Spinal Cord Injury

Introduction

Spinal cord injury (SCI), which results from sudden or sustained trauma or progressive neurodegeneration, is a devastating condition whereby sufferers experience significant functional and sensory deficits, as well as emotional, social, and financial burdens. They also have an increased risk of cardiovascular complications, deep vein thrombosis, osteoporosis, pressure ulcers, autonomic dysreflexia, and neuropathic pain. The estimated annual global incidence of SCI is 15 to 40 cases per million. In the United States (U.S.), approximately 1.275 million individuals are affected, with more than 12,000 new cases each year [13].

Despite advances in pre-hospital care, medical and surgical management, and rehabilitation approaches, many SCI sufferers still experience substantial neurological disability. Intensive efforts are underway to develop effective neuroprotective and neuroregenerative strategies. Given the debilitating consequences of SCI, and the significant advances in micro-neurosurgical techniques, image-guided surgery coupled with major progress in our understanding of the biology of cell-based therapeutic strategies for central nervous system (CNS) repair and regeneration through the large number of pre-clinical studies, as well as past clinical trials, we argue that the timing is optimal to commence and extend clinical trial investigations of cellular transplants for SCI. While it is acknowledged that much remains to be learned regarding the biology of transplantable cells, it is only with a balanced investigative approach, combining preclinical research with careful controlled clinical trials, that significant advances can be made in the field [4].

The significant advances that have been made on the basis of pre-clinical studies carried out in rodent models of SCI have enabled clinical trials demonstrating the safety of cell therapy for SCI to proceed, and have informed researchers of the knowledge gaps that remain to be addressed. Given that no experimental model, rodent or otherwise, will completely mimic the human condition, it is an unrealistic hurdle or criterion for all pre-clinical findings to be validated in larger animal models, such as pigs and primates. This would be financially and practically unfeasible for most laboratories and would place such a lengthy and unnecessary delay on the clinical translation of important developments already made, as to be detrimental to the field. Indeed, this has been recognized by the U.S. Food and Drug Administration (FDA) in allowing a clinical trial of embryonic stem cell-derived cells in SCI to proceed on the basis of pre-clinical safety demonstrated exclusively in rodent models (www.clinicaltrial.gov; clinical trial identifier: NCT01217008).

Moreover, it is important to note that rodent contusion/compression models of SCI are generally “incomplete” with partial sparing of motor and sensory functions. These models closely mimic most patients with severe, partial lesions with an American Spinal Injury Association (ASIA) impairment (ASI) scale rating of AIS B or C. Given that most trials of cell therapy have been carried out in AIS A patients (the safest to inject, but also the least likely to show cell therapy-induced benefit), there is a need for future clinical trials to include patients actually modeled in the laboratory. In addition, there is a compelling need for preclinical researchers to develop valid models of compressive/contusive cervical SCI, given that approximately 50 to 60% of human SCIs involve the cervical region [2, 3, 5].

Epidemiology: Incidence and Impact on Sufferers and Healthcare

The estimated annual global incidence of SCI is 15 to 40 cases per million. In the U.S., approximately 1.275 million individuals are affected, with more than 12,000 new cases each year [1, 6, 7]. The most common causes of traumatic SCI are road traffic accidents, falls, occupational mishaps, and sports-related injuries, resulting in contusion and compression of the spinal cord [1]. Approximately 55% of SCIs occur at the cervical level (C1 to C7-T1) with a mortality of 10% in the first year following injury and an expected lifespan of only 10 to 15 years postinjury, and thoracic (T1 to T11), thoracolumbar (T11-T12 to L1-L2), and lumbosacral (L2 to S5) injuries each account for approximately 15% of SCI [1]. Depending on the age of the patient, the severity, and the level of SCI, the lifetime cost of healthcare and other injury-related expenses can reach $25 million [8]. Cell therapy can potentially enhance the quality of life of those affected by SCI, while reducing their financial and practical dependence on other individuals and ultimately on the society as a whole.

Pathophysiology

SCIs involve a primary (the physical injury) and a secondary injury (the subsequent cascade of molecular and cellular events that amplify the original injury) [9]. The primary injury damages both upper and lower motor neurons and disrupts motor, sensory, and autonomic (including cardiac output, vascular tone, and respiration) functions. Pathophysiological processes occurring in the secondary injury phase, which are rapidly instigated in response to the primary injury in an attempt to homeostatically control and minimize the damage, are paradoxically, largely responsible for exacerbating the initial damage and creating an inhibitory milieu, which prevents endogenous efforts of repair, regeneration, and remyelination. These secondary processes include inflammation, ischemia, lipid peroxidation, production of free radicals, disruption of ion channels, axonal demyelination, glial scar formation (astrogliosis), necrosis, and programmed cell death. Nevertheless, endogenous repair and regenerative mechanisms are used during the secondary phase of injury to minimize the extent of the lesion (through astrogliosis), reorganize blood supply through angiogenesis, clear cellular debris, and reunite and remodel damaged neural circuits, and as such offer exploitable targets for therapeutic intervention. The spatial and temporal dynamics of these secondary mediators (for more detail see Figley et al. [10]) are fundamental to SCI pathophysiology.

A multitude of characteristics of cells tested both pre-clinically and clinically make them ideal for SCI repair and form the basis of the possible mechanisms by which they can promote functional recovery, which are anti-inflammatory, immunomodulatory [1113], anti-gliotic [14], pro-oligodendrogliogenic [15, 16], pro-neuronogenic [17], pro-axonogenic, and secrete various anti-apoptotic and pro-angiogenic neurotrophic factors. Given the pathophysiological targets of SCI [10], transplanted cells should: 1) enable regenerating axons to cross any cysts or cavities, 2) functionally replace dead or damaged cells, and/or 3) create an environment supportive of axonal regeneration and/or myelination. However, given the multifactorial nature of SCI and its dynamic pathophysiological consequences, the success of future clinical trials of cell therapy will likely depend on the informed co-administration of multiple strategies, including pharmacological and rehabilitation therapies [10].

Cell Therapy

Cell therapy is particularly well suited to addressing the multifactorial nature of the pathophysiology of secondary SCI, and as such its potential has been a focus in regenerative medicine for many years. Different sources and types of cells, including stem/progenitor cells (embryonic stem cells, neural progenitor cells, bone marrow mesenchymal cells) and non-stem cells (olfactory ensheathing cells [OECs] and Schwann cells) have been, and/or are being tested in clinical trials for SCI. Others are still in the pre-clinical stages of testing [18] (Table 1).

Table 1 A Comparison of the Different Cell Types and Sources Currently In* Consideration or being Considered for Clinical Trials for SCI

The advantages and disadvantages of each cell source and type being considered or already in clinical trials for SCI have been extensively described and compared elsewhere [10, 1922], and reflect their potential in the clinic (Table 1). A recent systematic review [23] has highlighted the paucity of data on the transplantation of human cells into the most clinically relevant models of SCI (chronic, cervical) especially in larger-than-rodent animals and nonhuman primates. However, while the latter would be useful for optimizing surgical and cell delivery parameters and strategies, we argue that is it unreasonable to require that all pre-clinical data be verified in these models. The reason for this is that they do not replicate all the details of the human condition, in addition to being prohibitively costly and unfeasible for many, if not most, laboratories that would contribute substantially to the field through the study of well-established and better characterized rodent models.

A systematic review of pre-clinical investigations of cell therapies for SCI has recently been conducted by a consensus panel convened by the senior author (MGF) [2325]. Although we acknowledge the urgent need for further investigations using cervical models to reflect the predominance of SCIs at this level, a considerable amount of information can be obtained from an examination of the published literature, especially in thoracic SCI models, particularly with regard to safety, even though the variation in the sources of each cell type, culture conditions, age of donor, and recipient can make comparisons between studies difficult. Indeed, so far, most patients being enrolled in the phase I safety trials (including the Geron trial on GRNOP1 cells [see below]), have been those who suffered thoracic SCIs with an AIS A rating for locomotor function.

Current Clinical Trials of Cell Therapy for SCI

There are currently more than a dozen recently completed, ongoing, or recruiting cell therapy clinical trials for SCI listed on clinicaltrials.gov [26], of which 2 do not involve stem cells per se (a phase I trial of OECs in thoracic AIS A subjects [clinical trial identifier: NCT01231893], and a phase II trial of activated macrophages in cervical and thoracic SCI [clinical trial identifier: NCT00073853], subsequently suspended due to financial constraints experienced by the sponsor). Most trials are phase I or I/II clinical safety and feasibility studies, indicating that cellular treatments for SCI developed in the laboratory are still in the very early stages of clinical translation.

These trials build on the development of micro-neurosurgical techniques and image-guided approaches that facilitate the safe targeting of the injured CNS with cell-based approaches. The safety of such techniques is borne out by the experience of the Florida group with fetal cell-based transplantation for SCI associated with syringomyelia in the 1990s [27], and the more recent efforts of MacKay-Sim and St. John [22] with OECs in patients with thoracic SCIs.

The following section discusses the different cell types and sources used in past and current clinical trials, and critically evaluates the scientific evidence justifying their use in the clinic.

Bone Marrow Mesenchymal/Stromal Cells

Bone marrow mesenchymal/stromal cells (BMSCs) are isolated from the stromal compartment of bone marrow, which also comprises hematopoietic stem cells by virtue of their adherence to tissue culture plastic and/or their expression of distinct cell surface antigenic markers and nonexpression of hematopoietic stem cells markers. They have anti-inflammatory and immunomodulatory effects, mesodermal differentiation potential, and secrete several neurotrophic factors, making them attractive candidates in CNS cell rescue and as autologous transplanted cellular sources of trophic support for endogenous and co-implanted cells. Contrary to recurring claims of their neurogenic differentiation potential in vitro or in vivo, there is no conclusive evidence to support these claims [28].

Most current and recently completed trials involve autologous BMSC transplantation, given the well-established safety record of using these cells as part of bone marrow transplants for leukemia, amyotrophic lateral sclerosis, and multiple sclerosis (MS) sufferers (see as follows). Furthermore, most studies of BMSCs have been carried out in rats (likely due to practical considerations due to size differences from mice and also due to closer modeling of the clinical situation), which were found to have a beneficial effect of BMSC administration after thoracic SCI, largely as a result of neurotrophic factor secretion [29, 30] and possibly also anti-inflammatory cytokine secretion. Intra-spinal (within or adjacent to the injury site), as well and intrathecal and systemic (intravenous) delivery have been successful in most studies [3133]. Porcine and nonhuman primate studies have been carried out to further support their clinical use [34, 35], and as in rodent studies, found that BMSCs promote a certain degree of axonal regrowth and sprouting, at least in transection models [36]. This strongly supports their use in trials of thoracic SCI.

However, the use of BMSCs in SCI does present certain issues, their migration beyond the injection site (for intra-spinally delivered cells) is limited, and inter-donor variability in efficacy and immunomodulatory potency might be reflected in variable clinical outcome [37], making their evaluation as therapy for SCI difficult. Studies of BMSCs in cervical contusion-compression models have yet to be carried out. BMSCs have, in all but 2 studies by the same group, been used in subacute and acute models [3841]. Based on the limited number of pre-clinical studies in chronic models, it is not yet possible to evaluate their efficacy in the latter. It is also not known whether BMSCs provide functional preservation of axons or de novo axonal regrowth across the lesion site in contusion-compression models, as these are more difficult to distinguish in these models, unlike in transection models [4246].

Ongoing clinical studies and those carried out to date have enrolled small patient numbers and have used autologous BM-derived cells rather than purified stromal cells (understandable from a practical perspective) [4750]. A recently published dose escalation trial examined autologous BMSCs in patients with chronic SCI [51]. Although BMSCs were safe, they were not found to be beneficial in this cohort of patients.

Having clearly established the safety and feasibility of the clinical use of BM-derived cells specifically for SCI in these trials, and as suggested by Tetzlaff et al. [23] in 2011 (coauthored by MGF) based on the pre-clinical literature, we now urge the immediate testing of BMSCs in properly designed and executed clinical trials using International Campaign for Cures of Spinal Cord Injury Paralysis (ICCP) guidelines in SCI patients carefully selected as the most likely to benefit (acute and subacute thoracic SCI with a rating of AISA A or B on the impairment scale).

It is worth mentioning the Canadian Bone Marrow Transplant (BMT) clinical trials for MS led by Freedman et al. [51]. These have shown that immunomodulation through an intravenous infusion of autologous bone marrow-derived cells can not only reduce, but also halt relapses and halt the formation of new MS lesions, reduce brain atrophy, and overall reduce disease burden [52]. The inflammatory component of SCI and subsequent demyelination of surviving axons are serious limiting factors in the efficacy of early cell therapy for SCI, as implanted cells are more likely to be eliminated by the host. It is therefore worth considering interventions to modify the inflammatory milieu to enhance donor cell survival and efficacy, as in the newly initiated clinical trial of autologous BMSCs in children suffering from SCI (Table 2) (clinical trial identifier: NCT01328860). On the other hand, and considering the 1 case of neoplasms in a recipient of fetal neural progenitor cells (NPCs) reported by Amariglio et al. [53], the question of whether this strategy may increase the oncogenic risk of cell therapy must be considered. It is noteworthy that no BMT-related tumors have been reported by the Canadian MS BMT study.

Table 2 Clinical Trials of Cell Therapy for SCI Listed on www.clinicaltrials.gov

OECs

OECs can be isolated autologously from the nerve fiber layer of the olfactory bulb, as well as the lamina propria of the olfactory epithelium. They support the constant regeneration of olfactory axons from the PNS (the olfactory mucosa) into the CNS (the olfactory bulb) [54], and, as such, represent cells potentially capable of creating a microenvironment permissive for axonogenesis across the inhibitory lesion site into levels of the spinal cord caudal to the injury site.

OECs have also been used in a few clinical trials [22, 55, 56] and smaller studies that do not necessarily conform to the strict criteria and protocols of formal clinical trials [ICCP guidelines]), with some measure of success, at least from the safety and feasibility perspective [22]. These trials were initiated on the basis of promising results reported in transection SCI models [5761], which have been variably replicated [6267]. There are several possible reasons for this, including known differences between sources of OECs [68], differences in culture conditions, and their changing phenotype in prolonged culture [69]. Studies are ongoing to address these differences. Nonetheless, beneficial effects have been reported in 2 separate clinical trials for chronic SCI, admittedly undertaken by the same group [60, 61]. Of note, these trials were carried out despite the fact that only 2 studies (by the same group) have examined human olfactory bulb OECs in SCI [70, 71], which tested fetal OECs implanted 1 week postinjury in a rat model of moderate-to-severe thoracic contusion SCI [70] and hemisection SCI [71]. Six weeks after transplantation, cavitation and gliotic scarring were reduced and functional recovery was superior in human OEC-treated rats. Furthermore, OECs alone have not been found to confer functional benefit in subacute or chronic thoracic contusion SCI, but can do so when combined with Schwann cells (SCs) [7274].

For balance, clinical trials have been carried out on the basis of a small, but compelling, body of scientific evidence [75, 76] for functional benefit in SCI, even though the patients selected for these trials have not always matched the pre-clinical models used. At least 2 of these trials [60, 61] have reported functional improvement in small patient cohorts, which may be confirmed by future pre-clinical studies on human OECs of various anatomical derivations to support future clinical trials. The safe use of OECs in clinical trials for other conditions, such as ischemic stroke, further strengthens the case for their use in SCI (clinical trial identifier: NCT01327768).

NPCs

NPCs can be generated from embryonic stem (ES) cells (ESCs), which are derived from the inner cell mass of the embryo. The latter have indefinite self-renewal capacity and are pluripotent, with the potential to generate all cell types of the body, making them a potentially limitless source of cells for therapy. However, they are not without their problems (Table 1), including and especially the moral dilemmas and the practical constraints of their embryonic derivation, their karyotypic instability with repeated freeze-thaw cycles [77, 78] and their teratogenicity in the host.

Pre-clinical studies have shown that animals transplanted with human ESC-derived oligodendrocytic progenitors (OPCs) show a marked improvement in functional recovery following SCI [7984]. After observing such promising pre-clinical data, extensive pre-clinical studies were conducted to characterize the safety and efficacy of these human ESCs exclusively in rodent models [85].Footnote 1 The Geron trial, which was originally approved by the FDA, but then halted due to concerns of abnormal cyst formation, was reinitiated and approved for phase I clinical trials in the U.S. in October 2010, using human ESC-derived OPCs implanted within 2 weeks into patients with thoracic SCI, after the FDA was satisfied with pre-clinical safety data generated by Geron, but not without considerable objection and controversy [8688]. In addition to the data from Keirstead’s group (which has not been independently verified), and the unpublished data generated by Geron for the FDA, the initiation of this clinical trial is further supported by behavioral and histological data from studies implanting glial restricted progenitors (GRPs) [89] and OPCs [90, 91] isolated from embryonic and postnatal rodents in SCI models, albeit indirectly. Although these predominantly show astroglial differentiation of GRP implanted within the blunt contusion-induced thoracic lesion site, there is a shift toward oligodendrocytic specification beyond the injury site correlated with the degree of functional improvement [9294], both of which can be enhanced by transduction of factors, such as D15A, BDNF, and/or NT-3 [95]. GRP implantation was also shown to be neuroprotective and inhibited neuropathic pain.

Although more studies are required on the effects of GRPs and OPCs of various derivations in subacute and chronic cervical models, the overall picture emerging from the current literature and reports of undisclosed safety studies by Geron on ESC-derived OPCs is one that favors the continuation of this clinical trial primarily on the grounds of verification of safety, and secondarily on the grounds of efficacy in humans.

Neural progenitor cells can also be derived from several regions of the fetal, postnatal, and adult CNS, including the subventricular zone of the brain, the central canal of the spinal cord, the hippocampus, and the cortex. They can be expanded in culture as nonadherent neurospheres, which have the potential to generate all 3 neural cell types with the appropriate conditions. The key advantage of this NPC source is the amenability to in vitro manipulation (including immortalization; see below) prior to implantation, as well as the lack of tumorigenicity. However, autologous derivation of the CNS NPCs would be unfeasible for cell therapy purposes.

On the basis of promising results in the highly and clinically relevant primate models [9698]), the canine cervical contusion models of SCI, and cell number-dependent locomotor recovery in acute, subacute, and chronic thoracic rodent models (stemcellsinc.com) [99102], a recently listed clinical trial in Switzerland sponsored by the biotechnology firm StemCells Inc. is the only one currently recruiting to treat SCI sufferers with nonimmortalized fetal human CNS stem cells (HuCNS-SC) (clinical trial identifier: NCT01321333). Despite an origin of the fetal brain, they are regulated and often referred to as adult NPCs, because they are nonembryonic (stemcellsinc.com). These have been previously tested for the fatal infantile demyelination disorder Pelizaeus-Merzbacher disease (clinical trial identifier: NCT01005004) and neuronal ceroid lipofuscinosis (also referred to as Batten disease) (clinical trial identifier: NCT00337636). In the latter, they have recently been shown to persist without causing harm for as long as 2.5 years after implantation into immunosuppressed patients (http://www.internationaldrugdiscovery.com/ViewArticle.aspx?ContentID=2474).

There is a lack of trials of NPCs in SCI, in spite of the bulk of pre-clinical findings to date in support of the potential of fetal and adult neural stem/progenitor cells (particularly the former) in experimental SCI models [23], which is likely to reflect ethical concerns as to their origins and practical issues hindering their isolation and directed differentiation. Another possible explanation for the absence of clinical trials of NPCs for SCI is that the mechanisms through which NPCs provide functional benefit (including immunomodulation and angiogenesis) are only now beginning to be understood, dismissing the absolute requirement in all cases for cellular replacement by highly purified populations of pre-differentiated NPC-derived neural cells. Also, aims of axonal regeneration through the injury site have been replaced pre-clinically by more realistic objectives of remyelination [23] and provision of trophic support for endogenous precursors and axons. This makes NPCs much more promising candidates for cell therapy for SCI and probably heralds their increased use in clinical trials.

In spite of the gaps in this field remaining to be filled by further pre-clinical research, there is a compelling amount and strength of evidence along with extensive safety and efficacy testing that has been undertaken by the biotechnology firms sponsoring such trials (including Geron and Stem Cells Inc.) to gain FDA approval. These support clinical trials of fetal, adult, and ESC-derived NPCs for SCI repair, although the considerable ethical and practical challenges and concerns imposed by their origins and derivation should not be ignored.

SCs

SCs myelinate axons of the peripheral nervous system and can also remyelinate spared CNS axons as occurs after SCI. They are promising in that they are obtainable autologously and are not tumorigenic. SCs are the longest studied cell type in the cell field of transplantation research [103, 104]. It has been shown that transplantation of SCs, but also OECs and BMSCs promote the invasion of host SCs into the lesion site after SCI [105, 106], and although this might be interpreted as endogenous SCs playing a role in repair, it might paradoxically hinder the endogenous repair process [107]. The consequences of schwannosis, referring to the invasion of host SCs into the post-SCI lesion site, need further study to elucidate the possible undesirable side effects of SC transplantation.

SC transplants have been studied the most in thoracic models of contusion-compression and transection SCI. Although they are demonstrably effective at promoting sensory axonal sprouting, when transplanted alone they are not capable of stimulating corticospinal tract regeneration or of allowing axons that have penetrated the graft site to grow across and re-enter the CNS parenchymal microenvironment.

Although neonatal SCs have not been shown to be effective in cervical crush SCI [108], subarachnoid neonatal rodent SC transplantation after clip compression-induced thoracic SCI has demonstrated dramatic functional improvement after injecting only 50,000 cells into young rats (100–140 g; 45–60 days old) [109].

Of the studies on adult nerve-derived SCs, only approximately one quarter have carried out BBB assessment of functional outcome. Of these, only 2 of 5 studies in subacute and chronic thoracic contusion SCI have reported improved functional motor outcome (BBB scores) with SC transplantation alone [73, 74]. The remaining studies, which were performed by the same group, combined SCs and either OECs or other molecules [72, 110113]. In studies of SCs in transection models of SCI, all the required controls often have not been included [114], making interpretation of their results difficult.

A phase I clinical safety and feasibility trial of purified autologous sural nerve-derived has already been conducted according to ICCP guidelinesFootnote 2 in 33 stable chronic mid-thoracic (T6-9) SCI patients with AIS motor scores A to C [115], despite a lack of studies on adult human SCs without co-administered therapies in cervical and thoracic contusion spinal cord lesion models [116118]. All patients received intensive rehabilitation pre-surgery and post-surgery. Four patients aged 22 to 43 were followed up for as long as 1 year post-transplant and so far have been reported. This trial found autologous adult SC transplantation to be safe (no pathological findings were reported), although a beneficial outcome was found (as rated by using the ASIA motor scale) in only 1 patient with incomplete SCI that could not be attributed to cell treatment since the donor cells could not be localized by magnetic resonance imaging.

Similarly, human SCs have been examined pre-clinically in spinal cord demyelination models [119, 120], and they have been implanted into MS patients, demonstrating safety and feasibility, as well as remyelination [121]. These trials offer partial support for clinical trials of SCs for SCI, from the safety aspect, as well as from the point of view of remyelinating the injured spinal cord. However, the deleterious consequences of contaminating fibroblasts demonstrated by pre-clinical studies [120] will greatly impact future clinical trials of SCs for SCI. It is clear that more clinically relevant studies are required to examine the potential of human SCs in SCI, even in conjunction with other cells and therapies. Nonetheless, preliminary data from the first trials are encouraging by helping to establish the safety of SC transplantation for SCI.

Promising Cell Sources Evaluated in Pre-Clinical Studies but Not Yet in Clinical Trials

In addition to the cell sources and types mentioned so far, there are other promising cell sources with the potential to be used in cell therapy clinical trials for SCI. However, apart possibly from fetal and immortalized NPCs, their safety and efficacy in clinically relevant experimental SCI models still remain to be more extensively studied.

Human Adult and Immortalized NPCs

As previously mentioned, NPCs can be derived from the adult, as well as the fetal CNS. It is generally recognized that adult-derived progenitor cells tend to fare less well than their fetal and embryonic counterparts in terms not only of proliferative and differentiation potentials, but also in regard to postimplantation survival, migration, and integration within the recipient CNS.

However, the serious and well-founded ethical and practical concerns of using material of fetal and embryonic derivation have limited their study at the clinical level, at least to date. From our own work and that of other groups, the predominating oligodendrogenic differentiation potential of adult rodent NPCs has been demonstrated following intra-spinal [122, 123] or intravenous [124] implantation into a range of rat and murine models (i.e., thoracic contusion), compression, and cervical transection SCI. The efficacy of NPCs can be enhanced by co-administered chondroitinase treatment [122, 123]. As with some other cell types, however, there have been indications of increased neuropathic pain, which remain to be confirmed [125].

The conditional immortalization of cells might provide a partial solution to limiting the use of fetal and embryonic tissue, overcoming the limited supply of cells irrespective of donor age, removing a source of variability between clinical trials and enabling extensive characterization of cell lines prior to clinical application. Clinically approved, conditionally immortalized fetal cortical NPCs (ReN001 or CTX0E03) of human origin are already in a phase I clinical trial for stroke in Scotland, UK (Pilot Investigation of Stem Cells in Stroke; clinical trial identifier: NCT01151124). These cells have been developed by the biotechnology firm ReNeuron using a karyotypically stable immortalization platform to enable cells to proliferate in culture and differentiate when induced without oncogenic transformation. Although this clinical trial will not include SCI patients, the results of this trial will be very interesting and relevant because stroke and SCI share many common pathophysiological mechanisms.

The large differences in cells (including age of donor, immortalization strategy, culture conditions) and the models used make a direct comparison between studies challenging. For instance, no study has yet compared the efficacy of adult versus fetal NPCs lines using the same immortalization strategy and in the same SCI model; in fact, no study has yet been published on immortalized adult NPCs in SCI, an approach which would limit the use of an ethically objectionable and difficult to procure posthumous source of cells in limited supply.

There are also no direct comparisons of functional outcome between adult and fetal NPCs in SCI, and between different times of implantations following SCI. It has been shown that implantation of fetal NPCs into the acutely compressed or contused (and microenvironment of the highly inflammatory) thoracic spinal cord is unsuccessful compared to a subacute (7–9 days post-SCI) implantation time point [126]. This suggests that implanting in the subacute stages of SCI would be more likely to provide benefit. On the other hand, immortalized fetal human NPCs (ReN001; ReNeuron) have been shown to have anti-inflammatory properties (to the extent that stroke patients in the Pilot Investigation of Stem Cells in Stroke trial are not being immunosuppressed following implantation of these cells). This suggests that NPCs could potentially be implanted in the acute stages of SCI. Clearly, more pre-clinical studies are required to investigate this to determine the likely outcomes of cell implantation at different times following SCI.

Finally, NPCs are a heterogenous population of cells, and differences in culture methods and durations are likely to create and amplify differences in efficacy, depending on the particular models used, as well as the design specifics of each study examining them. However, there has been extensive pre-clinical safety testing (reneuron.com; similar to Geron and its GRNOPC1 cell line, and Stem Cells Inc. and its HuCNS-SC adult NPC line) of at least 1 line of immortalized fetal NPCs carried out by ReNeuron prior to implantation into stroke patients. This addresses important questions regarding the safety and tolerability of their clinical use.

Human Wharton’s Jelly Cells/Umbilical Cord Matrix Cells

Nonembryonic tissues, such as bone marrow, peripheral and umbilical cord blood, and umbilical cord matrix [127, 128] represent plentiful, ethical, and easily accessible sources of mesenchymal stem cells (MSCs) for neural repair [129]. BMSCs, currently the main source of autologous stem cells (apart from adipose stem cells, incubated macrophages, olfactory mucosal cells and OECs, of which the latter three are not stem cells per se), are already in several clinical trials as mentioned.

The human umbilical cord consists of an outer layer of amniotic epithelial cells enclosing a gelatinous matrix known as Wharton’s jelly cells (WJCs) [130], which harbors a stem cell population of WJCs [15]. Although WJCs are similar to MSCs from other sources (umbilical cord blood, amniotic fluid, bone marrow, and fetal blood), there are several clinically important advantages to using WJCs for cell transplantation (for more detail see Vawda et al. [131, 132]). Most importantly, WJCs are more highly proliferative and can thus be more rapidly and extensively propagated than adult BMSCs [129, 133], but unlike the latter, WJCs can undergo repeated freeze–thaw cycles without a significant loss of viability, mesodermal differentiation potential, and without accumulating karyotypic abnormalities [129, 133]. WJCs are readily obtainable without ethical constraints after normal and Caesarean births. They are thought to be nonimmunogenic, and may even have the capacity to suppress the immune response [134], potentially making them suitable for allogeneic transplantation [125]. WJCs are highly pathotropic following transplantation [135]. Of greatest relevance to SCI, they secrete a wide range of trophic factors known to promote neural cell survival (including FGF2 and SDF-1a) [136], which would make them useful for cell rescue and as support cells. Unlike ES and induced pluripotent stem (iPS) cells [137], WJCs are nontumorigenic following transplantation [133, 135, 136], and they even exhibit anti-tumor properties [138, 139].

Two studies so far have examined the use of WJCs in SCI models, but were poorly conceived and designed. Nonetheless, they did indicate that WJC administration into SCI models can potentially promote repair and recovery through the release of trophic factors [140, 141].

Fetal Human Mesenchymal Progenitor Cells

Fetal human mesenchymal progenitor cells can be isolated by cardiocentesis from first trimester fetal blood and have been characterized by Campagnoli et al.’s group [142]. Fetal human mesenchymal progenitor cells are antigenically and morphologically similar to adult BMSCs and virtually indistinguishable from WJCs, and they have already been clinically used in Sweden for osteogenesis imperfecta [143]. Given the source of fetal human mesenchymal progenitor cells from elective abortions, there are serious ethical and practical issues constraining more widespread clinical use, but their potential in SCI cannot be ignored.

Immunosuppression might not be necessary for successful donor fetal MPC survival and engraftment, not only without rejection but leading to functional benefit. In 2005, Le Blanc et al.’s [143] team in Sweden transplanted allogeneic human leukocyte antigen-mismatched male fetal liver MSCs into an immunocompetent patient with osteogenesis imperfecta (a genetic bone defect caused by dysfunctional collagen) in utero at 32 weeks of gestation. This study found successful engraftment and osteogenic differentiation of donor cells, leading to functional recovery.

Skin-Derived Progenitors

Nearly a decade ago, Miller’s group described a population of multipotent progenitor cells residing within the adult dermis (termed skin-derived precursors [SKPs]) [144146] with the potential to generate myelinating cells [105, 147152] and enable axons to grow across the lesion site (unlike SCs) when injected intra-spinally after SCI [105]. On the basis of convincing histological and behavioral data in the latter study, the potential for autologous transplantation of treated SKPs for SCI is obvious. It is interesting to note the similarity between this study and one on BMSCs, which were found to further promote axonal sprouting when treated pre-implantation with growth factors [29]. However, more studies of the effects of SKP injection in direct comparison with SCs into more clinically relevant models of SCI, including cervical and chronic models, would be required to characterize the safety parameters of this cell source and to determine the optimal implantation criteria to maximize functional benefit post-SCI.

Induced Pluripotent Stem Cells

Recently, the discovery of iPS cells has opened a new potential therapeutic approach for regenerative neuroscience, although iPS cells have not yet been used clinically in SCI cell therapy [153155].

We recently reviewed the potential of iPS cells in SCI [154]. Given the novel nature of the technology and the safety and reliability of recent variations on the induction technique, these cells are not yet ready for use in clinical trials. As with ES cells, there are well-founded concerns in regard to the teratomatous potential of iPS cells in vivo. In addition, their controlled differentiation remains to be thoroughly compared to ES cells and conditions optimized for the subsequent derivation of each neural lineage. Of note, Tsuji et al. [156] recently derived “safe” mouse iPS cells and observed trilineage neural differentiation and functional recovery in a contusion model of SCI without teratoma formation [156].

A further concern is the inherent variability of the differentiation and proliferative potential of different iPS cell lines, an issue also raised by different ES cell lines and by the low transduction rates of current transfection strategies, and the requirement for extensive retrospective characterization to ensure ES cell-like phenotype. This needs to be resolved to be able to allow the “individualization” of iPS technology to derive patient-specific lines for autologous transplantation.

Even then, the time required to generate and characterize (and if required genetically manipulate) iPSC lines would render individualized iPSC-based treatment impractical. More likely is that a set of well-characterized iPSC lines will be generated and tested pre-clinically for subsequent administration into SCI patients. In this situation, the immunogenicity of iPSCs is a major obstacle that will need to be addressed [157]. As such, although clinical trials for these cells may be a possibility in time, the current state of the field would not warrant them.

Endogenous Progenitors

Stem/progenitor cells have been identified in the central canal adult of the mammalian spinal cord [158, 159]. They proliferate extensively following SCI [160] or in response to the infusion of exogenous growth factors into the fourth ventricle of the adult brain [161], which can also reduce inflammation, and generate astrocytes and oligodendrocytes. Although the potential of stimulating the proliferation and subsequent differentiation of endogenous NPCs to effect repair is clear, it needs to be seen alongside the possible risks that supra-physiological levels of neurotrophic factors within the cerebrospinal fluid might carry, including epileptogenesis and oncogenesis.

Recommendations for the Conduct and Evaluation of Pre-Clinical Studies and Clinical Trials of Cell Therapies for SCI

The absolute requirement for international peer-reviewed assessment, regulation, independent monitoring and duplication, complete transparency, and accurate record keeping of every step of the process of designing, initiating, and executing clinical trials cannot be over-emphasized [162]. In this regard, a series of guidelines and criteria has recently been published, and the principles for ethical implementation of clinical trials in patients with SCI have been established by the ICCP panel [163166]. These include a sound evidentiary basis and compelling clinical rationale for conducting a clinical trial, registration with clinicaltrials.gov, protection of the rights of participants and volunteers through informed consent, the absence of renumeration beyond basic expenses for patients and participating institutions, the absence of charges made to the patients for experimental treatments, which should not be misrepresented as established treatments, a prospective, controlled design, objective independent outcome assessments, adequate follow-up time to monitor neurological and safety issues, and oversight by an independent authority. This will minimize the likelihood of adverse effects of treatment, and if they do occur, the investigators will be informed as to possible ways to avoid them, overcome them, or manage them in future.

Amariglio et al. [53] published a report of a young boy with ataxia telangiectasia who had received a multi-donor-derived intrathecal injection of fetal human NPCs and developed a benign multi-focal brain and spinal cord glioneuronal neoplasm. It has been suggested that the dysfunctional immune system of ataxia telangiectasia patients increased the risk of this particular recipient to develop donor cell-derived neoplasms. There are serious doubts remaining in regard to the details of this clinical case, especially relating to the karyotypic status of the implanted cells after culture and immediately prior to implantation. This is a genuine concern, which deserves further investigation, as highlighted by the oncogenic potential of ES cell-derived neural cells reported by Roy et al. [167]. However, unfortunately, it is unlikely that this particular case will yield any useful information for future studies and trials, given the lack of details on the cells, their origin, their isolation, and the implantation procedure, and any co-administered treatments.

Although we advocate a combined strategy of careful exploration of cell-based therapies in phase I and II trials in humans, and intensive pre-clinical evaluation of these approaches in cervical models of SCI, it is important to emphasize that clinical investigation needs to follow sound ethical and scientific principles. One of the key principles involves establishing a sound scientific basis for the preclinical research [23, 24, 162, 168, 169].

Concluding Remarks

The National Institutes of Health translational roadmap (http://commonfund.nih.gov/clinicalresearch/overview-translational.aspx; http://CTSAweb.org/) was drafted to validate animal models and protocols developed in the laboratory, and to drive the implementation of pre-clinical studies and basic research, thus enabling scientists to begin addressing and overcoming the issues and challenges emanating from the clinic. In this regard, immediate, further, and larger clinical trials are justified for cells already proven safe and even effective through extensive laboratory testing, and/or through past and ongoing clinical trials, such as autologous BMSCs, OECs, olfactory mucosal cells, NPCs, and ESC-derived OPCs.

Given the paucity of studies carried out in more clinically relevant models of SCI, such as chronic cervical SCI [23], current clinical trials are justified primarily on the basis of pre-clinical studies conducted predominantly in thoracic models to determine the relative safety and feasibility of the cellular interventions currently at our disposal in thoracic SCI, and to iron out neurosurgical technical and logistical issues, and also to objectively evaluate functional outcome to determine efficacy using internationally recognized and established standards.

It is acknowledged that animal models are continuing to evolve and improve so as to more closely reflect the clinical situation [68, 170177], and that better and more accurate standardized analysis techniques (including the CatWalk system [178]) will enable more objective assessment of functional outcome. However, we argue that there is a real and urgent need to press ahead with human trials on the basis of relevant solid scientific pre-clinical evidence gathered from well-established models analyzed using thoroughly validated techniques. The feasibility of this approach is amply demonstrated by the ongoing trials of ESC-derived OPCs (GRNOPC1) by Geron.

Although the case in support of clinical trials of cell therapies for SCI generates vigorous debate and divergent views, it is our strong view that the field is now at a stage in which advancement into phase I and early phase II trials is justified and essential to move the field forward. Certainly, an important principle in considering the clinical translation of a therapeutic strategy is the need for replication of results. However, this principle may be more complex when a biotechnology company is driving the translational process, given the considerations of protecting intellectual property. Moreover, the need for data from 1 laboratory to be replicated using identical conditions and cells may be an unrealistic and unreasonable challenge to meet, given the pressure on researchers to publish novel rather than confirmatory data. Therefore, there will always be differences between studies, even on the same cell type (or types) or the same model (or models) that might make a direct comparison of results difficult. On the other hand, it is precisely those differences that are the most informative and pave the way ahead toward new developments and findings preceding clinical trials.

Taking into account the various issues at hand, it is our view that, while this is by no means a clear-cut situation, there are nonetheless sufficient pre-clinical and earlier clinical safety studies of satisfactory quality and reliability carried out so far to justify the immediate translation into the clinic of a greater number of pre-clinical findings, such as the clinical trial of OECs by Mackay-Sim et al.’s [55] group, and Mackay-Sim and St. John’s [179], the work on fetal cells by Giovanini et al.’s [27] group and the phase I Procord trial on macrophages [85].

Finally, no clinical intervention is 100% risk-free. The decision to forge ahead with clinical trials of cell therapy for SCI relies on striking a balance between the current and anticipated burden of SCI and the potential risks of cell therapy. Setbacks are an expected feature of novel therapies (as in the case of bone marrow transplantation, development of the polio vaccine, and initial gene therapy trials), and as difficult as setbacks can be to contemplate and overcome, they have been overcome, and they have even provided vital clinical and scientific information that has enabled subsequent therapies to be made safe and effective. Without the clinical translation of the firmly established basic scientific findings, however, we will be stuck at the preclinical level. Advances in the application of regenerative neuroscience to SCI can only be made with an investigative approach that balances excellent preclinical research with rigorous, ethical clinical trials.