Platelet-rich plasma-primed bone marrow mesenchymal stem cell-derived exosomes inhibit neuronal apoptosis and autophagy, and promote nerve regeneration via the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis after spinal cord injury Authors: Yao Yu, Man Fang, Cailiang Shen Published: 2026 ABSTRACT AbstractBoth platelet-rich plasma (PRP) and exosomes derived from bone marrow mesenchymal stem cells (BMSCs) exhibit therapeutic potential for spinal cord injury (SCI). However, the combined application of PRP and BMSCs-derived exosome (PRP-primed BMSCs-EXO, PRP-BMSCs-EXO) has not been reported. This study aimed to investigate whether the combined use confers superior therapeutic effects and to explore the underlying molecular mechanisms. First, we examined the effects of PRP priming on the proliferation and exosome secretion of BMSCs. High-throughput sequencing and RT-qPCR were used to analyze the differential miRNA expression profiles between exosomes from PRP- primed BMSCs (PRP-BMSCs-EXO) and naive BMSCs (BMSCs-EXO). Subsequently, a series of in vitro SCI models were established to evaluate the therapeutic effects of PRP-BMSCs-EXO on SCI. TargetScan, RNAhybrid, GO, KEGG, dual-luciferase reporter assay, RT-qPCR, western blotting (WB), and rescue experiments were performed to explore the roles of miRNAs in PRP-BMSCs-EXO and the associated signaling regulatory mechanisms in SCI. Finally, a series of in vivo SCI models were used to verify the therapeutic efficacy and mechanisms of PRP-BMSCs-EXO in vivo. PRP priming promoted the proliferation and exosome secretion of BMSCs. High-throughput sequencing and RT-qPCR revealed that multiple miRNAs were highly expressed in PRP-BMSCs-EXO compared with BMSCs-EXO, among which miR-29a-3p was the most abundant. In in vitro SCI models, PRP-BMSCs-EXO exerted better therapeutic effects than BMSCs-EXO and PRP. Mechanistically, PRP-BMSCs-EXO may enhance the viability and migration and inhibit the apoptosis and autophagy of pheochromocytoma 12 (PC12) cells by delivering miR-29a-3p, thereby promoting nerve regeneration and SCI repair. Bioinformatic analysis and luciferase assays indicated that PTEN may be a downstream target gene of miR-29a-3p. Rescue experiments confirmed that the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis may mediate the regulatory effects of PRP-BMSCs-EXO on neuronal apoptosis, autophagy, and nerve regeneration in SCI. These findings were further validated in in vivo SCI models. Compared with PRP and BMSC-derived exosomes alone, PRP-primed BMSC-derived exosomes provide superior neuroprotective effects against SCI. PRP-BMSCs-EXO inhibits neuronal apoptosis and autophagy and promotes nerve regeneration via the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis, representing a promising and effective therapeutic strategy for SCI.Keywords: Platelet-rich plasma, Exosomes, Spinal cord injuries, Apoptosis, Autophagy, miR-29a-3p/PTEN axis Introduction Spinal cord injury (SCI) is a severe central nervous system trauma with an annual incidence of approximately 8.9–14.6 per 100,000 people (“Global, regional, and national burden of spinal cord injury, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019,” 2023). It is typically caused by violent factors such as traffic accidents and falls from height, leading to various physical deficits, including permanent or progressive disabilities in motor, sensory, and autonomic functions (Li et al. 2024). This imposes a substantial economic burden on patients, their families, and society. The pathophysiology of SCI involves two distinct processes: primary injury, which is the initial mechanical disruption of neural tissue and blood vessels caused by external force, and secondary injury, a cascade of events mediated by the release of pro-inflammatory cytokines, production of reactive oxygen species (ROS), generation of chemokines, and activation of the complement system (Chen et al. 2024; Fuller et al. 2024; Jiang et al. 2025; Liu et al. 2024; Qian et al. 2024; Shen et al. 2024; Su et al. 2024). These events ultimately lead to spinal cord barrier disruption, glial scar formation, and local microenvironmental disturbance (Shu et al. 2024; Yu et al. 2017; S. Zhu et al. 2024a, b). Currently, there is no cure for SCI that can fundamentally repair damaged neural tissue or restore lost spinal cord function. Therefore, the search for novel therapeutic strategies is urgently needed. Recently, numerous studies have demonstrated that mesenchymal stem cells (MSCs) hold great potential as an emerging therapeutic approach for SCI (Akhlaghpasand et al. 2025; Cofano et al. 2019; K. Wang et al. 2025a, b; Wen et al. 2025; Zhang et al. 2020). However, MSC-based therapies face several challenges, including undirected differentiation (Yeng et al. 2016), lung tissue accumulation (Li et al. 2010), tumorigenicity (Mohseny and Hogendoorn 2011), and immune rejection (An et al. 2025). Emerging evidence suggests that the therapeutic effects of MSCs may be attributed to their paracrine activity, releasing trophic factors to inhibit inflammation, provide neurotrophic support, and promote nerve regeneration, rather than differentiating to replace damaged neural cells (Liau et al. 2020; Maldonado-Lasunción et al. 2018). Exosomes are nanoscale endosomal vesicles containing nucleic acids, lipids, and proteins that play crucial roles in cell differentiation, signal transduction, tissue repair, and regeneration (Cai et al. 2025; Khalatbary 2021). Compared with stem cells, exosomes offer advantages such as lower tumorigenicity (Wang et al. 2025a, b), higher safety, better stability, longer half-life, smaller diameter, and easier blood-brain barrier penetration (Cai et al. 2025; Khalatbary 2021; Ren et al. 2020). They can also be used as drug delivery vehicles (He et al. 2018; Li et al. 2018; Yom-Tov et al. 2022). MSCs-derived exosomes have been reported to exert anti-inflammatory and anti-apoptotic effects similar to their parent cells, promote neurovascularization (An et al. 2025; Li et al. 2022) and axonal regeneration (Huang et al. 2020; Sun et al. 2024; Wang et al. 2019, 2026), repair damaged spinal cord tissue structure, and treat SCI. Platelet-rich plasma (PRP) is a blood concentrate rich in growth factors, such as platelet-derived growth factor (PDGF) (Li et al. 2025), transforming growth factor-β (TGF-β) (Rath et al. 2025), vascular endothelial growth factor (VEGF) (Li et al. 2025), basic fibroblast growth factor (bFGF) (Cecerska-Heryć et al. 2022), insulin-like growth factor-1 (IGF-1) (Cecerska-Heryć et al. 2022), hepatocyte growth factor (HGF) (Cecerska-Heryć et al. 2022), epidermal growth factor (EGF) (Kia et al. 2018; Li et al. 2025; Marx 2004), nerve growth factor (NGF) (Farid et al. 2022), and brain-derived neurotrophic factor (BDNF) (Lai et al. 2018; Valente et al. 2022). These factors play important roles in tissue regeneration, cell proliferation, angiogenesis, and neuroprotection (Cecerska-Heryć et al. 2022; Kia et al. 2018). Studies have shown that PRP also contains various miRNAs, such as miRNA-21-3p (Babian et al. 2022), miRNA-155 (Vladulescu et al. 2023), and miRNA-223 (Li et al. 2019), and the miRNA content in PRP is more abundant compared with normal plasma (Cheng et al. 2013; McAlexander et al. 2013; Muth et al. 2018) . MiRNAs are short non-coding RNAs of approximately 22 nucleotides that bind to the 3’-untranslated region (3’-UTR) of target messenger RNAs (mRNAs), negatively regulating their expression through mRNA degradation or translational repression (Chekulaeva 2023). Studies have demonstrated that miRNAs are involved in regulating cell proliferation and apoptosis, promoting angiogenesis (Mani et al. 2024), tissue differentiation (Fu et al. 2024), and tissue regeneration (Pan et al. 2022; Yu et al. 2024a, b; Zhao et al. 2024). MiRNAs participate in multiple processes of SCI, such as the proliferation, migration, and differentiation of neural stem cells and progenitors (Genini et al. 2013, 2014; McNeill and Van Vactor 2012), and the expression of some miRNAs is downregulated after SCI (Hörauf et al. 2025). Due to their poor stability and susceptibility to RNase hydrolysis, miRNAs must be delivered via effective carrier systems. Exosomes, with their phospholipid bilayer membrane, can protect their internal cargo and safely deliver miRNAs to target cells (Yáñez-Mó et al. 2015) . Given the existing evidence that both MSC-derived exosomes and PRP have therapeutic effects on SCI but cannot achieve a complete cure, a combination of multiple approaches may improve efficacy. Therefore, in this study, we used hydrogen peroxide (H₂O₂) (300 µM) to induce PC12 cell injury to simulate SCI in vitro (Luo et al. 2019; Su et al. 2025), and established a rat SCI model in vivo using the modified Allen’s method (Cao et al. 2017; Kjell and Olson 2016). We examined the effects of PRP-BMSCs-EXO, BMSCs-EXO, and PRP on neuronal apoptosis, autophagy, and regeneration after SCI, and explored the underlying molecular mechanisms. Materials and methods Isolation and culture of rat bone marrow mesenchymal stem cells (BMSCs) All animal experiments were conducted in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH). The animal care and experimental protocols were approved by the Ethics Committee of Anhui Medical University (LLSC 20211493). All experimental rats were housed individually in cages under a controlled 12-hour light/dark cycle with free access to standard chow and water. Adult male Sprague-Dawley rats (200–250 g), purchased from Zhongkang Biotechnology (Suzhou) Co., Ltd., were euthanized by an overdose of 10% chloral hydrate (10–15 mL/kg). Following sterile preparation, the skin was incised, muscles were dissected, and the tibiae were isolated. After washing with phosphate-buffered saline (PBS, Cytiva, Uppsala, Sweden), the bone marrow cavity was flushed with Rat Primary Bone Marrow Mesenchymal Stem Cell Basal Medium (iCell, Shanghai, China) to collect the bone marrow aspirate into centrifuge tubes. A single-cell suspension was obtained after two rounds of centrifugation (1200 rpm, 5 min). The cell suspension was then seeded into T25 flasks (Nest, Wuxi, Jiangsu, China) and cultured at 37 °C in a humidified atmosphere containing 5% CO2. The culture medium was refreshed every 48 h. When the cells reached approximately 80% confluence and exhibited a characteristic vortex or radial arrangement (typically after 5–7 days), they were washed with PBS, detached using trypsin (Beyotime, Shanghai, China), and the digestion was terminated with fetal bovine serum (FBS, Gibco, Waltham, Massachusetts, USA). After centrifugation (1000 rpm, 4 min), the cells were passaged at a ratio of 1:2. Cell purity was significantly improved after the third passage (P3), indicating successful propagation. All BMSCs used in this study were at passage 3. Flow cytometry Approximately 1 × 106 rat BMSCs were resuspended in flow cytometry staining buffer and incubated with the following monoclonal antibodies: CD34-FITC (Santa Cruz, Dallas, Texas, USA), CD44-APC (Santa Cruz, Dallas, Texas, USA), CD45-PE (BioLegend, San Diego, California, USA), and CD90-FITC (BioLegend, San Diego, California, USA). After washing, the cells were resuspended in PBS and analyzed using a flow cytometer (BECKMAN, Suzhou, Jiangsu, China). Data acquisition and analysis were performed using FlowJo™ V10 software (BD Life Sciences, Franklin Lakes, New Jersey, USA). Adipogenic differentiation of BMSCs BMSCs were seeded in 6-well plates (Nest, Wuxi, Jiangsu, China) at a density of 1 × 105 cells/cm2 and incubated at 37 °C with 5% CO2. When the cells reached 80–100% confluence, the medium was replaced with a complete adipogenic induction medium (iCell, Shanghai, China). The cells were induced for 3 weeks at 37 °C. After induction, the medium was discarded, and the cells were rinsed with PBS, fixed with 4% paraformaldehyde (Ebiogo, Shanghai, China) for 30 min at room temperature, and rinsed again with PBS. The cells were then stained with a prepared Oil Red O working solution (iCell, Shanghai, China) for 30 min at room temperature. The staining solution was removed, and the cells were rinsed with PBS before observation under a light microscope. Osteogenic differentiation of BMSCs BMSCs were seeded in 6-well plates at a density of 2 × 104 cells/cm2 and cultured at 37 °C with 5% CO2. Upon reaching 80–100% confluence, the medium was replaced with a complete osteogenic induction medium (Pricella, Wuhan, Hubei, China), which was refreshed every 2–3 days for 3 weeks. The cells were then fixed with 4% paraformaldehyde for 30 min, rinsed with PBS, and stained with Alizarin Red S staining solution (Pricella, Wuhan, Hubei, China) for 30 min at room temperature. After washing with PBS, the formation of calcium nodules was observed under a light microscope. Chondrogenic differentiation of BMSCs Approximately 3 × 105BMSCs were pelleted by centrifugation (250 × g, 4 min, 20 °C) in a 15 mL conical tube. The supernatant was aspirated, and the cell pellet was resuspended in a complete chondrogenic induction medium (Pricella, Wuhan, Hubei, China). The tube was centrifuged again (150 × g, 5 min, 20 °C), the supernatant was removed, and fresh chondrogenic medium was added. After a final centrifugation (150 × g, 5 min, 20 °C), the supernatant was not discarded, and the tube cap was loosened to allow gas exchange. The cell pellets (micromass cultures) were incubated at 37 °C with 5% CO2 for 3 weeks, with medium changes every 2–3 days. The resulting chondrogenic pellets were fixed with 4% paraformaldehyde, embedded in paraffin, sectioned, deparaffinized, and stained with Alcian Blue staining solution (Pricella, Wuhan, Hubei, China) for 30 min to visualize glycosaminoglycan production. Preparation of rat platelet-rich plasma (PRP) PRP was prepared from the abdominal aorta blood of SD rats using a two-step centrifugation method (Z. Yu et al. 2024a, b). Briefly, whole blood was collected into anticoagulant tubes and initially centrifuged at 1100 rpm for 7 min. The upper plasma layer, containing a small number of red blood cells, was collected and subjected to a second centrifugation at 2400 rpm for 10 min. The upper layer of platelet-poor plasma (PPP) was discarded, and the remaining pellet was designated as PRP. To prepare activated PRP, a mixture of thrombin (60 U/mL; Yuanye, Shanghai, China; Cat. No.: S10117) and CaCl2 (0.1 g/mL; Yuanye, Shanghai, China; Cat. No.: S24110) was prepared. PRP was activated by mixing it with the thrombin/CaCl2 mixture at a volume ratio of 9:1 (PRP: mixture), incubated overnight at 4 °C, and then centrifuged at 7500 rpm for 10 min. The resulting supernatant was collected as the PRP extract for further experiments. Isolation and purification of BMSCs-EXO and PRP-BMSCs-EXO BMSCs were seeded in 6-well plates at a density of 1 × 105cells/mL and cultured overnight at 37 °C with 5% CO2. The medium was then replaced with either standard medium containing 10% FBS or medium supplemented with 10% PRP extract for 24 h. After incubation, the cells were washed with PBS to remove residual PRP-derived exosomes, and the medium was replaced with exosome-depleted medium (pre-cleared by ultracentrifugation). The cells were cultured for an additional 24 h, and the conditioned medium was collected. Exosomes were isolated via sequential centrifugation at 4 °C: 300 × g for 10 min (twice, pellets discarded), 10,000 × g for 30 min (pellet discarded), followed by ultracentrifugation at 120,000 × g for 80 min. The exosome pellet was resuspended in PBS and purified by a second round of ultracentrifugation under identical conditions. The final exosome pellet was resuspended in PBS for downstream experiments. Exosome protein quantification (BCA assay) The protein concentration of exosome preparations was determined using the Enhanced BCA Protein Assay Kit (Beyotime, Shanghai, China) according to the manufacturer’s instructions. A standard curve was generated using BSA standards (0, 0.025, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5 mg/mL) in a 96-well plate (Nest, Wuxi, Jiangsu, China). Exosome samples and BCA working solution were added to the plate, which was then incubated at 37 °C for 30 min. The absorbance at 562 nm was measured using a microplate reader (Rayto, Shenzhen, Guangdong, China), and the protein concentration of the samples was calculated based on the standard curve. Characterization of BMSCs-EXO and PRP-BMSCs-EXO The morphology of BMSCs-EXO and PRP-BMSCs-EXO was observed by transmission electron microscopy (TEM, Hitachi, Tokyo, Japan). The particle size distribution of BMSCs-EXO and PRP-BMSCs-EXO was analyzed by nanoparticle tracking analysis (NTA, Malvern Panalytical, Shanghai, China). Western blotting was performed to identify exosomal surface markers, including CD9, CD63, CD81, and TSG101. Western blotting Samples were lysed with RIPA lysis buffer (Beyotime, Shanghai, China), followed by centrifugation at 12,000×g for 15 min. The supernatant was collected, and the total protein concentration was determined using a BCA protein assay kit. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, Burlington, MA, USA). The membranes were incubated overnight at 4 °C with primary antibodies against CD9 (1:500, Rabbit), CD63 (1:2000, Rabbit), CD81 (1:5000, Rabbit), TSG101 (1:3000, Rabbit), FDX1 (1:2000, Rabbit), LIAS (1:2000, Rabbit), HSP70 (1:10000, Rabbit), LC3B (1:1000, Rabbit), P62 (1:5000, Rabbit), Beclin 1 (1:5000, Rabbit), BAX (1:5000, Rabbit), BCL-2 (1:1000, Rabbit), Caspase-3 (1:5000, Rabbit), PTEN (1:1000, Rabbit), PI3K (1:1000, Rabbit), p-PI3K (1:1000, Rabbit), Akt (1:1000, Rabbit), p-Akt (1:1000, Rabbit), mTOR (1:2000, Rabbit), p-mTOR (1:1000, Rabbit), and GAPDH (1:5000, Mouse). Subsequently, the membranes were incubated with secondary antibodies (1:20000) at room temperature for 1.5 h. Protein bands were detected using an ECL luminescence kit (Biosharp, Hefei, Anhui, China), and band intensity was analyzed using Image J software (NIH, Bethesda, MD, USA). Exosomal miRNA high-throughput sequencing Total RNA was extracted from exosome pellets using the miRNeasy Micro Kit (Qiagen, Valencia, California, USA) according to the manufacturer’s protocol. Small RNA libraries were constructed using the NEBNext Multiplex Small RNA Library Prep Kit (NEB, Ipswich, Massachusetts, USA). After size selection (135–155 bp) on a 3% agarose gel, the libraries were quantified using the Kapa Library Quantification Kit (Kapa Biosystems, Wilmington, Massachusetts, USA) on an ABI 9700HT Fast Real-Time PCR system (Thermo Fisher Scientific, Waltham, Massachusetts, USA). Clusters were generated using the TruSeq SR Cluster Kit v3 (Illumina, San Diego, California, USA) at a concentration of 10 pM, followed by 50-cycle sequencing on a HiSeq system using the TruSeq SBS Kit (Illumina). Raw sequencing data were processed, and clean reads were aligned to the rat reference genome (rn7). Differentially expressed miRNAs (DE-miRNAs) between the BMSCs-EXO and PRP-BMSCs-EXO groups were identified using the DESeq2 package (adjusted P-value < 0.05, |log2 fold change| ≥ 1). Target genes of DE-miRNAs were predicted using TargetScan 8.0 (https://www.targetscan.org/vert_80/) and RNAhybrid (https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid/), and filtered based on their relevance to spinal cord injury, apoptosis, and autophagy. Functional enrichment analysis (GO and KEGG) was performed on the predicted target genes to elucidate their biological functions. Selected DE-miRNAs were validated by RT-qPCR. Quantitative real-time RT-PCR (RT-qPCR) Total RNA was extracted from cells or tissues using TRIzol reagent (Life Technologies, Waltham, Massachusetts, USA). cDNA was synthesized from RNA using the PrimeScript™ RT Reagent Kit with gDNA Eraser (TaKaRa, Kusatsu, Shiga Prefecture, Japan) according to the manufacturer’s instructions. RT-qPCR was performed on a real-time PCR system (Thermo Scientific, Waltham, Massachusetts, USA) using gene-specific primers. The relative expression levels of target genes were calculated using the 2-△△Ct method, with GAPDH as the internal reference gene. The primer sequences used in this study are listed in Table 1. Table 1RNA primers for RT-qPCRGeneAmplicon size (bp)Forward primer (5’→3’)Reverse primer (5’→3’)β-actin150CCCATCTATGAGGGTTACGCTTTAATGTCACGCACGATTTCU694CTCGCTTCGGCAGCACAAACGCTTCACGAATTTGCGTrno-miR-31a-5p66ACACTCCAGCTGGGAGGCAAGATGCTGGCATGGTGTCGTGGAGTCGrno-miR-31a-5p-RTCTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCAGCTArno-miR-125a-5p67ACACTCCAGCTGGGTCCCTGAGACCCTTTAATGGTGTCGTGGAGTCGrno-miR-125a-5p-RTCTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCACAGGrno-miR-199-3p66ACACTCCAGCTGGGACAGTAGTCTGCACATTGGTGTCGTGGAGTCGrno-miR-199a-3p-RTCTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGTAACCArno-miR-29a-3p66ACACTCCAGCTGGGTAGCACCATCTGAAATTGGTGTCGTGGAGTCGrno-miR-29a-3p-RTCTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGTAACCGPTEN154GACATCAAGAAGGTGAAGGACCTTCATGGTCTTCATCCTGGTGTT RNA primers for RT-qPCR In vitro experimental design Experiment 1: neuroprotective effect of PRP-BMSCs-EXO PC12 cells were randomly divided into five groups to evaluate the neuroprotective effect of PRP-BMSCs-EXO against H₂O₂-induced injury: Control group: Untreated PC12 cells. Control group: Untreated PC12 cells. H2O2 group: Cells were treated with serum-free medium for 24 h, followed by exposure to H2O2 for an additional 24 h. H2O2 group: Cells were treated with serum-free medium for 24 h, followed by exposure to H2O2 for an additional 24 h. PRP group: Cells were pretreated with PRP extract for 24 h, then exposed to H2O2 for 24 h. PRP group: Cells were pretreated with PRP extract for 24 h, then exposed to H2O2 for 24 h. BMSCs-EXO group: Cells were pretreated with BMSCs-EXO for 24 h, then exposed to H2O2 for 24 h. BMSCs-EXO group: Cells were pretreated with BMSCs-EXO for 24 h, then exposed to H2O2 for 24 h. PRP-BMSCs-EXO group: Cells were pretreated with PRP-BMSCs-EXO for 24 h, then exposed to H2O2 for 24 h. PRP-BMSCs-EXO group: Cells were pretreated with PRP-BMSCs-EXO for 24 h, then exposed to H2O2 for 24 h. Experiment 2: mechanistic investigation via the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis To verify that PRP-BMSCs-EXO exert their neuroprotective effect through the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis, PC12 cells were divided into six groups: Control group: Untreated PC12 cells. Control group: Untreated PC12 cells. H2O2 group: Cells were treated with serum-free medium for 48 h, followed by exposure to H2O2 for 24 h. H2O2 group: Cells were treated with serum-free medium for 48 h, followed by exposure to H2O2 for 24 h. miR-29a-3p-inhibitor-NC group: Cells were transfected with miR-29a-3p inhibitor negative control (NC) for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-inhibitor-NC group: Cells were transfected with miR-29a-3p inhibitor negative control (NC) for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-inhibitor group: Cells were transfected with miR-29a-3p inhibitor for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-inhibitor group: Cells were transfected with miR-29a-3p inhibitor for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-mimic-NC group: Cells were transfected with miR-29a-3p mimic NC for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-mimic-NC group: Cells were transfected with miR-29a-3p mimic NC for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-mimic group: Cells were transfected with miR-29a-3p mimic for 48 h, then treated with H2O2 for 24 h. miR-29a-3p-mimic group: Cells were transfected with miR-29a-3p mimic for 48 h, then treated with H2O2 for 24 h. Labeling of BMSCs-EXO and PRP-BMSCs-EXO BMSCs-EXO and PRP-BMSCs-EXO were labeled with the lipophilic fluorescent dye DiI (Molecular Probes, Eugene, Oregon, USA) according to the manufacturer’s instructions. Briefly, DiI stock solution (4 mg/mL) was added to the exosome suspension in PBS and incubated for 1 h at 37 °C. Unincorporated dye was removed by ultracentrifugation (100,000 × g, 1 h, 4 °C). The labeled exosome pellet was washed with PBS and resuspended for subsequent experiments. The uptake of labeled exosomes by PC12 cells was observed using a laser scanning confocal microscope (Carl Zeiss Microscopy GmbH, Jena, Germany). In vivo imaging of the rat spinal cord injury lesions was performed 1 week after injection of the two exosomes using an in vivo small animal optical 3D imaging system (IVIS Spectrum, PerkinElmer Inc., USA). Cell viability assay (CCK-8) PC12 cells in the logarithmic growth phase were seeded into 96-well plates (Nest, Wuxi, Jiangsu, China) at a density of 5 × 10³ cells per well and incubated overnight at 37 °C with 5% CO₂. Following the respective interventions, CCK-8 solution (Biosharp, Hefei, Anhui, China) was added to each well, and the plates were incubated for an additional 2 h at 37 °C. The absorbance at 450 nm was measured using a microplate reader (Rayto, Shenzhen, Guangdong, China). Cell viability was calculated using the following formula: where the experimental wells contained cells, medium, and treatment; control wells contained cells and medium without treatment; and blank wells contained only medium. Cell migration assays (Transwell and scratch wound healing) Transwell assay After treatment, PC12 cells were trypsinized and resuspended in serum-free medium at a concentration of 2 × 105cells/mL. A 200 µL aliquot of the cell suspension was added to the upper chamber of a Transwell insert (8 μm pore size, Corning, New York, USA). The lower chamber was filled with 600 µL of complete medium containing 10% FBS as a chemoattractant. After incubation for 24 h at 37 °C with 5% CO₂, non-migratory cells on the upper surface of the membrane were removed with a cotton swab. Cells that had migrated to the lower surface were fixed with 4% paraformaldehyde for 30 min and stained with 0.5% crystal violet (SIOISCO, Lianyungang, Jiangsu, China) for 30 min. Excess stain was washed away with PBS, and the stained cells were observed and counted under an inverted microscope (OLYMPUS, Shinjuku-ku, Tokyo, Japan) at 100× magnification. Scratch wound healing assay Parallel lines were drawn on the back of 6-well plates as a reference. After intervention, PC12 cells were seeded into the plates at a density of 5 × 105cells per well and incubated overnight at 37 °C with 5% CO₂. Once a confluent monolayer was formed, a uniform scratch was created across the well using a 200 µL pipette tip (Dragon, Beijing, China) perpendicular to the reference lines. Detached cells were washed away with PBS, and images of the scratch at 0 h (T0) were captured using an inverted microscope. Serum-free medium was then added, and cells were incubated for 24 h at 37 °C with 5% CO2. Images of the scratch at 24 h (T24) were captured again. The scratch closure area was measured using ImageJ software (NIH, Bethesda, Maryland, USA). The migration rate was calculated as. Apoptosis assay (TUNEL) PC12 cells grown on coverslips after intervention were rinsed to remove the culture medium, fixed with 4% paraformaldehyde at room temperature for 20 min, and washed with PBS (paraffin sections of rat SCI tissue were deparaffinized to water and washed with distilled water). A hydrophobic barrier was drawn around the sample using a PAP Pen (Ebiogo, Shanghai, China). Proteinase K without DNase (Ebiogo, Shanghai, China) was added and incubated at 37 °C for 30 min, followed by PBS washing. TUNEL detection solution (Elabscience, Wuhan, Hubei, China) was prepared according to the manufacturer’s instructions, applied to the sample, and incubated at 37 °C for 60 min in the dark. After PBS washing, coverslips were mounted using an anti-fade mounting medium containing DAPI (Beyotime, Shanghai, China). Fluorescent sections were scanned using a digital slide scanner (3DHISTECH, Budapest, Hungary), and the mean optical density (MOD) value was measured. Immunofluorescence staining (BrdU, NeuN) PC12 cells on coverslips after intervention were fixed with 4% paraformaldehyde for 30 min and washed with PBS. Cells were permeabilized with 0.3% Triton X‑100 (Ebiogo, Shanghai, China) for 20 min and washed with PBS (paraffin sections of rat SCI tissue were deparaffinized to water, washed with distilled water, and subjected to antigen retrieval using citrate antigen retrieval solution at pH 6.0 (Zhongshan Golden Bridge, Beijing, China)). A hydrophobic barrier was drawn around the sample. Endogenous peroxidases were quenched with 3% hydrogen peroxide solution (Beyotime, Shanghai, China). Non‑specific binding sites were blocked with goat serum blocking solution (Thermo Fisher Scientific, Waltham, Massachusetts, USA). Primary antibodies were applied and incubated at 37 °C in the dark for 60 min (NeuN: Abcam, ab104224, mouse, 1:200; BrdU: Santa Cruz, sc‑32323, mouse, 1:300). After PBS washing, ready‑to‑use HRP‑conjugated anti‑mouse/rabbit secondary antibodies (Maixin, Fuzhou, Fujian, China) were applied and incubated at room temperature in the dark for 30 min. Following PBS washing, TSA fluorescent dye (Zhongkang, Suzhou, Jiangsu, China) was applied and incubated at room temperature in the dark for 10 min and washed with PBS. Coverslips were mounted with anti‑fade mounting medium containing DAPI (Beyotime, Shanghai, China), allowed to air‑dry overnight in the dark at room temperature, and then observed and photographed under a fluorescence microscope. miRNA transfection PC12 cells were seeded into 6-well plates and incubated for 48 h with miR-29a-3p mimic, miR-29a-3p mimic negative control (NC), miR-29a-3p inhibitor, or miR-29a-3p inhibitor negative control (NC). All transfections were performed using a transfection kit (Hanheng Biotechnology, Shanghai, China). After incubation, cells were collected for subsequent experiments. Dual-luciferase reporter assay The wild-type (WT) and mutant (Mut) PTEN 3’-UTR sequences containing the complementary binding site for rno-miR-29a-3p were cloned into the pmirGLO vector (Promega, Madison, Wisconsin, USA) to construct PTEN-WT-pmirGLO and PTEN-Mut-pmirGLO plasmids. HEK293T cells were co-transfected using a transfection kit with either rno-miR-29a-3p-mimic (or miR-29a-3p-mimic-NC) and either PTEN-WT-pmirGLO (or PTEN-Mut-pmirGLO) plasmid for 48 h. Luciferase activity was measured using the Promega Dual-Luciferase system (Promega, Madison, Wisconsin, USA) following the manufacturer’s protocol. Relative luciferase activity was calculated as the ratio of firefly luciferase activity (F) to Renilla luciferase activity (R). In vivo experimental design Spinal cord injury (SCI) model and treatment protocol The Allen’s weight-drop model was used to induce SCI in SD rats. After anesthesia, the skin was incised, muscles and ligaments were dissected, and the T10 lamina was removed to expose the spinal cord. A metal impactor rod was centered on the spinal cord and dropped from a height of 25 mm with a 20 g weight. Successful modeling was indicated by visible contusion at the impact site, twitching of both hindlimbs, and spasmodic tail movement. Postoperatively, penicillin was administered for 3 days, and manual bladder expression was performed three times daily until spontaneous urination recovered. Twenty-five SD rats were randomly divided into five groups (n = 5): Sham group: Underwent laminectomy only, no impact, followed by tail vein injection of 200 µL PBS immediately after surgery. Sham group: Underwent laminectomy only, no impact, followed by tail vein injection of 200 µL PBS immediately after surgery. SCI group: Received SCI and was injected with 200 µL PBS immediately after injury. SCI group: Received SCI and was injected with 200 µL PBS immediately after injury. PRP group: Received SCI and was injected with 200 µL PRP immediately after injury. PRP group: Received SCI and was injected with 200 µL PRP immediately after injury. BMSCs-EXO group: Received SCI and was injected with 200 µL BMSCs-EXO (200 µg) immediately after injury. BMSCs-EXO group: Received SCI and was injected with 200 µL BMSCs-EXO (200 µg) immediately after injury. PRP-BMSCs-EXO group: Received SCI and was injected with 200 µL PRP-BMSCs-EXO (200 µg) immediately after injury. PRP-BMSCs-EXO group: Received SCI and was injected with 200 µL PRP-BMSCs-EXO (200 µg) immediately after injury. Additional interventions (200 µL per dose) were administered via tail vein injection on postoperative days 1, 3, 5, and 7. Behavioral assessments Basso, Beattie, and Bresnahan (BBB) locomotor rating scale Rats were placed on a smooth, open plastic surface and observed for 10 min per rat. Two trained observers scored the rats independently, and the average score was used to minimize subjective bias. In cases of significant discrepancy, the rat was re-observed and a consensus score was reached. The scale ranges from 0 (complete paralysis) to 21 (normal locomotion), with detailed criteria based on hindlimb joint movement (hip, knee, ankle), paw placement, weight-bearing capacity, coordination, and trunk stability. Inclined plane test (Rivlin method) (Liu et al. 2023) The test was performed using a smooth, slightly non-slip wooden board (60 cm long, 40 cm wide) attached to a frame allowing adjustable angles from 0° to 90° with 1° precision, equipped with an angle dial. In a quiet, softly lit environment at 25 °C, rats were first allowed to acclimate on the board at 0° for 3 min. The rat was then placed in the center of the board with limbs extended and trunk parallel to the board. The angle was increased slowly (in 5° increments). The maximum angle at which the rat could maintain its position without slipping or limb sliding for 5 s was recorded. To minimize error, each rat was tested three times, and the average of the three trials was taken as the final maximum angle. Histological analysis (Hematoxylin and eosin, H&E staining) Twenty-eight days post-SCI, rats were euthanized via anesthetic overdose. The injured spinal cord tissue was collected, fixed in 4% paraformaldehyde, dehydrated, cleared, embedded in paraffin, and sectioned. Sections were deparaffinized, rehydrated, stained with hematoxylin (BASO, Zhuhai, Guangdong, China) for 5 min, differentiated and blued, then counterstained with eosin (BASO, Zhuhai, Guangdong, China) for 5 min. After dehydration, clearing, and mounting, sections were observed under a light microscope. Statistical analysis Statistical analysis and data visualization were performed using GraphPad Prism 6.0 software (GraphPad Software, LLC, San Diego, California, USA). Data are presented as mean ± standard deviation (SD). Differences between groups were analyzed using two-tailed Student’s t-test or one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. A P-value < 0.05 was considered statistically significant. Results Platelet-rich plasma (PRP) enhances the proliferation and exosome secretion of rat bone marrow mesenchymal stem cells (BMSCs) BMSCs were successfully isolated from the bone marrow of rat tibiae. At passage 3 (P3), the rat BMSCs exhibited a typical spindle-shaped morphology with plump cell bodies, uniform cytoplasm, and clear nuclei, arranged in a vortex or radial pattern, without obvious epithelial, adipogenic, or osteogenic differentiation (Fig. 1a). Flow cytometry analysis confirmed that these cells highly expressed the mesenchymal stem cell markers CD44 (100%) (Fig. 1e) and CD90 (98.3%) (Fig. 1g), while expressing very low levels of the hematopoietic stem cell markers CD34 (0.71%) (Fig. 1d) and CD45 (1.86%) (Fig. 1f). Characterization of rat BMSCs.a Morphology of P3 BMSCs. b–c Chondrogenic differentiation assessed by Alcian Blue staining. d–g Flow cytometry analysis of BMSC surface markers (CD34, CD44, CD45, CD90). h–j Adipogenic differentiation assessed by Oil Red O staining. k–m Osteogenic differentiation assessed by Alizarin Red S staining. (BMSCs = bone marrow mesenchymal stem cells; P3 = passage 3) After 3 weeks of induction, adipogenic differentiation was confirmed by Oil Red O staining, which revealed intracellular lipid droplets stained orange-red (Fig. 1 h–j). Osteogenic differentiation was verified by Alizarin Red S staining, showing the formation of red-yellow mineralized nodules (Fig. 1k–m). Chondrogenic differentiation was identified by Alcian Blue staining, which stained the extracellular matrix of the chondrogenic pellets bright blue (Fig. 1b–c). PRP solution was successfully prepared using a two-step centrifugation method, appearing as a pale yellow, high-purity solution with no contamination or precipitation (Fig. 2a). Compared with BMSCs cultured in medium containing 10% PBS, BMSCs incubated with medium supplemented with 10% PRP exhibited significantly enhanced proliferative capacity, as evidenced by a denser cell distribution and higher cell counts (Figs. 2b–d). Furthermore, the BCA assay confirmed that the protein concentration of exosomes isolated from the extracellular matrix of BMSCs cultured in PRP-supplemented medium was significantly higher, indicating a stronger exosome secretion capacity (Fig. 2e). PRP promotes BMSC proliferation and exosome secretion.a Successfully prepared PRP solution.b BMSCs cultured in medium containing 10% PBS. c. BMSCs cultured in medium containing 10% PRP. d Quantification of BMSC counts after incubation with the two types of medium. e Protein concentration of exosomes isolated from BMSCs cultured in the two types of medium, as determined by BCA assay. (PRP = platelet-rich plasma; PBS = phosphate-buffered saline; BCA = bicinchoninic acid assay; *P < 0.05, one-way ANOVA) Characterization of BMSCs-EXO and PRP-BMSCs-EXO and analysis of differentially expressed miRNAs (DE-miRNAs) Exosomes were successfully isolated from the extracellular matrix of BMSCs cultured in medium containing 10% PBS or 10% PRP using ultracentrifugation and designated as BMSCs-EXO and PRP-BMSCs-EXO, respectively. Transmission electron microscopy (TEM) revealed that both exhibited a typical round, double-membrane structure (Fig. 3a–b). Nanoparticle Tracking Analysis (NTA) showed that the average diameters of BMSCs-EXO and PRP-BMSCs-EXO were 67.3 ± 21.4 nm and 71.3 ± 24.1 nm, respectively (Fig. 3c–d). Western blotting (WB) confirmed that both exosome preparations significantly expressed the exosomal surface markers CD9, CD63, CD81, and TSG101 (Fig. 3e), verifying their identity as exosomes. Identification of extracellular matrix isolates from BMSCs cultured in medium containing 10% PBS or 10% PRP.a–b TEM images showing the morphology of exosomes (a: BMSCs-EXO, b: PRP-BMSCs-EXO). c–d NTA analysis of particle size distribution (c: BMSCs-EXO, d: PRP-BMSCs-EXO). e WB analysis showing the expression of exosomal surface proteins CD9, CD63, CD81, and TSG101 (A: BMSCs-EXO, B: PRP-BMSCs-EXO). (TEM = transmission electron microscopy; NTA = nanoparticle tracking analysis; WB = western blotting; EXO = exosomes) High-throughput sequencing was performed to analyze DE-miRNAs between the two exosome types. A total of 34 DE-miRNAs were identified. Compared with BMSCs-EXO, PRP-BMSCs-EXO exhibited upregulation of 20 miRNAs and downregulation of 14 miRNAs (Table 2). A volcano plot (Fig. 4a) and a hierarchical clustering heatmap (Fig. 4b) were generated to visualize the DE-miRNA expression profiles. The four most significantly upregulated miRNAs (rno-miR-125a-5p, rno-miR-199a-3p, rno-miR-29a-3p, rno-miR-31a-5p) were selected for validation by RT-qPCR. The results confirmed that rno-miR-29a-3p was the most significantly differentially expressed miRNA (P < 0.05) (Fig. 4c–f). Table 2Differentially expressed miRNA profilemiRNAFCLog2 (FC)p valueRegulationrno-miR-125a-5p80.0736.3236.877E-08Uprno-miR-199-3p15.5103.9550.006Uprno-miR-29a-3p14.5913.8670.008Uprno-miR-31a-5p14.0923.8160.003Uprno-miR-142-5p12.6553.6610.001Uprno-miR-128-3p10.6053.4066.053E-20Uprno-miR-30a-5p10.1153.3380.032Uprno-miR-25-3p9.6803.2750.0005852Uprno-miR-92a-3p7.1162.8310.0000109Uprno-miR-27a-3p5.3042.4072.683E-08Uprno-miR-22-5p4.6722.2240.013Uprno-miR-140-3p4.1192.0420.004Uprno-miR-26a-5p3.7931.9231.755E-20Uprno-miR-378a-3p3.6871.8820.013Uprno-miR-148a-3p3.6861.8828.585E-55Uprno-miR-30d-5p3.6521.8680.002Uprno-miR-24-3p3.6111.8526.756E-07Uprno-let-7i-5p2.9071.5392.825E-25Uprno-miR-191a-5p2.8461.5080.0000106Uprno-miR-23a-3p2.6381.3990.039Uprno-miR-423-3p0.497− 1.0060.03Downrno-miR-99a-5p0.444− 1.1697.446E-11Downrno-miR-328a-3p0.438− 1.1890.022Downrno-miR-6760.426− 1.2310.008Downrno-miR-455-5p0.246− 2.0190.048Downrno-miR-100-5p0.238− 2.0670.0000014Downrno-let-7c-5p0.227− 2.1345.144E-81Downrno-let-7b-5p0.223− 2.1625.589E-38Downrno-miR-122-5p0.222− 2.1652.607E-07Downrno-let-7e-5p0.207− 2.2670.001Downrno-miR-4860.136− 2.8680.001Downrno-miR-196a-5p0.092− 3.4360.0004Downrno-miR-125b-5p0.058− 4.0923.035E-15Downrno-miR-455-3p0.032− 4.9610.027Down Differentially expressed miRNA profile Analysis of DE-miRNAs between the two exosome types.a Volcano plot of DE-miRNAs. b Hierarchical clustering heatmap of DE-miRNAs. c–f RT-qPCR validation of the expression levels of four selected miRNAs. (DE-miRNAs = differentially expressed microRNAs; RT-qPCR = quantitative real-time reverse transcription polymerase chain reaction; *P < 0.05, one-way ANOVA) PRP-BMSCs-EXO protect spinal cord injury (SCI) function in vitro by enhancing neuronal viability and migration, inhibiting neuronal apoptosis and autophagy, and promoting neuronal regeneration After labeling BMSCs-EXO and PRP-BMSCs-EXO with DiI dye and co-incubating them with PC12 cells for 24 h, red fluorescence was observed in the cytoplasm of PC12 cells under a laser scanning confocal microscope, indicating successful uptake of both exosome types by PC12 cells (Fig. 5). Uptake of DiI-labeled BMSCs-EXO and PRP-BMSCs-EXO by PC12 cells.Laser confocal microscopy images of PC12 cells after exosome uptake (A: BMSCs-EXO, B: PRP-BMSCs-EXO). (DiI = 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate; EXO = exosomes) To investigate the protective effects of PRP-BMSCs-EXO, CCK-8, Transwell, and scratch wound healing assays were performed. The results showed that the neuronal viability and migration capacity in the PRP-BMSCs-EXO, PRP, and BMSCs-EXO groups were all significantly enhanced compared with the H2O2 group, with the PRP-BMSCs-EXO group showing the most pronounced enhancement (P < 0.05) (Fig. 6a–e). Effects of PRP-BMSCs-EXO on neuronal viability and migration capacity. a. Average cell viability determined by CCK-8 assay. b Average number of migrated cells determined by Transwell assay. c Average migration rate at 24 hours determined by scratch wound healing assay. d Microscopic images of the Transwell assay (×100 magnification). e Images of the scratch wound healing assay at 0 hours and 24 hours. (A: Control group, B: H₂O₂ group, C: PRP group, D: BMSCs-EXO group, E: PRP-BMSCs-EXO group; *P < 0.05, one-way ANOVA) TUNEL staining and immunofluorescence staining for BrdU and NeuN demonstrated that, compared with the H₂O₂ group, the PRP-BMSCs-EXO, PRP, and BMSCs-EXO groups had significantly fewer apoptotic neurons (TUNEL-positive), and significantly more surviving (NeuN-positive) and newly regenerated (BrdU-positive) neurons. Among these, the PRP-BMSCs-EXO group exhibited the least neuronal apoptosis and the most neuronal regeneration (P < 0.05) (Fig. 7a–f). Effects of PRP-BMSCs-EXO on neuronal apoptosis and regeneration.a MOD values of TUNEL staining. b MOD values of BrdU staining. c MOD values of NeuN staining. d TUNEL staining images. e BrdU staining images. f NeuN staining images. (A: Control group, B: H2O2 group, C: PRP group, D: BMSCs-EXO group, E: PRP-BMSCs-EXO group; MOD = mean optical density; TUNEL = terminal deoxynucleotidyl transferase dUTP nick end labeling; *P < 0.05, one-way ANOVA) WB analysis of apoptosis and autophagy markers revealed that, compared with the H2O2 group, the PRP-BMSCs-EXO, PRP, and BMSCs-EXO groups showed decreased expression of Bax and Caspase-3, increased expression of Bcl-2, decreased expression of Beclin-1 and the LC3B-II/LC3B-I ratio, and increased expression of P62. The PRP-BMSCs-EXO group displayed the most significant changes (P < 0.05) (Fig. 8a–h). Effects of PRP-BMSCs-EXO on the expression of neuronal apoptosis and autophagy markers.a–c Relative expression levels of Bax, Caspase-3, and Bcl-2. d–f Relative expression levels of Beclin-1, LC3B-II/LC3B-I ratio, and P62. g–h Representative WB bands for apoptosis and autophagy markers. (A: Control group, B: H2O2 group, C: PRP group, D: BMSCs-EXO group, E: PRP-BMSCs-EXO group; WB = western blotting; *P < 0.05, one-way ANOVA) To explore the mechanism underlying the neuroprotective effects of PRP-BMSCs-EXO, RT-qPCR was used to detect the expression of rno-miR-29a-3p. The expression of rno-miR-29a-3p was significantly higher in the PRP-BMSCs-EXO, PRP, and BMSCs-EXO groups compared with the H2O2 group, with the highest expression observed in the PRP-BMSCs-EXO group (P < 0.05) (Fig. 9). These results suggest that the neuroprotective effects of PRP-BMSCs-EXO may be associated with miR-29a-3p. Effect of PRP-BMSCs-EXO on the expression of miR-29a-3p in neurons. (miR-29a-3p = microRNA-29a-3p; *P < 0.05, one-way ANOVA) Prediction, validation, and enrichment analysis of miR-29a-3p target genes To investigate the potential regulatory functions of miR-29a-3p, its target genes were predicted using two online tools, TargetScan 8.0 and RNAhybrid. The results suggested that PTEN might be a direct target gene of miR-29a-3p, as its 3’-UTR sequence contains a potential binding site complementary to miR-29a-3p (Fig. 10a). Prediction, validation, and functional analysis of miR-29a-3p target genes.a Predicted binding sequence between rno-miR-29a-3p and Pten by RNAhybrid. b Relative luciferase activity (Rluc/Fluc) measured by dual-luciferase reporter assay. c Schematic diagram of the binding sites between rno-miR-29a-3p and rno-PTEN. d. GO bubble plot. e KEGG scatter plot. (miR-29a-3p = microRNA-29a-3p; PTEN = phosphatase and tensin homolog; GO = Gene Ontology; KEGG = Kyoto Encyclopedia of Genes and Genomes; *P < 0.05, one-way ANOVA) A dual-luciferase reporter assay was performed to validate this interaction. The results showed that rno-miR-29a-3p significantly downregulated the luciferase activity of the rno-PTEN-wt reporter vector compared with the negative control (NC) group (P < 0.05). However, rno-miR-29a-3p failed to downregulate the luciferase activity of the mutated rno-PTEN-mut reporter vector (P > 0.05), confirming that PTEN is a direct target gene of miR-29a-3p (Fig. 10b–c). Further Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses indicated that the target genes were associated with biological processes such as “spinal cord development,” “apoptosis,” and “autophagy.” A GO bubble plot (Fig. 10d) and a KEGG bar chart (Fig. 10e) were generated to visualize these enrichments. PRP-BMSCs-EXO exert neuroprotective effects on SCI by shuttling miR-29a-3p to inhibit PTEN and activate the PI3K/Akt/mTOR pathway To verify that PRP-BMSCs-EXO exert neuroprotective effects by delivering miR-29a-3p, PC12 cells were transfected with rno-miR-29a-3p-mimic, rno-miR-29a-3p-mimic-NC, rno-miR-29a-3p-inhibitor, or rno-miR-29a-3p-inhibitor-NC. CCK-8, Transwell, and scratch wound healing assays showed that cell viability and migration capacity were significantly better in the rno-miR-29a-3p-mimic group compared with the rno-miR-29a-3p-mimic-NC, rno-miR-29a-3p-inhibitor-NC, and H2O2 groups, while the rno-miR-29a-3p-inhibitor group exhibited the poorest outcomes (P < 0.05) (Fig. 11a–c). Validation of the protective effect of the miR-29a-3p/PTEN/ PI3K/Akt/mTOR axis on SCI in miRNA-transfected PC12 cells.a-1 Average cell viability after transfection determined by CCK-8 assay. a-2 Average number of migrated cells after transfection determined by Transwell assay. a-3 Average migration rate at 24 hours after transfection determined by scratch wound healing assay. b Microscopic images of the Transwell assay after transfection (×100 magnification). c Images of the scratch wound healing assay at 0 hours and 24 hours after transfection. d-1 MOD values of TUNEL staining after transfection. d-2 MOD values of BrdU staining after transfection. d-3 MOD values of NeuN staining after transfection. e-1 TUNEL staining images after transfection. e-2 BrdU staining images after transfection. e-3 NeuN staining images after transfection. f Relative expression levels of Bax, Caspase-3, and Bcl-2 after transfection. g Relative expression levels of Beclin-1, LC3B-II/LC3B-I ratio, and P62 after transfection. h Representative WB bands for apoptosis and autophagy markers after transfection. i–j Expression levels of miR-29a-3p, PTEN, PI3K/p-PI3K, Akt/p-Akt, and mTOR/p-mTOR after transfection. (A: Control group, B: H2O2 group, C: rno-miR-29a-3p-mimic-NC group, D: rno-miR-29a-3p-mimic group, E: rno-miR-29a-3p-inhibitor-NC group, F: rno-miR-29a-3p-inhibitor group). (A: Control group, B: H₂O₂ group, C: rno-miR-29a-3p-mimic-NC group, D: rno-miR-29a-3p-mimic group, E: rno-miR-29a-3p-inhibitor-NC group, F: rno-miR-29a-3p-inhibitor group; SCI = spinal cord injury; PI3K = Phosphatidylinositol 3-kinase; Akt = Protein kinase B; mTOR = mammalian target of rapamycin; *P < 0.05, one-way ANOVA) TUNEL staining and immunofluorescence staining for BrdU and NeuN demonstrated that the rno-miR-29a-3p-mimic group had fewer TUNEL-positive apoptotic neurons than the rno-miR-29a-3p-mimic-NC, H2O2, and rno-miR-29a-3p-inhibitor-NC groups, and significantly fewer than the rno-miR-29a-3p-inhibitor group (P < 0.05). Conversely, the rno-miR-29a-3p-mimic group had more BrdU-positive and NeuN-positive neurons than the other groups, with the rno-miR-29a-3p-inhibitor group having the fewest (P < 0.05) (Fig. 11d–e). WB analysis of apoptosis and autophagy markers showed that the rno-miR-29a-3p-mimic group exhibited lower expression of Bax, Caspase-3, Beclin-1, and the LC3B-II/LC3B-I ratio, and higher expression of Bcl-2 and P62 compared with the rno-miR-29a-3p-mimic-NC, H2O2, rno-miR-29a-3p-inhibitor-NC, and especially the rno-miR-29a-3p-inhibitor groups (P < 0.05) (Fig. 11f–h). RT-qPCR and WB analyses of miR-29a-3p, PTEN, PI3K/p-PI3K, Akt/p-Akt, and mTOR/p-mTOR revealed that the rno-miR-29a-3p-mimic group had significantly higher expression of miR-29a-3p, p-PI3K, p-Akt, and p-mTOR, and lower expression of PTEN compared with the other groups (P < 0.05) (Fig. 11i–j). Collectively, these results suggest that PRP-BMSCs-EXO may improve neuronal viability and migration, inhibit neuronal apoptosis and autophagy, and promote neuronal regeneration by shuttling miR-29a-3p to inhibit PTEN and activate the PI3K/Akt/mTOR pathway, thereby exerting neuroprotective effects on SCI. PRP-BMSCs-EXO promote functional recovery in rats with SCI via the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis After SCI induction and subsequent interventions, the neurological function of the rats gradually recovered. By 7 days post-injury, in vivo imaging showed that Dil-labeled BMSCs-EXO and PRP-BMSCs-EXO accumulated at the rat spinal cord injury lesions (Fig. 12a-1, 2). By 28 days post-injury, the Basso, Beattie, and Bresnahan (BBB) locomotor scores in the PRP-BMSCs-EXO group were significantly higher than those in the BMSCs-EXO, PRP, and SCI groups. The inclined plane test yielded consistent results, indicating that treatment with PRP-BMSCs-EXO achieved satisfactory outcomes for SCI (P < 0.05) (Fig. 12a-3, 4). Effects and mechanism of PRP-BMSCs-EXO on neurological function recovery in SCI rats.a-1~a-2 In vivo imaging (a-1: BMSCs-EXO group, a-2: PRP-BMSCs-EXO group); a-3 BBB scores. a-4 Inclined plane test angles. b Gross morphology and H&E staining of spinal cord tissue. c-1 MOD values of TUNEL staining in spinal cord tissue. c-2 MOD values of BrdU staining in spinal cord tissue. c-3 MOD values of NeuN staining in spinal cord tissue. d-1 TUNEL staining images of spinal cord tissue. d-2 BrdU staining images of spinal cord tissue. d-3. NeuN staining images of spinal cord tissue. e-1~ Relative expression levels of Bax, Caspase-3, and Bcl-2 in spinal cord tissue. e-4~e-6. Relative expression levels of Beclin-1, LC3B-II/LC3B-I ratio, and P62 in spinal cord tissue. f. Representative WB bands for apoptosis and autophagy markers in spinal cord tissue. g–h. Expression levels of miR-29a-3p, PTEN, PI3K/p-PI3K, Akt/p-Akt, and mTOR/p-mTOR in spinal cord tissue. (A: Sham group, B: SCI group, C: PRP group, D: BMSCs-EXO group, E: PRP-BMSCs-EXO group). (A: Sham group, B: SCI group, C: PRP group, D: BMSCs-EXO group, E: PRP-BMSCs-EXO group; BBB = Basso, Beattie, and Bresnahan; H&E = Hematoxylin and Eosin; *P < 0.05, one-way ANOVA) In terms of the gross morphology of injured spinal cord tissue in rats, the spinal cord structure at the injury site in the control group was normal with uniform thickness. In contrast, the spinal cord at the injury site in the SCI group was significantly atrophied. The spinal cords in the PRP, BMSCs-EXO, and PRP-BMSCs-EXO groups showed increased thickness compared to the SCI group, with the PRP-BMSCs-EXO group exhibiting the most pronounced thickening (Fig. 12b-1). Hematoxylin and eosin (H&E) staining of the injured spinal cord tissue showed normal spinal cord structure and intact, clear neuronal nuclei in the sham group. In contrast, the SCI group exhibited obvious hemorrhage, edema, inflammatory cell infiltration, and a significant reduction in the number of neurons at the injury site. The PRP, BMSCs-EXO, and PRP-BMSCs-EXO groups showed improvements in spinal cord tissue morphology, reduced hemorrhagic foci, and an increased number of neurons, with the PRP-BMSCs-EXO group showing the most significant improvements (Fig. 12b-2). TUNEL staining and immunofluorescence staining for BrdU and NeuN in spinal cord tissue revealed that the PRP-BMSCs-EXO group had significantly fewer TUNEL-positive cells and significantly more BrdU-positive and NeuN-positive neurons compared with the PRP, BMSCs-EXO, and SCI groups (P < 0.05) (Fig. 12c–d). WB analysis of apoptosis and autophagy markers demonstrated that the PRP-BMSCs-EXO group had lower expression of Bax, Caspase-3, Beclin-1, and the LC3B-II/LC3B-I ratio, and higher expression of Bcl-2 and P62 compared with the PRP, BMSCs-EXO, and SCI groups, indicating that PRP-BMSCs-EXO significantly inhibited neuronal apoptosis and autophagy in rats (P < 0.05) (Fig. 12e–f). RT-qPCR and WB analyses of miR-29a-3p, PTEN, PI3K/p-PI3K, Akt/p-Akt, and mTOR/p-mTOR in spinal cord tissue showed that the PRP-BMSCs-EXO group had significantly higher expression of miR-29a-3p, p-PI3K, p-Akt, and p-mTOR, and lower expression of PTEN compared with the PRP, BMSCs-EXO, and SCI groups (P < 0.05) (Fig. 12g–h). These results collectively indicate that PRP-BMSCs-EXO promote functional recovery in rats with SCI via the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis. Discussion Bone marrow is the most common source of stem cells for spinal cord injury (SCI) repair research. Previous studies have suggested that bone marrow mesenchymal stem cells (BMSCs) promote damaged axon regeneration and SCI repair primarily through differentiation to replace injured neurons, improving the microenvironment, and providing neurotrophic support (Assinck et al. 2017; Wan et al. 2024). Consequently, numerous studies have reported their functions in homing, proliferation, differentiation, secretion, and immunomodulation (L. Huang et al. 2022a, b; Yamada et al. 2019). However, accumulating evidence now indicates that BMSCs may exert their therapeutic effects predominantly through paracrine mechanisms, including anti-inflammatory, immunomodulatory, neuroprotective, and pro-angiogenic actions (Kadoya et al. 2016). Exosomes are currently the most extensively studied paracrine mediators, capable of exerting effects similar to their parent cells. This suggests that satisfactory therapeutic outcomes can be achieved without direct cell transplantation (Cantinieaux et al. 2013), making exosomes a promising candidate for SCI repair. Despite encouraging preclinical results, the clinical translation of exosome-based therapies faces significant challenges, primarily due to low production yields and difficulties in large-scale manufacturing. Studies have shown that platelet-rich plasma (PRP) not only promotes the repair of SCI tissue but also enhances the proliferation and migration of BMSCs in vitro (Wang et al. 2022). However, the exosomes secreted by BMSCs primed with PRP have not been previously investigated. In the present study, we added PRP to BMSC culture medium and confirmed that PRP enhances BMSC proliferative activity. The exosomes isolated from PRP-primed BMSCs (PRP-BMSCs-EXO) exhibited no significant differences in morphology (typical cup-shaped structure), particle size distribution (50–150 nm), or surface marker expression compared with exosomes from BMSCs cultured in standard medium (BMSCs-EXO). However, the protein concentration of PRP-BMSCs-EXO, as determined by BCA assay, was significantly higher, suggesting that PRP may promote the release of exosomes from BMSCs. Several studies have demonstrated that PRP can alter the miRNA expression profile of MSCs, for example, downregulating hsa-miR-212-5p (Chen et al. 2022), miR-101 (Konar et al. 2023), and miR-145 (Konar et al. 2023), and upregulating miR-146a, miR-199a (Hesari et al. 2020), miR-140 (Konar et al. 2023), miR-199 (Konar et al. 2023), miR-27 (Konar et al. 2023), and miR-146 (Konar et al. 2023). However, these reports only analyzed a limited number of miRNAs affected by PRP in MSCs, lacking a comprehensive analysis of the whole miRNA expression profile and, more importantly, the miRNA profile of their secretome (exosomes). Therefore, in this study, we performed high-throughput miRNA sequencing on PRP-BMSCs-EXO and BMSCs-EXO, identifying a total of 34 differentially expressed miRNAs (DE-miRNAs). The four most significantly upregulated miRNAs were selected for validation by RT-qPCR, which confirmed that miR-29a-3p was the most prominently upregulated miRNA in PRP-BMSCs-EXO. The pathophysiological mechanisms of SCI are highly complex, involving cellular apoptosis, neuroinflammation, vascular injury, reactive oxygen species (ROS) accumulation, excessive release of excitatory neurotransmitters, calcium overload, and mitochondrial dysfunction (Huang et al. 2017). Among these, cellular apoptosis and neuroinflammation are thought to play a dominant role in secondary SCI. Apoptosis is a cellular response to physiological or pathological stimuli, triggered by microenvironmental changes or adaptive damage (Hengartner 2000). The apoptosis of neurons and oligodendrocytes leads to the disruption of neural pathways and demyelination, which are critical causes of functional deficits after SCI (Profyris et al. 2004). The extent of apoptosis directly impacts functional recovery and plays a vital role in neuronal survival and axonal regeneration (He et al. 2024; Tang et al. 2014). The Bcl-2 family, Bax, and caspase family are the most common apoptotic markers involved in programmed cell death and directly influence functional recovery after SCI (Ning et al. 2024). Following SCI, the expression of pro-apoptotic proteins Bax and caspase-3 is upregulated, while the expression of the anti-apoptotic protein Bcl-2 is downregulated (Adams and Cory 1998; Han et al. 2022; Lin et al. 2006). MSCs-derived exosomes have been reported to possess anti-apoptotic effects (He et al. 2022), and similar to their parent cells, can promote angiogenesis and axonal regeneration (Huang et al. 2020; Wang et al. 2019). In the present study, compared with the SCI model group, treatment with BMSCs-EXO and PRP-BMSCs-EXO significantly enhanced neuronal viability and migration capacity, markedly decreased the expression of pro-apoptotic proteins (Bax, caspase-3), and significantly increased the expression of the anti-apoptotic protein (Bcl-2), with the most pronounced effects observed in the PRP-BMSCs-EXO group. Furthermore, TUNEL staining and immunofluorescence staining for NeuN and BrdU revealed that treatment with BMSCs-EXO and PRP-BMSCs-EXO significantly reduced neuronal apoptosis and increased the number of surviving and regenerating neurons, with the most significant changes occurring in the PRP-BMSCs-EXO group. These results, combined with the BBB scores, inclined plane test results, and histological staining, indicate that PRP-BMSCs-EXO can promote neurological functional recovery by inhibiting neuronal apoptosis and enhancing neuronal regeneration. Autophagy is a complex biological process involving autophagosome formation, fusion of autophagosomes with lysosomes, and degradation within autolysosomes (Zhou et al. 2017). Autophagy plays a crucial but controversial role in SCI. Some studies suggest that autophagy alters presynaptic structures and neurotransmission (Hernandez et al. 2012), limiting neural regeneration. In contrast, other studies indicate that autophagy, as a key regulator of normal axon guidance, vesicle release, dendritic spine formation and pruning, and synaptic plasticity (Hwang et al., 2019; Ko et al. 2020; Tran et al. 2020), can promote neural regeneration. Beclin-1, LC3B, and p62 are commonly used markers to assess autophagic activity in SCI. A study showed that the ratio of LC3-II to LC3-I significantly increased at 3 days post-SCI, peaked at 7 days, and then significantly decreased after 21 days (Zhang et al. 2014), indicating that autophagy is dynamically regulated after SCI, which may be beneficial for maintaining homeostasis. Neurons are particularly sensitive to autophagy, and alterations in autophagic activity may be associated with the location, type, and severity of SCI (Bisicchia et al. 2017; Lin et al. 2016; Lipinski et al. 2015; Ray 2020). Due to its potential positive implications for neuroprotection, autophagy has attracted increasing attention from researchers in the field of SCI. In the present study, WB analysis of Beclin-1, LC3B-II/LC3B-I ratio, and p62 expression in SCI models (in vitro and in animal models at 28 days post-injury) showed that treatment with BMSCs-EXO and PRP-BMSCs-EXO significantly decreased the expression of Beclin-1 and the LC3B-II/LC3B-I ratio, and significantly increased the expression of p62 compared with the SCI model group, with the most significant changes observed in the PRP-BMSCs-EXO group. These findings suggest that PRP-BMSCs-EXO exert neuroprotective effects by inhibiting neuronal autophagy. MiRNAs associated with various secondary injury processes, such as apoptosis, autophagy, inflammation, and oxidative stress, exhibit differential expression before and after SCI (Sabirzhanov et al. 2019). Liu et al. (Liu et al. 2009) identified approximately 300 miRNAs in the spinal cords of adult rats at 4 h, 1 day, and 7 days post-SCI using microarray analysis, of which 97 showed significant differential expression. In another study, researchers identified 1773 unique miRNA sequences in rat spinal cord tissue at 1, 3, and 7 days post-SCI, and characterized approximately 500 DE-miRNAs and 59 target genes of these DE-miRNAs (Yunta et al. 2012). Further validation confirmed that miR-21, miR-184, miR-340-5p, miR-369-3p, and miR-466b were consistent with their target genes. Strickland et al. (Strickland et al. 2011) analyzed 32 microarrays from injury sites at 4 and 14 days post-SCI and found that miR-124, miR-129, and miR-1 were significantly downregulated, while three miRNAs, including miR-21, were significantly upregulated and induced an anti-apoptotic response. Therefore, miRNAs likely play important roles in the secondary injury processes following SCI. In the present study, high-throughput sequencing and RT-qPCR indicated that miR-29a-3p was the most significantly upregulated DE-miRNA in PRP-BMSCs-EXO compared with BMSCs-EXO. Given the favorable therapeutic effects of PRP-BMSCs-EXO on SCI, we hypothesized that miR-29a-3p may be a key miRNA mediating the biological effects of PRP-BMSCs-EXO. TargetScan and RNAhybrid predicted that PTEN might be a target gene of miR-29a-3p, which was validated by a dual-luciferase reporter assay. In vitro and in vivo studies showed that treatment with BMSCs-EXO and PRP-BMSCs-EXO resulted in increased miR-29a-3p expression and suppressed PTEN expression compared with the SCI group, with the most significant effects observed in the PRP-BMSCs-EXO group. Further miRNA transfection validation experiments demonstrated that the rno-miR-29a-3p-mimic group exhibited effects similar to PRP-BMSCs-EXO treatment, including enhanced neuronal viability and migration, inhibited neuronal apoptosis and autophagy, promoted neuronal regeneration, and suppressed PTEN expression. In contrast, these effects were inhibited in the rno-miR-29a-3p-inhibitor group. These results suggest that PRP-BMSCs-EXO may exert their effects through the miR-29a-3p/PTEN axis. Currently, there are few studies on miR-29a-3p in SCI. Wang et al. (Wang et al. 2023) investigated the protective effects of exosomes derived from hypoxia-induced human umbilical cord MSCs (HucMSCs) on SCI. The results showed that HucMSC exosomes reduced spinal cord tissue damage, improved hindlimb motor function, and decreased the production of inflammatory cytokines IL-1β, IL-6, and TNF-α in vivo. In vitro, they promoted neuronal proliferation and reduced apoptosis. The underlying mechanism may involve circOXNAD1 in HucMSC exosomes sponging miR-29a-3p to increase the expression of FOXO3a. Multiple studies have shown that miR-29a-3p negatively regulates PTEN. Zhu et al. (Zhu et al. 2024a, b) found that miR-29a-3p can stimulate chondrocyte proliferation and inhibit apoptosis by suppressing PTEN expression. Liu et al. (Liu et al. 2025) demonstrated that miR-29a-3p is involved in the proliferation and invasion of pituitary adenomas by regulating the COL6A6/PTEN-PI3K/Akt/CUX1 pathway. Kang et al. (Kang et al. 2024) found that miR-29a-3p enhances the migration of human MSCs by inhibiting PTPRK and PTEN. Many studies have also shown that PTEN is a key negative regulator of the PI3K/Akt/mTOR pathway, which is a crucial regulator of cell growth, proliferation, metabolism, and viability. The deletion or inhibition of PTEN can promote axonal regeneration, neuronal survival, and functional recovery after central nervous system injury by activating the Akt/mTOR pathway (Geoffroy et al. 2015; Liu et al. 2010). Chen et al. (Chen et al. 2021a, b) showed that exosomes overexpressing miR-26a can downregulate PTEN and activate the PI3K/Akt/mTOR pathway to exert neuroprotective effects after SCI. Chen et al. (J. Chen et al. 2021a, b) found that neural stem cell-derived exosomes loaded with FTY720 can promote neurological functional recovery after SCI in rats through the PTEN/AKT signaling pathway. Huang et al. (Huang et al. 2022a, b) demonstrated that downregulating miR-429 in plasma exosomes from SCI rats affects the PTEN/PI3K/Akt pathway to promote neuronal apoptosis. In the present study, in addition to increased miR-29a-3p expression and suppressed PTEN expression, PRP-BMSC-EXO intervention also led to increased expression of p-PI3K, p-Akt, and p-mTOR, indicating activation of the PI3K/Akt/mTOR signaling pathway. Therefore, PRP-BMSCs-EXO may exert their protective effects on SCI through the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis. Although this study confirmed that PRP-primed BMSCs-EXO may have superior neuroreparative functions for SCI compared with PRP and BMSCs-EXO, and that the underlying mechanism may be related to the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis, several limitations should be acknowledged. First, PRP is rich in cytokines and trophic factors in addition to miRNAs, which may alter the protein cargo of BMSCs-EXO. This study did not investigate the proteomic differences between the two exosome types. Second, this study only investigated upregulated DE-miRNAs and did not analyze downregulated DE-miRNAs, nor the expression of long non-coding RNAs (lncRNAs) and circular RNAs (circRNAs). Finally, because exosomes are nanoscale vesicles that are difficult to count, estimating exosome secretion by measuring protein concentration via BCA assay may deviate from the actual secretion levels. Conclusion The results of this study indicate that PRP priming can promote BMSC proliferation and exosome secretion. Moreover, compared with PRP and BMSCs-EXO, PRP-primed BMSCs-EXO are more effective in promoting neurological functional recovery after SCI. The underlying mechanism may involve the inhibition of neuronal apoptosis and autophagy, and the promotion of neuronal regeneration through the miR-29a-3p/PTEN/PI3K/Akt/mTOR axis. These findings may provide new insights into the treatment of SCI. Acknowledgements This work was supported by grants from the Scientific Research Fund of Anhui Medical University (Grant Number: 2021xkj123). Author contributions Y.Y. contributed to the research design, acquisition, and analysis of the data, and manuscript writing. M.F. contributed to the manuscript writing, analysis, and interpretation of the data, and software used in the work. C.S. contributed to the conception and design of the work. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Conflict of interest The authors declare no conflict of interest. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.