Exercise-induced extracellular vesicles derived from platelet-rich plasma improved recovery after ischemic stroke

Abstract

Stroke remains a major global health burden, with limited treatments for chronic ischemic stroke necessitating novel therapies. This study explored the therapeutic potential of platelet-rich plasma (PRP)-derived extracellular vesicles (EVs) in stroke recovery, particularly in exercise-trained rats. PRP-derived EVs from treadmill-loaded and sedentary rats were designated athletes (aPRP-EVs) and non-athlete (nPRP-EVs), respectively. Both were administered to primary cortical neurons exposed to oxygen-glucose deprivation (OGD) and to adult male Wistar/ST rats subjected to permanent middle cerebral artery occlusion (MCAO). Exercise increased CD63, CD31, and transforming growth factor-β1 (TGF-β1) in PRP-derived EVs. In OGD-exposed neurons, aPRP-EVs enhanced viability, elevated phosphorylated neurofilament heavy chain, and reduced intracellular calcium. Canonical pathway analysis showed upregulated TGF-β/SMAD signaling in EV groups versus vehicle, while ‘Ca signaling’ was downregulated in aPRP-EVs versus nPRP-EVs. In MCAO rats, EVs improved neurological and motor function and reduced neuronal apoptosis at 28 days, with aPRP-EVs promoting earlier, greater recovery and infarct reduction. These effects correlated with TGF-β1 upregulation, SMAD4 nuclear translocation, reduced NMDAR2B expression, and enhanced axonal growth in the peri-infarct region. PRP-derived EVs, particularly from exercise-trained donors, enhance neuroregeneration and functional recovery in chronic ischemic stroke via TGF-β/SMAD and calcium signaling modulation.

Introduction

Stroke remains a primary cause of death and long-term disability globally. ¹ Although treatments, such as intravenous thrombolysis and mechanical thrombectomy, options for the chronic phase remain predominantly limited to rehabilitation. Many patients continue to experience enduring neurological deficits, highlighting the need for novel chronic stroke therapies. Recent regenerative strategies, including stem cell transplantation, have shown promise in bridging this gap. ² Neuroprotective agents are actively being investigated as potential therapeutic approaches. ³ However, despite encouraging preclinical results, clinical translation of neuroprotective therapies remains challenging.

Extracellular vesicles (EVs), including exosomes, are small, membrane-bound vesicles 40–100 nm in diameter. Once considered cellular waste, EVs are now recognized as key mediators of intercellular communication. ⁴ These vesicles carry biologically active molecules—lipids, proteins, and nucleic acids—and play roles in cell signaling, gene regulation, and immune modulation. Recent studies suggest that EVs contribute to essential biological processes, including neurogenesis, ⁵ immune regulation, ⁶ and inflammatory responses. ⁷ These findings highlight EVs’ potential for developing therapies for neurological and systemic diseases, including stroke.

Research shows the therapeutic efficacy of stem cell-derived EVs, particularly from bone marrow mesenchymal stem cells (BMSCs),^8,9 and human umbilical cord mesenchymal stem cells (HUBC-MSCs). ¹⁰ A major challenge is isolating and culturing these stem cells to obtain EVs. Previously, we showed that astrocyte-derived EVs promote axonal outgrowth and modulate post-stroke glial scar formation.^11,12 Nonetheless, the therapeutic potential of EVs from alternative sources remains largely unexplored.

This study focused on platelet-rich plasma (PRP)-derived EVs. PRP shows neurorestorative efficacy in a rat model of spinal cord injury. ¹³ Platelet-derived growth factors, elevated in exercise-loaded blood, ¹⁴ likely contribute to PRP’s therapeutic effects. ¹⁵ While PRP-derived EVs have been studied in orthopedic diseases, ¹⁶ their potential in neurological disorders, including stroke, remains largely unexplored. Unlike stem cells, which are difficult to source, store, and expensive to produce, PRP can be easily isolated from blood. PRP-derived EVs are generated by freezing at −80°C to activate platelets, enabling long-term storage before EV extraction. ¹⁷ Given these advantages, we investigated PRP-derived EVs from exercise-loaded blood as a novel therapy for chronic ischemic stroke recovery.

Material and methods

The Animal Care Committee of Juntendo University approved all experiments. Animals had ad libitum access to food and water and were housed following the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All animal experiments were conducted and reported in accordance with the ARRIVE guidelines 2.0.

Isolation of 2 types of PRP-derived EVs

We isolated two PRP-derived EV types: athlete PRP-derived EVs (aPRP-EVs) and non-athlete PRP-derived EVs (nPRP-EVs), as described below.

Adult male Wistar/ST rats (9 weeks old, ≈290 g; Charles River, Kanagawa, Japan) were used to avoid the potential influence of endogenous estradiol on cerebral ischemia in female rats. ¹⁸ Rats were housed on a 12 h light/dark cycle with free access to food and water in a controlled environment. ¹⁸ To generate athlete rats, we modified exercise protocols known for neuroprotection.^19,20 Rats underwent moderate treadmill exercise (15 m/min, 30 min per session, 3 times weekly) between 7 am–9 am from 5–9 weeks of age. Non-athlete rats were housed in standard cages without treadmill exposure during the same period. Whole blood was drawn from both groups via a 20 G needle inserted into the left ventricle under general anesthesia, and the rats were euthanized. Blood was centrifuged at 300 g for 15 min at 25°C to separate plasma. From this, 2.0 mL of PRP was extracted above the precipitated blood cells; the remainder was classified as platelet-poor plasma (PPP). PRP was centrifuged at 1400 g for 10 min at 25°C, and the pellet was diluted with 1 mL of PPP. Platelets were activated using 2 freeze-thaw cycles at −60°C. The sample was centrifuged at 1400 g for 10 min at 25°C, retaining only the supernatant. Finally, 10 mL PBS was added, and the mixture was ultrafiltered twice at 100 kDa and 2000 g at 4°C.

EVs were isolated by ultracentrifugation using a Himac CS 150 GXL (Hitachi Koki, Japan) with a P150AT angle rotor at 11,000 × g for 120 min at 4°C (Figure 1(a)). EVs were resuspended in RIPA Lysis and Extraction Buffer (Thermo Scientific, Rockford, IL, USA) for Western blotting and in PBS for other applications. These PRP-derived EVs were defined as aPRP-EVs and nPRP-EVs. The 2 EV types were characterized using NanoSight™ and EXOVIEW™ analysis and Western blot, as described below.

Figure 1.

Comparison of the 2 types of PRP-derived EVs. (a) Schematic representation of the protocol for isolating PRP-derived EVs from non-exercised (non-athlete) and exercised (athlete) rats. (b) Comparison of PRP-derived EV concentrations between groups using NanoSight™ particle tracking analysis (n = 8/group). (c) Size distribution analysis of nPRP-EVs (blue) and aPRP-EVs (red) using NanoSight™ particle tracking analysis (n = 5/group). (d) Tetraspanin expression on EVs assessed via EXOVIEW™. While CD63 Continued.expression was higher in aPRP-EVs, this difference did not reach statistical significance (P = 0.14, n = 5/group). (e) Representative Western blot images and densitometric analysis of PRP-derived EVs. aPRP-EVs exhibited increased CD63 expression among the 3 tetraspanins, along with elevated levels of the platelet marker CD31 and the platelet-derived growth factors VEGF and TGF-β1. Transferrin served as an internal control (n = 5/group). PRP: platelet-rich plasma; EVs: extracellular vesicles; nPRP-EVs: non-athlete PRP-derived extracellular vesicles; aPRP-EVs: athlete PRP-derived extracellular vesicles; FGF-β: basic fibroblast growth factor; PDGF-β: Platelet-Derived Growth Factor subunit B; TGF-β1: Transforming Growth Factor-beta1; VEGF: Vascular Endothelial Growth Factor. Data are expressed as mean (SD), n = 5–8/group. **P < 0.01, ***P < 0.001 versus nPRP-EVs. Unpaired Student’s t-test.

Nanosight™ analysis

EV concentration (n = 5/group) and size distribution (n = 8/group) were analyzed using the NanoSight LM10 platform (Malvern Panalytical, Kassel, Germany). Following previous studies with slight modifications, ¹² samples were diluted 1:2000 in PBS and measured in duplicate. A total of 400 µL of the diluted sample was injected into the chamber, and each sample was measured 5 times, recording 30-s videos per run.

EXOVIEW™ analysis

Tetraspanin expression in PRP-derived EVs was analyzed using the EXOVIEW™ system (Unchained Labs, Pleasanton, CA) following the manufacturer’s protocol and prior studies. ¹² Samples (35 µL, 1.0 × 10¹⁰ particles/mL) were incubated with capture antibodies on a microarray chip for 24 h. After washing, chips were incubated with fluorescent antibodies, rewashed, and analyzed using the ExoView R100 IMAGER. Data were processed using ExoViewer 2.7.9 (n = 5/group).

Primary cortical neurons

Cortical cells from embryonic day 17 Wistar rats (Charles River) were harvested following a published protocol with minor modifications. ²¹ Dissociated cells (6.0 × 10⁷ cells/mL) were cultured in a Neurobasal medium supplemented with L-glutamine and B27 (GIBCO, Grand Island, NY, USA). At 7 days in vitro, the medium was replaced with Ca²⁺- and Mg²⁺-free Hanks’ balanced salt solution, and neurons were incubated in 5% CO₂/95% N₂. After 3 h of oxygen–glucose deprivation (OGD), the culture medium was replaced with a Neurobasal medium supplemented with L-glutamine and B27, and neurons were cultured for 96 h in the following groups: (1) vehicle (no treatment), (2) nPRP-EVs, and (3) aPRP-EVs (n = 5/group).

Application of PRP-Derived EVs to primary neuronal cultures

To investigate the effects of PRP-derived EVs on neurons in the peri-infarct area of chronic ischemic stroke in vitro, 3 × 10⁹ EVs/mL were added per 12-well plate to primary cortical neuron cultures 96 h post-OGD, following previously described methods. ²²

Western blot

Cultured neurons were lysed in Cell Lytic Mammalian Tissue Lysis/Extraction Reagent (Sigma-Aldrich, St. Louis, MO, USA), while PRP-derived EVs were lysed in RIPA lysis buffer. Protein extraction and electrophoresis were performed, followed by transfer to polyvinylidene difluoride membranes. Membranes were blocked with 1.0% bovine serum albumin (BSA) for 60 min and incubated overnight at 4°C with primary antibodies. The primary antibodies used were rabbit anti-CD9 (1:1000; Abcam), mouse anti-CD63 (1:1000; Santa Cruz), mouse anti-CD81 (1:100; Invitrogen), rabbit anti-transforming growth factor-beta1 (TGF-β1) (1:1000; Proteintech), rabbit anti-vascular endothelial growth factor (VEGF) (1:500; Proteintech), rabbit anti-basic fibroblast growth factor (FGF-β) (1:1000; Bioss), rabbit anti-platelet-derived growth factor subunit B (PDGF-β) (1:1000; Abcam), mouse anti-microtubule-associated protein 2 (MAP-2) (1:5000; MilliporeSigma), mouse anti-SMI32 (1:5000; BioLegend), mouse anti-phosphorylated neurofilament heavy protein (pNFH) (1:5000; BioLegend), rabbit anti-N-methyl D-aspartate receptor subtype 2B (NMDAR2B) (1:200; Proteintech), rabbit anti-transferrin (1:5000; Abcam), and mouse anti-β-Actin (1:10000; Abcam). Membranes were incubated with peroxidase-conjugated secondary antibodies, and proteins were visualized using enhanced chemiluminescence (GE Healthcare, UK). Five samples per group were analyzed. Protein expression was quantified using densitometry in ImageJ and normalized to transferrin for EVs and β-actin for neurons. Relative expression in aPRP-EVs was compared to nPRP-EVs for EVs, while each treatment group was compared to the vehicle for neurons to calculate arbitrary units.

5 samples per group were analyzed. Protein expression was quantified using densitometry in ImageJ software and normalized to transferrin for EVs and β-actin for neurons. Relative expression in aPRP-EVs was compared to nPRP-EVs for EVs, while each treatment group was compared to the vehicle for neurons to calculate arbitrary units.

Cell viability assays

Neuronal cell viability was assessed using the Cell Counting Kit-8 (CCK-8) cytotoxicity assay (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). After 3 h of OGD treatment, neurons were plated in 12-well plates, and the medium was replaced with 500 µL of complete medium and 50 µL of CCK-8 solution. Plates were incubated at 37°C for 96 h, and absorbance was measured at 450 nm using an iMark microplate reader (Bio-Rad, Hercules, CA, USA). Experiments were conducted with samples from 4 groups, including the control group (n = 10/group).

Microarray data analysis

For mRNA analysis, total RNA was extracted from cultured cortical neurons using the RNeasy® Plus Mini Kit (Qiagen, Valencia, CA, USA; n = 4/group). Microarray analysis was conducted on samples from 4 groups, including the control group. The extracted mRNA was analyzed at Hokkaido System Science (Sapporo, Japan) using the SurePrint G3 Rat GE 8 × 60 K Ver. 2.0 platform (Agilent Technologies, Santa Clara, CA, USA), following the manufacturer’s protocol. Specifically, for canonical pathway analysis, a −log(P-value) > 2 was used as the significance threshold, and z-scores > 2 or < −2 were considered indicative of significant activation or inhibition, respectively. Consistency scores were calculated for regulatory effects and molecular networks, with higher scores reflecting greater accuracy. Upstream regulator analysis used a P-value of overlap <0.05 for the significance threshold. The algorithms for calculating z-scores and P-values of overlap were based on a previous study. ²³

Calcium (Ca) imaging of cultured cortical neurons with OGD challenging

Ca²⁺ imaging was performed as previously described ²⁴ with minor modifications. Dissociated cortical cells were cultured to confluence in 35 mm glass-bottom dishes (Matsunami Glass, Kishiwada, Osaka, Japan) and subjected to an OGD challenge, forming 3 intervention groups as described above. After 24 h, the neurons were washed 3 times with balanced salt solution (BSS) buffer (20 mM HEPES–KOH, 135 mM NaCl, 5.4 mM KCl, 2 mM CaCl₂, 10 mM glucose) (iNtRON Bio, Kirkland, WA, USA) and incubated with 0.01% F-127 (Biotium, Fremont, CA, USA) and 5 µM Fura2-AM (Dojindo, Kamimashiki, Kumamoto, Japan) for 30 min at 37°C in the dark. After incubation, neurons were washed twice with BSS buffer before observation. Data analysis was conducted using AquaCosmos 2.0 (Hamamatsu Photonics, Hamamatsu, Japan). Intracellular Ca²⁺ concentration was quantified by selecting regions of interest in 5 randomly chosen neurons per field of view, with 20 neurons per dish. The fluorescence ratio (F340/F380) was recorded every 0.2 s for at least 20 s. The average value per dish was calculated and compared among the 4 groups, including the control group (n = 5/group).

Inhibiting the TGF-β/SMAD pathway and stimulating NMDA receptor (NMDAR)

To investigate the mechanism of action of aPRP-EVs on neuroregeneration, galunisertib (a TGF-β receptor inhibitor; Selleck Chemicals) and NMDA (an NMDAR agonist; Selleck Chemicals) were administered to OGD-challenged cultured cortical neurons at concentrations of 0, 1, 3, and 10 µM. Neuronal survival rates and pNFH (a marker of axons) expression levels were compared across samples to determine the optimal concentrations of galunisertib and NMDA. Based on these findings, we conducted experiments to assess the effects of aPRP-EVs with or without these compounds. Primary cortical neurons were cultured as described above, subjected to OGD, and treated with aPRP-EVs only, aPRP-EVs + galunisertib, aPRP-EVs + NMDA, or aPRP-EVs + galunisertib + NMDA. Neuronal survival rates and the expression of pNFH, TGF-β1, and NMDAR2B were compared across 6 groups, including the control group (n = 5/group).

Generation of middle cerebral artery occlusion (MCAO) model rats

MCAO model rats were prepared using established protocols. ²⁵ Nine-week-old male Wistar/ST rats were anesthetized with 4.0% isoflurane (Abbott Japan CO., Ltd. Tokyo, Japan) and maintained at 1.0%–1.5% isoflurane in a 70% N₂O and 30% O₂ gas mixture. The rats underwent permanent left MCAO via insertion of a 4-0 surgical nylon suture with an expanded tip, created by heating with a lighter flame.

On days 3, 7, 14, 21, and 28 post-MCAO, rats were transcardially perfused under deep anesthesia with pentobarbital. Their brains were promptly removed, fixed in 4% paraformaldehyde for 24 h, and then immersed in 30% sucrose for at least 48 h. Brain coronal sections (20 µm thick) were prepared using a cryostat (Leica Biosystems, Wetzlar, Germany) for histological and immunohistochemical analyses, as described below.

Administration of PRP-derived EVs to MCAO rats

To examine the effects of PRP-derived EVs on the peri-infarct area in chronic ischemic stroke, PRP-derived EVs were administered intravenously via the tail vein at a dose of 100 µg protein equivalent, as reported in previous studies. ²⁶ EVs were given 7 days post-MCAO, based on our prior findings that astrocyte-derived EVs enhanced stroke recovery at this time point. ¹² Rats subjected to MCAO and PRP-derived EV treatment were divided into 3 groups: (1) vehicle (no treatment), (2) nPRP-EVs, and (3) aPRP-EVs.

Behavioral analyses

Modified neurological severity scores (mNSS) ²⁷ assess motor, sensory, reflex, and balance functions, ranging from 0 (normal) to 18 (severe deficits). For the rotarod test, rats were tested on a rotating cylinder (MK-660D, Muromachi Kikai Co., Ltd., Tokyo, Japan), and the time spent on the rotarod was recorded. The rotation speed increased from 10 to 40 rpm (0.005–0.08 g) over 120 s until the rat fell. Each rat underwent 6 trials, and the meantime on the rotarod during the second to fifth trials was calculated.^22,28 Evaluations for both mNSS and the rotarod test were conducted at 7, 14, 21, and 28 days post-MCAO, with n = 10/group. Survival counts and body weight measurements were also recorded on day 0 (pre-MCAO) and at 7, 14, 21, and 28 days post-MCAO.

Immunohistochemistry

Immunohistochemical staining was performed as previously described. ¹¹ Prior to transcardial perfusion, rats were deeply anesthetized with pentobarbital. The brains were immediately removed en bloc and fixed for 24 h in 4% paraformaldehyde in PBS at 4°C. After fixation, samples were soaked in 30% sucrose for at least 24 h. The brains were then frozen, and 20 µm-thick consecutive coronal sections were prepared using a cryostat. Sections were washed 3 times with PBS, blocked with 1.0% BSA for 60 min, and incubated overnight at 4°C with primary antibodies. The primary antibodies used were mouse monoclonal anti-MAP-2 (1:200; MilliporeSigma), mouse monoclonal anti-Neuronal nuclei (NeuN) (1:200; Sigma-Aldrich), rabbit polyclonal anti-integrin α_vβ₃ (1:100; Bioss), rabbit polyclonal anti-TGF-β1 (1:200; Proteintech), rabbit polyclonal anti-SMAD4 (1:75; Proteintech), rabbit polyclonal anti-NMDAR2B (1:200; Proteintech), and mouse monoclonal anti-pNFH (1:500; BioLegend). After PBS washes, sections were incubated with Cy3- or FITC-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) for 60 min at room temperature, shielded from light. Sections were covered with VECTASHIELD containing DAPI (Vector Laboratories). For each analysis, images were acquired from 3 randomly selected locations within layer VI of the cortex in the peri-infarct area, 300 µm from the ischemic core, using a 40× objective lens. Analyses were conducted on the following groups: (1) vehicle (no treatment), (2) nPRP-EVs, and (3) aPRP-EVs, with some including a sham-operated group (n = 5/group).

Comparison of cerebral infarct volume ratio

We performed hematoxylin-eosin (HE) staining on 6 representative 20 µm-thick coronal sections of rat brains, selected from the region spanning +1.6 mm to −4.0 mm relative to bregma, based on the Paxinos and Watson (2007) atlas (Paxinos, G., and Watson, C., 2007, The Rat Brain in Stereotaxic Coordinates, 6^th ed. San Diego, CA: Academic Press). Brain specimens were obtained from 3 experimental rat groups. The infarcted area in each section was determined using the following equation:

$$\mathrm{Infarct\hspace{0.17em}\hspace{0.17em}area=brain\hspace{0.17em}\hspace{0.17em}area\hspace{0.17em}\hspace{0.17em}of\hspace{0.17em}\hspace{0.17em}the\hspace{0.17em}\hspace{0.17em}unaffected\hspace{0.17em}\hspace{0.17em}}\mathrm{hemisphere\hspace{0.17em}\hspace{0.17em}–\hspace{0.17em}\hspace{0.17em}Non-infarcted\hspace{0.17em}\hspace{0.17em}area\hspace{0.17em}\hspace{0.17em}}\mathrm{of\hspace{0.17em}\hspace{0.17em}the\hspace{0.17em}affected\hspace{0.17em}\hspace{0.17em}hemisphere}$$

To estimate infarct volume, the infarct area of each section was multiplied by 20 µm, and the sum of the 6 sections was calculated. Similarly, the volume of the unaffected hemisphere was determined by multiplying the brain area of the unaffected hemisphere by 20 µm and summing the values across the 6 sections. The percentage of infarct volume was then calculated using the following formula:

$$\mathrm{Percentage}\mathrm{\hspace{0.17em}\hspace{0.17em}of\hspace{0.17em}\hspace{0.17em}}\mathrm{infarct}\mathrm{\hspace{0.17em}\hspace{0.17em}}\mathrm{volume}=\left(\mathrm{infarct}\mathrm{\hspace{0.17em}}\mathrm{volume}/\mathrm{volume}\mathrm{\hspace{0.17em}\hspace{0.17em}of\hspace{0.17em}\hspace{0.17em}}\mathrm{unaffected}\mathrm{\hspace{0.17em}\hspace{0.17em}}\mathrm{hemisphere}\right)\times 100$$

EVs tracking

To confirm that intravenously administered EVs successfully crossed the blood-brain barrier and were taken up by cells in the peri-infarct area, we labeled EVs from both the nPRP-EVs and aPRP-EVs groups using the ExoSparkler Exosome Membrane Labeling Kit-Red (Dojindo Laboratories, Kumamoto, Japan), according to the manufacturer’s instructions.

Briefly, Mem Dye-Red was dissolved in dimethyl sulfoxide (DMSO), and 2 µL of this solution was added to 100 µL of PBS containing 1 × 10¹⁰ EV particles. The solution was incubated at 37°C for 30 min, transferred to a filtration tube, and centrifuged at 3000 g for 5 min at room temperature. The centrifugation was repeated twice with 100 µL of PBS to remove excess dye. Finally, 50 µL of PBS was added, and the EVs were pipetted up and down 10 times to ensure thorough resuspension.

The labeled EVs were then administered via the tail vein of rats 7 days post-MCAO. Rat brains were extracted 24 h after administration, and frozen sections were prepared for observation.

Image acquisition and quantification

Coronal brain sections were examined using a Zeiss LSM 780 confocal laser scanning microscope (Carl Zeiss, Jena, Germany) and analyzed by immunohistochemical and TdT-mediated dUTP nick‑end labeling (TUNEL) staining. Three randomly selected sections from layer VI of the cortex within the peri-infarct region were evaluated per section at 40× magnification. The peri-infarct area was defined as the region 300 µm from the ischemic core in each HE-stained section. For sham-operated rats, the area corresponding to the peri-infarct region in MCAO rats was randomly selected for analysis.

Statistical analysis

Values are expressed as mean (standard deviation [SD]). Prior to applying parametric tests, the normality of data distribution was confirmed using the Shapiro–Wilk test. An unpaired Student’s t-test was used to assess significant differences between the 2 groups, while a one-way analysis of variance (ANOVA) with post hoc Tukey’s correction was applied for comparisons among 3 or more groups. Survival curves were generated using Kaplan–Meier analysis, with differences assessed using the log-rank test. Microarray data were evaluated for statistical significance using an unpaired Student’s t-test in GeneSpring software. Statistical significance was defined as P < 0.05. All statistical analyses were performed using JMP software (version 11.2; SAS Institute Inc.), and graphs were generated using GraphPad Prism (version 5.0; GraphPad Software, Inc.). Power estimates were calculated with α = 0.05 and β = 0.8 to ensure group sizes sufficient to detect effect sizes of 40%–50% in cell culture models and 30%–50% in in vivo models. Rats used for immunohistochemical analysis were randomly selected from those undergoing physiological assessments. All experiments and measurements were conducted in a blinded and randomized manner.

Results

Exercise-induced alterations in cargo proteins in PRP-derived EVs

EVs were extracted from PRP by ultracentrifugation from both athlete and non-athlete rats (Figure 1(a)). There were no significant differences in protein content between nPRP-EVs and aPRP-EVs, and NanoSight™ analysis showed no significant differences in EV concentration or particle size distribution (Figure 1(b) and (c)). EXOVIEW™ analysis revealed that aPRP-EVs contained more CD63⁺ EVs than nPRP-EVs, with a measurement of 1.969 (1.328) arbitrary units (Figure 1(d)). Western blot analysis confirmed higher CD63 protein levels in aPRP-EVs (3.390 [0.090] arbitrary units) than in nPRP-EVs. However, CD9⁺ and CD81⁺ levels measured by EXOVIEW™ and CD9 and CD81 levels from Western blotting did not differ between the 2 groups. Notably, CD31 (Platelet Endothelial Cell Adhesion Molecule 1), TGF-β1 (2.879 [0.817] arbitrary units), and VEGF (1.620 [0.101] arbitrary units) were more highly expressed in aPRP-EVs than in nPRP-EVs, while PDGF-β and FGF-β levels showed no significant differences (Figure 1(e)). These results align with previous studies demonstrating the positive effects of exercise on EV composition and function.^29–31 This suggests that exercise enhances PRP-derived EVs by increasing the surface expression cargo proteins such as CD63, TGF-β1, and VEGF—restorative components associated with PRP. ³²

Effects of PRP-derived EVs on cultured cortical neurons after ischemia

To assess the effects of PRP-derived EVs on ischemic neurons in vitro, cultured cortical neurons were divided into 4 groups: (1) control, without OGD or EVs; (2) vehicle, with OGD but no EVs; (3) nPRP, with OGD and nPRP-EVs; and (4) aPRP-EVs, with OGD and aPRP-EVs. Neuronal viability was significantly lower in the vehicle, nPRP-EVs, and aPRP-EVs groups compared with the control group. However, viability was higher in the nPRP-EVs group than in the vehicle group. The aPRP-EVs group showed significantly higher viability than the vehicle and nPRP-EVs groups (vehicle: 0.240 [0.064]; nPRP-EVs: 0.352 [0.054]; aPRP-EVs: 0.518 [0.111] arbitrary units; Figure 2(a)). These results indicate that while OGD reduced viability, PRP-derived EVs alleviated this effect.

Figure 2.

Effects of PRP-derived EVs on neuronal survival and related protein expression after OGD. (a) Representative image and comparison of neuron survival at 96 h post-OGD among treatment groups. Both PRP-derived EV-treated groups showed significantly higher neuron survival compared to the vehicle-treated group. Neuron survival was also significantly higher in the aPRP-EVs-treated group than in the nPRP-EVs-treated group. ***P < 0.001 versus control, ##P < 0.01, ###P < 0.001 versus vehicle-treated group; §§§P < 0.001 versus nPRP-EVs-treated group. Data are presented as the mean (SD) of n = 10/group—one-way ANOVA. Scale bar: 100 µm and (b) representative image and densitometric analysis of Western blot results for neurons at 96 h post-OGD. The expression of pNFH and TGF-β1 was significantly higher in the vehicle, nPRP-EVs-treated, and aPRP-EVs-treated groups, while the expression of NMDAR2B significantly decreased in the same order. *P < 0.05, **P < 0.01, ***P < 0.001 versus vehicle-treated group; #P < 0.05, ##P < 0.01, ###P < 0.001 versus nPRP-EVs-treated group. Data are presented as the mean (SD) of n = 5/group; one-way ANOVA. PRP: platelet-rich plasma; EVs: extracellular vesicles; nPRP-EVs: non-athlete PRP-derived extracellular vesicles; aPRP-EVs: athlete PRP-derived extracellular vesicles; OGD: oxygen-glucose deprivation; TGF-β1: transforming growth factor-beta1; NMDAR2B/NR2B: N-methyl‑D-aspartate receptor subtype 2B; MAP-2: microtubule-associated protein 2; pNFH: phosphorylated neurofilament heavy protein.

We then assessed axonal outgrowth following PRP-derived EV treatment in ischemic neurons. Western blot analysis showed a significant increase in pNFH in the nPRP-EVs group compared with the OGD group and in the aPRP-EVs group compared with both the OGD and nPRP-EVs groups (nPRP-EVs: 1.401 [0.268]; aPRP-EVs: 1.922 [0.290] arbitrary units). No significant differences were observed in MAP-2 across groups (Figure 2(b)). To explore the underlying molecular mechanisms, transcriptome analysis was conducted on the vehicle, nPRP-EVs, and aPRP-EVs groups. Canonical pathway analysis identified 2 key signaling pathways: the Inhibitor of Differentiation 1 (ID1) signaling pathway,’ involving TGF-β/SMAD signaling, upregulated in both nPRP-EVs and aPRP-EVs groups compared to the vehicle group, and ‘Ca signaling’, downregulated in the aPRP-EVs group relative to nPRP-EVs (Figure 3(a) to (c)).

Figure 3.

Effects of PRP-derived EVs on signaling pathways and intracellular calcium dynamics after OGD. (a–c) Predicted signaling pathways based on mRNA expression changes in OGD-treated neurons, analyzed using IPA software. Functional networks were generated via IPA (QIAGEN Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis). N = 4/group. (a) The ID1 signaling pathway, including TGF-β/SMAD signaling, was identified in comparisons between the vehicle-treated group and both the nPRP- EVs and aPRP-EVs-treated groups (left). (b–c) In comparisons between the nPRP-EVs and aPRP-EVs-treated groups, the calcium signaling pathway had the largest absolute value for a negative z-score and (d) representative calcium images of neurons at 24 h post-OGD and comparison of the fluorescence ratio (F340/380) across the 4 groups. The fluorescence ratios in the nPRP- and aPRP-EVs-treated groups were significantly lower than in the vehicle group. Furthermore, the aPRP-EVs-treated group exhibited significantly lower fluorescence intensity than the nPRP-EVs-treated group. **P < 0.01, ***P < 0.001 versus control, ###P < 0.001 versus vehicle-treated group, †P < 0.05 versus nPRP-EVs-treated group. Data are presented as the mean (SD) of n = 5/group, one-way ANOVA.

OGD: oxygen-glucose deprivation; IPA: Ingenuity Pathway Analysis; ID1: Inhibitor of Differentiation 1; TGF-β: Transforming Growth Factor-beta; PRP: platelet-rich plasma; EVs: extracellular vesicles; nPRP-EVs: non-athlete PRP-derived extracellular vesicles; aPRP-EVs: athlete PRP-derived extracellular vesicles.

Regulation of TGF-β/SMAD signaling and Ca signaling

Western blot was used to assess TGF-β/SMAD and Ca signaling pathways identified through Ingenuity Pathway Analysis (IPA). TGF-β1 levels were significantly higher in the nPRP-EVs group compared to the vehicle group and even higher in the aPRP-EVs group (nPRP, 1.393 [0.113]; aPRP, 1.707 [0.257] arbitrary units; Figure 2(b)). NMDAR2B, a subunit of the NMDA receptor involved in Ca signaling, was significantly lower in the nPRP-EVs group than in the vehicle group and further reduced in the aPRP-EVs group (nPRP-EVs: 0.673 [0.042]; aPRP-EVs: 0.394 [0.064] arbitrary units; Figure 2(b)). Ca²⁺ imaging showed OGD significantly raised Ca²⁺ levels despite PRP-derived EV treatment. However, PRP-derived EVs, especially those from aPRP, significantly lowered Ca²⁺ levels compared with the vehicle group (control: 0.352 [0.056]; vehicle: 0.840 [0.050]; nPRP-EVs: 0.594 [0.097]; aPRP-EVs: 0.488 [0.023]; Figure 3(d)). These results suggest that PRP-derived EVs improve neuronal viability and axonal outgrowth after ischemia, likely by modulating TGF-β/SMAD signaling and NMDAR activity, thereby reducing Ca²⁺ overload.

Effects of aPRP-EVs on post-ischemic neurons by modulating TGF-β and NMDARs

Based on the dose-response experiment results (Supplementary Figure), we administered 1 µM Galunisertib and 3 µM NMDA together with aPRP-derived EVs to cultured cortical neurons after OGD. The survival rate was significantly decreased in OGD-challenged groups, regardless of treatment with aPRP-EVs, Galunisertib, or NMDA, compared to the control group, but aPRP-EVs treatment showed a higher survival rate than the other groups (vehicle, 0.350 [0.051]; aPRP-EVs, 0.557 [0.079]; aPRP-EVs+Galunisertib, 0.412 [0.086]; aPRP-EVs+NMDA, 0.375 [0.072]; aPRP-EVs+Galunisertib+NMDA, 0.3626 [0.081], arbitrary units; Figure 4(a)). After OGD, pNFH expression was significantly higher in the aPRP-EVs-only group than combinations of aPRP-EVs with either Galunisertib or NMDA, or both, and the vehicle group (vehicle, 0.232 [0.071]; aPRP-EVs, 0.582 [0.082]; aPRP-EVs+Galunisertib, 0.378 [0.057]; aPRP-EVs+NMDA, 0.382 [0.055]; aPRP-EVs+Galunisertib+NMDA, 0.257 [0.069] arbitrary units; Figure 4(b)). Both Galunisertib and NMDA with aPRP-EVs further reduced pNFH levels compared to other groups, except the vehicle group. TGF-β1 expression was significantly decreased in aPRP-EVs with Galunisertib and aPRP-EVs with both Galunisertib and NMDA, and the vehicle group, compared to treatment with aPRP-EVs only and aPRP-EVs plus NMDA in OGD-challenged neurons (vehicle, 0.483 [0.123]; aPRP-EVs, 0.807 [0.092]; aPRP-EVs+Galunisertib, 0.504 [0.100]; aPRP-EVs+NMDA, 0.736 [0.128]; aPRP-EVs+Galunisertib+NMDA, 0.502 [0.079] arbitrary units; Figure 4(b)). NMDAR2B expression was significantly higher in aPRP-EVs with NMDA and aPRP-EVs with both Galunisertib and NMDA, as well as the vehicle group, compared to aPRP-EVs only and aPRP-EVs plus Galunisertib groups in OGD-challenged neurons (vehicle, 1.883 [0.185]; aPRP-EVs, 1.317 [0.133]; aPRP-EVs+Galunisertib, 1.335 [0.157]; aPRP-EVs+NMDA, 1.648 [0.156]; aPRP-EVs+Galunisertib+NMDA, 1.672 [0.157], arbitrary units; Figure 4(b)).

Figure 4.

Effects of aPRP-derived EVs on post-ischemic neurons by modulating TGF-β and NMDA receptors. (a) Comparison of neuron survival at 96 h post-OGD with the control group. The survival rates of all 5 post-OGD groups were significantly lower than those of the control group. However, the group treated with aPRP-EVs alone exhibited a significantly higher survival rate compared to the other 4 groups. ***P < 0.001 versus control, ###P < 0.001 versus vehicle-treated group, †P < 0.05 versus aPRP-EVs+Galunisertib (TGF-β receptor inhibitor)-treated group, ‡‡P < 0.01 versus aPRP-EVs+NMDA (NMDA receptor agonist)-treated group, ¶¶P < 0.01 versus aPRP-EVs+Galunisertib+NMDA-treated group. Data are presented as the mean (SD) of n = 5/group; one-way ANOVA and (b) representative image and densitometric analysis of Western blot results for neurons at 96 h post-OGD. pNFH expression was significantly higher in the aPRP-EVs-only group and in the aPRP-EVs plus Galunisertib‑ or NMDA-treated groups compared to the vehicle group. No significant differences were observed in the Galunisertib‑ and NMDA-treated group compared to the vehicle group. The aPRP-EVs-only group exhibited significantly higher pNFH expression than the aPRP-EVs plus Galunisertib‑ or NMDA-treated groups. TGF-β1 expression was significantly higher in the aPRP-EVs-only group and the aPRP-EVs plus NMDA-treated group than in the other OGD-challenged groups. NMDAR2B expression was significantly higher in the aPRP-EVs-only group compared to the aPRP-EVs plus Galunisertib-treated group and other OGD-challenged groups. Furthermore, NMDAR2B expression was significantly lower Continued.in the aPRP-EVs-only group and the aPRP-EVs plus Galunisertib-treated group than in other OGD-challenged groups. *P < 0.05, **P < 0.01, ***P < 0.001 versus control; #P < 0.05, ##P < 0.01, ###P < 0.001 versus vehicle-treated group; †P < 0.05, †††P < 0.001 versus aPRP-EVs + Galunisertib-treated group; ‡P < 0.05, ‡‡‡P < .001 versus aPRP-EVs+ NMDA-treated group; ¶P < 0.05, ¶¶P < 0.01, ¶¶¶P < 0.001 versus aPRP-EVs+Galunisertib+ NMDA-treated group. Data are presented as the mean (SD) of n = 5/group; one-way ANOVA.OGD: oxygen-glucose deprivation; PRP: platelet-rich plasma; EVs: extracellular vesicles; aPRP-EVs: athlete PRP-derived extracellular vesicles; pNFH: phosphorylated neurofilament heavy protein; TGF-β1: Transforming Growth Factor-beta1; NMDAR2B: N-methyl D-aspartate receptor subtype 2B.

PRP-derived EVs enhanced motor recovery in a rat MCAO model

PRP-derived EVs were administered 7 days post-MCAO via the tail vein (Figure 5(a)). At this time, integrin α_vβ_3–a ligand for CD31 highly expressed in PRP-derived EVs, particularly in aPRP-EVs—was upregulated in the peri-infarct area. ³³ The expression of integrin α_vβ₃ in neurons was significantly higher on day 7 post-MCAO compared to sham, 3 days, or 28 days post-MCAO (day 7, 5.852 [0.874]%; Figure 5(b)). Previous studies showed that astrocyte-derived EVs, when locally administered on day 7 post-MCAO, promoted secondary axonal regeneration and functional recovery through glial scar modulation. ¹² Based on this, we intravenously administered PRP-derived EVs on day 7, which facilitated neural regeneration and functional recovery. Prior research indicated adhesion between CD31 and α_vβ₃. ³⁴ In our study, neuronal integrin α_vβ₃ expression peaked on day 7 post-MCAO, with increased CD31 expression in aPRP-EVs and labeled EVs taken up by neurons, supporting day 7 as the optimal administration point. Labeled PRP-derived EVs were co-localized with NeuN⁺ neurons in the peri-infarct area of MCAO rats (Figure 5(c)).

Figure 5.

Effect of PRP-derived EVs on cerebral ischemia model in vivo. (a) Schematic diagram of the administration of PRP-derived EVs in an in vivo model of cerebral ischemia. (b) Temporal profile of integrin α_vβ₃ expression in neurons within the peri-infarct area compared to sham-operated rats, with the specific region of observation indicated by a solid rectangular outline in the day 7 tiling Continued.image. Integrin α_vβ₃ expression on neurons was significantly increased 7 days after MCAO compared to sham-operated rats and to days 3 and 28 post-MCAO. Scale bar: 50 µm, magnified image; 1 mm, tiling image. ***P < 0.001 versus sham; ###P < 0.001 versus day 3; ‡‡P < 0.01 versus day 28. Data are the mean (SD) of n = 5/group; one-way ANOVA. (c) Representative images of NeuN (green), ExoSparkler™ (red), and DAPI (blue) immunostaining in the peri-infarct area and the unaffected side at 24 h post-administration of PRP-derived EVs labeled with ExoSparkler™ via the tail vein. PRP-derived EVs were transferred into the cytoplasm of neurons. Scale bar: 50 μm. (d) Comparison of survival rate and temporal changes in body weight, mNSS score, and rotarod test score among vehicle- (black), nPRP- (blue), and aPRP-EVs-treated (red) rats. The aPRP-EVs-treated rats showed significantly better neurological outcomes, as assessed by mNSS, starting at an earlier time point than the nPRP-treated rats. n = 10/group. Data for body weight, mNSS, and rotarod test scores are presented as mean (SD) of the final surviving rats: n = 8 (vehicle and aPRP-EVs group) and n = 7 (nPRP group). *P < 0.05, ***P < 0.001 versus day 7, #P < 0.05 versus day 14 post-MCAO. (e) Representative images of brain coronal sections stained with hematoxylin and eosin from vehicle-treated rats and rats treated with nPRP-EVs and aPRP-EVs show the infarct volume ratio at 28 days post-MCAO. The aPRP-EVs-treated group demonstrated a significantly lower infarct volume ratio than the vehicle-treated group, and the nPRP-EVs-treated group showed a trend toward a lower infarct volume ratio compared to the vehicle group (P = 0.063). Data are presented as the mean (SD) of n = 8/group. *P < 0.05 versus vehicle. One-way ANOVA. Scale bar: 5 mm and (f) Representative image and intensity of TUNEL. TUNEL⁺ neuron density was higher in vehicle-, nPRP-, and aPRP-EVs-treated rats than in sham-operated rats. However, TUNEL⁺ neurons were significantly fewer in nPRP-EVs-treated rats than in the vehicle-treated rats, and even fewer in aPRP-EVs-treated rats than in both vehicle- and nPRP-EVs-treated rats. *P < 0.05, ***P < 0.001 versus sham-operated group; ##P < 0.01, ###P < 0.001 versus vehicle-treated group; †P < 0.05 versus nPRP-treated group. One-way ANOVA. Scale bar: 50 µm. PRP: platelet-rich plasma; EVs: extracellular vesicles; MCAO: middle cerebral artery occlusion; ITGα_vβ₃: integrin α_vβ₃; MAP-2: microtubule-associated protein 2; ExoSparkler: ExoSparkler Exosome Membrane Labeling Kit; NeuN: neuronal nuclei; DAPI: 4’,6-diamidino-2-phenylindole; nPRP-EVs: non-athlete PRP-derived extracellular vesicles; aPRP-EVs: athlete-PRP -derived extracellular vesicles; mNSS: modified neurological severity score; infarct volume ratio (%) = (Infarct Volume/volume of unaffected hemisphere) × 100; TUNEL: TdT-mediated dUTP nick‑end labelling.

No significant differences were observed in survival rate or body weight among vehicle, nPRP-EVs, and aPRP-EVs-treated rats (Figure 5(d)). The mNSS were significantly reduced in nPRP-EVs-treated rats by day 28 (day 28, 3.625 [1.188] versus 1.714 [0.951] points) and in aPRP-EVs-treated rats from days 7–28 (day 14, 5.500 [1.604] versus 3.875 [0.991]; day 21, 4.750 [0.707] versus 2.875 [0.991]; day 28, 3.625 [1.188] versus 0.750 [0.707] points; Figure 5(d)). In the rotarod test, both aPRP-EVs and nPRP-EVs groups showed a longer latency to fall compared to the vehicle group at day 28 (day 28, 63.93 [18.18] versus 85.18 [0.18] versus 89.57 [11.40] s; Figure 5(d)).

PRP-derived EVs-mediated neuronal apoptosis and enhanced axonal outgrowth

HE staining showed that although the nPRP-EVs group had a smaller infarct ratio than the vehicle group, the difference was not statistically significant. However, the aPRP-EVs group showed a significantly reduced infarct ratio (vehicle, 63.59 [5.861]%; nPRP-EVs, 56.15 [4.528]%; aPRP-EVs, 53.78 [7.722]%; Figure 5(e)). At 28 days post-MCAO, TUNEL⁺ neuron density in the peri-infarct area was higher in all treated groups compared with the sham group, significantly lower in nPRP-EVs–treated rats than in the vehicle group, and even lower in aPRP-EVs–treated rats (sham, 0.016 [0.009]%; vehicle, 2.440 [0.317]%; nPRP-EVs, 1.412 [0.589]%; aPRP-EVs, 0.758 [0.186]%; Figure 5(f)).

TGF-β1 expression in neurons was higher in the nPRP-EVs group than the vehicle group, with even greater levels in the aPRP-EVs group (vehicle, 0.806 [0.228]%; nPRP-EVs, 2.420 [0.316]%; aPRP-EVs, 3.352 [0.393]%; Figure 6(a)). SMAD4 expression in NeuN⁺ neurons was significantly higher in the sham, nPRP-EVs, and aPRP-EVs groups than in the vehicle group (sham, 0.180 [0.058]%; vehicle, 0.039 [0.028]%; nPRP-EVs, 0.125 [0.022]%; aPRP-EVs, 0.146 [0.043]%; Figure 6(b)). NMDAR2B expression was significantly lower in the nPRP-EVs group than in the vehicle group, with a further reduction in the aPRP-EVs group (vehicle, 11.24 [2.022]%; nPRP-EVs, 5.594 [0.596]%; aPRP-EVs, 2.048 [0.552]%; Figure 6(c)). pNFH levels were also significantly higher in the nPRP-EVs and aPRP-EVs groups than the vehicle group (vehicle, 1.898 [0.448]%; nPRP-EVs, 4.462 [1.066]%; aPRP-EVs, 5.184 [1.305]%; Figure 6(d)).

Figure 6.

Effect of PRP-derived EV administration on the MCAO model. Serial changes in neurons within the peri-infarct area. (a) Representative image and intensity of MAP-2/TGF-β1. TGF-β1 expression in neurons increased significantly in the vehicle, nPRP-, and aPRP-EVs-treated groups, with the highest levels in the aPRP-EVs-treated group. ***P < 0.001 versus sham-operated group; Continued.###P < 0.001 versus vehicle-treated group; §§§P < 0.001 versus nPRP-EVs-treated group. (b) Representative image and intensity of NeuN/SMAD4/DAPI per neuron. Nuclear expression of SMAD4 in neurons within the peri-infarct area was significantly higher in the nPRP- and aPRP-EVs-treated groups than in the vehicle-treated group. #P < 0.05, ##P < 0.01, ###P < 0.001 versus vehicle-treated group. (c) Representative image and intensity of MAP-2/NMDAR2B. NMDAR2B expression in neurons decreased significantly in the order of vehicle, nPRP-, and aPRP-EVs-treated groups. *P < 0.05, ***P < 0.001 versus sham-operated group; ###P < 0.001 versus vehicle-treated group; §§§P < 0.001 versus nPRP-EVs-treated group and (d) representative images and intensity of pNFH. pNFH expression within the peri-infarct area was significantly higher in the nPRP- and aPRP-EVs-treated groups than in the vehicle group, suggesting enhanced axonal outgrowth in the PRP-derived EVs-treated groups. ***P < 0.001 versus sham-operated group; ##P < 0.01, ###P < 0.001 versus vehicle-treated group. Scale bar: 50 μm. Data are presented as the mean (SD) of n = 5/group; one-way ANOVA. PRP: platelet-rich plasma; nPRP-EVs: non-athlete PRP-derived extracellular vesicles; aPRP-EVs: athlete PRP-derived extracellular vesicles; MCAO: middle cerebral artery occlusion; MAP-2: microtubule-associated protein 2; TGF-β1: transforming growth factor-beta1; NeuN: neuronal nuclei; DAPI: 4’,6-diamidino-2-phenylindole; SMAD4: SMAD family member 4; NMDAR2B/NR2B: N-methyl D-aspartate receptor subtype 2B; pNFH: phosphorylated neurofilament heavy protein.

Discussion

PRP is an autologous blood product from the plasma fraction of centrifuged whole blood. ³⁵ Its therapeutic benefits have been demonstrated in a rat spinal cord injury model, ¹³ and PRP-derived EVs have shown efficacy in treating musculoskeletal diseases. ¹⁶ In this study, PRP-derived EVs from treadmill-exercised donors were enriched in CD63, CD31, TGF-β1, and VEGF compared to non-exercised donors. Both PRP-derived EV types demonstrated significant therapeutic effects on neural restoration in a chronic ischemic stroke model. In both in vitro and in vivo models, PRP-derived EVs upregulated TGF-β1, downregulated NMDAR2B, and promoted axonal outgrowth, indicated by increased pNFH expression.

Elevated SMAD4 expression in NeuN⁺ neurons in the peri-infarct area suggests TGF-β/SMAD signaling activation. In vitro, OGD-challenged neurons treated with PRP-derived EVs showed improved survival and reduced Ca²⁺ levels, while in vivo, a reduction in TUNEL⁺ neurons in the peri-infarct area indicated decreased apoptosis. These findings highlight the neuroprotective effects of PRP-derived EVs, likely by mitigating neuronal death. In the MCAO model, neurological and motor functions improved significantly after PRP-EV treatment. Consistent with previous findings, mesenchymal stem cell-derived EVs suppress Ca²⁺ overload and reduce neuronal death via phosphatidylinositol 3-kinase/AKT signaling, ³⁶ and similarly, our findings suggest PRP-derived EVs may exert neuroprotective effects by regulating calcium dynamics. Exercise-loaded PRP-derived EVs promoted earlier neurological improvements and significantly reduced infarct volume, indicating therapeutic potential for chronic-phase ischemic stroke recovery. The enhanced efficacy of exercise-loaded PRP-derived EVs may stem from enriched platelet-derived growth factors that regulate TGF-β/SMAD and Ca signaling pathways.

TGF-β1, a platelet-derived growth factor abundant in PRP-derived EVs, ³⁷ promotes tissue repair and regeneration. ³⁸ It modulates pathways, including the TGF-β/SMAD, which regulates cell proliferation, differentiation, and survival.^39–41 Activation of TGF-β signaling in the peri-infarct area stimulates neurogenesis and angiogenesis, aiding recovery after ischemic stroke.^15,42 In TGF-β/SMAD signaling, the TGF-β superfamily binds its receptors, triggering phosphorylation of R-SMADs such as SMAD2 and SMAD3. These activated R-SMADs form complexes with SMAD4, translocate to the nucleus, and regulate gene transcription. ⁴³ Despite the upregulation of TGF-β1 expression in the aPRP-EVs group than in the nPRP-EVs group, SMAD4 nuclear translocation in neurons did not differ significantly. This inconsistency may reflect the regulation of SMAD4 by non-TGF-β pathways. Notably, nuclear respiratory factor 1 (Nrf1) can upregulate SMAD4 expression in neurons after MCAO or OGD through a TGF-β–independent mechanism. ⁴⁴ These alternative regulatory influences may have contributed to the SMAD4 expression patterns observed in the peri-infarct area in our study.

The NMDA receptor, especially its NMDAR2B subunit, is crucial in Ca²⁺ overload during excitotoxicity, contributing to ischemic stroke pathology. ⁴⁵ Overactivation of NMDA receptors leads to Ca²⁺ overload, triggering oxidative stress and inflammation, and resulting in neuronal death.^46,47 Downregulation of NMDAR2B in both in vitro and in vivo models, particularly in the aPRP-EVs-treated group, suggests a protective mechanism against excitotoxicity caused by abnormal Ca²⁺ influx. Evidence shows that modulating NMDA receptor subunits reduces excitotoxic damage^47,48 and improves outcomes after ischemic injury.^49,50

In our functional experiments using pathway-specific modulators, both Glunisertib and an NMDA agonist individually inhibited the neuroprotective effects of aPRP-EVs, particularly regarding neuronal survival and pNFH reduction. Their combined administration did not worsen survival outcomes but caused a greater pNFH decrease than either agent alone. These results suggest that TGF-β/SMAD and calcium signaling pathways may independently contribute to axonal regeneration, functioning in parallel rather than as a single, linear cascade. This implies possible partial cross-regulation or convergence between these two pathways, warranting further study.

Our findings suggest PRP-derived EVs exert therapeutic effects by promoting post-ischemic neuroregeneration and protecting against excitotoxic damage through dual regulation of the TGF-β/SMAD and Ca signaling. This aligns with current EV-mediated therapies, which emphasize modulating multiple pathways for optimal outcomes.^8,51,52 Additionally, the enhanced efficacy of aPRP-EVs highlights how exercise-induced changes in EV content can increase therapeutic potential. Physical activity in donors enriches EVs with neuroprotective factors like TGF-β1 and facilitates downregulation of harmful factors such as NMDAR2B. This supports evidence that EV bioactivity can be influenced by the donor’s physiological state.^14,53

In another study, the protective role of EVs released after exercise against cerebral ischemia was investigated. Plasma EVs released early after stimulation improved the survival of vascular endothelial cells exposed to OGD. ⁵⁴ Furthermore, EVs released after exercise accumulated in the ischemic area of a mouse model of stroke, and a trend toward reduced infarct volume was observed. The main difference between our study and theirs is the source of EVs: PRP in our study and plasma in theirs.

Collectively, this study presents original findings as follows: First, while most studies have used EVs from mesenchymal stem cells, we focused on PRP, which offers advantages such as low cost, ease of purification, and the ability to be frozen and stored. Second, exercise further enhanced the therapeutic effects of PRP-derived EVs on stroke recovery. Third, we found that PRP-derived EVs regulated TGF-β/SMAD and Ca signaling pathways and enhanced axonal outgrowth and functional recovery after stroke. These mechanisms were verified using a TGF-β inhibitor and NMDAR agonist.

We chose to administer PRP-derived EVs at 7 days after MCAO based on our previous study ¹² and the temporal changes in neuronal integrin α_vβ₃ observed in the present study. At this time, integrin α_vβ₃ was highly expressed in ischemic neurons in the peri-infarct area. Additionally, a previous study showed that integrin α_vβ₃ on vascular endothelial cells peaked 7 days after MCAO. ³³ CD31, shown as one of the ligands for integrin α_vβ₃, ³³ suggests that intravenous administration of PRP-derived EVs at 7 days after MCAO may be optimal. However, earlier therapeutic time points for PRP-derived EVs may also need evaluation.

This study had several limitations. First, PRP-derived EVs may exhibit heterogeneity due to their ability to reflect changes in gene expression influenced by environmental factors such as circadian rhythms and stress. Ensuring uniform exercise intensity across all animals in the models is also challenging, which could contribute to variability. Additionally, stress and anxiety reduced by treadmill exercise were not quantified. Electrical stimulation during treadmill exercise can alter receptor expression, ⁵⁵ and its impact on animals that underwent exercise and PRP collection is difficult to assess. Alternative methods for reproducing exercise-induced EV changes are needed to translate exercise-loaded PRP-derived EVs from the bench to the bedside,. Second, the neurological and motor assessments were limited to the mNSS and rotarod tests, highlighting the need for more comprehensive behavioral evaluations. Third, the timing and dosage of PRP-derived EVs administration were fixed at 7 days after MCAO and 100 µg, respectively, based on this study and previous reports. This fixed approach may obscure potential differences in therapeutic effects at different timings or dosages. ²⁶ Although microarray analysis suggested the involvement of TGF-β and neuroinflammation signaling pathways, the precise molecular mechanisms, including the role of microRNAs, remain unclear. Fourth, the restorative effects observed may not be solely due to direct impacts on brain cells, as they may also involve indirect effects mediated through other organs. Therefore, further investigation into EV release, transport, and targeting at the single-vesicle level using advanced imaging techniques is necessary. Fifth, the effect of aging must be considered. Cerebral infarction is more common in older adults, and the effects of exercise and the composition of PRP-derived EVs may change with age. It is also possible that the effects of PRP-derived EVs after cerebral infarction vary with age; therefore, the impact of aging must be taken into account. Further study is warranted to account for these limitations.

In conclusion, this study provides strong evidence that PRP-derived EVs, especially those from exercise-loaded donors, offer a promising therapeutic approach for chronic ischemic stroke. The regulation of TGF-β/SMAD and Ca signaling is a key mechanism underlying their efficacy, suggesting the potential of integrating exercise to optimize PRP-derived EV-based therapies. Future research should focus on elucidating these molecular mechanisms and exploring their clinical potential.

Data availability statement

Raw data were generated at Juntendo University School of Medicine. The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.