The pathophysiology underlying SCI occurs in two phases, the primary and secondary phases of injury. The primary injury phase is caused by the initial mechanical forces delivered to the spinal cord at the time of injury (5, 6). These mechanical forces cause direct structural damage to the surrounding neuronal tissue and vasculature tissue resulting in acute cell dysfunction and cell death (5, 6, 13). Broadly, four types of primary injury mechanisms have been described: (a) impact plus persistent compression; (b) impact alone; (c) distraction; and (d) laceration/transection (5, 6). Among the four mechanisms of primary injury, impact plus persistent compression is most common and frequently occurs via burst fractures (5, 6). In contrast to the primary phase of injury which occurs within a short window of time, studies have shown that the secondary phase of injury continues for days to weeks or even months after SCI (12–14). The secondary phase of injury propagates the acute cell dysfunction and cell death initially caused by the primary injury. Several secondary injury mechanisms ( Figure 1 ) have been shown to contribute to propagation of acute cell dysfunction and cell death including neuroinflammation, ischemia, free radical formation, lipid peroxidation, blood CNS barrier break down, edema, release of proteases, and excitotoxicity (12–14). Importantly, the immune response plays a profound role in the propagation of secondary injury after SCI.

Following SCI, three zones that differ in tissue quality form including: (1) Zone 1, which is mainly a product of the initial trauma and is defined by regions of necrosis, inflammation, and cysts; (2) Zone 2, is characterized by regions of incomplete injury that are still accompanied by immune cell infiltration, axonal swelling, and Wallerian degeneration; and (3) Zone 3, which are histologically intact areas (15, 16). Among the zones, Zone 1 is the region that has the highest level of damage with the lowest chance of neuron survival/recovery (12, 15, 16). Neuroinflammation has been shown to be a critical secondary injury mechanism that directly alters the progression of lesions and cell function across all three zones (12). Although surgical decompression is the first line therapy for spinal cord compression following SCI, studies have shown that the severity of injury is dependent on the extent of secondary damage (10, 12, 17). Therefore, therapeutic strategies have focused on strategies to modulate the immune response after SCI in order to promote neuro-regeneration and neuroprotection. This review will focus on examining the immune response underlying the pathophysiology of SCI and assesses both current and future neuro-immunological therapies.

After SCI, both immune depression and autoimmunity commonly co-occur (18, 19). The CNS exerts control over the immune system through several pathways including the hardwired fibers of the autonomic nervous system (19, 20). Throughout the body, both central and peripheral autonomic nervous system sensors function to relay information about the status of the body’s immune system (19, 20). Due to SCI, the CNS interaction with the immune system is disrupted resulting in a significant systemic decrease in immune function with reports of several different functional alterations in macrophages, T and B cells, and natural killer (NK) cells (20, 21). This systemic decrease in immune function after SCI is termed spinal cord injury-induced immune deficiency syndrome (SCI-IDS) (12, 21). Although the precise mechanism(s) causing SCI-IDS are not completely clear, patients with SCI-IDS are at a significantly higher risk of developing complications such as infections, pneumonia, and urinary tract infections (12, 21).

One recent study suggested that the mechanism underlying post-SCI immune suppression may be due to the fact that after SCI, significant plasticity develops below the injury site within the autonomic spinal circuitry resulting in a sympathetic anti-inflammatory reflex (22). Furthermore, the authors showed that chemogenetic silencing of this reflex prevents post-SCI immune suppression and could be a promising therapeutic approach in the future (22). Although there are several negative effects of SCI-IDS, immunosuppression following SCI may actually be advantageous and serve a protective function (12, 23). SCI damages the blood-spinal cord barrier and spine parenchyma, which exposes cells of the adaptive immune system including B and T lymphocytes to CNS antigens (12, 23). Exposure of the adaptive immune system to CNS antigens can then trigger autoimmune responses and the production of autoantibodies that have been shown to worsen pathology within the spinal cord (12, 23). Notably, studies in mice suggest that SCI especially impairs the production of new antibody responses but preserves existing immunity (18, 24). Thus, autoantibodies identified after SCI may have existed prior to the injury (18). This hypothesis was supported by a study that demonstrated a set of naturally occurring auto-antibodies expanded following SCI (18). Although it is clear that trauma induced autoimmunity has the potential to contribute pathologically, the triggering mechanisms and molecular signatures will require further investigation in the future.

Following CNS injury, neuroinflammation is a result of innate immune system activation that is mediated by cytokines and chemokines released by astrocytes, resident microglia, endothelial cells, and peripherally derived immune cells (14). Several studies have clearly demonstrated that within hours of CNS injury, the inflammatory cytokines IL-1, IL-6, and TNF are all upregulated (7, 25). Upregulation of these potent inflammatory cytokines subsequently triggers substantial infiltration of macrophages, microglia, and neutrophils (7). Furthermore, these infiltrating immune cells continue to produce and secrete additional chemokines and cytokines that modulate the immune response. The size of the primary insult is a critical factor in determining the magnitude of neuroinflammation (26). In the case of SCI, additional cell death is often caused by overactivation of the immune response. Therefore, in order to develop therapies that can selectively dampen the overactive immune response, the immune response following SCI must be thoroughly characterized. In this review, we briefly review the literature on the systemically upregulated cytokines derived from the blood/cerebrospinal fluid (CSF), which may have clinical utility as biomarkers and focus on the timeline of the local cytokine/chemokine cascade and cell infiltration patterns ( Figure 2 ) (32).

Overall, the patterns of cytokine and chemokine upregulation following SCI are generally similar in the serum and CSF compared to the patterns within the injured tissue itself. Understanding the tempo and the major players of the immune response in SCI could allow for the identification of biomarkers from patient’s serum and CSF that could guide management of SCI. Within 6 hours after SCI, a rat model showed that CSF levels of IL-1, IL-10, IL-17a, IFN-γ, and TNFα were all significantly elevated (7, 33). Furthermore, the serum levels of IL-1β, IL-6, and TNFα, remained elevated throughout the first week after SCI in rats (7, 34, 35). A clinical study examining changes in serum biomarkers found that within 24 hours after SCI, there were significant elevations in structural proteins including tau, S100β, and GFAP as well as elevations in the cytokines IL-6, IL-8, and MCP-1 (7, 36). One study reported decreased serum levels of IL-1β and IL-10 after 14 days following SCI while another study found persistently elevated IL-1β and no change in IL-10 28 days after SCI (7, 36). Notably, this study (36) also reported persistent serum elevations in TNFα 28 days after SCI and downregulation of IL-4 at 28 days. Importantly, clinical studies have revealed that lower levels of serum and CSF IL-6, IL-8, tau, S100b, GFAP, and MCP-1, are associated with significantly higher AIS grade improvement and neurological recovery compared to patients that did not improve (37, 38). In these studies, the most severely injured patients typically had the highest levels of NSE tau, S100b, and GFAP (37, 38). In the future, additional clinical studies will be needed to further validate these proteins as true biomarkers that could enhance clinical decision-making.

Below, we discuss the timeline of the inflammatory cascade following SCI including cytokine/chemokine regulation and cell infiltration patterns ( Figure 2 ).

Within minutes after the primary SCI, secondary injury mechanisms begin to cause significant damage (7, 47, 48). More specifically, edema, oxidative damage, cell permeabilization, excitotoxicity (glutamate), ischemic injury due to microvascular supply destruction, and proapoptotic signaling among other factors all contribute to substantial cell dysfunction and cell death (7, 47, 48). Dysregulation in intracellular calcium also causes injury and further cell death as dysregulated intracellular calcium results in calpain activation causing mitochondrial dysfunction (7, 48–50).

Although in the early phases of injury the resident immune cells of the CNS, microglia, respond in a protective manner, quickly microglia morph into proinflammatory cells that secrete cytokines that trigger peripheral immune cell infiltration (27, 51–53). For example, one study showed that microglia and astrocytes start expressing TNFα mRNA and IL-1β within a half hour following SCI (25). Notably, the number of TNFα positive cells peaked 1 hour following SCI (25).

Within 1-3 hours following tSCI, TNFα, IL-1β, and IL-6 are believed to be the key mediators. mRNA colocalization studies have demonstrated that microglia, neurons, and astrocytes synthesize TNFα, IL-1β, and IL-6 during the early hours after SCI in a mouse model (7, 25). In five other studies, the levels of TNFα in the early hours after SCI were also substantially increased (39–41, 54, 55). Three of these studies showed a significantly elevated level of IL-1β during the early hours of SCI; however, two studies using enzyme-linked immunosorbent assays (ELISAs) reported that there was not an increase in IL-1β during the early hours of injury (34, 39–41, 54). Hellebrand et al. (7) suggested that this inconsistency may be due to the time necessary to synthesize the full protein product as well as the time required for caspase 1 to proteolytically process it to its active form. Studies assessing IL-6 upregulation observed similar results to the IL-1β pattern (39–41, 54).

Although there appear to be a few differences in the adaptive immune system among rodents, a majority of literature has shown a significant increase in proinflammatory cytokines (7, 25, 40, 56–59). More specifically, within the window of 3-6 hours after SCI, several studies have shown that TNFα, IL-1β, and IL-6 are the primary cytokines significantly elevated (7, 25, 40, 56–59). In addition, most studies agree that within the 3-6 hours window there is a slight lag in upregulation of IL-1β and IL-6 relative to TNFα and IL-1a (7, 25, 40, 41, 57–59). The increase in cytokine production in microglia and astrocytes also results in increased peripheral immune cell invasion. For instance, neutrophils start to arrive within 4-6 hours after SCI and function to prepare the region for repair through the production of proteolytic and oxidative enzymes (28, 29). The overwhelming number of neutrophils, however, can cause significant tissue damage (28, 29).

At 3 hours following SCI, the number of cells expressing TNFα mRNA within the center of the injury is significantly increased relative to control cells; however, the number of cells expressing TNFα mRNA was 66% lower compared to the 1 hour time point (25). Additional evidence suggests that TNFα levels return closer to baseline following rapid onset as two studies reported no significant increase at 4 hours following SCI (40, 54).

Even though anti-inflammatory cytokines could decrease the levels of proinflammatory cytokines, studies suggest that following SCI, the levels of anti-inflammatory cytokines are either low or absent (7). Theoretically, IL-10 is an anti-inflammatory cytokine that could reduce pro-inflammatory cytokine levels and IL-4/IL-13 are anti-inflammatory cytokines that can trigger alternative macrophage activation (7, 60, 61). Studies using rodent models have shown that IL-13 was increased at 4 hours following SCI in rats but not in mice, and IL-10 was increased in mice but not in rats (40, 57, 58, 60–64). Finally, Xiong et al (65) reported increases in 4-hydroxynonenal (4-HNE), which is formed during lipid peroxidation, starting at 3 hours following SCI (7, 65).

Within the 6-12 hour time period, the proinflammatory environment continues facilitating increased microglia proliferation/recruitment and increased peripheral immune cell invasion (7). Studies agree that from 6-12 hours after SCI, TNFα, IL-1β, and IL-6 continue to be upregulated while IL-4, IL-10, and IL-13 remain at baseline levels (25, 33, 39–45). In addition, other immune mediators that are significantly increased during the 6-12 hour window include macrophage inflammatory protein 1-alpha (MIP-1a), monocyte chemoattractant protein 1 (MCP-1), C–X–C motif chemokine ligand 1 (CXCL1), growth-related oncogene (GRO), and RANTES (Regulated upon Activation, Normal T Cell Expressed and Presumably Secreted) (33, 43, 46).

Initially, neutrophils are the most common infiltrating cell type (30). Overactivation of neutrophils can be quite destructive to tissue as these cells release substantial amounts of neurotoxic substances including chemokines, enzymes, reactive oxygen species (ROS), and reactive nitrogen species (RNS) (9, 29, 66, 67). At approximately 8 hours after SCI, apoptosis from the inflammatory response peaks in neuronal cells (68, 69).

During the 12-24 hour window a majority of studies show: (a) upregulated levels of the proinflammatory cytokines TNFα, IL-1β, and IL-6; (b) elevated levels of IL-7, GRO, MIP-1α, and MCP-1; and (c) baseline levels of TGF-β1, IL-4, IL-10, and IL-13 (39, 45, 46, 56, 70–72). At 24 hours following SCI, the level of neutrophils peak ( Figure 2 ) (31). Neutrophils infiltrating the site of injury function to clear debris, proteases release reactive oxygen species, and neutrophils secrete myeloperoxidase, elastase, and proteases (31). In glial cells, apoptosis peaks at 24 hours following SCI (68, 69).

One study in rats showed that lymphocytes start to collect near blood vessels within gray matter as early as 6 hours after SCI (52). Lymphocytes that infiltrate the site of injury produce IL-1β, IL-6, TNFα, and LIF at 12 hours post injury (25). The levels of the neurotoxic 4-HNE peak at 24 hours and for two weeks remained elevated (65).

In a rat model of SCI, flow cytometry showed that the initial phase of cellular inflammation at 1 day after SCI consisted of an early neutrophil peak; at 7 days after SCI a peak of macrophages/microglia; and at 9 days after SCI a peak of T-cells (73). In mice, although neutrophils enter the site of injury at 6 hours after SCI, levels do not peak until 14 days after injury and remain for up to two weeks (27). At day 5 after SCI, phagocytic macrophages are most commonly localized in areas of necrosis while microglia are mainly at the margins (7, 74). Microglia are highly dynamic and during the first week after SCI have been shown to proliferate extensively (7, 74). Furthermore, at the border of the site of injury between infiltrating peripherally-derived macrophages and astrocytes, microglia form a dense cellular interface (7, 74). Due to the initial injury, axons start to retract at day 2 and macrophages also trigger a late phase of axon retraction (75–77).

After 2 days following SCI, co-localization studies showed that TNFα mRNA levels had returned to a level that was not significantly different from the level of mice that had received control laminectomies (7, 25). The average number of cells expressing IL-6 mRNA decreased from 24 hours-4 days after SCI to an almost zero cells 7 days after injury (7, 25). Similarly, the number of IL-1β positive cells decreased at 2, 4, and 7 days following SCI (7, 25). Notably, although the number of cells expressing the above transcripts decreased, several studies in both rats and mice have reported increased levels of TNFα, IL-1β, and IL-6 at 1, 3, and 7 days following SCI (7). In addition, GRO has been reported to be significantly increased at 1, 3, and 7 days following SCI (46, 71, 78).

During the first week of injury, studies have shown that there is a significant increase in production of chemokines that recruit T-cells, monocytes, and dendritic cells to the lesion (7). Previous studies have shown upregulation of MIP-1α at 1 and 3 days following SCI, CXCL1 expression is increased at 3 and 7 days following SCI, and RANTES is significantly upregulated at 3 days following SCI (43, 71, 78–81). While there have been some conflicting results, a majority of studies show that MCP-1 mRNA is increased in mice 1 day after SCI with return to baseline 7 days following SCI (82, 83). Among rats and mice, it is important to note that there have been some contrasting behaviors (7). Cytokines that have been reported to have at least somewhat contrasting behaviors among rats versus mice include IL-2, IL-4, IL-5, IL-13, and IFN-γ (7, 70, 80, 81, 84). A majority of studies have suggested that IL-10 is not significantly increased until 3 days following SCI although some reports have not observed changes in IL-10 levels after SCI in both mice and rats (40, 42, 79). Finally, TGF-β1 was reported to be upregulated at 3 and 7 days following SCI (7, 39).

Macrophages and peripherally derived macrophages peak at 7 days following SCI and persist for months after the injury in mice, rats, and humans (15, 60, 73, 74, 85). At 14 days following SCI injury, Mukhumedshina et al. (78) found that IL-1a, IL-2, MIP-1a, and GRO were significantly increased in rats while IL-2, IL-5, IL-13, IL-17a, IL-18, and GM-CSF were significantly decreased (78). In mice on the other hand, IL-1a, IL-4, IL-7, IL-12, IL-15, MIP-1a, MCP-1, and RANTES were significantly upregulated (79, 84, 86). Mice also were found to have a second surge of cells expressing IL-1β and TNFα at 14 days after SCI (25). Rats, however, did not undergo a second surge at 14 days after SCI (78, 87). Future studies examining the cytokines profiles after 14 days in SCI are necessary as there is less data compared to other time frames.

Astrocytes are a subtype of CNS glial cells that function to maintain neurons and the blood spinal cord barrier (7, 88). Initially, reactive astrocytes flock to the injury site to aid in tissue repair but eventually become scar forming astrocytes (7, 88). These scar forming astrocytes form a glial scar surrounding the injury site. The layer of astrocytes around the lesion are defined by increased expression of glial fibrillary acidic protein (GFAP)/intermediate filaments, cellular hypertrophy, and process extension (7, 88). Overall, the glial scar has immune cells in the center with macrophages interacting with pericytes, which are the main scar connective tissue source around the edges and astrocytes that surround the periphery (74, 89).

The question of whether T cells mediate wound healing or result in secondary degeneration is controversial (7, 15, 90–92). T cells mediate adaptive immunity and appear to infiltrate into the spinal cord at somewhat different times depending on the species of animal (7). Studies have shown that chemokines CXCL10 and RANTES are critical for the proliferation and cytokine production of T cells (7, 93). Recent evidence suggests that T cells appear to play more of a destructive role after SCI as Gonzalez et al. (93) demonstrated that tissue preservation and functional recovery were improved when T cell infiltration was decreased by neutralizing the chemoattractant CXCL10. In a rat model, cytotoxic T-cells were predominant compared to regulatory T cells (>90% versus 10%, respectively) further suggesting a negative role of T cells (94). In addition, secondary injury was aggravated after SCI by T cell perforin as perforin destroyed the blood spinal cord barrier triggering infiltration of inflammatory cytokines (95).

To date, we do not fully understand the clinical utility and efficacy of several immunotherapies and pharmacological agents following tSCI. These immunotherapies have the potential to improve outcomes in tSCI patients by modulating the immune response to injury and promoting regeneration, however, further preclinical and clinical studies are needed. Although our current understanding of the precise immunological mechanisms by which some of these therapies may potentially improve functional outcomes is limited, here we provide the state of the field and future directions of pharmacological therapies in tSCI.

The clinical utility of methylprednisolone in the treatment of traumatic SCI is controversial (48, 96–98). Theoretically, corticosteroids could prevent secondary damage after SCI and be beneficial by preserving the ultrastructure of the spinal cord by decreasing the injury induced, decreasing free radical catalyzed lipid peroxidation, enhancing impulse conduction and neuronal excitability, improving blood flow, reducing oxidative stress, and modulating the immune response (99, 100). Three landmark multicenter, double-blinded, randomized controlled trials, National Spinal Cord Injury Study I (NASCIS-I), NASCIS-II, and NASCIS-III were performed in 1984, 1990, and 1997 (96, 101, 102). NASCIS-I sought to assess the methylprednisolone dosage following SCI, but there was no placebo group in this trial due to the belief that methylprednisolone was beneficial, and it would be unethical to withhold methylprednisolone therapy (101). Significant debate has been centered around the results of the NASCIS-II which randomized 487 patients with acute traumatic SCI to receive an initial bolus of 30 mg/kg of methylprednisolone followed by an infusion of 5.4 mg/kg per hour for 23 hours or treatment with placebo or naloxone (48, 96–98). Although there was no significant benefit in the neurological outcomes among the 162 patients receiving methylprednisolone within 12 hours, analysis of a group of 65 of these patients that received methylprednisolone within 8 hours after SCI significantly showed improved neurological function at 6 months (96, 103). Therefore, the authors concluded that treatment with methylprednisolone within 8 hours improved recovery of neurological function after SCI (96). Advocates of this study tend to highlight the lack of high-quality evidence currently available within the literature while critics point to the potential arbitrary 8-hour time point, the magnitude of treatment effects, and the potent issues with losses to follow-up (98, 103–108). A prospective randomized clinical trial demonstrated no neurological benefits (109). Several class III medical evidence studies have demonstrated the beneficial effects of methylprednisolone therapy in SCI (96, 101, 102, 110, 111). Critics of these studies, however, suggest that these studies are limited by small sample sizes derived retrospectively from larger study populations and/or incomplete data sets (112–115). Another potential concern of using methylprednisolone in the treatment of acute SCI is the potential for harmful side effects (96, 101, 102, 110, 111, 116). More specifically, three Class I studies reported wound infection, hyperglycemia requiring insulin administration, and GI hemorrhage as statistically significant side effects (96, 101, 102, 110, 111, 117, 118).

Although the clinical utility of these trials remains controversial, the methylprednisolone treatment protocols outlined in NASCIS II and III are still used today (96, 102). Early critics of the NASCIS advocated that methylprednisolone therapy be considered “a treatment option” rather than “standard of care” (96, 97, 101, 102). In 2013, the routine administration of methylprednisolone in the management of acute traumatic SCI was not recommended by “Guidelines for the Management of Acute Cervical Spine and Spinal Cord Injuries” (97, 119–121). The use of methylprednisolone among surgeons has decreased based on studies surveying the practice habits of surgeons, a 2018 study suggested that a majority of surgeons worldwide (52.9%) still use methylprednisolone in the treatment of acute SCI (122–127). In the future, further studies assessing the effect of steroid therapy on the neurological outcomes of SCI patients will be necessary in order to determine its clinical utility.

The use of non-steroidal anti-inflammatory drugs (NSAIDs) following acute SCI injury appears promising as NSAIDs have been suggested to decrease spinal cord inflammation and edema, improve motor function with minimal side effects, and increase axonal sprouting (128). The effectiveness of NSAIDs in acute SCI is likely due to their ability to inhibit Rho-A (128). In animal models after traumatic SCI, two non-steroidal anti-inflammatory drugs that can be utilized in order to maintain spinal cord blood flow are ibuprofen and meclofenamate (129, 130). Similar results have been reported with the use of both a prostacyclin analogue and a thromboxane inhibitor (131, 132). Other studies have suggested that COX-2 expression is significantly increased within the damaged rat spinal cord tissue after contusion injury, and inhibition of COX-2 enhances functional outcomes (129, 133). Small sample size and heterogeneity creates problems with studies, and the therapeutic effects seen in animals may not be the same in patients (128, 131). Given the potential for improved outcomes and the low risk for adverse side effects, well-designed prospective studies evaluating ibuprofen and indomethacin are needed in the future.

Minocycline is a chemically modified second generation tetracycline antibiotic that has the ability to penetrate the blood brain barrier (134, 135). Minocycline has been shown to provide neuroprotective effects and has been shown to have anti-inflammatory, antioxidant, and anti-apoptotic properties (131, 134, 136). In animal models of SCI, TBI, and cerebral ischemia, minocycline has been shown to protect white matter structures, motor neurons, and oligodendrocytes (137, 138). Studies suggest that minocycline following SCI had significant anti-inflammatory and anti-apoptotic effects by preventing the activation of TNFα, IL-1β, COX-2, and MMPs decreasing caspase-1 and caspase-3 levels (131, 139, 140). Minocycline has also been reported to alleviate excitotoxicity and has been suggested to inhibit the p38 mitogen activated protein kinase pathway in microglial cells (137, 141, 142).

Tissue sparing, reduction of the lesion, inhibiting oligodendroglia and neuronal apoptosis, reducing inflammation, and enhancing neurological and histological outcomes could possibly be achieved with minocycline treatment (131, 143). Studies in mice have suggested that minocycline therapy is associated with substantial improvements in long term functional outcomes in which administration of minocycline over a 4 week recovery time course increased Beattie, Basso, and Bresnahan scores at 20 days following injury (144). A 2008 study in rats is not in agreement with other preclinical and clinical studies such as Pinzon et al. (143, 145) which found that minocycline therapy did not generate tissue sparing or improved behavioral outcomes (131, 143). In clinical studies, minocycline treatment improves outcomes in acute incomplete cervical SCI patients (131, 135). A randomized, double bind, placebo-controlled, single center phase II clinical trial examined treatment with IV minocycline (n=27) for 7 days versus placebo (n=25) treatment in patients with SCI. This study showed that minocycline therapy was safe and feasible. In addition, there was a trend toward improvements in motor scores especially in tetraplegic patients. Currently, a multicenter phase III randomized controlled trial (NCT01828203) is assessing the impact minocycline therapy versus placebo therapy on acute (<12 hours) SCI patient recovery at 3 months and one year after injury. Minocycline therapy, therefore, appears to be promising for SCI, as it works via several mechanisms; however, more well-designed trials are necessary to definitively conclude the role of minocycline in SCI management.

The axons of patients with SCI do not regenerate appropriately (74, 89, 131, 146). Glial reaction occurs at the site of injury, resulting in formation of a glial scar (74, 89, 131, 146). Furthermore, microglia, oligodendrocytes, and their precursors (which include astrocytes, meningeal cells, and myelin fragments) are recruited due to the glial reaction (14, 69, 89, 131, 146). Many of these cells release factors that halt axonal regeneration (74, 89, 131, 146). Chondroitin sulfate proteoglycans (CSPG) are one major class of axonal growth inhibitors that may play a key role in regeneration failure (131, 146–148). Importantly, chondroitinase ABC (ChABC) is an enzyme that could be critical in overcoming CSPG mediated inbition as ChABC digests glycosaminoglycans on CSPGs (146–149).

Animal studies have demonstrated that treatment with ChABC after SCI can dramatically improve functional recovery and regeneration of both sensory axons and the corticospinal cord (131, 150, 151). Mechanistically, ChABC most likely results in growth promoting effects through causing increased germination of spare axons, formation of new synaptic connections beneath the sites of injury, and removing perineuronal nets (152). In addition to these mechanisms, ChABC also has been shown to have immunoregulatory activity (153). More specifically, Didangelos et al. showed that ChABC increased expression of the anti-inflammatory cytokine IL-10 and reduced expression of the pro-inflammatory cytokine IL-12B (153).

In rat models, administration of ChABC promoted both improved recovery of proprioceptive and locomotor behaviors (151). In addition, below the lesion treatment with ChABC resulted in restoration of postsynaptic activity following electrical stimulation of corticospinal neurons (151). The benefits of ChABC for SCI have been clearly demonstrated in rodent models of: (a) SCI, (b) stroke, and (c) nigrostriatal injury as well as in cats with SCI (131, 154–156). A recent study in dogs with severe chronic SCI strongly supported a beneficial effect of intraspinal injection of chABC (157). Importantly, there was no evidence of long-term adverse effects with this therapy (157). Given the impressive improvements of functional outcomes with ChABC treatment and the lack of reported harmful effects, human SCI ChABC trials should be conducted in the future.

Glycosphingolipids exert several critical functions in cellular differentiation, interaction and immune response (158). Theoretically, the benefit of glycosphingolipid GM-1 stems from the ability of GM-1 to stimulate tyrosine kinase receptors, resulting in improved regeneration and enhanced neuronal plasticity (100, 131). A randomized, double blind, prospective phase II trial demonstrated significant improvement of 1-year ASIA motor scores in patients that were treated with daily GM-1 ganglioside for 18-32 days following SCI (135, 159). A randomized, double-blind, sequential multicenter phase III clinical trial examined the effect of two doses of Sygen (monosialotetrahexosylganglioside GM1 sodium salt) compared to placebo in 797 patients with acute SCI (160). Although there appeared to be accelerated bladder, bowel, and motor recovery following Sygen treatment within the first three months, ultimately there was no significant effects after the study ended (135, 160). The time of administration of GM-1 ganglioside has been postulated to be a key factor in the observed variation among the two trials (135). Furthermore, a Cochrane Database Systematic Review described significant weaknesses in data collection and presentation methodologies in these trials (135, 161). This review concluded that there is no available evidence to support treatment with gangliosides to reduce the death rate or improve the quality of life in SCI and to date no follow-up studies have been performed (161).

Neuroimmunophilin ligands are a class of compounds that have great potential for the treatment for SCI. To date, however, the results of neuroimmunophilin ligand therapy for SCI have been conflicting. Cyclosporin A and FK-506, which act as neuroimmunophilin ligands, have shown protective properties in ischemia, neurodegenerative disorders, and in trauma (131, 162, 163). In neuronal tissues, NIL-A, which is currently in Phase II clinical trials for SCI, binds with FK506 binding protein (FKBP)-12 (131, 164). Neurophilin ligand V10367 binds to FK-506 binding protein and was investigated in a preclinical model (131, 165). V10367 resulted in increased neuroregeneration in the CNS and peripheral nervous system (165). Tacrolimus in addition to its immunosuppressive effects also has neurotrophic activity and has been shown to enhance nerve regeneration in peripheral nerve injury (166, 167). Due to the potential of neuroimmunophilin ligands in the treatment of nerve injury, further preclinical studies are needed to clarify their potential role in the treatment of tSCI.

Nogo-A is neuritite growth inhibitory myelin protein which plays a key role in inhibition of neurite outgrowth and limits neurological recovery (146, 168–170). Treatment with anti-Nogo-A antibody after SCI in rat models has been shown to significantly improve recovery of locomotor training and results in superior spinal cord reorganization and regeneration (171, 172). A macaque model showed similar results (131, 173). Another study examined the therapeutic outcomes of anti-Nogo-A antibody therapy on the corticospinal tract after acute, 1 week, or 2 week delayed intrathecal anti-Nogo-A antibody infusions (131, 174). In this study, Gonzenbach et al. (174) find that treatment of SCI with anti-Nogo antibodies in rodents is limited to less than 2 weeks. A multicenter Phase I open-label cohort study assessed intrathecal administration of the human anti-Nogo-A antibody ATI355 in acute, complete traumatic paraplegia and tetraplegia (175). Administration of anti-Nogo antibodies was well tolerated and in 7 of 19 patients with tetraplegia conversion from complete to incomplete SCI injury occurred (175).

The Rho signaling pathway plays a key role in neuronal growth inhibition as well as regulating the cytoskeleton and motility (176). C3 transferase, an enzyme from Clostridium botulinum, inhibits Rho signaling by locking RhoA in an inactive state (176). In animal models of SCI, local application of C3 transferase induced long distance cortico-spinal regeneration and improved functional outcomes (176, 177). C3 transferase therapy with wild-type C3 transferse was limited, however, by its very low levels of cell penetration (177, 178). Thus, a recombinant version that readily crosses the dura was developed BA-210 (Cethrin) (177). In addition to promoting axonal regeneration and neuroprotection, Cethrin also appears to modify the adverse immune reaction following SCI (177–179). A multicenter phase I/IIa clinical trial involving cervical and thoracic SCI patients taking 0.3 mg to 9 mg BA-210 (later VX-210) showed that BA-210 treatment significantly improved motor recovery (178, 179). More specifically, the largest changes in motor recovery were observed in cervical injury patients taking 3 mg BA-210 in which 31% of cervical injury patients, 66% of patients taking 3mg BA-210,and 6% of thoracic injury patients converted from ASIA A to AISA C or D (178, 179). A phase Iib/III study assessing whether application of VX-210 to the overlying dura would improve motor outcomes following cervical SCI was stopped prematurely as the primary efficacy end-point was not met (180).

The CD11/CD18 family of integrins are localized on the surface of leukocytes and have been shown to bind both human intercellular adhesion molecule-3 (ICAM-3) and human/rat vascular cell adhesion molecule-1 (VCAM-1) (181–184). Following injury, damaged endothelial cells upregulate VCAM-1 and ICAM-3 which facilitates leukocyte binding and migration to the site of injury (184, 185). Preclinical studies have demonstrated the promising potential of a monoclonal antibody, 217L, which binds to the CD11d subunit of the CD11/CD18 integrins (184). Initial preclinical studies in rats demonstrated that anti-CD11d therapy significantly decreased number of both neutrophils and macrophages within the cord following SCI (186, 187). Furthermore, anti-CD11d therapy substantially decreased secondary damage and resulted in superior autonomic and locomotor recovery as well as decreased neuropathic pain (188–191). Similarly, other studies have also suggested that decreasing the number of neutrophils is neuroprotective (29, 46, 85, 192–194). A CD11d monoclonal antibody study in mice, however, reported that depletion of neutrophils after SCI through the use of Ly-6G (RB6-8C5) mAb impeded rather than enhanced functional recovery (195). A follow-up study examining the therapeutic effectiveness of anti-CD11d treatment in mice demonstrated that neutrophil infiltration was reduced by 61% at 72 hours after SCI and significantly improved functional outcomes and decreased secondary damage (184).

In addition to the role of anti-CD11d therapy in decreasing the infiltration of neutrophils, studies have suggested that anti-CD11d therapy improves neurological outcomes, provides neuroprotective effects decreases pain, significantly reduces histopathological damage, and improves intraspinal serotonergic innervation patterns (190, 196). A recent study also suggested that anti-CD11d therapy decreases secondary damage by significantly reducing free radical formation (197). Given the promising results from preclinical studies, future clinical trials assessing the efficacy of anti-CD11d in humans are needed. Importantly, the mechanism of injury and injury severity may be key determinants of the efficacy of anti-CD11d therapy following SCI (198). More specifically, using a rat model Geremia et al. (198) suggested that anti-CD11d therapy is most efficacious in cases of modest injury severity with minimal non-penetrating injury and frank hemorrhage into the spinal cord (198). Thus, clinical studies assessing efficacy of anti-CD11d therapy with careful consideration of the injury severity and mechanism of injury are necessary in the future.

Following SCI, a robust B cell response results in the production of pathogenic antibodies that impedes neurological recovery (199, 200). The role of B cells in the pathogenesis of SCI is further supported by the fact that functional outcomes are improved in B cell deficient RAG knockout mice following SCI (201). Based on this principle, selective B-cell depletion may be neuroprotective following SCI (201, 202). Casili et al. (202) assessed the effects of antibody mediated B cell depletion in mice using a glycoengineered anti-muCD20 antibody (18B12). Casili et al. (202) demonstrated that antibody mediated B cell depletion in mice improved functional recovery, slowed neuronal death and hindlimb motor dysfunction, and substantially inhibited the NFkB-dependent production of pro-inflammatory molecules. In the future, clinical studies are necessary in order to assess efficacy of anti-CD20 antibodies including rituximab or obinutuzumab after SCI (202, 203).

After SCI, T cells can directly or indirectly effect neurons, glia, or other CNS cells through the production of tumor necrosis factor (TNF)-α, interleukins (ILs), or other proinflammatory cytokines that play key roles in cytotoxic cell damage (92, 202, 204). Studies have demonstrated that therapeutic strategies targeting T cell infiltration and trafficking are neuroprotective (205). More specifically, antagonism or neutralizing antibodies against CXCL10, a known T cell recruiter, has been shown to improve functional outcomes, decrease T cell infiltration, decrease neuronal death, and significantly increase regeneration of axons (8, 206, 207). Another promising immunosuppressive therapy following SCI, involves the use of fingolimod, which targets sphingosine 1-phosphate receptor (S1P1), induces internalization of S1P1, decreases the number of circulating lymphocytes, and prevents trafficking of lymphocytes into tissues by sequestering lymphocytes within lymph nodes (208, 209). Locally administered fingolimod has been shown to result in superior functional recovery, decreases reactive gliosis, and prevents neuronal cell death (203, 210). Systemic treatment of fingolimod after SCI has been shown to prevent neuroinflammation and significantly improve both bladder and motor function (211, 212). Given the promising results of T cell targeting therapies, further clinical studies are necessary in order to assess the efficacy of CXCL10 antagonisms and fingolimod after SCI.

Previously, it has been shown that the macrophage immune response to injury was both delayed and blunted in the CNS of mature rodents compared to the more regenerative peripheral nervous system (213–215). Rapalino et al. (216) showed that local injection of homologous macrophages activated ex vivo through incubation with homologous peripheral nerves triggered partial motor recovery in rats with complete spinal cord transection (216, 217). Incubation of macrophages with autologous skin has been shown to result in equivalent effects and is thought to produce an “alternatively activated wound-healing phenotype” (213, 218). More specifically, in rats with a contused spinal cord, injection of macrophages activated with skin resulted in decreased spinal cyst formation and enhanced motor recovery (213, 218). This study also demonstrated that the skin activated macrophages had increased levels of cell surface markers that are characteristic of antigen presenting cells and macrophage secretion of brain-derived neurotrophic factor and interleukin-1b (213, 218). Autologous macrophage therapy has exciting potential due to the key functions macrophages play including clearing tissue debris from the site of injury, ability to modulate the immune system and influence spinal cord neurons, glial cells, and immune cells (217, 218). Safety studies in animals reported no short-or long-term adverse effects or toxicity in animals treated with macrophages to the Food and Drug Administration under Investigational New drug application No. 8427 (217).

Based on preclinical studies, an open-label phase I clinical trial assessing the safety and tolerability of treatment with incubated autologous macrophages in acute complete SCI patients was conducted (217). In this trial 3 of 8 subjects demonstrated conversion from AIS A to AIS C and overall the therapy was well tolerated (217). A single-blinded primary outcome randomized controlled trial involving six treatment centers in the United States and Israel, however, did not demonstrate a significant difference in the primary outcome among the autologous incubated macrophage treatment group (n=26) and the control group (n=16) (213). In the future, additional studies and further protocol optimization will be necessary to assess the efficacy of treatment of SCI with autologous incubated macrophages.

Hepatocyte growth factor (HGF) binds to c-Met and plays key roles in protection, regeneration, homeostasis, and exerts anti-inflammatory effects in several organs (219). Studies in rodent models have suggested that treatment with HGF results in the neuroprotective effects, enhanced angiogenesis, reduction of the lesion size, improved neuronal survival, and a decrease in production of oligodendrocytes (131, 220, 221). Furthermore, injection of HGF expression vector into rat spinal cords resulted in higher rates of survival, regeneration of axons and oligodendrocytes, and increased angiogenesis (131, 222). In a primate model of cervical SCI, significant improvements in hand dexterity were seen after HGF treatment (221, 223). Another study showed that exogenous treatment of HGF decreased astrocyte activation, decreased glial scar formation, decreased leukocyte infiltration, and produced anti-inflammatory effects (222). A recent double blind, randomized phase I/II clinical trial demonstrated the safety of treatment with recombinant human HGF and a larger phase III trial is necessary in the future in order to assess efficacy (224).

Neurotrophic factors are molecules that control the growth, survival, proliferation, and differentiation of neurons (225). In addition, neurotrophic factors play key roles in modulating immune responses and autoimmunity (226). Neurotrophic factors are often delivered with the use of collagenases (152). Delivering neurotrophic factors via collagenases has been reported to be more effective compared to delivery via direct injection, growth factor-saturated gel, or continuous injection (131, 152, 227). Ex vivo therapy can also be used when a patient’s cells are removed, genetically modified to synthesize specific neurotrophic factors, expanded in culture, and retransplanted into the patient (131, 152). The benefit of this route is that it provides localized, high-dose growth factors on delivery; however, initial studies suggest that ex vivo therapy is not ideal at stimulating distal axonal growth following initial axon growth (227).

Fibroblast growth factors are a family of growth factors that are present in both the central and peripheral nervous system that play key roles in determining cell fate, differentiation, migration, and display neuroprotective effects (228–230). In addition to their neuroprotective effects, FGFs also may have some immunomodulatory effects as one approach utilizing acidic FGF (aFGF) combined with peripheral nerve grafts in rats following transection SCI increased levels of IL-4, IL-10, and IL-13 (9, 231). Other studies have shown FGFs protect against excitotoxicity and prevent the production of free radicals (131, 232). Treatment with FGF1 and FGF2 has been shown to significantly increase the growth and survival of various types of neurons including dopaminergic, cerebellar, hippocampal, and isolated sensory neuronal cells (143, 228, 233–235). Furthermore, both keratinocyte growth factor and basic FGF treatments appears to provide neuroprotection following SCI (131, 236). A prospective, open label uncontrolled clinical study demonstrated the safety, feasibility, modest improvements in functional outcomes in chronic SCI patients at 48 months that were treated with survival nerve grafts with fibrin glue containing acidic FGF (aFGF) (131, 237). Another clinical trial involving nine patients with cervical SCI treated with direct implantation of fibrin glue with aFGF for over six months showed significantly improved ASIA motor and sensory scale scores between the preoperative and follow-up (6-months postoperative) (131, 238). In addition, a prospective, open-label, uncontrolled clinical trial with 60 recruited patients with SCI demonstrated substantial improvement in AISA motor and sensory scale scores after 24 months following FGF therapy (239).

Granulocyte colony-stimulating factor (G-CSF) is a cytokine that in addition to its role in inducing proliferation, survival, and development of granulocyte lineage cells has been shown to play key roles in functional recovery following neuronal injury (240, 241). More specifically, studies have shown that G-CSF decreases both neuron and oligodendrocyte apoptosis, suppresses inflammatory cytokines (reduces TNFα and IL-1β in the spinal cord lesion), inhibits glutamate excitotoxicity, enhances angiogenesis, and promotes proliferation/mobilization of neuronal cells (131, 242–245). Based on these studies, phase I/IIa clinical trials were conducted and confirmed both the feasibility and safety of G-CSF therapy following SCI (246–248). The sentence should now read: "A phase II multicenter, prospective, non-blinded, nonrandomized clinical trial assessing patients with acute cervical SCI (within 48 hours after injury) receiving G-CSF therapy versus control therapy showed significant improvement (P<0.01) of motor paralysis and ASIA scores even at 1 year after injury in patients receiving G-CSF therapy compared to patients receiving control therapy (249). A recent phase 3 prospective, randomized, double-blind, and placebo controlled comparative trial was conducted and failed to show significant changes in ASIA motor scores from baseline to three months after G-CSF therapy (245). Additional randomized controlled trials will be needed in the future in order to determine the therapeutic benefits of SCI. Immunological pharmaceutical clinical studies are summarized in Table 1 .

Clinical trials, preclinical studies, and meta-analyses have shown that although the efficacy of stem cell therapy is promising, many challenges remain which will require further optimization and innovative solutions (293). One of the biggest and most concerning challenges of stem cell therapy for spinal cord injury moving forward is the potential for adverse effects. Although many clinical trials have been reported as safe without mortalities or severe morbidity, tumorigenicity, immunogenicity, and cell therapy-related immunotoxicity are potentially serious consequences of stem cell therapy (294). More specifically, to date there have been at least 28 kinds of adverse effects reported (293). Some side effects reported are likely secondary to the delivery procedures rather than the cells (294). Side effects secondary to the delivery procedure itself likely include subarachnoid hemorrhage, cerebrospinal fluid leaks, and transient deterioration in sensorimotor symptoms (25, 217, 280, 294–297). Moving forward, innovative delivery strategies will need to be investigated and employed to reduce side effects and further improve efficacy. In addition, side effects not associated with the transplanted cells and/or procedures were reported and include but are not limited to vomiting, pulmonary thromboembolism, fever, body aches, urinary tract infection (217, 262, 273, 291, 292, 294, 296, 298). Unfortunately, comprehensive assessment of safety and efficacy of stem cell therapy has been challenging for SCI patients as clinical trials in the past have been limited by relatively low sample sizes and insufficient control groups (299). In the future, large preclinical studies and early clinical studies are necessary in order to assess the safety of transplantation of human stem cells. Without rigorous assessment of the safety and potential adverse effects, the widespread adaptation of stem cell therapy will continue to be hampered.

Next, the properties and characteristics of the stem cells themselves must be rigorously assessed in order to continue to improve efficacy. Arguably the most critical factor in response to stem cell therapy is the origin of the stem cells used in the therapy (269, 300, 301). In addition to the studies examining MSCs, studies have also suggested potential feasibility and safety of human neural stem cell transplantation, but the clinical efficacy of such therapies requires further study. Similarly, studies examining autologous human Schwann cell transplantation have demonstrated feasibility, although currently they show unclear benefit (296, 302). Multiple studies have also demonstrated safety and feasibility of olfactory mucosa autografts (297, 303). To date, factors such as cell survival and integration are hurdles that potentially explain some of the heterogenous results among clinical trials must be improved through more carefully standardizing experimental designs.

The dose of stem cells also appears to be an especially critical factor and has varied significantly among trials from 10⁶ to 10¹⁰ cells (250, 304). The route of administration of stem cells also appears to have a significant impact on therapy as intraspinal approaches result in the highest level of cell engraftment but is highly invasive compared to intravenous approaches which have been suggested to result in the lowest number of cells within the damaged lesion (250, 264). Furthermore, the optimal timing of stem cell transplantation in SCI remains to be determined. One pitfall to stem cell therapy moving forward is that a majority of clinical trials are single-centered investigator-oriented trials and therefore, further standardization including standardization between agencies will be necessary moving forward (250). Stem cell clinical studies are summarized in Table 2 . In the future, clear and standardized transplantation strategies, cell numbers, treatment times, storage protocols, and consistent generation protocols which produce genetically stable and consistent therapeutic cells will be critical to unlock the potential efficacy of stem cell therapy for spinal cord injury

Importantly, recent evidence suggests that the biological and therapeutic effects of stem cell therapy including MSCs are contained within extracellular vesicles and secreted factors (305–307). Next, we review the exciting potential of EV therapy.

Arguably the cell source of the EVs is the most critical factor in manufacturing EVs for therapeutic purposes as the phenotype of EVs including the cell-surface proteins, cytosolic proteins, and composition of the nucleic acids that are available to be packaged into EVs are at least partly dictated by the cell source (339–342). For instance, the EV content of one critical class of small ncRNAs and miRNAs is significantly different among cell sources suggesting the importance of cell source as an EV design criterion (343, 344).

The cell culture conditions and microenvironment are also critical in order to tune the therapeutic bioactivity of EVs (319). Parameters including cell density seeding, cell age/passage, and the collection frequency of EVs have been reported to result in significant changes in production rates and the therapeutic activities in several different cell types (319, 345). For instance, EVs isolated from stressed astrocytes that were oxygen and glucose deprived resulted in neuroprotection in vitro most likely due to an increase in EV associated miR-92b-3p content compared to astrocytes that were not oxygen and glucose deprived (346, 347). The microenvironment cells are grown in also appears to be a critical factor as neural progenitor cells grown in proinflammatory conditions caused significant changes in RNA and protein levels with one study demonstrating enrichment of membrane bound IFN-γ and IFN-γ receptor 1 (IFNGR1) complexes that can induce key regulators of neuroinflammation and cell proliferation such as IFNGR1 and STAT1 (348). The employment of 3D printing could also be critical as 3D printing biomaterials will facilitate construction of complex precisely controlled microenvironments that could optimize EV production (319, 349, 350).

EVs are advantageous for therapeutic purposes as they protect nucleic acids and other bioactive components from degradation over long distances while also ensuring delivery to the cytosol of the target cell without triggering an auto-immune response (319). Although EVs hold great promise for transporting small interfering RNAs (siRNAs), miRNAs, mRNAs, and antisense oligonucleotides, in the case of some critical ncRNAs, the loading can be relatively low with as few as one bioactive copy of a given miRNA per EV (319, 351–353). Thus, it may be necessary to use high or repeated EV doses or employ cargo loading control strategies in situations where the dose of bioactive cargo is insufficient to adequately modulate gene expression (319, 351–353). Multiple studies have reported methods to load CNS-specific miRNA cargo into non-resident CNS cell types using genetic modification and subsequent over expression (14, 319). Other studies have used sonication, coincubation of EVs with desired cargo, and lipid or cholesterol functionalization (354–356).

One of the most important limitations of EVs to address for therapeutic purposes is the isolation protocols and scalability. Currently, several EV isolation strategies, each with their own advantages, have been utilized including differential centrifugation, ultrafiltration, precipitation, size exclusion chromatography, immunoisolation, and density gradient separation (357, 358). Ideally, an EV isolation protocol would be readily scalable to produce EVs with high purity for therapeutic purposes. The issue is that methods that result in a high purity can often have a relatively low yield and protocols with relatively high yields often co-isolate various protein aggregates, lipoproteins, and other impurities (357, 358). Therefore, the isolation protocol must always be optimized for yield and purity when manufacturing EVs for therapeutic purposes. Currently in the field, most groups are isolating EVs by differential centrifugation followed by ultracentrifugation (359). In situations where high purity is needed, further density-based purification using a sucrose or iodixanol centrifugation gradient can be used (360).

Finally, the route of administration appears to be a critical factor in EV therapy. More specifically, a 2021 study demonstrated that intralesional injection of EVs in a SCI rat model resulted in a more robust improvement of BBB score and sub-score compared to EVs delivered intravenously (288). Although intravenous injection is less invasive, delivering EVs intravenously will require higher doses of EVs as EVs can accumulate at off-target sites including the liver, spleen, and lungs, and systemic delivery of EVs intravenously results in delayed delivery of therapeutic EVs to the site of injury (288, 319, 333). The half-life of EVs is also an important to consider as a study observed a half-life of approximately 30 minutes in vivo in EVs loaded with dye after intravenous delivery (358). Therefore, intralesional injection could facilitate improved outcomes by more rapidly delivering a higher dose of therapeutic EVs to the lesion while minimizing off target accumulation and systemic degradation (7, 47, 48).