Stem cell-based therapies for neural regeneration and repair garnered attention after the identification of specific regions of the adult human brain capable of maintaining the capacity for neuroregeneration throughout the human adult lifespan (6, 77, 93–95).

Stem cell-based techniques have been increasingly innovative, with relatively rapid advances enabling the potential to combine stem-cell therapies with previously explored pharmacological, structural, and even other cell-based methods (96–99). For example, stem cells could be modified to deliver biomolecules or to replace damaged neurons, astrocytes, oligodendrocytes, etc. and thereby act directly and/or indirectly, as noted above (100).

As illustrated in Table 1, embryonic stem cells (ESCs), mesenchymal stem cells (MSCs), neural stem/progenitor cells (NSCs), and induced pluripotent stem cells (iPSCs) have all been explored for use in cell therapies for neuroregeneration in a variety of models and applications.

ESCs were first cultured and isolated from mice (134), but can now be derived from donated human blastocysts following in-vitro fertilization (IVF) procedures (135, 136), somatic cell nuclear transfer (137), human or mice fetal brains (120, 122), or existing hESC lines (there are currently 390 NIH-approved hESC and 70 unapproved cell lines¹

ESCs are pluripotent and can proliferate almost indefinitely in vitro (135, 138, 254). Furthermore, ESCs have potential to differentiate into any cell type, including neurotransmitter or growth factor-secreting cells, neural stem cells (NSCs) and neural progenitor cells that can be further differentiated into neuronal subtypes, and/or glia (e.g., oligodendrocytes, astrocytes) capable of effecting roles in facilitating neural repair and/or regeneration (117, 120, 121, 139, 254, 255).

Early preclinical studies employing in vivo mouse models demonstrated the ability of hESC-derived neural progenitor cells to integrate into host parenchyma, migrate along established pathways in the brain, and differentiate according to region-specific cues (254). Various cell transplantation applications of hESC-derived, as well as mouse or human fetal-derived NSCs, in animal models of TBI suggest the potential of these cells to migrate to injured regions of the brain, differentiate into neurons and neuronal subtypes, and improve cognitive and motor functional recovery in the injured brain (121, 122, 139). Transplanted ESC-derived cells in ischemic animal models (e.g., rats subject to middle cerebral artery occlusion (MCAO)) have also demonstrated the ability to differentiate in vivo and to improve structural, functional, behavioral, and motor and sensory repair (123–125). NSCs and NPCs derived from ESCs have also been applied in preclinical animal models of stroke (126–131) with marked improvements in the size of the infarct area, the level of differentiation into neurons and neuronal cell types post-transplantation, and improved behavioral deficits (256).

Transplanted ESC-derived NSCs have demonstrated functional and structural improvement in animal models of SCI, as well (101, 102, 104–107, 257, 258). An ongoing phase I/II study (NCT02302157) is investigating the application of human ESC-derived oligodendrocyte progenitor cells (OPCs) in subjects with subacute cervical SCI; this clinical trial is supported, in part, by preclinical evidence of the safety and ability of these cells to promote neurite outgrowth in vitro and to facilitate myelination in vivo in rodent models of thoracic SCI (103).

Risks of inappropriate differentiation (i.e., teratoma and tumor formation), as well as technical issues arising from the possibility of immunological and graft rejection of the implanted tissues are major challenges that are still to be overcome (140–142). As well, ethical controversies surrounding the source of ESCs has led to alternative sources of pluripotent stem cells being explored (143, 144).²

Fetally-derived NSCs can be isolated from fetal brain tissue, with minute quantities also reported in bone marrow and umbilical cord blood (259). While these cells readily give rise to neurons, astrocytes, and oligodendrocytes, they are less versatile and proliferative than ESCs because they maintain some of the characteristics of the region of the brain from which they were derived³ (100, 145). Transplantation of human fetal stem cells and their derivatives have shown promise in: increasing neuronal survival and decreasing lesion size in animal models of TBI through increased angiogenesis and reduced astrogliosis (122, 147); and, neuronal regrowth, functional and structural improvement following transplant in animal models of SCI (108–110). Clinical trials involving fetal stem cells or their derivatives for the treatment of SCI and stroke are completed and ongoing. A Phase I study is underway that is investigating the safety and efficacy of human fetal spinal cord-derived NSCs (NSI-566) for the treatment of chronic SCI (NCT01772810), and a Phase II study investigating the effect of human central nervous system stem cells (HuCNS-SCs) on patients with thoracic spinal cord injury has been completed (NCT01321333). Yet another Phase I/II clinical trial is ongoing that is evaluating the effect of NSI-566 cells on motor deficits in stroke patients (NCT03296618).

Despite such progress in preclinical and clinical fields, ESC- and fetally-derived stem cells have given rise to ethical concerns regarding the source and process of their procurement, technical considerations focal to possible risks incurred during transplantation, and unregulated tissue growth in situ [for overview, see (260)].

NSCs are found in the developing and adult CNS in a number of brain regions [i.e., cortex, hippocampus, olfactory bulb, striatum, ventricles, midbrain, cerebellum, spinal cord, and retina, (31, 93, 94, 157–161). NSCs can be derived from ESCs, fetal tissues (as discussed in prior section), iPSCs, and adult brain tissue¹ (160, 261). NSC differentiation (e.g., into neurons, astrocytes, and/or oligodendrocytes) is dependent, in part upon exposure to particular growth and environmental factors.

The use of adult NSCs/NPCs (derived from biopsy tissue from brain or spinal cord) in cell therapies could, if validated in preclinical proof-of-concept studies, allow for autologous cell transplantation and repair by endogenous NSCs, thereby potentially minimizing risks of incompatibility. Early studies in animal models of SCI and TBI have suggested the potential viability of this approach (isolation, in vitro expansion, and transplantation), either alone or in combination with other methods (e.g., bone marrow-derived MSCs) for provision of neuroprotection, immunomodulation and facilitation of re-myelination and functional recovery of the injured spinal cord (149–151, 153, 154, 156).

In transplantation-based therapy, NSCs or NPCs are implanted within defined areas of the CNS¹. Transplanted NSCs and NPCs facilitate neuroregeneration and therapeutic plasticity in the injured CNS via processes of neuroprotection [i.e., the “bystander” “mechanism” in SCI, see (262) for overview; (132, 155) for ischemic stroke], neurotrophic support (262); immunomodulation (164), and cell replacement and integration (111). Broadly speaking, neuroprotection is the activation of pathways that prevent further neuronal cell death, while neurogenesis involves proliferation and differentiation of endogenous neural stem and progenitor cells (259, 262), or implanted NSCs capable of cell replacement. Lu et al. (113) found improved motor functionality following SCI in rodents, when rat- and human-derived NPCs, human ESC-derived NSCs, and human iPSC-derived NSCs were implanted at injury sites (112, 114, 115, 263). Specifically, it was observed that these cells were able to extend multiple axons over long distances, and form synapses with host neurons (112, 264). Research suggests that endogenous NSCs and the immune system share secreted mediators (chemokines, cytokines) and receptors relevant both to inflammation and the maintenance, structure, and function of endogenous neural stem cell niches in the brain (165, 166, 265). Transplanted exogenous NSCs may modulate local immune responses and secrete growth factors that are conducive to neuronal growth (132, 133). Indeed, both in vitro and in vivo proof-of-concept studies have demonstrated the potential of NSCs as a cell-based therapy for SCI (266–268), stroke (269), and TBI (266, 268, 269), based, at least in part, upon the capability for defined differentiation and target specific integration and functional activity of these cells when introduced at or proximate to sites of neural injury (259).

However, NSCs and NPCs from adult tissues cannot be easily engaged to generate certain types of neurons [e.g., dopamine or motor neurons; (162)]², nor are they easily isolated or expanded from the regions of the human brain that contain NSCs [i.e., the subgranular zone of the hippocampus, and the subventricular zone of the striatum; (94)], in part, because they are present in rather low numbers (100). Moreover, despite the viability and potential value of transplanted NSCs as a therapeutic approach, gaps exist in knowledge about the activity and effect(s) of NSCs in vivo with regard to migration mechanisms and the optimum number of NSCs needed to facilitate regeneration in different types of lesions (166, 259). The method and timing of delivery, and potential interactions between the transplanted NSCs and the host immune system may also impact the induction and extent of neuroregenerative effects produced by these cells (160, 166).

The CNS environment at the site of neural injury in SCI, TBI, or stroke is not conducive to neuroregeneration (270). This is due to the activation of inhibitory pathways, glial scar formation, and the lack of guiding astrocytes essential for axonal regrowth (6). As a result, many applications of NSCs have resulted in poor cell survival, failed integration into host tissue, and poor, if not uncontrolled, differentiation of cells (6, 153, 271). Additionally, while NSC transplants derived from ESCs have shown increased risk for formation of tumor growth compared to NSCs that have been derived from either fetal or adult brain tissues, that is not to say that these cells do not also have associated risks of tumor formation¹ (100, 260). Furthermore, given that a source of NSCs is from ESCs and fetal tissues, research and potential therapies using these stem cells generate concerns about derivation, cost, supply, and accessibility of these cells (11, 272, 273). In sum, these challenges, as well as mixed evidence of safety and efficacy in animal studies (152, 153, 274), have hindered translation of adult brain-derived NSCs/NPCs into clinical therapies (100, 259, 260).

Researchers successfully demonstrated that induced pluripotent stem cells (iPSCs) could be generated from mouse embryonic fibroblasts (190) and from adult fibroblasts (189, 191, 192, 199) that are cultured in the presence of some combination of four transcription factors (Oct4, Sox2, Klf2, and c-Myc; or Oct4, Sox2, Nanog, and Lin28)—rather than from embryonically- or fetally-derived cells (11, 144, 190, 191). Since then, iPSCs have become an attractive source of pluripotent stem cells for use in neural replacement and regeneration therapies, given their potential to generate a “virtually limitless supply” of autologous iPSCs—thus resolving ethical concerns regarding the source of stem cells and risks associated with immunological rejection in transplant recipients (11, 196). In neural therapies, iPSCs and their progeny can facilitate regeneration of neurons, induce axonal remyelination, provide trophic support and immunomodulation, and modify the extracellular microenvironment (105, 167, 177, 187).

The first step in generating iPSCs is to select an appropriate type of donor cell (201). Currently, the majority (i.e., 80%) of published studies of cellular reprogramming employ fibroblasts in that they are easy to obtain, purify, maintain, and have significant potential for autologous transplantation (11, 191, 273). Other somatic donor cells include: melanocytes, adipocytes, CD34+ cells, mesenchymal stem cells, human primary keratinocytes, umbilical cord blood CD133+ cells, and peripheral blood mononuclear cells (199, 208, 273, 304–307).

Each of these cell types has a characteristic differentiation profile related to its epigenetic memory (201, 308, 309). Such “memory” (e.g., particular patterns of DNA methylation, histone acetylation and phosphorylation) can influence patterns of expression and differentiation profiles of the resultant pluripotent stem cells (201, 309). Cells of different origins can also differ in their reprogramming efficiencies (11, 310). When designing cell therapies for neurotrauma, it is important to consider the donor cell source in light of the ease of isolation/expansion, the differentiation profile, the pathology of the CNS injury of interest, and the intended destination of the cell therapy in the injured patient.

Skin fibroblasts, for instance, are easily obtained but can require more time, resources, and manipulation to generate iPSCs due to their low programming efficiency (311, 312). Given the increased time required for reprogramming, they also have potential for increased mutations in vitro. Keratinocytes and melanocytes show therapeutic potential as sources of iPSCs: both are relatively easy to obtain from skin biopsies or plucked hairs, have relatively short reprogramming periods and high reprogramming efficiencies (203). Some research suggests the potential for these cells to be reprogrammed to neural stem cells, though further research is needed to improve the reprogramming process and validate its potential for differentiation and neuroregenerative applications (11, 313). Use of CD133+ cells from umbilical cord blood has been explored due to these cells ‘relatively high reprogramming efficiency and pluripotency (205); however, their use in therapies for CNS insult remains questionable, as there remains a need to characterize their NSC differentiation potential and optimal culture conditions (11, 314). CD34+ cells, derived from bone marrow and umbilical cord blood, are not recommended for clinical use due to their low availability and low reprogramming efficiency (206, 207). Stem cells derived from adipose tissue show notable promise as a reliable source of iPSCs: they are easily obtained during liposuction procedures, require fewer genomic changes to generate iPSCs, and the reprogramming period is both relatively short and efficient (208, 315).

Since the publication of studies by (189), a number of protocols for the development of iPSCs have become available that use these various adult-derived donor cells (some autologous) and employ different combinations of transcription factors (TFs) and delivery methods—with the goal of finding a protocol that optimally balances efficiency, safety and related risks (201, 311–321). Some TF delivery methods involve the use of viruses, such as lentiviruses, to integrate viral DNA into replicating regions of the host genome (322). The use of viral vectors to deliver reprogramming factors can, however, lead to chromosomal disruption, expression of transgenes, and/or mutations (322). More recent “scarless” technologies have been introduced that use episomal vectors, synthetic mRNAs and Sendai viruses that do not integrate to the genome during reprogramming of adult somatic cells into iPSCs (198, 323–325). Kim et al. (199) developed a vector-free method to generate human iPSCs directly from human fibroblasts through delivery of reprogramming proteins (Oct4, Sox2, Klf4, and c-Myc) fused to cell penetrating peptides capable of crossing the cell membrane (199). These “scarless” and vector-free systems reduce potential risks of chromosomal disruption associated with DNA transfection and genome manipulation.

Studies in animal models of SCI that involve transplantation of human NSCs derived from iPSCs also suggest that difficulties remain in ensuring in vivo differentiation of cells into appropriate neural cell lineages and with tumorigenicity—both of which may hinder successful cell engraftment and functional recovery (326, 327). Evidence suggests that the local CNS microenvironment and pro-inflammatory factors (and associated neuroinflammation) can influence transplanted cell engraftment, differentiation and survival (326). In light of this, optimization of protocols for human cell reprogramming and the generation of iPSCs for use in clinical trials remains a work in progress in an attempt to achieve balance in safety and efficiency when applied in humans for certain types of neurotrauma (11, 144, 328).

Cell reprogramming studies published in the late twentieth and early twenty-first centuries that demonstrated transdifferentation challenged the widespread notion of predetermined cell fates (329–332). Indeed, NPCs/NSCs can be generated from iPSCs, but can also be directly generated from somatic cells (e.g., fibroblasts) via a process known as direct reprogramming (333, 334). Lujan et al. (335) showed the direct conversion of MEFs to NPCs using reprogramming factors (Sox2, FoxG1, and Brn2). Somatic cells can also be transdifferentiated to become mature neuronal cells [e.g., neurons, glia; (336, 337)]. Vierbuchen et al. (336), for instance, demonstrated that mouse embryonic fibroblasts (MEFs) could be directly converted to neurons. The same and other iterative combinations of factors (e.g., Musashi-1 (Msi-1)⁵) and techniques (e.g., silencing of donor cell-specific transcriptional programs, chromatin remodeling) to optimize cell transdifferentiation, generation and stability have been further identified (339–343).

Direct reprogramming via chromatin remodeling involves opening the chromatin in somatic cells, and facilitating sustained exposure to a set of transcription factors that effectively converts the somatic cells to NSC/NPCs; chromatin remodeling enables NSCs/NPCs to maintain their new identities and bypass the intermediate pluripotent state (337, 344, 345). For direct reprogramming approaches, bypassing the intermediate pluripotent state reduces the associated reprogramming risks of genetic and epigenetic abnormalities (346, 347). Continuing research is further exploring those combinations of TFs that could be most effective in direct reprogramming somatic cells into neuronal subtypes (e.g., dopaminergic, GABA-ergic, glutamatergic, and motor neurons), as well as distinct types of glial cells [i.e., oligodendrocytes, astrocytes, and Schwann cells; (325, 348)].

Several important questions remain regarding reprogramming of human cells that must be considered pursuant to their use in clinical applications. For example, it will be important to discern to what extent iPSCs retain epigenetic memory of their donor cell types, and whether the induced pluripotency is equivalent to that seen in ESCs (144, 309, 351). It is also necessary to determine the extent to which different reprogramming methods (choice of donor cell and iPSC line, cell selection during propagation, and culture conditions) influences genetic variability and functionality of the differentiated cells (144, 310). While genetic variability between iPSC lines could theoretically be resolved via gene editing, human pluripotent stem cells have demonstrated resilience to conventional gene targeting methods, and even those methods that have worked (352–354) have been shown to be both inefficient and time consuming (144). Advancements made in the use of novel gene-editing systems, such as CRISPR/Cas-9 and its variants, may offer a possible solution to this demonstrated resilience in iPSCs to targeted gene-editing (144, 355).

Concerns regarding mutations and tumorigencitiy are focal to any consideration of the development and use of reprogrammed, differentiated iPSCs. iPSCs can acquire mutations during both the reprogramming process and in vitro expansion; furthermore, some protocols for generation of iPSCs call for the use of c-Myc—a proto-oncogene associated with various types of cancers (325, 356, 357). Risk of tumorigenicity and accumulation of genetic alterations might be reduced by selecting appropriate cell source(s) for the generation of iPSCs (i.e. adipose cells, keratinocytes, or melanocytes) (11, 310). Direct reprogramming may also offer a way to reduce these risks, as reprogrammed cells do not require an intermediate pluripotent stage (329–332).

Yet, even with sufficient controls for mutation during these phases of development, there remains a definitive need for an efficient technique for evaluating differentiation and screening for mutagenesis and potential tumor- and/or teratoma formation both prior to and following transplantation. Antibodies that target high-risk cell subpopulations and selective in vivo ablation of transplanted cells are potential solutions being explored to reduce this risk (358), and further studies are needed to more fully define the effectiveness of this, and other methods of insuring stability of these cells.

A growing body of literature describes the numerous pathways of communication that exist between the immune system and the CNS—namely, the role that such communication has in the CNS' ability to maintain homeostasis and to react to neurological injury and trauma [e.g., neuroinflammation, (359–362)]. Communication between these two vital systems is mediated primarily through immune regulatory molecules that are released by cells residing in the CNS, such as microglia, mast cells, astrocytes, and oligodendrocytes (361–365).

Following neurotrauma, the CNS undergoes pathophysiological changes that both affect and are affected by the innate and adaptive arms of the immune system. After acute stroke, for instance, brain ischaemia, damage to tissue and subsequent release of pro-inflammatory factors (e.g., matrix metalloproteinase-3 (MMP-3), ATP, alpha-synuclein, and neuromelanin, (366, 367)) can activate innate immune (via leukocytes, toll-like receptors, and the lectin pathway) and neuroinflammatory responses, such as the release of cytokines and activation of resident microglial cells (359, 365, 368). T and B cells mediate the adaptive immune responses to neurological damage, providing both potentially protective and potentially damaging autoimmune/reactive effects on neural tissue, if inflammation is prolonged and unregulated [(362, 365, 369–371); for a thorough overview of the innate and adaptive immune responses to neurotrauma, see (359, 371)].

For cell therapies intended to treat the effects of neurotrauma, neuroinflammation and the potential for an immune response must be considered. In some cases, the administration of antibiotics, immunosuppresants, or other small-molecule anti-inflammatory agents [e.g., N-palmitoylethanolamine (PEA)] may be required, particularly following or alongside transplant or injection of tissue/cells into injured patients (362, 372, 373). However, one must also consider the effect that administration of such immunosuppressants may have on the stem cell therapies, cell differentiation and cell engraftment, as well as the optimum time for delivery of these substances (374).

In theory, the use of autologous adult somatic cells to generate iPSCs and progeny would avoid or at least significantly reduce immunogenicity in stem cell transplant therapies. However, Zhao et al. (375) suggested that transplantable stem cells could be subject to aberrant DNA methylation during the reprogramming process, and that such changes to the DNA might trigger an immune response after transplantation in the host. Other potential sources of immunogenicity may include culture reagents, protein expression, maturity of transplanted cells, or factors of the micro- or macro-environment at the site of transplantation (376). Others have explored the potential immunogenicity of autologous iPSCs using mouse (377, 378) and primate models (379), and these studies appear to contradict the results reported by Zhao et al. (375). Such ambiguity establishes the need to more stringently assess potential sources of immunogenicity in iPSC-derived cell therapies prior to their clinical use.

Interest in stem cell research is sustained by the potential contribution that these cells and their progeny can make to transplant-based therapies for neurological repair and regeneration (260, 380, 381). Indeed, research suggests that stem cell transplantation affords potential as a therapeutic strategy for central neural injury and disease (260, 382, 383). Initially focused on use of ESCs as sources of pluripotent stem cells, the field was soon rife with ethical debate related to controversies surrounding the sources of ESCs, implications relevant to the status of early-stage human embryos (380), limited supply, high cost, and associated risks of tumorigenicity and immune rejection after transplantation.

In light of this, more recent basic and preclinical studies of regenerative therapies for CNS insult have shifted toward use of iPSCs and their derivatives that are generated from adult somatic cells. As noted above, iPSC-derived and directly reprogrammed neuronal and glial cell types can be generated from cell sources that are not (or at least less) controversial and more readily available. Such cell sources can be autologous—derived from tissue samples directly from patients—and thus, can eliminate or significantly reduce risks of immunological rejection after transplantation, and may appeal to current incentives for personalized and precision medicine. As well, developments have been made in reducing time spent in vitro during the reprogramming processes, improving selection of cell type, and increasing efficiency of methods to remove harmful cells pre- and post-transplantation, which, when taken together, can reduce risks of tumorigencity in iPSC transplant-based therapies.

To be sure, autologous reprogrammed cells have a number of attributes that make them attractive and potentially viable alternatives (to ESCs and fetal stem cells) in cell-based therapeutic approaches for particular types of CNS injury and disease. Yet, while this approach has potential to fulfill therapeutic goals of enabling partial to complete recovery of defined functions, ethico-legal and social issues persist. As described in detail elsewhere (260, 384, 385) such issues involve the unknowns of intermediate and long-term use of new techniques and technologies in practice. For example, (1) how such unknowns impact the validity and contingencies of informed consent; (2) the need for continuity of research and clinical care (to address emerging issues consequential to sustained use-in-practice); (3) legal liabilities and process of claims if/when adverse events occur; and (4) provision/distribution of resources and services (to furnish cells, as well clinical interventions, and ongoing support) in and across various populations in need.

We have proposed a risk assessment approach to more specifically define and address the issues, questions and problems generated by possible uses of emerging neuroscientific and neurotechnological developments (386, 387); but address does not assure resolution. While some concerns relate to safety and effectiveness (e.g., possibilities of unanticipated consequences, runaway effects) with regard for patients' welfare (i.e., via integrity of informed consent and availability of sustainable clinical care), others reflect recognition of problematic distribution of medical goods [for overview, see (388)]. Both safety and distribution can be met, to some extent, by rigorous quality control and expediency in the manufacture and delivery of cells. Still, it remains to be determined if the type and extent of broad-based medical care (i.e., trained personnel and institutional resources) can and/or will be made available to perform the procedures and insure sustained post-implantation evaluations and care.

We have opined, and re-iterate here, that any consideration of the ethics of biomedicine must regard economic factors, and in many cases, economics of biomedical research and care are determined by policy and law (389, 390). Here discussion of the viability and value of NSCs—or any biomedical technique and/or technology—centers upon the interactive roles of regulatory policy in establishing both standards of care and fiscal (i.e., insurance) subsidy of these approaches in patient care. Suffice it to note that the gravitas—and international relevance and importance—of this situation is such that a complete discussion of economic and policy issues focal to NSCs or translational neuroscience would be significant, and is therefore beyond the scope of this paper.

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