cRGD-functionalized exosome–Mn₃O₄ nanoplatform enhanced ischemic lesion accumulation and ros scavenging-mediated neuroprotection in ischemic stroke

Abstract

The therapeutic efficacy of ischemic stroke (IS) treatment is severely limited by insufficient accumulation of therapeutic agents within ischemic lesions and persistent secondary injury after ischemia–reperfusion. Herein, we report a cRGD-functionalized exosome-based nanoplatform that enhances ischemic lesion-associated accumulation and antioxidative neuroprotection for the treatment of IS. Neural stem cell-derived exosomes were functionalized with cyclic RGD peptides (cRGD) and subsequently loaded with Mn₃O₄ nanoparticles to construct a hybrid nanosystem (cRGD-Exo@Mn₃O₄). The engineered exosomes preserve intrinsic brain tropism, while cRGD modification promotes preferential accumulation in ischemic regions, potentially through interaction with α_vβ₃ integrin that is upregulated in ischemic lesions. The incorporated Mn₃O₄ nanoparticles confer robust reactive oxygen species (ROS) scavenging capability, thereby mitigating oxidative stress in ischemic microenvironments. In vitro and in vivo studies demonstrate that cRGD-Exo@Mn₃O₄ exhibits enhanced accumulation in ischemic regions compared with non-modified counterparts. The nanosystem effectively attenuates oxidative stress and neuroinflammation, leading to reduced infarct volume, alleviation of cerebral edema, and improved neurological function in MCAO/R mice. Mechanistically, transcriptomic analysis suggests that the therapeutic effects are associated with modulation of inflammation-and cell death-related pathways, including suppression of the RIPK1/RIPK3/MLKL signaling cascade. Collectively, this study presents a rationally designed exosome-based nanoplatform integrating ischemic lesion-associated accumulation with ROS-scavenging capability.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04922-0.

Introduction

Ischemic stroke (IS), accounting for more than 80% of all stroke cases, is a leading cause of mortality and long-term neurological disability worldwide [1]. It is characterized by sudden interruption of cerebral blood flow, resulting in oxygen deprivation, metabolic dysfunction, and irreversible neuronal injury [2]. Despite advances in clinical management, including intravenous thrombolysis with recombinant tissue plasminogen activator (rt-PA, alteplase) or tenecteplase and mechanical thrombectomy, therapeutic outcomes remain limited due to the narrow therapeutic window, risk of hemorrhagic transformation, incomplete microvascular reperfusion, and subsequent ischemia–reperfusion injury [3, 4]. Therefore, current therapeutic strategies primarily focus on rapid restoration of blood flow, while effective approaches for mitigating secondary injury after reperfusion remain insufficient. In particular, the efficient delivery and accumulation of neuroprotective agents within ischemic lesions remain challenging due to the complex brain microenvironment and the presence of the blood–brain barrier (BBB) under physiological conditions [5, 6]. Thus, the development of targeted nanotherapeutic systems capable of improving delivery to ischemic regions and alleviating secondary pathological cascades represents a promising strategy to complement current reperfusion therapies.

Oxidative stress is a central contributor to ischemia–reperfusion injury. Excessive production of reactive oxygen species (ROS), including superoxide anions (•O₂⁻), hydrogen peroxide (H₂O₂), and hydroxyl radicals (•OH), induces extensive oxidative damage to lipids, proteins, and nucleic acids, while simultaneously amplifying inflammatory responses and neuronal death pathways [7–9]. Therefore, targeted scavenging of ROS within the ischemic microenvironment has emerged as an attractive therapeutic strategy [10]. However, conventional antioxidant agents often suffer from poor stability, rapid degradation, and limited delivery efficiency, which restrict their therapeutic potential [11–13]. Nanozymes with intrinsic enzyme-like catalytic activities have recently attracted considerable attention as effective antioxidant platforms [14, 15]. Among them, Mn₃O₄ nanoparticles possess abundant Mn²⁺/Mn³⁺ redox couples, enabling multiple enzyme-mimetic activities, including superoxide dismutase (SOD)-like, catalase (CAT)-like, and glutathione peroxidase (GPx)-like activities, thereby providing broad-spectrum ROS scavenging capability [16, 17]. Previous studies have demonstrated that Mn₃O₄-based nanomaterials can effectively alleviate oxidative stress-associated tissue damage and promote regeneration [18, 19]. However, the clinical translation of naked Mn₃O₄ nanoparticles is limited by poor aqueous stability, rapid systemic clearance, and insufficient accumulation at pathological sites.

Exosomes are nanoscale extracellular vesicles (30–150 nm) that transport bioactive cargos, including proteins, nucleic acids, and lipids, and play essential roles in intercellular communication [20]. Owing to their excellent biocompatibility, low immunogenicity, and intrinsic biological properties, exosomes have emerged as promising drug delivery vehicles for neurological disease therapy. Notably, human induced neural progenitor cells exosomes (hiNPCs-Exos) possess inherent neuroprotective and immunomodulatory functions, carrying bioactive molecules involved in neuronal survival, angiogenesis, and tissue repair [21]. Previous studies have demonstrated that hiNPCs-Exos exhibit enhanced accumulation in injured brain regions and exert beneficial effects against ischemic injury through anti-inflammatory, neuroprotective, and neurogenic mechanisms [22–24]. However, the therapeutic efficacy of endogenous exosomes remains limited by insufficient lesion-specific accumulation and relatively weak antioxidative capacity, highlighting the need for rational engineering strategies to enhance their therapeutic performance.

To improve the targeting capability of exosome-based delivery systems, various surface modification strategies have been developed. Among them, cyclic RGD peptides (cRGD) have attracted considerable attention due to their high affinity toward α_vβ₃ integrin, which is markedly upregulated in ischemic brain lesions but minimally expressed in normal brain tissue. In the ischemic microenvironment, α_vβ₃ integrin is highly expressed in vascular endothelial cells, activated astrocytes, and M1-type microglia within IS lesions, making it an attractive target for lesion-specific delivery [25, 26]. Previous studies have demonstrated that cRGD modification can enhance nanoparticle accumulation in ischemic regions, likely through interactions with α_vβ₃ integrin-expressing cells, thereby improving therapeutic outcomes [27]. Therefore, integrating cRGD-mediated targeting with exosome-based delivery may provide an effective strategy for improving the accumulation of therapeutic cargos at ischemic lesions.

Although cRGD-engineered exosomes, exosome-based therapeutics, and Mn₃O₄ nanozymes have each shown promise for IS treatment, previous studies have generally focused on improving either delivery efficiency or antioxidant therapy independently. A multifunctional platform that integrates ischemic lesion-associated accumulation, endogenous exosome-mediated neuroprotection, and nanozyme-based ROS scavenging within a single system remains largely unexplored. In this study, we hypothesized that engineering hiNPCs-derived exosomes with cRGD peptides and loading them with Mn₃O₄ nanozymes could synergistically enhance ischemic lesion-associated accumulation, improve antioxidant capacity, and alleviate secondary injury after IS. To test this hypothesis, we constructed a cRGD-functionalized exosome-based nanoplatform (cRGD-Exo@Mn₃O₄) by conjugating cRGD peptides onto hiNPCs-derived exosomes and incorporating Mn₃O₄ nanoparticles through membrane fusion (Scheme 1). This multifunctional platform integrates the intrinsic neuroprotective properties and biocompatibility of hiNPCs-derived exosomes, the enhanced ischemic lesion-associated accumulation conferred by cRGD through interactions with α_vβ₃ integrin-expressing cells, and the broad-spectrum ROS-scavenging activity of Mn₃O₄ nanozymes. The delivery behavior, therapeutic efficacy, and underlying molecular mechanisms of cRGD-Exo@Mn₃O₄ were systematically evaluated using in vitro models and a middle cerebral artery occlusion/reperfusion (MCAO/R) mouse model. Our results demonstrate that this engineered nanoplatform enhances accumulation in ischemic brain regions, attenuates oxidative stress and neuroinflammation, improves neurological recovery, and suppresses RIPK1/RIPK3/MLKL-mediated necroptosis following IS.

Scheme 1.

Schematic illustration of the fabrication of cRGD-Exo@Mn₃O₄ and its proposed mechanism for enhanced delivery to ischemic brain tissue, ROS scavenging, inflammation inhibition, and neuroprotection following MCAO/R

Materials and methods

Reagents

Manganese acetate tetrahydrate, oleic acid, and n-hexane were purchased from Aladdin Reagents (China); oleamide and 1-octadecene were purchased from McLean Reagents (China); DSPE-PEG2000, DBCO-NHS, c(RGDyK)-N₃, and Cy5.5 were purchased from MedChemExpress (USA); Dialysis bags (7000 Da) were purchased from Solabio (China); human induced neural progenitor cells (hiNPCs) were generated through in vitro directed differentiation of human peripheral blood mononuclear cells (PBMCs). The sample collection and stem cell induction protocols were approved by the Ethics Review Committee (Approval No. Ethics-KY-GZR-2022-082); DMEM/F12, Neurobasal, GlutaMAX, B-27, N-2, LIF, and penicillin/streptomycin antibiotic mixture were purchased from Thermo Fisher (USA); SB-431,542 and Laduviglusib were purchased from MedChemExpress (USA); Matrigel matrix was purchased from Corning (USA); ACCUTASE cell digestion solution was purchased from StemCell (Canada); TSG101, CD63, and Calnexin antibodies were purchased from Abcam (UK); RIPK1, p-RIPK1, RIPK3, p-RIPK3, MLKL, p-MLKL, and GAPDH antibodies were purchased from Cell Signaling Technology (USA), Zenbio (China), and Proteintech (China); H₂O₂ and LPS were purchased from Aladdin Reagents (China) and Sigma-Aldrich (USA); DiO, DiR, Hoechst 33,342, dihydroethidium (DHE) fluorescent probes, and live/dead cell staining kits were purchased from Beyotime (China); The CCK-8 assay kit was purchased from Meilun Biotechnology (China); the mouse TNF-α, IL-1β, and IL-6 ELISA kits were purchased from Meimian (China); all other reagents used in the experiments were of analytical grade.

Cell culture

Human induced neural progenitor cells (hiNPCs) were generated from peripheral blood mononuclear cells (PBMCs) isolated from healthy human donors through reprogramming and directed differentiation. These cells were cultured in a specialized neural stem cell medium consisting of the following components: 24 mL DMEM/F12, 24 mL Neurobasal, 500 µL GlutaMAX, 50 µL human leukemia inhibitory factor (LIF), 500 µL N-2 Supplement, 500 µL B-27 Supplement, 10 µL SB-431542, and 15 µL Laduviglusib. The culture process was conducted in T25 flasks pre-coated with diluted Matrigel matrix and incubated at 37 °C in 5% CO₂. To verify successful hiNPCs induction, immunofluorescence staining was performed to detect hiNPCs marker expression. As shown in Fig. S1, the induced cells exhibited strong positive expression of Nestin, SOX2, and PAX6, which are widely recognized markers of neural stem/progenitor cells, confirming the successful induction of PBMC-derived hiNPCs.

The murine brain microvascular endothelial cell line (bEnd.3), hippocampal neuronal cell line (HT22), and microglial cell line (BV2) were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. All cells were maintained in a humidified incubator at 37 °C with 5% CO₂. Cells were subcultured upon reaching 80%–90% confluence, and those in the logarithmic growth phase were utilized for subsequent experiments.

Experimental animals

Male C57BL/6J mice (7–8 weeks old, 20–25 g) were obtained from the Laboratory Animal Center of Guangxi Medical University. The animals were housed under specific pathogen-free (SPF) conditions at a controlled temperature of 25 ± 2 °C and a relative humidity of 50%–60% under a 12 h light/dark cycle. All mice had ad libitum access to standard food and water. All animal experimental procedures were reviewed and approved by the Laboratory Animal Ethics Committee of Guangxi Medical University (Approval No. 202504009) and were performed in strict accordance with the guidelines for the care and use of laboratory animals.

Synthesis of cRGD-Exo@Mn₃O₄

Mn₃O₄ nanoparticles were synthesized using a previously reported thermal decomposition method with slight modifications [18]. Manganese acetate tetrahydrate (0.49 g) was mixed with oleic acid (1.3 mL), oleamide (7.4 mL), and 1-octadecene (30 mL), followed by stirring at room temperature for 4 h. Subsequently, 0.38 mL of a 0.05 M aqueous manganese acetate solution was added dropwise before the reaction mixture was gradually heated to 100 °C. After reaction at 100 °C for 10 min, the mixture was naturally cooled to room temperature. The reaction mixture was precipitated with 300 mL of anhydrous ethanol, and the resulting precipitate was collected and washed several times with anhydrous ethanol/n-hexane to obtain brownish-black Mn₃O₄ nanoparticles. To improve the aqueous dispersibility and colloidal stability of the hydrophobic Mn₃O₄ nanoparticles, surface modification with DSPE-PEG₂₀₀₀ was performed according to a previously reported protocol with minor modifications [28]. Mn₃O₄ nanoparticles (4 mg) were dispersed in 3 mL of chloroform by sonication for 10 min, while DSPE-PEG₂₀₀₀ (10 mg) was dissolved in 3 mL of chloroform. Equal volumes of the two solutions were mixed and magnetically stirred at room temperature for 2 h. The organic solvent was then evaporated under a nitrogen stream to form a thin film, which was subsequently hydrated with 7 mL of deionized water by sonication. The resulting suspension was purified using a 7 kDa dialysis membrane for 3 days to obtain PEGylated Mn₃O₄ nanoparticles.

The culture supernatant of hiNPCs was collected under sterile conditions and stored at − 80 °C. Before exosome isolation, the supernatant was thawed at 4 °C. Exosomes were isolated by differential ultracentrifugation according to our previously reported method [24]. Briefly, the supernatant was centrifuged at 600 g for 10 min at 4 °C to remove cells and debris, then centrifuged at 10,000 g for 40 min to remove large vesicles; subsequently, centrifuged at 100,000 g for 1 h in an ultracentrifuge to precipitate exosomes, discard the supernatant, and resuspend the exosome pellet in PBS. Surface functionalization of exosomes with cRGD peptides was achieved via copper-free click chemistry according to a previously reported protocol [29]. Exosomes were diluted in PBS to a protein concentration of 1 mg/mL, followed by the addition of DBCO-NHS dissolved in DMSO to a final concentration of 6 µM. The mixture was gently rotated at room temperature for 4 h and then purified three times using a 100 kDa ultrafiltration tube to remove unreacted DBCO-NHS, yielding DBCO-modified exosomes (DBCO-Exo). Subsequently, c(RGDyK)-N₃ was added to the DBCO-Exo suspension at a final concentration of 0.6 µM and incubated with gentle shaking at 4 °C for 12 h. Unbound peptides were removed by ultracentrifugation at 100,000 g for 1 h, and the resulting pellet was resuspended in PBS to obtain cRGD-modified exosomes (cRGD-Exo). Membrane fusion between cRGD-Exo (or Exo) and PEGylated Mn₃O₄ nanoparticles was performed using a previously reported freeze–thaw method with slight modifications [30]. Briefly, cRGD-Exo (or Exo) and Mn₃O₄ nanoparticles were mixed at a mass ratio of 5:2, frozen at − 80 °C for 15 min, and subsequently thawed at 37 °C for 15 min. This freeze–thaw cycle was repeated three times to obtain cRGD-Exo@Mn₃O₄ (or Exo@Mn₃O₄). The obtained cRGD-Exo@Mn₃O₄ (or Exo@Mn₃O₄) was finally dispersed in sterile PBS to form a homogeneous aqueous suspension for subsequent physicochemical characterization, in vitro experiments, and intravenous administration in vivo. Unless otherwise specified, the suspension was stored at 4 °C and used within one week.

Structural characterization of Mn₃O₄ nanoparticles

The crystalline structure of Mn₃O₄ nanoparticles was characterized using X-ray diffraction (XRD) [31]. Dried powder samples were spread evenly on a glass slide, and the resulting diffraction patterns were compared with the Mn₃O₄ standard PDF card (#24–0734). X-ray photoelectron spectroscopy (XPS) was used to determine the chemical valence state of Mn on the surface of Mn₃O₄ nanoparticles. The samples were mounted in conductive adhesive and analyzed under ultra-high vacuum conditions; the spectra were energy-corrected, and peak fitting was performed using Advantage software.

Morphological characterization

The morphology of Mn₃O₄ nanoparticles, Exo, cRGD-Exo, and cRGD-Exo@Mn₃O₄ was observed by transmission electron microscopy (TEM) according to a previously reported procedure. Samples were deposited onto carbon-coated copper grids and allowed to stand for 1–2 min before excess liquid was removed using filter paper. Exosome-containing samples were negatively stained with 2% phosphotungstic acid for 1 min, air-dried, and imaged by TEM.

Physicochemical characterization

A Malvern particle size and zeta potential analyzer was used to measure the hydrodynamic diameter, polydispersity index (PDI), and zeta potential of each sample. Samples were appropriately diluted with PBS, thoroughly mixed, and transferred to cuvettes or folded capillary electrophoresis chambers for dynamic light scattering (DLS) and zeta potential measurements at 25 °C.

Verification of cRGD modification

To verify successful surface modification of exosomes, c(RGDyK)-N₃ peptides were labeled with Cy5.5, whereas the exosome membrane was labeled with DiO. Fluorescence overlap analysis was performed following a previously reported method. Briefly, DBCO-Exo was reacted with Cy5.5-labeled c(RGDyK)-N₃, and excess dye was removed by ultracentrifugation. Subsequently, DiO (1 µM) was added and incubated at 37 °C for 20 min to label the exosome membrane. After washing, samples were mounted on glass slides and observed under a fluorescence microscope to evaluate fluorescence colocalization.

Western blotting was performed to verify the biochemical identity and purity of the isolated vesicles by probing for canonical exosomal marker proteins (CD63 and TSG101) and ruling out cellular contaminants.

Determination of encapsulation efficiency and Mn₃O₄ loading capacity

The Mn₃O₄ content in cRGD-Exo@Mn₃O₄ was quantified by inductively coupled plasma optical emission spectrometry (ICP-OES). Free Mn₃O₄ nanoparticles were removed by ultrafiltration centrifugation. The collected fractions were digested with concentrated acid and diluted to a final volume of 10 mL using ultrapure water. After instrument calibration and parameter optimization, a standard calibration curve was established using manganese standard solutions. The amount of Mn₃O₄ was calculated based on the measured manganese (Mn) content. The encapsulation efficiency (EE) and loading capacity (LC) of Mn₃O₄ were subsequently calculated according to the Mn₃O₄ content determined by ICP-OES in combination with the total exosomal protein content quantified using the bicinchoninic acid (BCA) protein assay. Determination of encapsulation efficiency and Mn₃O₄ loading capacity was calculated using the following formula: EE (%) = (encapsulated Mn₃O₄ / total Mn₃O₄added) × 100% and LC (%) = (loaded Mn₃O₄ / total exosomal protein) × 100%.

ROS scavenging activity

The scavenging activities of cRGD-Exo@Mn₃O₄ against three major reactive oxygen species (ROS), namely superoxide radical (•O₂⁻), hydrogen peroxide (H₂O₂) and hydroxyl radical (•OH), were determined [32]. All experiments were carried out following the manufacturer’s instructions for the corresponding assay kits. The scavenging activity against •O₂⁻ was tested using a superoxide dismutase activity assay kit (Solarbio, China). The H₂O₂-scavenging activity was measured using a catalase activity assay kit (Solarbio, China). The scavenging capacity of •OH was detected by a hydroxyl radical assay kit (Nanjing Jiancheng Bioengineering Institute, China).

In vitro cytotoxicity assay

Before therapeutic evaluations, the cytocompatibility profile of cRGD-Exo@Mn₃O₄ was assessed using a Cell Counting Kit-8 (CCK-8) assay. bEnd.3, HT22, and BV2 cells were selected as representative in vitro models of the neurovascular unit. Cells were seeded in 96-well plates (5 × 10³ cells /well) and allowed to adhere for 24 h. The culture medium was replaced with complete medium containing graded concentrations of cRGD-Exo@Mn₃O₄ (10, 20, 40, 60, 80, and 100 µg/mL). After 24 h of incubation, 10 µL of CCK-8 reagent was added to each well, and cells were incubated in the dark at 37 °C for 1 h. The absorbance at 450 nm was measured via a microplate reader to calculate relative cell viability.

In vitro uptake by bEnd.3 cells

An oxygen-glucose deprivation/reoxygenation (OGD/R) model was established in bEnd.3 cells to evaluate the effect of cRGD modification on cellular uptake. bEnd.3 cells were seeded at 1 × 10⁴ cells per well in a 48-well plate. After overnight culture, the cells were washed three times with sugar-free DMEM medium, then resuspended in fresh sugar-free medium and incubated in a hypoxic incubator (1% O₂, 5% CO₂, 94% N₂) for 2 h (OGD treatment); cells in the control group were cultured in a standard incubator. After OGD, the medium was replaced with complete medium containing 100 µg/mL DiO-labeled Exo@Mn₃O₄ or cRGD-Exo@Mn₃O₄, and the cells were incubated in a standard incubator for 4 h. Cells were washed three times with PBS, incubated with 25 µM Hoechst 33,342 in the dark for 15 min, and washed again with PBS, observe and capture images under a fluorescence microscope, and quantify the fluorescence intensity using ImageJ software.

In vitro transendothelial delivery assay

An in vitro endothelial–neuronal co-culture model was established using a Transwell system to evaluate the transendothelial delivery capability of the engineered exosomes. bEnd.3 cells were seeded at a density of 1 × 10⁵ cells per well in the upper chamber of a 24-well Transwell plate (pore size 0.4 μm) and cultured until a tight monolayer was formed. HT22 cells were seeded in the lower chamber (3 × 10⁴ cells/well) and cultured overnight. Before the experiment, cells were subjected to 4 h of OGD treatment. Subsequently, 100 µg/mL of DiO-labeled Exo@Mn₃O₄ or cRGD-Exo@Mn₃O₄ was added to the upper chamber, and the system was incubated for 12–24 h in a standard incubator. After incubation, collect the HT22 cells from the lower chamber, wash them three times with PBS, stain the cell nuclei with Hoechst 33,342, and observe and quantify the fluorescence intensity within the HT22 cells under a fluorescence microscope. The fluorescence intensity of HT22 cells was quantified to evaluate the relative transendothelial delivery efficiency of different formulations in the endothelial–neuronal co-culture model.

In vitro uptake by HT22 cells

HT22 cells were seeded at 8 × 10³ cells per well in a 48-well plate and cultured overnight; 100 µg/mL of DiO-labeled cRGD-Exo@Mn₃O₄ was added, and the cells were incubated for 6, 12, and 24 h, respectively. After washing with PBS, cell nuclei were stained with Hoechst 33,342, and cell uptake was observed under a fluorescence microscope. Fluorescence intensity was quantified using ImageJ software.

Assessment of ROS scavenging capacity

An H₂O₂-induced oxidative stress model was established in HT22 cells. HT22 cells were seeded in 96-well or 48-well plates and cultured for 24 h. Except for the normal control group, the other groups were treated with 800 µM H₂O₂ for 2 h; Subsequently, Mn₃O₄, Exo, Exo@Mn₃O₄, and cRGD-Exo@Mn₃O₄ (exosome protein concentration 100 µg/mL) were added, respectively, and the cells were incubated for an additional 24 h. Cell viability was assessed using the CCK-8 assay; cells were incubated with 10 µM DHE fluorescent probe for 30 min, washed with PBS, and then examined under a fluorescence microscope to observe intracellular ROS levels; fluorescence intensity was quantified using ImageJ software.

In vitro detection of inflammatory cytokines in microglia

To model stroke-induced neuroinflammation and evaluate the capacity of the nanoplatform to modulate microglial polarization, a lipopolysaccharide (LPS)-induced inflammatory model was utilized in BV2 cells. BV2 microglia were seeded in 48-well plates at a density of 2 × 10⁴cells per well and cultured overnight. To trigger a pro-inflammatory response, cells were stimulated with 1 µg/mL LPS and simultaneously treated with Exo, Mn₃O₄, Exo@Mn₃O₄, or cRGD-Exo@Mn₃O₄ for 24 h. The cell culture supernatant was collected, and the concentrations of tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), and interleukin-6 (IL-6) in the supernatant were measured using commercial ELISA kits in accordance with the manufacturer’s protocols.

Live/dead cell staining

To further corroborate the neuroprotective potency of the hybrid nanoplatform under ischemia-reperfusion-like conditions, a Calcein-AM/Propidium Iodide (PI) live/dead cell staining assay was performed on an OGD/R neuron model. An OGD/R injury model was established in HT22 cells. HT22 cells were seeded in 48-well plates. After 24 h of culture, the medium was discarded, and the cells were washed three times with glucose-free DMEM. The medium was then replaced with sugar-free DMEM, and the plates were incubated in an anaerobic incubator (1% O₂, 5% CO₂,94% N₂) for 4 h. Upon completion of OGD, the medium was replaced with complete medium containing Exo, Mn₃O₄, Exo@Mn₃O₄, or cRGD-Exo@Mn₃O₄, and the cells were incubated for an additional 24 h in a standard incubator. The cells were then stained using a Calcein-AM/PI cell viability kit. Live cells (green fluorescence) and dead cells (red fluorescence) were imaged via fluorescence microscopy and counted using ImageJ software to determine the cell survival rate.

Establishment of the mouse MCAO/R model

An MCAO/R model was established in male C57BL/6J mice (7–8 weeks old, 20–25 g) using the transient intraluminal filament occlusion method, which closely recapitulates clinical embolic stroke and spontaneous or therapeutic recanalization. Briefly, mice were anesthetized via an intraperitoneal injection of sodium pentobarbital and maintained in a supine position on a thermostatically controlled heating pad to regulate body temperature. The right common carotid artery, external carotid artery, and internal carotid artery were carefully exposed through a midline cervical incision. A poly-L-lysine-coated nylon monofilament was gently inserted through the internal carotid artery to occlude the origin of the middle cerebral artery. After 2 h of transient ischemia, the monofilament was carefully withdrawn to initiate reperfusion. Mice in the sham group underwent identical surgical procedures without monofilament insertion. Therapeutic formulations were administered via tail vein injection at a standardized dose of 100 µg exosomal protein per mouse at 1 h post-modeling.

In vivo brain accumulation analysis

Exosomes were labeled with the DiR fluorescent dye by incubating 1 mg of exosomes with 2 µM DiR at 37 °C for 20 min, followed by ultracentrifugation to remove unbound dye. One hour after MCAO/R reperfusion in mice, 100 µg of DiR-labeled Exo, Exo@Mn₃O₄, and cRGD-Exo@Mn₃O₄ were administered via the tail vein. Twelve hours after administration, the anesthetized mice were euthanized, and their brains were harvested. In vivo fluorescence imaging of the ex vivo brain tissue was performed using a small-animal in vivo imaging system to analyze fluorescence distribution and intensity. Fluorescence intensity in the ischemic and contralateral hemispheres was quantified to evaluate the relative accumulation of different formulations in ischemic brain tissue.

Twenty-four hours after MCAO/R reperfusion, anesthetized mice were euthanized, and their brains were removed. The brains were frozen at − 20 °C for 20 min and sectioned into 2-mm-thick continuous coronal slices. The brain sections were placed in a 2% TTC solution and incubated at 37 °C in the dark for 15 min, followed by fixation with 4% paraformaldehyde. Normal brain tissue appears red, while the infarct area appears pale. After photographing the sections, the ImageJ software was used to measure the infarct volume and calculate the percentage of infarct volume.

Water contents in the brain

The wet-to-dry weight method was used to determine the water content of brain tissue. Twenty-four hours after MCAO/R reperfusion, the mice were euthanized, and their brains were removed. Surface moisture was blotted with filter paper, and the wet weight was immediately recorded. The brain tissue was then placed in an oven at 100 °C and dried for 24 h until constant weight was achieved, at which point the dry weight was recorded. Brain tissue water content was calculated using the following formula: Water content (%) = (Wet weight − Dry weight) / Wet weight × 100%.

Histological staining

24 h after reperfusion, the mice were euthanized, and their brains were removed. The brains were fixed in 4% paraformaldehyde for 24 h, dehydrated through a graded series of ethanol, cleared in xylene, embedded in paraffin, and sectioned into 4-µm-thick coronal sections. H&E staining and Nissl staining were performed separately. Pathological morphological changes, neuronal survival status, and Nissl body changes in the cerebral cortex and hippocampus were observed under a light microscope.

Detection of inflammatory cytokines in brain tissue

Brain tissue was harvested from the right ischemic hemisphere of mice. RIPA lysis buffer containing protease inhibitors was added at a 1:10 (w/v) ratio, and the sample was homogenized on ice. The mixture was centrifuged at 4 °C and 12,000 rpm for 15 min, and the supernatant was collected. An ELISA kit was used to detect the expression levels of TNF-α, IL-1β, and IL-6 in the brain tissue.

Neurological severity score

Using the Zea-Longa scoring system, a researcher unaware of the group assignments scored the mice for neurological deficits. The scoring criteria were as follows: 0 points, no neurological deficits; 1 point, inability to fully extend the forepaw on the paralyzed side; 2 points, circling toward the paralyzed side while walking; 3 points, leaning toward the paralyzed side while walking; 4 points, inability to walk spontaneously and loss of consciousness; 5 points, death.

Open field test

Before the test, the mice were transferred to the behavioral laboratory to acclimate to the environment for 30 min. The mice were then gently placed in the fixed area of the open-field chamber and allowed to move freely for 10 min. A video tracking system recorded the mice’s movement trajectories, and the total distance traveled, average speed, distance traveled in the central area, and time spent in the central area were analyzed. The experimental apparatus was thoroughly cleaned with 75% ethanol before and after the test to eliminate odor interference.

Transcriptome sequencing

Brain tissue was collected from the right ischemic hemisphere of mice in the PBS group and the cRGD-Exo@Mn₃O₄ group. Total RNA was extracted using the Trizol method, and RNA integrity was assessed using the Agilent 2100 Bioanalyzer. Samples with a RIN value ≥ 7.0 were selected for high-throughput sequencing. Raw sequencing data were processed through quality control to obtain clean reads, which were then aligned to the mouse reference genome. Gene expression levels were calculated and normalized using FPKM. The DESeq2 software was used to identify differentially expressed genes (DEGs) with the following screening criteria: FDR < 0.05 and |log2FC| ≥ 1. KEGG pathway enrichment analysis was performed on the DEGs, and gene set enrichment analysis (GSEA) was conducted on the whole-genome expression data. Enrichment results were considered statistically significant at an FDR ≤ 0.05.

Western blot analysis

Brain tissue was harvested from the right ischemic hemisphere of mice. RIPA lysis buffer containing protease inhibitors and phosphatase inhibitors was added, and the tissue was homogenized on ice. The mixture was centrifuged at 4 °C and 12,000 rpm for 15 min, and the supernatant was collected. Protein concentration was quantified using the BCA assay. Equal volumes of protein were subjected to SDS-PAGE electrophoresis, transferred to a membrane, and blocked. The membranes were incubated with primary antibodies against p-RIPK1, RIPK1, p-RIPK3, RIPK3, p-MLKL, MLKL, and GAPDH at 4 °C overnight. After washing, the membranes were incubated with corresponding secondary antibodies at room temperature for 1 h. Develop the membrane using ECL, the image was captured, and quantitative analysis of the band gray values was performed using ImageJ software. The expression levels of phosphorylated proteins are expressed as a ratio relative to the corresponding total protein, with total protein normalized to GAPDH.

In vivo safety evaluation

Blood samples were collected from the orbital veins of mice. Whole blood was used for routine blood tests to analyze red blood cell count (RBC), hemoglobin (HGB), white blood cell count (WBC), and platelet count (PLT); After allowing the blood samples to stand, centrifuge to separate the serum and test for the liver function markers alanine aminotransferase (ALT) and aspartate aminotransferase (AST), as well as the kidney function markers blood urea nitrogen (BUN) and creatinine (CREA). Concurrently, major organs, including the heart, liver, spleen, lungs, and kidneys, were collected, fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, and stained with H&E. Histopathological changes in the tissues of each organ were then examined under a light microscope.

Statistical analysis

All experiments were repeated at least three times independently, and results are expressed as mean ± standard deviation (mean ± SD). Statistical analysis was performed using GraphPad Prism 9.5 software: Student’s t-test was used for comparisons between two independent samples; one-way analysis of variance (ANOVA) combined with Tukey’s post-hoc test was used for comparisons among multiple groups. Statistical significance was set at *p < 0.05, **p < 0.01, and ***p < 0.001.

Results and discussion

Fabrication and physicochemical characterization of the cRGD-Exo@Mn₃O₄

A cRGD-functionalized exosome-based nanosystem loaded with Mn₃O₄ nanoparticles (cRGD-Exo@Mn₃O₄) was successfully constructed. Mn₃O₄ nanoparticles were first synthesized via thermal decomposition, exhibiting a uniform size of ~ 10 nm with good dispersity, as confirmed by TEM (Fig. 1A, Fig. S2A). XRD patterns revealed characteristic diffraction peaks that were essentially consistent with the Mn₃O₄ standard card (PDF#24–0734) (Fig. S2B). XPS analysis further confirmed the coexistence of mixed Mn²⁺/Mn³⁺ valence states in the nanoparticles, providing a structural basis for their enzyme-mimetic antioxidant activity (Fig. S2C) [31]. This mixed-valence structure endows Mn₃O₄ with multiple enzyme-mimetic activities, including SOD-like and CAT-like catalytic activities, enabling efficient ROS scavenging [32]. cRGD peptides were conjugated to the surface of exosomes via click chemistry, and Mn₃O₄ nanoparticles were loaded into the cRGD-modified exosomes via membrane fusion. TEM images showed that free exosomes exhibit a typical cup-shaped morphology (Fig. 1B). The overall morphology of cRGD-Exo, after cRGD peptide modification, did not undergo significant disruption and retained an intact vesicular structure, indicating that the cRGD modification process had minimal impact on the exosome membrane structure (Fig. 1C). After loading with Mn₃O₄ nanoparticles, the particle size of cRGD-Exo@Mn₃O₄ was approximately 140 nm, and it retained an intact spherical vesicle structure. Compared with unloaded exosomes, cRGD-Exo@Mn₃O₄ exhibited a distinct core-shell structure, indicating that cRGD-Exo had successfully enveloped the Mn₃O₄ nanoparticles (Fig. 1D). The synthesized Mn₃O₄ nanoparticles exhibited a relatively uniform spherical morphology with a primary particle size of approximately 10 nm, which minimizes potential shape-dependent effects on biological performance.

Fig. 1.

Synthesis and characterization of cRGD-Exo@Mn₃O₄. (A–D) TEM images of Mn₃O₄ (A), Exo (B), cRGD-Exo (C), and cRGD-Exo@Mn₃O₄ (D). (E) Hydrodynamic diameters of Mn₃O₄, Exo, cRGD-Exo, and cRGD-Exo@Mn₃O₄. (F) Zeta potentials of Mn₃O₄, Exo, cRGD-Exo, and cRGD-Exo@Mn₃O₄. (G) Western blot analysis of exosome markers (TSG101, CD63) and a negative control (Calnexin). (H) Fluorescence image of cRGD-Exo@Mn₃O₄ showing the targeting peptide (green: DiO-labeled exosome membrane; red: Cy5.5-labeled cRGD peptide). (I–K) Changes in particle size (I), Zeta potential (J), and PDI (K) of cRGD-Exo@Mn₃O₄ during 7-day storage at 4 °C. (L) Hemolysis ratio of cRGD-Exo@Mn₃O₄ at different concentrations. (M–O) Enzyme-mimetic ROS scavenging activities of cRGD-Exo@Mn₃O₄, including SOD (M), CAT (N), •OH scavenging activity (O). Data are presented as mean ± SD from three independent experiments (n = 3)

Dynamic light scattering (DLS) results showed that the particle sizes of Mn₃O₄ nanoparticles, Exo, cRGD-Exo, and cRGD-Exo@Mn₃O₄ were 19.97 ± 3.11 nm, 101.1 ± 4.33 nm, 128.4 ± 2.71 nm, and 160.5 ± 6.47 nm, respectively (Fig. 1E). The gradual increase in hydrodynamic diameter after cRGD modification and Mn₃O₄ incorporation confirmed the successful engineering of the exosomal nanoplatform. The larger particle sizes measured by DLS compared with TEM were expected because TEM measures the dehydrated core size under vacuum conditions, whereas DLS measures the hydrodynamic diameter of nanoparticles in aqueous suspension, including the hydration layer and surface modifications. The zeta potential of Mn₃O₄ nanoparticles, Exo, cRGD-Exo, and cRGD-Exo@Mn₃O₄ was − 5.55 ± 0.31 mV, − 12.23 ± 0.32 mV, − 16.37 ± 0.84 mV, and − 14.30 ± 0.79 mV, respectively (Fig. 1F). The slight shift in surface charge after Mn₃O₄ loading indicated successful incorporation of Mn₃O₄ nanoparticles into the exosomal structure. Importantly, the preserved negative surface charge is favorable for maintaining colloidal stability and reducing nonspecific adsorption of serum proteins, which may reduce nonspecific protein adsorption and contribute to colloidal stability during systemic circulation.

To quantitatively evaluate the loading efficiency of Mn₃O₄ nanoparticles into exosomes, ICP-OES was employed to determine the manganese content in cRGD-Exo@Mn₃O₄ samples. ICP-OES analysis showed that three independent batches of cRGD-Exo@Mn₃O₄ were analyzed, and the manganese content was converted to Mn₃O₄ content based on the molecular composition of Mn₃O₄. Combined with the total exosomal protein content determined by the BCA assay, the encapsulation efficiency and drug loading capacity were calculated. As summarized in Table S1, the encapsulation efficiency and Mn₃O₄ loading capacity of three independent batches were 39.33 ± 2.28% and 13.86 ± 0.81%, respectively. The relatively low batch-to-batch variation indicates the good reproducibility of the freeze–thaw loading strategy. These results demonstrate that Mn₃O₄ nanoparticles were successfully incorporated into the exosomal vesicles with stable loading performance, confirming efficient and reproducible loading of Mn₃O₄.

To verify the preservation of exosomal characteristics after engineering, Western blot analysis was performed to detect exosomal marker proteins in cRGD-Exo@Mn₃O₄. The results showed that cRGD-Exo@Mn₃O₄ expressed the exosomal markers TSG101 and CD63, whereas the endoplasmic reticulum marker Calnexin was undetectable (Fig. 1G). These results indicate that exosomal characteristics were retained after cRGD modification and Mn₃O₄ loading, with minimal contamination from cellular components. To confirm the surface conjugation of cRGD peptides onto exosomes, fluorescence labeling and colocalization analysis were performed. The DiO-labeled exosome membrane (green) exhibited substantial fluorescence overlap with Cy5.5-labeled cRGD peptides (red), and a distinct yellow signal was observed in the merged images, confirming the successful association of cRGD peptides with the exosome surface (Fig. 1H).

Considering that exosome-based injectable formulations are generally prepared shortly before administration to preserve biological activity, short-term storage stability is important for subsequent biological applications. Therefore, the stability of cRGD-Exo@Mn₃O₄ was monitored over 7 days under storage conditions relevant to subsequent biological experiments. The stability results showed that after being stored in an aqueous solution for 7 days, the hydrodynamic diameter, zeta potential, and PDI of cRGD-Exo@Mn₃O₄ remained largely unchanged, indicating that this nanosystem exhibits good short-term colloidal stability under aqueous conditions (Fig. 1I–K). Hemolysis assays demonstrated that even at a high concentration of 400 µg/mL, the hemolysis rate remained below the 5% safety threshold for biomaterials, confirming that cRGD-Exo@Mn₃O₄ has minimal impact on red blood cell membrane integrity and exhibits good hemocompatibility (Fig. 1L).

ROS scavenging capability of cRGD-Exo@Mn₃O₄

Ischemia-reperfusion injury is characterized by excessive ROS generation, where the oxidative cascade involving •O₂⁻, H₂O₂, and hydroxyl radicals (•OH) plays a central role in neuronal damage. Given that Mn₃O₄ nanoparticles have been reported to exhibit multiple enzyme-mimetic activities for effective ROS scavenging, we evaluated the SOD-like activity, CAT-like activity, and hydroxyl radical scavenging capacity of cRGD-Exo@Mn₃O₄ to verify its core antioxidant functionality. The results demonstrated that cRGD-Exo@Mn₃O₄ exhibited a significant concentration-dependent scavenging efficiency against the three primary ROS involved in the pathological progression of IS. Specifically, the •O₂⁻ scavenging rate (Fig. 1M) increased from 14.7% at 20 µg/mL to 80.1% at 100 µg/mL, indicating robust superoxide anion scavenging capability, while the CAT-like activity (Fig. 1N) revealed that the H₂O₂ scavenging rate rose sharply from 18.7% to 88.4% over the same concentration range, demonstrating exceptional hydrogen peroxide decomposition performance. Furthermore, the nanosystem exhibited an effective scavenging effect on •OH (Fig. 1O), with the clearance rate ascending from 29.4% at 20 µg/mL to 74.2% at 100 µg/mL. In parallel, the cRGD-Exo control without Mn₃O₄ loading was evaluated (Fig. S3A–C), showing negligible scavenging of •O₂⁻ and H₂O₂ and only marginal activity against •OH, which is insufficient to reverse oxidative stress. Together, these results indicate that the superior antioxidant performance of cRGD-Exo@Mn₃O₄ is primarily derived from the encapsulated Mn₃O₄ nanoparticles.

To comprehensively assess the biocompatibility of the cRGD‑Exo@Mn₃O₄ nanoplatform with the representative cellular components of the neurovascular unit, we selected bEnd.3, HT22, and BV2 as representative cell lines. Cytotoxicity assays revealed that even at a concentration of 100 µg/mL, cell viability remained above 85% for all three cell types, indicating that cRGD‑Exo@Mn₃O₄ possesses favorable cytocompatibility (Fig. S4A–C).

The above results demonstrate the successful fabrication of a cRGD-Exo@Mn₃O₄ engineered nanoplatform with appropriate particle size, favorable dispersibility, stability, and good hemocompatibility. In vitro assays further confirmed that this nanoplatform exhibited multiple enzyme-mimicking antioxidant activities and efficient ROS scavenging capacity. These findings suggest that cRGD-Exo@Mn₃O₄ may serve as a potential therapeutic platform for ischemic stroke treatment. Furthermore, the exosome-based formulation provides enhanced biocompatibility and aqueous dispersibility, supporting its subsequent evaluation for ischemic lesion targeting and neuroprotective effects.

In vitro ischemia-targeting and transendothelial transport apability of cRGD-Exo@Mn₃O₄

To evaluate the effect of cRGD modification on cellular targeting capability under ischemic conditions, the uptake efficiency of the nanosystem was first investigated in bEnd.3 cerebral microvascular endothelial cells subjected to OGD/R treatment. The results showed that under normal conditions, weak fluorescence was observed in all groups; however, following OGD/R treatment, the intracellular fluorescence intensity in the cRGD-Exo@Mn₃O₄ group was significantly enhanced, approximately 2.53 times that of the Exo@Mn₃O₄ group (Fig. 2A, B), indicating that cRGD modification enhances the uptake of the nanosystem by ischemic endothelial cells, possibly through interaction with α_vβ₃ integrin, which is upregulated under ischemic conditions. Through interaction with α_vβ₃ integrin receptors, cRGD may contribute to preferential recognition and accumulation of the nanosystem within the ischemic microenvironment [33, 34].

Fig. 2.

Evaluation of the in vitro ischemic endothelial targeting and transendothelial delivery capability of cRGD-Exo@Mn₃O₄. (A) Fluorescence images of Exo@Mn₃O₄ and cRGD-Exo@Mn₃O₄ uptake by bEnd.3 cells under OGD/R conditions (green: DiO-labeled exosomes; blue: Hoechst 33342-labeled cell nuclei). (B) Quantitative analysis of fluorescence intensity for the uptake of Exo@Mn₃O₄ and cRGD-Exo@Mn₃O₄ by bEnd.3 cells under OGD/R conditions. (C) Schematic diagram of an in vitro BBB model. (D) Fluorescence images of Exo@Mn₃O₄ and cRGD-Exo@Mn₃O₄ uptake by HT22 cells at the bottom after transendothelial delivery in the Transwell model. (E) Quantitative analysis of fluorescence intensity for Exo@Mn₃O₄ and cRGD-Exo@Mn₃O₄ uptake by HT22 cells at the bottom after transendothelial delivery. (F) Fluorescence images of uptake by HT22 cells after co-incubation with cRGD-Exo@Mn₃O₄ for different durations. (G) Quantitative analysis of fluorescence intensity of uptake by HT22 cells after co-incubation with cRGD-Exo@Mn₃O₄ for different durations. Data are presented as mean ± SD from three independent experiments (n = 3)

Furthermore, a Transwell-based endothelial-neuronal co-culture model was established to evaluate the transendothelial delivery capability of cRGD-Exo@Mn₃O₄ under ischemic conditions (Fig. 2C). To mimic the ischemic pathological state, the established BBB model was subjected to OGD/R treatment before the transport experiments. The integrity and barrier functionality of the endothelial monolayer were monitored by measuring transendothelial electrical resistance (TEER) throughout the experimental period. As shown in Fig. S5, the TEER values of the endothelial monolayers were maintained above 200 Ω·cm² before nanoparticle incubation, confirming the successful establishment of a functional barrier model. OGD/R stimulation induced a moderate decrease in TEER values, reflecting impaired endothelial barrier function under ischemic conditions; however, the TEER values remained above the threshold indicative of a severely disrupted barrier. Notably, cRGD-Exo@Mn₃O₄ treatment partially restored TEER values during the incubation period, suggesting a protective effect on endothelial barrier integrity. These results indicate that the enhanced transendothelial transport of cRGD-Exo@Mn₃O₄ was unlikely to be solely caused by ischemia-induced barrier disruption. Instead, the increased transport efficiency may be attributed to cRGD-mediated interactions with ischemia-activated endothelial cells, while the detailed mechanisms underlying receptor-mediated transport require further investigation. Following incubation, fluorescence signals were detected in HT22 neuronal cells located in the lower chamber, and the fluorescence intensity gradually increased with prolonged incubation time. At 24 h, HT22 cells treated with cRGD-Exo@Mn₃O₄ exhibited significantly higher fluorescence intensity than those treated with Exo@Mn₃O₄, with approximately a 2.12-fold increase (Fig. 2D, E). Additionally, HT22 neuronal uptake of cRGD-Exo@Mn₃O₄ displayed a time-dependent pattern (Fig. 2F, G), further confirming efficient delivery of the nanoplatform in this in vitro endothelial-neuronal model. Previous studies have suggested that exosomes can undergo receptor-mediated interactions with brain endothelial cells, contributing to cellular uptake and transport across endothelial barriers [35]. In the present study, cRGD modification further enhanced endothelial cell recognition and internalization under ischemic conditions, resulting in improved transendothelial delivery efficiency of cRGD-Exo@Mn₃O₄ in the established in vitro model. Nevertheless, it should be noted that the Transwell co-culture system represents a simplified in vitro BBB model and does not fully recapitulate the structural and functional complexity of the neurovascular unit. Moreover, OGD/R treatment may partially compromise endothelial barrier integrity, which could also contribute to the increased transport observed. Therefore, our findings support enhanced transendothelial delivery of cRGD-Exo@Mn₃O₄ in the established in vitro BBB model but do not conclusively demonstrate active receptor-mediated BBB transcytosis. Further mechanistic studies using more physiologically relevant BBB models will be valuable to clarify the relative contribution of active targeting to brain delivery.

cRGD-Exo@Mn₃O₄ alleviates oxidative stress, suppresses neuroinflammation, and protects neurons in vitro

Oxidative stress is a major contributor to neuronal death after IS. In this study, H₂O₂ was used to stimulate HT22 cells to simulate oxidative damage during ischemia-reperfusion, and intracellular ROS levels were assessed using the DHE (dihydroethidium) fluorescent probe and the CCK-8 assay. Following H₂O₂ induction, the PBS group exhibited intense red fluorescence, indicating that H₂O₂ induced the production of a large amount of ROS. In contrast, both Exo and Mn₃O₄ treatments reduced ROS levels to varying degrees. Intracellular ROS levels were further reduced in the Exo@Mn₃O₄ group, and the cRGD-Exo@Mn₃O₄ group exhibited the weakest ROS fluorescence, indicating that this nanosystem has an enhanced effect on ROS scavenging. Quantitative fluorescence analysis revealed that the average fluorescence intensity in the PBS group increased to 2.7 times that of the normal control group, whereas the average fluorescence intensity in the cRGD-Exo@Mn₃O₄ group decreased to near-normal levels, significantly lower than that of the PBS group and superior to the other treatment groups (Fig. 3A, B). Consistent with the DHE staining results, following H₂O₂ treatment, the viability of HT22 cells in the PBS group dropped sharply to 24.3%, suggesting that the oxidative stress model was successfully established. Compared with the PBS group, both the Exo and Mn₃O₄ treatment groups could improve cell viability to some extent, while the Exo@Mn₃O₄ group showed significantly higher cell viability than either the Exo or Mn₃O₄ group alone, suggesting a synergistic protective effect upon their combination. Notably, cell viability in the cRGD-Exo@Mn₃O₄ group recovered significantly to 83.4%, significantly outperforming the Exo@Mn₃O₄ group, indicating that it possesses the optimal cell protection capacity under oxidative damage conditions and effectively alleviates oxidative damage-induced neuronal death following neuronal oxidative damage (Fig. 3C).

Fig. 3.

Evaluation of the antioxidant, anti-inflammatory, and neuroprotective effects of cRGD-Exo@Mn₃O₄ in vitro. (A) Fluorescence images of DHE-stained HT22 cells in each group following H₂O₂ treatment. (B) Quantitative analysis of DHE fluorescence intensity. (C) Viability of HT22 cells in each group following H₂O₂ treatment. (D–F) ELISA-detected levels of pro-inflammatory cytokines TNF-α (D), IL-1β (E), and IL-6 (F) secreted by BV2 cells in each group following LPS stimulation. (G) Viability of HT22 cells in each treatment group in the OGD/R model. (H) Calcein-AM/PI staining images of HT22 cells in each group in the OGD/R model. (I) Quantitative analysis of survival rates of HT22 cells in each group in the OGD/R model following Calcein-AM/PI staining. Data are presented as mean ± SD from three independent experiments (n = 3)

In the aftermath of IS, microglia, as the resident innate immune cells of the CNS, undergo rapid activation and shift to a pro-inflammatory state, releasing large quantities of TNF-α, IL-1β, and IL-6 that potentiate neuroinflammatory injury and neuronal loss [36]. To evaluate the ability of nanomedicines to regulate inflammatory responses in the microenvironment, this study utilized the mouse microglial cell line BV2. Lipopolysaccharide (LPS) stimulation was employed to simulate an in vitro neuroinflammatory environment, and the expression levels of TNF-α, IL-1β, and IL-6 in the cell culture supernatant were measured. In the LPS-stimulated BV2 microglial cell inflammation model, the concentrations of TNF-α, IL-1β, and IL-6 in the BV2 cell supernatant all exhibited a significant increase; cRGD-Exo@Mn₃O₄ significantly inhibited the release of the pro-inflammatory factors TNF-α, IL-1β, and IL-6 (Fig. 3D–F).

To evaluate the neuroprotective potential of cRGD-Exo@Mn₃O₄ under ischemia-reperfusion conditions, this study established an OGD/R injury model using HT22 cells. Cell metabolic viability was assessed using the CCK-8 assay, and cell survival was visually evaluated using Calcein-AM/PI dual staining. Compared with the Control group, cell viability in the OGD/R-treated model group (PBS) dropped sharply to approximately 29%, indicating that OGD/R caused severe neuronal cell death and that the in vitro model was successfully established. Following treatment with Exo or Mn₃O₄ nanoparticles alone, cell viability showed a certain degree of recovery, suggesting that both agents possess inherent neuroprotective effects. Compared with the non-targeted Exo@Mn₃O₄ group, the cRGD-Exo@Mn₃O₄ group exhibited a more significant neuroprotective effect, with cell viability recovering to over 80% (Fig. 3G). The Calcein-AM/PI dual-staining results further confirmed its significant inhibitory effect on neuronal death. The field of view in the blank control group was filled with dense green fluorescence and almost no red fluorescence, with cells appearing plump, indicating good cellular growth. In the PBS group, green fluorescence was sharply reduced, cells became shrunken and rounded, and extensive red fluorescence denoting dead cells emerged, reflecting loss of membrane integrity and confirming severe OGD/R-mediated neuronal death. As the treatment strategy was optimized from single-component to loaded-component approaches, red fluorescence gradually decreased, and green fluorescence density gradually recovered. The staining results of the cRGD-Exo@Mn₃O₄ group were closest to those of the blank control group. A large number of green fluorescent cells were visible in the field of view, with only a very small amount of red fluorescence. Quantitative statistical analysis based on the fluorescence images showed high consistency with the CCK-8 assay results (Fig. 3H, I). The proportion of live cells in the cRGD-Exo@Mn₃O₄ group was significantly higher than that in the PBS group and other treatment groups, confirming the superior efficacy of this nanosystem in counteracting OGD/R-induced neuronal damage and maintaining neuronal survival.

cRGD-Exo@Mn₃O₄ exhibits enhanced accumulation in ischemic brain regions and alleviates brain injury in the MCAO/R model

To evaluate the in vivo therapeutic efficacy and brain accumulation of cRGD-Exo@Mn₃O₄, a middle cerebral artery occlusion/reperfusion (MCAO/R) mouse model was established. One hour after reperfusion initiation, PBS, Exo, Mn₃O₄, Exo@Mn₃O₄, and cRGD-Exo@Mn₃O₄ were administered via tail vein injection (Fig. 4A). To further characterize the dynamic biodistribution profile of cRGD-Exo@Mn₃O₄, longitudinal fluorescence imaging was performed at multiple time points after intravenous administration. The fluorescence signals gradually accumulated in the brain region, reaching a relatively high level at 12 h, followed by a gradual decline over time (Fig. 4B). Compared with Exo@Mn₃O₄, cRGD-Exo@Mn₃O₄ exhibited enhanced brain-associated fluorescence signals, suggesting improved accumulation efficiency after cRGD functionalization. Ex vivo fluorescence imaging of major organs further revealed that the liver and spleen were the primary distribution sites, consistent with the clearance characteristics of exosome-based nanomaterials. Notably, cRGD-Exo@Mn₃O₄ showed enhanced fluorescence enrichment in ischemic brain tissue compared with the unmodified control group (Fig. 4C), supporting its improved ischemic lesion-associated accumulation capability. Ex vivo fluorescence imaging demonstrated that both Exo and Exo@Mn₃O₄ exhibited a certain degree of accumulation in the brain, whereas cRGD-Exo@Mn₃O₄ displayed the strongest fluorescence signal in the ischemic cerebral hemisphere. At 12 h after administration, the fluorescence intensity of the cRGD-Exo@Mn₃O₄ group was approximately 2.98-fold and 2.92-fold higher than that of the Exo and Exo@Mn₃O₄ groups, respectively. Furthermore, the fluorescence intensity ratio between the ischemic and contralateral hemispheres was significantly increased in the cRGD-Exo@Mn₃O₄ group, reaching approximately 2.08-fold that of the Exo@Mn₃O₄ group, suggesting that cRGD functionalization promoted increased retention and accumulation of the nanoplatform in ischemic brain tissue (Fig. 4E, F). These findings are consistent with the enhanced cellular uptake observed in vitro and further support the potential contribution of cRGD-mediated interactions to ischemic brain accumulation. However, ischemia–reperfusion injury is accompanied by substantial BBB disruption, increased vascular permeability, inflammatory activation, and altered vascular clearance, all of which may contribute to nanoparticle accumulation in the ischemic hemisphere. In addition, the interaction between cRGD and ischemia-associated α_vβ₃ integrin expression may further facilitate lesion-associated retention of the nanoplatform. Therefore, the present fluorescence imaging results demonstrate enhanced accumulation of cRGD-Exo@Mn₃O₄ in ischemic brain regions, but they do not independently prove active BBB transcytosis or definitive integrin-mediated transport in vivo. Further studies, such as α_vβ₃ integrin-blocking experiments and quantitative biodistribution analyses, are warranted to clarify the precise targeting mechanism.

Fig. 4.

cRGD-Exo@Mn₃O₄ enhances ischemic brain accumulation and ameliorates MCAO/R-induced neurological injury. (A) Experimental design and timeline. (B) Representative whole-body fluorescence images of mice intravenously injected with Exo, Exo@Mn₃O₄, or cRGD-Exo@Mn₃O₄ at different time points (6, 12, and 24 h). (C) Ex vivo fluorescence imaging of isolated brains collected at 6, 12, and 24 h after administration. (D) Quantitative analysis of fluorescence intensity in the brain regions. (E) Quantification of fluorescence signals in the brain tissues. (F) The ratio of brain fluorescence intensity to the corresponding whole-body fluorescence intensity. (G) TTC staining of brain tissue from MCAO/R mice following different treatments. Red and white represent normal and infarcted regions, respectively. (H) Quantitative analysis of infarct volume in brain tissue from MCAO/R mice in each group. (I) Neurological deficit scores of mice in different treatment groups based on Zea–Longa evaluation. (J) Measurement of brain water content in different groups. (K–M) ELISA analysis of TNF-α (K), IL-1β (L), and IL-6 (M) levels in ischemic brain tissues from different treatment groups. Data are presented as mean ± SD from three independent experiments (n = 3)

TTC staining results showed that brain sections in the Sham group appeared uniformly red overall, with no signs of infarction. In contrast, the PBS model group exhibited extensive infarct areas in the right cerebral hemisphere, with an infarct volume as high as 42.96%, confirming the successful establishment of the MCAO/R model and extensive ischemic injury. All treatment groups demonstrated varying degrees of reduction in infarct area. Among these, the infarct size in the single-component group was reduced compared to the PBS group, but the effect was limited. The infarct area in the Exo@Mn₃O₄ group was further reduced. The cRGD-Exo@Mn₃O₄ group demonstrated the most superior therapeutic effect, with brain sections showing only minimal pale areas and the infarct volume reduced to 8.03% (Fig. 4G, H). Neurological function was further evaluated using Zea–Longa scoring by an investigator blinded to treatment groups. Compared with the PBS group, mice treated with Exo@Mn₃O₄ and cRGD-Exo@Mn₃O₄ exhibited significantly improved neurological performance. Notably, the neurological deficit score in the cRGD-Exo@Mn₃O₄ group was reduced by 63.49% compared with that in the PBS group (Fig. 4I), indicating marked functional recovery after treatment. In addition, cRGD-Exo@Mn₃O₄ treatment markedly reduced cerebral edema in MCAO/R mice. Brain water content increased from 76.97% ± 0.37% in the Sham group to 84.25% ± 0.79% in the PBS group, indicating severe edema formation after ischemia–reperfusion injury (Fig. 4J). Treatment with Exo, Mn₃O₄, and Exo@Mn₃O₄ partially reduced brain water content, with varying degrees of reduction. In comparison, the cRGD-Exo@Mn₃O₄ group exhibited a significant reduction in brain water content to 78.37% ± 0.37%, approaching the level observed in the Sham group. These findings suggest that cRGD-Exo@Mn₃O₄ effectively alleviated ischemia-associated cerebral edema.

To further evaluate inflammatory responses in vivo, ELISA assays were performed to measure TNF-α, IL-1β, and IL-6 levels in ischemic brain tissues. Compared with the Sham group, MCAO/R injury markedly increased the expression levels of these pro-inflammatory cytokines in the PBS group. In contrast, treatment with cRGD-Exo@Mn₃O₄ significantly reduced TNF-α, IL-1β, and IL-6 levels compared with other treatment groups (Fig. 4K–M), suggesting effective suppression of ischemia-associated inflammatory responses. These in vivo findings were also consistent with the anti-inflammatory effects observed in LPS-stimulated BV2 cells in vitro. Collectively, the enhanced accumulation of cRGD-Exo@Mn₃O₄ in ischemic brain regions was accompanied by significant neuroprotection and attenuation of inflammatory responses following MCAO/R injury.

cRGD-Exo@Mn₃O₄ alleviates oxidative stress in ischemic brain tissue

Since excessive ROS accumulation is a critical contributor to ischemia–reperfusion injury, we further evaluated oxidative stress levels in ischemic brain tissues following different treatments. SOD activity and MDA content were measured as indicators of antioxidant defense and lipid peroxidation, respectively. Compared with the Sham group, MCAO/R markedly reduced SOD activity and increased MDA levels in the ischemic hemisphere, confirming severe oxidative stress after ischemic injury. In contrast, cRGD-Exo@Mn₃O₄ treatment significantly restored SOD activity and decreased MDA accumulation compared with the PBS group (Fig. S6), indicating that the nanoplatform effectively attenuated oxidative damage in vivo. These findings are consistent with the intrinsic ROS-scavenging activity of Mn₃O₄ nanozymes observed in vitro and further support the antioxidant mechanism underlying the neuroprotective effects of cRGD-Exo@Mn₃O₄.

cRGD-Exo@Mn₃O₄ alleviates pathological damage in the brain tissue of MCAO/R mice and promotes neurological recovery

To further evaluate the protective effects of cRGD-Exo@Mn₃O₄ on ischemic brain injury, histopathological analyses were performed using H&E and Nissl staining. H&E staining results showed that brain tissues in the Sham group exhibited intact cortical and hippocampal structures with well-preserved neuronal morphology and no obvious pathological abnormalities (Fig. 5A). In contrast, the PBS-treated MCAO/R group displayed severe pathological damage in the ischemic hemisphere, including disorganized tissue architecture, neuronal loss, cell shrinkage, and nuclear pyknosis, which are characteristic features of ischemia–reperfusion injury [37]. Compared with the PBS group, treatment with cRGD-Exo@Mn₃O₄ markedly alleviated histopathological damage in ischemic brain tissues. The cortical and hippocampal regions in the cRGD-Exo@Mn₃O₄ group retained relatively intact tissue structures with improved neuronal morphology and reduced cellular degeneration, showing morphological features more comparable to those observed in the Sham group. Nissl staining further demonstrated substantial neuronal damage in the PBS-treated MCAO/R group, as indicated by decreased numbers of Nissl bodies and disrupted neuronal morphology in the hippocampal region (Fig. 5B). In contrast, the cRGD-Exo@Mn₃O₄-treated group retained a relatively higher density of Nissl bodies with more intact neuronal morphology and more uniform cytoplasmic staining, suggesting reduced neuronal degeneration following treatment. In addition, open-field behavioral tests were conducted to assess locomotor activity and exploratory behavior in MCAO/R mice. Compared with the PBS group, cRGD-Exo@Mn₃O₄ treatment significantly improved total travel distance, average movement speed, and exploration behavior in the central area, including both movement distance and residence time (Fig. 5C–G). The behavioral performance of mice treated with cRGD-Exo@Mn₃O₄ was closer to that of the Sham group, suggesting improved neurological recovery.

Fig. 5.

cRGD-Exo@Mn₃O₄ mitigates histopathological damage and improves functional recovery after MCAO/R. (A) Representative H&E staining images of brain sections from different groups, showing pathological changes in the cortex and hippocampus. (B) Representative Nissl staining images of brain sections from different groups. (C) Representative movement trajectories and heatmaps of mice from different groups during the open-field test. (D–G) Quantitative analysis of open-field behavioral performance, including total travel distance (D), average movement speed (E), distance traveled in the central area (F), and residence time in the central area (G). Data are presented as mean ± SD from three independent experiments (n = 3)

cRGD-Exo@Mn₃O₄ attenuates ischemic injury by suppressing RIPK1/RIPK3/MLKL-mediated necroptotic signaling

To elucidate the underlying molecular mechanisms by which cRGD-Exo@Mn₃O₄ exerts its therapeutic effects in IS, this study performed transcriptomic sequencing analysis on brain tissues from mice in the PBS group and the cRGD-Exo@Mn₃O₄ treatment group. The gene expression profiles of the cRGD-Exo@Mn₃O₄ group and the PBS group exhibited distinctly different clustering patterns, with significant differences in gene expression patterns between the groups, indicating that cRGD-Exo@Mn₃O₄ treatment induced substantial transcriptional remodeling in ischemic brain tissue. Compared with the PBS group, 762 genes were significantly downregulated, and 55 genes were upregulated in the treatment group (Fig. 6A–C). KEGG enrichment analysis revealed that differentially expressed genes were primarily enriched in pathways related to inflammatory immune regulation, cell death, and stress response (Fig. 6D). Among these, the TNF signaling pathway and the NOD-like receptor signaling pathway are key upstream pathways regulating neuroinflammation and the release of pro-inflammatory factors, while necroptosis is one of the primary forms of programmed neuronal death following ischemia. These enriched pathways are closely associated with inflammatory regulation, programmed cell death, and neuronal survival after ischemic injury [38]. GSEA analysis further confirmed that cRGD-Exo@Mn₃O₄ treatment significantly influenced the expression patterns of relevant gene sets. In the TNF signaling pathway, NOD-like receptor signaling pathway, and necroptosis pathway, the normalized enrichment scores (NES) were − 1.986, − 1.907, and − 1.586, respectively, all showing significant negative enrichment, indicating that cRGD-Exo@Mn₃O₄ intervention significantly inhibits the abnormal activation of these three pathways (Fig. 6E–G). The number of differentially expressed genes enriched in the TNF signaling pathway, NOD-like receptor signaling pathway, and necroptosis pathway were 38, 42, and 21, respectively (Fig. 7A–C). The Venn diagram showed that there were 5 differentially expressed genes common to all three pathways, including Ripk1, Ripk3, Casp8, Il-1β, and Birc3 (Fig. 7D, E). Among these, RIPK1 and RIPK3 are core components of the RIPK1/RIPK3/MLKL signaling axis, a classic pathway mediating necroptosis. Previous studies have reported that ischemia-reperfusion injury can activate the RIPK1/RIPK3/MLKL pathway, and inhibition of this pathway can significantly reduce cerebral infarct volume and improve neurological deficits [39].

Fig. 6.

Transcriptomic analysis reveals gene expression changes and pathway enrichment induced by cRGD-Exo@Mn₃O₄. (A) Heatmap of differentially expressed genes between the PBS group and the cRGD-Exo@Mn₃O₄ group. (B) A volcano plot showing upregulated and downregulated genes. (C) Quantitative statistics of differentially expressed genes between the cRGD-Exo@Mn₃O₄ group and the PBS group. (D) KEGG pathway enrichment analysis. (E–G) Gene set enrichment analysis plots for the TNF signaling pathway (E), NOD-like receptor signaling pathway (F), and necroptosis pathway (G)

Fig. 7.

cRGD-Exo@Mn₃O₄ attenuates ischemia-induced activation of the RIPK1/RIPK3/MLKL necroptosis pathway. (A–C) Heatmaps of differentially expressed genes in the TNF signaling pathway (A), NOD-like receptor signaling pathway (B), and necroptosis pathway (C). (D) Venn diagram of differentially expressed genes in the TNF signaling pathway, NOD-like receptor signaling pathway, and necroptosis pathway. (E) Heatmap of differentially expressed genes in the intersection of the TNF signaling pathway, NOD-like receptor signaling pathway, and necroptosis pathway. (F) Western blot analysis of the expression of RIPK1, RIPK3, MLKL, and their phosphorylated forms (p-RIPK1, p-RIPK3, p-MLKL) in brain tissues from mice in each group. (G) Schematic diagram illustrating the mechanism by which cRGD-Exo@Mn₃O₄ blocks necroptosis by inhibiting the phosphorylation-mediated activation of the RIPK1/RIPK3/MLKL pathway. (H–J) Western blot analysis of protein expression levels of p-RIPK1 (H), p-RIPK3 (I), and p-MLKL (J) in each group. Data are presented as mean ± SD from three independent experiments (n = 3)

In the classical necroptosis pathway, binding of a death ligand, such as TNF-α, to their cognate receptors induces the recruitment of RIPK1 and the formation of signaling complexes. When caspase-8 activity is inhibited or insufficient, RIPK1 interacts with and activates RIPK3 through phosphorylation, leading to the formation of the necrosome/necroptosome complex. Activated RIPK3 subsequently phosphorylates the downstream effector molecule MLKL. Phosphorylated MLKL oligomerizes and translocates to the plasma membrane, where it forms membrane pores and disrupts membrane integrity, ultimately resulting in necroptotic cell death [40]. Increasing evidence indicates that necroptosis is critically involved in the pathological progression of ischemic stroke [41]. Following cerebral ischemia-reperfusion injury, oxidative stress, calcium overload, and inflammatory activation contribute to RIPK1/RIPK3/MLKL pathway activation, thereby exacerbating neuronal loss and neurological dysfunction [42, 43]. Moreover, tnecroptotic cell death promotes the release of damage-associated molecular patterns (DAMPs), further amplifying neuroinflammation and establishing a vicious cycle between neuronal injury and inflammatory responses [44]. To further validate the transcriptomic findings at the protein level, Western blot analysis was performed to evaluate the expression and phosphorylation status of key proteins involved in the necroptosis pathway, including RIPK1, RIPK3, MLKL, p-RIPK1, p-RIPK3, and p-MLKL, in ischemic brain tissue. Compared with the Sham group, MCAO/R markedly increased the phosphorylation levels of RIPK1, RIPK3, and MLKL, indicating activation of the RIPK1/RIPK3/MLKL signaling axis. In contrast, cRGD-Exo@Mn₃O₄ treatment significantly reduced the phosphorylation levels of these proteins without substantially altering their total protein expression (Fig. 7F–J, Fig. S7A–C). These results suggest that cRGD-Exo@Mn₃O₄ alleviates ischemia–reperfusion-induced neuronal injury, at least partly, through suppression of RIPK1/RIPK3/MLKL-mediated necroptotic signaling.

To further assess neuronal membrane damage associated with ischemic cell death, PI staining combined with neuronal marker immunofluorescence was performed in ischemic brain sections. Compared with the Sham group, MCAO/R injury resulted in a significant increase in PI-positive signals within NeuN-positive neurons in the ischemic region, indicating increased loss of neuronal membrane integrity after ischemia–reperfusion injury. Notably, cRGD-Exo@Mn₃O₄ treatment markedly reduced PI-positive neuronal signals compared with the PBS group (Fig. S8), suggesting improved preservation of neuronal membrane integrity and reduced necrotic cell death. Although PI staining alone cannot specifically distinguish necroptosis from other forms of regulated necrotic cell death, the reduced neuronal PI uptake, together with the inhibition of RIPK1/RIPK3/MLKL phosphorylation and transcriptomic pathway enrichment results, provides additional evidence confirming the involvement of necroptotic signaling in the neuroprotective effects of cRGD-Exo@Mn₃O₄.

In vivo biosafety of cRGD-Exo@Mn₃O₄

To evaluate the short-term in vivo biocompatibility of cRGD-Exo@Mn₃O₄, the major organs, hematological parameters, and liver and kidney function were examined following intravenous administration. H&E staining revealed no obvious pathological abnormalities in the heart, liver, spleen, lungs, or kidneys of mice from any treatment groups compared with the Sham group (Fig. S9). Complete blood count analysis showed that red blood cell count (RBC), hemoglobin (HGB), white blood cell count (WBC), and platelet count (PLT) remained within the normal physiological range in all groups (Fig. S10A–C). Serum biochemical analysis further demonstrated that alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (CREA) levels were not significantly different among the treatment groups (Fig. S11A–C), indicating that cRGD-Exo@Mn₃O₄ did not induce detectable acute hepatic or renal toxicity under the experimental dose and administration regimen. Collectively, these findings indicate that cRGD-Exo@Mn₃O₄ exhibits favorable short-term systemic compatibility at the administered dose, supporting its potential for further therapeutic development. Nevertheless, long-term biosafety, including manganese accumulation, repeated-dose toxicity, immunogenicity, and long-term neurological outcomes, warrants further investigation before clinical translation.

Conclusion

In summary, we developed a multifunctional cRGD-Exo@Mn₃O₄ nanoplatform by integrating cRGD-functionalized neural progenitor cell-derived exosomes with Mn₃O₄ nanozymes for ischemic stroke therapy. The engineered nanosystem exhibited uniform morphology, favorable colloidal stability, and excellent hemocompatibility. cRGD functionalization enhanced the interaction of exosomal nanoparticles with ischemia-associated endothelial cells and facilitated increased accumulation within ischemic brain regions. In vitro studies demonstrated that cRGD-Exo@Mn₃O₄ improved cellular uptake under ischemic conditions, enhanced transendothelial delivery in an endothelial-neuronal model, and provided effective protection against oxidative stress and inflammatory injury. Benefiting from the intrinsic biological activity of neural progenitor cell-derived exosomes and the multi-enzyme mimetic antioxidant properties of Mn₃O₄ nanoparticles, cRGD-Exo@Mn₃O₄ effectively scavenged excessive reactive oxygen species, suppressed microglial-mediated neuroinflammation, and alleviated neuronal damage. In the MCAO/R model, systemic administration of cRGD-Exo@Mn₃O₄ resulted in enhanced enrichment in ischemic brain tissue and significantly improved neurological outcomes, as evidenced by reduced infarct volume, decreased cerebral edema, attenuated inflammatory cytokine production, and improved behavioral recovery. Mechanistically, transcriptomic and molecular analyses revealed that cRGD-Exo@Mn₃O₄ exerted neuroprotective effects, at least partly, through inhibition of RIPK1/RIPK3/MLKL pathway activation and suppression of ischemia-reperfusion-induced necroptotic signaling. Preliminary biosafety evaluation further demonstrated that cRGD-Exo@Mn₃O₄ exhibited good short-term in vivo tolerability without detectable systemic toxicity at the therapeutic dose. Collectively, this study presents an exosome-based nanozyme platform that integrates ischemic lesion-associated enrichment, antioxidative regulation, anti-inflammatory activity, and neuronal protection, providing a promising strategy for the treatment of ischemic stroke. Further investigations focusing on the detailed targeting mechanisms, neurovascular unit interactions, long-term biodistribution, and safety profiles will be essential to advance the translational potential of this nanoplatform.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

X. Deng and X.X. Xie contributed equally to this work. W. Chen and C.X. Chen conceived and designed the experiments. X. Deng and X.X. Xie performed most of the experiments. X. Zhou assisted with the cell experiments. L.F. Jiang assisted with the characterization of nanomaterials. T. Zhu assisted with data analysis. R.X. Su and C.M. Chen assisted with the animal experiments. X. Deng, X.X. Xie, and C.X. Chen analyzed the data and wrote the manuscript. All authors read and revised the manuscript.

Funding

This work was supported by the Guangxi Natural Science Foundation Project (Nos. 2026GXNSFD202160007 and 2023GXNSFAA026138), the National Nature Science Foundation of China (No. 82260258), the Guangxi Key Research and Development Plan (Guike AB25069064), the National Key Laboratory of Neuroscience and Oncology Drug Research Open Projects (SIML-202402), and the Zhong Nanshan Youth Science and Technology Innovation Award Fund of the China Youth Entrepreneurship and Employment Foundation (No. 240607).

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no conflict of interest.