Abstract Traumatic brain injury (TBI) leads to secondary injuries, such as neuroinflammation and brain dysfunction, which is a critical challenge in clinical treatment. The use of bone marrow mesenchymal stem cells (BMSCs) is one of the potential strategies to treat TBI by alleviating inflammation, reducing neuronal loss, and promoting brain function recovery. Extracellular vesicles (EVs) released by BMSCs are regarded as an ideal alternative to cell therapy. This study showed that hypoxia significantly enhanced the release of EVs from BMSCs, and hypoxia- preconditioning (H-EVs) treatment significant effects on promoting microglial M2 polarization, improving endothelial cell activity, and inhibiting the formation of neutrophil extracellular traps, ultimately accelerating brain function recovery. Mechanistically, single-cell sequencing revealed a significant reduction in specificity protein 1 (SP1) expression and a change in the proportion of infiltrating inflammatory cell subsets in brain tissues after the H-EVs treatment. Hypoxia-preconditioning changed the miRNA microarray analysis results in H-EVs, such that miR-145-5p negatively regulated nuclear factor kappa-B (NF-κB) by targeting SP1, induced microglial M2 polarization, alleviated endothelial cell dysfunction, and promoted brain function recovery. Intranasal delivery of hypoxia-induced BMSC-EVs showed great potential in the treatment of secondary TBI and revealed a novel mechanism by which miR-145-5p regulates inflammatory response and intercellular communication by inhibiting the SP1/NF-κB axis. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-025-04572-3. Highlights Hypoxic preconditioning enhances the secretory capacity of BMSCs for extracellular vesicles, thereby conferring a more potent anti-inflammatory effect. Intranasal H-EVs delivery significantly promotes the activation of M2 microglia after TBI, thereby alleviating the apoptosis of endothelial cells and neurons. This is the first study to show that H-EVs can regulate microglial polarization via the miR-145-5p/SP1/NF-κB pathway, alleviate endothelial cell dysfunction, inhibit the formation and recruitment of neutrophil extracellular traps in the brain injury areas, and promote the recovery of cortical cerebral blood flow. This is the first instance of using single-cell sequencing to reveal the differences between the H-EVs and normoxia-preconditioning (N-EVs) in TBI treatment. Behavioral experiments demonstrated that H-EVs treatment significantly improved the neurological and motor function recovery of TBI rats. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-025-04572-3. Graphical Abstract [48]graphic file with name 13287_2025_4572_Figa_HTML.jpg Keywords: Traumatic brain injury, Mesenchymal stem cells, Neuroinflammation, Extracellular vesicles, Microglia Introduction Traumatic brain injury (TBI) is a neurological affliction precipitated by external trauma and has ascended to become one of the leading causes of mortality and long-term disability across the globe. Annually, more than 50 million new cases are documented worldwide, with its incidence escalating at an average annual rate of 4.67%. Concurrently, the global financial burden of TBI treatment now exceeds 70 billion US dollars each year [[49]1–[50]4]. The treatment of TBI is highly challenging, mainly due to the heterogeneity of trauma itself as well as the complex pathological mechanism. There are currently no accepted guidelines for the medical treatment of severe TBI beyond craniotomy and basic vital sign support [[51]5]. The mechanical damage caused by TBI is called “primary brain injury”. It triggers a series of harmful biochemical cascades collectively referred to as “secondary brain injury“ [[52]6, [53]7]. Secondary brain injury after TBI is a complex pathophysiological process, and neuroinflammation is its most direct cause [[54]4]. Central nervous system injury triggered by TBI activates resident microglia and peripheral immune cells and induces an inflammatory response. Microglia can sense damage-associated molecular patterns (DAMPs) and respond rapidly. For example, the production or impairment of adenosine triphosphate or intracellular proteins released by dying cells can be recognized by microglia and trigger the corresponding inflammatory response [[55]8]. The initiation of DAMP receptor signaling drives classical M1-phenotype microglia to secrete proinflammatory cytokines, such as TNF-α, IL-6, and IL-1β, which are destructive to neurogenesis. In contrast, M2-phenotype microglia secrete anti-inflammatory cytokines, such as TGF-β, IL-10 and IL-4, which contribute to promoting nerve regeneration and repair [[56]9]. Local production of cytokines and chemokines directly regulates the inflammatory response in injured tissues and provides key mediators for peripheral immune cell infiltration [[57]8]. Therefore, regulating microglial activation, polarization, and downstream inflammatory signaling cascade may become an effective way to treat secondary brain injury and improve brain dysfunction after TBI. Mesenchymal stem cells (MSCs) are pluripotent stem cells used to stimulate axonal growth and promote nerve repair in damaged nerve tissues due to their self-renewal ability and multi-directional differentiation potential [[58]10]. Nevertheless, the inherent limitations and immunological hurdles of MSCs transplantation into target tissues remain undeniable. Emerging evidence reveals that the post-transplantation survival rate of MSCs is remarkably low, with a concomitant loss of certain functional attributes. Moreover, the procedure carries significant risks, including immune rejection, cellular dedifferentiation, and tumorigenesis, all of which substantially curtail the clinical viability of direct MSCs transplantation [[59]11, [60]12]. In recent years, some scholars have suggested that the paracrine mechanism may represent a possible MSC-based treatment for a variety of diseases, and its therapeutic effect may be related to MSC-secreted extracellular vesicles (MSC-EVs) [[61]13, [62]14]. MSC-EVs therapy offers a multitude of advantages. As signaling molecules, MSC-EVs exhibit functional parity with MSCs, and their more stable membrane structure can protect the integrity of their contents. With their small size, high activity, wide distribution, and low immunogenicity properties, MSC-EVs can safely and effectively participate in information exchange, especially given their ability to cross the blood-cerebrospinal fluid barrier, making them potential carriers for drug delivery to the brain in neurological diseases [[63]15]. Previous studies have illuminated the considerable therapeutic potential of BMSC-EVs in neurological disorders; when administered via the retro-orbital route in TBI mouse models, these vesicles effectively quell early neuroinflammatory responses and profoundly enhance neurological recovery and overall prognosis [[64]16]. In addition, several trials have confirmed that activated microglia are polarized toward M1 after TBI, leading to severe neuroinflammation. BMSC-EVs can not only reduce neuroinflammation after TBI, but also play a neuroprotective role by promoting neurovascular repair [[65]17]. Therefore, regulating the phenotypic microglial shift after TBI may be an effective way to alleviate neuroinflammation. Another study further showed that BMSC-EVs significantly increased the number of endothelial cells and neurons at the injury site and hippocampus area in a TBI animal model and effectively promoted neurological function recovery [[66]18]. In addition, BMSC-EVs can also up-regulate the expression of a variety of cytokines, such as vascular endothelial growth factor, and promote angiogenesis. These factors increase blood vessel density and improve local blood perfusion by stimulating the proliferation, migration, and differentiation of endothelial and adventitial cells [[67]19]. Recent studies have shown that the secretory function of MSCs and their secretion components are closely related to their microenvironment. Among them, oxygen concentration plays a key role in MSCs self-renewal, enhancing the biological activity of EVs and promoting the paracrine effect [[68]20, [69]21]. MSCs were generally exposed to normoxic conditions (21% O[2]) during in vitro culture, which was significantly different from the in vivo oxygen concentration. The majority of MSCs in vivo reside in a hypoxic environment (≤ 1–8% O[2]). In a recent study, EVs derived from MSCs cultured in a serum-free (0% fetal bovine serum (FBS)) and hypoxic (1% O[2]) environment were found to be enriched in a variety of factors that may have pro-angiogenic effects on ischemic tissues [[70]20]. Another study revealed that hypoxic conditions did not markedly affect the marker expression or differentiation potential of MSCs; instead, they augmented their paracrine activity and profoundly enhanced cardiac function [[71]22]. Indeed, hypoxic MSCs preconditioning can enhance their biological functions and activities, thereby enhancing the efficacy of MSCs transplantation in various disease models [[72]23]. Although most current studies on EVs focus on the analysis of their protein and RNA composition, the miRNA contents in H-EVs and their potential repair mechanisms remain unclear. Furthermore, it is unknown whether MSCs are more effective in promoting functional recovery after TBI under hypoxic conditions and whether this enhanced effect is mediated by EVs. This is the first instance of using single-cell sequencing to reveal the differences between the H-EVs and normoxia-preconditioning (N-EVs) in TBI treatment. This study hypothesizes that intranasally delivered hypoxia-conditioned BMSC-derived EVs enhance miR-145-5p expression and markedly alleviate the inflammatory cascade during both the acute and subacute stages post-TBI. Moreover, miR-145-5p further suppressed the transcriptional activity of NF-κB by targeting SP1, thereby enhancing microglial polarization toward the M2 phenotype and ameliorating endothelial dysfunction—ultimately contributing to the attenuation of neuroinflammation and the restoration of neurological function following TBI. Materials and methods Animals and study design Sprague Dawley (SD) male rats (8–10 weeks old, weight: 250 ± 20 g) used in the study were provided by the Animal Experimental Center of Fuzhou General Clinical Medical College, Fujian Medical University. Cesarean section was performed on female SD rats on day 18 of pregnancy to obtain fetal rats for primary neuronal cell extraction. At the end of the study, all rats were humanely euthanized via intraperitoneal injection of an overdose of sodium pentobarbital (150 mg/kg). Death was confirmed by the complete cessation of respiration and heartbeat. All procedures were carried out in strict accordance with the institutional guidelines for the care and use of laboratory animals, and were approved by the Animal Ethics Committee of Fuzhou Clinical Medical College, Fujian Medical University (Approval No. 2024-05, approved on May 22, 2024). The work has been reported in line with the ARRIVE guidelines 2.0. The study schematic is represented in Fig. [73]1. Fig. 1. [74]Fig. 1 [75]Open in a new tab A schematic illustrating the timing and analysis of various in vitro and in vivo experiments conducted in this study. The sequence of procedures encompasses the induction of controlled cortical impact injury (CCI), nasal administration of hyaluronidase immediately following TBI, and intranasal (IN) administration of BMSC-EVs 30 min post-TBI. The timeline of both in vivo and in vitro studies after TBI is also delineated BMSCs isolation, culture, identification, and hypoxia BMSCs isolation and culture In brief, under strictly sterile conditions, the tibias and femurs were carefully harvested from three-week-old male Sprague-Dawley (SD) rats, with all adherent muscle and connective tissues meticulously removed [[76]24]. The bones were immersed in pre-chilled phosphate-buffered saline (PBS; Gibco, NY, USA). Using a 1 mL syringe fitted with a 22G needle, 5 mL of ice-cold complete culture medium—DMEM/F12 (Gibco, NY, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, NY, USA), 100 µg/mL penicillin, and 100 µg/mL streptomycin (Invitrogen, CA, USA)—was repeatedly flushed through the bone marrow cavity to extract the marrow, and the resulting suspension was collected into centrifuge tubes. Following centrifugation at 1000 rpm for 5 min, the supernatant was discarded, and the cell pellet was gently resuspended in fresh complete medium. The suspension was passed through a 70 μm cell strainer to eliminate residual tissue debris. The filtered cells were subsequently transferred into 25 cm² culture flasks (Corning, NY, USA) and incubated at 37 °C in a humidified atmosphere containing 5% CO₂. After 24 h, the culture medium was first replaced to remove non-adherent cells, followed by medium changes every three days. Within 7 to 10 days, cells exhibiting the characteristic spindle-like morphology and swirling growth pattern of bone marrow-derived mesenchymal stem cells (BMSCs) became evident. BMSCs from passages 3 to 8 were employed for subsequent experimental procedures. BMSCs marker identification According to previous research [[77]25], third-passage BMSCs were collected by first discarding the spent medium and gently rinsing the adherent cells with 2 mL of PBS. Following aspiration of the PBS, 2 mL of 0.25% trypsin-EDTA was added, and enzymatic digestion was conducted at 37 °C for 3 min. The digestion was promptly neutralized with an equal volume of complete culture medium. The resulting cell suspension was subjected to centrifugation at 1000 rpm for 5 min. The supernatant was removed, and the cell pellet was gently resuspended in PBS supplemented with 1% FBS, followed by three thorough washes to eliminate residual enzymatic activity. Cell density was then adjusted to 1 × 10⁶ cells/mL, and the suspension was evenly distributed into eight Eppendorf tubes, each containing 200 µL. The following fluorochrome-labeled antibodies were introduced to the respective tubes: PE-IgG and FITC-IgG (isotype controls), FITC-CD105, FITC-CD29, FITC-CD45, PE-CD90, PE-CD11b, and PE-CD34 (all procured from BD Biosciences, San Jose, CA, USA). After gentle agitation, the samples were incubated at 4 °C in darkness for 20 min. Upon completion, cells were washed with PBS and centrifuged at 3000 rpm for 5 min to eliminate any unbound antibodies. This wash was repeated three times to ensure purity. Finally, the cells were resuspended in 300 µL of PBS and subjected to flow cytometric analysis within one hour. Data interpretation and phenotypic profiling were conducted using FlowJo software. Osteoblast differentiation of BMSCs BMSCs were seeded into six-well plates at a density of 2 × 10⁴ cells/cm² and cultured in standard complete medium. Once the cells reached approximately 70% confluence, the medium was gently aspirated, and each well was replenished with 2 mL of OriCell^® Rat BMSC Osteogenic Differentiation Medium (RAXMX-90021). To sustain the osteogenic microenvironment, the induction medium was refreshed every three days. After 2 to 4 weeks of induction, abundant calcium nodule formation became evident under light microscopy. The medium was carefully removed, and each well was rinsed twice with PBS to eliminate residual medium. Subsequently, 2 mL of 4% paraformaldehyde was added per well to fix the cells for 1 h, preserving cellular morphology and the integrity of mineralized nodules. After fixation, the wells were again gently washed twice with PBS. Alizarin Red working solution was then applied to each well and incubated at room temperature for 10 min to selectively stain the calcium deposits indicative of osteogenic differentiation. Upon completion, the staining solution was aspirated, and wells were rinsed three times with PBS. Finally, 2 mL of PBS was added to each well, and the stained osteogenic features were visualized and documented under a light microscope. Adipocyte differentiation of BMSCs BMSCs were seeded into six-well plates at a density of 2 × 10⁴ cells/cm², with 2 mL of standard complete medium added to each well. Once the cells reached full confluency, the culture medium was carefully aspirated and replaced with 2 mL of OriCell^®(RAXMX-90031) Rat BMSC Adipogenic Differentiation Medium A to initiate adipogenic induction. After three days of induction, Medium A was removed and substituted with 2 mL of Adipogenic Differentiation Medium B. Following one day of incubation in Medium B, the medium was again replaced with Medium A. This alternating regimen of A and B media was continued until intracellular lipid droplets of sufficient quantity and appropriate size were observed.Upon completion of adipogenic induction, the medium was aspirated, and the cells were gently rinsed 2–3 times with 1× PBS. Each well was then fixed with 2 mL of 4% paraformaldehyde solution for 1 h at room temperature. After fixation, the wells were again gently rinsed 2–3 times with 1× PBS to remove any residual fixative. Subsequently, 2 mL of freshly prepared Oil Red O staining solution was added to each well and incubated at room temperature for 30 min. Following staining, the dye was aspirated, and the cells were gently washed three times with 1× PBS. Finally, 2 mL of 1× PBS was added to each well, and the stained lipid droplets were examined under a microscope. Representative images were captured to document the adipogenic differentiation outcome. Chondroblast differentiation of BMSCs A total of 4 × 10⁵ BMSCs were transferred into a 15 mL centrifuge tube and subjected to centrifugation at 250 × g for 5 min at 20 °C. Following removal of the supernatant, 0.5 mL of OriCell^® Rat BMSC Chondrogenic Differentiation Premix Solution (RAXMX-90041) was added to the cell pellet. The suspension was thoroughly resuspended and centrifuged again at 150 × g for 5 min. Subsequently, 20 µL of OriCell^® Rat BMSC Chondrogenic Differentiation Supplement II was mixed with 2 mL of the chondrogenic premix to prepare the complete chondrogenic induction medium. The resulting cell pellet was resuspended in 2 mL of the complete medium and centrifuged once more at 150 × g for 5 min at 20 °C.To ensure adequate gas exchange, the cap of the centrifuge tube was slightly loosened, and the tube was positioned upright in the incubator. Once the formation of compact cell aggregates—chondrogenic spheroids—was observed, gentle tapping at the base of the tube dislodged them, allowing the spheroids to remain in suspension within the medium. The chondrogenic medium was replaced every 2 to 3 days. The induction was continued until the spheroids reached a diameter of approximately 1.5–2 mm, at which point they were processed for histological staining. Following sectioning and drying, Alcian Blue staining solution was applied to the sections, followed by incubation at 37 °C for 1 h. Slides were then rinsed under running water for 5 min and air-dried. The stained sections were examined under a microscope, and images were captured for documentation. Areas exhibiting positive Alcian Blue staining—characterized by vivid blue coloration—indicated the presence and distribution of acidic mucopolysaccharides within the chondrogenic tissue. Hypoxic induction of BMSCs BMSCs were meticulously seeded into six-well plates at a density of 2 × 10⁵ cells per well and incubated under normoxic conditions (37 °C, 5% CO₂) for 24 h to ensure robust cellular adherence. Thereafter, the conditioned medium was gently aspirated and replaced with 2 mL of pre-warmed, fresh complete medium supplemented with extracellular vesicle-depleted FBS (System Biosciences, Mountain View, USA), thereby minimizing contamination from exogenous vesicular components. A hypoxic microenvironment was established using a tri-gas incubator pre-equilibrated for no less than four hours to attain an internal atmosphere of 1% O₂, 5% CO₂, at 37 °C. The culture plates were subsequently transferred into this chamber and maintained for 48 h, eliciting the hypoxia-induced adaptive response in BMSCs. Upon completion of the hypoxic incubation, the plates were swiftly retrieved, and all ensuing procedures were conducted atop an ice pack to avert biochemical perturbations arising from reoxygenation stress. Isolation, identification, and quantification of EVs When BMSCs reached 80% confluence, their medium was replaced with EVs-free FBS medium and cultured under normoxic (21% O[2]) or hypoxic (1% O[2]) conditions for 48 h. The medium was then collected and centrifuged at 300 × g for 10 min at 4 °C followed by 2,000 × g for 10 min. After centrifugation, the cell supernatant was filtered using a 0.22-µm sterile filter (Steritop™ Millipore, Burlington, MA). The filtered supernatant was then added to the upper compartment of an Amicon Ultra-15 centrifuge filter (Millipore, USA) and centrifuged at 4,000 × g until the volume in the upper compartment was reduced to 200 µL. To purify EVs, the liquid was loaded into a mat of 30% sucrose/D2O and ultracentrifugation was performed at 100,000 × g (Beckman Coulter) for 60 min at 4 °C. EVs were then resuspended in phosphate-buffered saline (PBS) and stored at − 80 °C for subsequent experiments [[78]26]. WB was used to identify surface markers of specific extracellular vesicles, such as TSG101, CD9, and CD63. BMSC-EVs protein concentration was determined using a bicinchoninic acid protein assay (BCA; Thermo Fisher Scientific, USA). The NanoSight LM10 system (NTA, NanoSight Ltd., Navato, CA) was employed to assess the size distribution and particle concentration of N-EVs and H-EVs.The concentration of EV samples was meticulously adjusted to fall within the optimal analytical range of 1 × 10⁸ to 1 × 10¹¹ particles/mL, ensuring suitability for NTA measurement. The diluted extracellular vesicle suspensions were gently introduced into the sample chamber of the NTA device, guaranteeing uninterrupted flow and the absence of air bubbles. The instrument harnesses a laser beam to illuminate the particles, enabling visualization of their Brownian motion through dynamic light scattering (DLS) technology. Transmission electron microscopy (TEM, Tecnai12; Philips, Best, The Netherlands) was utilized to exquisitely delineate the morphology and ultrastructural features of EVs. A droplet of the EVs suspension was delicately applied onto a pristine copper TEM grid, followed by fixation with 2.5% glutaraldehyde at room temperature for 10–15 min. The grids were then immersed in PBS and gently rinsed three times, each for 5 min, to eliminate residual fixative, before being air-dried. Subsequently, the EVs samples were treated with 1% phosphotungstic acid for approximately 30 min to further preserve structural integrity. The grids were then stained with 2% uranyl acetate for 5–10 min to enhance electron density, followed by a gentle rinse with distilled water. To augment contrast, a 5-minute treatment with sodium tungstate was performed. Finally, the grids were placed into the TEM chamber, where imaging parameters were meticulously adjusted to achieve optimal resolution, enabling detailed visualization of the vesicles’ morphology, including contour, dimensionality, and other discernible ultrastructural characteristics. Uptake efficiency of PKH67-labeled EVs by microglial cells EVs were labeled and traced using the lipophilic green fluorescent dye PKH67 (Sigma-Aldrich, USA). Briefly, EVs were resuspended in 1 mL of diluent C with 4 µL of PKH67 and incubated for 4 h at room temperature. The reaction was subsequently terminated by addition of an equal volume of EV-free FBS. EVs were washed twice with FBS/RPMI-1640 to remove excess PKH67. The labeled EVs were then resuspended in PBS. Fluorescently labeled EVs prepared as described above were added to 80% confluent microglia and incubated for 12 h. The plates were then fixed with 4% paraformaldehyde for 15 min at room temperature. The nuclei were stained with DAPI. The uptake of PKH67-labeled N-EVs and H-EVs by microglia was observed using laser confocal microscopy. ImageJ software was used to evaluate the fluorescence intensity of PKH67 at different time points within the two groups [[79]27]. Establishment of TBI rat model TBI model was established using controlled cortical injury (CCI) [[80]28]. The SD rats were anesthetized and fixed using a brain stereotaxic apparatus and the skull tops were exposed with animals remaining in the prone position. The isoflurane concentration was adjusted to 0.5–1%, the flow rate was set to 1 L/min to maintain anesthesia, and the rat body temperature was maintained at 37 ± 0.2 °C with a heating pad. A midline scalp incision was made to expose the skull. Craniotomy was performed with a high-speed dental drill 2.5 mm posterior to the rat anterior fontanel and 2.5 mm lateral to the right of the midline, with a bone window of 4 mm in diameter. Impact parameters were as follows: impact head diameter: 3.5 mm; depth: 3 mm; speed: 5 m/s; and residence time: 200 ms. After the blow, hyaluronidase (1 mg/mL) was administered nasally (Sigma-Aldrich, USA). Magnetic resonance imaging of the lesion site in rats subjected to TBI In this study, a 9.4 T small-animal magnetic resonance imaging (MRI) scanner (Bruker, Germany) was employed to perform serial imaging of rats subjected to traumatic brain injury (TBI) at distinct postoperative intervals—24 h, 72 h, 7 days, and 14 days. Anesthesia was induced using a mixed gas comprising 4% isoflurane, 25% oxygen, and 75% nitrogen, with an induction time of approximately 45 s. Following induction, anesthesia was maintained with 1.8% isoflurane delivered in a continuous flow of the gas mixture (25% oxygen, 75% nitrogen) at a rate of 1 L/min. Initial imaging involved T1-weighted sequences (T1WI) to acquire anatomical views of the rat brain in sagittal, coronal, and axial planes. This was followed by high-resolution T2-weighted imaging using a rapid acquisition with relaxation enhancement (RARE) 3D sequence. The scanning parameters were as follows: repetition time (TR) = 3000 ms, echo time (TE) = 33 ms, number of excitations = 2, field of view (FOV) = 40 × 40 mm², slice thickness = 0.8 mm, number of slices = 28, and matrix size = 256 × 256 mm²—ensuring the acquisition of finely detailed T2-weighted images [[81]29]. Laser speckle imaging system measurement of cortical cerebral blood flow after TBI Laser Speckle Contrast Imaging (LSCI, RWD, China) was employed to monitor cerebral perfusion, utilizing a 785 nm continuous-wave laser directed perpendicularly onto the TBI lesion site, with reflected signals captured via a CCD imaging system. The LSCI apparatus comprised a laser source, an Olympus ZS61 microscope, a high-resolution camera, and a computer workstation. Image exposure time was set to 8 ms to facilitate continuous, real-time monitoring of cerebral blood flow dynamics. Relative perfusion values of the lesion and surrounding vasculature were recorded at baseline (pre-TBI), as well as 3 h, 24 h, and 48 h post-injury. For each time point, 600 consecutive frames were acquired and averaged to derive the relative cerebral blood flow (rCBF). Special attention was given to the injury epicenter, where localized hemodynamic changes were scrutinized to evaluate the therapeutic efficacy of different EV treatments. Regions of interest (ROIs) were delineated on the images to quantify mean speckle contrast variations, which serve as proxies for blood flow velocity. To ensure precision in ROI selection, dynamic circular ROIs were drawn intraoperatively using the system’s integrated software, targeting characteristic vascular structures within the lesion zone and enabling direct temporal comparison of hemodynamic alterations. Modified neurological severity score and behavioral experiments The Modified Neurological Severity Score (mNSS) was used to assess the recovery of neurological function 1, 3, 7, 14, 21, and 28 days after TBI (n = 5 for each group). Rats were tested and evaluated before surgery to confirm normal scores (0). The mNSS were subsequently analyzed at different time points after TBI. Day 14 marks a pivotal transitional window from the subacute to the chronic phase of TBI, during which neuroinflammation begins to subside and the processes of neuroplastic repair gradually emerge—most notably, the restoration of hippocampus-related cognitive functions begins to assume a central role [[82]30, [83]31]. Morris water maze (MWM) test was used to evaluate the spatial learning and memory ability of TBI rats after different treatments on day 14 (n = 5 for each group). The escape latency, number of crossing stations, and time spent in the target quadrant were recorded and analyzed. Open Field Test (OFT) was used to evaluate the exploratory behavior and motor ability of TBI rats after different treatments on day 14 (n = 5 for each group). Histological staining of brain tissue The brain tissue was fixed with 4% paraformaldehyde and then dehydrated in 15% sucrose solution. It was then embedded, sectioned (4 μm), and dewaxed. Hematoxylin and eosin (HE) staining was used to observe the tissue structure and cell morphology, while Nissl staining was used to visualize neuronal Nissl bodies to explore the degree of neuronal damage. For immunohistochemistry analysis, CD68 (CST) antibody was used to stain the proliferating microglia in the cortex and hippocampal DG area, while NeuN (Abcam) antibody was used to stain the neuronal damage in the hippocampal DG and CA3 regions. All images were captured using an optical microscope (BX53, Olympus, Japan) and analyzed using ImageJ software. Western blotting As previously described by Fathi et al. [[84]32]. RIPA buffer (Beyotime) containing PMSF was used to lyse cells for 30 min at room temperature. Then, the samples were centrifuged at 4 °C and 13,000 rpm for 20 min and the supernatant was collected. The entire process was performed on ice. The BCA protein assay kit (Thermo Fisher Scientific, USA) was used according to the manufacturer’s instructions to determine the protein concentration. Next, the samples were boiled at 95 °C for 5 min. The obtained protein samples were separated using 10% SDS-PAGE and then transferred to a PVDF (Millipore) membrane. After blocking with 5% BSA for 2 h, the primary antibody was incubated at 4 °C overnight, followed by incubation with HRP-conjugated secondary antibody at room temperature for 2 h. Following secondary antibody incubation, the membranes were once again rinsed with TBST to eliminate excess antibodies. Enhanced chemiluminescence (ECL) substrate was then applied to the PVDF membranes for 30 s to initiate signal development. The resulting bands were captured using a chemiluminescent imaging system and subsequently subjected to densitometric analysis via ImageJ software. Quantification of the target protein expression was performed by calculating the ratio of the optical density of the target protein band to that of the GAPDH internal control. The sample size was n = 3 per group of experiments for in vitro studies and n = 5 per group of experiments for in vivo studies.The following antibodies were used: SP1 (Abcam), NF-κB (Abcam), GAPDH (CST), iNOS (CST), NLRP3 (CST), Arg1 (CST), MyD88 (Abcam), ASC (Abcam), pp-65 (Abcam), Bax (Abcam), Bcl-2 (CST) Abcam), caspase-3 (Abcam), TLR4 (CST), sydencan-1 (CST), TNF-α (Abcam), IL-6 (Abcam), IL-1β (Abcam), IL-4 (Abcam), IL-10 (Abcam), and TGF-β (Abcam). Immunofluorescence As previously described by Fathi et al. [[85]33]. Cells or tissue sections were fixed in 4% formaldehyde for 15 min at ambient temperature, followed by permeabilization with 0.1% Triton X-100. Afterward, samples were blocked with 10% bovine serum albumin (BSA) for 2 h to prevent nonspecific binding and subsequently incubated overnight at 4 °C with primary antibodies. Following three thorough washes with PBS, samples were exposed to the appropriate secondary antibodies at room temperature for 2 h within a staining cassette. After another series of PBS washes, nuclear counterstaining with DAPI was performed, and immunofluorescent signals were visualized under a fluorescence microscope (BX53, Olympus, Japan). Fluorescence intensity and image analyses were conducted using ImageJ software. Specifically, a consistent thresholding strategy was applied to eliminate background interference, while the Regions of Interest (ROIs) tool was employed to define cellular areas for quantitative fluorescence measurements. Auto-thresholding algorithms were utilized to further distinguish specific signals from background fluorescence. Image acquisition and data interpretation were independently carried out by separate investigators blinded to experimental grouping, thereby minimizing subjective bias. All data were derived from a minimum of three independent replicates to ensure the robustness and reproducibility of the findings. The following antibodies were used: IBA-1 (CST), iNOS (Abcam), Arg1 (Abcam), CD31 (CST), α-SMA (Abcam), MAP2 (Abcam), and MPO (Abcam). Pre-stimulation of cells by lipopolysaccharide RAW264.7 cells were obtained from the Scientific Experimental Center of Guilin Medical University. Microglia were extracted from rat brain tissue. RAW264.7 cells were treated with LPS (500 ng/mL) for 24 h. Microglial cells were treated with lipopolysaccharide (LPS) (1 µg/mL) for 24 h to induce the pro-inflammatory phenotype. The EVs concentration was 10 µg/mL in cellular experiments [[86]34] and 200 µg/100 µL in animal experiments [[87]35]. MicroRNA sequencing analysis in EVs MicroRNA was extracted from N-EVs and H-EVs using Trizol. Fragments ranging from 18 to 30 nt were selected by agarose gel electrophoresis and cut. The 3’ and 5’ adapters were ligated and the small RNA connected to the two adapters was reverse transcribed and amplified by polymerase chain reaction (PCR). Finally, the approximately 140-bp band was recovered and purified by agarose gel electrophoresis to complete the library construction. Agilent 2100 and quantitative PCR (qPCR) were used for quality control of library construction, and Illumina Novaseq6000 was used for sequencing. Real-time quantitative qPCR Total RNA samples were extracted from tissues or EVs using Trizol reagent (Takara, Japan). Small RNA was extracted from N-EVs and H-EVs using the Exosome RNA Purification Kit (ESscience, Beijing, China), in strict accordance with the manufacturer’s protocol. For quantitative real-time PCR analysis of mRNA expression, complementary DNA (cDNA) was synthesized utilizing the PrimeScript™ RT Reagent Kit (TaKaRa Co., Kyoto, Japan). Amplification was carried out using Fast SYBR Green Master Mix (Thermo Fisher, Waltham, MA, USA) in conjunction with gene-specific primers, as detailed in Supplementary file1. Relative mRNA expression levels were normalized to β-actin using the 2^–ΔΔCT method. For the quantification of mature miRNAs, cDNA synthesis and subsequent real-time PCR were performed using the All-in-One™ miRNA qRT-PCR Detection Kit (GeneCopoeia, Guangzhou, China). Rnu6 small nuclear RNA (snRNA) served as the endogenous reference. miRNA expression levels were likewise normalized to Rnu6 using the 2^–ΔΔCT comparative method. All primers for miRNA analysis were procured from GeneCopoeia (Guangzhou, China) [[88]26]. Single cell sequencing of injured tissue in TBI rats On the third day of treatment, TBI rats were divided into the NC (n = 3), TBI + PBS (n = 3) and TBI + H-EVs (n = 3) groups. Three samples from the injury site of each group of rats were mixed into one sample (one sample each in NC, TBI + PBS, and TBI + H-EVs groups) and a single-cell suspension was prepared immediately. Single-cell libraries were constructed according to the 10XChromium protocol. Each library was sequenced using Illumina Novaseq6000. The Seurat and Tidyverse toolkits were used to normalize the scRNA-seq data. The top 2,000 genes with the most variation was selected to construct a principal component map. The top 15 PCs were selected for dimensionality reduction, clustering, gene mapping, and differential expression analysis. The Monocle toolkit was used to sort individual cells and construct a pseudo-temporal trajectory map of cells in chronological order. The CellChat toolkit was employed to study cell-to-cell communication to systematically analyze and understand the signal transduction network between cells and reveal the interaction patterns and regulatory mechanisms between different cell types. Dual luciferase reporter assay Full-length overlap primers were designed based on the PCR-based Accurate Synthesis method. Wild-type (WT) SP1 ([89]NM_012655.2) and mutant-type (MUT) SP1 ([90]NM_012655.2) sequences were synthesized by two rounds of PCR. The recombinant vectors SP1 ([91]NM_012655.2)-WT and SP1 ([92]NM_012655.2)-MUT were ligated into the psicheck2.0 vector. Then, rno-miR-145-5p mimics were synthesized in vitro. The cells were co-transfected into 293T cells by Lipofectamine 3000 reagent (Invitrogen) for 24 h. Firefly and renilla luciferase signals were determined using the Dual Luciferase^® Assay Kit (Promega, Madison, WI, USA). Target gene interference by siRNA, lentivirus, mimics, and inhibitors rno-miR-145-5p-mimic and rno-miR-145-5p Inhibitor vectors were constructed (GenePharma, Shanghai, China). BMSCs were infected by lentiviral vectors at an appropriate multiplicity of infection when they reached 50% confluence. Primary microglia were transfected with lentiviral vectors LV-SP1 (Hanheng, Shanghai, China) and si-SP1 (GenePharma Shanghai, China) for 24 h to achieve SP1 overexpression and knockdown. Transfection was performed using Lipofectamine 3000 reagent (Invitrogen) according to the manufacturer’s instructions. Statistical analysis All experiments were performed in at least three independent biological replicates. Data are shown as mean ± SD. GraphPad software 7.0 and SPSS22.0 were used for statistical analysis. The Student’s t-test was employed for comparisons between two groups, while one-way or two-way ANOVA was utilized for comparisons involving more than two groups, to calculate P-values. CBF at different time points in the LSCI was compared using ANOVA of repeated measures data. Statistical significance was present when P-values were < 0.05. Results Hypoxia promoted EVs secretion by BMSCs BMSCs were isolated from rat bone marrow (Figure [93]S1A) and subsequently cultured to the third generation for subsequent experiments (Figure [94]S1B). BMSCs were subjected to hypoxia (Fig. [95]2A). Both H-MSCs (hypoxia- preconditioning BMSCs) and N-MSCs (normoxia-preconditioning BMSCs) were able to differentiate into osteogenic, adipogenic, and chondrogenic cells (Figure [96]S1C). N-MSCs and H-MSCs were phenotyped using flow cytometry. Both H-MSCs and N-MSCs highly expressed markers CD29, CD90, and CD105, and showed a low expression of CD11b, CD45, and CD34 (Figure [97]S1D). The results revealed that hypoxia- preconditioning did not alter the marker expression or differentiation potential of BMSCs. Previous studies have shown that MSCs function depends on paracrine EVs, whose main activity is to mediate cell communication and interaction. EVs content depends on its cellular origin and external physiological conditions under which it is generated [[98]36]. Changes in oxygen concentration enhance many of the unique characteristics of stem and progenitor cells. Therefore, hypoxic MSCs pretreatment under conditions that increase their regenerative capacity may result in EVs with enhanced regenerative and anti-inflammatory abilities [[99]37]. On this basis, the present study sought to verify whether hypoxic MSCs conditions affect EVs released by them. TEM observation showed that both N-EVs and H-EVs had bilayer, cup, or oval structure (Fig. [100]2B). NTA particle size detection found that N-EVs and H-EVs had a similar distribution size with an average of 130.3 nm and 123.7 nm, respectively (Fig. [101]2C). In addition, the BCA concentration assay found that H-EVs exhibited a higher particle concentration compared to N-EVs (Fig. [102]2D-E). Western blot analysis was conducted to assess the expression of negative EVs surface markers Calnexin and GM130, alongside the positive markers TSG101, CD9, and CD63. The results revealed that both N-EVs and H-EVs exhibited minimal expression of Calnexin and GM130, while H-EVs displayed markedly elevated levels of TSG101, CD9, and CD63 compared to N-EVs (Fig. [103]2F-I). RAW264.7 cells were treated with LPS (500 ng/mL) for 24 h. Then, N-EVs and H-EVs were added to evaluate their anti-inflammatory ability. The results showed that IL-6 and TNF-α pro-inflammatory factor levels in the H-EVs treatment group were significantly lower than those in the N-EVs treatment group (Fig. [104]2J). Following PKH67 labeling of EVs, their biodistribution within the brain was assessed 24 h post-intranasal administration in TBI rats via immunofluorescence staining, focusing on uptake by microglia, neurons, and astrocytes. The results demonstrated that while both neurons and astrocytes were capable of internalizing EVs, the majority were predominantly taken up by microglia. Moreover, EVs internalization was markedly enhanced in the H-EVs group (Fig. [105]2K–L). In vitro, EVs were labeled with PKH67 and co-cultured with microglial cells for 24 h to evaluate differential uptake. The results revealed that EV internalization was significantly greater in the H-EVs group compared to the N-EVs group, with a marked statistical difference emerging as early as 8 h post-incubation. These findings suggest that EVs derived under hypoxic conditions are more readily internalized by microglia (Fig. [106]2M–N). Collectively, the data imply that hypoxia may potentiate the paracrine activity and anti-inflammatory capacity of MSCs. Fig. 2. [107]Fig. 2 [108]Fig. 2 [109]Open in a new tab Isolation, comprehensive characterization, and cellular internalization of BMSC-derived extracellular vesicles. (A) A schematic illustration depicting the cultivation of BMSCs, their hypoxic preconditioning, and the subsequent extraction of EVs via ultracentrifugation. (B) Representative transmission electron microscopy images depicting the morphological architecture of N-EVs and H-EVs. (C) Nanoparticle tracking analysis revealed that the particle sizes of both N-EVs and H-EVs predominantly range from 50 to 150 nanometers. (D) Quantitative analysis of particle concentrations in N-EVs and H-EVs. (E) Assessment of protein concentration in N-EVs and H-EVs using the BCA assay kit. (F) The expression levels of the negative EV markers Calnexin and GM130 were examined across different groups via Western blot analysis. (G) ImageJ was employed to quantify and compare the differential expression of Calnexin and GM130 among the N-MSC, H-MSC, N-EVs, and H-EVs groups. The sample size of each group was n = 3. (H) Detection of the expression of specific EVs markers in N-EVs and H-EVs via WB analysis. (I) Statistical analysis of WB results showed that H-EVs marker levels were significantly higher than those of N-EVs. The sample size of each group was n = 3. (J) ELISA analysis of anti-inflammatory N-EVs and H-EVs effects in LPS-stimulated RAW264.7. (K) Immunofluorescence staining was utilized to visualize the differential uptake of EVs by distinct cell types within the brain. (L) Fluorescence intensity reflecting EVs uptake by microglia, neurons, and astrocytes was quantitatively assessed using ImageJ to elucidate cell type-specific differences. (M) EVs were labeled with the lipophilic dye PKH67 and subsequently co-cultured with microglial cells. The process of EVs uptake and cellular metabolism was meticulously captured using Fluorescence microscope. (N) Fluorescence intensity of EVs internalized by microglial cells in the N-EVs and H-EVs groups was analyzed using ImageJ software. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ****P < 0.001; ***P < 0.005; **P < 0.01; *P < 0.05 H-EVs alleviated brain injury and promoted brain function recovery in vivo To elucidate the impact of H-EVs on brain injury and neurological dysfunction following TBI, hyaluronidase was administered intranasally immediately after TBI, facilitating a reduction in intercellular matrix viscosity and enhancing tissue fluid permeability. Treatment was followed by intranasal delivery of EVs. First, the injured area of TBI rats was observed using the MRI T2 sequence. The results showed that the N-EVs and H-EVs groups had a significantly reduced edema starting on the third day compared to the PBS group. Interestingly, the treatment effect of the H-EVs group was superior compared to that of the N-EVs group on day 14 (Fig. [110]3A-D). In addition, MRS analysis showed that the NAA (N-acetylaspartate) /Cr (Creatine) value of the H-EVs group was significantly higher than that of the N-EVs and PBS groups, suggesting that the neuronal activity of the H-EVs group was recovered better. The Glx (Glutamine)/Cr value of the H-EVs group was significantly lower than that of the N-EVs and PBS groups. This indicated that the degree of ischemia and hypoxia in brain tissue was lower in the H-EVs group (Figure [111]S2A). The HE and Nissl staining results showed that both the N-EVs and H-EVs groups had varying degrees of neuronal loss and brain tissue edema. However, the treatment effect of the H-EVs group was significantly better than that of the N-EVs group compared to the effect observed in the PBS group (Fig. [112]3B-C). LSCI results demonstrated that CBF in the injured region and the surrounding cortex was severely compromised in the PBS group, whereas a remarkable recovery in CBF was observed following H-EVs treatment. Moreover, CBF recovery in the H-EVs group was more pronounced compared to that in the N-EVs group (Fig. [113]3E-F). This suggests that the H-EVs treatment can restore cerebral perfusion and improve oxygen deficit after TBI. In addition, inflammation evaluation in the injured area and surrounding tissues revealed that the H-EVs treatment inhibited the release of pro-inflammatory factors TNF-α, IL-6, and IL-1β and promoted the expression of anti-inflammatory factors IL-4, IL-10, and TGF-β (Fig. [114]3G-I). Simultaneously, Microtubule-Associated Protein 2 (MAP2) staining was employed to assess neuronal damage in brain tissue across different treatment groups following TBI (Fig. [115]3J-K). The results showed that neuronal damage loss in the N-EVs and H-EVs groups was improved compared to that in the TBI group, but the effect of the H-EVs group was more significant. Finally, neurological function recovery in different treatment groups was initially evaluated using the mNSS. The N-EVs group showed notable improvement in neurological function on day 28 compared to that in the PBS group. However, the H-EVs group demonstrated noteworthy improvement in neurological function starting on day 3 compared to that in the PBS group, (Fig. [116]3L). The above results suggest that H-EVs treatment can significantly promote the recovery of brain injury and the recovery of neurological function after TBI. Fig. 3. [117]Fig. 3 [118]Fig. 3 [119]Open in a new tab H-EVs treatment ameliorated brain injury and facilitated the restoration of brain function. (A) MRI scans were employed to observe the size and edema of the injury sites during the acute and subacute phases in the Control, TBI + PBS, TBI + H-EVs, and TBI + H-EVs groups. (D) A comparative analysis was conducted to examine the differences in the size and edema of the injury sites during the acute and subacute phases across the Control, TBI + PBS, TBI + H-EVs, and TBI + H-EVs groups. (B) Nissl staining was employed to analyze the differential expression of Nissl bodies in neurons across various treatment groups, reflecting the extent of neuronal injury. (C) H&E staining was employed to observe the morphological differences in tissue architecture at the injury sites across the various treatment groups. (E) Laser Speckle Contrast Imaging (LSCI) was utilized to observe the CBF in the TBI injury regions across different treatment groups and time points. (F) Repeated measures analysis of variance (ANOVA) was employed to compare the differences in CBF recovery across the various groups. (G) Elisa was utilized to assess the expression of anti-inflammatory and pro-inflammatory factors across the different treatment groups. (H) WB was utilized to assess the expression of anti-inflammatory and pro-inflammatory factors across the different treatment groups. (I) Image J was used to analyze protein expression differences among different groups. The sample size of each group was n = 5. (J) MAP2 staining of neurons was employed to examine neuronal damage in the various treatment groups following TBI. (K) ImageJ analysis was employed to assess the differences in neuronal survival at the injury sites across the various treatment groups. (L) Neurological function recovery in each group of TBI rats from the time of injury to day 28 was evaluated based on mNSS. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 H-EVs induced M2 polarization of microglia and inhibited neutrophil extracellular trap formation to alleviate inflammatory infiltration in vivo and in vitro Microglia are the most important type of resident immune cells in the central nervous system (CNS). They exert pro-inflammatory (M1) and anti-inflammatory (M2) functions via different polarization states and can respond quickly to brain injury or infection. To evaluate the effect of EVs on polarization of reactive microglia after TBI, samples were obtained on day 7 after modeling and immunological staining was performed to assess microglial polarization. N-EV and H-EV immunofluorescence staining of brain tissue significantly reduced the percentage of iNOS-positive M1 microglia and increased the proportion of Arg1-expressing M2 microglia in Iba-1-positive cells compared to those in the PBS group. H-EVs treated fewer M1 microglia and increased the proportion of M2 microglia at the injury site (Fig. [120]4A-C). Fluorescence staining for myeloperoxidase (MPO), a marker of neutrophil extracellular traps (NETs), was also performed to assess the extent of NET infiltration in the injured area. The results showed that MPO expression in the N-EVs and H-EVs groups was significantly lower than that in the PBS group. MPO fluorescence intensity was lower in the H-EVs-treated group than in the N-EVs-treated group at the injury site (Fig. [121]4D-E). Western blot experiments were used to analyze the effects of N-EVs and H-EVs on classical inflammatory pathway proteins. The results showed that the levels of classical pro-inflammatory factors, such as TLR4, NLRP3, pp-65, p65MyD88, ASC, and NETs marker MPO, in the N-EVs and H-EVs groups were decreased to varying degrees compared to those in the PBS group. Moreover, the decrease was more obvious in the H-EVs group (Fig. [122]4F-G). In addition, differences in microgliosis and neuronal loss in the cortex and deep hippocampus between treatment groups were identified by immunohistochemistry. Compared to the PBS group, H-EVs significantly reduced the proliferation of microglia (CD68+) in the cortex and hippocampal DG areas, and the loss of neurons in the hippocampal CA3 and DG areas was less pronounced and more closely arranged (Fig. [123]4H-I). Peripheral blood samples were obtained from the medial canthus vein on day three post-injury to assess the distribution of inflammatory cells between the different treatment groups. The results showed that the number of inflammatory cells in the peripheral blood of the H-EVs group was significantly increased compared to that in the control group. However, the number of inflammatory cells in the peripheral blood of the H-EVs group was significantly decreased compared to that in the PBS and N-EVs groups (Table [124]S1). Finally, microglial cells were stimulated with LPS to induce an inflammatory state in vitro. The results showed that iNOS was significantly down-regulated and Arg1 was up-regulated in the H-EVs group compared to the levels in the PBS group. In addition, the expression of pro-inflammatory factors in H-EVs was decreased, and anti-inflammatory factors were significantly up-regulated compared to those in the PBS group (Fig. [125]5A-E). These results suggest that H-EVs can inhibit inflammatory cell infiltration and reduce neuroinflammation by inducing microglial M2 polarization both in vivo and in vitro. Fig. 4. [126]Fig. 4 [127]Fig. 4 [128]Open in a new tab H-EVs inhibited inflammatory cell infiltration and alleviated neuronal apoptosis in vivo. (A-B) Immunofluorescence was utilized to examine the expression of iNOS and Arg1 in microglial cells across the different treatment groups. (C) ImageJ software was employed to analyze the differential expression of IBA-1 and iNOS double-positive cells, as well as IBA-1 and Arg1 double-positive cells, across the various treatment groups. (D) Immunofluorescence was employed to observe the expression of MPO, a marker of neutrophil extracellular traps (NETs) activation, in the infiltrated neutrophils within the injury regions across the various treatment groups. (E) ImageJ was utilized to analyze the differential fluorescence intensity of MPO-positive expression following NETs activation in the infiltrated neutrophils across the various treatments. (F) WB was employed to assess the expression levels of the classical inflammatory proteins TLR4, NLRP3, pp-65, p-65, MyD88, ASC, and MPO in the injured brain tissue across the various treatment groups. (G) ImageJ software was employed to quantitatively assess the expression of inflammatory proteins across the groups, followed by statistical analysis. The sample size of each group was n = 5. (H) Immunohistochemical analysis of CD68 + microglia in the cortex and DG regions and NeuN staining to observe neuronal loss in CA3 and DG regions. (I) Expression of CD68 + microglia and neuronal marker NeuN in each group was analyzed by ImageJ. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 Fig. 5. [129]Fig. 5 [130]Fig. 5 [131]Open in a new tab In vitro, H-EVs treatment effectively attenuates the expression of inflammatory proteins and fosters the polarization of microglial cells toward the M2 phenotype. (A) Microglial cells were activated using LPS and subsequently subjected to interventions with variously preconditioned EVs. The expression levels of iNOS and Arg1 were then elegantly delineated across treatment groups via WB. (B) ImageJ was employed to quantitatively assess the differential expression of iNOS and Arg1 proteins in microglial cells across the various treatment groups. The sample size of each group was n = 3. (C, D) Microglial cells were stimulated with LPS, followed by a series of therapeutic interventions. Subsequently, the expression levels of IL-1β, IL-6, TNF-α, TGF-β, IL-10, and IL-4 were assessed via ELISA and Western blot analysis. (E) Expression differences in inflammatory factors in each group were analyzed using ImageJ. The sample size of each group was n = 3.All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 H-EVs alleviated endothelial cell dysfunction and promoted cerebrovascular recovery by regulating microglial polarization The above data showed that treatment with H-EVs restored partial cortical CBF (Fig. [132]3E). Microglia are important inflammatory cytokine-secreting cells. After TBI, their activation triggers the secretion of a variety of pro-inflammatory factors, which directly damages cerebral vessels and endothelial cells, leading to vascular dysfunction, blood-brain barrier destruction, and decreased endothelial cell activity [[133]38–[134]41]. Next, it was investigated whether H-EVs could regulate microglial polarization to alleviate endothelial dysfunction and promote cerebrovascular recovery. LPS-stimulated microglial M1 polarization was followed by co-culture with brain microvascular endothelial cells and EVs addition (Fig. [135]6A). Flow cytometry apoptosis assay showed that the proportion of apoptotic endothelial cells in the N-EVs and H-EVs groups was significantly reduced compared to that in the PBS group (Fig. [136]6B). Tube formation assay was used to evaluate the function of endothelial cells and showed that the tube length and number of branch points were significantly increased in the H-EVs group compared to the values in the PBS group. Interestingly, no difference was observed between the N-EVs treatment and PBS groups. It is possible that some special H-EVs molecules play an important role in rescuing endothelial cell function (Fig. [137]6C-D). Next, the levels of iNOS protein (M1), Arg1 protein (M2), and sydecan-1 (endothelial cell marker) in the microglia stimulated above were detected using WB. The results suggested that compared to the PBS-treated group, N-EVs and H-EVs significantly inhibited the microglial M1 polarization, promoted the M2 polarization, and significantly increased the expression of sydecan-1 in endothelial cells, suggesting the improvement in endothelial cell function (Fig. [138]6E-F). Furthermore, TEM analysis of brain tissue sections from TBI rats revealed severe mitochondrial damage in endothelial cells within the PBS group, characterized by disrupted outer and inner membranes and the complete loss of cristae. In contrast, mitochondria in the H-EVs treatment group remained relatively intact (Fig. [139]6G). ELISA assays were employed to evaluate the expression of vascular functional markers—VEGF, eNOS, and Ang-1—within the injured brain region. The results revealed that, in comparison to the PBS and N-EVs groups, the H-EVs group exhibited a pronounced upregulation of these vascular indicators, reflecting a substantial enhancement in vascular function (Fig. [140]6H). Immunofluorescence staining for endothelial and smooth muscle cell markers in cortical cerebral vessels showed that the level of CD31 (a marker protein of endothelial cells in blood vessels) was significantly up-regulated in the H-EVs group compared to that in the PBS group, while the level of α-SMA, which is mainly expressed in smooth muscle cells in blood vessels and is responsible for regulating blood vessel contraction, was significantly up-regulated in the H-EVs group compared to the level in the PBS group. Further evidence confirmed that H-EVs treatment facilitated the restoration of CBF, inducing a transition of blood vessels from a constricted to a dilated state. This may be the main reason for the recovery of CBF after the H-EVs treatment (Fig. [141]6I, K). As the H-EVs treatment significantly restored CBF, it also indirectly restored tissue oxygen content in the injured and surrounding areas. To this end, TUNEL staining of neurons was performed in the tissue surrounding the injury, and the results were consistent with the above outcomes. The neuron loss in the H-EVs group was significantly reduced compared to that in the PBS group (Fig. [142]6J, L). Finally, CD31, sydecan-1, and α-SMA content in the tissues was evaluated using WB, and the results were consistent with the above findings. CD31 and sydecan-1 were significantly up-regulated, while α-SMA was down-regulated in the H-EVs group (Fig. [143]6M, N). These findings suggest that H-EVs may alleviate endothelial cell dysfunction by regulating microglial polarization, thereby restoring cerebrovascular function. Fig. 6. [144]Fig. 6 [145]Fig. 6 [146]Open in a new tab H-EVs regulate microglial M2 polarization and alleviate endothelial cell dysfunction in vitro and in vivo. (A) Schematic illustration of microglial-endothelial co-culture following LPS-induced activation. (B) Flow cytometric assessment of endothelial cell apoptosis following co-culture with microglia treated with EVs derived from distinct preconditioning protocols. (C) Tube formation ability of endothelial cells in different treatment groups was analyzed using tube formation assay. (D) Tube length and number of branch points among the groups were analyzed by ImageJ. (E) WB analysis of iNOS and Arg1 expression in microglia across various preconditioning groups during co-culture, along with the evaluation of syndecan-1 expression in endothelial cells as a marker of glycocalyx integrity. (F) ImageJ analysis of iNOS, Arg1, and sydecan-1 expression differences in each group. The sample size of each group was n = 3. (G) Mitochondrial damage in endothelial cells in the injured tissue of TBI rats in different treatments was observed by TEM. (H) ELISA assays were conducted to assess the differential expression of vascular functional markers VEGF, eNOS, and Ang-1 within the injured regions across various treatment groups. (I)Immunofluorescence was used to determine the expression of CD31 and α-SMA in the injured tissue of TBI rats after different treatments. (K) ImageJ was employed to quantify the differential fluorescence intensities of CD31 and α-SMA within injured tissues across various treatment groups following TBI. (J) TUNEL staining was used to observe neuronal apoptosis in the injured tissue of TBI rats after different treatments. (L) Quantitative analysis of neuronal apoptosis within the injured regions across distinct treatment groups was performed using ImageJ based on TUNEL staining. (M) TBI tissues from rats in each treatment group were harvested, and the expression levels of CD31, α-SMA, and syndecan-1 were examined via WB. (N) ImageJ analysis of CD31, α-SMA, and syndecan-1 expression differences in each group. The sample size of each group was n = 5. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 Hypoxic preconditioning significantly promoted miR-145-5p expression in BMSC-EVs Previous studies have shown that miRNA are important components of BMSC-EVs and are involved in cell communication, inflammation regulation, and tissue regeneration. To evaluate the effect of hypoxia on the miRNA profile in BMSC-EVs, miRNA microarray analysis was performed on BMSC-EVs under hypoxic and normoxic conditions. Compared to N-EVs, H-EVs exhibited an upregulation of 54 miRNAs and a downregulation of 23 miRNAs (Fig. [147]7A-B). Since EVs carry miRNA that mediate their biological effects, the study focused on miRNAs that are up-regulated in H-EVs compared to N-EVs. Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway enrichment analysis showed that the target mRNAs of the differentially expressed miRNAs were enriched in the NF-κB signaling pathway, TGF-beta signaling pathway and Toll-like receptor signaling pathway that could regulate inflammatory response (Fig. [148]7C). Furthermore, Gene Ontology (GO) analysis revealed that the identified target mRNAs play a pivotal role in regulating cellular behaviors by influencing key biological processes, including IκB kinase/NF-κB signaling, nervous system development, and endothelial cell proliferation (Fig. [149]7D). The resulting histogram shows the top 15 miRNA significantly up-regulated in H-EVs as determined by microarray analysis (Fig. [150]7E). Microarray analysis revealed that miR-411-5p showed the most significant difference, while external environmental factors, such as pH and temperature, may affect vesicle integrity and miRNA stability. Thus, it is uncertain whether EVs are still dominated by miR-411-5p after entering microglia. The 15 miRNA significantly up-regulated in microarray analysis were verified by qPCR in microglia treated with N-EVs and H-EVs. The results showed that miR-145-5p was the most significantly up-regulated miRNA in the H-EVs group compared to the N-EVs group (Fig. [151]7F). Previous studies have shown that miR-145-5p has anti-inflammatory effects by regulating inflammation-related signaling pathways, which is suitable for the treatment of a variety of diseases. For instance, MSC-EVs enriched with miR-145-5p modulate the TLR4/NF-κB pathway, thereby mitigating inflammation in spinal cord injury. Additionally, they attenuate inflammation and apoptosis in myocardial ischemia-reperfusion injury by suppressing NADPH oxidase homologs. By targeting CXCL16, they regulate the AKT/GSK pathway, effectively curbing RMC proliferation and inflammation. Moreover, geniposide safeguards PC12 cells from LPS-induced inflammatory damage through the upregulation of miR-145-5p [[152]42–[153]45]. These findings indicate that miR-145-5p may serve as a crucial mediator of the anti-inflammatory and therapeutic properties of H-EVs. Fig. 7. [154]Fig. 7 [155]Open in a new tab Hypoxia significantly altered the miRNA array of BMSC-EVs. (A) Heat map showing 54 up-regulated and 23 down-regulated miRNAs in H-EVs compared to N-EVs, with the difference of ≥ 1.5-fold. (B) Volcano plot showing significant miRNA up-regulation (red dots) and down-regulation (green dots) in H-EVs compared to N-EVs. (C) Enriched pathways for target genes of miRNAs enriched within H-EVs in KEGG pathways. (D) GO analysis of target mRNAs of differentially expressed miRNAs. (E) The histogram illustrates the top 15 miRNAs most significantly upregulated in H-EVs compared to N-EVs. (F) qPCR was used to verify up-regulation levels of the top 15 miRNAs in microglia after N-EVs and H-EVs treatment. n = 3. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ****P < 0.001; ***P < 0.005; **P < 0.01; *P < 0.05 Inhibition of miR-145-5p expression reversed anti-inflammatory and therapeutic H-EVs effects in vitro To ascertain whether the biological activity of H-EVs is contingent upon miR-145-5p, microglia were first stimulated with LPS and subsequently treated with EVs. The expression levels of miR-145-5p in each group were quantified via qPCR after 24 h. The results showed that the level of miR-145-5p was significantly up-regulated in the LPS + H-EVs group compared to the level in the other groups (Fig. [156]8A). Next, miR-145-5p Inhibitor was transfected into BMSCs and EVs were collected for further experiments 24 h later. The qRT-PCR results showed that the miR-145-5p level was significantly reduced in EVs after Inhibitor transfection (Fig. [157]8B). Subsequently, Western blot and ELISA analyses were performed to assess the inflammatory profile of microglia. The findings revealed a marked upregulation of pro-inflammatory cytokines and a concurrent suppression of anti-inflammatory mediators in the LPS + Inhibitor (miR-145-5p) + H-EVs group, in stark contrast to the LPS + NC + H-EVs group (Fig. [158]8C-E). This indicates that miR-145-5p inhibition in H-EVs can significantly reduce the anti-inflammatory effect of H-EVs. Next, different treatment groups of microglial cells were co-cultured with endothelial cells to evaluate the effect of inhibiting miR-145-5p expression on endothelial cells in co-culture. Western blot results showed that the level of microglial Arg1 in LPS + Inhibitor + H-EVs group was significantly decreased, while iNOS level was significantly increased compared to that in the LPS + NC + H-EVs group. In endothelial cells, sydecan-1 levels were decreased in the LPS + Inhibitor + H-EVs group compared to those in the LPS + NC + H-EVs group (Fig. [159]8F-G). Flow apoptosis analysis of endothelial cells showed that the proportion of apoptosis in the LPS + Inhibitor + H-EVs group was significantly higher than that in the LPS + NC + H-EVs group (Fig. [160]8H). The endothelial cell tube formation assay results confirmed that the tube formation length and number of branch points were significantly reduced in the LPS + Inhibitor + H-EVs group compared to those in the LPS + NC + H-EVs group (Fig. [161]8I-J). Finally, microglial cells from different treatment groups were co-cultured with primary neurons (Figure [162]S2B). The immunofluorescence results suggested that the positive number of NeuN and synapses in the LPS + Inhibitor + H-EVs group was significantly reduced compared to that in the LPS + NC + H-EVs group (Fig. [163]8K-L). These results suggested that inhibition of miR-145-5p expression significantly reduced the anti-inflammatory effect of H-EVs and the functional recovery of endothelial cells in vitro. Fig. 8. [164]Fig. 8 [165]Fig. 8 [166]Open in a new tab Inhibition of miR-145-5p expression reversed the therapeutic effect of H-EVs in vitro. (A) qPCR was used to detect the expression level of miR-145-5p in each treatment group after LPS stimulation of microglia. (B) Inhibitor miR-145-5p was added to BMSCs and the Inhibitory effect of miR-145-5p in EVs was determined by qPCR. (C) WB analysis was employed to assess the expression of IL-1β, IL-6, TNF-α, TGF-β, IL-10, and IL-4 in microglial cells across different groups following LPS stimulation. (D) ImageJ was utilized to analyze and statistically compare the differential expression of IL-1β, IL-6, TNF-α, TGF-β, IL-10, and IL-4 across the various groups. The sample size of each group was n = 3. (E) Elisa was employed to quantify the inflammatory cytokine profiles of microglial cells across the various treatment groups. (F)Endothelial cells were co-cultured with microglia from different treatment groups, and expression of iNOS and Arg1 in microglia and sydecan-1 in endothelial cells was observed using WB. (G) ImageJ was employed to statistically evaluate the differential expression of iNOS, Arg1, and Syndecan-1 across the various groups. The sample size of each group was n = 3. (H) Flow cytometry was employed to assess the apoptosis rate of endothelial cells following a 48-hour co-culture with microglial cells subjected to various interventions. (I) Tube formation ability of endothelial cells in different treatment groups was analyzed via tube formation assay. (J) Tube formation length and number of branch points among groups were analyzed using ImageJ. (K) Microglial cells, stimulated with LPS, were co-cultured with neurons to observe neuronal survival and alterations in synaptic morphology. (L) ImageJ was used to analyze differences in NeuN fluorescence intensity expression among groups. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 miR-145-5p in H-EVs inhibited inflammatory infiltration and promoted brain function recovery in TBI rats in vivo To elucidate the pivotal role of miR-145-5p in mediating anti-inflammatory effects and promoting neural recovery following H-EV treatment for TBI, Antagomir-miR-145-5p was employed to suppress its function. Preliminary evaluation of the effect using MRI, MRS, histological staining, and LSCI showed that Antagomir significantly reversed the therapeutic effect of H-EVs. Compared to the TBI + NC + H-EVs group, the TBI + Antagomir + H-EVs group had more obvious edema in the injured brain area on day 14 (Fig. [167]9A-B). With reduced neuronal activity recovery, greater cerebral ischemia and hypoxia (Figure [168]S2C), and destruction of cellular structures in and around the injured tissue, there was substantial neuronal loss (Figure [169]S3A-B) and CBF recovery was inhibited (Fig. [170]9C-D). The integrity of the blood-brain barrier across different treatment groups was evaluated via Western blot analysis. Results demonstrated that the expression levels of key tight junction proteins Claudin-5, Occludin, and ZO-1 were markedly elevated in the TBI + NC + H-EVs group compared to the TBI + Antagomir + H-EVs group (Fig. [171]9E-F). Similarly, immunofluorescence analysis showed that Arg1 was significantly down-regulated, while iNOS was significantly up-regulated in the TBI + Antagomir + H-EVs group compared to the TBI + NC + H-EVs group (Fig. [172]9G-I). Subsequently, immunohistochemical analysis of microgliosis and neuronal loss in the cortex and hippocampus showed that the microgliosis marker CD68 + was significantly up-regulated in the cerebral cortex and hippocampal DG region in the TBI + Antagomir + H-EVs group compared to the TBI + NC + H-EVs group. Moreover, neuron loss in the hippocampal DG and CA3 regions was more serious (Fig. [173]9J-K). Western blot results showed that the levels of major inflammatory pathway marker proteins, such as TLR4, NLRP3, pp-65, p-65, MyD88, and ASC, in the TBI + Antagomir + H-EVs group were significantly increased compared to those in the TBI + NC + H-EVs group. There were also increases in the M1 microglial marker iNOS and NET marker MPO (Figure [174]S3C-F). The differences in blood vessel function were further analyzed using immunofluorescence staining. The α-SMA expression was up-regulated, constriction degree of blood vessels was increased, and CBF was decreased in the TBI + Antagomir + H-EVs group compared to those in the TBI + NC + H-EVs group (Fig. [175]9L-M). Consistent with the above outcomes, the WB results showed that α-SMA expression was up-regulated in the TBI + Antagomir + H-EVs group, while the endothelial cell markers CD31 and sydccan-1 were down-regulated (Fig. [176]9N-O). In addition, TEM observations showed that compared to the H-EVs group, the TBI + Antagomir + H-EVs group demonstrated more severe damage to mitochondrial function due to the breakage of the outer and inner endothelial cell membranes and the disappearance of cristae on the inner membrane (Figure [177]S3G). Finally, the neurological function of TBI rats was evaluated based on the mNSSs. The results showed that the mNSS was higher in the TBI + Antagomir + H-EVs group compared to that in the TBI + NC + H-EVs group, and the difference became apparent on day 14 (Figure S4A). The Morris water maze results suggested that the rats in the TBI + Antagomir + H-EVs group took longer to find the platform in the navigation test compared to those in the TBI + NC + H-EVs group. The rats in the probe trial had a shorter trajectory in the target quadrant and lower frequency of crossing the platform (Figure S4B-D). The Open Field Test showed that the rats in the TBI + Antagomir + H-EVs group spent less time in the central area and ran a shorter distance on the track compared to those in the TBI + NC + H-EVs group (Figure S4E-F). These results indicated that Antagomir-miR-145-5p inhibited neurological function and spatial learning and memory and promoted anxiety behavior in rats. Western blot analysis of the expression of apoptotic protein caspase-3 and Bax and anti-apoptotic protein Bcl-2 in the injured area on day 14 showed that caspase-3 and Bax were significantly up-regulated, while Bcl-2 was down-regulated in the TBI + Antagomir + H-EVs group compared to those in the TBI + NC + H-EVs group (Figure: S4G-H). The above experimental results collectively demonstrated that miR-145-5p in H-EVs inhibited inflammatory infiltration and promoted brain function recovery in TBI rats in vivo. Fig. 9. [178]Fig. 9 [179]Fig. 9 [180]Fig. 9 [181]Open in a new tab Antagomir-miR-145-5p significantly reversed the therapeutic effect of H-EVs on inflammatory response and brain function recovery in TBI rats in vivo. (A) MRI scans were employed to observe the size and edema of the injury sites during the acute and subacute phases in the Control, TBI + PBS, TBI + Antagomir + H-EVs, and TBI + NC + H-EVs groups. (B) A comparative analysis was conducted to examine the differences in the size and edema of the injury sites during the acute and subacute phases across the Control, TBI + PBS, TBI + Antagomir + H-EVs, and TBI + NC + H-EVs groups. (C) Laser Speckle Contrast Imaging (LSCI) was utilized to observe the CBF in the TBI injury regions across different treatment groups and time points. (D) Repeated measures analysis of variance (ANOVA) was employed to compare the differences in CBF recovery across the various groups. (E) Western blot analysis was performed to examine the expression of blood-brain barrier integrity markers Claudin-5, Occludin, and ZO-1 across the various treatment groups. (F) ImageJ was utilized to quantitatively analyze the intergroup variations across the different treatment groups. The sample size of each group was n = 5. (G),(H)Immunofluorescence staining was employed to visualize the expression of Arg1⁺ and iNOS⁺ microglia within TBI-lesioned brain tissues of rats subjected to various therapeutic interventions. (I) ImageJ analysis of Arg1⁺ and iNOS⁺ fluorescence intensity expression differences among groups. (J) Immunohistochemical staining was employed to examine the expression of CD68-positive microglia within the cortical and DG regions, as well as NeuN-positive neurons in the CA3 and DG areas of TBI-lesioned rats across different treatment groups. (K) ImageJ was utilized to quantitatively assess intergroup disparities in the expression of CD68- positive microglia and NeuN-positive neurons across distinct brain regions. (L) Immunofluorescence analysis of CD31 and actin α-SMA expression in vascular endothelial cells of TBI rats during different treatments. (M) ImageJ was used to analyze expression differences in CD31 and α-SMA fluorescence intensity among groups. (N) CD31, α-SMA, and sydecan-1 expression in brain tissues of TBI rats in different treatment groups were analyzed using WB. (O) ImageJ was used to analyze expression differences in CD31, α-SMA, and sydecan-1 among groups. The sample size of each group was n = 5. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 Single-cell sequencing illuminated the intricate mechanisms underpinning H-EVs therapy To further investigate the specific regulatory mechanisms underlying H-EVs’ therapeutic effects in TBI rats, and particularly to pinpoint the pathway through which upregulated miR-145-5p exerts its beneficial influence, scRNA-Seq analysis was conducted on the injured tissues from various treatment groups. Data sets were combined and filtered based on transcript count values between 200 and 30,000, gene numbers between 200 and 5,000, mitochondrial proportion of < 20%, and hemoglobin proportion of < 5% (Figure S5A). The filtered data were subsequently standardized and the top 2,000 genes with the largest differences were subjected to principal component analysis. The top 15 PCs were selected for tSNE dimensionality reduction clustering to obtain 36 cell subsets (Figure S5B-D). Annotation by cell markers (Figure S5E) revealed that the major cell clusters were endothelial cells, microglia, neutrophils, T cells, pericytes, B cells, astrocytes, and oligodendrocytes (Fig. [182]10A). The H-EVs group had fewer microglia and more endothelial cells and neutrophils compared to the PBS group (Fig. [183]10B). Microglia are a major type of immune cells in the central nervous system. They are responsible for recognizing pathogens and injury signals and upon activation secrete pro- and anti-inflammatory factors. However, their excessive activation may trigger neuroinflammation and damage neurons and the blood-brain barrier, which in turn aggravates neurological disease progression. For this purpose, 9,871 microglial cells from the PBS group and 7,365 microglial cells from the H-EVs group were individually extracted for dimensionality reduction and clustering analysis (Fig. [184]10C). Gene mapping revealed that the anti-inflammatory genes Arg1, CD206, and TGF-β in microglia in the H-EVs group showed high expression, while the pro-inflammatory genes CD86, CD80, and MHC-II demonstrated low expression compared to those in the PBS group (Fig. [185]10D). In addition, neutrophils have a dual role in the inflammatory process, with N1 type playing a pro-inflammatory role and N2 type playing an anti-inflammatory role. Thus, neutrophils from the two groups were extracted separately for dimensionality reduction clustering (Fig. [186]10E). The results showed that N2 anti-inflammatory cells were significantly up-regulated in the H-EVs group compared to those in the PBS group (Fig. [187]10F). After TBI, microglia were rapidly activated and secreted pro-inflammatory factors, such as TNF-α, IL-1β, and IL-6, which attract neutrophil aggregation and damage endothelial cells. Cell communication among microglia, endothelial cells, and neutrophils was attenuated in the H-EVs group compared to that in the PBS group (Fig. [188]10G). Finally, differential microglial gene screening between the PBS and H-EVs groups showed that the reduction in transcription factor SP1 level was the most significant in the H-EVs group (Fig. [189]10H-I). The injured brain tissues in the two groups were evaluated using western blot detection, which showed that SP1 level in the H-EVs group was significantly decreased compared to that in the PBS group (Fig. [190]10J-K). These results indicated that the H-EVs treatment induced microglial M2 polarization, enhanced anti-inflammatory function, and significantly suppressed transcription factor SP1 expression. Since H-EVs are mainly miR-145-5p in play a role, the treatment resulted is a significant reduction in transcription factor SP1 level. Therefore, it was necessary to determine whether miR-145-5p directly targeted SP1, resulting in its suppression. Fig. 10. [191]Fig. 10 [192]Fig. 10 [193]Open in a new tab Molecular mechanism of anti-inflammatory H-EVs therapy revealed using single-cell sequencing. (A) Single-cell sequencing analysis of the main cell population and proportion in the injured tissue in PBS and H-EVs treatment groups. (B) Clustering of cells in damaged tissues after PBS and H-EVs treatments based on cell marker genes. (C) Microglia were extracted for dimensionality reduction cluster analysis to uncover differentially expressed genes. (D) Inflammation-associated genes—including CD80, CD86, MHC II, ARG1, CD206, and TGF-β—were annotated in microglia following treatment with PBS and H-EVs. (E) Neutrophils infiltrating the injured brain tissues of TBI rats from distinct treatment groups were extracted and subjected to dimensionality reduction and clustering analyses. (F) Marker genes were used to annotate neutrophil proportion in different polarization states N1 and N2. (G) Alterations in intercellular communication within the injured brain tissues of TBI rats following PBS and H-EVs treatment were meticulously examined. (H), (I) In comparison to the PBS group, microglia from TBI rats treated with H-EVs exhibited a distinct transcriptional profile, marked by the upregulation and downregulation of specific gene sets. (J) SP1 expression in TBI-injured brain tissues was extracted and validated via Western blot, revealing differential levels between the PBS-treated and H-EVs-treated groups. (K) ImageJ analysis of SP1 expression difference between PBS and H-EVs groups. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005 miR-145-5p in H-EVs targets SP1 and regulates NF-κB expression to inhibit inflammatory response Activated NF-κB following TBI enhances the transcription of pro-inflammatory genes, thereby triggering the release of TNF-α, IL-1β, IL-6, and other inflammatory mediators, which in turn exacerbates the inflammatory response. NF-κB enhances the expression of chemokines, attracts neutrophils and macrophages to the injured area, and enhances local inflammation. NF-κB-activated pro-inflammatory factors can also damage cerebrovascular endothelial cells, increase the permeability of blood-brain barrier, and further aggravate inflammation and brain edema. In the present study, the differential pathways were enriched in the NF-κB signaling pathway in both KEGG and GO analyses (Fig. [194]7). Interestingly, single-cell results showed that SP1 was significantly decreased in microglia after the H-EV treatment. Previous studies have shown that the decrease in the transcription factor SP1 level inhibited the transcription of NF-κB, resulting in the reduced NF-κB expression [[195]46, [196]47]. It was hypothesized that miR-145-5p can directly target SP1 and inhibit the expression of NF-κB to alleviate inflammation [[197]46, [198]47]. Using an online target prediction tool, a binding target was identified between miR-145-5p and SP1 in the 3’ untranslated region (UTR) based on the principle of complementary base pairing (Fig. [199]11A). Then, dual luciferase reporter gene assay was used to verify whether miR-145-5p directly targeted SP1 3’ UTR. The results demonstrated that the miR-145-5p transfection mimic markedly diminished the luciferase activity of Sp1-WT, while exerting no significant effect on the luciferase activity of the Sp1-MUT group (Fig. [200]11B), thereby confirming that miR-145-5p directly targets the SP1 3’ UTR. Next, microglia were treated with LPS, and WB results showed that SP1 was significantly decreased in the H-EVs group compared to that in the PBS group (Fig. [201]11C-D). Following transfection of small microglia with lentiviral LV-SP1 overexpression and si-SP1 knockdown, the results showed that SP1 was significantly downregulated in the si-SP1 group compared to PBS, H-EVs, and LV-SP1 (Fig. [202]11E-F). Furthermore, by WB, the inflammation-related proteins were detected, and the results showed that the pro-inflammatory proteins INOS, NLRP3, MyD88, and ASC were significantly downregulated in the si-SP1 group compared to PBS, H-EVs, and LV-SP1, while Arg1 was upregulated (Fig. [203]11G-H). Finally, the above results were verified by qPCR. In microglia, the mRNA SP1 level was significantly decreased in the H-EVs treatment group compared that in the PBS treatment group (Fig. [204]11I). After SP1 overexpression or knockdown in microglia, the results showed that the mRNA SP1 level in the si-SP1 treatment group was significantly lower than that in the PBS, H-EVs, and LV-SP1 treatment groups. The NF-κB mRNA level was also inhibited with the decrease in SP1. This is consistent with the protein experiment results described above (Fig. [205]11J-K). Taken together, these results suggest that miR-145-5p in H-EVs promoted the polarization of M2 microglia and alleviated the progression of inflammation after TBI by targeting SP1. thereby inhibiting the level of NF-κB (Fig. [206]12). Fig. 11. [207]Fig. 11 [208]Fig. 11 [209]Open in a new tab miR-145-5p targets SP1 and then regulates NF-κB expression to inhibit inflammatory response. (A) Predicted binding targets of miR-145-5p and SP1. (B) Luciferase reporter assay was performed to confirm whether SP1 was the target gene of miR-145-5p. (C) Microglia were treated with LPS to detect the expression of SP1 in PBS and H-EVs groups. (D) ImageJ analysis revealed differential SP1 expression across the various treatment groups. (E) Following LPS preconditioning of microglia, WB analysis was conducted to assess the expression levels of SP1 and NF-κB across the LPS + PBS, LPS + H-EVs, LPS + H-EVs + LV-SP1, and LPS + H-EVs + si-SP1 groups. (F) Expression differences of SP1 and NF-κB in the four groups were analyzed using ImageJ. The sample size of each group was n = 3. (G) Western blot analysis was employed to assess the expression levels of inflammatory proteins, including iNOS, NLRP3, Arg1, MyD88, and ASC, in microglia from different treatment groups. (H) ImageJ was used to analyze expression differences in each inflammatory protein in the four groups. The sample size of each group was n = 3. (I) Microglia were stimulated with LPS followed by PBS and H-EV treatment and the difference in SP1 expression was detected by qPCR. (J) Microglia were pretreated with LPS, followed by qPCR analysis to assess the expression of SP1 in the LPS + PBS, LPS + H-EVs, LPS + H-EVs + LV-SP1, and LPS + H-EVs + si-SP1 groups. (K) Microglia were pretreated with LPS, followed by qPCR analysis to evaluate the expression of NF-κB in the LPS + PBS, LPS + H-EVs, LPS + H-EVs + LV-SP1, and LPS + H-EVs + si-SP1 groups. All experiments were performed in triplicate. Data are presented as mean ± SEM. Student’s t-test: ns, no significance; ***P < 0.005; **P < 0.01; *P < 0.05 Fig. 12. [210]Fig. 12 [211]Open in a new tab Regulatory mechanism schematic shows how nasal miR-145-5p delivery in hypoxia-induced BMSC-EVs alleviated the inflammatory response and aided in brain dysfunction recovery after TBI by targeting and negatively regulating the expression of SP1 in microglia, thereby inhibiting NF-kB transcription. Discussion TBI is currently the leading cause of disability and death. The heterogeneity of pathological anatomical subtypes and diversity in pathogenesis and injury degree are important factors leading to the differences in the course and prognosis of TBI [[212]7]. TBI can be divided into primary injury and secondary injury. Primary injury includes diffuse axonal injury, brain contusion and laceration, and intracranial hemorrhage, while secondary injury involves neuroinflammation, neurodegeneration, and delayed neuronal death. Secondary injury may further damage brain structure and function in the non-acute phase of TBI, leading to severe motor and cognitive impairment and irreversible neurological damage [[213]48]. At present, TBI treatment typically includes surgery, electrical nerve stimulation, drugs, and cell therapy, but the effect of surgery and drugs on brain injury improvement is limited. To effectively treat secondary brain injury and the associated pathophysiological disturbances following TBI, the therapeutic application of stem cells and their derived EVs is regarded as a promising approach to ameliorate secondary brain injury during the acute phase post-TBI. In the current investigation, BMSCs were exposed to hypoxic conditions to generate H-EVs. Subsequently, hyaluronidase was intranasally administered to rats for 30 min to facilitate the opening of the nasal-brain pathway. Finally, different BMSC-EVs treatments were delivered via the nasal cavity to bypass the systemic circulation and act directly on the TBI injury site. The results showed that compared to the other treatment groups, the inflammatory response, neurological function, and rat behavior in the H-EVs group were significantly improved. Previous studies have suggested that paracrine mechanisms may contribute to the therapeutic effects of transplanted MSCs [[214]49, [215]50]. EVs are important components of paracrine signaling because they can mediate communication between different cells by transferring RNA and proteins. Previous studies have confirmed that EVs transplantation produces therapeutic effects and functional properties similar to cell transplantation, while avoiding adverse reactions and improving clinical efficacy [[216]51]. Due to their nanoscale size and unique ability to cross the blood-brain barrier, EVs may play an important role in treating neuroinflammation secondary to TBI [[217]52]. In recent years, many studies have confirmed that MSC-EVs can alleviate neuroinflammation and brain function recovery by regulating microglial polarization state in the treatment of TBI and prognosis. Li et al. showed that local injection of EVs secreted by human deciduous tooth stem cells improved inflammatory responses and promoted functional recovery after TBI by altering the M1/M2 polarization of rat microglia [[218]53]. Another study showed that early treatment with a single dose of MSC-derived EVs modulated brain transcriptome to produced neuroprotective changes in a porcine model of TBI and hemorrhagic shock [[219]54]. Thus, MSC-derived EVs may be an effective way to treat the central nervous system. Remarkably, mounting evidence confirms that extracellular vesicles (EVs) derived from mesenchymal stem cells (MSCs) under various preconditioning regimens exhibit markedly enhanced therapeutic efficacy [[220]22, [221]55, [222]56]. Among these, hypoxic preconditioning stands out as a particularly straightforward and potent approach to augment the reparative potential of MSC-EVs, bolstering their immunomodulatory capacity while preserving intrinsic biological characteristics [[223]57]. Hypoxia has been shown to amplify the therapeutic properties of EVs derived from BMSCs, adipose tissue, umbilical cord, and hair follicle stem cells across diverse pathological contexts—ranging from myocardial infarction and spinal cord injury to enhanced angiogenesis, dampened inflammation, and accelerated bone fracture healing [[224]57–[225]61]. Nonetheless, whether hypoxia augments the reparative effects of BMSC-derived EVs specifically in the context of traumatic brain injury (TBI) and its sequelae remains unexplored and warrants further investigation. In our study, hypoxia-exposed BMSCs retained their osteogenic, adipogenic, and chondrogenic differentiation capacities, with no discernible alterations in biological features compared to their normoxic counterparts (S1C–D), aligning with previous findings [[226]57]. Notably, hypoxic conditions stimulated a greater release of EVs from BMSCs, corroborating prior research wherein hypoxia-inducible factor HIF-1α facilitated MSC-EV-mediated angiogenesis [[227]62]. Further, EVs harvested from hypoxia-pretreated MSCs have been shown to modulate microglial M1/M2 polarization, thereby supporting spinal cord injury repair [[228]23]. Consistent with this, our findings revealed that hypoxia-conditioned EVs (H-EVs) significantly promoted microglial M2 polarization post-TBI, surpassing the efficacy of normoxic EVs (N-EVs). Additionally, H-EVs effectively suppressed the formation of neutrophil extracellular traps (NETs), accompanied by a reduction in the proportion of circulating inflammatory cells in peripheral blood.Another study demonstrated that EVs from hypoxia-preconditioned MSCs enhanced cerebral angiogenesis, restored endothelial function, and promoted neurological recovery following focal cerebral ischemia in murine models [[229]63, [230]64]. Similarly, our investigation confirmed that H-EVs facilitated endothelial and neural functional restoration more efficiently than N-EVs. Intriguingly, H-EV treatment significantly downregulated the expression of α-smooth muscle actin (α-SMA), suggesting diminished vasoconstrictive tone and accelerated cerebral blood flow recovery—findings substantiated by laser speckle contrast imaging (LSCI) analysis.Moreover, it has been reported that MSC-derived EVs mitigate neuronal loss in hippocampal CA3 and dentate gyrus (DG) regions and attenuate CD68⁺ microglial proliferation in the cortex and DG after TBI [[231]35]. In harmony with these insights, H-EVs in our study yielded pronounced neuroprotection in CA3 and DG areas and curtailed CD68⁺ microglial expansion in both the cortex and DG. Beyond mitigating neuronal attrition, H-EVs enhanced neurofunctional outcomes, as evidenced by elevated N-acetylaspartate/creatine (NAA/Cr) ratios, and ameliorated cerebral hypoxia, reflected in reduced Glx/Cr levels. In summary, our findings underscore the therapeutic promise of hypoxia-induced BMSC-derived EVs in mitigating neuroinflammation during both the acute and subacute phases following TBI, while concurrently fostering the restoration of neurological function. Through orchestrated modulation of microglial polarization, vascular dynamics, and neuronal resilience, H-EVs emerge as compelling candidates in the pursuit of innovative neuroregenerative therapies. Since a superior therapeutic effect of H-EVs compared to that of N-EVs was observed, our study sought to explore the underlying mechanism of the difference between H-EVs and N-EVs. The miRNA is non-coding microRNA and a key mediator of the therapeutic EV function [[232]65, [233]66]. It has been shown that the beneficial effects of hypoxia-preconditioned MSCs in the treatment of myocardial infarction models are mediated by EVs, and that miR-210 in EVs directly targets AIFM3 to alleviate myocardial infarction [[234]67]. It was also found that EVs secreted by BMSCs contained miR-199a-5p and directly targeted BIP to prevent I/R [[235]68]. However, the differential analysis of the miRNA array between H-EVs and N-EVs and the miRNA-mediated role in alleviating neuroinflammation and aiding in brain function recovery in TBI remains to be investigated. In the present study, miRNA sequencing demonstrated differences in miRNA profiles between H-EVs and N-EVs. Specifically, miR-145-5p was highly expressed in H-EVs compared to N-EVs, and miR-145-5p in EVs was promoted and transferred to target microglia more efficiently after treatment with hypoxia. Therefore, miR-145-5p was enriched in H-EVs, which may be the main reason for the biological differences between H-EVs and N-EVs. In addition, the inhibition of miR-145-5p in H-EVs reversed the beneficial effects of H-EVs in the treatment of secondary brain injury after TBI. It has been reported that miR-145-5p overexpression can alleviate myocardial ischemic injury by inhibiting hypoxia-induced inflammation and injury in cardiomyocytes [[236]69]. Previous studies have shown that MSC-derived EVs containing miR-145-5p attenuate inflammation in SCI by regulating the TLR4/NF-κB signaling pathway [[237]42]. Ma et al. demonstrated that geniposide protected PC12 cells from LPS-induced inflammatory injury by upregulating miR-145-5p [[238]45]. Another report showed that miR-145-5p inhibited the proliferation and migration of vascular smooth muscle cells by directly targeting Smad4 [[239]70]. These results further validate the reliability of the present study results on neuroinflammation relief via miR-145-5p in H-EVs and α-SMA inhibition in vascular smooth muscle cells. This suggests that H-EVs have a strong anti-inflammatory effect that is essential for the inflammatory response, brain function recovery, and nerve regeneration in TBI and may be achieved by delivering miR-145-5p in H-EVs. H-EVs can be used as a biological carrier of functional miR-145-5p to the recipient microglia. Single-cell RNA sequencing is increasingly used in TBI research in recent years, providing unprecedented detail and depth for revealing the complex communication mechanisms between cells and dynamic changes in various cells during injury. Using this technique, researchers can accurately analyze the gene expression profile of specific cell populations, such as microglia, and reveal their communication patterns with other cell types in inflammatory response, tissue repair, and pathological processes. This provides important technical support for in-depth understanding of the molecular mechanism of TBI [[240]71–[241]75]. Moreover, scRNA-seq possesses the remarkable capacity to discern subtle state transitions in microglia following TBI with exceptional precision. As the resident immune cells of the central nervous system, microglia are rapidly activated after TBI and dominate the local inflammatory response. Using scRNA-seq, researchers can determine microglial activation status at different stages and identify the pro- or anti-inflammatory factors secreted by microglia in order to better understand their dynamic role in the pathological process of TBI [[242]73]. Inflammatory factors do not only regulate the local immune response but also affect the overall brain injury repair process by interacting with other cells, such as neurons, astrocytes, and endothelial cells [[243]72, [244]76]. The scRNA-seq can reveal the complex communication networks among these cells and clarify the signaling pathways through which microglia regulate the behavior of surrounding cells, thereby providing important insights into how inflammatory responses promote or inhibit tissue repair [[245]77, [246]78]. To the best of our knowledge, this is the first study to use hypoxia-pretreated BMSC-EVs for TBI treatment and to reveal the potential mechanism of BMSC-EVs in TBI treatment by performing scRNA-seq analysis on the brain tissue of TBI rats after treatment. The scRNA-seq results showed that compared to the PBS group, the anti-inflammatory genes ARG1, CD206, and TGF-β in the H-EVs group microglia were highly expressed, while the pro-inflammatory genes CD86, CD80, and MHC-II were significantly down-regulated, which also verified the anti-inflammatory effect of H-EVs in microglia. Interestingly, neutrophil infiltration was increased in the H-EVs treatment group, which seemed to be paradoxical. Therefore, a secondary dimensionality reduction clustering of neutrophils found that although the number of neutrophils in H-EVs increased, the increased portion of the neutrophils was of the N2 type neutrophils with anti-inflammatory effect. At the same time, cell communication among microglia, endothelial cells, and neutrophils was weakened in the H-EVs treatment group, which may be due to the strong anti-inflammatory ability of H-EVs. Finally, the transcription factor SP1 in microglia was significantly down-regulated in the H-EVs group. SP1 plays an important role in the regulation of inflammatory response, especially by inhibiting the expression of NF-κB [[247]47, [248]79, [249]80]. NF-κB is a key pro-inflammatory factor that usually exacerbates inflammatory responses by activating inflammation-related genes, such as TNF-α and IL-1β [[250]39, [251]81, [252]82]. The results of our study indicated that miR-145-5p targeted SP1 to inhibit the transcriptional level of NF-κB. By interfering with SP1 knockdown and overexpression, SP1 was found to be required for the therapeutic effect of miR-145-5p. Therefore, our results demonstrated that miR-145-5p targeted SP1 to inhibit NF-κB transcription, thereby alleviating neuroinflammation and brain function recovery after TBI. In conclusion, this study showed that H-EVs exhibited more significant efficacy in inhibiting neuroinflammation and promoting brain function recovery after TBI compared to N-EVs. In addition, we explored the potential H-EVs mechanisms for inhibiting inflammatory response and promoting brain function recovery. Hypoxic preconditioning effectively up-regulated the expression of miR-145-5p in H-EVs and significantly enhanced the therapeutic effect of H-EVs on secondary injury after TBI. H-EVs stimulated the M2 polarization of microglia, enhanced the anti-inflammatory response, promoted the survival of neurons and endothelial cells, and inhibited the formation of NETs by alleviating the inflammatory microenvironment after TBI. This process promoted the recovery of brain function after TBI by activating the miR-145-5p/SP1/NF-κB axis. Therefore, we believe that H-EVs-based therapy is potential strategy to deal with secondary brain injury after TBI. Further studies on the spatial and temporal distribution of H-EVs and optimization of their delivery mode are necessary to fully understand their pharmacokinetic characteristics before achieving clinical application. In addition, determining the optimal H-EVs dose and establishing a feasible method for large-scale production are key to promoting a successful clinical translation. Conclusion Hypoxic preconditioning serves as a potent strategy to enhance the functional attributes of BMSC-derived EVs. Intranasal administration of hypoxia-pretreated BMSC-EVs emerges as a compelling therapeutic avenue for mitigating secondary injury following TBI. H-EVs have demonstrated a remarkable capacity to facilitate microglial M2 polarization and attenuate neuroinflammatory responses post-TBI. Moreover, H-EVs actively modulate a spectrum of pathological cascades in the aftermath of TBI and markedly accelerate neurological functional recovery. In conclusion, intranasal administration of H-EVs exerts a salutary effect on alleviating neuroinflammation during the acute and subacute phases of TBI, while enhancing cerebral functional recovery, thereby highlighting their potential for clinical translation. Supplementary Information Below is the link to the electronic supplementary material. [253]Supplementary Material 1^ (209KB, xlsx) [254]Supplementary Material 2^ (68.6MB, pdf) [255]Supplementary Material 3^ (18.9MB, docx) Acknowledgements