Macrophages
Due to the immune privilege, recruitment of macrophages is limited in CNS and the resident microglia cells are the main immune cells that are activated after SCI [
19]. It has been shown that controlled boosting of local immune response by delivering of autologous macrophages, which were alternatively activated to a wound-healing phenotype, can promote recovery from the spinal cord injury. Initial experiments with implantation of macrophages activated by preincubation with peripheral nerve fragments lead to partial recovery of paraplegic rats [
32]. Improved motor recovery and reduced spinal cyst formation of rats was also observed by implantation of macrophages activated by incubation with autologous skin [
33]. The postulated mechanisms are activation of infiltrating T cells, and increased production of trophic factors such as brain-derived neurotrophic factor (BDNF) [
33,
34] leading to removal of inhibitory myelin debris [
32]. Promotion of a permissive extracellular matrix containing laminin is another observation [
34]. Following these and subsequent positive results from animal experiments, autologous macrophages activated by incubation with autologous skin, under the brand name of ProCord, were entered into a multicentric clinical trial. The results of phase I studies show that out of eight patients in the study, three recovered clinically significant neurological motor and sensory function. Also, it has been shown that this cell therapy is well tolerated in patients with acute SCI [
35].
Dendritic cells
In animal model studies, transplantation of dendritic cells into the injured spinal cord of mice led to better functional recovery as compared to controls [
36]. The implanted dendritic cells induced proliferation of endogenous neural stem/progenitor cells (NSPCs) and led to de novo neurogenesis. This observation was attributed to the action of secreted neurotrophic factors such as neurotrophin-3, cell-attached plasma membrane molecules, and possible activation of microglia/macrophages by implanted dendritic cells [
36].
Dendritic cells pulsed (incubated) with encephalitogenic or non-encephalitogenic peptides derived from myelin basic protein when administered intravenously or locally to the site of injury, promoted recovery from SCI [
37]. The mechanisms proposed to explain this phenomenon is based on presentation of the loaded antigen to the naïve T cells by dendritic cells. The stimulated T cells start a cascade of events leading to "beneficial autoimmunity". They may secrete growth factors that protect the injured tissue. Also, they lead to a transient reduction in the nerve's electrophysiological activities, decreasing nerve's metabolic requirements and thus preserving neuronal viability [
38]. This explanation is in line with the finding that in those rats, which are unresponsive to myelin self-antigens, the outcome of CNS injury is worse than normal rats [
39].
Olfactory ensheathing cells (OECs)
Olfactory ensheathing cells (OECs) are glial cells ensheathing the axons of the olfactory receptor neurons. These cells have properties of both Schwann cells and astrocytes, with a phenotype closer to the Schwann cells [
40]. OECs can be obtained from olfactory bulb or nasal mucosa (lamina propria). Cells from both sources have been used for treatment of spinal cord injury in animal models. Those from olfactory bulb origin lead to axonal regeneration and functional recovery after transplantation to animals with transected [
41,
42], hemisected [
43,
44] or contused [
45] spinal cords. Similar results were also obtained by transplantation of OECs isolated from lamina propris in both transected [
46] and hemisected [
47] models. It has been shown that these cells are able to retain their regenerative ability after cryopreservation [
48] and after establishment of a clonally derived cell line [
49]. Boosting of regenerative capability of OECs by overexpression of brain-derived neurotrophic factor (BDNF) [
50] or glial cell line-derived neurotrophic factor (GDNF) [
51] was also tried successfully in animal models.
OECs migrate after implantation [
52], decrease neuronal apoptosis [
53] and secrete a number of extracellular matrix molecules such as type IV collagen, and the chondroitin sulfate proteoglycan NG2 [
54]. They also secrete trophic factors such as vascular endothelial growth factor (VEGF) [
54], nerve growth factor (NGF), and BDNF [
55]. Remyelination is also increased after transplantation of OECs [
56‐
58]. A comparison of acute versus delayed transplantation of OECs has shown that acute transplantation leads to earlier recovery and better functional and histological results [
59]. The efficacy and behavior of olfactory bulb-derived cells were compared with lamina propria (LP)-derived cells after implantation. LP-derived cells showed superior ability to migrate within the spinal cord, and reduce the cavity formation and lesion size, but they enhanced autotomy [
60]. All the above properties can explain the observed histological and functional improvements following transplantation of olfactory ensheathing cells to the site of injury.
According to the promising results obtained from animal experiments, several clinical trials have been started. In a large series more than 400 patients underwent transplantation of fetal olfactory bulb-derived cells, of which the results of 171 operations were published [
61], showing functional recovery, regardless of age and as early as the first day after implantation [
61]. But, an independent observational study of 7 cases from this series did not report any clinically useful sensorimotor, disability, or autonomic improvements [
62]. In a recent case report, a rapid functional recovery was noted within 48 hours of transplantation of olfactory bulb-derived cells [
63]. This reemphasizes the need for further studies into the mechanism of action of these cells, as according to the animal studies, such a rapid start of improvement is not expected. Nasal mucosal-derived OECs were also used in a phase I clinical trial conducted on 3 patients who were followed for one year after transplantation [
64]. The results confirm the safety and feasibility of this approach.
Schwann cells (SCs)
Schwann cells originating from dorsal and ventral roots are one of the cellular components that migrate to the site of tissue damage after spinal cord injury [
65‐
68]. The remyelinating capability of Schwann cells has been demonstrated in a number of studies [
66,
69] and the functioning status of this myelin in conduction of neural impulses was confirmed [
70,
71]. SCs promote axonal regeneration by secretion of adhesion molecules such as L1 and N-CAM, extracellular matrix molecules such as collagen [
72] and laminin (see Chernousov and Carey [
73] for review), and a number of trophic factors such as FGF-2 [
74], nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF) and NT3 (see Mirsky
et al [
75] for review). In addition to their on neural regeneration and remyelination, a number of unwanted effects were also reported following the use of these cells. It has been shown that when SCs come into contact with CNS astrocytes, their migration into the CNS is stopped [
76]. Also, corticospinal tracts (CST) show a delayed and poor regenerative activity in response to Schwann cells implantation when compared with OECs [
77]. The other unwanted issue in regard to SCs is that the damaged axons, which are stimulated by these cells to regenerate, grow into the grafted population of Schwann cells, but there is little evidence to support that they leave these cells and re-enter their original white matter pathways [
70]. When combining SCs transplantation with delivering of neurotrophic factors [
78] or OECs plus chondroitinase [
79] exit of regenerating axons could be observed from the transplanted population of grafted cells.
In animal model studies, Schwann cells are isolated from either newborn or adult sciatic nerve and cultured in the presence of mitogens. Upon transplantation to the damaged spinal cord of adult animals, they stimulate tissue repair by causing regenerating axons and astroglia to express developmentally related molecules. When compared with the effects of OECs in an acute SCI setting, it was concluded that the degree of functional recovery achieved by SCs is less than OECs [
80]. It has been shown that delayed transplantation leads to a higher survival of SCs in host tissue as compared with acute transplantation; meanwhile, implanted Schwann cells cause extensive infiltration of endogenous SCs to the site of injury [
81]. Schwann cells are usually transplanted by direct injection to the site of injury, which can add to the inflammatory process in the region. Recently, as an alternative route, transplantation to the subarachnoid space was tried and led to a favorable outcome [
82]. The results of a phase I human clinical trial in patients with chronic SCI will be presented in the next annual meeting of the Congress of Neurological Surgeons in Chicago [
83].
Neural stem cells in CNS
Neural stem cells (NSCs) are present in adult and developing central nervous system of mammals and can be isolated and expanded
in vitro [
84]. Neurosphere technique is the most common method for isolation of NSCs. Using this technique, stem cells have been isolated from developing spinal cord [
85], cerebral cortex [
86] and brain [
87], and from adult subependymal, subventricular zone of the lateral ventricle [
88,
89], cerebral cortex [
90] and spinal cord [
91]. Also, there was a widely held assumption that dentate gyrus of the hippocampus contains neural stem cells in adults. But, it has been shown recently that dentate gyrus is a source of neural-restricted progenitors (NRPs) and not multipotent stem cells [
92]. NRPs are different from neural stem cells as they are committed to neural lineage at time of isolation. It has been shown that NSCs differentiate to neural and glial cells both
in vitro [
93,
94] and
in vivo [
93‐
95]. Also, following a clonal study, it has been reported that neural stem cells from the adult mouse brain can contribute to the formation of chimeric embryos and give rise to cells of all germ layers [
96].
The fate of
in vivo differentiation of neural stem cells depends on the niche they have been transplanted to. When transplanted into a neurogenic region e.g. dentate gyrus [
95,
97] or subventricular zone [
97], they will differentiate into neurons. Transplantation into other, so called, non-neurogenic regions, such as spinal cord [
94], will induce them to differentiate into glial cells. Although a few studies report limited differentiation in non-neurogenic regions [
84,
85], most reports are consistent with differentiation into glial fate. This shows the importance of environmental cues in directing the differentiation of NSCs. NRPs isolated from fetal spinal cord were transplanted into normal and injured spinal cord and differentiated into neurons in normal cords. But, the injured spinal cord niche restricted their differentiation and the cells remained undifferentiated or partially differentiated in this niche [
98]. In an interesting study, a mixed population of NRPs and GRPs were transplanted into the injured spinal cord. The mixed population was provided by either direct isolation from fetal spinal cord or pre-differentiation of NSCs
in vitro. This approach resulted in generation of a microenvironment that led to an excellent survival, migration out of the injury site and differentiation of the cells into both neural and glial phenotypes [
99,
100]. Functional improvements have been reported after transplantation of NSCs derived from embryonic spinal cord [
85] and brain [
101], adult brain [
102] and spinal cord [
103], and a mixed population of NRPs and GRPs isolated from fetal spinal cord [
104].
Hematopoietic stem cells and marrow stromal cells
As hematopoietic stem cells (HSCs) and marrow stromal cells (also known as mesenchymal stem cells) (MSCs) are more accessible than other cells mentioned in this review, they have attracted much attention as the potential cell sources in management of spinal cord injury. Bone marrow is a rich source of these cells; although, HSCs have also been obtained from umbilical cord blood [
105] and fetal tissues [
106].
Much of the evidence used to support the potential of HSCs and MSCs to differentiate into neural and glial cells comes from
in vivo studies. Transplantation of unfractioned bone marrow has led to detection of bone marrow-derived cells that expressed neural markers in CNS, in both animal models [
107‐
109] and humans [
110,
111]. In a recent clinical trial [
112] bone marrow cells were delivered to patients with acute and chronic SCI intravenously or via vertebral artery. The study demonstrated the safety of the procedure. Partial improvement in the ASIA score and partial recovery of electrophysiological recordings of motor and somatosensory potentials have been observed in all subacute patients (n = 4) who received cells via vertebral artery and in one out of four subacute patients who received cells intravenously. Improvement was also found in one out of two chronic patients who received cells via vertebral artery. In another clinical trial unfractioned bone marrow cells were transplanted in conjunction with the administration of granulocyte macrophage-colony stimulating factor (GM-CSF) in six complete SCI patients and followed for 6–18 months. The procedure was safe and led to sensory improvements immediately. Also, AIS scores improved in 5 patients [
113].
As unfractioned bone marrow is a mixture of different progenitor cells that might show different behavior in the same condition, more detailed studies have been performed on isolated fractions of HSCs and MSCs. Derivation of cells which have been phenotypically defined as neurons [
106,
114] and glial cells [
105,
106] has been reported after
in vitro differentiation of HSCs. But, the point to be remembered is the fact that subsets of hematopoietic stem cells express neuronal and oligodendroglial marker genes [
115,
116] and this should be considered in interpretation of results of any differentiation study.
It was reported that transplanted hematopoietic stem cells transdifferentiate
in vivo into neurons and glial cells without fusion [
117]. But, dissimilar results were obtained from
in vivo transdifferentiation studies. For example Koshizuka
et al [
118] have shown that HSCs only differentiate into glial cells not neurons. Lack of transdifferentiation into neurons, which is a matter of controversy [
119‐
121], was also reported by Wagers
et al [
122] and Castro
et al [
123]. A recent electrophysiological study on neuron-like cells derived from HSCs failed to detect generation of action potentials in these cells [
124]. But, locomotor improvement has been reported in the mice with contused spinal cord after transplantation of hematopoietic stem cells [
118,
125]. Also, it was shown that implantation of HSCs into developing spinal cord lesion of chicken embryos directs these cells to differentiate into neurons with no apparent fusion to the host cells [
126]. These apparently disparate findings may be due to the issues such as the employed technique, the subpopulation of the HSCs used, and the experimental model. A phase I clinical trial in which CD34+ cells were delivered into the injured spinal cord via lumbar puncture technique demonstrated feasibility and safety of the procedure after 12 weeks of follow up [
127].
The capacity of marrow stromal cells (MSCs) to differentiate
in vitro into cells expressing neuronal markers have been shown in a number of studies [
128,
129], and the potential of these cells to generate voltage-sensitive ionic current was confirmed by electrophysiological recording [
130].
In vitro differentiation into glial cells was also reported [
131].
In vivo differentiation into neurons [
132,
133] and glial cells [
134‐
136] has been reported in a number of studies. But a few studies have failed to demonstrate this transdifferentiation [
125,
137,
138]. Fusion is another observation that needs to be considered. The question that bone marrow cells may adopt the phenotype of other cells by cell fusion was raised by
in vitro observations [
139,
140] and tested in an
in vivo model in which fusion of marrow stromal cells with Purkinje neurons was detected [
141]. It has been shown that transplanted cells are capable not only to migrate in the injured tissue [
135,
142] but also to attract host cells to the site of transplantation [
137]. Also, they form cell bridges within the traumatic cavity [
134,
137]. To address the best rout of delivery of these cells, chronic paraplegic rats received MSCs either locally or intravenously and it was concluded that transplantation of the cells to the spinal cord leads to superior functional recovery [
143]. Locomotor improvements have been reported in most of the above studies even in those that did not detect transdifferentiation. This observation was attributed to secretion of cytokines and growth factors from MSCs [
138,
144], which might be subjected to batch-to-batch variation [
138]. The point to be considered is that in most studies locomotor function was assessed by the Basso-Beattie-Breshnahan (BBB) test, which is a subjective test. More objective tests such as electrophysiological studies should be considered for achieving to more conclusive results. To the author's knowledge, no peer-reviewed clinical trial using MSCs for SCI patients has been published yet. But, a clinical trial involving transplantation of
in vitro expanded MSCs to the spinal cord of the patients with amyotrophic lateral sclerosis revealed that the procedure is safe and feasible [
145].
Embryonic stem (ES) cells
Embryonic stem (ES) cells are pluripotent cells derived from inner cell mass of the blastocyst, an early embryonic stage. It has been known for many years that pluripotent embryonic stem cells can proliferate indefinitely
in vitro and are able to differentiate into derivatives of all three germ layers [
146].
Neural stem cells derived from ES cells can lead to behavioral improvement after transplantation to the site of injury in the spinal cord [
147]. It has been shown that after prolonged
in vitro expansion of ES cells-derived neural stem cells, they remain able to differentiate into neurons and astrocytes both
in vitro and upon transplantation into brain [
148]. Transplantation of motor neuron-committed ES cells to the injured spinal cord combined with pharmacological inhibition of myelin-mediated axon repulsion and provision of attractive cues within the peripheral nerves led to extension of transplanted axons out of the spinal cord. The axons reached the muscle, formed neuromuscular junctions and their functionality was confirmed by electrophysiological studies [
149]. Transfection of ES cells with MASH1 gene is another strategy that caused ES cells to differentiated into motor neurons lacking Nogo receptor after transplantation into the transected spinal cord of mice and led to functional improvements confirmed by electrophysiological assessment [
150]. Myelination was also addressed in a number of studies; for example, it was shown that neural cells derived
in vitro from ES cells can myelinate the demyelinated rat spinal cord upon transplantation [
151]. Oligodendrocyte-restricted progenitor cells were also derived from ES cells and were able to enhance remyelination and led to functional improvements after transplantation into a rat model of acute spinal cord injury [
152].