Umbilical mesenchymal stem cell-derived exosomes facilitate spinal cord functional recovery through the miR-199a-3p/145-5p-mediated NGF/TrkA signaling pathway in rats

Background Although exosomes, as byproducts of human umbilical cord mesenchymal stem cells (hUC-MSCs), have been demonstrated to be an effective therapy for traumatic spinal cord injury (SCI), their mechanism of action remains unclear. Methods We designed and performed this study to determine whether exosomes attenuate the lesion size of SCI by ameliorating neuronal injury induced by a secondary inflammatory storm and promoting neurite outgrowth. We determined the absolute levels of all exosomal miRNAs and investigated the potential mechanisms of action of miR-199a-3p/145-5p in inducing neurite outgrowth in vivo and in vitro. Results miR-199a-3p/145-5p, which are relatively highly expressed miRNAs in exosomes, promoted PC12 cell differentiation suppressed by lipopolysaccharide (LPS) in vitro through modulation of the NGF/TrkA pathway. We also demonstrated that Cblb was a direct target of miR-199a-3p and that Cbl was a direct target of miR-145-5p. Cblb and Cbl gene knockdown resulted in significantly decreased TrkA ubiquitination levels, subsequently activating the NGF/TrkA downstream pathways Akt and Erk. Conversely, overexpression of Cblb and Cbl was associated with significantly increased TrkA ubiquitination level, subsequently inactivating the NGF/TrkA downstream pathways Akt and Erk. Western blot and coimmunoprecipitation assays confirmed the direct interaction between TrkA and Cblb and TrkA and Cbl. In an in vivo experiment, exosomal miR-199a-3p/145-5p was found to upregulate TrkA expression at the lesion site and also promote locomotor function in SCI rats. Conclusions In summary, our study showed that exosomes transferring miR-199a-3p/145-5p into neurons in SCI rats affected TrkA ubiquitination and promoted the NGF/TrkA signaling pathway, indicating that hUC-MSC-derived exosomes may be a promising treatment strategy for SCI. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-021-02148-5.


Introduction
Acute traumatic spinal cord injury (SCI) has a prevalence rate of 10-83 per million [1] worldwide each year and usually results in devastating functional loss below the level of the injured spinal cord. The pathophysiology of SCI consists of primary mechanical injury and a subsequent cascade of inflammation, ischemia, and secretion of cytotoxic substances that aggravates SCI damage [2][3][4].
Neurons play a central role in transducing biological signals [5], but they fail to receive and send electrical and chemical signals following injury. Even though SCI patients achieve certain recovery of locomotion [6,7], injured axons rarely reconnect to the same extent as axons in the normal spinal cord, and patient motor function never returns to normal [8]. A failed neuronal reconnection of two severed axons mainly occurs for the following reasons [8]: (a) loss of intrinsic outgrowth capacity of neurons, (b) deterioration of the microenvironment, and (c) inhibition of fibrous scar and glial scar formation. Therefore, promoting neurite outgrowth and improving the microenvironment of the remaining neurons should be essential for locomotor recovery in SCI patients.
TrkA, a critical member of the receptor tyrosine kinase (RTK) family composed of TrkA, TrkB, TrkC, and p 75 NTR, can be phosphorylated and induced to form intracellular dimers in the structural domain by coupling with NGF, which subsequently initiates signaling cascades through phosphorylating Ras protein.
Based on previous studies on the mechanism of action of NGF/TrkA, which is indispensable in the development and maturation of the nervous system [9]. Exosomes, important paracrine factors of stem cells, are nanosized particles with a diameter of 20 to 150 nm and consist of a lipid bilayer that encapsulates RNAs, DNAs, and soluble proteins [10,11]. Currently, an increasing number of researchers have demonstrated that stem cell-derived exosomes share similar therapeutic effects as stem cells, such as inhibition of proinflammatory cytokines [12], inactivation of A1 astrocytes [13], and conversion of macrophages to the M2 phenotype [14].
Through preliminary experiment, we demonstrated that UC-MSC exosomes can help to stimulate the NGF/ TrkA signaling pathway and promote the neuronal markers expression (NF-H, Neu-N, and β-tubulin-III) inhibited by LPS in PC12 cells. Subsequently, we conducted this study to confirm the specific mechanisms of exosomes in inducing neurite outgrowth in vivo and in vitro and the efficacy of exosomes in treating SCI.

Ethical statement
Ethical approval for the experiment was obtained from the Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University, China (Ethic number: IACUC-F3-19-0401). All experiments were performed following relevant laws and guidelines and the institutional guidelines of the Third Affiliated Hospital of Sun Yat-sen University. This article does not involve any human studies.

Animal experimental protocol
We purchased adult female Sprague-Dawley rats (6-7 weeks old), all specific pathogen-free (SPF), from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China). All rats were housed in an SPF environment (temperature, 22 ± 1°C; humidity, 65-70%) with free access to food and water, and five rats were assigned per cage. All animal experiments were approved by the animal ethics committee of the Third Affiliated Hospital of Sun Yat-sen University and in accordance with the guidelines included in the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Eighth Edition). In total, seventy-six rats were sacrificed by CO 2 suffocation (detailed animal information is provided in Additional file 1).

Exosome purification and characterization
Exosomes were isolated from the supernatants of passage 5 hUC-MSCs according to the Total Exosome Extraction Reagent (Invitrogen, Carlsbad, California, USA) instructions. Before being subjected to transmission electron microscopy (TEM; Tokyo, Japan) and the qNano nanoparticle analysis system (Izon, Christchurch, New Zealand), exosomes in suspension were filtered using a 220-nm filter and stained with 3% aqueous phosphotungstic acid. Exosomes were lysed with radioimmunoprecipitation assay (RIPA) buffer (Beyotime, Shanghai, China) supplemented with a protease and phosphataseinhibitor cocktail (CST, Boston, MA, USA). The protein concentration of exosomes was quantified using the BCA Protein Assay Kit (Thermo Fisher, Boston, MA, USA). Primary antibodies against CD63 (Abcam, Cambridge, UK), CD9 (Abcam, Cambridge, UK), and β-actin (CST, Boston, MA, USA) were used to characterize exosomes (detailed methods and materials are provided in Additional file 2).

Statistical analysis
Statistical analysis was performed, and graphs were generated by using Graph Pad Prism 8.0 (San Diego, CA, USA). All data are presented as the mean ± standard deviation (SD) and the range from min to max, and differences were analyzed using one-way analysis of variance (one-way ANOVA). P < 0.05 was considered statistically significant.

Results
Exosomal miRNA profile analyzed by absolute quantitative sequencing We obtained hUC-MSCs from Guangzhou Selera Stem Cell Technology Co., Ltd., and the cells that presented a long spindle shape were identified by flow cytometry. Stem cell markers (CD73, CD90, CD105) ( Fig. 1a) were highly expressed, while non-stem cell markers (CD45, CD34, HLA-DR) (Fig. 1b) were expressed at extremely low levels on the cell surface. Then, we purified exosomes from the culture supernatants of MSCs by the Total Cell Exosome Extraction Kit. Analysis of the morphology and surface markers confirmed that the obtained pellet were exosomes. First, CD9 and CD63 were detected in these particles, which did not express β-actin (Fig. 1c). Nanoparticle tracking analysis (NTA) revealed the concentration and size of the particles ranging from 30 to 150 nm (Fig. 1d). Furthermore, a typical cup-shaped morphology of these nanoparticles was observed by transmission electron microscopy (TEM) (Fig. 1e).
miRNAs are the most abundant nucleic acids in MSCderived exosomes (MSC-Exo) and play important roles in regulating cell activities. To obtain a better understanding of the exosomal miRNA profile, we performed small RNA absolute quantitative sequencing (see Additional file 6) of MSC-Exo, and the results are deposited in the GEO database of NCBI (GEO database: GSE 159814). The sequence results are provided in Additional file 6. Small RNA sequencing identified approximately 990 miRNAs in MSC-Exo. Further analysis showed a linear relationship between log-transformed standard deviation (SD) and transcripts per million (TPM) values (see Additional file 7A).
Interestingly, we found different miR-199a-3p/145-5p expression profiles between primary neurons, astrocytes, endothelial cells, and meningeal fibroblasts. Among these four types of primary cells, neurons had the highest levels of miR-145-5p and the lowest levels miR-199a-3p (see Additional file 8B). All primary cells were isolated from rats (see Additional file 8C). Moreover, through QRT-PCR analysis, we also discovered more miR-199a-3p transcripts than miR-145-5p transcripts in PC12 cells (See Additional file 8D).
Furthermore, we pretreated PC12 cells, which were subsequently stimulated by NGF and LPS, with MSC-Exo to investigate whether exosomal miR-199a-3p/145-5p were indispensable in NGF-induced PC12 differentiation. Our results showed that neuronal marker expression (NF-H, β-tubulin-III, Neu-N) suppressed by LPS could be upregulated by Exo; however, this protective effect was partially abolished by knockdown of miR-199a-3p/145-5p in MSC-Exo (Fig. 2g-i). The results also showed that the downregulation of p-Akt and p-Erk induced by LPS was also reversed by increasing doses of MSC-Exo (Fig. 2k, l). Moreover, the inhibition of PC12 neurite outgrowth induced by LPS, which was partially reversed by MSC-Exo, was closely related to exosomal miR-199a-3p/145-5p expression (Fig. 2m, n). More importantly, we observed that neurons were greatly preserved at 3-and 35-days post injury (DPI) as demonstrated by Neu-N staining (Fig. 3). The compromised effect of Exo-K on the promotion of neurite outgrowth, which was inhibited by LPS, was also observed in primary neurons in vitro (see Additional file 9A). Thus, the presence of miR-199a-3p/145-5p recapitulated the neuroprotective effects of MSC-Exo, while miR-199a-3p/ 145-5p knockdown suppressed these effects. Moreover, we found that miR-199a-3p/145-5p levels in MSC-Exo were correlated with increased endothelial cell migration and tube formation abilities, which were suppressed by LPS (see Additional file 9B-E), while these effects were compromised by MSC-Exo-K. This finding may explain why exosomes are often administered in the early stage of SCI, as the blood supply is potentially impacted as well. In summary, miR-199a-3p/145-5p participated in MSC-Exomediated neurite outgrowth in vitro and in vivo.
miR-199a-3p and miR-145-5p targeted the Cblb and Cbl genes, respectively Through bioinformatics prediction, we hypothesized that Cblb and Cbl were direct targets of miR-199a-3p and miR-145-5p, respectively (Fig. 6a, b). To investigate whether miR-199a-3p regulated Cblb gene expression at the transcriptional or posttranscriptional level, we transfected both miR-199a-3p/145-5p mimics and inhibitors into PC12 cells. Western blotting and QRT-PCR showed that Cblb, and Cbl protein and mRNA expressions decreased with increasing doses of the mimics and increased with increasing doses of the inhibitors (Fig. 6c-e). These results suggested that miR-199a-3p/145-5p suppressed Cblb and Cbl expression in PC12 cells at the transcriptional and posttranscriptional levels.
Furthermore, we cloned the two predicted wild-type binding sites (WT-1/-2) and the mutant sites into plasmids and cotransfected the pGL3-reporter and pRL-TK vector with miRNA mimics or inhibitors into HEK293 cells. A previous study predicted a relationship between miR-199a-3p and Cblb, while our results demonstrated a direct interaction between them. Significant changes in relative luciferase activity in HEK293 cells transfected Fig. 3 Preserved neurons in four groups of rats identified by Neu-N. Representative immunofluorescence images of Neu-N/GFAP Exo-K in shamoperated rats, SCI rats, SCI rats treated with Exo, and SCI rats treated with Exo-K at 3 DPI and 35 DPI. Scale bar = 500 μm, Scale bar = 100 μm with scramble sequence (Si NC group) were not observed, while decreased luciferase activity along with increasing doses of the miR-199a-3p mimic was determined ( Fig. 6f, g). Increased luciferase activity in HEK293 cells was also observed when cells were transfected with 150 nm miR-199a-3p mimics (Fig. 6f, g). Regardless of whether the cells were transfected with miR-199a-3p mimics or inhibitors, the luciferase activity of HEK293 cells transfected with a vector carrying mutant sites was not affected (Fig. 6f, g).
Regardless of whether the cells were transfected with miR-145-3p mimics or inhibitors, the luciferase activity of HEK293 cells transfected with a vector carrying the mutant site was not significantly changed (Fig. 6h). The scramble control did not influence luciferase activity in HEK293 cells. These results suggested that miR-199a-3p/145-5p suppressed Cblb and Cbl expression via direct binding to their predicted mRNA seed regions.

Cblb and Cbl regulated the NGF/TrkA pathway through modulation of TrkA ubiquitination
To further explore how Cblb and Cbl interact with the NGF/TrkA pathway, we knocked down or overexpressed these two genes in PC12 cells. First, we designed three siRNAs against the Cblb and Cbl genes, two of which (Si 10D-E). When Cblb and Cbl were knocked down by the siRNAs, total ubiquitinated proteins in PC12 cells were greatly reduced; however, a marked increase in TrkA was observed compared with that in nontreated cells (Fig. 7a, b). Single knockdown of Cblb or Cbl caused upregulation of p-Erk and p-Akt, and NGF activated p-Erk and p-Akt (Fig. 7c, d), subsequently promoting NGFinduced neurite outgrowth in PC12 cells (see Additional file 10C; 10F).
Then, we successfully upregulated Cblb and Cbl expression in PC12 cells (see Additional file 11). Overexpression of Cblb and Cbl led to higher ubiquitination levels of total proteins and lower expression of TrkA (Fig. 7e, f). NGF stimulation accelerated the degradation process of TrkA when Cblb and Cbl were overexpressed, but this process was inhibited by the proteasome inhibitor MG-132 (Fig. 7e, f). We also found that NGF-activated p-Erk and p-Akt were further upregulated in Cblb-and Cbl-overexpressing cells by the protease inhibitor MG-132 (Fig. 7g, h). More importantly, we pulled down Cblb and Cbl proteins in PC12 cells with knockdown and overexpression of Cblb and Cbl by using a primary antibody against TrkA, further confirming the direct interaction between Cblb and TrkA (Fig. 7i, j).
Exosomal miR-199a-3p/145-5p alleviated damage to the lesion site and facilitated locomotor function in SCI rats In vivo, we injected the rats with PBS, Exo, and Exo-K via the tail vein. The high expression of miR-199a-3p and miR-145-5p in rats receiving Exo injection was partially downregulated in rats receiving Exo-K injection, which implied the successful delivery of the miRNAs by exosomes (Fig. 8a). Furtherly, track of PKH26-labeled exosomes by immunofluorescence in vivo proved this point (see Additional file 12). We also observed that the increase in TrkA in rats receiving Exo injection was partly downregulated in rats receiving Exo-K injection at 3 DPI (Fig. 8b). Exo injection exerted an antiapoptotic effect on the lesion site, while this effect was abrogated with the inhibition of miR-199a-3p/145-5p (Fig. 8c, d). Decreased inflammatory levels in vivo were also partially Fig. 6 miR-199a-3p and miR-145-5p targeted the Cblb and Cbl genes, respectively. Predicted binding sites of miR-199a-3p (a) and miR-145-5p (b) on Cblb and Cbl mRNAs. c-e Cblb and Cbl gene expressions in PC12 cells transfected with different doses of miR-199a-3p/145-5p mimics and inhibitors assayed by western blot and QRT-PCR (n = 3). Data are represented as beanplot exhibiting values from 3 independent experiments. Luciferase reporter assay of the first predicted binding site (f) and the second predicted site (g) of miR-199a-3p in HEK293 cells. Data are represented as mean ± SD. h Luciferase reporter assay of the predicted binding site of miR-145-5p in HEK293 cells (n = 3). Data are represented as mean ± SD. *p < 0.05; **p < 0.01; #p > 0.05. WT wild type, MT mutant, QRT-PCR quantitative real-time polymerase chain reaction attributed to exosomal miR-199a-3p/145-5p levels (see Additional file 13).
The Basso, Beattie and Bresnahan (BBB) scale was used to assess the locomotor function of the hindlimbs of rats, and the values were recorded from onset to 35 DPI. Strongly impaired locomotor function of the hindlimbs of rats was observed on the 1st DPI, and significant improvement was detected on the 7th DPI (Fig. 8e). Hematoxylin-eosin (HE) staining showed that the damage at the lesion site of SCI rats was smaller than that of SCI + Exo rats. Interestingly, this protective effect was partially suppressed when miR-199a-3p/145-5p were inhibited in MSC-Exo (Fig. 8f, g). We further detected spinal cord damage in vivo, and the MRI results indicated that MSC-Exo greatly decreased the lesion size of the injured spinal cord, while Exo-K exerted weaker protective effects on the damaged spinal cord (Fig. 8h). These data indicated the pivotal roles of miR-199a-3p/ 145-5p in MSC-Exo-mediated treatment of SCI in vivo (Fig. 9). Cbl and Cbl inhibition by exosomal miR-199a-3p/145-5p emphasized the neuroprotective effect of MSC-Exo.
Discussion miRNAs, noncoding RNAs of approximately 22 nucleotides, depending on the exosomal origins, are markedly enriched in exosomes and target mRNAs by binding to their 3′UTRs [16,17]. The hUC-MSC-derived exosomal miRNA expression profiles first identified by Fang using QRT-PCR (miR-21, -23a, -125b, -145) [18] and then quantified by us using absolute quantitative sequencing shared similar results. Cargo miRNAs, as the most abundant nucleotides in exosomes, are receiving increasing attention from researchers, mainly owing to the improved understanding of miRNAs and advances in the techniques used to investigate miRNAs.
Findings showing attenuated central inflammation in rats through detection of IL-1, IL-6, and TNF-α; upregulation of proinflammatory cytokines (IL-4, IL-10); and reduction in apoptotic neurons in the early process of exosomal treatment [14] were consistent Fig. 8 Function of exosomal miR-199a-3p/145-5p in vivo in SCI rats. a Relative miR-199a-3p/145-5p expression levels in SCI and SCI of MSC-Exoand Exo-K-treated rats at 3 DPI (n = 3). Data are represented as mean ± SD. b Representative images and quantitative analysis of TrkA in SCI, SCI treated with Exo, and SCI treated with Exo-K rats at 3 DPI (n = 3). Data are represented as mean ± SD. c, d Representative images and quantitative analysis of TUNEL staining in sham-operated rats, SCI rats, SCI rats treated with Exo, and SCI rats treated with Exo-K 3 DPI (n = 3). Scale bar = 500 μm, scale bar = 100 μm. Data are represented as mean ± SD. e BBB scores of the four groups of rats (sham, SCI, SCI + Exo, and SCI + Exo-K) on the 1st, 3rd, 5th, 7th, 14th, 21st, 28th, and 35th DPI (n = 10/group). Data are represented as mean ± SD. f, g Quantitative analysis and representative images of the lesion size (HE staining) in the four groups of rats at 35 DPI (n = 10). Data are represented as mean ± SD. Scale bar = 1000 μm. h Representative MRI images of the lesion area (red arrow) of rats at 35 DPI. *p < 0.05; #p > 0.05. SCI, spinal cord injury; DPI, day post injury; MSC-Exo, MSC-derived exosomes; Exo-K, MSC-Exo with miRNA inhibition; BBB, Basso, Beattie and Bresnahan; HE, hematoxylin-eosin with our results. Improvement of the microenvironment at the lesion site enhanced neurite outgrowthrelated p-Akt/p-Erk pathway activation by NGF in PC12 cells and decreased secondary apoptosis of neurons. However, there is less evidence that exosomal miRNAs are involved in the protection of neurons at the early stage of SCI, and the underlying mechanism remains unclear. The injury model of PC12 cells was established by using LPS instead of other inflammatory cytokines as previously reported [19] because both LPS model and other inflammatory models truly reflect the damaged status of neurons under inflammation [20,21]. Through knockdown of mir-199a-3p/ 145-5p in exoxomes, we demonstrated the crucial role of mir-199a-3p/145-5p in the neuronal differentiation promoted by exosomes. Furtherly, we predicted that Cblb and Cbl genes are targeted by miR-199a-3p and miR-145-5p.
The protooncogenes Cbl and Cblb have been suggested to have a negative effect on RTKs, and the regulation of both genes is analogous to that of T cell receptor and epidermal growth factor (EGFR) [22][23][24][25]. As demonstrated for RTKs, E3 ubiquitin ligases such as NEDD4-2, tumor necrosis factor receptor-associated factor 6 (TRAF6), Cbl, and Cblb facilitate TrkA ubiquitination, internalization, and turnover [26][27][28]. Ubiquitins (Cbl and Cblb) are conjugated onto TrkA along with the onset of self-phosphorylation of intracellular peptide fragments of TrkA, which can be suppressed by exosomal miR-199a-3p/145-5p in vivo and in vitro. Through an in vitro study, we determined that the Cbl mRNA 3′ UTR was the target of miR-145-5p and that the Cblb mRNA 3′UTR was the target of miR-199a-3p. The predicted biding site of Cbl mRNA for miR-145-5p is first reported and proved by us; interestingly, the predicted biding site of Cblb mRNA for miR-199a-3p was previously reported but not examined in RAW264.7 cells by Rong [29].
In the absence of the Cbl and Cblb genes, an activation of the p-Akt/p-Erk pathways and increased neurite growth stimulated by NGF were observed. However, both the p-Akt/p-Erk pathways were soon downregulated, which was associated with rapid degradation of TrkA in response to NGF stimulation and the overexpression of Cblb and Cbl genes. Similar to Takahashi's results [27][28][29][30], our co-IP data showed that TrkA ubiquitination and degradation were statistically affected by Cbl, but the domain of TrkA required for ubiquitination by Cbl remains unclear. The mechanisms of the TrkA-Cblb interaction may involve the TKB domain (an SH2like domain) promoting the binding of phosphorylated tyrosine on activated TrkA and Src homology 2 adapter Fig. 9 Signaling pathways affected by exosomal miR-199a-3p/145-5p. Exogenous exosomes from human umbilical cord mesenchymal stem cells are taken up by damaged neurons and release abundant levels of miRNA-199a-3p/145-5p, which target Cblb and Cbl mRNAs to inhibit the ubiquitination-mediated degradation of TrkA. Sustained activation of p-Erk and p-Akt leads to persistent expression of Neu-N, NF-H, β-tubulin-III, and neurite outgrowth protein 2 (SH2B2, also known as APS) [31]. We showed from two aspects that the levels of Cblb and Cbl proteins, which could be clearly reduced by exosomal miR-NAs, were associated with neurite outgrowth in PC12 cells and were potential influencing factors for SCI by affecting the turnover and/or activation of TrkA and maintaining sustained activation of p-Erk/p-Akt and neurite outgrowth. Intriguingly, this theory is potentially suitable for other E3 ubiquitin ligases, other Trk members regulating neurite outgrowth, and these E3 ubiquitin ligases are underlying targets of exosomes that are enriched for regulatory noncoding RNAs. Furthermore, knockdown of both miRNAs in exosomes directly led to lower expression of TrkA in vivo and impaired neurological function in SCI rats, which emphasized the effects of miR-199a-3p/145-5p in vivo.
Even though multiple drugs, such as nonsteroidal antiinflammatory drugs (NSAIDs), anti-CD11d antibodies, minocycline, progesterone, estrogen, riluzole, polyethylene glycol (PEG), astatin and inosine, are already used in human clinical applications, satisfying outcomes are rarely obtained [32]. The use of exosomes in treating SCI is mainly due to the therapeutic effect of stem cells, which is thought to be a result of paracrine mechanisms, and an increasing number of researchers speculate that exosomes are a promising next-generation cell-free agent. However, rigorous assessment of their efficacy and safety and detailed mechanisms of action are urgently needed before or in conjunction with clinical application.

Conclusions
In summary, miR-199a-3p/145-5p derived from hUC-MSCs could promote locomotor functional recovery of SCI rats potentially by targeting the Cblb and Cbl genes, which resulted in modulation of the turnover and/or activation of TrkA. Furthermore, exosomes tended to promoted neurite outgrowth by decreasing the inflammatory level at the lesion site of the spinal cord, creating a favorable environment for neurite outgrowth.