The leading edge: Emerging neuroprotective and neuroregenerative cell-based therapies for spinal cord injury

1Traumatic spinal cord injuries (SCIs) result in tremendous lifelong disability and financial burden for millions of patients and caregivers worldwide. Cell‐based therapies have emerged as an exciting neuroprotective and neuroregenerative strategy for SCI. This review highlights key preclinical and clinical data in cell therapy with an emphasis on the pathobiology and mechanisms of recovery. Also discussed are adjunct treatments to maximize the efficacy of the grafts. Finally, important translational considerations such as clinical‐grade scale‐up and delivery techniques are discussed. The article succinctly provides readers with a working knowledge of SCI and cell therapies at the leading edge of research.

2Traumatic spinal cord injuries (SCIs) result in tremendous lifelong disability and financial burden for millions of patients and caregivers worldwide. Cell‐based therapies have emerged as an exciting neuroprotective and neuroregenerative strategy for SCI. This review highlights key preclinical and clinical data in cell therapy with an emphasis on the pathobiology and mechanisms of recovery. Also discussed are adjunct treatments to maximize the efficacy of the grafts. Finally, important translational considerations such as clinical‐grade scale‐up and delivery techniques are discussed. The article succinctly provides readers with a working knowledge of SCI and cell therapies at the leading edge of research.

3Spinal cord injuries (SCIs) have tremendous physical, social, and financial consequences for over 1 million North Americans and their families. 1, 2 Direct lifetime costs of care range from $1.1 to $4.7 million per person not including lost wages and productivity. 2 Rapid delivery of specialized medical and surgical care has significantly reduced mortality; however, long‐term functional recovery remains limited. 3, 4, 5, 6 Cell‐based therapies have emerged as an exciting strategy to neuroprotect and regenerate the injured cord through multiple mechanisms such as immunomodulation, paracrine signaling, extracellular matrix (ECM) modification, and lost cell replacement. 7, 8 Herein, we summarize the most promising preclinical and clinical cell therapies, adjunct strategies to enhance transplant success, as well as key translational considerations such as sex and age. Throughout, we emphasize the fundamental biology of stem cells, critical features in the pathophysiology of spinal cord injury and provide meaningful discussions on the strengths and limitations of each therapeutic approach.

4The epidemiology of SCI is an important consideration when designing clinical trials. Traumatic SCI is more common in males (79.8%) than females (20.2%). Most injuries are cervical (~60%) followed by thoracic (32%) and lumbosacral (9%). 9 There is a bimodal age distribution with one peak occurring from 15 to 29 years of age and a second, smaller but growing peak, occurring after age 50. 10, 11 High‐energy motor vehicle collisions (MVCs) and sports‐related injuries disproportionately affect younger individuals. Low‐energy trauma, such as falls, are more common in those over 60 years old where underlying degenerative spinal conditions, such as degenerative cervical myelopathy, are more prevalent. 11, 12 Interestingly, MVCs account for a declining majority (38%) of SCIs in North America, 9 whereas falls are increasing and account for 31% of injuries followed by sports‐related impacts at 10% to 17%. 11, 12

5The initial traumatic event causes permeabilization of cell membranes, ion and small molecule dysregulation, and ischemia due to damage to the sensitive microvascular supply. 13, 14 Together, these events initiate a secondary injury cascade which generates further permanent damage (Figure 1A). Over several hours, progressive edema and hemorrhage cyclically add to the harsh postinjury milieu. The compromised blood‐spinal cord barrier (BSCB) exposes the vulnerable cord to inflammatory cells, vasoactive peptides, and cytokines such as tumor necrosis factor and interleukin‐1β. 16 Ongoing cell death releases DNA, ATP, and K+ into the microenvironment; microglia respond by secreting additional pro‐inflammatory cytokines and promoting the infiltration of large numbers of macrophages, neutrophils, and nearby microglia. This activates astrocytes and endothelial cells which further secrete factors such as BMPs, TGF‐β, and Notch activating ligand, Jagged. Activated phagocytes can clear myelin debris within the injury but also produce oxygen free radicals (eg, O2−, peroxynitrite and hydrogen peroxide) and cytotoxic by‐products which generate additional cell death through lipid peroxidation, protein oxidation, and DNA damage. 17, 18 Extracellular glutamate accumulates as neurons die and astrocytes' reuptake capacity is lost. 19, 20 This leads to excitotoxic cell death of the remaining neurons through NMDA, kainate, and AMPA receptor overactivation combined with ATP‐dependent ion pump dysfunction and subsequent sodium dysregulation (Figure 1B). 21, 22

6At a systemic level, poor respiratory function can cause hypoxia whereas loss of sympathetic innervation to the vasculature can result in profound hypotension. Combined with the impaired autoregulatory capacity of the cord, this can contribute to ongoing ischemia for days to weeks postinjury. 23 The multiple causes of acute and subacute cell death in this injury cascade represent important targets for cell‐based neuroprotective approaches.

7In the intermediate‐chronic phase, acute inflammation subsides and the cord undergoes alterations in ECM composition, attempts at remyelination, and remodeling of neural networks. 24 Although this can result in limited recovery, multiple barriers to local circuit and long‐tract regeneration persist.

8Neuroglial cell death and degeneration in the early phase disrupts the cord's structural framework and leads to ex vacuo formation of microcystic cavitations containing extracellular fluid with thin bands of connective tissue. 25 These cavities coalesce into larger collections which lack substrate for directed axonal regrowth and regenerative cell migration. 26, 27 Additionally, oligodendrocytes are susceptible to necrotic and apoptotic cell death. The denuded axons they leave behind cannot utilize rapid saltatory conduction and are particularly susceptible to nonfunctional electrogenesis which further contributes to poor recovery. 28

9Early after injury, astrocytes also proliferate within the perilesional zone and tightly interweave an irregular mesh of processes to sequester the injured region (Figure 1C). Resident neural stem and progenitor cells surrounding the central canal can also differentiate to astrocytes and contribute to this astrogliosis. The astrocytes, pericytes, and ependymal cells in the region generate dense deposits of chondroitin sulfate proteoglycans (CSPGs), NG2, and tenascin which form the fibrous component of the glial scar. 29, 30, 31, 32 Although literature exists supporting the beneficial aspects of scar, the balance of evidence suggests that chronic scarring potently inhibits axonal regeneration and neurite outgrowth by acting as a physical barrier and tightly binding transmembrane protein tyrosine phosphatase receptors. 33, 34, 35

10Furthermore, CNS myelin‐ and neuron‐associated ligands, such as myelin associated glycoprotein (MAG), oligodendrocyte myelin glycoprotein (OMgp), neurite outgrowth inhibitor (NOGO), and semaphorin 3A/4D, bind NOGO receptor‐p75 complexes (NgR) and plexins to activate Rho GTPase and its downstream effector, Rho‐associated protein kinase (ROCK). 36, 37, 38, 39 This results in a change in actomyosin contractility and collapse of the axonal growth cone and further inhibition of regeneration. There are other potent inhibitors of axonal regeneration such as Repulsive Guidance Molecule A which are upregulated in the injured cord. 40, 41, 42

11Multiple cell strategies discussed below aim to preserve and/or regenerate functional, myelinated neural circuits to enhance functional recovery.

12Cell‐based therapies include cell transplantation and harnessing the potential of endogenous neural precursor cells. Cell treatments can immunomodulate, alter the microenvironment, and replace lost cells depending on the cell type, cell state, delivery route (eg, systemic vs local), and timing of administration. 43, 44, 45, 46 The most commonly studied and promising cell types include mesenchymal stem cells (MSCs), neural stem cells (NSCs), oligodendrocyte progenitor cells (OPCs), Schwann cells (SCs), and olfactory ensheathing cells (OECs). 47, 48, 49, 50 This section outlines the mechanisms of action for each and summarizes progress along the translational research spectrum. Key preclinical studies are highlighted in Table 1, and completed and ongoing clinical trials are summarized in Tables 2 and 3, respectively.

13MSCs, SCs, and OECs can all be harvested from an adult allogeneic source to generate standardized stocks depending on the success of proliferation. MSCs, SCs, and OECs can also be derived directly from the patient to avoid post‐transplant immunosuppression. 84 However, autologous primary cells are typically more costly requiring harvest surgery, in vitro expansion and extensive characterization prior to transplant.

14CNS cells, such as NSCs, OPCs, astrocytes and microglia, are more challenging to isolate from adult allogeneic donors, and the performance of a line is influenced by donor age, genetics, and harvest conditions. 85, 86, 87 Furthermore, autologous CNS tissue is inaccessible. As a result, these cells are often derived from embryonic stem cell (ESC) sources. 88, 89 ESCs can be propagated indefinitely and can generate cells of any germ layer. However, ESC‐derived grafts have ethical issues surrounding their use and may show karyotypic instability or hold the potential for tumorigenesis due to incomplete or aberrant differentiation. More recently, induced pluripotent stem cells (iPSCs) have allowed derivation of NSCs and OPCs from autologous, accessible cells such as bone marrow and skin fibroblasts. This has been further adapted to allow direct reprogramming of adult somatic cells to multipotent neuroglial cells while bypassing the pluripotent state. 7 In more recent protocols, it has also become possible to convert easily accessible somatic cells directly into neurons, 90, 91 neuronal subtypes, 92, 93 and oligodendrocytes progenitors. 94 Some limitations associated with these approaches such as reprogramming efficiency, line variability, lineage‐specific differentiation, and retention of epigenetic memory are being investigated.

15NSCs are tripotent, self‐renewing cells which have attracted great interest as they can potentially replace the neurons, oligodendrocytes, and astrocytes lost after injury. 88, 95, 96, 97 During embryological development, NSCs are found throughout the neural tube where they acquire unique identities based on their position and temporal exposure to patterning morphogens. 98, 99, 100, 101 In adults, they are found in a more limited number of regions such as the subventricular zone in the brain 95, 96, 97 and around the central canal in the spinal cord. 102, 103, 104, 105 There are two distinct NSC populations that can be isolated from the adult spinal cord: (a) primitive NSCs (pNSCs) and (b) the definitive NSCs (dNSCs) they give rise to (Figure 2). 106, 107 pNSCs are rare cells expressing pluripotency marker, Oct4, and are responsive to leukemia inhibitory factor in vitro. 108, 109, 110 dNSCs are more abundant in adults, express astrocyte marker, GFAP, and respond to epidermal and fibroblast growth factors (EGF and FGF) in vitro. Both populations can proliferate and generate neurons, astrocytes and oligodendrocytes.

16Numerous small and large animal preclinical studies have provided evidence supporting the therapeutic efficacy of transplanting NSCs derived from ESCs or allogeneic adult sources. 60, 62, 84 More recently, autologous iPSC‐derived and directly reprogrammed NSCs have emerged as translationally relevant options which mitigate immunorejection concerns and generate highly pure populations of cells over several weeks. A 2016 meta‐analysis (N = 74 studies) found all sources of NSCs to provide significant motor recovery in models of SCI (pooled SMD = 1.45; 95% confidence interval [CI]: 1.23‐1.67; P <.001). 111 Although the mechanisms underlying recovery continue to be investigated in preclinical studies, it is likely that trophic signaling (eg, BDNF, GDNF, IGF‐1, etc.), remyelination of denuded axons, partial regeneration and remodeling of neural circuitry, and ECM deposition play a role. 7, 84, 112 Recent work demonstrating the self‐assembly of grafted spinal cord NSCs into organotypic, dorsal horn‐like domains highlights their plasticity and the potential for further optimization as a therapy. 113

17Currently, a phase I/II study in China (N = 30; NCT02688049) is comparing 1 × 106 MSCs vs NSCs supported by a linearly ordered collagen biomaterial, NeuroRegen scaffold, in individuals with chronic AIS Grade A C5‐T12 injuries. The study is expected to conclude in 2020 and report on AIS grade, somatosensory evoked potentials, motor evoked potentials, Functional Independence Measure, and MRI assessments. Neuralstem is also conducting a trial (N = 8; NCT01772810) of NSI‐566, human fetal spinal cord NSCs, delivered through six intraspinal injections into patients with 1 to 2 year old thoracic injuries. In 2016, a pair of phase II trials led by Stem Cells Inc were terminated prior to completion. The studies were assessing the effects of human CNS stem cell transplants for thoracic (NCT01321333) and cervical (NCT02153876) SCI. 114 A recently published report on the studies found that escalating doses up to 4.0 × 106 were well tolerated with no significant increase in serious adverse events related to the cells or free‐hand manual injection technique. 83 Based on emerging preclinical data, it is likely that further modifying the cells in vitro and/or altering the local environment will be necessary to enhance functional recovery.

18OPCs are tripotent cells which predominantly differentiate into oligodendrocytes to remyelinate axons. They are suited for regeneration in SCI as they respond rapidly to injuries, can self‐renew, and address an important component of the injury pathophysiology. Animal studies have shown that transplanted OPCs can promote white matter preservation, increase the number of surviving endogenous oligodendrocytes, and reduce cavity volume, resulting in enhanced motor recovery. The underlying mechanism may be a combination of remyelination, local immunomodulation, trophic factor secretion, and provision of a physical scaffold to support growing axons. 69, 70, 115, 116 The cells also possess a favorable secretome, 117 consisting of growth factors, neurotrophins, chemokines and cytokines, and can form glutamatergic synapses with neurons, an area of ongoing discovery. 118 It remains unclear whether improved outcomes are due to these factors or direct remyelination of denuded axons, and to what extent remyelination can itself enhance function after injury. 119, 120

19A large clinical trial by Geron Inc assessing human ESC‐derived OPCs was discontinued for financial reasons. Renewed funding allowed Asterias Biotherapeutics Inc to extend the trial of these cells, termed AST‐OPC1, in a phase I/IIa open‐label dose‐escalation study (N = 25; NCT02302157). 2 × 106, 1 × 107, or 2 × 107 cells were transplanted 21 to 42 days after SCI patients with AIS grade A or B injuries at C4 to C7. The study completed in 2019 and BioTime, which later changed its name to Lineage Cell Therapeutics, acquired the company. Key findings were no increase in significant adverse events and evidence of regeneration on MRI at 12 months. Ninety‐five percent of patients improved by at least 1 AIS grade and 32% of patients improved by at least 2 AIS grades by 12 months after receiving the 1 × 107 and 2 × 107 doses. 114 Longer term follow‐up is continuing and the company plans to further optimize the cell manufacturing process. 121

20OECs are specialized glial cells that encircle olfactory neurons and clear bacteria and debris at the CNS‐nasal mucosa transition. 122, 123, 124, 125 They also secrete neurotrophic factors 126 and maintain a favorable environment for neuronal function. When the olfactory nerve or epithelium is damaged, OECs support the growth of olfactory epithelium‐derived neurons into the olfactory bulb. They differ from typical glia but share a number of morphological and molecular markers with astrocytes and SCs. After transplantation, OECs form a cellular substrate through which injured axons can regenerate across a spinal cord transection lesion site. 127 OECs can be harvested from the nasal mucosa or directly from the olfactory bulb and transplanted into the cord parenchyma as a bridging, nonrelay approach. 75 OECs have also been shown to enhance neurite outgrowth, promote remyelination, neuroprotect, provide guidance cues, and locally immunomodulate. 49, 128, 129

21Several clinical trials have transplanted OECs for subacute and chronic SCI. Early work confirmed the safety of a purified population of OECs; however, subsequent studies using mucosal tissue reported conflicting results. 130, 131, 132 To clarify the discrepancy, a meta‐analysis of key clinical trials (pooled N = 1193) was conducted which found no statistically significant increase in serious adverse events; however, efficacy has not been definitively established due to technical and methodological concerns with existing studies. 133 Recently, a small phase I trial (N = 6) of autologous mucosal OECs and olfactory fibroblasts demonstrated sensorimotor improvements after transplant into patients with AIS grade A injuries; however, a larger sample size and extended follow‐up will be required to confirm safety and efficacy. 45, 129, 131

22SCs are myelinating cells in the peripheral nervous system (PNS) and are an important component of the robust, spontaneous regeneration observed in the PNS. They provide a structural scaffold acting as a conduit to guide growing axons. They also produce a favorable environment by expressing growth factors and extracellular proteins. In animal models of SCI, they have been shown to promote tissue sparing, reduce cystic cavitation, enhance CNS axon regeneration, remyelinate axons, and enhance endogenous SC myelination resulting in sensorimotor recovery. 134, 135, 136, 137, 138 The growth‐promoting properties of SCs are also being exploited in combinatorial therapies. For example, neuroprotectant D15A (a modified human NT3 that can activate both TrkB and TrkC receptors) has been combined with phosphodiesterase‐4 inhibitor, rolipram, and SCs to enhance axonal sparing and growth of serotonergic fibers into and beyond the SC graft. 139 SCs have also been combined with NSC, 140 OEC, 74 and BM‐MSC 141 transplants to enhance cell survival and promote additional remyelination.

23In humans, The Miami Project to Cure Paralysis conducted a phase I, open‐label study (N = 8; NCT02354625) of autologous SC transplants in individuals with chronic AIS grade A‐C C5‐T12 injuries. The study completed in August 2019 with results pending. The same group conducted an open‐label, nonrandomized, noncontrolled, dose‐escalation phase I study (N = 6) of autologous SCs transplanted into the lesion epicenter for subacute‐intermediate AIS grade A thoracic injuries. SCs were harvested from the sural nerve within 5 to 30 days of injury, expanded in vitro, and transplanted within 4 to 7 weeks of injury. After 1 year, there were no medical, surgical, or neurological complications to indicate that the treatment was unsafe. 50

24MSCs are multipotent, self‐renewing connective tissue progenitor cells found throughout the body, particularly in the perivascular region. They have the capacity to regenerate muscles (myocytes), cartilage (chondrocytes), bone (osteoblasts), and fat (adipocytes). 142, 143 They also have favored properties for conducting a clinical trial. Multipotent MSCs expand rapidly, 144 remain viable after −80°C or liquid nitrogen cryostorage, 145 demonstrate minimal immunoreactivity after allogenic transplant, 146 and can be harvested from accessible tissue such as fat, bone marrow, and skeletal muscle. 147 This has encouraged their translation in multiple fields such as sepsis, 148 multiple sclerosis, 149 and arthritis. 150 In SCI, MSCs have been shown to promote angiogenesis and significantly enhance tissue sparing through neurotrophic signaling and immunomodulation. 8, 56, 151, 152 There is also evidence that MSCs transplanted directly into the spinal cord can modulate activation of macrophages and promote tissue sparing. 153, 154 However, MSC transplants have shown considerable variability with some studies showing positive effects whereas other studies have shown no benefit (Table 1).

25Multiple adipocyte‐derived MSC clinical trials are ongoing to assess safety, dosing, and efficacy. The Mayo Clinic is conducting a phase I study (N = 10; clinicaltrials.gov identifier NCT03308565) of 1 × 108 autologous, adipose‐derived MSCs delivered intrathecally into patients with American Spinal Injury Association (ASIA) Impairment Scale (AIS) grade A/B/C injuries from 2 weeks to 1 year prior to transplant. The study will be completed by 2023.

26Bone marrow‐derived MSCs are also being studied. A phase II/III trial (N = 32; NCT01676441) conducted by Pharmicell Co is assessing 1.6 × 107 intraparenchymal and 3.2 × 107 intrathecal autologous bone marrow‐derived MSCs in chronic cervical AIS grade B patients with 12 month follow‐up of ASIA motor scores, MR Diffusion Tensor Imaging, and electrophysiological parameters. The study is expected to conclude in 2020.

27The umbilical cord is an alternate source of MSCs (UC‐MSCs). UC‐MSCs can be isolated from cord tissue, cord blood, or Wharton's jelly, a gelatinous substance within the umbilical cord that insulates the blood vessels. Umbilical cord tissue is readily accessible and frequently discarded, and MSCs from the umbilical cord are less prone to rejection, as evidenced by a lower risk of developing graft vs host disease. 155 Compared with adult sources, the number of MSCs obtained from cord blood or placental tissues is small, although they can be readily expanded and tissue can be frozen and used later for isolation. 156 MSCs derived from the umbilical cord have also been shown to have immunomodulatory properties. 157 Several recently registered phase I/II trials (NCT03505034; NCT02481440; NCT03521323) are recruiting to assess allogeneic UC‐derived MSCs for subacute and chronic SCI. A phase I/II open‐label study (N = 14; NCT02981576) in Jordan directly compared adverse effects and AIS improvement with intrathecal adipose‐ vs bone marrow‐derived MSCs in AIS A/B/C patients 2 or more weeks postinjury. The study completed in January 2019 with results pending.

28An alternative approach to promote neural repair is to harness the potential of resident stem cells, such as NSCs within the injured CNS. 158, 159 This approach circumvents immunorejection, cell delivery challenges and logistical hurdles such as good manufacturing practice (GMP) scale‐up, cell storage, and transport. Within the spinal cord, it has been demonstrated that NSCs respond to injury by proliferating 160 and migrating to the lesion, although the effect is likely insufficient to generate recovery. 29, 159 Recently, small molecules, such as the FDA‐approved drugs cyclosporin A and Metformin, have been found to increase the size of the endogenous pNSC and dNSC pool to augment this intrinsic injury response. 161, 162 Cyclosporin A is an immunosuppressive medication that has been shown to enhance survival of NSCs and promote recovery within the brain. 161, 162 Activation of NSCs with cyclosporin A has also been observed in the spinal cord and ongoing investigations aim to elucidate its effect following SCI. 161, 162 Metformin, a drug commonly used to treat type II diabetes, has also been shown to activate endogenous NSCs and direct their differentiation toward neurons and oligodendrocytes. 110, 159 Although molecular mechanisms underlying these effects have not been elucidated, metformin administration has been shown to improve functional recovery after insult to the brain. 110, 159 Studies of metformin in SCI animal models are ongoing. Additional targets being explored to enhance the endogenous NSC response are the C‐Kit and ErbB2 signaling pathways. 163

29Transplanted cell survival has historically been low in animal models of SCI. 164 This is typically overcome by delivering excess numbers of cells to compensate for losses; however, this approach introduces variability into the therapy, contributes cytotoxic by‐products of cell death to the microenvironment, and becomes infeasible when large numbers of surviving cells are required. Ongoing work seeks to enhance cell survival, migration, and axonal outgrowth by overcoming key barriers in the SCI microenvironment.

30To support graft survival, growth factors (eg, PDGF, EGF, and IGF‐1), neurotrophins (eg, BDNF, NT3, NGF), and anti‐inflammatory agents (eg, minocycline) have all been successfully delivered via intrathecal injections and pumps. 63, 164, 165, 166 Unique biomaterials have also been engineered to gradually deliver key factors to support grafts. 167 Growth factors and treatments such as ferritin are also associated with enhanced endogenous OPC proliferation and oligodendrogenesis. However, pumps are prone to failure, require refilling and explant procedures, and expose growth factors to mammalian body temperatures for prolonged periods. As a result, alternate approaches continue to be investigated such as the codelivery of cells with biomaterials capable of slowly releasing growth factors directly into the environment. This unique strategy is discussed further in the Biomaterials section below. Another approach is the in vitro genetic modification of cells or the in vivo transfection of endogenous cells to secrete the necessary factors. For example, MSCs have been successfully engineered to express bFGF, 168 HGF, 169 NT3, 170 BDNF, 171 and GDNF 172 in vivo for various applications. SCs have also been transduced to overexpress BDNF and NT3 simultaneously. 173 Similarly, safe and highly efficient methods of engineering human iPSCs, ESCs, and NSCs are currently being developed.

31An often overlooked method of promoting endogenous trophic factor release and long‐term cell survival noninvasively is rehabilitation. Physical rehabilitation, with or without electroceutical augmentation, is an integral component of the care plan for individuals with SCI; however, it is underrepresented in preclinical trials. Whether the rehabilitation entails forced treadmill training, free swimming, or task‐specific tests such as forelimb reaching, the functional benefits can be significant. 174 In addition to enhancing cardiorespiratory and musculoskeletal function, treadmill locomotor training has been found to enhance transplanted NSC survival by more than fivefold through increased IGF‐1 signaling. 175 This finding underscores the value of multimodality, interdisciplinary care in SCI.

32Biomaterials can enhance cell‐based approaches for SCI in several ways. Scaffolds derived from either natural or synthetic polymers have been implanted within the lesion cavity to bridge the gap to serve as a substrate for axonal growth and cell migration. 176, 177, 178, 179, 180 InVivo Therapeutics' Neuro‐Spinal Scaffold is a porous bioresorbable polymer scaffold shown to promote appositional healing, white matter sparing, and normalization of intraparenchymal tissue pressure in preclinical models of SCI. 181 A recent case study at 6 month follow‐up from a patient enrolled in the clinical trial (NCT02138110) reported no adverse effects related to acute scaffold implantation. 182

33Scaffolds can also provide a physical substrate for seeded cells and provide directional guidance for axons. Moreover, biomaterials can also be used as vehicles to deliver cells and release growth factors to aid in graft cell survival, integration, and differentiation. Injectable in situ polymerizing hydrogels can deliver cells and factors directly into a lesion site with less invasive surgical interventions. For example, a polymer blend of hyaluronan/methylcellulose (HAMC) is injectable, in situ gelling, biodegradable, and noncytotoxic. 183 HAMC modified with PDGF‐AA, to enhance graft survival and oligodendrocyte differentiation of cotransplanted rat brain‐derived NSCs, 184, 185 promoted host oligodendrocyte sparing and improved fine motor function. 186 Further modification of the HAMC hydrogel with RGD peptide promoted the survival, integration, and differentiation of human iPSC‐derived OPCs. 187

34Another approach has been the use of fibrin scaffolds which have been shown to promote the survival of transplanted stem cells after SCI and, when codelivered with growth factors, have been used to direct differentiation and enhance recovery. 188, 189 QL6 is an exciting peptide biomaterial which self‐assembles to form a lattice‐like structure at physiological temperatures. QL6 injected with NSCs improved graft survival, reduced glial scarring and inflammation, and improved forelimb function in cervical models of SCI. 190, 191, 192

35Electrical fields (EFs) are a physical environmental cue present within living tissue. During development, multipotent cells rely on these fields for appropriate migration and differentiation. If the fields are disrupted, severe defects can result. 193 EFs have also been shown to guide cells in adults after injury. 194, 195 Therapeutic galvanotaxis (the directed migration of cells in an electric field) exploits the electrosensitivity of cells to promote migration using externally applied EFs. 196 This has been shown to be feasible with SCs, 197 NSCs, 198 and many other cell types. 199 Both endogenous and transplanted NSCs, but not their differentiated progeny, have been shown to migrate with transcranial direct current electrical stimulation; however, directed migration will require further optimization to establish the ideal current, voltage, phase, lead placement, and timing of EF application. 200, 201

36The glial/CSPG scar is well established in chronic injury and limits axon regeneration through the lesional/perilesional region (Figure 1C). A meta‐analysis of NSC treatments found motor function recovered to a greater extent with cell delivery in the acute phase (SMD = 1.80; 95% CI: 1.36‐2.24) or subacute phase (SMD = 1.38; 95% CI: 1.08‐1.67) than in the chronic phase (SMD = 1.04; 95% CI: 0.47‐1.60; P =.03) of injury 111 highlighting the difficulty in regenerating the chronically injured spinal cord.

37As a result, numerous labs have focused on methods to disrupt the glial/CSPG scar during or prior to cell transplant. Chondroitinase ABC (ChABC) is a bacterial enzyme which rapidly degrades the long glycosaminoglycan (GAG) side chains of CSPGs 202 In SCI, intrathecal or intraparenchymal ChABC treatments have been shown to promote neurite outgrowth and enhance anatomic plasticity, by degrading CSPGs within the perineuronal nets, resulting in sensorimotor behavioral recovery. 203, 204, 205 This exciting approach is under further development to address key limitations and better combine with cell transplants. First, thermostabilized variants of ChABC are being developed to retain activity for longer periods at mammalian body temperatures. 206 Second, ChABC is being delivered via novel vehicles such as affinity‐release biomaterials or lentivirus transfections of host cells. 207, 208 Finally, alternate human enzymes with GAG or CSPG core protein degrading activity are being studied to mitigate immunogenicity risks associated with using a bacterial protein. 209

38Another exciting approach is to inhibit the association of the protein tyrosine phosphatase σ receptor with its CSPG ligand. 210 An example is Intracellular Sigma Peptide (ISP) which is administered subcutaneously after injury, crosses the BSCB, and results in significant axonal regrowth within the injured cord. 34 Recently, ISP has also been shown to indirectly immunomodulate when combined with leukocyte antigen‐related receptor blockade. 211

39A major challenge in translating cell therapies to clinic is assessing and optimizing graft survival. Culture and storage conditions, characterization pipelines, and transplant conditions can be significantly different in the laboratory than in clinical trial or routine use. 212 Furthermore, commonly employed graft assessment techniques such as immunohistochemistry and bioluminescent tagging are typically not possible in humans. Strategies discussed previously such as increased trophic support, timed rehabilitation, and codelivering biomaterials may prove to be important adjuvants in advancing cell therapy for humans. These are in early deployment such as Neuro‐Spinal Scaffold (NCT02138110) by InVivo Therapeutics Inc or the phase I/II study of MSCs and NSCs using the NeuroRegen scaffold (NCT02688049). Assessing grafted cells in vivo may also become viable as novel MRI‐based cell trackers are developed. 213, 214

40Another important translational hurdle is understanding and overcoming potential immunorejection within the CNS. The extent and temporality of cell graft rejection within the human CNS is currently unknown. Additionally, the cell source (eg, autologous, allogenic, genetically modified, etc.) may significantly affect downstream graft survival in humans. 215, 216 Enhancing endogenous cell proliferation is one strategy to avoid immunorejection; however, endogenous cell pools may still be limited and optimal methods to drive their differentiation and migration have not yet been established. 217 Recently, genetic techniques which modify major histocompatibility complexes and CD47 have been shown to generate immune‐evasive iPSC lines. 218 This may become an important strategy in the future to protect grafts from immune cells.

41Effective translation of cell therapies into clinical trial and beyond requires careful planning of GMP facilities and scale‐up strategies. Many challenges of GMP‐grade production are not typically encountered in a research setting and are associated with significant financial costs. For example, all manual handling must be highly reproducible between facilities and occur in stringent GMP‐grade clean rooms. Standard operating procedures must be established and compliance must be traceable throughout the production chain. 212 Cultures are typically free of animal products (eg, proteins, feeder cells, etc.), unless more suitable options are not available, and all reagents require qualification as meeting the standards for GMP‐grade culture. 212 Common viral transduction and nonviral transfection (eg, electroporation, lipofection) techniques require additional steps to validate that risks inherent to viral particles or foreign DNA have been mitigated. Furthermore, master and working cell banks must be established with extensive testing for bacteria, viruses, fungi, mycoplasma, and endotoxins at each step of preparation. It is also important to note that during this translational process, the research cell line with which preclinical efficacy was validated, may not be the same as the final clinical cell line for trial which may necessitate additional efficacy testing in animal models. 219

42The importance of cell line and subline testing is apparent when addressing the potential for tumor formation as graft‐derived tumors can occur in rigorously tested and banked lines due to the multiple potential etiologies of tumorigenicity. 220, 221 As a living therapy, even small variations in transport, subculture, or transplant technique at individual centers hold the potential to alter the phenotype of the graft. Furthermore, lines are typically characterized by sampling a portion of the larger population; however, tumor formation can occur due to a single aberrant cell. Advanced technologies, such as environment‐controlled cell culture robots and high‐throughput total population screening techniques, seek to address these challenges. 222, 223

43Fortunately, the US FDA and European Medicines Agency have provided regulatory frameworks to begin approaching the complexities of GMP cell manufacturing. 224, 225 Additionally, networks of GMP‐grade facilities, such as the CellCAN Regenerative Medicine and Cell Therapy Network in Canada, have been established to aid in navigating these challenges. 226

44Conducting clinical trials or treatments across multiple sites requires a coordinated storage and transportation approach capable of accommodating international shipping delays and unexpected package handling conditions. For example, a study of human MSCs found that cells stored at 2°C to 8°C were sensitive to 25 Hz vibrations, 227 leading to cell death and increases in MSC marker expression such as CD29 and CD44. 228 Similarly, temperature and repeated freeze‐thaw cycles have a significant effect on cell viability. Cryoprotectant toxicity, rapid osmotic shifts, ice crystal formation, and activation of apoptosis are key mechanisms underlying cell death during the process. 229 Early studies achieved human cryopreserved pluripotent cell survival rates of only 30% or less. 230 As a result, other cold storage techniques have been developed such as vitrification, the rapid cooling of cells in high concentrations of cryoprotectants to inhibit ice formation. These were found to provide >75% survival but add technical complexity and may be limiting in the production of large‐scale banks. 229 Furthermore, the storage solution (eg, DMSO, polyethylene glycol, etc.) and additives (eg, ROCK‐inhibitor, trehalose, poly‐L‐lysine, etc.) exert their own effects on the survival and differentiation of stored cells. 229 Therefore, each cell type and cell line requires optimization to establish ideal conditions for a functional and reliable supply chain.

45Establishing the ideal patient population for cell transplant in SCI is challenging. Given the currently high cost and limited availability of GMP‐grade cell therapies, one strategy is to focus on populations with the greatest potential gains. Cervical cord injuries are the most common and can result in devastating impairments in activities of daily living (eg, feeding, grooming, transferring) and are often associated with respiratory and autonomic complications. 47 Even modest improvements in key muscle groups such as grip and elbow flexion can have profound benefits for quality of life making this an important population for inclusion in cell‐based studies. It is, however, important to balance these inclusion benefits with potential risks. The cervical SCI population can be more expensive to study in trial as hospital stay, treatment, and rehabilitation costs are higher. 2 Additionally, limiting studies to highly specific study regions can make recruitment more challenging.

46Another key consideration is transplant timing. Due to differences in physiology, injury etiology, cord architecture, patient comorbidities, and a host of microenvironmental cell signals, the optimal timing for transplant in animal models likely differs from humans. 15 Furthermore, even with established infrastructure, there is often a lead time associated with delivering banked cells (eg, allogeneic therapy) or generating a cell line (eg, autologous therapy). 231 Therefore, many trials are now conducted in the intermediate to chronic phases of injury where the patients' condition and neurologic status are better established and study recruitment is less complex. However, overcoming regenerative barriers in chronic SCI such as the glial scar and cystic cavitation may require some of the innovative strategies discussed previously.

47Adapting delivery techniques is a key facet of translating cell therapies. Systemic or intrathecal treatments avoid many challenges by using existing, well‐established medical techniques; however, the distribution of cells is poorly controlled. For intraparenchymal treatments (Figure 3), trials are increasingly utilizing standardized, tightly controlled delivery systems which must be scaled for human doses, provide high reliability, allow sterilization, and have undergone regulatory approval as a medical device. 232

48The two main classes of injectors currently in use are table‐mounted and spine‐mounted systems. Table‐mounted injectors 233, 234 provide a high degree of injection cannula stability along all three axes but do not account for cord movement due to ventilation, cardiovascular pulsations, or other patient movement. Spine‐mounted devices are typically immobilized on pedicle screws or by clamping the posterior vertebral arch to account for respiration but may not account for all cord pulsations. Recently, a floating cannula system has also been developed which compensates for these natural pulsations to further improve targeting. 235 A third class of devices is also under development which utilizes tools mounted on surgical robots, such as the da Vinci Surgical System (Intuitive Surgical Inc), for highly precise localization.

49The multifaceted pathophysiology of SCI and the complexities of neural repair and regeneration necessitate novel approaches to treatment. Cell‐based therapies continue to be very attractive and hopeful strategies for repair of SCI and the rapid pace of innovation continues to increase as our understanding of fundamental cell biology deepens. We predict that as the timing, dose, and delivery of adult‐ and pluripotent stem cell‐derived treatments are optimized, increasing numbers of cell‐based therapies will be translated to humans. It is highly likely that successful approaches will integrate strategies to enhance and support cells, such as genetic engineering, biomaterials, galvanotaxis, and scar degradation to maximize clinical outcomes. Ongoing preclinical and clinical trials highlight the excitement and tremendous progress that has been made in the field and underscore the importance of the collaborative work being conducted by researchers, clinicians, stakeholders, and funding agencies worldwide.

50Michael G. Fehlings declared advisory role with Fortuna Fix. The other authors declared no potential conflicts of interest.

51C.S.A., A.M.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; M.K., J.H.B., E.A.G., D.V.D.K., C.M.M., C.T.: manuscript writing, final approval of manuscript; M.G.F.: conception and design, financial support, manuscript writing, final approval of manuscript.