Emerging molecular therapeutic targets for spinal cord injury

Abstract

Introduction:

Spinal cord injury (SCI) is a complicated and devastating neurological disorder. Patients with SCI usually have dramatically reduced qualify of life. In recent years, numerous studies have reported advances in understanding the pathophysiology of SCI and developing preclinical therapeutic strategies for SCI, including various molecular therapies, and yet there is still no cure.

Areas covered:

After SCI, tissue damage, responses and repair involve interactions among many cellular components, including neurons, axons, glia, leukocytes and other cells. Accordingly, numerous cellular genes and molecules have become therapeutic targets for neural tissue repair, circuit reconstruction and behavioral restoration. Here, we review the major recent advances in biological and molecular strategies to enhance neuroprotection, axon regeneration, remyelination, neuroplasticity and functional recovery in preclinical studies of SCI.

Expert opinion:

Researchers have made tremendous progress in identifying individual and combined molecular therapies in animal studies. It is very important to identify additional highly effective treatments for early neuroprotective intervention and for functionally meaningful axon regeneration and neuronal reconnections. Because multiple mechanisms contribute to the functional loss after SCI, combining the most promising approaches that target different pathophysiological and molecular mechanisms should exhibit synergistic actions for maximal functional restoration. [Databases searched: PubMed; inclusive dates: 6/27/2019]

1. Introduction

After initial traumatic injury to the spinal cord (i.e., primary injury), ongoing pathological changes cause secondary damage, most notably, progressive hemorrhaging, intravascular thrombosis, vasospasm, ischemia and edema of the injured spinal cord (Fig. 1). Such pathological events result in free radical formation, lipid peroxidation, excitotoxicity, apoptotic and necrotic cell death, inflammation and immunological responses around the lesion. The secondary cascades dramatically expand the primary injury into adjacent tissues and spinal cord segments, causing further neuronal and glial loss, cavitation, migration and proliferation of reactive cells, and reactive astrogliosis, which reduces spread of secondary injury, but also blocks regeneration of injured axons by forming physical and chemical barriers to axon elongation. In the chronic stage, patients with spinal cord injury (SCI) usually suffer from severe functional deficits largely due to anatomical disconnection and conduction failure of axon tracts between the spinal cord segments rostral and caudal to the lesion.

Figure 1. Pathophysiological events at different stages after SCI.

Researchers have designed various therapeutic strategies to target different pathophysiological changes after SCI. In acute post-injury stage, treatments focus on preventing ongoing direct damage to the spinal cord and neuroprotection to minimize the secondary injury. During the later phase, the strategies usually target neural repair, regeneration, rehabilitation, and complications due to SCI, such as pain, pressure ulcers and infections. Recently, scientists have made tremendous progress in developing therapeutic strategies from basic and preclinical studies. Here, we summarize the major advances in identifying therapeutic approaches to SCI in mammalians, including neuroprotection, CNS regeneration, remyelination and neuroplasticity, by targeting the discovered molecular and cellular mechanisms.

2. Neuroprotective approaches.

Both initial trauma (such as contusion and/or compression) and secondary injury contribute to tissue damage after SCI, including disruption of white matter tracts. Secondary injury dramatically increases the lesion volume several days after SCI and represents an important therapeutic target for neuroprotection. In the early phase (< 48h after SCI), vascular disruption, hemorrhage and ischemia result in microcirculatory failure, loss of ionic homeostasis, excitotoxicity, excessive production of free radicals and inflammatory responses, all of which result in neuronal and glial loss around the lesion ¹. Two days after SCI, these pathological changes disrupt the integrity of blood brain barrier (BBB) and increase its permeability. With help from resident microglia, neutrophils and other immune cells penetrate damaged BBB and enter the lesioned spinal cord, producing cytokines and chemokines that spread the inflammatory response. Neutrophils infiltrate the lesioned spinal cord within the first 24h post-injury. Monocytes penetrate by approximately 48h, and at ~72h, are transformed into macrophages, which secrete both pro-inflammatory and anti-inflammatory factors. Activated microglia and macrophages are beneficial or detrimental largely depending on their subtypes: M1 phenotypes reflect an inflammatory state, while M2 phenotypes are neuroprotective. M1 macrophages exert their detrimental effects by generating pro-inflammation cytokines [e.g., interleukin 6 (IL-6), IL-1β, tumor necrosis factor α (TNFα)] and free radicals (e.g., reactive oxygen and nitrogen species) ². M2 macrophages produce anti-inflammatory cytokines (e.g., IL-4 and IL-13), mitigate the inflammatory responses, clear dead cells and promote wound healing. The predominance of M1 macrophages in the acute stage contributes to the secondary damage. Many neurons around the lesion die largely through necrosis ³, but many oligodendrocytes (OL) along injured white matter tracts die by apoptosis primarily mediated by Fas and p75 receptors, causing subsequent demyelination ⁴. Days to weeks after SCI, the damaged spinal cord gradually transitions from a “destructive” to a “rehabilitative” state. Activated microglia and macrophages clean the cytosolic debris. Reactive glia, especially astrocytes, proliferate and migrate to form scar tissues, which strongly suppress axon regeneration by forming a physical barrier and highly upregulating axon growth inhibitors ⁵.

By targeting the pathological changes during the acute stage, researchers have developed numerous neuroprotective strategies to prevent progressive damage to spinal cord (Fig. 2). Early surgical decompression appears to improve neurological function in both preclinical and clinical studies ⁶. Methylprednisolone (MP) is the only FDA-approved drug for treating SCI. In early preclinical studies, MP appeared to improve neurological recovery after SCI by downregulating inflammatory cytokines (such as TNFα, IL-1, and IL-6) ^{7, 8}. In the clinical trials of national acute SCI studies (NASCIS), an MP regimen beginning 24h post-injury had no effect on long-term neurological recovery when all post-injury time points are considered, but subjects receiving the same MP regimen within 8h of injury showed additional motor recovery compared with controls. Though MP was used extensively for SCI over the past decades, its appropriateness is still controversial. Major concerns about the NASCIS trials include their reliance on the subgroup analysis, limited amount of neurological improvement, and potentially serious adverse events. Later, scientists developed other neuroprotectants by targeting secondary injury mechaniams, including loss of ionic hemostasis (Na⁺, K⁺ and Ca²⁺ channel blockers), excitotoxicity (NMDA and non-NMDA receptor antagonists), oxidative stress, inflammation, mitochandrial dysfunction and protein degradation.

Figure 2. Pathophysiological events after SCI and the major neuroprotective strategies to target them.

Various secondary destructive mechanims contribute to continous damages to the spinal cord hours to weeks after the primary mechanical trauma and thus increase the lesion volume by causing further tissue damages. The strategies to target the secondary injury mechanisms may protect tissues and reduce the lesions.

2.1. Inflammation-based strategies.

Inflammation is a major determinant of consequent neurological outcomes after SCI. Inflammatory cells are thus targets for designing therapies, including regulating their recruitment to the area of injury and the functions of their released chemicals.

2.1.1. Macrophages and their associated cytokines.

Macrophages are the next immune cells present in the lesion area following neutrophil infiltration and they reach peak levels at ~7 d after SCI. They are derived predominantly from myeloid monocytes of circulating blood although some are from native CNS microglia. The blood-derived macrophages phagocytose cellular debris and secrete pro and anti-inflammatory cytokines in the lesion epicenter, while microglia-derived macrophages are distributed around the lesion border to prevent spread of the damage ⁹. Macrophages are classified into M1 and M2 subtypes based on the soluble factors they generate: pro-inflammatory cytokines [Inducible nitric oxide synthase (iNOS), IL-6, TNFα and IL-1β] and reactive oxygen species (ROS) by M1, and anti-inflammatory cytokines (IL-10, TGFβ and IGF-1) by M2 ¹⁰. M1 macrophages generally produce harm by inducing inflammation, neuronal and glial loss, demyelination and regeneration failure, whereas M2 cells protect by reducing pro-inflammatory reactions. Many chemicals, including lipopolysaccharide (LPS) and IFN-γ, stimulate differentiation of Ly6c ^hi CX3CR1^lo monocytes into M1 macrophages, which usually express high levels of MHC II and CD68 marker and are CD68+/CD80+/CD163−/^{low 11}. Some cytokines, including IL-4 and IL-1, stimulate Ly6c ^lo CX3CR1 ^hi monocytes to differentiate into M2 macrophages, the CD68+CD163+IL-12^lowIL-10^highIL-1RA^high cells. The overall M2 population is low after SCI.

The strategies that suppress M1 macrophage activities and enhance M2 macrophage functions are beneficial for tissue repair and recovery. Because the microenvironment around lesions usually favor M1 activation, transplantation of M2 macrophages into lesioned spinal cord preserves neurons, glia and myelin and enhances functional recovery in SCI rodents ¹². Because M1 and M2 are activated by different chemicals (e.g., M1 by IFN-γ, LPS and MCP-1 vs. M2 by IL-4 and IL-10), modifying the macrophage milieu by altering the concentrations of these chemicals in the lesion area may facilitate neural repair. Liposomes, vesicles with a lipid bilayer, have been used to deliver certain chemicals for modifying macrophage functions because macrophages engulf liposome-conjugated chemicals and release them intracellularly. Liposomal prednisolone phosphate suppresses the M1 phenotype, promotes the M2 phenotype, and protects injured tissues ¹³. Deletion of CD36, the most abundant lipid transporter in macrophages, decreases macrophage lipid content and lesion size and improves locomotor recovery in SCI rodents ¹⁴. Systemic delivery of maresin-1, a pro-resolving lipid mediator, also reduces macrophage accumulation and lesion pathology and promotes functional recovery.

The pattern recognition receptors of macrophages are therapeutic targets for SCI. These receptors act as sensors for detecting pathogen, inducing macrophage responses and damaging related molecules. Among them, several toll-like receptors (TLR) induce a protective phenotype in macrophages. TLR4 activation in macrophages by its agonists E6020 and LPS facilitate myelin clearance, iron uptake, axon preservation, OL formation and remyelination in rodent models ¹⁵. Selective TLR2 agonists Pam2CSK4 and Pam3CSK4 reprogram endogenous macrophage responses, diminishes axon dieback, OL loss and lesion area in SCI rodents. Intraocular zymosan, an activator for TLR2 and dectin-1 (a pattern recognition receptor), induces a protective macrophage response and regeneration of injured retinal ganglion cells (RGC). In contrast, inhibition of CX3CR1, a receptor on resident microglia, attenuates inflammatory reactions and promotes neural repair after CNS injuries.

Strategies manipulating macrophage phenotypes show protection in SCI models. Because iron accumulation in macrophages activates M1 and upregulates TNFα, iron chelators promote M1 to M2 transition, tissue repair and functional recovery in SCI rodents ^{16, 17}. Higenamine, a frequently used food ingredient in sports and weight loss dietary supplements, promotes neural repair and functional recovery in SCI mice by activating M2 cells and upregulating IL-4 and IL-10 ¹⁸. Azithromycin, a commonly used macrolide antibiotic, improves locomotor recovery by inactivating pro-inflammatory macrophages in SCI mice ¹⁹. Moreover, blocking chemokine receptor 5 (CCR5), a macrophage receptor, reduces CD11b+ macrophages, enhances M2 marker arginase 1, and promotes locomotor recovery in SCI mice ²⁰.

Blocking the major cytokines generated by M1 macrophages is also neuroprotective after SCI. Because TNFα contributes to neuronal and glial loss after SCI, its antagonists enhance recovery after SCI. Etanercept, a recombinant dimer of TNFα receptor, functions as a TNFα antagonist and promotes functional recovery after SCI in both preclinical ²¹ and clinical ²² studies although there is no such an effect if treatment is delayed for 14 d after SCI ²³. Epigallocatechin gallate, an abundant catechin in green tea, prevents tissue damage after CNS injury by altering the levels of TNFα and other cytokines ²⁴. Suppressing several ILs also reduces inflammation and tissue damage and promotes neural tissue repair after CNS injury. An antibody against IL-6, a product of M1 macrophages, increased the number of M2 macrophages around the lesion, spared myelin and improved recovery of locomotor functions ²⁵. Microglia generate IL-1α several hours after SCI, thus recruiting leukocytes, producing IL-1β and amplifying post-injury inflammation. Intrathecal anakinra, a recombinant IL-1-R1 antagonist, promotes functional recovery in SCI rodents. IL-1α inhibition also protects OLs after SCI by upregulating transcriptional factor Tox3. IL-33, another member of the IL-1 family, and its associated chemokines are also therapeutic targets in SCI. IL-33 treatment promotes an M2 macrophage response, prevents demyelination, reduces lesion size and improves locomotor function in SCI rodents. Injured OLs could secrete IL-33 in SCI animals, which upregulates chemotactic proteins (such as CCL2, CCL5, CXCL1, CXCL2 and CXCL10) in astrocytes and microglia.

2.1.2. Inflammation mediating micro-RNAs (miRNA or miR).

MiRNAs regulate inflammatory pathology after SCI and are also therapeutic targets. MiRNAs, the small non-coding RNAs, control post-transcriptional gene expression by regulating mRNA stability or translation. SCI alters the levels of numerous miRNA-targeting proteins that regulate inflammation and apoptosis ²⁶. The serum level of miR-223 is a biomarker for neutrophil infiltration in mice with acute SCI. Some miRNAs (e.g., miR-155, miR-136-5p and miR-99b-5p) are harmful while others (miR-124, miR-136, miR-137, miR-30a-5p and miR-146a) are beneficial for recovery. MiR-155 enhances inflammatory reactions and its deletion reduces macrophage mediated damage and promotes axon regeneration and functional recovery in SCI rodents ²⁷. Upregulating miR-136-5p raises the levels of various inflammatory factors (NF-κB, MIP2, IL-6, MCP-1, TNFα and IL-1β) and increases tissue damage and its suppression prevents these detrimental effects. In contrast, miR-124, miR-137 and miR-30a-5p suppress activities of various inflammatory factors and are neuroprotective to lesioned spinal cord.

2.1.3. Neutrophil inhibition and intravenous immunoglobulin G (IVIG).

Neutrophils, the most abundant type of granulocyte, are the first immune cells recruited into the lesion epicenter after SCI. Chemo-attractants, such as IL-8, IFNγ and other activators (e.g., CXCL12-CXCR4 and CXCL1/2-CXCR2), attract neutrophils across the disrupted BBB and into the lesioned spinal cord. Infiltrated neutrophils create a deleterious environment by releasing pro-inflammatory cytokines, free radicals and proteases, but also have beneficial effects by phagocytosing cellular debris and creating a favorable condition for regeneration. Suppressing neutrophil recruitment and activities reduces immune reactions, alleviates apoptosis and oxidative stress, and improves motor function. However, depleting neutrophils by targeting Ly6G, a marker for granulocytes, using a selective antibody shows both positive and negative effects after SCI: inhibiting astrogliosis and augmenting effects of transplanted neural stem cells ²⁸, but worsening functional outcome. Neutrophils and astrocytes promote neural repair partly by secreting leukocyte protease inhibitor ²⁹. Thus, it appears important to design strategies to fine tune the functions of infiltrated leukocytes. IVIG, originally used to treat various inflammatory/autoimmune diseases by antibody replacement, improves recovery in SCI models by targeting detrimental inflammatory response within the lesioned spinal cord ³⁰. IVIG reduces activation of complement products, especially C5a and C3b/iC3b, which boost secondary inflammatory reactions after SCI ³¹.

2.2. Mitochondrial damage and biogenesis.

CNS injuries disrupt mitochondrial functions and ATP-dependent cellular homeostasis (e.g., calcium overload, excitotoxicity and oxidative stress) around the lesion. Neurons normally consume a lot of energy and are very sensitive to oxidative stress, and their high metabolic function heavily relies on ATP generation by mitochondria ³². After acute SCI, neurons depolarize, opening membrane ion channels, thus accumulating intracellular Ca²⁺ and releasing intracellular glutamate ³³. Although mitochondria and other organelles buffer excess Ca²⁺ under physiological condition ³⁴, after SCI, the overwhelming Ca²⁺ influx opens the mitochondrial permeability transition pores, impairs ATP synthesis, ruptures mitochondrial outer membrane, and releases Ca²⁺, ROS and pro-apoptotic proteins into the cytoplasm ^{35, 36}, eventually damaging the mitochondria and their host cells ³⁷.

Repairing damaging mitochondria is potentially important for treating acute SCI and several agents have been employed to maintain mitochondrial function. Immunosuppressant cyclosporin A (CsA) inhibits activation of mitochondrial permeability transition pores in CNS neurons ⁵ and shows neuroprotection in brain injury and stroke models ^{3, 38}, but its therapeutic potential for SCI remains controversial largely due to its high toxicity ^{39, 40}. The less toxic derivative of CsA, NIM811, has neuroprotective effects in animal models of SCI ⁴¹. Treatment with acetyl-L carnitine, a mitochondrial inner membrane component, increases the number of healthy mitochondria and their membrane potential, and alleviates SCI-induced apoptosis in rats ⁴². Acetyl-L carnitine provides the acetyl group for synthesizing acetyl-CoA, a necessary chemical in the citric acid cycle. Another compound N-acetylcysteine amide, a variant of glutathione precursor (an FDA-approved drug), when applied immediately after SCI, increases mitochondrial level and functional recovery in SCI rodents ⁴³. Glutathione is an intracellular thiol compound that protects against damage from ROS, lipid hydroperoxides and electrophiles ⁴⁴.

Mitochondrial division inhibitor 1 (Mdivi-1), a selective inhibitor of dynamin related protein 1 (Drp1), promotes recovery in SCI animals by enhancing mitochondrial membrane potential and ATP production, and reducing cell apoptosis ⁴⁵. Mitochondrial fission and fusion are required to replace damaged mitochondria in post-mitotic cells (including neurons) and several genes are critical for mediating this process, including proteins involved in mitochondrial fusion (mitofusin 1 and 2) and fission (mitochondrial fission 1 and Drp1). SCI causes dysfunctional mitochondrial fission and fusion by altered expression of these proteins, including upregulation of Drp1 12-24 hrs after SCI ^{46, 47}. Because increased Drp1 contributes to apoptotic cell loss by reducing mitochondrial membrane potential, releasing cytochrome c and upregulating caspase 3, inhibition of Drp1 promotes neural repair after SCI.

Strategies to promote mitochondrial biogenesis appear promising for treating SCI. Systemic treatment with formoterol, a selective and potent β2-adrenoreceptor agonist approved by FDA for mitochondrial biogenesis, increases white and gray matter sparing and locomotor recovery in SCI mice ⁴⁸. It also would be interesting to evaluate other reagents that promote mitochondrial biogenesis in SCI models, e.g., 5-HT receptor agonists, or resveratrol, a phytoalexin produced by several plants ^49,50. Exogenous mitochondria have been transplanted into injured rat spinal cord, but failed to show neuroprotection and improved recovery ⁵¹.

2.3. Autophagy.

Autophagy is crucial in maintaining cellular homeostasis by degrading and recycling damaged organelles, unfavorable proteins and toxins via an autophagosome-lysosome fusion mechanism ⁵². Generally, autophagy is neuroprotective and favors functional recovery after CNS injuries ⁵³ although excessive activation of autophagic machinery may cause an unusual form of cell death called autosis. The overall role of autophagy in mediating outcomes after SCI is controversial largely depending on the types of injuries. Autophagy is impaired after severe SCI, but enhanced in moderate injury ^54,55. Moderate SCI activates autophagic pathways generally and promotes recovery by removing damaged organelles and toxins in injured cells ⁵⁶.

Enhancing autophagy seems neuroprotective in general ⁵⁷, but drug therapies to target autophagy have yielded inconsistent results ⁵⁸. The mTOR inhibitor rapamycin activates autophagy by increasing autophagosome formation ⁵⁵, but has different effects on SCI recovery. Subcutaneous rapamycin at a high dosage exhibits detrimental outcome of SCI rodents while its intraperitoneal application shows neuroprotection ⁵⁹. Consistently, over-activation of autophagy by locally applying rapamycin induces cell loss by apoptosis ⁶⁰. On the other hand, a recent study indicates that autophagy suppresses necrosis because a lysosomal inhibitor increases the levels of necrotic proteins RIPK1 and RIPK3 ⁶¹. In addition, OLs require moderate autophagy for their survival after SCI because blocking autophagy in OLs by deleting Atg5, an autophagy-required protein that complexes with Atg12, impairs locomotor recovery in SCI rodents ⁶².

2.4. Other neuroprotectants.

2.4.1. Neurosteroids.

Neurosteroids are steroid hormones synthesized in brain or transported to CNS through the BBB. Neurosteroid progesterone, which is in phase II/III clinical trials for brain injury, promotes recovery after SCI by preserving white matter, upregulating BDNF, increasing NG2+ cells and inactivating astrocytes and microglia ⁶³. Long term treatment with progesterone at a high dose further improves locomotor functions by reducing damaged spinal cord ⁶⁴. Moreover, progesterone suppresses the development of allodynia by downregulating pro-inflammatory factors COX-2 and iNOS ⁶⁵. Consistently, dehydroepiandrosteron, another neurosteroid, protects the injured spinal cord in rodents.

2.4.2. Agents to target the microbiome.

Many pathological conditions, including SCI, alter the composition of the gut microbiome. SCI increases intestinal permeability and bacterial translocation from the gut, activating immune cells in the gut-associated lymphoid tissues ⁶⁶. Forced dysbiosis (i.e., microbial imbalance) before SCI by oral delivery of broad spectrum antibiotics worsened primary spinal cord damage and functional recovery ⁶⁷. Melatonin, a neurohormonal peptide, promoted SCI recovery probably by improving intestinal barrier function, altering composition of gut microbiota and relieving inflammation and oxidative stress ^{68, 69}.

2.4.3. Ketogenic diet.

Ketogenic diet (KD) includes high fat (80-90%), low carbohydrate and appropriate protein consumption and is able to increase ketone body (i.e., β-hydroxybutyrate) production ⁷⁰. KD was originally used to treat drug resistant epilepsy patients ⁷¹, but has also been shown neuroprotection in many neurological models, including SCI ⁷². In clinical trials for SCI, KD increased the levels of ketone bodies and improved motor functions with minimal side effects ⁷³. The mechanisms underlying KD treatment may include increased oxidative metabolism, subsequent ATP synthesis in mitochondria and reduced ROS formation ⁷⁴. Ketone bodies reduce oxidative stress probably by inhibiting histone deacetylase (HDAC) and altering transcription of its related genes after SCI ⁷⁵.

3. Neural repair and regeneration.

Most SCI patients suffer from life-long functional deficits largely due to disconnections of whiter matter tracts between the rostral and caudal spinal cord though complete spinal cord transection is clinically uncommon. Biological and functional repair of SCI requires axon regrowth, reconnections onto caudal/rostral neurons and formation of new neural circuits to restore signal conduction ⁷⁶. Reduced intrinsic growth capacity of mature neurons and a no-permissive environment around the lesion contribute most to failure of axon regeneration and repair in the CNS of adult mammals ⁷⁷. Most therapeutic strategies developed recently target these mechanisms accordingly (Fig. 3).

Figure 3. Major molecular mechanisms and strategies for spinal cord regeneration and repair.

Summary of various major molecular strategies for axon regeneration and neural repair after SCI. Some of them primarily target intrinsic neuronal growth capacity in adult CNS while others aim to repair lesioned spinal cord locally. Acutally, many molecular therapties involve cell growth programs at both spinal cord and supraspinal levels to promote regeneration, repair and neuroplasticity.

3.1. Enhance neuronal growth capacity.

During development, differentiated neurons establish axon growth programs, extend axons to their cellular targets and form synaptic connections with proper cellular targets. Mature CNS neurons have reduced growth capacity, as observed in cultured RGCs ⁷⁸, and their axons fail to regrow after axotomy. In recent years, many molecules have been reported to regulate axon growth ability in mature neurons by targeting numerous signaling pathways and mechanisms, thus becoming molecular targets for neural repair after CNS injuries, including SCI.

3.1.1. PI3K and mTOR pathway.

During development, the PI3K/Akt pathway is essential for regulating axon formation and extension ⁷⁹. Because PTEN suppresses Akt activity by dephosphorylating phosphoinositide substrates, its inhibition should increase neuronal growth by activating PI3K/Akt signaling. Accordingly, using transgenic and pharmacological approaches, many groups confirm the role of PTEN in restricting regeneration of axons in multiple types of mature CNS neurons. PTEN deletion in mice activates the Akt pathway and stimulates dramatic regeneration of α-RGCs and corticospinal tract (CST) axons ⁸⁰. Combined PTEN and Socs3 deletion plus ciliary neurotrophic factor (CNTF) treatment induces greater regeneration than individual approaches. Systemic treatment with selective PTEN antagonist peptides initiated 2 d after injury promotes regrowth of descending motor axons in the caudal spinal cord of mice and recovery of locomotor function several weeks after SCI by increasing locomotor scores and stride length of hindlimbs and reducing grid walk errors ⁸¹. Post-injury treatment with bisperoxovanadium, a general inhibitor of PTEN and other protein tyrosine phosphatases (PTP), exhibits beneficial effects on recovery after CNS injuries ⁸². Several signals have been shown to mediate axon growth downstream of PTEN. Among them, ribosomal S6 kinase, however, was reported to inhibit axon growth because its selective inhibitor enhanced axon outgrowth into and beyond the lesion and improved function recovery after CST axon transection ⁸³. Previously, this kinase was considered a positive regulator for axon elongation.

3.1.2. Liver kinase B1 (LKB1) and adenosine monophosphate-activated kinase (AMPK) pathway.

LKB1 is a master kinase to target multiple other enzymes, including AMPK and AMPK-related protein kinases. LBK1 is critical for mediating tumor suppression, cell growth, metabolism and cell polarity, including axon formation and branching during development. Activating LKB1 is a highly effective strategy for promoting CNS axon regeneration. Overexpressing LKB1 by AAV2 vector in sensorimotor cortex promotes dramatic regeneration of CST axons in adult mice with SCI ⁸⁴. Upregulating LKB1 specifically in neurons by systemic injection of a mutant-AAV9 vector stimulates long distance regeneration of injured CST fibers into caudal spinal cord in adult mice and regrowth of descending serotonergic and tyrosine hydroxylase immunoreactive axons. Either intracortical or systemic viral delivery of LKB1 promotes recovery of locomotor functions in SCI mice. Further mechanistic experiments demonstrate AMPKα, NUAK1 and ERK as the major downstream effectors of LKB1 in adult CNS. Thus, LKB 1 is an important therapeutic molecular target for SCI.

3.1.3. Transcriptional factors (TF).

Manipulating a number of TFs, including ATF3, cAMP, c-myc, HIF-1α, c-Jun, Kruppel-like factors (KLFs), Smad1, Sox11, STAT3, arginase 1 and growth-associated proteins (GAP-43, CAP-23 and SPRR1a), alters expression of regeneration associated genes and promotes regeneration of mature neurons ⁸⁵. Some KLFs (e.g., KLF6/7), stimulate axon regeneration, while others (KLF4/9) impede it. Genetic knockdown of KLF4 and 9 enhances axon regeneration after optic nerve injury (ONI) in adult mice. Proto-oncogene c-myc activates a number of genes for cell growth and is downregulated after ONI. C-myc upregulation stimulates RGC regeneration after injury and further increases the distance of regenerated axons when combined with exogenous CNTF and co-deletion of PTEN and Socs3. These combinations stimulate injured optic axons to pass optic chiasm and enter the ipsi- and contralateral optic tracts ⁸⁶.

Dual leucine zipper kinase (DLK) is part of the conserved MAPK pathway and forms heterodimers with leucine zipper containing transcription factors, such as CREB and myc. PKA, the cAMP effector kinase, can directly activate DLK ⁸⁷. Axotomy results in Ca influx, disrupts microtubules (MT), and upregulates and activates axonal DLK, which triggers multiple transcriptional responses and downstream effectors. By interacting with JNK1, DLK also regulates MT stability, which is necessary for axon regeneration. Deleting DLK increases RGC survival, but blocks optic nerve regeneration induced by PTEN deletion, suggesting the complicated roles of DLK in mature CNS neurons. Fine tuning DLK activity may become an effective approach for regenerating injured CNS axons.

Recent studies support the critical roles of several other TFs in regulating regrowth of CNS neurons in spite of incomplete understanding of their neuronal functions. STAT3 activation promotes optic nerve regeneration probably by regulating transcription of growth associated genes and cytoskeletal proteins ⁸⁸. STAT3 binds and blocks MT-destabilizing protein stathmin and stimulates MT elongation in neurons. Co-deleting PTEN and SOCS3, the STAT3 inhibitor, displays synergy on the number and distance of regenerating axons after ONI. SOX11, another TF that promotes both PNS and CNS regeneration, reprograms adult non-α-RGCs and CST neurons to a growth-capable state ^{89, 90}, but its upregulation promotes death of α-RGCs, the subtype that survives ONI in wildtype mice and regenerates its axon after PTEN deletion. Notably, the TFs that promote neuronal growths may interact with each other; e.g., p53 binds STAT3, KLF6, Myc, ATF3, CREB, HIF1A and SMAD1, and STAT3 binds KLF4, ATF3, SMAD1, p53 and HIF1A. KLF4 suppresses axon regrowth probably by interacting with and inhibiting pro-regenerative STAT3.

3.1.4. Transmembrane and membrane-bound proteins.

Activating lysophosphatidic acid receptor 1 by its ligand lysophosphatidic acid suppresses axon growth. In contrast, activating lipid phosphate phosphatase-related protein type 1 (LPPR1), which also interacts with lysophosphatidic acid, stimulates axon growth and sprouting of intact CSTs into the denervation side of the spinal cord after unilateral pyramidotomy ⁹¹. Deleting Rab27b, a membrane protein that binds Rab GTP and regulates axon transport, significantly promotes RGC regeneration after ONI, and sprouting of raphespinal tracts and recovery of motor function after spinal cord transection ⁹². The Alpha2delta2 subunit of voltage-gated calcium channels, which is encoded by Cacna2d2 gene, suppresses axon outgrowth. Deletion or pharmacological inhibition of Alpha2delta2 promoted neurite extension in vitro and axon regeneration in adult mice with SCI ⁹³.

3.1.5. Inflammation related regenerative strategies.

Targeting innate immune responses is also promising for developing therapeutic regenerative strategies. Inflammation associated signals alter intrinsic growth capability at least in some types of mature neurons. Intravitreal zymosan (a macrophage activator) or lens injury induces regeneration of injured optic nerve, which is probably mediated by activating dectin-1, a receptor expressed on resident microglia and infiltrated leukocytes. Consistently, intravitreal β-glucan, a dectin-1 ligand, promotes optic nerve regeneration in adult mice ⁹⁴. Several molecules derived from activated immune cells also increase the growth ability of adult CNS neurons. Oncomodulin, a calcium-binding protein secreted by macrophages, enhances regrowth of injured RGCs. Certain cytokines generated around the lesion, e.g., CNTF and LIF, boost axon regeneration after CNS injury. CNTF and LIF probably mediate optic nerve regeneration after lens injury because co-deleting both in mice abolished the axon growth promoting effects ⁹⁵. The JAK/STAT pathway, which is activated by multiple immune cytokines and critical for mediating inflammatory responses, promotes regeneration of mature CNS neurons. Activation of JAK signaling by applying cytokines IL-6 and CNTF and deleting Socs3 enhanced regeneration of injured CNS axons ⁹⁶.

3.1.6. Cytoskeletal proteins and mitochondrial transport.

Several studies emphasize the therapeutic potential of targeting cytoskeletal and mitochondrial molecules for CNS lesions. The extracellular and intracellular molecules that regulate cell growth eventually converge on cytoskeletal proteins actin and MT. Several MT stabilizing compounds or severing proteins mediate CNS axon regrowth. Fidgetin, a MT-severing protein, inhibits axon outgrowth by preventing elongation of MT labile domains. Its deletion or knockdown overcomes growth suppression by chondroitin sulfate proteoglycans (CSPGs) and promotes injured sensory axons to cross the dorsal root entry zone into spinal cord ⁹⁷. Inhibiting non-muscle myosin II facilitates reorganization of actin and MTs in growing axons and enables them to extend over the inhibitory substrates (CSPGs and myelin inhibitors) in vitro ⁹⁸, and may promote axon regeneration after SCI. Applying taxol (an FDA-approved anti-cancer drug) to the lesion area improves spinal cord axon regeneration and functional recovery by reducing TGFβ signaling and scar formation. Systemic administration of epothilone B (another clinical anti-cancer drug) promotes axon elongation and functional recovery by inducing concerted MT polymerization into axon tip and suppressing scar formation by abolishing polarization and migration of scar-forming fibroblasts. Because axon regrowth requires energy generated locally, mitochondrial transport proteins are also targets for regeneration. Overexpression of Armcx1 improves mitochondrial transport, survival and regeneration of RGCs in adult mice ⁹⁹. HDAC6-dependent deacetylation of MIRO1 (mitochondrial Rho GTPase 1) inhibits axon growth in vitro by decreasing mitochondrial transport ¹⁰⁰.

3.1.7. Combined regenerative approaches.

Because many molecules contribute to failure of CNS regeneration through different mechanisms, the combination strategies to target multiple factors should show synergy for promoting CNS regeneration. Co-expression of osteopontin and IGF-1/BDNF by viral vectors mimic the effects of PTEN deletion. Overexpression of osteopontin, IGF-1 and CNTF further enhance axon regeneration ¹⁰¹. Co-deleting PTEN and SOCS3 or PTEN deletion combined with oncomodulin or zymosan (inflammatory factors) and cell permeable cAMP induced robust regeneration of injured optic nerve although only a small number of regenerating axons reached the chiasm ¹⁰². Application of zymosan and cAMP to PTEN-deleted mice promoted dramatic growth of optic axons and a small number reached the lateral geniculate nucleus and superior colliculus to promote partial recovery of vision ¹⁰³, although another group failed to reproduce these results ¹⁰⁴.

3.2. Target extrinsic factors and their downstream signals.

Many inhibitory components around the lesion are important therapeutic targets for CNS regeneration. After SCI, the local environment around the lesion is generally non-permissive for axon regrowth, in contrast to that during development. Embryonic glial processes and adjacent growing axons guide developing axons, but injured axons in adult spinal cord typically encounter various barriers preventing elongation, including damaging inflammatory responses, physical glial and fibroblast scar tissues, lesion cavities and various inhibitory cues produced by reactive glia, OLs and other cells. In acute and subacute stages, huge numbers of macrophages, activated microglia and other inflammatory cells inhabit the lesioned spinal cord and generate numerous cytokines and chemokines. The injury pathology results in cavitation inside and around the lesion, activates migration and proliferation of reactive astrocytes, fibroblasts and other cells, eventually generating sturdy scar tissue. Importantly, reactive astrocytes and other cells highly upregulate scar-sourced inhibitors, especially CSPGs around the lesion, and potently block axon regrowth. Adult OLs and CNS myelin express several axon growth-inhibitors, including Nogo, MGA and OMgp, which further hamper regeneration of injured spinal axons. Furthermore, glial cells around lesion produce repulsive axon guidance cues, including ephrin B3 and A4, semaphorin 3A and 5A and Wnts.

3.2.1. CSPG receptors and scar tissues as therapeutic targets.

Overcoming potent inhibition of axon regrowth by scar-sourced inhibitors is an important goal for CNS regeneration. After SCI, scar tissues form both physical and chemical barriers of axon regeneration. Particularly, upregulation of CSPGs strongly suppresses axon regeneration and neural repair. The main CSPGs expressed in the CNS comprise lecticans (neurocan, versican, aggrecan and brevican), phosphacans and NG2. Sulfation pattern of GAG chains is essential for CSPG suppression and digestion of the glycosaminoglycan side chains by locally applied chondroitinase ABC has been frequently used to promote axon regeneration and sprouting and recovery of various functions after SCI ^{105, 106}. OL precursor cells (OPC) express high levels of NG2 and antibody perturbation of NG2 improves axon regrowth and locomotor recovery ¹⁰⁷.

CSPG receptor antagonism is promising for treating SCI. CSPGs inhibit neuronal growth by several mechanisms, including interactions with specific functional receptors on neuronal membrane, formation of non-permissive perineuronal nets that limit synaptogenesis, and support of some chemo-repulsive cues. LAR subfamily transmembrane receptors, PTPσ and LAR phosphatase, bind CSPGs with high affinity and mediate CSPG suppression of axon growth ^{108, 109}. CSPGs also bind two receptors for myelin-associated inhibitors, Nogo receptor (NgR) 1 and 3 ¹¹⁰. Blocking LAR with systemic application of selective antagonist extracellular or intracellular peptides promotes regrowth of descending 5-HT fibers in spinal cord 5-7 mm caudal to the lesion and also locomotor functional recovery ¹⁰⁹. Furthermore, systemic treatment with an intracellular PTPσ peptide enhances regrowth of serotonin axons and recovery of locomotor and bladder functions in rats with contusion SCI. Notably, PTPσ has complicated functions regulating activities of multiple cell types, including OLs and immune cells ^{111, 112}.

It is also important to design therapies targeting the molecules essential for controlling astrogliosis. Because scar tissues highly suppress axon extension after CNS injury ^113-115, intervening scar formation should facilitate injured axons to regrow for neural repair. After SCI, upregulation of N-cadherin increases reactive astrocyte adhesion to strengthen intercellular contacts ¹¹⁶. Perturbation of either N-cadherin or β1 integrin represses N-cadherin expression and the conversion of reactive astrocytes into scar astrocytes. Anti-β1 and anti-N-cadherin antibodies applied 9-13 d after SCI enhance the numbers of descending serotonin and tyrosine hydroxylase fibers in the caudal spinal cord and support behavioral recovery in SCI mice ¹¹⁶. In addition, multiple cells in the scar tissues, including meningeal cells, pericytes and fibroblasts, upregulate semaphorin 3, a major inhibitory axon guidance cue during development, and its selective inhibitor SM-216289 preserves injured axons and enhances regeneration, myelination and functional recovery ¹¹⁷.

3.2.2. CNS myelin inhibitors and their receptors.

Suppressing signaling pathway of myelin associated inhibitors and their receptors are effective for enhancing CNS axon regeneration and sprouting. OLs express several axon growth inhibitors, especially Nogo-A, MAG, OMgp and netrin-1, which are present in the CNS myelin. The former 3 proteins suppress axon elongation by binding NgR1, which signal is conveyed by several co-receptors, including p75, leucine-rich repeat protein Lingo1 and/or Troy. Selective antagonists targeting Nogo A or NgR1 enhance regrowth of CST and 5-HT axons and functional recovery in adult rodents with SCI ^118-120. Treatment with IN-1, an antibody against NogoA, also promotes axon sprouting, regeneration and neuroplasticity in rodents and primates with SCI or other CNS injuries ¹²¹. IN-1 is currently in phase III clinical trials for SCI. Notably, a recent study shows that CNS myelin also contains the components that promote axon growth by interacting with neuronal growth regulator 1 generated from neural progenitor cells ¹²².

3.2.3. Convergent signals downstream of axon growth inhibitors.

The extrinsic axon growth factors usually regulate cell growth by converging on shared intracellular signaling pathways (e.g., RhoA, Akt and GSK-3β), which are targets for neural repair. Both CSPGs and myelin inhibitors activate RhoA and Rho associated kinase, which trigger actin de-polymerization and axon growth cone collapse. Treatments with Rho inhibitors, including C3 transferase and certain non-steroidal anti-inflammatory drugs, promote axon regrowth and behavioral recovery in SCI rodents ^{123, 124}. A phase I/IIa clinical trial of a RhoA inhibitor VX-210 (called BA-210 or Cethrin) shows a good safety profile and possibly increased motor recovery ¹²⁵. The phase IIb/III trials of VX-210 are ongoing to evaluate its efficacy in SCI patients. CSPG inhibition on neuronal growth is partly mediated by GSK-3β and GSK-3β inhibitors, particularly lithium, are beneficial after CNS injuries ^{109, 126}. Lithium was in phase I/II clinical trials to evaluate its efficacy on chronic SCI patients ¹²⁷.

3.3. Epigenetic factors.

Epigenetic regulation of transcription and expression of heritable DNAs includes modifying DNAs and their wrapping histone proteins, ATP-dependent chromatin remodeling and non-coding RNAs. Some of these regulators may become therapeutic targets for neural repair after SCI.

3.3.1. Histone modifications.

Both histone acetyltransferases (HATs) and histone deacetylases (HDACs) are molecular targets for CNS axon regeneration. Post-translational histone modifications regulate gene expression and various cellular functions, including cell growth programs. Histone acetylation by HATs and deacetylation by HDACs are important to regulate gene transcriptions. HATs add acetyl groups to lysines and result in open chromatin for gene transcription while HDACs condense chromatin and silence genes by eliminating acetyl groups. Several HATs, including p300, CREB-binding protein (CBP) and P300/CBP-associated factor, regulate neurite outgrowth dependently or independently of p53. Overexpression of P300 enhances H3K18 acetylation and the number of regenerating axons after ONI in rodents. Systemic treatment with the CBP activator TTK21 conjugated to carbon nanospheres to cross the BBB enhances regeneration and sprouting of sensory and serotonin axons and functional recovery in SCI rodents ¹²⁸. Both general HDAC1/2 inhibitor Trichostatin A and HDAC1 selective inhibitor MS-275 promote neurite outgrowth in mouse dorsal root ganglion (DRG) cultures. Treatments with CI-994 (a class I HDAC inhibitor), RGFP966 (a HDAC3 inhibitor) and valproic acid (a general HDAC inhibitor) promote tissue repair and functional recovery after SCI ¹²⁹, ¹³⁰.

3.3.2. DNA methylation.

DNA methylation directly regulates gene transcription and expression and its regulators are targets for axon regeneration. DNA methyltransferases include three groups based on the chemical reactions they catalyze: m6A, m4C and m5C, all transferring methyl groups to DNA. Mammalian maintenance methyltransferases include Dnmt1, Dnmt3b and Dnmt3c, which methylate DNA when one strand is already methylated. Because DNA hyper-methylation represses gene expression, DNA demethylation generally reprograms mature mammalian neurons and promotes axon regrowth. Ten-eleven translocation methylcytosine dioxygenases (Tets) removes DNA methylation independent of cell cycle by oxidizing 5-methylcytosine to 5-hydroxymethylcytosine or other forms ¹³¹. Tet1 contributes to RGC regeneration due to PTEN deletion. DNA demethylation by Tet3 is necessary for functional axon regeneration of adult DRGs by upregulating regeneration associated genes ¹³². DNA methylation in DRG contributes to nerve injury induced chronic pain ¹³³ and may become a target for treating neuropathic pain after SCI. Moreover, methylation of histones modifies chromatin structure, gene expression and cell growth.

Enzymes for acetylation or methylation may regulate axon growth also by targeting other molecules, such as cytoskeletal proteins. HDAC class I and HDAC6 could deacetylate α-tubulin and p300 could acetylate α-tubulin, in addition to KLF4. Interactions between Dnmts and UHRF1 (ubiquitin-like containing PHD ring finger 1) mediate regeneration of injured PNS axons by suppressing PTEN and the transcription regulator REST ¹³⁴. Furthermore, retrograde transported BMP and Smad signals along axons regulate neuronal regeneration by targeting multiple factors, including gene transcription and cytoskeletal dynamics ¹³⁵.

3.3.3. MiRNAs mediated growth.

MiRNAs may also become therapeutic targets to promote axon regeneration. MiRNAs are highly abundant in nervous system and associated with multiple neural functions during development, including axon elongation, branching and guidance. Many miRNAs are known to contribute to axon regeneration, including let-7, miR-21, miR-26a, miR-34a, miR-29c, miR-132 and miR-222. Let-7 negatively regulates axon regeneration in anterior ventral MT neurons in C. elegans by interacting with its upstream and downstream regulators lin28 and lin41 ¹³⁶. Lin28 expression promotes axon regeneration of both DRGs and RGCs after injury at least partly mediated by Akt/mTOR pathway ¹³⁷. Lin28 overexpression in amacrine cells significantly enhances spontaneous RGC activity and RGC regeneration induced by IGF-1 ¹³⁸. Intravitreal miR-135s stimulates RGC regeneration in adult mice with ONI partly by blocking KLF4 ¹³⁹. MiR-124 contributes to recovery after SCI by targeting neurotrophic signaling pathway, microglial activation and differentiation of implanted stem cells ^{140, 141}. Treatment with docosahexaenoic acid, an omega-3 fatty acid, supports functional recovery by increasing miR-21 expression and suppressing PTEN activity ¹⁴². MiR-26a-GSK3β-Smad1 signaling pathway regulates DRG axon regeneration ¹⁴³. In addition, it would be interesting to dissect potential roles of other non-coding RNAs (e.g., PIWI interacting RNAs and long non-coding RNAs) in neuronal growth ¹⁴⁴.

3.3.4. Nuclear receptors.

Several nuclear receptors, including retinoic acid receptors (RAR), peroxisome proliferator-activated receptor (PPAR) and retinoid X receptor (RXR), are also therapeutic targets for treating SCI and other CNS disorders. Retinoids play important roles in neurodevelopment by binding RAR and PPAR and regulating gene transcription, lipid metabolism and inflammation. RAR and PPAR act as ligand-activated transcription factors and heterodimerize with RXR to form active retinoid complex. Upregulation of RARbeta2 enhances regeneration of both sensory and motor neurons in rodents with SCI ^{145, 146}. Several molecules that interact with these receptors, including polyunsaturated fatty acids (such as omega-3), lipid vitamins and PPAR agonists, are also promising for treating CNS disorders by interacting with RAR/RXR and PPAR/RXR signals. Omega-3 and Maresin 1, a docosahexaenoic acid derivative, protected injured spinal cord and improved functional recovery largely by suppressing inflammatory responses ^{142, 147, 148}.

3.4. Interneuron mediated neuroplasticity.

Reorganizing local inter-neuronal networks is a very promising approach to promote recovery after SCI. Both animals and patients exhibit spontaneous functional recovery after incomplete SCI, which is at least partly due to reorganization of propriospinal neuron (PSN) networks ^{149, 150}. PSN relay connections bypass the lesion and form detour neuronal circuits in rodents after hemi-transection SCI. The unique structures of the spinal cord with both gray and matters appear especially appropriate for such relays. Using different approaches (e.g., molecular modification of interneuron activities, electrical or magnetic stimulation and rehabilitation), many groups have reported reorganization of different axonal tracts (e.g., CST, reticulospinal, rubrospinal, serotonergic and sensory axons) with local interneurons and improved functional recovery after SCI.

CST axons are important for controlling voluntary movements, but are high refractory to regeneration after injury. It is thus important to reconnect this pathway by facilitating collateral sprouting to innervate relays by interneurons and axonal pruning (i.e., eliminating unnecessary collaterals). Neuropilin-1, a receptor for Sema 3A, mediates CST pruning and functional recovery after SCI. Its deletion or knockdown prevents locomotor recovery after SCI¹⁵¹. Blocking Ryk, a Wnt receptor upregulated after SCI, with antibodies enhances locomotor recovery after incomplete SCI probably by increasing CST sprouting and PNS-mediated neuroplasticity ¹⁵². In addition, a recent study indicated that combination approaches stimulated PNS axons to pass into the spared neural tissue, exhibit synaptic structures and convey electrophysiological conduction across the lesions ⁷⁷. Particularly, enhancing growth capacity using osteopontin, IGF1 and ciliary-derived neurotrophic factor before SCI combined with various growth factors (fibroblast growth factor 2, epidermal growth factor and glial-derived neurotrophic factor) delivered after SCI promoted robust regrowth of PSN axons through astrocytic scar borders. Also, viral delivery of neurotrophin-3 to demyelinated sciatic nerve that retrogradely transports to lumbar spinal motor neurons promotes sprouting and synaptic interactions of descending serotonergic, dopaminergic and PSNs on lumbar motor neurons and behavioral recovery in SCI rodents ¹⁵³.

Reducing inhibitory interneuron activity by activating KCC2, a potassium-chloride co-transporter, also stimulates functional recovery in a mouse model of staggered bilateral hemi-sections. Spinal inhibitory interneurons appear to restrict integration of descending inputs into relay circuits after SCI and regulating their functions may promote functional recovery. KCC2 overexpression in inhibitory neurons or activation by its agonist CLP290 reduces excessive inhibitory signaling and promotes activity of PSNs and functional recovery in SCI mice ¹⁵⁴.

3.5. Remyelination after SCI.

Remyelinating demyelinated and/or regenerated axons is an important goal for successful neural repair after SCI. A great number of OLs die through apoptosis several hours to weeks after SCI (peak level: 3-7 d after SCI). OLs are highly vulnerable to various pathological conditions, such as ischemia, oxidative damage, inflammation and excitotoxicity. Though SCI may trigger spontaneous myelin repair process by OPCs, OLs and migrated Schwann cells from PNS, demyelination occurs after SCI and partly contributes to signal conduction failures around lesion and functional deficits. Remyelination requires OPC activation, recruitment, differentiation and myelination and is tightly controlled by many extrinsic and intrinsic regulators, such as cytokines, neurotrophins, TFs and myelin regulatory factor, which are potential therapeutic targets.

3.5.1. Cytokines and neurotrophins.

Many cytokines and trophic factors facilitate OPC proliferation, OL differentiation and remyelination. Generally, post-injury inflammation clears myelin debris and stimulates remyelination. Activation of some cytokines (such as IL-1β and TNFα) and TLRs enhances OPC recruitment, differentiation and remyelination around lesion. In addition to glia, leukocytes and microglia produce numerous neurotrophic factors, including neuregulin 1 (Nrg1), oncomodulin, osteopontin, platelet-derived growth factor, epidermal growth factor, fibroblast growth factor-2, CNTF, activin-A, glial-derived growth factor, endothelin-2, IGF-1 and BDNF. Among them, Nrg-1 interacts with ErbB2/3 receptors on various types of glia and regulates their functions by Erk1/2 and Stat3 signaling. Delivery of Nrg-1 immediately after SCI preserves neural tissues around the lesion and promotes functional recovery in rodents by upregulating anti-inflammatory factors IL-10 and arginase-1 and downregulating pro-inflammatory elements IL-1β, TNFα, nitric oxide and MMP-2 and 9 ¹⁵⁵. Consistently, intraspinal delivery of an Nrg-1 peptide with microparticles stimulates OPC differentiation into OLs and reduces neuroinflammation and glial scar formation in SCI rodents ¹⁵⁶.

3.5.2. Transmembrane receptors.

Many extracellular and cell surface components (e.g., axonal guidance cues and receptors for myelin axon growth inhibitors) regulate OPC/OL functions and remyelination. Interactions between Sema4D and its receptor plexin B1 suppress OPC differentiation and blocking the former with its antibody (called VX15/2503) promotes OL formation and remyelination. Phase I clinical trials of multiple sclerosis (MS) patients show its safety profile. Because Sema7A appears to suppress OL migration into the lesion area, blocking its function may also facilitate remyelination. Lingo-1, a transmembrane leucine-rich repeat protein known as co-receptor for NgR1, represses OPC maturation and myelination. Lingo-1 antibody shows various beneficial effects in preclinical studies ¹⁵⁷, including enhancing remyelination in different CNS lesion models. Clinical trials for optic neuritis with Lingo-1 antibody was promising, but its phase II trials in MS failed to meet the primary end points. AMIGO3, an analog to Lingo-1, is also upregulated after SCI and may become a molecular target for axon regeneration and remyelination. Canonical notch signaling blocks OPCs differentiation during development, but its potential role in CNS remyelination is unclear. It is likely that the interactions between notch receptor on OPCs and Jagged1 expressed by reactive astrocytes repress OPC differentiation and OL genesis after SCI. Quercetin, a notch inhibitor that targets γ-secretase-mediated cleavage, improves remyelination in different CNS demyelination models.

3.5.3. Epigenetic factors.

Various epigenetic factors regulate function of myelinating glia in development and could act as therapeutic targets for demyelinating diseases. In spite of crucial roles of HDACs in OL differentiation during development, temporarily blocking HDAC is generally neuroprotective after CNS injuries. HDAC inhibition seems to reduce inflammation and to increase remyelination in demyelination models. HDAC inhibitors valproate and L-thyroxine (a T3 precursor) suppress infiltration of CD4+ T cells into CNS and clinical symptoms in rodents with experimental allergic encephalomyelitis (EAE). Trichostatin A, another broad HDAC inhibitor, shows similar beneficial effects in EAE rodents. HDAC inhibitor sodium butyrate also prevents OL loss in brain ischemia models by suppressing inflammation and enhancing OL genesis. SIRT1, a class III HDAC, also regulates OPC proliferation and regeneration in white matter after neonatal brain injury. Because HDAC subtypes have both overlapping and distinct functions, dissecting the roles of individual HDACs and identifying selective inhibitors for HDAC subtypes are important to boost therapeutic efficacy and to minimize off-target effects. Current use of some HDAC inhibitors for treating cancer patients may facilitate their translation into CNS disorders.

Because numerous miRNAs regulate OL differentiation and myelination by targeting diverse genes, they are also therapeutic targets to remyelinate injured CNS ¹⁵⁸. Dicer1 deletion interrupts normal CNS myelination mediated by miR-219 and miR-338. MiR-219 halts proliferation and prompts differentiation of OPC by silencing several TFs (e.g., SOX6, FOXJ3 and zinc finger protein 238). MiR-338 stimulates OPC differentiation into mature OLs by repressing HES5, SOX4 and SOX6. MiR-219 and miR-338 could synergically enhance OPC differentiation into mature OLs by targeting inhibitory Lingo1 and Etv5 ¹⁵⁹. MiR-297c-5p supports cell cycle arrest and OPC differentiation by targeting cyclin T2, an OL maturation inhibitor ¹⁶⁰. EAE mice exhibit downregulated miR-19b and its application upregulates myelin proteins and reduces demyelination. Also, suppressing miR-146a attenuates demyelination and axon loss.

Other epigenetic factors are also molecular targets for remyelination. ATP-dependent chromatin remodeling regulates OL differentiation and genesis by using ATP to relocate or remove nucleosomes and then to access chromatin for gene transcription and expression. Deleting Brahma-related 1 (Brg-1), a member in the ATP-dependent chromatin remodeling complexes, prevents OPC differentiation into OLs in development ¹⁶¹. Brg1 and Olig2 cooperate in promoting OL lineage progression by activating another ATP-dependent remodeler CHD7, which regulates OPC differentiation and activation by collaborating with Sox10 and Sox2 ^{162, 163}. Activating retinoid X receptor γ, a heterodimeric nuclear receptor, potentiates OPC differentiation during remyelination and its small activator, IRX4204, is effective for remyelinating CNS axons in lesion models.

It is also interesting to study therapeutic potential of other molecules for SCI. Several FDA-approved clinical drugs, including anti-muscarinic drugs benztropine and clemastine and NSAID indomethacin, promote OPC proliferation, OL differentiation and remyelination in demyelinated rodent models through various molecular mechanisms. Upregulation of myelin regulatory factor in OPCs or OLs appears also to facilitate remyelination of CNS axons.

4. Conclusion.

In recent years, researchers have made significant progress in further understanding the pathophysiology of SCI and in developing effective preclinical therapeutic strategies for SCI by targeting diverse cellular and molecular mechanisms. Most treatments for acute SCI target prevention of ongoing direct damage to the spinal cord and neuroprotection to minimize the secondary injury, while most therapeutic strategies for chronic SCI focus on neural repair, regeneration, rehabilitation and therapy for complications. Preclinical reports show that some therapeutic approaches are very promising for treating SCI, including those that modulate inflammatory reactions, enhance mitochondrial biogenesis and facilitate axon regeneration and neuronal reorganizations in adult CNS. Targeting multiple genes has been shown to improve functional recovery, but it is difficult to compare the extent of their effects because diverse outcome measures and model systems were employed, including different SCI lesions and animal species/ages. Among them, several molecular therapies have been moved to clinical trials for SCI, such as ketogenic diet, Rho inhibitor VX-210, NogoA antibody IN-1 and NgR antagonist. Table 1 summarizes the currently active clinical trials for SCI using pharmaceutical strategies reported in clinicaltrials.gov. In the future, it is essential to develop additional highly effective molecular and cellular strategies that target different mechanisms and to combine the most promising therapies for maximal functional recovery.

Table 1.

Currently active clinical trials for SCI using pharmaceutical strategies.

Drug name	Major mechanism and function	Study title
4-Aminopyridine	A potassium channel antagonist to restores axonal conduction	High Doses of 4-aminopyridine in Clinically Complete Chronic Spinal Cord Injury Patients
AXER-204 [NgR(310)ecto-Fc]	A soluble ectodomain of Nogo receptor as its antagonist to increase neurite growth	AXER-204 in Participants with Chronic Spinal Cord Injury (RESET)
Buspirone and levodopa-carbidopa	Buspirone acts as a 5-HT1A receptor agonist to decrease serotonin levels while increasing dopamine levels. Carbidopa inhibits biosynthesis of levodopa to dopamine.	Acute Effects of Pharmacological Neuromodulation on Leg Motor Activity in Patients with SCI Treated with EES (STIMO-PHARMA)
Cannabinoids and anti-inflammatory diet	A psychoactive drug in cannabis plus a diet to reduce neuropathic pain and inflammation	Cannabinoids and an Anti-inflammatory Diet for the Treatment of Neuropathic Pain After Spinal Cord Injury
Cethrin (VX-210)	A Rho inhibitor to suppress axonal dieback and to promote axon regrowth	A Safety Study for Cethrin (BA-210) in the Treatment of Acute Thoracic and Cervical Spinal Cord Injuries (completed)
D-Cycloserine	Partial agonist at glycine recognition site of NMDA receptor to enhance neural connections	Lower Limb Function After Spinal Cord Injury
Fibroblast growth factor	Growth factors to promote cell proliferation, neurogenesis and axon growth	Safety and Efficacy of SC0806 (Fibroblast Growth Factor 1 and a Device) in Traumatic Spinal Cord Injury Subjects
Glyburide (glibenclamide)	A SUR1-TRPM4 channel inhibitor to reduce cytotoxic edema and cell lysis	Spinal Cord Injury Neuroprotection with Glyburide (SCING)
Hyperbaric oxygenation (HBO)	Hyperbaric Oxygenation to reduce lipid oxidation	Hyperbaric Oxygenation (HBO) in Traumatic Spinal Cord Injury. (EOS) - Pilot Study
Imatinib	A tyrosine kinase inhibitor to reduce inflammation	Treatment of Cervical Spinal Cord Injury with Imatinib - a Safety and Feasibility Study
Ketogenic diet	Ketogenic diet contains high fat, low carbohydrate and appropriate protein consumption to increase ketone body production, oxidative metabolism and ATP synthesis, and to reduce ROS formation	Ketogenic Diet to Improve Neuro-recovery
Lexapro	A selective serotonin reuptake inhibitor to modulate motor activity	Serotonergic Modulation of Motor Function in Subacute and Chronic SCI
Lithium carbonate with umbilical cord blood cell transplant	Lithium is a GSK-3bβ inhibitor and also interacts with several neurotransmitters and receptors. It is used primarily for bipolar disorder, but also increases neurite growth.	Umbilical Cord Blood Cell Transplant into Injured Spinal Cord with Lithium Carbonate or Placebo Followed by Locomotor Training
Midodrine	A vasopressor to increase blood pressure	Blood Pressure, Cerebral Blood Flow and Cognition in Spinal Cord Injury
Minocycline	A regulator for MAPK, PI3 kinase and NMDA receptor to reduce inflammation and apoptosis	Minocycline in Acute Spinal Cord Injury (MASC)
NeuroAiD	A traditional Chinese medicine to stimulate neural cell growth and connections	Spinal Cord Injury - Assessing Tolerability and Use of Combined Rehabilitation and NeuroAiD
NG-101 (IN-1 antibody)	A Nogo A antibody to promote neurite growth	NISCI - Nogo Inhibition in Spinal Cord Injury (NISCI)
PMZ-1620 (sovateltide)	An endothelin-B receptor agonist to increase neurogenesis and blood vessel formation	PMZ-1620 (Sovateltide) in Patients of Acute Spinal Cord Injury
Riluzole	An inhibitor for tetrodotoxin-sensitive sodium channel and NMDA receptor to protect neural tissues and to reduce excitotoxicity	Riluzole in Spinal Cord Injury Study (RISCIS);Riluzole in the Treatment of Spasticity in the Traumatic Chronic Spinal Cord Injury Condition (RILUSCI)
Testosterone plus finasteride	Testosterone is male sex hormone to increase masculine growth. Finasteride is 5α-reductase inhibitor to reduce production of androgen dihydrotestosterone from its precursor testosterone.	Higher-Than Replacement Testosterone plus Finasteride Treatment After SCI
Tiotropium bromide	A muscarinic receptor antagonist to relief respiratory functional impairment	24hr Effects of Tiotropium Bromide in Tetraplegia

5. Expert opinion.

It is critical to identify additional highly effective therapies and to translate them into clinics successfully. Pathology of SCI is very complicate, including secondary damages to neural tissues after a primary injury, neuronal and glial loss, scar formation, and permanent disconnections of axon tracts around the lesion. Recent advances in understanding the mechanisms of secondary injury, such as inflammation, autophagy and mitochondrial dysfunctions, may help develop novel and effective neuroprotective strategies. The major challenges for neuroprotective treatments include limited effects of targeting individual mechanisms, need for earliest intervention after SCI and their translation into human use. In recent years, researchers have made remarkable progress in developing various therapeutic strategies, but few have reached the clinics. SCI patient treatments is limited to early surgical decompression, strict blood pressure control and optional use of steroids.

Axon regeneration and rewiring of damaged neural circuits around the lesion are critical for restoring lost functions. Disconnection of white matter tracts underlies the primary functional deficits after SCI. Manipulating multiple genes could stimulate axon regeneration and sprouting, but it is still necessary to identify new targets to stimulate more robust and long distance regrowth in adults and even elders, such as new molecules and pathways critical for controlling growth capacity. Targeting both intrinsic and extrinsic growth modulators appears necessary to achieve dramatic and long distance regeneration of adult CNS axons. Future strategies should fine-tune the timing and degree of axon regeneration, such as activating neuronal growth programs for a certain period after injury and terminating it when regenerating axons reach and connect with their cellular targets.

It is important to identify strategies promoting functionally meaningful axon regeneration. It is critical to define the ultrastructure and function of regenerated axons, such as expression and distribution of various ionic channels and myelination, which are essential for signal conduction. Targeting several growth related signals simultaneously induces robust RGC regeneration, but regenerated axons fail to myelinate for functional recovery. Because the environment for axon regrowth in adult CNS is very distinct from that during neurodevelopment, another major challenge is guiding regenerating axons to functionally meaningful cells for proper synaptogenesis. Many regenerating RGC, spinal cord and PNS axons do not follow their original projection trajectories. Because of the relatively slow regrowth rate of injured axons and long distance requirement to reach their proper cellular targets, forming relay pathways by targeting interneurons, especially PSNs, is a very promising approach for neuroplasticity and functional recovery.

Combination strategies may maximize functional recovery by enhancing cell survival and axonal rewiring (i.e., axon regeneration, sprouting or relay circuits around the lesion). Combinations to treat both secondary injury mechanisms for neuroprotection and neuronal growth failure for regeneration may exhibit synergistic actions for functional restoration. Because multiple time-dependent mechanisms contribute to axon growth failure and functional loss, combined strategies to target several of them should be more effective than individual ones, such as enhancing intrinsic growth capacity, blocking local growth inhibitors, applying chemotropic gradients, and transplanting various types of cells or biomaterials to bridge the lesion gap. Appropriate rehabilitative therapies, such as electrical/magnetic stimulation and task-dependent training, should further improve clinical outcome. Additionally, the optimal strategies to repair different types of SCI might be variable, such as incomplete vs. complete, high cervical vs. low lumbar injuries, and juveniles vs. elders.

It is extremely important to translate the most promising individual and/or combined strategies into clinical use for neuroprotection, regeneration and more importantly functional recovery. Currently, there are no approved therapies for recovery of motor, sensation and autonomic neural functions in SCI patients. Though there are many ongoing clinical trials for SCI, the translational potential for most promising therapies remains unclear. It is important to further understand cellular and molecular mechanisms for the newly identified effective strategies and to reproduce promising results with different model systems, including use of various animal species. Among various strategies, non-invasive approaches, such as systemic administration of compounds, are more attractive than invasive ones, such as surgical transplants or drug delivery. Though most animal species used for SCI models have similar pathobiology to humans and share homologs in molecules, the requirements for neural repair are actually very different among them, such as the distance required for regeneration and the types and extent of functional recovery after CNS lesions. Because most targeted molecules for therapies are expressed in multiple cell types and have diverse roles, manipulating their functions may also result in off-target effects, such as tumorigenesis due to cell overgrowth and neuropathic pain due to extensive sprouting of sensory axons.

Article highlights.

• SCI patient treatments is currently limited to early surgical decompression, strict blood pressure control and optional use of steroids.

• Preclinical studies show promising results in developing therapeutic strategies for repairing injured spinal cord.

• Neuroprotective strategies targeting inflammation, autophagy, oxidative stress and metabolic dysfunctions may reduce tissue damage and promote recovery.

• Rewiring disconnected axon circuits around the lesion is critical for functional restorations after SCI.

• Combined strategies to target multiple responsible factors may promote robust CNS repair, axon regeneration and functional recovery.