Regulation of Inflammatory Cytokines for Spinal Cord Injury Repair Through Local Delivery of Therapeutic Agents

Abstract The balance of inflammation is critical to the repair of spinal cord injury (SCI), which is one of the most devastating traumas in human beings. Inflammatory cytokines, the direct mediators of local inflammation, have differential influences on the repair of the injured spinal cord. Some inflammatory cytokines are demonstrated beneficial to spinal cord repair in SCI models, while some detrimental. Various animal researches have revealed that local delivery of therapeutic agents efficiently regulates inflammatory cytokines and promotes repair from SCI. Quite a few clinical studies have also shown the promotion of repair from SCI through regulation of inflammatory cytokines. However, local delivery of a single agent affects only a part of the inflammatory cytokines that need to be regulated. Meanwhile, different individuals have differential profiles of inflammatory cytokines. Therefore, future studies may aim to develop personalized strategies of locally delivered therapeutic agent cocktails for effective and precise regulation of inflammation, and substantial functional recovery from SCI.


Introduction
The spinal cord is part of central nervous system (CNS), extending from medulla oblongata to the lumbar region in the vertebral canal. It plays a crucial role in the transmission of motor and sensory information, as well as in the coordination of certain reflexes. The importance of its functions makes SCI one of the most disastrous traumas. SCI is the damage to the spinal cord that results in alterations of its functions, including loss of sensory, motor and visceral functions. [1] SCI often ends up with paralysis. [2] The incidence of SCI varies from 3.6 to 195.4 patients per million in different countries worldwide. [3] For example, in the United States, an estimated 12 500 new cases occur every year, and as many as 276 000 persons are living with SCI. For each patient, the estimated lifetime costs directly associated with SCI range from $1 113 990 to $4 724 181, not including any indirect costs such as loss of wages, fringe benefits and productivity, which average $71 961 per year. [4] The high prevalence and lifetime costs demonstrate the urgency for therapy development. However, there is currently no effective cure that leads to functional recovery.
After SCI, a series of inflammatory responses are activated. The disruption of the blood-spinal cord barrier (BSCB) leads to the accumulation of various inflammatory cytokines. [1] Inflammatory cytokines are a broad category of soluble small proteins that are the mediators and modulators of the complicated functional interactions and reactions of the immune system. [5] The enormous influence of the inflammatory cytokines on SCI repair has been demonstrated by a large body of research, which will be reviewed here.
Therapeutic agents can be used for the regulation of inflammatory cytokines to promote repair from SCI. However, the fact that the therapeutic agents have difficulty crossing the BSCB restricts the treatments of spinal cord diseases. The BSCB consists of tight junctions and capillary endothelial cells without intracellular fenestrations [6] and dramatically limits compounds entering the spinal cord from the blood. [7] There are three main approaches to achieving entry of the agents into the CNS: 1) tailoring therapeutic agents to take advantage of the natural permeability of the BSCB; 2) disruption of the BSCB; and 3) local delivery. [8] In the field of SCI repair, much research focuses on the utilization of local delivery methods. Therefore, in this review, we will also summarize and discuss The balance of inflammation is critical to the repair of spinal cord injury (SCI), which is one of the most devastating traumas in human beings. Inflammatory cytokines, the direct mediators of local inflammation, have differential influences on the repair of the injured spinal cord. Some inflammatory cytokines are demonstrated beneficial to spinal cord repair in SCI models, while some detrimental. Various animal researches have revealed that local delivery of therapeutic agents efficiently regulates inflammatory cytokines and promotes repair from SCI. Quite a few clinical studies have also shown the promotion of repair from SCI through regulation of inflammatory cytokines. However, local delivery of a single agent affects only a part of the inflammatory cytokines that need to be regulated. Meanwhile, different individuals have differential profiles of inflammatory cytokines. Therefore, future studies may aim to develop personalized strategies of locally delivered therapeutic agent cocktails for effective and precise regulation of inflammation, and substantial functional recovery from SCI.
G-CSF is a 19.6 kDa cytokine [99] that has been used in clinical applications for the treatment or prevention of chemotherapyinduced neutropenia, bone-marrow harvest, and antiinfection treatment. [39] Urdzíková et al. first found that G-CSF improved functional recovery after SCI in rats [100] and concluded that the effect was due to the mobilization of bone marrow stem cells. Later, Yamazaki et al. first applied G-CSF in a mouse SCI model and reported hindlimb functional recovery. [38,101] G-CSF was discovered to attenuate neuronal death [38] and enhance connectivity. [39] Moreover, G-CSF protected oligodendrocytes in SCI repair. [37] These neuroprotective effects might be attributed to the promotion of angiogenesis [43] and autophagy [40] and alternative activation of microglia. [42] Chen et al. investigated a different delivery method. [41] Compared to methylprednisolone (MP) administration, direct intrathecal administration of G-CSF suppressed the expressions of TGF-β1, CSPGs, and TNF-α.

Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF)
GM-CSF is a 14.2 kDa hematopoietic factor [102] which is essential for proliferation and differentiation of mature granulocytes and macrophages. [103] Ha et al. first demonstrated that GM-CSF improves functional outcome after rat contusive SCI. [46] The improvement is probably due to the prevention of apoptosis of the cells, including neurons, [46] via reduction of the expression of the proapoptotic proteins p53, p21, and Bax, and induction of nucleophosmin-1 [44] and the antiapoptotic protein B-cell lymphoma 2 (Bcl-2). [45] Additionally, GM-CSF increased BDNF expression by macrophages, and subsequently stimulated axonal regeneration. [49] Moreover, GM-CSF suppressed glial scar formation, [47,48] activated dendritic-like cells and neural stem cells (NSCs), [53] and improved the survival of transplanted NSCs. [50,52] GM-CSF was also found to improve sensory function by minimizing the abnormal sprouting of sensory nerves. [51]

Interferon-Gamma (IFN-γ)
IFN-γ, one of the type II interferons, has the major activity of immunoregulation. It can be produced by mitogenically or antigenically stimulated lymphocytes.
IFN-γ was first found to enhance hindlimb function in a mouse contusive SCI model. [60] IFN-γ facilitated the secretion of IL-10 from T helper 1 cells (Th1) and microglia/macrophages, [58] decreased CSPG expression from reactive astrocytes, and increased the expression of neurotrophic factors, including glial cell line-derived neurotrophic factor and insulin-like growth factor-1. [60] IFN-γ is also crucial to the immunological plasticity of the choroid plexus epithelium, allowing the regulated entry of T cells and monocytes, which support CNS repair. [59]

Interleukin-4 (IL-4)
IL-4 is an antiinflammatory cytokine produced by activated T-lymphocytes. [106] IL-4 is associated with the Th2 immune response and the activities of other hematopoietic lineage cells.
Yaguchi et al. demonstrated in a mouse SCI model that IL-12, secreted from DCs, promoted functional recovery. [75] The recovery might be because IL-12 increased the number of activated microglia/macrophages and DCs, as well as BDNF expression, and subsequently improved neurogenesis and remyelination.
Gonzalez et al. first showed that antibody neutralization of CXCL10 enhanced functional recovery in a mouse dorsal hemisection SCI model. [124] The neutralization reduced apoptosis, [122] inhibited T-lymphocyte invasion [124] and inflammation, [125] and thus enhanced tissue sparing. [123] CXCL10 neutralization also enhanced angiogenesis and the subsequent axon sprouting.

Interleukin-17 (IL-17)
IL-17, also known as IL-25, is mainly produced by T cells and their precursors. IL-17 is closely related to the origin of a type 2 immune response, which is characterized by the differentiation of Th2 and the production of the type 2 cytokines. [160,161] Hill et al. first showed that IL-17 hindered functional recovery in a mouse contusive SCI model. [137] This might be because IL-17 recruited B cells, dendritic cells, neutrophils, [137] and Th cells, [138] and increased lesion size and demyelination.

Local Delivery of Therapeutic Agents Regulates Inflammatory Cytokines and Promotes SCI Repair
Local delivery means that the therapeutic agents are administered directly into the lesion site or that an agent administration system is constructed for local application. In comparison with systemic administration, local delivery overcomes the BSCB and thus improves the concentration of the local agent with less damage to other sensitive organs. [166,167] Here, we classify local delivery into 4 subunits: direct administration, sustained release, cell transplantation, and transgene, and discuss their promotion of SCI recovery through regulation of inflammatory cytokines in animal models (Figure 1).

Direct Local Administration
Administrating agents around or in the lesion site can achieve sufficient therapeutic effects of regulating inflammatory cytokines and promoting injury repair. The directly administered agents include inhibitors, cytokines and other therapeutic agents.

Inhibitors
Several investigations have used antibodies to directly inhibit the levels of certain inflammatory cytokines. Local applications of anti-TNF-α antibodies reduced apoptosis, [141,143] and attenuated BSCB disturbances and cord pathology, [164] thus enhancing neurorepair. [168] The suitable combination of antibodies against TNF-α and NOS also improved neuroprotection and functional recovery. [164,168] In addition to anti-TNF-α antibodies, CCL20neutralizing antibodies have been found to promote function recovery as well. [169] They reduced the levels of IL-1β, IL-6, TNFα, and IL-17, and decreased Th17 cell recruitment.
Antagonists are also important inhibitors. Intrathecal injections of IL-1ra decreased apoptosis after rat SCI. [126] The antagonists against TNF-α increased levels of IL-4 and IL-10 and decreased the level of IL-1β as well as tissue damage, improving functional outcome. [170,171]

Cytokines
In the SCI epicenter, directly administered cytokines include EPO, G-CSF, and VEGF. The intrathecal injection of recombinant human EPO (rhEPO) was demonstrated to enhance the SDF-1/CXCR4 axis in transplanted bone marrow stromal cells (BMSCs), thus increasing their migration toward the lesion site. [26] rhEPO also reduced the apoptosis of BMSCs and improved neurological outcome. G-CSF administered intrathecally had antiinflammatory and anti-apoptotic effects by directly regulating the activity of MAPK and Akt pathways, suppressing the expression of TGF-β1 and TNF-α, and improving neurological function in both ischemic and contusive SCI models. [41,172] The direct injection of VEGF was also found to regulate inflammatory cytokines post-SCI. [173] VEGF decreased the levels of IL-1β and TNF-α, reduced motor neuron loss, and improved function.

Other Therapeutic Agents
Moreover, there are other therapeutic agents, including chondroitinase ABC, Annexin A1 (ANXA1) and 11-dehydrosinulariolide. Multiple lumbar puncture injections of chondroitinase ABC promoted IL-10 expression and inhibited IL-12b expression after SCI. [174] ANXA1 is suggested to be an endogenous neuroprotective agent that mediates the antiinflammatory actions of glucocorticoids. [175] The intraspinal injection of ANXA1 significantly decreased IL-1α expression, inhibited the inflammatory response, and reduced tissue damage. The intrathecal injection of the coral-derived 11-dehydrosinulariolide was found to attenuate SCI-induced upregulation of TNF-α, increase M2 polarization, and improve functional recovery. [176]

Sustained Release
Sustained release forms are designed to release agents at a controllable rate in order to maintain a constant drug concentration for a period of time. To regulate inflammatory cytokines after SCI, sustained agent release can be achieved mainly through two methods: biomaterial systems and pumps.

Biomaterial Systems
In SCI models, the topically applied biomaterial systems involve gelatin sponges, hyaluronan methylcellulose (HAMC), and biomaterial particles.
Gelatin Sponge: Gelatin sponges are sterile sponges with porous structure that work primarily as hemostatic agents. They are pH-neutral and biodegradable and, therefore, also used for local delivery. [177,178] After SCI, topically implanted gelfoam sponges soaked with IL-12 increased the number of activated ameboid-type microglia/macrophages and their BDNF expression, accompanied by increased remyelination and promotion of functional recovery. [75] In another study, gelfoam sponges loaded with GM-CSF modulated apoptosis and promoted neuroprotection. [179] Further, Li et al. developed an NT-3/fibroincoated gelatin sponge scaffold, which could continually release NT-3 for 28 days. [180] Its local application decreased TNF-α level, CD68-positive cells and cavity formation and improved neural regeneration.
Gelatin Sponge: HAMC: The University of Toronto developed a new biomaterial called hydrogel of HAMC as a drug carrier for SCI treatment. [189] Compared with intrathecal or intraperitoneal  EPO injection, HAMC EPO delivery had better results in tissue sparing and functional recovery post-SCI. [190] Interestingly, HAMC had some therapeutic functions on its own, [189,191] such as improving the outcome of post-traumatic syringomyelia. [191] Intrathecal injection of HAMC reduced IL-1α levels and lesion volume and finally promoted neurobehavioral recovery from arachnoiditis caused by SCI.

Osmotic Pumps
Osmotic pumps are small, infusion pumps for administrating agents in animals. They are convenient, cost-effective and capable of maintaining drug levels. [192] These pumps have been widely employed to release agents for regulation of inflammatory cytokines post-SCI, including cytokines, inhibitors, and other therapeutic agents.
Cytokines: Intrathecal administration of SDF-1α by osmotic pumps reduced the levels of inflammatory cytokines, [78] including IL-18, IL-1β, and TNF-α, protected neurons, [77] and promoted sprouting of corticospinal tract axons and motor recovery. [78,84] Inhibitors: In rat SCI models, EGFR blockers from osmotic pumps depressed IL-1β and TNF-α production, inhibited microglia/astrocyte activation and glial scar/cavity formation, and promoted axonal outgrowth with functional recovery. [193] Central administration of a soluble TNF blocker reduced spinal cord damage and improved locomotor function. [171] Etanercept, a blocker of TNF-α, was administered into the SCI site by pumps, attenuated microglial activation and prevented the development of mechanical hypersensitivity. [194] Local pumping of IL-1ra decreased IL-1β, [159] blocked its combination with its receptor, and completely abolished the increased apoptosis after SCI. [128] Other Therapeutic Agents: Metabotropic glutamate receptor 5 (mGluR5) has neuroprotective properties. mGluR5 agonist treatment through pumps reduced TNF-α, microglial activation and lesion volume and improved functional recovery. [195] Agomir-210, a chemical mimic of microRNA-210, attenuated the expression of proinflammatory TNF-α and IL-1β, upregulated IL-10, and inhibited apoptosis in SCI models. [196] Pumping a connexin43 mimic into the injured spinal cord decreased TNF-α and IL-1, reduced secondary damage and eventually promoted hindlimb function. [197]

Local Cell Transplantation
Recently, cell transplantation has been one of the hotspots in SCI repair. Transplanted cells can release cytokines/chemokines and influence the inflammatory response. [198] Various investigations in SCI models have shown that local cell transplantation regulates inflammatory cytokines and facilitates recovery. Donor cells for local transplantation involve stem cells and somatic cells.

Somatic Cells
Olfactory Ensheathing Cells (OECs): OECs replace necrotic or apoptotic neural cells and help remyelination and neurotrophin secretion. [218,219] Transplanted OECs increased the reactivity of GFAP, tomato lectin, IL-1β, and iNOS in the lesioned cord of SCI rats at 7 days, which were all reduced at 14 days post-injury. [220] While Schwann cells (SCs) promote axonal regeneration by secreting basement membrane components type IV collagen and laminin, [221] OEC and SC cotransplantation reduced astrocyte/microglia/macrophage infiltration and inhibited proinflammatory factor (IL-6 and TNF-α) secretion, increased the levels of anti-inflammatory factors (IL-10 and IL-13), repaired cystic cavities, and improved functional recovery. [222] Other Somatic Cells: Skin-coincubated macrophages injected into the caudal border of the lesion elevated the secretion of IL-1β and BDNF, reduced the secretion of TNF-α and showed meaningful recovery in rats. [223] In addition, transplantation of Wharton's jelly cells (WJCs), which are isolated from umbilical cord Wharton's jelly tissue, inhibited IL-1β expression, promoted nerve growth factor (NGF) expression in spinal cord tissues, and improved neurological function recovery. [224]

Local Gene Therapy
Gene therapy refers to the techniques and strategies to genetically transform or modify cells for healing or attenuating disease conditions. Various animal experiments have shown that regulation of inflammatory cytokines and repair from SCI is facilitated by local gene therapy, including liposome-mediated and virus-mediated gene therapy.

Liposome-Mediated Gene Therapy
Ito et al. built a liposome-mediated IFN-β gene delivery system and applied it locally after SCI. This gene therapy inhibited glial scar formation and promoted functional recovery by deactivating the MEK-ERK pathway. [57] Liposomes were also used to deliver miR-199b. [225] Intrathecal delivery of miR-199b reversed the upregulation of IL-1β and TNF-α, and attenuated acute SCI.

Virus-Mediated Gene Therapy
Viral vector injections are also used for expressing specific proteins in targeted areas. [226] To regulate inflammatory cytokines post-SCI, herpes simplex virus (HSV), lentivirus, adenovirus and poliovirus have been locally applied in animal experiments.
HSV is neurotropic, lives without replication, has a large genome and is easily operated in tissue culture. [227] HSV was used as a vector to express IL-10 and EPO locally in SCI mice. Local injection of IL-10 vector was discovered to promote survival of neurons. [66] Wang et al. made an HSV vector coding EPO and injected it locally after SCI. Their study showed an enhanced tissue sparing, preserved axons and promotion of synaptogenesis, in accordance with a diminution of the injury size and a significant functional improvement. [34] Lentivirus is the most widely used virus in gene therapy studies aimed at inflammatory cytokines after SCI. Local applications of lentivirus-mediated regulation of both TNF-α [228] and IL-10 [71] promoted functional recovery. In addition to inflammatory cytokines, local lentivirus-mediated regulation of other proteins also affected inflammation and facilitated neurorepair. Ji et al. edited a lentivirus BDNF-overexpressing vector and injected it into the SCI epicenter. [229] The injection resulted in reduced IL-1β and TNF-α, elevated IL-10 and IL-13, and promotion of M2 polarization. This prevented CST retraction and improved motor function. Local lentivirus-mediated RNA interference (RNAi) of the regulator of calcineurin 1 suppressed the increase in IL-1β and TNF-α and improved behavioral performance in injured rats. [230] Local transfection of siRNA against the TNF-like weak inducer of apoptosis decreased IL-1β and TNF-α, and enhanced locomotor functional recovery in injured mice. [231] Adenovirus is also widely applied for local inflammatory cytokines in SCI. Local adenovirus-mediated RNAi of IL-1β decreased TNF-α and neuronal loss, and improved locomotor function. [232] Adenovirus-mediated local overexpression of BMP and activin membrane-bound inhibitor decreased IL-1β, IL-6, and TGF-β, and facilitated functional recovery. [233] Poliovirus vectors have the capacity for self-amplification and for expressing foreign proteins without infecting other cells. Jackson et al. intrathecally injected poliovirus vectors to express IL-10 in SCI animals. [70] This treatment inhibited microglia activation and improved functional recovery.

Clinical Studies
Quite a few promising clinical studies have tested different treatments for SCI. [234] As inflammatory cytokines play important roles in SCI, there are some studies focusing on inflammatory cytokines for SCI therapy. We found 24 registered clinical trials (clinicaltrials.gov) and 14 published clinical reports on inflammatory cytokines in SCI. The United States and China have the most registered clinical trials, with 6 and 4 trials, respectively. Canada, Iran and Japan have the most published clinical reports. Each of the 3 countries has 4 reports ( Figure 3A). There is a rising trend in the number of registered clinical trials and published clinical reports, with 62.5% of the registered trials and 85.7% of the published reports coming after 2010, showing the current interest for the potential of regulation of inflammatory cytokines for SCI healing ( Figure 3B).
Here, we summarized and discussed all of the relevant clinical studies of SCI repair through regulation of inflammatory cytokines ( Table 3). According to the method of agent administration, we divide these trials into four types: subcutaneous injection, intravenous injection, intramuscular injection and local delivery.

Subcutaneous Injection
Subcutaneous injection is a shot given into the subcutis. This form of administration is used to give small amounts of certain kinds of medicine. At present, clinical trials with subcutaneous injections of inflammatory cytokines, including G-CSF and EPO, have shown promise.

G-CSF
Derakhshanrad et al. recruited 19 patients, all of whom received G-CSF (5 µg kg −1 day −1 ) subcutaneously for 5 days. [235] Some patients showed mild side effects (rash, fever, bone pain, neuropathic pain, and spasticity) that disappeared in a week. The results demonstrated that subcutaneous injection of G-CSF is safe for SCI patients. After a year, this team recruited more patients. After receiving the identical G-CSF administration for 7 consecutive days, patients with motor-incomplete SCI showed better improvement, including improved American Spinal Cord Injury Association (ASIA) motor score and light touch and pinprick sensory scores. [236]

EPO
In another case, a 42-year-old woman showed low response to routine drug treatment after cervical discectomy that directly Adv. Sci. 2018, 5, 1800529 led to SCI. [237] Then, a derivative of EPO was injected subcutaneously (100 µg day −1 ) for 3 days as an addition to her medications. The next day, her pain was relieved, and lower limb strength recovered gradually. After one year, she walked without help and returned to her previous job.
Additionally, rhEPO administration has had a positive impact on SCI patients with pressure ulcers (PUs) associated with Anemia of Chronic Disease (ACD). [238,239] Four patients received rhEPO (75 IU kg −1 ) subcutaneously three times a week for 6 weeks. [238] After the treatment, ulcer status was improved in all patients, as shown by the reduction in ulcer depth, decrease in the wound surface areas and improvement in wound appearance and damage degree. However, in another trial, rhEPO administration at a dose of 75 IU kg −1 was not enough to significantly improve patient prognosis. [239]

Intravenous Injection
Intravenous injection is a method that infuses liquid substances directly into the vein, through which they are then carried by blood circulation and brought into effect systemically. Currently, intravenously administered agents being examined in clinical trials include G-CSF and EPO.

G-CSF
In 2012, Sakuma et al. reported that G-CSF administration was safe and had neuroprotective effects. [240] A total of 5 patients received G-CSF (5 µg kg −1 day −1 ) intravenously for 5 days, while another 10 patients received 10 µg kg −1 day −1 . None of the 15 patients showed side effects, and treatment resulted in neurological improvements in both motor and sensory functions. Along with the above research, this team also obtained similar results in patients with acute SCI caused by falls and road trauma, as well as athletic injury. [241] In consideration of the influence of different injury sites, this group continued a further trial only on patients with cervical injury. [242] The ASIA motor scores collected at one week, three months and one year after primary injury were dramatically increased in the G-CSF group compared to those in the control group, suggesting that intravenous injection of G-CSF improved the muscle strength of the upper and lower extremities.
In addition, compared with high-dose methylprednisolone sodium succinate (MPSS) treatment, intravenous administration of G-CSF produced significantly higher ASIA motor scores. [243] Furthermore, less severe side effects such as pneumonia were observed in patients treated with G-CSF, indicating that intravenous administration of G-CSF is safer and more effective.

EPO
In clinical studies, EPO has also been administered intravenously. Ten paraparetic patients with malignant extradural spinal cord compression (MESCC) received 1500 IU kg −1 EPO by chemotherapy infusion pump over 30 min and showed improved motor function and shorter recovery time, including functional class and ambulation. [244] Additionally, Costa et al.
found that intravenous administration of EPO resulted in a meaningful clinical improvement in ASIA Impairment Scale scores compared with MP treatment. [245] Meanwhile, the combination of EPO and MP has been confirmed in clinical trials to be beneficial for SCI treatment. [246] ASIA scores were dramatically higher in patients treated with the combination than in those treated with MPSS only. However, in another trial, administration of rhEPO together with MP did not improve the functional outcome of patients with traumatic cervical SCI, which might be due to insufficient dosage (500 IU kg −1 ) of rhEPO, expression of the EPO receptor or the route of administration. [247]

Intramuscular Injection
Intramuscular injection is the injection of a substance directly into a muscle where medicines are absorbed quickly by abundant blood vessels. Intramuscular injection of EPO is also used in SCI treatment.
Xiong et al. recruited 63 patients with spinal cord ischemiareperfusion (I-R) injury and found that combined administration of 1000 IU kg −1 EPO injected intramuscularly and MP injected intravenously markedly improved AISA score and the activity of daily living compared to MP administration alone. [248]

Local Delivery
Two methods of local delivery have been applied in clinical trials of SCI: intrathecal delivery and intraspinal injection. These trials have been registered (clinicaltrials.gov), but currently they have no posted results.
Intrathecal delivery is a very important means of local administration and is relatively convenient as it can be done by lumbar puncture. BMSCs, adipose-derived MSCs and umbilical cord-derived MSCs have been utilized respectively in the intrathecal delivery of 6 trials (NCT02481440, NCT01186679, NCT03308565, NCT01393977, NCT02482194, and NCT02570932). The outcome measures include TNF-α, TGF-β, IL-1β, IL-6, iNOS, IL-10, subsets of T-lymphocytes, and C-reactive protein.
There are two clinical trials using intraspinal injection (NCT01046786 and NCT02917291). The investigators at the Chinese University of Hong Kong and The University of Hong Kong slowly injected cord blood mononuclear cells into the posterior grey matter after laminectomy and opening of the dura. In the other trial, the intramedullary injection of allogeneic adipose-derived MSCs was used.

Future Directions
Current studies demonstrate that regulation of inflammatory cytokines plays a significant role in SCI repair both in animal models and clinical trials. Nevertheless, two problems exist: enabling the efficacy of agents at small dosages and reducing side effects. Local delivery is showing outstanding advantages in solving these problems. Local delivery of therapeutic agents and personalized therapy have great potential in the future.

Local Delivery Enables the Efficacy of Agents at Small Dosages
We have learned that subcutaneous or intravenous injections of inflammatory cytokines are widely used for the clinical treatment of SCI. As mentioned above, EPO administration of 1500 [244] and 1000 IU kg −1 [248] can improve functional recovery in SCI patients, while it fails to produce apparent therapeutic effects in a dose of 500 IU kg −1 . [247] Thus, we propose that local delivery of a small-dose administration may have a better effect, and we hope to apply these local delivery methods to clinical studies in the future.
Moreover, a large number of animal experiments have revealed that local delivery of therapeutic agents can regulate inflammatory cytokines and promote functional recovery, which further illustrates that local agent delivery can promote SCI repair in the clinical setting. Local, but not systemic administration of TNF inhibitor, was therapeutic for traumatic SCI in mice. [171] Central administration of XPro1595 reduced damage to the lesions and improved locomotor function, while peripheral administration of anti-TNF therapies was inefficient.
At the same time, SCI could occur in any spinal cord site, so local delivery is more accurate and economic than systemic administration.

Local Delivery Reduces Side Effects
In systemic agent delivery, high doses of agents produce side effects. In clinical trials of SCI, the restriction of the blood-brain barrier and short half-life of IL-10 required a high dose of IL-10 in systemic delivery, [249] which resulted in increased susceptibility to Klebsiella pneumonia, [250] Listeria monocytogenes, [251] and Streptococcus pneumonia infections. [252] In human beings, systemic administration of MP also induced side effects, including wound infection, hyperglycemia, severe pneumonia, and delayed wound healing. [253] Importantly, local delivery indeed reduced the MP-induced side effects in animal experiments. [254] Further, the emergence of local delivery methods in more and more clinical trials (NCT02481440, NCT01186679, NCT03308565, NCT01393977, NCT02482194, NCT02570932, NCT01046786, and NCT02917291) is also showing their great potential. Thus, enabling efficacy at small dosages and reducing side effects, local delivery of therapeutic agents is very promising for the future.

Future Personalized Strategies of Locally Delivered Therapeutic Agent Cocktails
Local delivery of a single agent only affects a small portion of the inflammatory cytokines, and different individuals have differential profiles of inflammatory cytokines. Huang et al. found that inflammatory cytokine profiles were related to monocyte (MO) phenotypes in SCI patients. [255] According to the expression of CD14 and CD16, the inflammatory MO subpopulation contained two subgroups: M1-dominant subpopulations (CD14 low /CD16 + ) and M2-dominant subpopulations (CD14 high / CD16 + ). They found that the M1-dominant subpopulation showed higher levels of IL-12 and CXCL10 in the plasma, while the M2-dominant subpopulation presented with more IL-10, IL-15, and IL-7. Therefore, future studies may aim to develop personalized strategies for locally delivered therapeutic agent cocktails for effective and precise regulation of inflammation and substantial functional recovery from SCI.