Mesenchymal Stem Cells for Spinal Cord Injury: Current Options, Limitations, and Future of Cell Therapy

1Spinal cord injury (SCI) constitutes an inestimable public health issue, with an incidence of 40–80 per million people per year. Generally, young adults are involved, where the burden of permanent neurological damage is unbearable for patients, for their caregivers, and for the health system. Prevention plays of course a key role, such as in cases of road accidents, criminal acts, or secondary causes (tumors, degenerative diseases); however, the real challenge involving scientists is about therapy, given the absence of a gold standard or effective treatment. Most of the post-traumatic degeneration of the nervous system is caused by multifactorial secondary damage including different molecular processes such as inflammation, neuronal death, ionic dysregulation, free radicals and lipid peroxidation, disconnection of normal nerve pathways, blood–brain barrier dysfunction, apoptosis, and necrosis, followed by cavitation processes and retrograde degeneration. In traumatic SCI, an early surgical decompression seems to be important in preventing secondary damage, in the range between 8 and 24 h after injury, together with spinal fixation to allow correct nursing and rehabilitation.

2One of the most important and promising tested strategies involves the use of stem cells. Among them, mesenchymal stem cells (MSCs) are particularly appealing and showed hopeful promise in preclinical research, followed by weak and conflicting results in clinical trials. In this review, the therapeutic role of MSCs is discussed, together with their properties, application, limitations, and future perspectives.

3The spinal cord of mammals is organized in ten laminae of neurons, named dorsoventrally, according to the Rexed description (1952 and 1954). The neurons are mostly multipolar and vary in size. In the dorsal laminae, sensory neurons are found which receive inputs from the dorsal root ganglion cells and project to other spinal levels or to the upper centers of the sensory pathways. In the ventral laminae, cholinergic large motoneurons are devoted to the control of muscle contraction with motor axons. Somewhat in between, interneurons of different morphologies receive the descending projections and recurrent axonal fibers from spinal motoneurons, and influence motoneuron activity (Mai and Paxinos, 2011). Spinal cord neurons form intraspinal circuits which are controlled by descending pathways. The reflex arc is the most elementary intraspinal circuit.

4The acute phase of SCI depends on the mechanism of trauma, which could be caused by contusion, laceration, stretch, compression, or direct massive destruction. The events related to the trauma constitute the primary injury, with disruption of neuronal pathways. During the immediate phase (occurring within the first two hours) (Rowland et al., 2008), neurons and glial cells at the lesion site die either by necrosis or by apoptosis (Zhang et al., 2012). Therefore, spinal cord repair should aim first to restore intraspinal circuits, and then to obtain regrowth of descending pathways to regain voluntary control of these intraspinal circuits.

5It is well known that the most crucial phase in the pathophysiological process of SCI concerns the secondary injury, which is the uncontrolled and destructive cascade occurring later with aberrant molecular signaling, inflammation, vascular changes, and secondary cellular dysfunctions.

6In the injured spinal cord, to a greater or lesser extent depending on the primary injury, a large amount of destructive processes upset the environment. Taking into account the vascular scenario, a global reduction of blood flow is observed, as a result of vasospasm, together with focal microhemorrhages or thrombosis, causing a global disfunction of the blood–spinal cord barrier. The cascade of events also affects electrolytic homeostasis around cellular membranes and their ion pumps/transporters. Potassium (K+) increases its extracellular concentration, while sodium (Na+) and calcium (Ca2+) concentrations increase intracellularly. This leads to the blockage of neuronal transmission. The influx of water caused by acidosis promotes cytotoxic edema followed by cellular death. Many molecules are released, such as free radicals and neurotransmitters. The inflammatory process involves an immune response mediated by cellular invasion after disruption of the blood–spinal cord barrier. T cells, macrophages, microglia, and neutrophils infiltrate the neuronal tissue, acquiring a proinflammatory phenotype. The environment develops with the production of cytokines such as interleukin-1 beta (IL-1β), interleukin-1 alpha (IL-1α), tumor necrosis factor alpha (TNF-α), and interleukin-6 (IL-6), recruiting more cells in loco promoting neurodegeneration.

7The chronic phase is characterized by scar formation after gliosis and deposit of the extracellular matrix. Molecules with growth-inhibitory effects are released and target neuronal receptors. Oligodendrocyte death in the primary injury seems to be a crucial point in the SCI, because myelin debris contains inhibitory molecules preventing axonal growth in animal models, such as Nogo-A protein or myelin associated glycoprotein (MAG). Proteoglycans are also involved in the chronic phase and show a different pattern of functions in the pathophysiological process; while most of them constitute a limitation for axonal regrowth with their inhibitory features, others seem to border and limit the scar, preventing further amplification of tissue damage.

8In this scenario, the removal of cellular debris and the cell environment is a key point for neuroregeneration; the modulation of macrophages, with their different phenotypes (M1 and M2) and effects in supporting neuroprotection or boosting inflammation, is then a multifactorial and crucial step in determining final outcomes.

9Noteworthy, the regeneration of neurons within the injured spinal cord seems a pipe dream in mammals, but it is innate in the axolotl (salamander) where specific molecules may regulate glial reaction after SCI and promote the proliferation and migration of glial cells to replace the missing neural tube and stimulate axonal growth. The identification of the cellular mechanisms which control neural regeneration is fundamental to promoting spinal cord repair after injury. The modulation of intraneuronal signaling networks and of the extracellular milieu is pivotal to enhance axonal regeneration, thus stimulating the regrowth of intraspinal circuits and of the descending and ascending pathways of the spinal cord. The modulation of glial scar formation and of the alterations in the perineuronal nets, and the control of neuroinflammation following SCI are mandatory for spinal cord repair, even though far from being achieved. Finally, axonal sprouting, synapse plasticity, and remodeling, in part cell-autonomous, may be differently regulated by many cells and molecules in the different compartments of the lesioned spinal cord.

10Stem cell division gives birth to an asymmetrical offspring with an additional progenitor cell and a daughter stem cell. A stem cell is able to differentiate into different phenotypes, thereby determining its potency. Totipotency is defined in the case where all terminal cell populations could be achieved, while multipotency describes the possibility to pursue a more restricted pattern of phenotypes.

11The promotion of synapse formation or axon elongation by transplanted neuronal progenitors after damage was described in animal models. Direct modulation in the differentiation of stem cells into terminal phenotypes expanded the focus of research, while promising studies showed encouraging recovery of neurological deficits after transplantation of derived cellular populations from embryonic stem cells (ESCs) in rodents after SCI.

12Since then, researchers multiplied their fields of interest, ranging from the modulation of phenotypic pathways and optimization of transplant techniques, to imaging techniques in order to obtain spatial and temporal information on the grafts, up to clinical studies starting with the Geron clinical trial which promoted the use of human ESC-derived oligodendrocyte progenitor cells (OPCs) in the site of injury. Although mechanisms are far from being elucidated, stem cell functions seem linked mostly to their paracrine effects and trophic support as shown in other neurological degenerative diseases. Given by a relatively high number of studies focusing on SCI and neuronal repair, both in vivo and in vitro, evidence shows that a combinatory strategy involving not only stem cells, but also gene therapy, biomolecular targets and drugs, and biomaterials as scaffolds could dramatically improve the functional outcomes after SCI.

13In this charming landscape, mesenchymal stem cells (MSCs) gained attention because of their easy isolation (from different sources) and preservation, raising no ethical concern, and of the limited risk of developing tumors. In the case of ESCs, indeed many ethical controversies limited their application because of the problems related to the violation of a human embryo. Additionally, MSCs maintain their regenerative potential even after cryopreservation at 80° C. Their proliferation is very rapid, and a high multilineage differentiation can be obtained. Immunoreactivity or a reaction versus hosts is minimal or absent.

14Finally, MSCs show properties of “homing”, being able to migrate toward the site of lesion (Figure 1). According to other authors, we previously observed this phenomenon in SCI experimental models both by immunofluorescence reactions and in MRI experiments. Many authors demonstrated that this cellular behavior is mediated by several inflammatory or chemotactic factors; for example, the vascular endothelial growth factor and hepatocyte growth factor, released at the injury level, can actually attract MSCs. Additionally, the SDF-1α/CXCR4 (stromal cell-derived factor-1α/C–X–C chemokine receptor 4) axis plays an important role in these mechanisms; the impairment or the upregulation of this axis can respectively affect or increase the MSC homing ability. Other factors able to positively influence MSC migration include substance P and the granulocyte colony-stimulating factor. However, the precise mechanisms justifying the homing MSC ability are still largely unknown.

15Although some differences were reported depending on the source, MSCs show a remarkable autocrine and paracrine activity (Figure 2).

16Through their secretome, MSCs can stimulate proliferation and differentiation of different cell types, including themselves. Notably, it was demonstrated that the release of growth factors, cytokines, and interleukins can also influence MSC migration (see also “homing” mechanism above), via an autocrine loop; indeed, when exposed to conditioned medium (i.e., the medium where MSCs are cultured), the MSC expression of Aquaporin 1 and CXCR4 (two membrane proteins involved in cell migration) significantly increased, by activating Akt and Erk intracellular signal pathways, and caused an enhancement of MSC migration.

17Moreover, the MSC secretome can also exert immunomodulatory, anti-inflammatory, neurotrophic/neuroprotective and angiogenetic effects on the host microenvironment (as necessary in case of SCI).

18The immunomodulation is realized thanks to the expression of the major histocompatibility complex-I on the MSC surface, in this way preventing T-cell recognition and induction of a host immune response. Moreover, MSCs are able to inhibit the proliferation, the activation, and differentiation of T cells.

19Concerning their anti-inflammatory potential, MSCs can secrete a variety of soluble molecules; among the anti-inflammatory cytokines, we can include tumor necrosis factor (TNF) β1, interleukin (IL)-13, IL-18 binding protein, ciliary neurotrophic factor (CNTF), neurotrophin 3 factor (NT-3), IL-10, IL-12p70, IL-17E, IL-27; moreover, MSCs can also modulate cytokine production of the host, for example, by inhibiting the release of pro-inflammatory cytokines (as interferon-γ and tumor necrosis factor α) or increasing the release of anti-inflammatory IL-10.

20To exert neuroprotection, MSCs secrete a number of neurotrophic factors, as brain-derived growth factor (BDNF), glial-derived growth factor (GDNF), nerve growth factor (NGF), NT-1, NT-3, CNTF, and basic fibroblast growth factor (bFGF); through these factors, MSCs can, on one side, prevent nerve degeneration and apoptosis, and, on the other, support neurogenesis, axonal growth, re-myelination, and cell metabolism.

21MSCs can also induce angiogenesis, an important process by which new vasculature sprouts from pre-existing blood vessels; to this aim, MSCs secrete the tissue inhibitor of metalloproteinase-1, vascular endothelial growth factor, hepatocyte growth factor (HGF), platelet-derived growth factor (PDGF), IL-6, and IL-8. The production of these factors is particularly important for supporting the wound healing processes.

22MSCs can be obtained from different sources, each of which bears intrinsic characteristics differences, as shown below (Figure 2; Table 1).

23These cells are found within the adult bone marrow, where they contribute to hematopoiesis and bone regeneration. BM-MSCs can not only be obtained from humans, rodents, or primates, but also from several animal species such as sheep, dogs, cats, and bovines (Figure 3).

24The possibility to differentiate into cells of mesodermal origin and to adhere to plastic distinguishes BM-MSCs from hematopoietic cells. Their range of differentiation is larger than expected, including not only mesenchymal cells such as osteocytes, chondrocytes, and adipocytes, but also a broad range of lineages expressing non-mesenchymal markers. Pre-clinical studies collected promising results (Table 2) Wislet-Gendebien et al. addressed the question of differentiation of MSCs in vitro trying to identify neuronal phenotypes. A series of markers were evaluated. Authors found Nestin expression in some groups of cells, a marker for the responsive characteristic of MSCs to extrinsic signals. In Nestin-positive cells, they also registered an overexpression of proteins like sox2, sox10, pax6, fed, erbB2, and erb4. These cells showed a neuron-like conduction, responding to several neurotransmitters (GABA, glycine, glutamate). Compared to neurons, however, trains of action potentials or synaptic activities in co-cultured Nestin-positive MSCs were not observed.

25BM-MSCs can not only be transplanted directly into the damaged spinal cord, but also infused with intravenous injections because of the aforementioned homing properties. Deng et al. transplanted BM-MSCs two weeks after dorsal SCI in monkeys. A partial functional improvement was noticed in terms of a slight motor recovery (active joints movements in the study group) and in electrophysiological studies with evoked potentials. Monkeys were then studied after three months with characterization of the scar. No neuronal cells were found. The analysis revealed the presence of markers such as neuron-specific enolase (NSE), the neurofilament (NF), and the glial fibrillary acidic protein (GFAP) in approximately 10% of the cells. The true blue, originally injected at the caudal side of injuries, was at the end traceable in the rostral thoracic spinal cord, red nucleus, and sensorymotor cortex. Zurita et al. transplanted BM-MSCs three months after dorsal SCI in pigs. Authors used a motor score (0–10, with 10 considered as animals without deficits). Twelve weeks after transplantation, pigs that underwent stem cell therapy showed a mean score of 6.2 on the motor function scale. Some of the treated animals were even able to get up spontaneously. Recovery of evoked potentials was also noticed.

26Many authors investigated BM-MSC transplantation in dogs. Among them, Ryu et al. recorded improved neurological outcomes in MCSs groups after acute transplantation (one week after trauma), since all dogs had purposeful hind limb motion. He also showed that some MSCs expressed markers for neurons (NF160), neuronal nuclei (NeuN) and astrocytes (GFAP). NF160- and NeuN-positive neurons were found, and GFAP-positive reactive astrocytes were observed more often in the control group than in MSCs groups.

27Hofstetter et al. in 2002 studied stem cell therapy after both iperacute and acute (one week) transplantation. Populations of neuron-like cells with the presence of neural markers were found, but they were not able to depolarize their membrane-like mature neurons. No clinical benefits were recorded. In the acute SCI group, a population of neuronal progenitors and astrocytes of the host were found in tissues after introduction of BM-MSCs in the lesion site.

28Other preclinical studies are reported in Table 2.

29Jeon SR et al. described one of the first applications of these cells in patients with cervical SCI (Table 3). In this case, cells were isolated from iliac bones and then subjected to intramedullary or intradural introduction after expansion in a subacute and chronic state. After six months, most patients showed a slight improvement of motor function in the upper limbs, while magnetic resonance imaging (MRI) showed changes at the level of treatment in terms of the disappearance of the cavity margin and the presence of fiber-like streaks. No evidence of neoplasm growth was observed even at three years follow-up This study and others showed promising but very limited results (Table 3). Dai et al. tested BM-MSCs in a randomized study with complete and chronic SCI patients. Neurological functions were evaluated with AIS grading, ASIA score, residual urine volume, and neurophysiological examination. In the treatment group (N = 20), 10 had clinical improvement. Mean motor improvement with AIS grading was 0.9 ± 1.07, that with the ASIA score was 11.5 ± 17.07, that with the sensory prick score was 5.2 ± 7.78, and that with the sensory light touch score was 5.4 ± 8.22. Residual urine volume (mL) was decreased with a mean of 61.55 ± 77.43. Patients were followed up for six months after an interval between the injury and stem cell therapy of 51.9 ± 18.3 months. No details about clinical improvements before stem cell therapy or other therapies were mentioned.

30In a phase I/II controlled single-blind clinical trial, El-Kehir et al. showed functional improvements over patients in the control group of stem cell therapy and physical therapy using AIS grading and ASIA scores in about half of the cases (46%), especially in patients with thoracic injuries with shorter durations of injury and smaller cord lesion. Motor recovery was recorded and promising but still qualitatively limited. Geffner et al. described a partial efficacy of stem cell therapy with some improvements in ASIA, Barthel (quality of life), Frankel, and Ashworth scoring in eight cases (four acute, four chronic). Karamouzian et al. described the results of a nonrandomized clinical trial of transplantation of BM-MSCs in 11 complete SCI patients against 20 in the control group. Results showed improvements of 45.5% of patients (a two-grade improvement from baseline, i.e., from ASIA A to ASIA C) in the study group vs. 15% in the control group, but were not statistically significant (p = 0.095). The heterogeneity and small number of the patients did not allow a reliable analysis. Mendonca et al., Park et al., and Cheng et al. (NCT01393977) described a slight functional improvement in small groups of patients treated with stem cells. Other papers and reviews questioned the extent of improvement and the correct timing of treatment. Park et al. only described six cases with some improvements in motor function and changes in cord enhancement with the MRI. Sykova et al. reported data from 20 patients with complete SCI who received transplants 10 to 467 days post-injury. Patients were then evaluated at three, six, and 12 months after implantation with ASIA protocol, the Frankel score, motor and somatosensory evoked potentials, and MRI evaluation of lesion size. Authors registered improvement in motor and/or sensory functions within three months in five of six patients with intra-arterial application, in five of seven acute patients, and in one of 13 chronic patients. Transplantation of cells appeared safe but there was no evidence that the observed beneficial effects were linked to cell therapy.

31Therefore, new trials are needed, given the absence of protocols and the poor knowledge about mechanisms and outcomes of BM-MSC transplantation. In the ongoing trials (phase I and II), hundreds of patients should be enrolled, thus trying to improve quality of evidence.

32Some studies are trying to explore benefits of different combinatory strategies involving not only BM-MSCs, but also technological tools such as virtual reality or exoskeletal stimulation to face the challenge with more holistic approaches.

33BM-MSCs showed a very promising anti-inflammatory effect on cell environment. In animal models (rats), they promoted anti-inflammatory phenotypes of macrophages (M2) and suppressed lymphocyte proliferation before sustaining regeneration. Furthermore, molecules such as vascular endothelial growth factor (VEGF), the glial cell-derived neurotrophic factor (GDNF), the nerve growth factor (NGF), and the brain-derived neurotrophic factor (BDNF) could be produced by MSCs and are currently related to the ability of MSCs to provide trophic support, studied in vivo with animal models. This is probably why the genetically modified over-expression of these factors could improve clinical outcomes. Finally, the homing properties of BM-MSCs could sustain targeted delivery of drugs acting like specific vectors.

34In clinical trials involving SCI patients, BM-MSCs were injected with an intrathecal approach in about half of the cases, while, in the remaining studies, other routes of administration were used (in situ as grafts or with scaffold, intravenous or intramuscular).

35These cells are obtained from cord blood or the umbilical cord, and grow in colonies with the support of growth factors. The risk of graft rejection using these cells is very low as confirmed by studies demonstrating their hypoimmunogenicity. In preclinical studies, using SCI animal models with rats or mouse, UC-MSCs showed a promising profile of neurotrophic, anti-apoptotic, and anti-inflammatory effects. Molecular markers and neuron-like characteristics were observed after homogeneous maturation of UC-MSCs. Despite the aforementioned results in pre-clinical studies, only few clinical trials were published describing minor improvements in some SCI patients. Kang et al. described a case report of a young female patient with slight improvements after acute transplantation. In the study of Yao et al., 25 patients with traumatic SCI (injury time >6 months) were treated with human umbilical cord blood stem cells via intravenous and intrathecal injection. The follow-up period was 12 months after transplantation. Results reported some autonomic restoration and changes in somatosensory evoked potentials. The trial of Zhu et al. showed more promising, although limited results after transplantation in the chronic phase: 13 out of 20 patients improved their motor and sphincteric functions. Five out of 20 converted from complete to incomplete (two sensory, three motor; p = 0.038) SCI. A phase II ongoing multicenter, randomized, sham-controlled trial (NCT03521336) recently started with patient enrollment, trying to evaluate efficacy of intrathecal transplantation of UC-MSCs. Completion of the study is expected in 2022.

36AD-MSCs represent an appealing source of transplantable MSCs, given the remarkable population of somatic stem cells and the availability of adipose tissue. The ability of AD-MSCs to secrete growth factors, proteases, cytokines, extracellular matrix molecules, and immunomodulatory factors supports their potential of neuroregenerative, anti-apoptotic, angiogenetic, and wound healing actions. Cellular survival pathways and repairing mechanisms in pre-clinical studies involved the upregulation of kinase proteins like ERK1/2 and Akt. AD-MSC transplantation was studied in animal models showing no adverse effects but often unsatisfactory functional results. Biomolecular and histological analysis revealed promising details. Kolar et al. studied the effect of transplantation in rats with SCI. AD-MSCs were transplanted into the lateral funiculus 1 mm rostral and caudal to the C3–C4 lesion. In animals treated with cyclosporine, BDNF, vascular endothelial growth factor, and fibroblast growth factor-2 were expressed for about three weeks. An extensive ingrowth of 5HT-positive raphespinal axons was noticed in the trauma zone with some terminal arborizations reaching the caudal spinal cord. Furthermore, sprouting of raphespinal terminals in C2 contralateral ventral horn and C6 ventral horn on both sides was observed. Relative to the lesion scar, astrocytic processes extended into the middle of the trauma zone in association with regenerating axons. Menezes et al. described an abundant deposition of laminin at the lesion site and spinal midline, the appearance of cell clusters composed of neural-like precursors in the areas of laminin deposition, and the appearance of blood vessels. Kim et al. showed a modification of the inflammatory environment after transplantation of AD-MSCs, with decreased astrogliosis-related signal molecules such as phosphorylated signal transducer and activator of transcription. Furthermore, markers like Tuj-1, Nestin, microtubule-associated protein 2, and neurofilament M were expressed as shown in other aforementioned studies. In clinical studies, a slight sensory improvement was recorded in the majority of patients after intrathecal transplantation, but longitudinal clinical trials with concrete motor responses are still lacking. Coadministration of other compounds, such as 17b-estradiol, and overexpression of Bcl-2 or chondroitinase ABC, were able to enhance therapeutic efficacy in dog models. On this premis, a series of ongoing trials of AD-MSC transplantation is currently underway (Table 3).

37These cells can be obtained from amniotic membrane or amniotic fluid. Several features are attributed to AF-MSCs such as their multipotency, ease of isolation, and ability of proliferation, together with a low immunogenicity. Despite this, only few pre-clinical studies in animal models were performed. They showed preliminary and limited results in terms of reduced inflammation and apoptosis, promoted angiogenesis, and provided trophic support. No effective clinical trials followed pre-clinical investigations.

38Due to technological advances, researchers started to investigate biomaterials with the aim of promoting tissue repair, improving stem cell survival, and supporting their functions. This strategy could be pursued using biomaterials as carriers, thereby ensuring stem cell biofactor delivery, or as a scaffold, offering a structural support for tissue regeneration.

39Among synthetic polymers, biodegradable hydrogels (such as polylactic acid (PLA), polyglycol acid (PGA), and polyethylene glycol (PEG)) were developed to promote cellular survival and carry several advantages. They easily fill the lesion cavity after injection and show high flexibility, gas permeability, no toxicity, and a favorable mechanical profile. Drugs, biomolecules, and biofactors can be loaded and released locally by hydrogels. To improve proper micro-structure and ensure correct support, three-dimensional (3D) printing nano-architecture was developed to recreate a sustainable and attractive stem cell niche. Usually, hydrogels are injected at the site of the lesion. The possibility to use them with a minimally invasive injection reduces the risk of a surgical procedure. Furthermore, they have the ability to load hydrophilic drugs and biomolecules with controlled release. Among the disadvantages, it is important to highlight that the kinetics and delivery of drugs could be inadequate because of uncontrolled diffusion or an unfavorable environment. For example, molecules with low steric hindrance cannot be controlled easily and might diffuse without reproducible control. Furthermore, the loading of hydrophobic drugs with a reduced affinity for the aqueous environment constitutes a real limitation. Bonds between drugs molecules could be built to increase or reduce the rate of release, depending on the ease of breakage of links, thereby offering controlled stem cell biofactor delivery.

40Among natural scaffolds, a variety of materials were evaluated. Because of its biocompatibility, plasticity, and flexibility, fibrin was shown to promote regeneration and delay accumulation of astrocytes at the site of the injury. Enriched with stem cells or growth factors, fibrin improved survival and migration of transplanted cells, also increasing neural fiber density. Collagen and hyaluronic acid were proposed and used because of their elasticity, time of degradation, and ability to support cell adhesion and migration. Time of degradation plays a key role in letting the matrix produced by transplanted cells progressively replace the scaffold during the repairing process. An example of an innovative composite implant was described by Rochkind et al., cross-linked hyaluronic acid with growth factors and the adhesive molecule laminin (NVR-Gel), showing promising results. Natural scaffolds were able to reduce neuro-inflammation in the acute stage and support synaptic plasticity, as well as axonal outgrowth. That said, despite this promising neuroengineering background, no concrete results were observed in clinical trials. The field of polymeric scaffolds was developed to support stem cell survival and efficacy after transplantation, but favorable biomechanical properties did not translate encouraging pre-clinical studies into clinical success, highlighting the necessity for new comprehensive experimentations.

41Stem cell technology is a growing and evolving field with an unquestionable appeal, as testified by many research papers and state-of-the-art reviews published in the English literature. In our experience with MSCs or other stem cells, like in many other aforementioned papers, tests with animal models showed promising results. Despite this, many caveats arise and, thus, elicit caution against inordinate enthusiasm.

42First of all, studies involving animal models are usually performed applying standardized protocols of lesions, treatments, and specific timings of transplantation in each group of investigation. These conditions are often inimitable in human patients with SCI, when timing and treatments are dependent on chance and emergency setting, or where lesions at the cord site could differ a lot from tailored laboratory damage. Most in vivo studies are necessarily performed with rodents, and, despite many anatomical or behavioral correspondences, human clinical trials should be the unavoidable aim of stem cell research.

43Therefore, completed human trials showed only limited results. On the one hand, the use of MSCs in SCIs seems caused no harm; different trials described, above all, the safety of stem cell therapy showing no adverse reactions or side effects. On the other hand, results in terms of clinical outcomes were poor compared to expectations. Among the others, few studies seemed particularly to encourage cell therapy.

44Dai et al. tested BM-MSCs in a randomized study with complete and chronic SCI patients. Neurological functions were evaluated with AIS grading, ASIA score, residual urine volume, and neurophysiological examination. In the treatment group (N = 20), 10 had clinical improvement. As already mentioned, patients were followed up for six months and details about clinical improvements before stem cell therapy or other therapies were not mentioned. Outcomes seemed limited, even if promising.

45In the trial of El-Kehir et al., functional improvements were noted but were confined and particularly involved patients with smaller and thoracic lesions. Geffner et al. described partial efficacy of stem cell therapy with some improvements in ASIA, Barthel (quality of life), Frankel, and Ashworth scoring in eight cases (four acute, four chronic). The other clinical studies confirmed the trend of confined (Mendonca et al., Park et al., and Cheng et al. (NCT01393977)) or no significant improvements (Karamouzian et al.). Many other trials failed to report satisfactory outcomes (NCT01186679, NCT02027246, NCT01769872, NCT01873547, NCT01624779, NCT01328860, NCT02237547, NCT01694927, and NCT01730183). There is a marked lack of large phase III trials of therapeutic efficacy, due to financial, ethical, and logistics reasons. The phase III study of Oh et al. showed weak efficacy in functional recovery, although some limitations could have compromised clinical results. For instance, only a single administration was given because of a restrictive government policy.

46Finally, even if immunochemistry, molecular markers, and morphological tracts show that MSCs, once transplanted, present neuron-like characteristics, it is hard to consider them as such; indeed, the expression of neuronal antigens can be simply due to the extremely immature nature of MSCs. Moreover, cell fusion (between MSCs and neurons) was sometimes documented; furthermore, when forced to transdifferentiate by chemical means (such as DMSO), MSCs showed evident morphological changes, which finally were simply attributed to cell shrinkage and changes in the cytoskeleton. More sophisticated protocols are continuously being developed in order to differentiate MCSs into neurons. However, currently, the efficacy of MSCs still seems in particular related to their paracrine activity, rather than to cellular replacement mechanisms.

47Ongoing trials (Table 4) will probably help researchers improve knowledge about the clinical impact of stem cell therapy. Encouraging data from preclinical experiments were not concretely translated into clinical practice. This probably reflects the multifactorial and complex physiopathology of SCI, requiring a multimodal therapeutic approach. As a matter of fact, many points need to be further clarified and depicted, as listed below. The optimal therapeutic protocols regarding the preparation, type, and number of stem cells transplanted;The timing of transplantation and route of administration;The paracrine effects and their influence on functional recovery;The importance of biomaterials and scaffold;The importance of microenvironment;The plasticity and ability to recreate connections of neuronal cells.Additionally, logistics, ethical, and financial problems related to this field of research constitute a real challenge to face in order to channel basic science studies into clinical practice.

48The optimal therapeutic protocols regarding the preparation, type, and number of stem cells transplanted;

49The timing of transplantation and route of administration;

50The paracrine effects and their influence on functional recovery;

51The importance of biomaterials and scaffold;

52The importance of microenvironment;

53The plasticity and ability to recreate connections of neuronal cells.

54Additionally, logistics, ethical, and financial problems related to this field of research constitute a real challenge to face in order to channel basic science studies into clinical practice.

55MSC therapy represents an intriguing field of research trying to face the burden of SCI. MSCs of different origin, together with scaffolds, can release immunomodulating and neuroprotective factors which may support neuron survival, axonal growth, and control of glial scarring in absence of significant side effects. Despite promising preclinical findings, clinical trials failed to keep their promises and are still far from obtaining functional recovery and restoring neural circuits. Further studies are needed to improve our knowledge on their mechanisms of action and on the cellular mechanisms preventing restoration of neural circuits after SCI, while combinatory strategies involving stem cells, biomaterials, and modifications of cell environment could be the key to translate fascinating premises into clinical practice. A better relationship between preclinical and clinical studies with a back-and-forth approach is mandatory to enhance the efficacy of cell therapy. Nevertheless, stem cell therapy in SCI injury remains an experimental therapy, possibly in association with others, and should be tested and provided at no cost for the patient. Moreover, patients should be aware of the poor clinical results obtained thus far in clinical trials to prevent exaggerated expectations and dramatic psychological consequences in the case of failure to obtain significant results.