According to the etiology, spinal cord injury (SCI) can be divided into traumatic and nontraumatic. Traumatic SCI is often caused by severe damage to the spinal cord due to external physical impacts (for example, car accidents, falls, sports-related injuries, or violence). Nontraumatic SCI often occurs during acute or chronic diseases (for example, tumors, infections, disc herniation, or vertebral fracture-dislocations), causing spinal cord compression to produce primary and secondary injuries. SCI not only affects the quality of life of patients but also poses a heavy burden on their families. Therefore, it is essential to prevent the occurrence of SCI; for unpreventable SCI, it is critical to develop effective treatments (Ahuja et al., 2017).

A central concept in managing any SCI patient has been “time is the spine” (Ahuja et al., 2017). SCI is characterized by the progressive loss of neurologic function within a few hours. Therefore, it is essential to rapidly diagnose patients and provide neuroprotective interventions in the acute injury stage. Because of these treatments including hemodynamics (Ryken et al., 2013), hormonal therapy (Hurlbert et al., 2015), and surgical decompression (Ramakonar and Fehlings, 2021), long-term functional recovery is improved in patients. Management of patients with SCI is complex and involves multiple stages of care, often lasting several years after the initial injury. Thus, later rehabilitation is also an integral part of the treatment process (Gómara-Toldrà et al., 2014; van der Scheer et al., 2021).

However, these treatments are not adequate for long-term functional recovery in SCI patients. In recent years, various major breakthroughs have been made in cell therapy to protect and regenerate the damaged spinal cord via various mechanisms such as immune regulation, paracrine signaling, extracellular matrix (ECM) modification, and lost cell replacement. Among them, the most commonly studied and promising cell types include induced pluripotent stem cells (iPSC), mesenchymal stem cells (MSCs), neural stem cells (NSCs), oligodendrocyte progenitor cells (OPCs), Schwann cells (SCs), and olfactory ensheathing cells (OECs; Harrop et al., 2012; Shao et al., 2019; Ahuja et al., 2020). Nevertheless, many recent studies have shown that the cell therapy has many disadvantages, such as tumorigenicity, low survival rate, and immune rejection. Because of these disadvantages, the clinical application of cell therapy is limited (Feng et al., 2021; Liu J. et al., 2021).

The remarkable effect of cell therapy can be attributed to its dominant paracrine effect. Exosomes, an intercellular communication tool, affect normal and pathological conditions. In recent years, the role of exosomes in various diseases and their therapeutic potential have attracted much attention. The same is true for exosomal noncoding RNAs (ncRNAs), which do not encode proteins but affect transcriptional and translational processes by targeting specific mRNAs. Their diverse functions have attracted much interest (Colombo et al., 2014; Quinzaños-Fresnedo and Sahagún-Olmos, 2015; Shi et al., 2018; Dutta et al., 2021). The types of source cells can influence the heterogeneity of exosomes, which have different contents and specific markers from varied cell sources. For example, ERBB2 is specifically expressed in breast cancer cell-derived exosomes, and TSPAN8 is a specific marker of epithelial cell-derived exosomes. The inherent biology and the microenvironment of the cells can also give exosomes distinct functions such as uptake by specific cells and tropism to certain organs (Kalluri and LeBleu, 2020). Especially, exosomal miRNAs have attracted substantial attention because of their various functions in the context of SCI treatment. They vary widely with different cell sources (Cho et al., 2019). For example, neuron-derived exosomes were enriched for miRNA-383, whereas glial cells-derived exosomes were not (Pomper et al., 2020). It is noteworthy that a study directly compared the efficacy of human pluripotent stem cells (hPSCs)-derived and MSCs-derived exosomes in an animal model of ischemic stroke. Results showed that hPSCs-derived exosomes were more effective than MSCs-derived (Webb et al., 2018).

MSCs-derived exosomes are the most widely studied to date in the treatment of SCI. However, no studies have directly compared the efficacy of exosomes from different cell sources in the treatment of SCI (Dutta et al., 2021). Therefore, it is essential to directly compare the therapeutic potential of exosomes from varied cell sources in the treatment of SCI. This review focuses on the mechanism of action and therapeutic potential of exosomes and exosomal ncRNAs from different cell sources in the treatment of SCI. This review also discusses the future opportunities and challenges, proposing that exosomes and exosomal ncRNAs from different cell sources might be promising tools for the treatment of SCI.

According to its pathological process, traumatic SCI has a complex pathophysiological process, and it can be broadly classified into primary and secondary injuries. The primary damage can immediately cause mechanical destruction and dislocation of the spine, causing spinal cord compression or transection. Subsequently, the primary injury leads to a continuous cascade of secondary injuries, causing further damage to the spinal cord and neurological dysfunction. Then, this injury causes damage to myelin, axons, and neurons. Moreover, it also disrupts the blood-spinal cord barrier. The primary damage is often irreversible, while the secondary injury usually causes more severe damage than the primary injury (Tator, 1995; McDonald and Sadowsky, 2002; Ahuja et al., 2017).

SCI is divided into five stages depending on the pathophysiological process and the time of injury: hyperacute phase (0–2 h), acute phase (2–48 h), subacute phase (2–14 days), intermediate phase (2 weeks to 6 months), and chronic phase (>6 months; Rowland et al., 2008). In the hyperacute phase (0–2 h), it is characterized by traumatic axons, hemorrhagic necrosis of gray matter, and microglial activation releasing proinflammatory cytokines. The proinflammatory cytokines TNFα and IL-1β can be immediately released by the microglial after injury (Donnelly and Popovich, 2008). In the acute phase (2–48 h), it is characterized by vasogenic and cytotoxic edema, persistent hemorrhage and necrosis of glutamate-mediated excitotoxicity, disruption of blood-spinal cord barrier (BSCB) permeability, early demyelination, axonal swelling, and neuronal death. The BSCB can remain disrupted even at 28 days after SCI and spread along the entire length of the cord (Whetstone et al., 2003). The complex network of tight junction (TJ) proteins, which are the major protein component of the BSCB, can be modulated by inflammatory cytokines (Lee et al., 2012; Kumar et al., 2017). In the subacute phase (2–14 days), it is characterized by persistent edema, thrombosis, and vasospasm aggravate ischemia. A continual inflammatory cell infiltration causes further cell death, forming cavities. In addition, reactive astrocytes act as a barrier to prevent damage from aggravating but secrete some inhibitory ECM molecules around the lesion. Chondroitin sulfate proteoglycans (CSPGs), which are a key component of the ECM, can inhibit axon regeneration through binding to their major cognate receptor, Protein Tyrosine Phosphatase Sigma (PTPσ; Sakamoto et al., 2019). Therefore, blocking the combination of them can effectively promote axonal regeneration. In the intermediate phase (2 weeks to 6 months), the axons continue to degenerate, and reactive astrocyte scars mature, becoming an effective regenerative inhibitor, and the cysts merge to limit axonal regeneration and cell migration. The scars have two distinct components after SCI: the lesion core, which primarily includes macrophages and fibroblasts, is generally considered as the fibrotic scar, and the lesion border, which is predominantly composed of microglia, reactive astrocytes, and NG2+ oligodendrocyte progenitor cells, is commonly regarded as a glial scar (Bradbury and Burnside, 2019; Tran et al., 2021). In the traditional concept, the glial scar is considered to exert a detrimental function for neurological recovery, which not only secretes some inhibitory extracellular matrix molecules, cytokines, and oxidative stress products but also inhibits axonal regeneration as a chemical barrier (Silver and Miller, 2004). However, many studies have confirmed that glial scar plays a vital role in neuroprotection. It can limit the spread of inflammation at the injury site to the surrounding injury as a barrier. And it can also secrete neurotrophic factors (nerve growth factor and fibroblast growth factor) and extracellular matrix proteins (laminin and fibronectin; Lukovic et al., 2015). Therefore, an increasing number of studies are focusing on the beneficial function of the glial scar for neurological recovery. In the chronic phase (>6 months), the Wallerian degeneration process of severed axons continues, and the severed axons and their cell bodies may take years to be completely removed (Ehlers, 2004). Unfortunately, neurological dysfunction and neuropathic pain will be caused by the formation of syringomyelia (Todor et al., 2000). Therefore, therapeutic strategies aim to improve axonal degeneration and demyelination by drug or cell transplantation.

Exosomes are small extracellular vesicles (EVs) of 40–150 nm, endosome-derived, secreted by most cells. They have been isolated from many biological fluids, including blood, urine, semen, and cerebrospinal fluid (Kalra et al., 2016). During the formation of exosomes, the first phase is the early invagination of the endosome membrane allows intracellular components to be engulfed in early sorting endosomes (ESE), with the participation of mitochondria, Golgi apparatus, and the endoplasmic reticulum (Hessvik and Llorente, 2018). Then, the ESEs can mature into the late sorting endosomes (LSEs) with the participation of the endosomal-sorting complex necessary for transport (ESCRT) proteins (Vietri et al., 2020). Eventually, the LSEs generate multivesicular bodies (MVBs) after the selective integration of substances. The MVBs contain several vesicular intraluminal vesicles (ILVs; van Niel et al., 2018). Then, MVBs fuse with the plasma membrane and then release ILVs as exosomes into the extracellular space. Also, they can fuse with lysosomes or autophagosomes to be degraded (Thery et al., 2002; Colombo et al., 2014; van Niel et al., 2018; Jeppesen et al., 2019; Figure 1) Moreover, exosomes can be taken up by recipient cells through phagocytosis, direct fusion, endocytosis, and ligand-receptor interactions. Then, the contents of the exosomes can be deposited into the cytoplasm (Kalluri and LeBleu, 2020).

Because exosomes are derived from endosomes, these substances, including the proteins involved in MVB formation (Alix and TSG101), membrane transport and fusion (annexins, GTPases), adhesion (integrins), tetraspanins (CD9, CD63, CD81), antigen presentation [major histocompatibility complex (MHC) class molecules], heat shock proteins (HSP70, HSP90), and other related proteins, are commonly used to identify exosomes regardless of the cell type of origin. In addition to proteins, exosomes are rich in specific lipids, mainly containing ceramide, cholesterol, and sphingolipids. And the lipids contribute to the formation and structural stability of exosomes (Mashouri et al., 2019). Exosomes also contain surface polysaccharides and glycans, mainly containing mannose, α-2,6-sialic acid, and polyglactin. They present at the plasma membrane of exosomes, contributing to the docking and attachment of these exosomes to recipient cells, especially the Glypican 1 (Melo et al., 2015). Exosomes have been reported to carry DNA and RNAs, including mRNAs and some ncRNAs (Lotvall et al., 2014; Kalra et al., 2016; Thery et al., 2018; Figure 2) Although the ability of exosomes to contain DNA remains controversial, there have been many pieces of research showing that they are used for identification (Thakur et al., 2014; Hagey et al., 2021). Exosomal RNAs are secreted to regulate intercellular communication; and miRNAs, in particular, play a vital role in various biological mechanisms (Treiber et al., 2019). However, the composition and function of exosomes remain to be fully elucidated.

Around 98% of all the genomic output is ncRNAs in the data from genome-wide transcriptional analysis in humans (Mattick, 2001). According to the size of ncRNAs, ncRNAs are mainly divided into two groups: long ncRNAs with more than 200 nucleotides and small ncRNAs with no more than 200 nucleotides (Mercer et al., 2009). Interestingly, although ncRNAs do not code for proteins, they have diverse functions in physiology and development of the organisms (Amaral and Mattick, 2008). For this reason, an increasing number of studies have shown that many ncRNAs, particularly noncoding small RNAs (microRNAs), long ncRNAs (lncRNAs), and circular RNAs (circRNAs) become differentially expressed after SCI. In the treatment of SCI, these ncRNAs regulate the translation and transcription mainly by targeting specific mRNAs to affect neuronal survival, axonal regeneration, and glial cell phenotype (Zhou et al., 2016; Bie et al., 2021). MicroRNAs (miRNAs) are 20–24 nucleotide RNA molecules with the function of regulating the protein expression levels by affecting mRNA. Their pivotal role in SCI can be attributed to individual miRNAs that can target the translation of many mRNAs (Bhalala et al., 2013). There are two modes of action for miRNAs binding to the 3’-untranslated region (3’-UTR) of target mRNA. One is that the perfect binding of miRNAs to targets induces mRNA degradation, and the other is that imperfect binding of miRNAs to targets represses translation (Shahzad et al., 2021). The two modes of action will prevent protein accumulation by an unknown mechanism. Furthermore, lncRNAs will act as endogenous RNA to compete for miRNAs binding to regulate gene expression. lncRNAs are defined as thousands of RNA transcripts of more than 200 nucleotides in length without protein-coding potential, which has attracted much attention in various fields (Shi et al., 2018). A study showed that the effect of miR-203 and miR-101, which can down-regulate the expression of BARD1 protein, can be counteracted by a novel lncRNA (BARD1 9’L). These findings subvert our perception of the biological function of ncRNAs, from thinking that they are nonfunctional transcriptional junk to gaining insight into their involvement in the pathogenesis of various diseases (Lee, 2012). It is reported that the exosomes play a role in intercellular communication. The attachment of RNA-induced silencing complexes (RISCs) to the ESCRT components makes the ncRNAs recruit to the exosomes (Sato-Kuwabara et al., 2015) and the ncRNAs, which are one of the enriched cargo in exosomes, can be exported outside cells to target specific mRNAs. Ceramide-dependent machinery controls the release of exosomal ncRNAs (Kosaka et al., 2010). For this reason, an increasing number of studies are focusing on the role of the exosomal ncRNAs as potential therapeutic strategies in SCI (Pant et al., 2021). Exosomes derived from different cells exhibit their functions, both in vivo or in vitro, and in healthy or disease states (Lotvall et al., 2014; Thery et al., 2018). Their differential expression in diverse states can be used as a particular biomarker, provide new therapeutic targets for diseases, and even used as a therapeutic approach to replace or assist cell therapy (Chen et al., 2017). Therefore, we summarize the respective roles of exosomes and exosomal ncRNAs obtained from different cell sources in SCI treatment with great confidence for their promotion of functional recovery after SCI (Figure 3).

MSCs-Exos were separated and concentrated by ultracentrifugation and differential centrifugation in many studies (Huang et al., 2017, 2020a; Lankford et al., 2018; Wang et al., 2018; Guo et al., 2019; Ji et al., 2019; Kang et al., 2019; Li et al., 2019; Li C. et al., 2020; Li et al., 2021; Chen et al., 2021; Cheng et al., 2021; Han et al., 2021; Jia et al., 2021; Jiang and Zhang, 2021; Liu W. Z. et al., 2021; Liu et al., 2022; Nakazaki et al., 2021; Nie and Jiang, 2021; Noori et al., 2021; Sheng et al., 2021; Xiao et al., 2021; Xin et al., 2021; Zhang A. et al., 2021; Liang et al., 2022; Zhou et al., 2022) and by precipitation kits/polymer (PEG or others) in some studies (Li et al., 2018; Ren et al., 2019; Xu et al., 2019; Yu et al., 2019; Zhao et al., 2019; Fan et al., 2021; Jia et al., 2021a, b; Kang and Guo, 2022). To achieve better specificity of MSCs-Exos separation, many researchers isolated MSCs-Exos using ultrafiltration-centrifugation combined with density-gradient ultracentrifugation (Liu et al., 2020; Shao et al., 2020; Chang et al., 2021; Luo et al., 2021; Huang et al., 2022). Furthermore, some researchers used one or more techniques following the ultracentrifugation, such as density-gradient ultracentrifugation (Liu et al., 2019), size-exclusion chromatography (Li L. et al., 2020), and magnetic sorting (Kim et al., 2018). NSCs-Exos were separated and concentrated by ultracentrifugation (Ma et al., 2019), ultrafiltration-centrifugation (Zhong et al., 2020), and by kits methods (Kang et al., 2020). To achieve better specificity of NSCs-Exos separation, some researchers isolated NSCs-Exos using ultrafiltration-centrifugation combined with density-gradient ultracentrifugation (Rong et al., 2019). SKP-SCs-Exos were separated and concentrated by kits methods (Wu et al., 2020). SC-Exos (Pan et al., 2021), HOECs-Exos (Tu and Hsueh, 2020), and PMs-Exos (Zhang B. et al., 2021) were separated and concentrated by ultracentrifugation. BMDMs-Exos were separated and concentrated by ultracentrifugation and kits methods (Wang et al., 2020). There are no specific separation and concentration methods for exosomes from different cell sources. However, as reviewed in the 2018 guidelines, different isolation methods have their advantages and disadvantages (Thery et al., 2018). For example, the ultracentrifugation method has intermediate recovery and intermediate specificity. The kit method has high recovery but low specificity. Moreover, ultrafiltration-centrifugation combined with density-gradient ultracentrifugation has low recovery but high specificity. Although there are no high recovery and high specificity exosome isolation methods, dendritic cells (DCs)-derived exosomes have been applied in clinical trials to treat patients with malignant melanoma and non-small cell lung carcinoma and achieved some efficacy (Nikfarjam et al., 2020).

Traditional drugs have numerous disadvantages: poor water solubility, rapid in vivo clearance, poor biocompatibility, unsatisfactory in vivo distribution, and low permeability. New drug carriers are being continuously developed to optimize and improve the bioavailability to solve these problems. In recent years, exosomes have also been developed for drug loading, which can improve the stability of drugs and exhibit natural targeting ability based on donor cells. Because it is a nanomolecule with cell surface substances, it can readily and selectively penetrate biological barriers (Batrakova and Kim, 2015; Antimisiaris et al., 2018). At present, there are two main methods of exosomes transplantation in the treatment of SCI (Table 1), including intrathecal injection and tail vein injection. Moreover, a meta-analysis showed that intrathecal therapy seems to be more effective than tail vein injection therapy (Yi and Wang, 2021). Interestingly, a study showed that exosomes could also be used to treat SCI by intranasal injection. More specifically, intranasal exosomes led to significant locomotor recovery as compared to the intrathecal exosomes group (Shahzad et al., 2021). However, so far, there is no uniform standard for the isolation methods of exosomes. Exosomes are obtained using different isolation methods with different purities and specificities. At present, the most used isolation method for exosomes is ultracentrifugation, but the purity of obtained exosomes is low. Therefore, the isolation methods of exosomes still should be developed to obtain a higher purity and specificity. For questions regarding exosome preservation and transport, refer to the 2018 International Association for Extracellular Vesicles statement (Gardiner et al., 2016; Thery et al., 2018). When exosomes are used as drugs or carriers for treatments, the dosage, timing, and administration route are still not known. Therefore, it is essential to assess their half-life and in vivo distribution characteristics in advance (Smyth et al., 2015; Yi and Wang, 2021). Regarding the content and function of exosomes, as mentioned earlier, the results obtained from different cell sources and culture conditions are different, and the diversity in their therapeutic effects remains to be studied. Specific exosomes can be selected according to their studies. Moreover, more exosomes obtained from cell sources and culture conditions can be sought.

Currently, in the field of treatment of SCI, most studies on exosomal ncRNAs focused on miRNAs, lncRNAs, and circRNAs. In the future, more RNAs should be developed, such as rRNA, tRNA, and piRNA (Chandran et al., 2017; Jogia and Ruitenberg, 2020). For the targeted therapy of ncRNAs, the biological regulatory pathway is very complex, and a single targeted axis might play only a limited role. Therefore, to clarify the regulatory network, multicellular, multitarget, and multipathway validation is one of the directions to elucidate the specific regulatory mechanism for future research (Li X. et al., 2020; Wang W. Z. et al., 2021). Nevertheless, ncRNA-targeted therapy has many disadvantages. For example, it is easy to miss the target. Even it has a low transfection efficiency and a short half-life. What’s worse, it is difficult to overcome the limitations of blood-spinal cord barrier (Shahzad et al., 2021). Fortunately, targeting and half-life can be improved by the carrier delivery of drugs such as viruses, siRNAs, lipids, polyethylene glycol, and exosomes (Guo et al., 2019; Xu et al., 2019; Li L. et al., 2020; Segel et al., 2021). What’s more, as mentioned earlier, many studies have achieved some results using ncRNAs in combination with exosomes for the treatment of SCI. Because exosomes have the ability to penetrate the BSCB, they can deliver ncRNAs to the lesioned area of SCI to enhance efficacy (Ding et al., 2019). Moreover, exosomes have a lower risk of tumorigenicity, toxic effects, and autoimmune responses (Lai et al., 2013). However, the clinical application of exosomes remains problematic. It has been shown that the amount of exosomes produced by cellular secretion is small (Katsuda et al., 2013), and it is hard to meet clinical needs. Therefore, regulating exosomes release is particularly important. It is reported that the pH of the microenvironment has a role in the secretion of exosomes (Parolini et al., 2009). Regulating the pH of the microenvironment may result in increased exosomes production. Furthermore, most studies have focused on the mechanism of miRNAs in combination with exosomes for the treatment of SCI, and only a few studies have used other RNAs.

In summary, the treatment of SCI is still a huge concern, and no effective methods are available to promote neurological recovery. SCI is the result of multiple factors, hindering the development of rehabilitation because of its complexity. Therefore, understanding the pathomechanism will facilitate better treatment of SCI. As a mediator of intercellular communication, exosomes are advantageous in treating SCI. Exosomal ncRNAs have also been shown to contribute to nerve regeneration. Exosomes derived from different cells play very much the same role, and exosomal ncRNAs such as miRNAs, lncRNAs, and circRNAs have tremendous therapeutic potential in SCI. We must optimize and enrich exosomes and exosomal ncRNAs of various cellular sources and combine both of them effectively to improve their therapeutic efficacy in SCI. Then, more studies are needed to elucidate the specific mechanism of action of exosomes and exosomal ncRNAs from different cell sources in SCI. These searches provide a comprehensive theoretical basis for the clinical translation of exosomes and exosomal ncRNAs from different cell sources in SCI treatment and provide great hope for the clinical treatment of SCI.

All the listed authors directly, substantially, and intellectually contributed to the preparation of this manuscript and approved the publication. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.