Exosomes Derived from Tanshinone IIA-Pretreated Umbilical Cord Mesenchymal Stem Cells Repair Traumatic Spinal Cord Injury by miR-223–5p/USP8/NLRP3 Axis

Abstract

Spinal cord injury (SCI) lacks effective therapies, and umbilical cord mesenchymal stem cell (UCMSC)-derived exosomes show promise but require efficacy optimization. This study explored tanshinone IIA (TSA)-pretreated UCMSC exosomes (T-Exos) to enhance neuroprotection and functional recovery post-SCI. T-Exos were isolated from the TSA-treated UCMSCs. SCI mice received T-Exos, with motor function assessed behaviorally. In vitro and in vivo, microglial NLRP3 inflammasome activity, polarization (M1/M2), and cytokine release were analyzed. miRNA profiling identified key miRNAs in T-Exos; miR-223–5p’s role was validated via inhibition. Mechanistic studies linked miR-223–5p to USP8/NLRP3 using Western blot, qPCR, and co-IP assays. T-Exos improved motor recovery in SCI mice by suppressing NLRP3 inflammasome activation, shifting microglia to anti-inflammatory M2 phenotypes, reducing the level of IL-1β/IL-18, and alleviating oxidative stress. miR-223–5p was highly enriched in T-Exos, and its inhibition reversed therapeutic effects. Mechanistically, miR-223–5p targeted USP8, a deubiquitinase stabilizing NLRP3. By inhibiting USP8, T-Exos reduced NLRP3 levels, dampening neuroinflammation. TSA pretreatment enhances exosomal miR-223–5p, which blocks USP8-mediated NLRP3 stabilization, driving microglial M2 polarization and neuroprotection. These findings highlight T-Exos as a targeted therapy for SCI, offering insights into the exosome-mediated modulation of inflammatory microenvironments for neural repair.

Introduction

Traumatic spinal cord injury (SCI) is a serious disorder affecting the central nervous system that causes irreversible neurological impairments, such as paralysis or death, and places a heavy financial burden on patients and their families. The pathophysiology of SCI is characterized by the occurrence of both primary and secondary injury processes. Primary SCI refers to the immediate mechanical damage to neurons and axons, while secondary SCI is characterized by subsequent neuroinflammation within the injured spinal parenchyma, contributing to ischemia, edema, and progressive neuronal and glial cell loss at the site of injury. Current therapeutic strategies predominantly focus on mitigating the effects of secondary injury; however, their effectiveness is constrained by the restricted capacity of drugs and other interventions to traverse the blood-spinal cord barrier (BSCB). Consequently, few treatments have demonstrated significant efficacy in promoting functional recovery in clinical settings. Therefore, understanding the molecular and cellular mechanisms of SCI is essential. This knowledge will enable the development of therapies that overcome BSCB to improve treatment efficacy and advance rehabilitation protocols for SCI patients.

Mesenchymal stem cell (MSC) transplantation has shown considerable therapeutic potential for SCI. However, the clinical application of direct MSC transplantation is limited by several risks, including cell dedifferentiation, immune rejection, and tumorigenesis. Emerging evidence suggests that the regenerative potential of MSCs is primarily mediated through paracrine mechanisms, with exosomes identified as a key component of this effect. Exosomes facilitate intercellular communication and homeostasis by transporting intracellular cargo, such as proteins and nucleic acids, to recipient cells while also protecting these biomolecules from degradation. Importantly, Exos-based therapies provide therapeutic benefits comparable to those of direct MSC transplantation but without the associated adverse effects. UCMSCs’ capacity to self-renew, accessibility, and variety of biological activity make them a perfect source of exosomes. − Research has indicated that MSC-derived exosomes have therapeutic potential for improving tissue regeneration and controlling inflammatory conditions in the central nervous system. − Furthermore, these nanoscale exosomes can cross the BSCB, thereby enhancing neural recovery following SCI. − To improve therapeutic results, preconditioning using cytokines, pharmaceuticals, hypoxic conditions, or physical stimuli has become a crucial topic of study in exos engineering. , The lipophilic diterpene tanshinone IIA, which is isolated from Salvia miltiorrhiza roots, has anti-inflammatory, antioxidant, antiapoptotic, and BSCB protecting qualities. − In animal models of stroke, multiple sclerosis, and other neuroinflammatory diseases, Tanshinone IIA (TSA) has been demonstrated to provide neuroprotection, encourage M2 microglial polarization, and prevent NLRP3 inflammasome activation in macrophages. − However, it remains uncertain whether exosomes derived from TSA-preconditioned UCMSCs confer enhanced therapeutic benefits for SCI, and if so, whether these benefits are mediated by the genetic material encapsulated within the exosomes.

Activated microglia, the spinal cord’s resident immune cells, assemble at the site of injury following SCI. Here, they remove cellular debris and release neurotrophic factors to aid tissue repair. Microglia are highly responsive to changes in the spinal microenvironment and are critical in regulating neuroinflammatory responses. , The two primary phenotypic states that these cells can polarize into are the M1 (pro-inflammatory) and M2 (anti-inflammatory) states. − M1-polarized microglia secrete proinflammatory cytokines that include interleukin-1β (IL-1β), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α) after SCI, which exacerbate inflammation. , Additionally, increased permeability of the BSCB enhances the release of chemokines by M1 microglia, facilitating the recruitment of peripheral leukocytes and amplifying inflammatory pathways, which eventually leads to excitotoxic neuronal death. In contrast, M2-polarized microglia generate anti-inflammatory cytokines for tissue repair, neuroprotection, and inflammation suppression, for example transforming growth factor-beta (TGF-β), interleukin-4 (IL-4), and interleukin-10 (IL-10). An important factor in inflammation of the central nervous system (CNS) is the NLRP3 inflammasome. When it is activated, it initiates caspase-1 cleavage, which subsequently activates gasdermin D (GSDMD). Pro-inflammatory cytokines can be released through transmembrane holes created by activated GSDMD, which exacerbates neuroinflammation. Furthermore, NLRP3 inflammasomes are essential regulators of microglial polarization. By promoting the phenotypic transition of microglia from the M1 to the M2 state, inhibiting NLRP3 inflammasome activation reduces neuroinflammation and enhances outcomes after SCI.

In this investigation, we found that, compared to exosomes (Exos) derived from untreated UCMSCs, exosomes from TSA-pretreated UCMSCs (T-Exos) exhibited no significant alterations in physicochemical characteristics. Furthermore, T-Exos markedly suppressed microglial NLRP3 inflammasome activation, promoted M2 polarization of microglia, and enhanced neuromotor recovery post-spinal cord injury through miR-223–5p delivery. Mechanistically, miR-223–5p within T-Exos inhibited USP8 expression by targeting its 3′ untranslated region (3′UTR), thereby suppressing USP8-stabilized NLRP3 inflammasome activation and modulating microglial pyroptosis following spinal cord injury. Our research offers a potential and successful therapeutic method to treat SCI by elucidating a novel mechanism underpinning TSA-pretreated UCMSC-derived exosomes.

Methods and Materials

Experimental Animals

In this study, we chose healthy male C57BL/6 mice, 8 to 12 weeks old, as experimental models to ensure the consistency and reliability of our results. The mice were obtained from the Animal Research Center at Nanjing Medical University and kept under standardized conditions. We strictly followed established breeding protocols to provide a stable environment. The animals were exposed to a 12 h light cycle followed by 12 h of darkness daily to mimic natural circadian rhythms. The facility’s temperature was maintained at 22 ± 2 °C, and the relative humidity was kept at 45% ± 5% to promote the well-being of the mice and minimize environmental variables that could affect experimental outcomes. All procedures were approved by the Animal Ethical and Welfare Committee of KdC of Nanjing Medical University (Approval No. KD03220001) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Cell Culture

The BV2 murine microglial cell line was acquired from Shanghai Cell Research Center in China. Cells were cultured in a high-glucose DMEM medium, which included 4.5 g/L glucose, along with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin to supply essential nutrients and antimicrobial protection. The cells were maintained at 37 °C in a humidified atmosphere with 5% CO₂ to provide optimal conditions for growth. Upon reaching roughly 80% confluence, the cells were digested with trypsin to facilitate detachment and collection, after which they were passaged for subsequent experimental procedures.

The human-umbilical-cord-derived mesenchymal stem cells (UCMSCs) used in this study were obtained from Cyagen Biosciences, located in Guangzhou, China. These stem cells were cultured in a stable 37 °C environment with 5% CO₂, using a specific mesenchymal stem cell medium (MSCM) enriched with 10% Exos-depleted FBS to prevent Exos-related interference and retain stem cell properties. Prior to medium collection, UCMSCs from passages 3 to 5 underwent a 24 h pretreatment with 10 μmol/L TSA to promote uniformity and stability for subsequent analyses.

Plasmids and Cell Transfection

Plasmids encoding the miR-223–5p mimic, miR-223–5p inhibitor, and their negative control, Flag-tagged USP8, Flag-tagged USP8 C786A, and Myc-tagged NLRP3 were constructed by Nanjing KeyGEN Biotech Corp., Ltd. Plasmids encoding HA-tagged Ubiquitin-Lys63 and HA-tagged Ubiquitin-Lys48 were provided by Genebay Biotech. According to the manufacturer’s instructions, transfections were performed using Lipofectamine 3000 reagent (Invitrogen) with 500 ng, 1 μg, or 2 μg of plasmid.

Exos Isolation and Characterization

Exosomes were isolated following established protocols: cell supernatant was collected and subjected to sequential centrifugation at 500g, 2000g, 10 000g, and finally 100 000g to obtain the Exos. Exos morphology was visualized by TEM, while DLS was used to measure particle size, concentration, and zeta potential. The presence of the Exosomal markers Alix, CD9, CD63, and CD81 was confirmed by Western blot analysis.

SCI Model

In this study, a contusive spinal cord injury was produced at the T9 thoracic level in mice using the Louisville Injury System Apparatus (LISA) impactor. In brief, mice received pentobarbital anesthesia (40 mg/kg, intraperitoneally) and were placed on a sterile surgical platform. The spinal cord was exposed by a T9 laminectomy, and a 0.6 mm impact was applied to the exposed cord by using the LISA impactor. After injury, the muscle and skin layers were closed with sutures and then maintained in temperature- and humidity-controlled housing until they were fully recovered from anesthesia. Post-operatively, mice were given daily intraperitoneal injections of gentamicin for 1 week and had manual bladder expression performed twice daily until the bladder reflex returned.

Behavioral Assessments

Gait Photography Analysis

Kinematic gait video recordings of mice were captured at specified post-injury intervals to evaluate motor performance. These recordings enable a thorough analysis of gait parameters, such as stride length, paw placement, and coordination.

Footprint Analysis

To assess stride length and width, the forepaws and hindpaws of mice were coated with blue and red dyes, respectively. Footprint patterns were documented as the mice ran at a steady pace.

Swimming Test

Swimming trials were conducted in a controlled tank to assess the swimming ability. Mice were trained to traverse the tank, and their performance was evaluated by using the Louisville Swim Score. This assessment focused on parameters including forelimb dependency, hindlimb movement and alternation, trunk stability, and overall body posture.

Western Blot Analysis

Protein extraction was performed using RIPA lysis buffer, and the protein concentration was determined using a BCA assay kit from Thermo Fisher Scientific. The proteins were then separated on a 10% or 12.5% SDS-PAGE gel and transferred to a PVDF membrane. The membrane was blocked with 5% bovine serum albumin (BSA) at room temperature for 1 h. Subsequently, the sample was subjected to overnight incubation at 4 °C with the respective primary antibody. After three washes with 1% TBST, the membrane with the appropriate secondary antibody was incubated for 1 h. Finally, Thermo Fisher Scientific enhanced chemiluminescence reagents were used to visualize the protein bands. NLRP3 (abcam, ab270449, 1:1000), GSDMD/GSDMD-N (abcam, ab209845, 1:1000), Il-1β (abcam, ab234437, 1:1000), iNOS (abcam, ab3523, 1:1000), Arg1 (CST, 93668, 1:1000), USP8 (proteintech, 67321–1-Ig, 1:5000), Caspase-1/p20 (AdipoGen, AG-20B-0042-C100, 1:1000), and GAPDH (abcam, ab9485, 1:2500).

Immunoprecipitation

Following the manufacturer’s guidelines, the Pierce Co-IP Kit (Thermo Fisher Scientific) was used for immunoprecipitation (Co-IP) in this study. To begin, specific antibodies were attached to agarose beads to promote efficient interactions with the intended target protein. Next, cell lysates derived from either microglial cells or HEK 293T cells were introduced to the pre-prepared antibody-agarose bead complexes. This mixture was then incubated overnight at 4 °C under gentle rotation to promote the formation of antigen–antibody complexes. After the incubation period, the beads underwent multiple washes with lysis buffer to remove any nonspecifically bound proteins, thereby improving the binding specificity and minimizing background interference. Following the washing steps, an elution buffer was applied to release the target protein from the agarose beads into the solution. To complete the protein detection and characterization process, the collected eluate was subsequently subjected to Western blot analysis to verify the presence and expression level of the target protein.

Ubiquitylation Assays

To comprehensively evaluate the ubiquitination level of NLRP3, an anti-NLRP3 antibody was initially used in an immunoprecipitation assay to selectively enrich the NLRP3 protein. Following immunoprecipitation, Western blot analysis was carried out using an antiubiquitin (Ub) antibody on the immunoprecipitated sample, allowing both detection of NLRP3 ubiquitination and assessment of the types and extent of ubiquitin chain linkages. To further examine the effect of USP8 overexpression on NLRP3 ubiquitination, a cotransfection experiment was performed in HEK 293T cells using multiple expression plasmids. These constructs included Myc-tagged NLRP3 (Myc-NLRP3), Flag-tagged USP8 (both wild-type and C786A mutant Flag-USP8), and HA-tagged ubiquitin (HA-Ub). After transfection, total protein lysates were prepared from HEK 293T cells, followed by immunoblotting with specific antibodies.

Isolation of RISC-Associated RNA

BV2 microglial cells were transfected to overexpress either miR-223–5p or a negative control (miR-NC). Initially, the treated cells were fixed with 1% formaldehyde, which facilitated cross-linking between proteins and nucleic acids. This was followed by chromatin shearing to generate fragments of appropriate size for immunoprecipitation. Cell lysis was conducted using NETN buffer to guarantee a thorough release of cellular contents. For the immunoprecipitation process, Dynabeads Protein A (Invitrogen) served as the magnetic bead platform and specific protein-RNA complexes were captured using either IgG control antibodies or pan-AGO antibodies. After the immunoprecipitation complexes were established, proteinase K was applied to digest the complexes, thus, releasing the RNA from the bound complexes. The extracted RNA underwent purification via a phenol-chloroform-isopropanol method to eliminate proteins and other contaminants, thereby improving sample clarity. Subsequently, the RNA was concentrated by glycogen-assisted ethanol precipitation, and any residual DNA was removed by DNase I treatment, ensuring a high RNA purity for downstream analyses.

RT-qPCR

Total RNA was extracted from both cells and exosomes using TRIzol reagent (Invitrogen), selected for its high efficiency in comprehensive RNA extraction. The extracted RNA was then processed for cDNA synthesis. For miRNA cDNA synthesis, the Hairpin-it miRNA qPCR Quantitation Kit (GenePharma, China) was used, designed to ensure precise reverse transcription of miRNA for expression analysis. For mRNA cDNA synthesis, the PrimeScript RT Reagent Kit (Takara, Japan) was applied to produce high-quality mRNA cDNA, providing a reliable foundation for quantitative measurements. Subsequently, qRT-PCR was carried out using the TB Green Premix Ex Taq kit (Takara, Japan) to evaluate the expression levels of mRNA and miRNA. To ensure data accuracy, relative expression levels were calculated using the 2^–ΔΔCT method with GAPDH and U6 as internal control genes for mRNA and miRNA normalization. The mRNA and miRNA primer sequences were as follows: TNF-α: Forward CCCTCAC­ACTCAGAT­CATCTTCT, Reverse GCTACGA­CGTGGG­CTACAG; IL-1β: Forward GCAACTG­TTCCTGAA­CTCAACT, Reverse ATCTTTT­GGGGTCC­GTCAACT; iNOS: Forward GTTCTCAG­CCCAACAA­TACAAGA, Reverse GTGGAC­GGGTCGA­TGTCAC; Arg1: Forward CTCCAAG­CCAAAGTCC­TTAGAG, Reverse AGGAGCT­GTCATTA­GGGACATC; CD206: Forward CTCTGTTC­AGCTATT­GGACGC, Reverse CGGAATTT­CTGGGATT­CAGCTTC; YM1/2: Forward CAGGGT­AATGAGT­GGGTTGG, Reverse CACGGCA­CCTCCT­AAATTGT; GAPDH: Forward AGGTCGG­TGTGAAC­GGATTTG, Reverse TGTAGACC­ATGTAGTT­GAGGTCA; miR-223–5p: GCGCGC­GTGTATTT­GACAAG­CTGAGTT; miR-223–3p: CGGGCTGT­CAGTTTGT­CAAAT; miR-21–5p: TAGCTTAT­CAGACTGA­TGTTGA; miR-16–5p: CCGTAGC­AGCACGT­AAATATT­GGCG; miR-150–5p: AGTGCTG­TCTCCC­AACCCTT; U6: CCGAGAGAA­GATTAGCAT­GGCCCCTG.

Luciferase Reporter Assay

To detect potential binding sites between miR-223–5p and the 3′ UTR of USP8 mRNA, TargetScan (https://www.targetscan.org/vert_80/) was utilized to predict interaction sites. This analysis pinpointed a specific binding region for miR-223–5p within USP8. To evaluate the functionality of this predicted site, both the wild-type sequence and a mutated version of the binding region were inserted into the luciferase reporter vectors, allowing for the assessment of regulatory effects at this site. These luciferase reporter constructs were then cotransfected with either a miR-223–5p mimic or a negative control (miR-NC) into HEK 293T cells to investigate miR-223–5p’s regulatory impact on the USP8 3′ UTR. Following transfection, cells were incubated for 48 h to facilitate adequate interaction between miR-223–5p and its target mRNA. Upon completion of this incubation period, cell lysates were obtained, and luciferase activity was determined by a dual-luciferase reporter assay, offering insights into miR-223–5p’s regulatory role in USP8 expression.

Immunofluorescence Staining

Following cardiac perfusion, spinal cord tissue was harvested and fixed in 4% paraformaldehyde. Tissue sections were prepared, blocked, and incubated overnight at 4 °C with primary antibodies against Neun, NF200, NLRP3, Caspase-1, IBA-1, iNOS, and Arg1. The following day, sections were incubated with the appropriate secondary antibodies at room temperature for 1 h. Nuclei were counterstained with DAPI. For cell samples, cells were fixed with 4% paraformaldehyde and rendered permeable with 0.05% Triton X-100. After blocking, primary antibodies for NLRP3, Caspase-1, and IBA-1 were applied overnight at 4 °C. On the subsequent day, secondary antibodies and DAPI were applied. Fluorescence images were captured by using a fluorescence microscope. Neun (abcam, ab177487, 1:200), NF200 (abcam, ab8135, 1:5000), iNOS (abcam, ab3523, 1:50), Arg1 (CST, 93668, 1:200), NLRP3 (abcam, ab270449, 1:50), and IBA-1 (abcam, ab178846, 1:1000).

Statistical Analysis

Comparisons among multiple groups were assessed using one-way or two-way analysis of variance (ANOVA), whereas comparisons between two groups were performed using the two-tailed t-test. All data were presented as the mean ± standard deviation (SD). Consequently, statistical evaluations were conducted via GraphPad Prism 8.0.2 software (GraphPad Software, USA). The statistical significance of these effects was determined by setting the significance level at (*) P < 0.05, (**) P < 0.01, and (***) P < 0.001, respectively.

Results and Discussion

Characterization and Comparison of Exos and T-Exos

To take advantage of TSA’s biological characteristics, we employed a pretreatment approach that produced T-Exos and improved the intrinsic functionality of individual exosomes. Umbilical cord mesenchymal stem cells (UCMSCs) were identified by flow cytometry; they were CD45 and CD34 negative but CD90, CD105, and CD73 positive (Figure A). Exos and T-Exos were found to have similar spherical or cup-shaped structures by transmission electron microscopy (TEM) (Figure B). Further validation with dynamic light scattering (DLS) indicated comparable size distribution profiles for Exos and T-Exos nanoparticles (Figure C), with no significant differences in mean particle size, nanoparticle concentration per milliliter, or zeta potential between the groups (Figure D). The absence of the non-Exosomal biomarker Calnexin confirmed sample purity, while Western blot analysis identified key Exosomal markersAlix, CD9, CD63, and TSG101in both Exos and T-Exos (Figures E and F). The evaluation of cellular uptake was conducted by co-culturing Dil-labeled exosome samples with immortalized mouse microglial cells (BV2) for a duration of 24 h. Confocal microscopy images verified that the target cells had effectively internalized the exosomes (Figures G and H). The in vivo imaging results indicate no significant difference in the distribution of Exos and T-Exos at the spinal cord injury site (Figure S1A–S1D). As miRNAs serve as primary messenger molecules carried by exosomes, we hypothesized that the miRNA content of T-Exos may differ from that of Exos. Bioinformatics analysis revealed that 12 miRNAs were upregulated in T-Exos relative to Exos (Figure S2A). We subsequently gathered the target genes for these 12 up-regulated miRNAs in T-Exos from the TargetScan and miRDB databases. According to KEGG enrichment analysis, the target genes were involved in the regulation of axon guidance and regeneration (Figure S2B). In addition, GO analysis revealed that these target genes were related to a number of immune-related biological processes, including lymphocyte migration, cytokine-mediated signaling pathways, response to lipopolysaccharide, and regulation of cytokine synthesis (Figure S2C).

1.

Characterization and comparison of Exos and T-Exos. (A) Flow cytometric characterization of UCMSCs; (B) morphological assessment of Exos and T-Exos via TEM; (C) Exos and T-Exos particle distribution by DLS analysis (scale bar = 200 nm); (D) mean particle size, particle concentration per mL, and surface zeta potential of exosomes (n = 6); (E, F) Western blot analysis of Exosomal biomarkers Alix, CD9, CD63, and CD81, alongside the negative control calnexin (n = 3); (G, H) uptake of red fluorescent dye Dil-labeled exosomes by BV2 microglia (n = 3).

Administration of T-Exos Promotes Better Neurological Functional Recovery after SCI in Mice

In this study, miRNA sequencing and KEGG enrichment analysis indicated that T-Exos play a further role in regulating axon guidance and regeneration (Figure S2B). To investigate whether T-Exos facilitate axonal growth guidance and enhance neurological functional recovery, we conducted in vivo experiments. T-Exos were administered to mice, followed by a series of assessments conducted over 28 days post-spinal cord injury to test this hypothesis (Figure A). As measured by Basso Mouse Scale (BMS) scores, mice treated with T-Exos showed noticeably higher increases in walking capacity, hind paw location, and hind limb mobility over the 4-week recovery period following SCI (Figure B). Footprint analysis further confirmed that T-Exos treatment accelerated gait recovery and improved motor coordination more effectively than Exos treatment (Figures C and D). Gait photography provides additional corroboration for the aforementioned results (Figure E). Additionally, the rotarod test showed that the T-Exos group recovered better than the Exos group in terms of the hind limb and tail balance (Figure F). Similarly, mice treated with T-Exos showed faster hind limb swimming movements, maintained a lower body-to-water angle, and depended less on their forelimbs, according to data from the Louisville Swim Scale (Figures G and H). Electrophysiological analysis revealed that T-Exos-treated mice exhibited higher motor-evoked potential (MEP) amplitudes with shorter latencies than those treated with Exos (Figures I and J). Furthermore, the lesion area in the T-Exos group significantly decreased compared with the PBS and Exos groups, according to a macroscopic analysis of the thoracic region (Figure K). Collectively, these behavioral assessments demonstrate that the administration of T-Exos significantly enhances neurological functional recovery following SCI, offering superior therapeutic outcomes compared to Exos.

2.

T-Exos enhances the recovery of function in mice with SCI. (A) In vivo experimental setup diagram; (B) BMS scores in mice treated with PBS, Exos, and T-Exos during the 28-day postinjury recovery period (n = 6); (C, D) gait analyses via footprint tests for the three experimental groups (n = 6); (E) locomotor photography assessment for the three groups (n = 6); (F) statistical evaluation of rotarod test performance for the three groups (n = 6); (G, H) quantification of LSS swimming scores reflecting the functional status of mice 28 days post-injury (n = 6); (I , J) electrophysiological analysis using MEP at day 28 post-injury (n = 6); (K) morphometric analysis of spinal cord gross anatomy and injury site; (L, M) IF staining and quantitative analysis of NF200 expression in spinal cord tissue across the three groups (n = 6, scale bar = 500 μm); (N, O) NeuN IF staining and corresponding statistical analysis (n = 6, scale bar = 200 μm).

To investigate the structural basis for the observed functional improvements, we examined the axonal density and condition within the core of the spinal cord lesions. Neurofilaments, which provide structural integrity to neurons and facilitate axonal transport, were marked by neurofilament-positive (NF⁺) axons, serving as indicators of axonal regeneration. The T-Exos group had significantly more NF⁺ axons in the lesion center than the Exos group had at 28 days after the damage (Figure L-M). NeuN labeling was used to detect surviving neurons at particular sites (Z1–Z4) at different distances around the lesion margin to validate these findings. According to analysis, there were noticeably more NeuN+ neurons in areas Z1 and Z2 in the T-Exos group than in the Exos group (Figure N–O). In conclusion, these results highlight the therapeutic benefit of T-Exos by showing that they perform better than Exos in stimulating axonal regeneration, as well as improving the recovery of function following SCI.

Administration of T-Exos Alleviates Secondary Inflammatory Responses Following SCI

GO analysis demonstrated that T-Exos are predominantly involved in the regulation of the immune microenvironment, particularly through processes such as cytokine production regulation, response to lipopolysaccharides, and cytokine-mediated signaling pathways (Figure K). The inflammatory response linked to secondary damage SCI is triggered in large part by microglia. Exosomes have been shown to modulate NLRP3 inflammasome activation in microglia in previous studies. − Building on this, we investigated whether T-Exos has improved regulatory effects on NLRP3 inflammasome activation in SCI. Lipopolysaccharide (LPS) and ATP were introduced to cell cultures prior to the administration of PBS, Exos, or T-Exos to mimic NLRP3 activation in vitro. Initially, immunofluorescence (IF) was employed to evaluate the activation of the NLRP3 inflammasome in BV2 microglial cells. The findings suggested that Exos treatment suppressed NLRP3 activation, and this inhibitory effect was further amplified by T-Exos treatment (Figures A and B). We further validated these observations by Western blot (WB) analysis of proteins associated with NLRP3 inflammasome activation (including NLRP3, GSDMD, GSDMD-N, Caspase-1, p20, and muture IL-1β) (Figures C and D). To determine if the therapeutic effects of T-Exos reported in vitro were also evident in vivo, evaluation of NLRP3 activation levels in microglia following SCI was conducted through IF. On day 7 post-injury, microglia from mice treated with T-Exos expressed considerably lower levels of NLRP3 and Caspase-1 than those treated with Exos (Figure E and F). We further examined the expression of NLRP3 activation markers in the injured spinal cord through WB. Exos and T-Exos treatments both reduced NLRP3 inflammasome activation in comparison with the control group (PBS-treated), and T-Exos had a greater inhibitory impact (Figure G and H). These findings indicate that T-Exos more effectively regulate NLRP3 inflammasome activation in microglia following SCI, aligning with the results obtained in vitro.

3.

T-Exos regulates NLRP3 inflammasome activation and the polarization of microglia in mice with SCI. (A, B) IF staining and quantification of NLRP3 and Caspase-1 in BV2 microglia exposed to PBS, Exos, and T-Exos (n = 3, scale bar = 20 μm); (C, D) Identification of proteins associated with NLRP3 inflammasome activation-related genes in BV2 microglia across the three groups using Western blot and quantitative analysis (n = 3); (E, F) Representative immunofluorescence images depicting IBA-1 (green) and NLRP3/Caspase-1 (red) expression at day 7 post-injury, along with quantitative analyses (n = 3, scale bar = 10 μm); (G, H) identification of NLRP3 inflammasome activation-related proteins in injured spinal cord tissue via Western blot and quantitative analyses (n = 3); (I, J) DHE staining of injured spinal cords from PBS, Exos, and T-Exos animals on day 7 post-injury to assess ROS accumulation (n = 3, scale bar = 10 μm); (K) statistical evaluation of pro-inflammatory and anti-inflammatory cytokine levels in the three groups’ mice’s injured spinal cords on day seven after the injury via ELISA (n = 3); (L) quantitative analysis of M1- and M2-related marker gene expression via qRT-PCR (n = 3); (M, N) Western blot and quantitative analysis to identify proteins associated with M1- and M2-related genes (n = 3); (O, P) representative immunofluorescent images of IBA-1 (green) and iNOS (red) expression at day 7 after injury with quantitative analysis (n = 3, scale bar = 10 μm); (Q, R) representative immunofluorescent images of IBA-1 (green) and iNOS (red) expression at day 7 after injury with quantitative analysis (n = 3, scale bar = 10 μm).

The relationship between the microglia polarization and NLRP3 inflammasome activation following SCI is well-established. , Therefore, we investigated how T-Exos affected microglial polarization and inflammatory cytokine levels in the injured spinal cord. To assess oxidative stress in the damaged area, DHE staining was used, revealing that Exos treatment reduced reactive oxygen species (ROS) accumulation, an effect that was even more pronounced with T-Exos (Figures I and J). ELISA results indicated that both Exos and T-Exos treatments, relative to the PBS control group, reduced pro-inflammatory factors (TNF-α, IL-1β, and IL-6) and increased anti-inflammatory factors (TGF-β, IL-4, and IL-10), with T-Exos showing greater effects (Figure K). The results of qRT-PCR analysis indicated a significant upregulation of M2-related genes (Arg1, CD206, and YM1/2) as well as a downregulation of M1-related genes (iNOS, TNF-α, and IL-1β) among both Exos and T-Exos groups, compared to that of the PBS group. T-Exos induced a more significant shift toward the M2 phenotype (Figure L). Western blot analysis further supported these findings (Figures M and N). Immunofluorescence labeling for iNOS (M1 marker), Arg1 (M2 marker), and IBA-1 (microglial marker) was also performed to assess microglial polarization post-SCI. Compared with animals treated with Exos, those given T-Exos had a large increase in Arg1-positive microglia and a significant decrease in iNOS-positive microglia in the damaged spinal cord on day 7 (Figures O–R).

Exosomes derived from preconditioned cells have demonstrated enhanced therapeutic effects and transplantation efficacy. , For instance, exosomes produced by nerve cells pretreated with ginsenoside Rg1 promote neural functional recovery by modulating the immune microenvironment. Studies have revealed that exosomes from hypoxia-preconditioned MSCs improve SCI healing by increasing the polarization of M2 microglia. Another study indicated that exosomes generated by melatonin-preconditioned MSCs improved motor performance in animals with SCI, enhanced the spinal cord microenvironment, and encouraged axonal regeneration. Tanshinone IIA is known to have a range of pharmacological properties, including antioxidant, anti-inflammatory, and antiapoptotic effects, which exert their functional effects through multiple mechanisms. − For example, TSA has been shown to promote autophagy through the AMPK-mTOR pathway, thereby inhibiting neuronal damage induced by intermittent hypoxia. Research has also shown that TSA treats Alzheimer’s disease by specifically degrading endogenous Tau protein through the ubiquitin-proteasome pathway. Shan et al. reported that TSA alleviates pulmonary fibrosis by modulating glutamine metabolism. Zeng et al. demonstrated that TSA sodium sulfonate improves neuropathic pain by regulating microglial polarization and neuroinflammation. Wang et al. demonstrated that TSA prevents lipopolysaccharide-induced brain damage by maintaining the blood-brain barrier and decreasing oxidative stress and inflammatory responses. However, few studies have explored the use of exosomes derived from TSA-preconditioned MSCs, highlighting the need for further research, particularly in the context of central nervous system injuries. Our findings, which are consistent with previous investigations, show that administering exosomes can improve cerebral motor function recovery by suppressing microglial inflammasome activation and increasing microglial polarization toward the M2 phenotype. Notably, we found that T-Exos exhibit more potent therapeutic effects than conventional Exos, prompting us to investigate the potential mechanisms underlying the differences between these two types of exosomes.

miR-223–5p Is Expressed at High Levels in T-Exos and May Regulate the Immune Microenvironment after SCI

The findings indicate that T-Exos is more effective than the Exos treatment group in modulating the immune microenvironment after SCI, thereby promoting functional recovery to a greater extent. As miRNA is the primary messenger molecule carried by exosomes, we hypothesize that the miRNAs with elevated expression in T-Exos may serve as the main functional components responsible for their enhanced activity. Bioinformatics analysis revealed that 12 miRNAs were upregulated in T-Exos relative to Exos, with miR-223–5p identified as the most significantly upregulated miRNA (Figure S2D). To validate the bioinformatics analysis results, the top five up-regulated miRNAs (miR-223–5p, miR-223–3p, miR-21–5p, miR-16–5p, and miR-150–5p) were selected, and their in vitro expression levels were examined by qRT-PCR. Of which, miR-223–5p was found to be the highest upregulation in T-Exos, compared to Exos (Figure S2E). Based on these bioinformatics and qRT-PCR results, we focused on miR-223–5p, which was found to have the highest expression in T-Exos. To clarify the role of Exosomal miR-223–5p in influencing NLRP3 activation in microglia following SCI, we created miR-223–5p overexpression (miR-223–5p^OE) and knockdown (miR-223–5p^KD) UCMSCs using lentiviral vectors, as well as corresponding negative controls (miR-NC^OE and miR-NC^KD). The efficacy of the transfection was validated by qRT-PCR (Figure A). Exosomes were then isolated from these four UCMSC groups and labeled as miR-NC^KD-T-Exos, miR-223–5p^KD-T-Exos, miR-NC^OE-T-Exos, and miR-223–5p^OE-T-Exos, respectively. Compared with miR-NC^KD-T-Exos, miR-223–5p expression was significantly reduced in miR-223–5p^KD-T-Exos and markedly increased in miR-223–5p^OE-T-Exos relative to miR-NC^OE-T-Exos (Figure B). These exosomes were then co-cultured with microglia. In comparison to the miR-NC^KD-T-Exos group, microglia treated with miR-223–5p^KD-T-Exos displayed a significantly lower miR-223–5p level, while microglia treated with miR-223–5p^OE-T-Exos showed an elevated miR-223–5p level relative to that of the miR-NC^OE-T-Exos group (Figure C). These findings suggest that the upregulation and transfer of miR-223–5p to microglia via T-Exos is a crucial mechanism underlying the therapeutic benefits of T-Exos after spinal cord injury.

4.

T-Exos modulates the immune microenvironment and NLRP3 inflammasome activation post-SCI through miR-223–5p delivery. (A) miR-223–5p overexpression and knockdown in UCMSCs with efficiency validated by qRT-PCR (n = 3); (B) relative miR-223–5p expression levels in miR-NC^OE-T-Exos, miR-223–5p^OE-T-Exos, miR-NC^KD-T-Exos, and miR-223–5p^KD-T-Exos (n = 3); (C) relative miR-223–5p expression in target primary microglia and BV2 microglia from the four groups (n = 3); (D) ROS accumulation in injured spinal cords of mice administered with miR-NC^OE-T-Exos, miR-223–5p^OE-T-Exos, miR-NC^KD-T-Exos, and miR-223–5p^KD-T-Exos detected by DHE staining on day 7 post-injury (n = 3, scale bar = 10 μm); (E) ELISA measurement of pro- and anti-inflammatory cytokine levels in injured spinal cords from the four mouse groups with statistical analysis (n = 3); (F) mRNA expression of M1- and M2-related genes assessed by qRT-PCR and statistical evaluation (n = 3); (G–I) protein expression of M1- and M2-related genes analyzed by Western blot and quantification (n = 3); (J, rK) Representative immunofluorescence images of IBA-1 (green) and iNOS (red) expression at day 7 post-injury with quantification (n = 3, scale bar = 10 μm); (L, M) Representative immunofluorescence images of IBA-1 (green) and Arg1 (red) expression at day 7 post-injury with quantification (n = 3, scale bar = 10 μm); (N) Western blot and quantification of NLRP3 inflammasome activation-related proteins in injured spinal cords (n = 3); (O–R) representative immunofluorescence images of IBA-1 (green) and NLRP3/Caspase-1 (red) expression at day 7 post-injury with quantification (n = 3, scale bar = 10 μm).

Then, we investigated miR-223–5p’s impact on microglial polarization in vivo to better understand its function in modulating the impacts on the immune microenvironment following SCI. The results revealed that miR-223–5p^OE-T-Exos were significantly more effective than miR-NC^OE-T-Exos in inhibiting ROS accumulation in the SCI region, while the effectiveness of miR-223–5p^KD-T-Exos was notably reduced (Figure D). Importantly, ELISA, qRT-PCR, and Western blot studies indicated that miR-223–5p^OE-T-Exos significantly increased M2 microglial polarization, but miR-223–5p^KD-T-Exos decreased M2 microglial polarization (Figure E-I). Furthermore, immunofluorescence analysis provided additional corroboration of these findings (Figure J–M). Further examination of miR-223–5p’s role in regulating NLRP3 activation was carried out via Western blot. The results demonstrated that miR-223–5p overexpression enhanced the ability of T-Exos to suppress NLRP3 inflammasome activation after spinal cord injury, whereas miR-223–5p knockdown attenuated this suppressive effect (Figure N). Immunofluorescence staining revealed that in the miR-223–5p^OE-T-Exos group, IBA-1⁺ microglia exhibited lower fluorescence intensity for NLRP3 and Caspase-1, in comparison to the miR-NC^OE-T-Exos group. Conversely, in the miR-223–5p^KD-T-Exos group, IBA-1⁺ microglia showed increased fluorescence for NLRP3 and Caspase-1, in comparison to the miR-NC^KD-T-Exos group (Figures O–R). Overall, these data imply that miR-223–5p is required for T-Exos to modulate the immune microenvironment and activate the NLRP3 inflammasome following SCI.

T-Exos Enhances Functional and Behavioral Recovery after SCI via Transferring miR-223–5p

In addition, the role of miR-223–5p in promoting the therapeutic effects of T-Exos on the functional recovery after SCI was investigated. Exosomes from each group-miR-NC^KD-T-Exos, miR-223–5p^KD-T-Exos, miR-NC^OE-T-Exos, and miR-223–5p^OE-T-Exos were administered to SCI mice. Behavioral assessments showed that suppressing miR-223–5p significantly reduced T-Exos’ efficacy in improving motor function recovery. In contrast, miR-223–5p overexpression improved the recovery effects of T-Exos, as indicated by improvements in BMS scores, Gait photography, rotarod tests, and electrophysiological tests (Figure A–E). Additionally, compared to mice treated with miR-NC^OE-T-Exos, those receiving miR-223–5p^OE-T-Exos showed a higher count of NeuN⁺ neurons and NF200⁺ axons. In contrast, mice treated with miR-223–5p^KD-T-Exos displayed fewer NeuN⁺ neurons and NF200⁺ axons, compared to the miR-NC^KD-T-Exos group (Figures F–I).

5.

T-Exos enhance functional and behavioral recovery and promote axonal regeneration post-SCI through miR-223–5p delivery. (A) Locomotor function assessment using BMS scores in mice administered with miR-NC^OE-T-Exos, miR-223–5p^OE-T-Exos, miR-NC^KD-T-Exos, and miR-223–5p^KD-T-Exos during the 28-day recovery period post-injury (n = 6); (B) gait analysis via photography for the four mouse groups (n = 6); (C) rotarod performance evaluation for the four mouse groups (n = 6); (D, E) MEP evaluation of the four mouse groups at day 28 post-injury (n = 6); (F–I) IF staining and statistical analysis of NeuN⁺ neurons and NF200⁺ axons in spinal cord tissue from the four groups of mice 28 days after SCI (n = 6, scale bar = 200 μm).

Exosomes have been shown to carry certain miRNAs and deliver them to target cells. For instance, after SCI, Treg cells distribute miR-709 and miR-2861 to target cells via exosomes to control neuroinflammation and repair the blood-spinal cord barrier. , Moreover, Liu et al. revealed that exosomes produced from iPSC-NSCs transport let-7b-5p, which improves motor function following spinal cord damage by targeting LRIG3. As a result, we investigated the function of the miRNA cargo in the positive effects of T-Exos. The miRNA sequencing investigation of exosomes indicated an elevation of miR-223–5p in T-Exos, which regulates the immunological milieu after spinal cord damage. We also discovered that the absence of miR-223–5p in exosomes substantially mitigated the positive effects of the T-Exos. However, since multiple miRNAs are upregulated in T-Exos and this study focused solely on miR-223–5p, we cannot exclude the potential roles of other functional miRNAs in regulating NLRP3 inflammasome activation in microglia after spinal cord injury. Interestingly, the GO analysis of miRNA target genes in T-Exos indicated that their functions are associated with ubiquitin-like protease binding. We validated that miR-223–5p regulates microglial NLRP3 inflammasome activation by targeting the deubiquitinating enzyme USP8 through luciferase reporter assays, RNA-ChIP experiments, and rescue experiments. Nevertheless, miR-223–5p may have multiple target genes, and we do not exclude the possibility that it regulates NLRP3 inflammasome activation by targeting other genes. Additionally, aside from miR-223–5p, exosomes include a variety of other substances, such as proteins, RNA, and lipids. As a result, future research should find further pathways by which miR-223–5p exerts its therapeutic effects as well as investigate other potential novel therapeutic factors supplied by exosomes for the treatment of SCI and other disorders.

miR-223–5p Regulates Microglia NLRP3 Inflammasome Activation by Targeting USP8

In order to conduct a more comprehensive investigation into the mechanism by which miR-223–5p regulates the activation of the NLRP3 inflammasome in microglia, this study identified and investigated the target genes of miR-223–5p. Recent studies have emphasized the pivotal role of ubiquitin-related enzymes in the activation of the NLRP3 inflammasome. − Our GO analysis also suggested that the miRNA is associated with ubiquitin-like protease binding (Figure K). Using the TargetScan mRNA prediction database, we analyzed potential targets of miR-223–5p and identified USP8 as a candidate, previously linked to both inflammation and ubiquitination processes. We hypothesized that miR-223–5p specifically binds to the 3′ untranslated region (3′-UTR) of USP8. To test this, we constructed both wild-type (WT) and mutant (MUT) 3′-UTR sequences of USP8 based on predicted binding sites and co-transfected them with miR-223–5p into 293T cells (Figure A). Results from luciferase reporter assays showed a significant decrease in luciferase activity when WT-USP8–3′UTR was co-transfected with miR-223–5p^OE, compared to its co-transfection with miR-NC^OE. However, no substantial change was detected with the MUT-USP8–3′UTR, demonstrating that miR-223–5p selectively interacts with the USP8 3′ UTR (Figure B). USP8 mRNA incorporation into the Ago2/RNA-induced silencing complex (RISC) was quantitatively measured by RNA-ChIP analysis subsequent to miR-223–5p overexpression. USP8 into RISC was higher in cells overexpressing miR-223–5p (Figure C). Additionally, USP8 mRNA and protein levels decreased when miR-223–5p was overexpressed but increased when miR-223–5p was knocked down, confirming the function of miR-223–5p as a downstream target (Figures D–F).

6.

miR-223–5p inhibits NLRP3 inflammasome activation by modulating USP8 expression. (A) miR-223–5p directly targets the 3′-UTR of USP8 to regulate its expression; (B) Luciferase reporter assay confirmed that USP8 is a target gene of miR-223–5p (n = 3); (C) Ago2/RISC complex immunoprecipitation was carried out with a pan-Ago2 antibody in BV2 microglial cells overexpressing miR-NC or miR-223–5p (IgG served as a negative control, and GAPDH served as an internal control (n = 3)); (D) USP8 mRNA levels were quantified in BV2 cells treated with miR-223–5pOE-T-Exos and miR-223–5pKD-T-Exos (n = 3); (E, F) USP8 protein levels in BV2 cells were measured following treatment with miR-223–5pOE-T-Exos and miR-223–5pKD-T-Exos, with quantitative analysis provided (n = 3); (G–K) rescue experiments for miR-223–5p overexpression were performed by upregulating USP8 in microglia (NLRP3 inflammasome activation was analyzed by Western blot and immunofluorescence (n = 3, scale bar = 20 μm)); (L–P) Downregulation of USP8 expression in microglia was used to rescue miR-223–5p inhibition (NLRP3 inflammasome activation was assessed using Western blot and immunofluorescence (n = 3, scale bar = 20 μm)).

To confirm the connection between miR-223–5p and USP8, we conducted rescue experiments to confirm that T-Exos’s regulation of NLRP3 inflammasome activation in microglia via miR-223–5p is mediated by modulating USP8. First, we overexpressed USP8 in microglia through lentiviral transfection and cotreated them with miR-223–5p^OE-T-Exos. The Western blot analysis demonstrated that USP8 expression led to a significant enhancement in NLRP3 inflammasome activation in microglia (Figures G and H). This finding was further corroborated by immunofluorescence staining (Figures I–K). Conversely, shRNA was utilized to knock down USP8 in microglia and cotreated these cells with miR-223–5p^KD-T-Exos. Western blot results revealed that USP8 knockdown reduced the level of NLRP3 inflammasome activation in microglia (Figures L and M), corroborated by immunofluorescence staining (Figures N–P). These findings support our hypothesis, demonstrating that miR-223–5p in T-Exos regulates the activation of NLRP3 inflammasome in microglia via its target, USP8.

Ubiquitination, a post-translational modification that has demonstrated significant biological importance, is a key player in diverse biological processes involving protein metabolism, cell cycle progression, cell death, and transcription. , Deubiquitinating enzymes (DUBs) are classified as a broad group of proteases that oppose the protein ubiquitination effect by removing ubiquitin moieties from target proteins. They play critical roles in regulating cellular protein homeostasis and participating in cellular processes such as immune responses, particularly by enhancing protein stability. , Numerous studies conducted in the field provide substantial evidence that ubiquitination is essential for the regulation of NLRP3 induced inflammasome activation. According to Liu et al., for example, the E3 ubiquitin ligase MARCH2 prevents cardiac ischemia-reperfusion injury by preventing the myocardium’s NLRP3 inflammasome from activating. Ha et al. demonstrated that the initiation of NLRP3-mediated inflammasome activation is initiated by SERTAD1 through the limitation of NLRP3 polyubiquitination. Other studies indicate that USP22 can deubiquitinate and stabilize ATG5, negatively regulating NLRP3. Liu et al. also reported that USP19 regulates NLRP3 function through autophagy, inhibiting inflammation and promoting M2-like macrophage polarization. In this study, we demonstrate that deubiquitinating enzyme USP8 is a miR-223–5p target gene that regulates NLRP3 inflammasome activation in microglia.

USP8 Interacts with and Stabilizes the NLRP3

To further elucidate the potential mechanism by which USP8 promotes NLRP3 inflammasome activation, we conducted co-immunoprecipitation (Co-IP) coupled with MS (IP/MS) experiments to determine which protein binds to USP8. IP/MS identified NLRP3as a putative USP8-interacting protein (Figure A). Co-IP confirmed the interaction, demonstrating that USP8 specifically precipitates NLRP3 in microglia, while the control group using IgG did not show such a precipitation. Reverse Co-IP further validated that NLRP3 can also significantly precipitate USP8 in microglia (Figure B). Co-IP was performed on HEK 293T cells using Flag-tagged USP8 and Myc-tagged NLRP3. Flag-tagged USP8 was coprecipitated with Myc-tagged NLRP3, as seen in Figure C–D. These findings support the notion that USP8 and NLRP3 interact. After this binding was established, we explored at how USP8 knockdown affected the levels of NLRP3 expression in BV2 microglia. While NLRP3 mRNA levels were unaffected by USP8 silencing, NLRP3 protein levels were considerably decreased (Figures E and F), suggesting that USP8 controls NLRP3 protein levels, as opposed to the mRNA level. The proteasome inhibitor MG132 could be used to counteract the decrease in NLRP3 protein levels seen following USP8 deletion (Figure F). Next, we examined whether USP8 stabilizes NLRP3, finding that overexpression of USP8 resulted in increased NLRP3 protein levels, while the noncatalytic C786A mutant of USP8 had no effect on NLRP3 protein levels (Figure G). In this study, the impact of USP8 inhibition on endogenous NLRP3 protein stability was investigated in the presence of a protein synthesis inhibitor, cycloheximide (CHX). To assess this impact, the half-life of the NLRP3 protein was measured. The results of the study revealed that USP8 inhibition resulted in a significant reduction in the half-life of the NLRP3 protein in microglia, in comparison to that in control cells (Figures H and I).

7.

USP8 interacts with and stabilizes NLRP3. (A) IP/MS analysis to identify proteins binding to USP8 revealed NLRP3 as an interacting protein; (B) endogenous interactions were confirmed in BV2 microglia lysates through immunoprecipitation with anti-USP8 or anti-NLRP3 antibodies, followed by immunoblotting with anti-NLRP3 or anti-USP8, respectively (n = 3); (C, D) exogenous overexpression of USP8 and NLRP3 in HEK 293T cells similarly demonstrated their binding interaction (n = 3); (E) NLRP3 mRNA levels in BV2 microglia transfected with shCtrl or shUSP8 were measured (n = 3); (F) NLRP3 and USP8 protein levels in BV2 microglia transfected with shUSP8, with and without treatment with MG132 (n = 3); (G) Flag-tagged USP8 (WT or C786A mutant) was transfected into BV2 microglia at increasing concentrations. Cell lysates were analyzed by immunoblotting with anti-NLRP3 and anti-Flag antibodies (n = 3); (H, I) NLRP3 protein levels in shCtrl and shUSP8 BV2 microglia were evaluated by immunoblotting with anti-NLRP3 and anti-USP8 antibodies in the presence of CHX (10 μg/mL) at various time points (n = 3); (J, K) Lysates from BV2 microglia transfected with shCtrl or shUSP8 and treated with MG132 prior to harvest were immunoprecipitated, and relative ubiquitin-NLRP3 levels were quantified; (L, M) following transfection with Flag-tagged USP8 (WT) or Flag-tagged USP8 C786A, HEK 293T cell lysates were immunoprecipitated with anti-Myc antibodies and immunoblotted with anti-HA and anti-Myc antibodies, in addition to HA-tagged Ub and Myc-tagged NLRP3 (relative Ub-NLRP3 level quantification is displayed (right, n = 3)); (N) spinal cord lysates from mice in the specified groups after injury were immunoprecipitated with anti-NLRP3 antibodies, followed by immunoblotting with anti-Ub and anti-NLRP3 antibodies (relative Ub-NLRP3 levels are quantified (right, n = 3)); (O) HEK 293T cells were co-transfected with Flag-USP8, Myc-NLRP3, and the indicated HA-Ub, Lys0, Lys48-only, or Lys63-only plasmids, and the NLRP3 ubiquitylation linkage was assessed (n = 3); (P) HEK 293T cells transfected with Ub WT or Ub Lys48R were cultured for 72 h in the presence of shCtrl or shUSP8. Immunoblotting of cell lysates was performed with anti-NLRP3 and anti-USP8 antibodies (n = 3).

It was then investigated whether USP8’s ability to stabilize NLRP3 is related to its function as a deubiquitinating enzyme. When USP8 was knocked down, NLRP3 ubiquitination increased significantly, while protein levels decreased compared to the shCtrl group (Figures J and K). In order to gain further insight into USP8’s regulatory effect on NLRP3 ubiquitination, HEK 293T cells were co-transfected with Flag-USP8 (either wild-type or the C786A mutant), Myc-NLRP3, and HA-Ub. Furthermore, it was shown that the overexpression of wild-type USP8 led to a reduction in NLRP3 ubiquitination, while the C786A mutant exhibited no significant effect on its ubiquitination levels (Figures L and M). Next, the effect of miR-223–5p on NLRP3 ubiquitination was investigated in a mouse model of SCI. Following SCI, spinal cord samples from the miR-223–5p^OE-T-Exos treatment group exhibited significantly increased NLRP3 ubiquitination and degradation compared to the miR-NC^OE-T-Exos treatment group. In contrast, spinal cord samples from the miR-223–5p^KD-T-Exos treatment group showed significantly reduced NLRP3 ubiquitination and degradation, in comparison to those from the miR-NC^KD-T-Exos treatment group (Figure N). Polyubiquitin chains are generally classified into two primary types: Lys48-linked and Lys63-linked. Our findings indicate that USP8 specifically cleaves Lys48-linked polyubiquitin chains from NLRP3, without significantly affecting Lys63-linked polyubiquitin chains (Figure O). Furthermore, we introduced a Lys48-resistant (Lys48R) variant of ubiquitin into HEK 293T cells with USP8 knockdown, which confirmed that NLRP3 degradation relies on Lys48-linked polyubiquitination. This intervention successfully averted the drop in NLRP3 protein levels caused by USP8 knockdown (Figure P).

USP8 cleaves ubiquitin from proteins such PD-L1, TβRII, GPX4, and DDX3X, which are involved in tumor microenvironment remodeling, tumor growth, inflammatory responses, and autoimmune reactions. − However, USP8’s function in neuroscience, notably in regulating NLRP3 inflammasome activation in microglia after spinal cord damage, remains unknown. We discovered that NLRP3 is a USP8-interacting protein in microglia by using IP/MS to investigate the likely mechanisms and protein–protein interactions underlying USP8’s control of NLRP3 inflammasome activation. , NLRP3 expression and activity are subject to regulation by a variety of mechanisms, including transcriptional, post-transcriptional, and translational processes. In the ordinary course of events, the levels of protein and the degree of transcriptional activity of NLRP3 remain at relatively low levels. Recent research has revealed many deubiquitinating enzymes (DUBs) that interact with NLRP3, preventing its ubiquitination and consequent destruction. Yalcinkaya et al. demonstrated that BRCC3-mediated NLRP3 deubiquitination promotes inflammasome activation and atherosclerosis. Song et al. reported that the UAF1 deubiquitinase complex enhances NLRP3 inflammasome activation by promoting NLRP3 expression. Therefore, uncovering novel DUBs that directly modulate NLRP3 ubiquitination and proteasomal degradation is essential. In this study, we demonstrate that USP8 interacts with NLRP3, diminishing its ubiquitination and consequently reducing its proteasomal degradation in microglia. These findings suggest that USP8 regulates the inflammatory response by modulating the NLRP3 stability. In addition, the result showed that miR-223–5p overexpression promotes NLRP3 ubiquitination and degradation via suppression of USP8 expression. Since T-Exos are the main focus of our current work, USP8 knockout (USP8 KO) and USP8 transgenic (USP8 TG) mice can be used to further study and validate the precise functional role of USP8.

Conclusion

In this study, a TSA pretreatment strategy was employed to optimize the biological functions of umbilical cord mesenchymal stem cell-derived exosomes, systematically elucidating their critical role and molecular mechanisms in neural repair following spinal cord injury. Experimental results demonstrated that T-Exos significantly inhibited NLRP3 inflammasome activation in microglia, promoted their polarization toward the anti-inflammatory M2 phenotype, and markedly reduced the release of pro-inflammatory cytokines. These effects collectively remodeled the inflammatory microenvironment in the spinal cord injury area, facilitating neural tissue regeneration and functional recovery. Further investigation revealed that miR-223–5p, highly enriched in T-Exos, served as a core molecular mediator of its therapeutic effects: miR-223–5p targeted the deubiquitinating enzyme USP8, blocking its stabilization of NLRP3, thereby accelerating NLRP3 ubiquitination-mediated degradation. This process ultimately suppressed neuroinflammatory cascades and alleviated oxidative stress damage. These findings not only deepen the understanding of the mechanisms underlying stem cell-derived exosome therapy but also establish a theoretical foundation for developing precision therapeutic strategies based on engineered exosomes.

miRNA sequencing data were obtained from GSE212320.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.5c09766.

• In vivo imaging and bioinformatics analysis results of exosomes (PDF)

The experimental design and artwork were collaboratively accomplished by J.G. and B.W. J.G. prepared the initial draft of the paper. G.S. executed the animal model establishment and neurobehavioral assessments, while J.W. managed cellular experiments including isolation and purification of exosomes. The molecular biology investigations involving immunoblotting, quantitative polymerase chain reaction, and cytochemical staining were jointly completed by Z.Q. and Y.C. Critical manuscript evaluation and academic editing were undertaken by Z.J and S.C. Final approval for publication was obtained through unanimous consent from all contributors.

#.

Authors Guoqiang Shen and Minhao Liu contributed equally to this work.

This study was supported by Strengthening Health Career through Science and Education of Suzhou (No. MSXM2024068), Science and Technology Development Program of Suzhou (Nos. SKY2023102, SYWD2024080, SKYD2022024), Renaissance Health Service by Education and Science in SuZhou and Wujiang (No. WWK202307), Suzhou and Wujiang key subject.

The study titled “Mechanism of Exosomes Derived from Mesenchymal Stem Cells in the Treatment of Spinal Cord Injury” was approved by the Animal Ethical and Welfare Committee of Kangda College, Nanjing Medical University (Approval Reference No. KD03220001) on 22 October 2022.

The authors declare no competing financial interest.