DMEM with 10% FCS and 0.5% Pen/Strep. After centrifugation at 300× g for 5 min at 4 °C, the pellet was washed two times with 10 mL ice-cold DPBS and resuspended in RIPA+. Incubation on ice for 30 min on a shaker with vortexing every 10 min was followed by a centrifugation step at 18,000× g for 30 min at 4 °C. For protein isolation of the EVs obtained from ultracentrifugation, pellets were resuspended directly in RIPA+ and vortexed without centrifugation. Protein concentrations were determined using a BCA Protein Assay Kit (Thermo Scientific) following the manufacturer’s instructions. The absorbance was measured at 562 nm with i-control 2.0 software using the Tecan infinity plate reader (Tecan, Männedorf, Switzerland). For the sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), 20 µg of protein for BV-2 cells and EVs, and 15 µg of protein for primary astrocytes were loaded and separated by 12% (v/v) gel electrophoresis. Proteins were transferred to PVDF membranes and blocked with 5% milk in Tris-buffered saline (TBS). The membranes were incubated for 24 h at 4 °C with the respective primary antibody, see

Table 2. Used primary and secondary antibodies with manufacturer, host, clonality and applied concentration. Rb = rabbit, AH = Armenian hamster, M = mouse, Gt = goat, (m) = monoclonal, (p) = polyclonal.

Antibody	Manufacturer	Host, Clonality	Concentration
Anti-CD9	Invitrogen, SA-35-08	Rb (m)	1:1000
Anti-CD81	Santa Cruz, Dallas, TX, USA, sc18877	AH (m)	1:500
Anti-GM130	Bioscience, San Diego, CA, USA, 610822	M (m)	1:1500
Anti-IL-1β	Abcam, ab9722	Rb (p)	1:500
Anti-LCN2/NGAL	Cloud clone, Katy, TX, USA, AB388Mu01	Rb (p)	1:500
Anti-β-Actin	Santa Cruz, sc47778	M (m)	1:5000
Anti-ALDH1L1	Abcam, ab87117	Rb (p)	1:600
Goat Anti-Rabbit IgG (H + L) Alexa Fluor 488	Invitrogen, A-11008	Gt (p)	1:500
Goat Anti-Mouse IgG	Sigma, A4416	Gt (p)	1:4000
Goat Anti-Rabbit IgG (H + L)-HRP	BIO-RAD, Hercules, CA, USA, 170-6515	Gt (p)	1:5000
Mouse Anti-arm-hamster IgG-HRP	Santa Cruz, sc2789	M (m)	1:5000

. After washing, the membranes were incubated with the appropriate secondary antibody for 2 h at room temperature. For visualization, the enhanced chemiluminescence method (Fusion SL, Vilber, Germany) was used according to the manufacturer’s protocol. For evaluation, densitometric quantification was performed. The intensity of the specific bands was measured and normalized to β-actin as a loading control and quantified using ImageJ [52].

2.6. Mass Spectrometry

Exosomal proteins were extracted using a lysis buffer containing 5% SDS, 0.1 M triethylammonium bicarbonate (TEAB), and 10 mM tris(2-carboxyethyl)phosphine (TCEP) at pH 8.5, followed by sonication (20 cycles, Bioruptor, Diagenode, 4 °C). Protein reduction was performed at 95 °C for 10 min, and cysteine alkylation was achieved with 20 mM iodoacetamide for 30 min at room temperature in the dark. A total of 100 μg of protein from each sample was purified and digested using S-Trap cartridges (Protifi, Fairport, NY, USA) according to the manufacturer’s protocol. The resulting peptides were vacuum-dried and resuspended in 0.1% formic acid for liquid chromatography–mass spectrometry (LC–MS/MS) analysis. Peptide separation was carried out using an Ultimate 3000 nano-LC system coupled to a Q Exactive HF mass spectrometer (Thermo Fisher Scientific) with a C18 reverse-phase column (75 μm × 30 cm, 1.9 μm ReproSil-Pur C18-AQ resin, Dr. Maisch GmbH) maintained at 60 °C. Peptides were eluted at 0.3 μL/min using a linear gradient from 2% to 95% acetonitrile (0.1% formic acid) over 72 min. Data were acquired in data-independent acquisition (DIA) mode with a total cycle time of ~3 s. MS1 scans were collected at a resolution of 120,000 (FWHM at 200 m/z) across 350–1600 m/z, with an AGC target of 3 × 10⁶ and a maximum injection time of 100 ms. MS2 scans were recorded with 18 m/z isolation windows, normalized HCD energy of 28%, and a resolution of 30,000 (FWHM at 200 m/z). Raw spectra were analyzed using Spectronaut v17.1 (PTM workflow) against the Mus musculus UniProt database (release 22 February 2022), assuming full tryptic digestion with up to three missed cleavages. Carbamidomethylation (C) was set as a fixed modification, and methionine oxidation and N-terminal acetylation were set as variable modifications.

2.7. Proteomics Analysis

Proteomic data were processed and filtered to ensure high quantitative accuracy and reproducibility. A Q-value threshold of 0.01 was applied during the first data import to control the false discovery rate (FDR). Non-proteotypic peptides and proteins identified by only a single feature (combination of peptide, precursor charge, and fragment ion) were excluded. To increase data robustness, only proteins supported by at least six cumulative features per experimental group were retained for downstream analysis. Pairwise comparisons with infinite fold-change values were considered only if fewer than 25% of samples were missing in the detected group. Data normalization was performed using the Equalize Medians method to correct for systematic intensity variations across runs. Statistical analyses, including differential abundance testing, were conducted using the MSstats R package (v4.7.3) within the R environment (v4.3.1) [53]. Only proteins meeting the established quality control (QC) criteria and exhibiting consistent quantification across biological replicates were included in the final dataset.

Finally, for precise evaluation a q-value threshold of 0.05 was considered to analyze the proteomics data, and proteins with Log2FC ≥ 1.5 or ≤−1.5 (uniquePeptideCount > 1) were selected. Subsequently, a Venn diagram was constructed for the up-and down-regulated proteins using the Draw Venn Diagram web tool (https://bioinformatics.psb.ugent.be/webtools/Venn/ (accessed on 31 March 2025)). KEGG pathway enrichment analysis was performed using Enrichr (https://maayanlab.cloud/Enrichr/ (accessed on 23 March 2025)) [54] for proteins uniquely upregulated in each group, with a cut-off p-value of >0.05 for the results from Enrichr. The protein–protein interaction (PPI) network for these unique proteins was constructed using STRING v12 (https://string-db.org/ (accessed on 25 March 2025)) [55]. Finally, the PPI networks were analyzed using Cytoscape v3.10.3 software.

2.8. Statistical Analysis

Statistical analysis was performed using GraphPad Prism 10. If not stated otherwise, the results are presented as arithmetic means ± SEM with n given in the figure legend. Data from RT-qPCR were tested for normal distribution using the Shapiro–Wilk test and, if so, analyzed using a two-way analysis of variance (ANOVA) with Tukey multiple comparisons test as a post hoc test. If the normal distribution could not be assumed, data were transformed by the Box-Cox transformation using an optimal λ. p < 0.05 was considered statistically significant. The following symbols were used to indicate the level of significance: * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

3. Results

3.1. BV-2-Derived Extracellular Vesicle Characterization and Uptake by Astrocytes

To explore intercellular communication via EVs between microglia and astrocytes under neuroinflammatory conditions, we utilized BV-2 cells, a murine microglial cell line known for EV production. Polarization of BV-2 cells towards an activated M1 phenotype using LPS treatment was verified by increased expression of M1 markers Nitric Oxide Synthase 2 (Nos2) and Tnf-α (Figure 1A,B). Ionized calcium binding adapter molecule 1 (Iba-1) was used as microglia marker and the mRNA expression did not differ between groups (Figure 1C), which may be attributable to the 24 h time point selected after stimulation. M2 markers Mannose receptor C-type 1 (Mrc1), Arginase 1 (Arg1), and Interleukin-4 (Il-4) were significantly decreased following stimulation with LPS (Figure 1D–F). BV-2-derived extracellular vesicles (BEVs) were collected through a two-step ultracentrifugation protocol, as detailed in the methods section, and analyzed using NTA. This process yielded two distinct vesicle populations: BV-2-derived microvesicles (BMVs), collected first, and BV-2-derived exosomes (BEXs), collected second. Those vesicle populations were also collected from LPS-activated BV-2 cells, resulting in activated BV-2-derived microvesicles (ABMVs) and activated BV-2-derived exosomes (ABEXs). No significant size differences were observed between these populations (Figure 2A–F). However, the concentration of vesicles was notably higher in the EXs groups compared to MV populations, whereas LPS stimulation (ABMVs and ABEXs) did not affect the concentration compared to the respective control population (BMVs and BEXs) (Figure 2C). Exosomal markers CD9 and CD81 were detected in protein extracts from BEXs, though CD81 was absent in BV-2 cell lysates (Figure 2G). The Golgi marker GM130 was not detected in EXs, ruling out cytosolic contamination. Additionally, EXs were visualized during NTA (Figure 2H). To evaluate the uptake of microglia-derived EVs by astrocytes, isolated BEVs were labelled with a fluorescent membrane dye and subsequently added to primary astrocyte cultures. Fluorescence intensity around astrocytes progressively increased during the first five hours of incubation, indicating active vesicle binding and internalization (Figure 3A). Confocal 3D reconstruction further confirmed the intracellular localization of labelled vesicles, as evidenced by the presence of red fluorescence within the astrocytic cytoplasm (Figure 3B,C), thereby validating effective EV internalization by recipient astrocytes.

3.2. Activated BV-2-Derived Exosomes Induce an Astrocyte Phenotype Shift

We next investigated the potential of microglia-derived EVs to influence astrocyte phenotype polarization under inflammatory conditions. Astrocytes are known to shift their activation state when stimulated, which can be characterized by distinct gene expression profiles. The reactive phenotype markers Complement Component 3 (C3) and serine protease inhibitor A3N (Serpina3n) were significantly upregulated following stimulation with LPS and treatment with ABEXs (Figure 4A,B). Notably, Serpina3n expression was also elevated in astrocytes treated with ABMVs. Interestingly, BEXs cells were also capable of inducing Serpina3n expression, suggesting that even in the absence of direct inflammatory priming, basal EX cargo can modulate astrocytic activation toward an A1-like profile.

As reactive phenotype markers described for anti-inflammatory states sphingosine kinase 1 (Sphk1) and S100 calcium-binding protein A10 (S100a10) were included. Treatment with BEXs and ABEXs significantly increased Sphk1 expression (Figure 4C), independent of LPS exposure. In contrast, S100a10 expression showed a downward trend in LPS-treated groups compared with controls; however, this change did not reach statistical significance (Figure 4D). Together, these findings indicate that ABEXs differentially modulate astrocytic gene expression, promoting features of reactive phenotypes depending on their origin and cargo composition.

3.3. Astrocytes Express Inflammatory Mediators in Response to Activated BV-2-Derived Exosome Stimulation

To further characterize the inflammatory response of astrocytes, we examined the expression of key pro-inflammatory mediators. Lipocalin-2 (LCN2), a marker of reactive astrocytes, showed significantly elevated mRNA levels following LPS stimulation (Figure 5A). Interestingly, BEXs were sufficient to induce Lcn2 expression in astrocytes even in the absence of LPS, and this effect was further amplified when astrocytes were treated with ABEXs. Supporting this, LCN2 protein could be detected after LPS, ABMV, BEX, and ABEX stimulation (Figure 5B). Similarly, interleukin-1 beta (Il1b) mRNA—normally synthesized as an inactive precursor and cleaved by the inflammasome complex into mature IL-1β—was strongly induced by LPS in astrocytes (Figure 5C). ABEXs also significantly upregulated Il1b expression. Western blot analysis confirmed the presence of pro-IL-1β in all samples, while both LPS and ABEXs promoted the appearance of the mature IL-1β form (Figure 5D), indicating functional activation of the inflammasome pathway. The chemokines C-C motif chemokine ligand 2 (Ccl2) and Ccl5 displayed a similar pattern, with robust induction following LPS treatment and a tendency toward higher expression in astrocytes treated with ABEXs (Figure 5E,F). C-X-C motif chemokine ligand 10 (Cxcl10) expression was also enhanced by LPS stimulation (Figure 5G). Notably, BEXs were sufficient to elevate Cxcl10 expression to levels comparable to ABEXs, which was significant compared to control (Figure 5G). In summary, these results demonstrate that LPS drives an inflammatory astrocytic phenotype, characterized by increased expression of Lcn2, Il1b, Ccl2, Ccl5, and Cxcl10. Importantly, inflammatory activation induced by ABMVs was less pronounced than that triggered by direct LPS stimulation. In contrast, ABEXs recapitulated the pro-inflammatory signature of LPS, while BEXs partially upregulated inflammation-associated genes, highlighting their intrinsic modulatory potential even under basal conditions.

3.4. Different Extracellular Vesicles Have Distinct Protein Cargo

To investigate the molecular basis of the differential effects of microglia-derived EVs on astrocyte activation, we subsequently conducted proteomic profiling of BMVs/ABMVs and BEXs/ABEXs (n = 4 for each group). By profiling the protein cargo of these vesicle subtypes, we sought to identify the molecular components and signaling pathways that regulate astrocytic polarization and inflammatory responses. Quantitative LC–MS/MS profiling yielded comparable proteome depth across samples, with ~5000–6000 protein identifications per sample before specific cutoffs (see Supplementary Material), and revealed vesicle subtype- and stimulation-dependent patterns of differential protein abundance. After applying our filtering criteria—using a q-value cutoff of 0.05 and selecting proteins with |log2FC| ≥ 1.5 supported by >1 unique peptide—approximately half of the identified proteins were excluded, and downstream analyses were subsequently performed on the remaining high-confidence dataset. This stringent approach is commonly employed in extracellular vesicle proteomics, where samples often contain a large proportion of fragmented, low-abundance, or non-functional protein cargo that likely represents nonspecific or “garbage-loaded” material rather than biologically active components. Our findings suggest that the EV proteome dynamically reflects the activation state of the parent microglia and may underlie their selective effects on recipient astrocytes (Figure 5A and Figure 6A). Comparative pathway analysis revealed that BEXs contained 220 downregulated proteins compared to BMVs (Figure 5A), primarily enriched in pathways related to endocytosis, protein processing in the endoplasmic reticulum, ribosome biogenesis, and metabolic processes (Figure 5B). Network analysis identified several key hub proteins–Hsp90aa1, Kras, Clpp, Mrps5, Grb2, and Itgb1–which displayed high interaction degrees and may represent central regulators of EV-mediated signaling (Figure 6D,E). These proteins are known to influence mitochondrial function, protein folding, and cellular stress responses. Conversely, in the upregulated proteome, 17 proteins were specifically upregulated in BEXs compared to BMVs (Figure 7A).

3.5. Proteomic Signatures Depleted in Activated Microglia-Derived Extracellular Vesicles

Across all comparisons in the downregulated proteome (Figure 6A), 634 proteins were shared among BEXs and ABEXs, representing a conserved core proteome likely essential for EXs biogenesis and basic cellular communication. Among these, Prohibitin 2 (PHB2) was uniquely detected across all subtypes except ABMVs, suggesting a pivotal role in maintaining mitochondrial integrity, protein folding, and vesicle assembly. Given PHB2’s known functions in mitochondrial dynamics and stress signaling, its presence across activated ABEVs supports its importance in sustaining glial cell homeostasis and intercellular signaling under stress conditions. In contrast, ABEXs exhibited 24 downregulated proteins compared with ABMVs, primarily associated with pyruvate metabolism, phagosome formation, and endocytosis (Figure 6C).

3.6. Proteomic Signatures Enriched in Activated Microglia-Derived Extracellular Vesicles

Proteomic enrichment analysis of upregulated proteins revealed distinct activation-dependent changes in the exosomal cargo. Across all comparisons within the downregulated proteome (Figure 6A), 405 proteins were commonly identified in both BEXs and ABEXs, representing a conserved exosomal core proteome that likely underpins exosome biogenesis, structural stability, and fundamental roles in intercellular communication. Compared with BEXs, ABEXs displayed 16 upregulated proteins, notably enriched in the NOD-like receptor signaling pathway (IFNAR2, CYBA), which is central to innate immune activation (Figure 7A,B). In contrast, 165 proteins were upregulated in ABEXs relative to ABMVs, predominantly involved in ribosome biogenesis, spliceosome assembly, RNA transport, and cell cycle regulation, indicating an enhanced capacity for protein synthesis and RNA processing (Figure 7C). Network topology analysis (Figure 7D,E) identified highly connected hub proteins–Rpf2, Gtpbp4, Lsg1, Pes1, and Wdr12–which are key regulators of ribosomal maturation and nucleolar function. Moreover, Chromatin Accessibility Complex Subunit 1 (Chrac1) was a unique overexpressing factor among ABEXs. Chrac1, a nucleosome-remodeling protein, facilitates chromatin accessibility and DNA repair, suggesting a potential role in transcriptional reprogramming or DNA damage response within EV-producing cells.

4. Discussion

This study aimed to elucidate the effects of ABEVs on astrocyte activation, a crucial component in neuroinflammatory processes associated with NDDs. The findings provide significant insights into the role of EVs in modulating the astrocyte phenotype, particularly their potential to induce a pro-inflammatory astrocyte state. The successful isolation and characterization of MVs and EXs from BV-2 cells align with existing studies [56,57]. Despite no significant differences in size and concentration between EXs and MVs observed through NTA, the functional analysis revealed that EXs exerted more pronounced effects on astrocyte activation compared to MVs.

The activation of astrocytes towards a reactive state, indicated by the significant upregulation of markers such as C3, Serpina3n, and Cxcl10, suggests that ABEXs might push astrocytes into a pro-inflammatory phenotype. This finding is consistent with previous research demonstrating the characteristics of reactive, A1-like astrocytes [23,58]. The presence of active IL-1b, a potent pro-inflammatory cytokine, exclusively in astrocytes treated with ABEXs, further supports the hypothesis that these vesicles might trigger inflammasome activation, contributing to a sustained inflammatory response in the CNS. The observed pro-inflammatory effects of EXs on astrocytes align with several studies that highlight the role of EVs, particularly EXs, in mediating neuroinflammation. For instance, Kumar et al. [32] and Chen et al. [59] demonstrated that ABEVs can induce a pro-inflammatory response in recipient cells, including microglia and astrocytes, further emphasizing the role of EVs in intercellular communication during inflammation. Interestingly, while BEXs showed less pronounced effects on astrocyte activation than ABEXs, they were not entirely inert, suggesting that even without an inflammatory stimulus, EXs might carry signals capable of modulating astrocyte function. This is contrary to some studies where BEXs did not elicit significant responses [57,59], indicating that the context and origin of EXs, as well as the experimental conditions, might influence their impact on recipient cells.

However, the results of this study underscore the complexity of astrocyte activation in the CNS and the pivotal role of microglia-derived EVs in this process. Moreover, the potential role of EXs in complement-mediated synapse elimination, as indicated by the upregulation of C3 in astrocytes, highlights a possible mechanism through which chronic neuroinflammation could lead to synaptic loss and neuronal dysfunction in diseases such as AD and MS [60,61]. The finding that ABEXs induce IL-1b activation further suggests that these EVs might contribute to the chronic activation of the inflammasome pathway, exacerbating neuroinflammatory conditions. These data suggest that LPS activation drives a substantial remodeling of exosomal proteomes, favoring pathways associated with innate immune signaling and translational control, thereby supporting a specialized role of exosomes in microglia-mediated neuroinflammatory communication.

Neuroinflammation represents a tightly regulated but double-edged process within the central nervous system (CNS), where glial cells–particularly microglia and astrocytes–play pivotal roles in both neuroprotection and neurodegeneration. In recent years, EVs have emerged as crucial mediators of glia-to-glia communication, transporting diverse bioactive molecules that shape the neuroinflammatory microenvironment [33,62,63]. Our study demonstrates that ABEXs can induce a pro-inflammatory astrocytic phenotype and stimulate cytokines such as LCN2 and IL-1β, even in the absence of direct inflammatory stimuli. These findings are consistent with previous reports showing that microglial activation leads to the release of EVs enriched in immune-related proteins and microRNAs that propagate inflammatory signaling to surrounding neural cells [7,40].

Proteomic profiling of BEVs revealed marked differences in protein composition between microvesicles (MVs) and exosomes (EXs), as well as between control and LPS-stimulated conditions. Comparative analysis showed that in control (unstimulated) cells, BEXs contained 220 downregulated proteins—predominantly involved in endocytosis and protein synthesis in the endoplasmic reticulum (ER)—and 17 upregulated proteins enriched in NOD-like receptor signaling pathways compared with MVs. Upon LPS stimulation, the protein landscape shifted substantially: 24 proteins were downregulated in exosomes (mainly associated with endocytosis and pyruvate metabolism), whereas 165 proteins were upregulated, largely linked to ribosome biogenesis and spliceosome assembly. These results suggest that microglial activation dynamically alters the protein cargo of EVs, reshaping their functional potential to modulate recipient cells. The enrichment of ribosomal and spliceosomal components in ABEXs indicates increased involvement in post-transcriptional and translational regulation, while the upregulation of proteins related to NOD-like receptor signaling suggests a role in innate immune activation. Collectively, these proteomic alterations support the concept that BEXs act as specialized mediators of neuroinflammatory signaling, transmitting activation-specific molecular signatures that can influence astrocyte responses.

Beyond descriptive findings, our results hold broader biological and therapeutic implications. The enrichment of translation- and inflammation-related proteins in ABEXs implies that microglial activation states may propagate through vesicle-mediated signaling, amplifying astrocytic responses during neuroinflammation. This supports the concept that EVs are active participants in CNS immune communication rather than by-products of cellular activity [57]. Therapeutically, targeting exosome biogenesis, cargo sorting, or uptake represents a promising avenue for attenuating maladaptive glial activation and chronic neuroinflammation. Strategies that disrupt microglial EX formation or prevent their internalization by astrocytes may effectively decouple upstream microglial activation from downstream astrocytic reactivity, mitigating glia-driven neurotoxicity.

5. Study Limitations

Despite these insights, several limitations should be acknowledged. Our in vitro approach, relying on BV-2 microglial cells and primary astrocyte cultures, cannot fully replicate the in vivo microenvironment. Future studies using animal models of neuroinflammation or neurodegeneration (e.g., LPS injection or APP/PS1 mice) are needed to validate these findings under physiological conditions. Furthermore, the classification of astrocytes into A1 and A2 subtypes remains a topic of debate. As suggested by Escartin et al. [22], a more nuanced understanding of astrocyte states beyond the binary A1/A2 classification is necessary, especially in the context of chronic neurodegeneration where astrocytes exhibit diverse and context-dependent phenotypes. Additionally, while mass spectrometry identified distinct protein cargo profiles, the potential contribution of non-protein components—particularly microRNAs and circular RNAs—remains to be clarified. Functional experiments involving selective cargo inhibition or gene silencing will be instrumental in establishing causal relationships between exosomal components and astrocytic outcomes. Investigating the temporal dynamics of EV release and uptake under varying inflammatory conditions, or how external modulators such as aging, oxidative stress, or pharmacological interventions reshape EV signaling, could provide crucial mechanistic insights.

6. Conclusions

Together, these findings highlight that microglial activation markedly alters the proteomic composition of exosomes versus microvesicles, suggesting subtype-specific functional specialization and supporting their differential roles in mediating glia-to-glia communication during neuroinflammation. In conclusion, our findings establish that ABEXs selectively modulate astrocytic polarization and cytokine expression through their protein cargo, emphasizing their role in glia-to-glia communication during neuroinflammation. These results advance our understanding of EV-mediated neuroimmune crosstalk and highlight microglial exosomes as promising targets for therapeutic modulation in neurodegenerative and neuroinflammatory diseases.