SUMMARY

INTRODUCTION

Research over the last several decades has revealed numerous molecular mechanisms in the environment of the adult central nervous system (CNS) that contribute to the failure of axonal regeneration after injury, including myelin-associated proteins that inhibit axonal growth (He and Koprivica, 2004; Buchli and Schwab, 2005), the deposition of inhibitory extracellular matrix molecules around injury sites (Fawcett, 2006; Fitch and Silver, 2008), and the lack of positive environmental stimuli such as growth factors (Tuszynski and Lu, 2008). The observation that at least some classes of adult CNS axons can grow over long distances in peripheral nerve bridges supports the view that the adult CNS environment is inhibitory (David and Aguayo, 1981; Houle et al., 2006). However, some studies indicate that neuron-intrinsic mechanisms also contribute to axonal growth failure in the adult CNS (Filbin, 2006; Kadoya et al., 2009). Indeed, the extent to which neuron-intrinsic mechanisms alone can overcome the inhibitory growth environment of the adult CNS is actively debated. To address this question, we grafted either freshly dissociated neural stem cells/progenitors from green fluorescent protein (GFP) - expressing rat embryos (Bryda et al., 2006; Baska et al., 2008; Mayer-Proschel et al., 1997), or cultured human stem cells derived from two different sources (HUES7 cells and 566RSC cells from NeuralStem Inc.) to the adult lesioned spinal cord. Expression of the GFP reporter gene in all cells provides an unprecedented opportunity to track the fate, integration, process extension and differentiation of grafted cell types within the inhibitory milieu of the adult injured spinal cord.

We now report a remarkable capability of early stage neurons from different sources and species to survive, integrate, extend axons over very long distances and form functional relays in the lesioned adult CNS. These findings indicate that, despite the inhibitory milieu of the adult CNS, neuron-intrinsic mechanisms are sufficient to support remarkably extensive axonal growth and synapse formation after spinal cord injury, resulting in formation of novel neuronal relays that restore electrophysiological activity and behavior. Moreover, stem cells across species exhibit these properties, supporting the intrinsic capabilities of these cells and suggesting translational relevance.

RESULTS

Fischer 344 adult rats underwent T3 complete spinal cord transection. Two weeks later, a clinically relevant time point, we dissected embryonic day 14 (E14) spinal cords from Fischer 344 rats ubiquitously expressing the GFP reporter gene. Grafts were trypsinized and implanted as suspensions (Giovanini et al., 1997), but survived only at the host/lesion margins and failed to fill the complete transection site. To enhance graft survival and filling of the lesion, in new experiments the E14 neural stem cells were embedded into fibrin/thrombin matrices containing a cocktail of growth factors (N=26 rats; see Methods). Animals then underwent functional and electrophysiogical studies and were perfused 9 weeks after the initial injury.

Anatomical analysis revealed that grafted cells completely and consistently filled the lesion cavity when assessed seven weeks post-grafting (Fig. 1A–1B). Grafted cells were not observed to migrate into the host beyond the immediate region of the graft/lesion site. Grafted cells primarily differentiated into neurons (27.5 ± 2.7% of all GFP-labeled cells), oligodendroglia (26.6 ± 3.9%) and astrocytes (15.9 ± 1.6%) (Fig. 1C–1H and Suppl. Fig. 1). Numerous large GFP-labeled cells co-localized with the mature neuronal markers NeuN (Fig. 1C–1H), βIII tubulin (Tuj1) and MAP2 (Suppl. Fig. 2A–B). In addition to expressing mature neuronal markers, many grafted cells also expressed choline acetyltransferase (ChAT), characteristic of spinal motor neurons, and the inhibitory neuronal marker glutamic acid decarboxylase 67 (GAD67) (Suppl. Fig. 2C–D).

Open in a separate window

Figure 1

Survival, Filling and Differentiation of Neural Stem Cell Grafts in T3 Complete Transection Site

(A–B) Overview of GFP and GFAP fluorescent immunolabeling in a horizontal section demonstrates excellent graft survival, integration and filling of T3 complete transection site, seven weeks post-grafting. (C–D) GFP and NeuN labeling confirm extensive neuronal differentiation/maturation of grafted rat neural stem cells. (E–F) Higher magnification from c showing excellent integration and transition from host (h) neurons to grafted (g) neurons (dashed lines) (E: GFP, NeuN; F, NeuN alone). (G–H) Higher magnification from center of graft showing high density of NeuN-labeled neurons (inset) (G: GFP, NeuN; H, NeuN alone). Scale bar: A–D, 320 μm; E–H, 48 μm. Also see Figure S1 and S2.

Graft-derived spinal cord neurons extended large numbers of axons into the host spinal cord in both rostral and caudal directions over remarkably long distances (Fig. 2). Emerging GFP-labeled processes from grafts were axons, as demonstrated by co-labeling for the axonal marker neurofilament (Fig. 2A). In the cephalad (rostral) direction, graft-derived axons extended to C6 in all subjects examined (N = 6), and a distance of 25 mm from the T3 transection cavity into the C4 cervical segment in 5 of 6 subjects (Fig. 2E–F). Graft-derived axons also extended caudally to the lower thoracic spinal cord (T10) in all subjects (N = 6), continuing to the upper lumbar spinal cord, a distance of more than nine spinal segments (25 mm), in 4 of 6 subjects (2A–D, G–H). All animals with caudal extension of axons to L1 also exhibited rostral extension of axons to C4, indicating that long-distance, bi-directional axonal growth occurred commonly. To estimate the magnitude of axon outgrowth that can be achieved by implantation of neural primordia into sites of spinal cord injury, we stereologically (Rosenzweig et al., 2010) quantified 29,000 GFP-labeled axons emerging from the graft in the caudal direction (0.5 mm caudal to the T3 transection site). This number likely represents an underestimation, as dense bundles of axons that could not be individually resolved were counted as single axons. The density of axons emerging from grafts was highly consistent among all six subjects examined (Suppl. Fig. 3). In 12 additional subjects, progressive axon emergence from the lesion/graft site was serially examined at time points of 1, 2, 3 and 7 days post-grafting (N=3 rats per time point). Axons emerged from the graft/lesion site by day 2 after grafting, and extended at a rapid rate of 1–2mm/day (Suppl. Fig. 4). The visualization of axon growth over time documents the extension of new axons and excludes the possible artifact that grafted cells fused with host neurons. Thus, axons extended in remarkably large numbers and over long distances, including extension through adult white matter. Indeed, time course studies indicated that initial axon outgrowth preferentially occurred through white matter (Suppl. Fig. 4).

Open in a separate window

Figure 2

Extensive Long-Distance Axonal Outgrowth from Neural Stem Cell Grafts

(A) GFP and NeuN immunolabeling reveals that GFP-expressing neural stem cell grafts robustly extend axons into the host spinal cord rostral and caudal to the T3 complete transection site (caudal shown) over the 12mm length of the horizontal section. Extensive regions of the host spinal cord contain graft-derived projections in white matter and gray matter. Inset shows that GFP-labeled projections arising from grafts express neurofilament (NF), confirming their identity as axons. (B) Light-level GFP immunolabeling also clearly reveals the density and distribution of axons extending from the lesion site. Boxes are shown at higher magnification in C, D. (C) Dense numbers of GFP-labeled axons are present in host white matter (WM) and gray matter (GM) 2mm caudal to the lesion, and (D) large numbers of axons remain at the end of the block of tissue, 6mm caudal to the lesion. (E) A high density of GFP-labeled axons is present in the lateral portion of C8, three spinal segments above the lesion, and (F) remain detectable at C4, 7 spinal segments above the lesion. Asterisk indicates locations of higher magnification view. (G) GFP-labeled axons also extend in dense numbers caudally to T6 white matter and gray matter (shown) and (H) caudally to L1. Overall, axons extend at least 25 mm in each direction. Scale bar: A–B, 600 μm; B–C, 40 μm; D–G, 10 μm. Also see Figure S3 and S4.

Graft-derived axons in host white matter were frequently myelinated by host oligodendrocytes (i.e., cells myelinating graft-derived axons were not labeled for GFP; Fig. 3A–B). In addition, grafted cells that differentiated into neurons extended axons from the spinal cord lesion site into adjacent host ventral roots in every subject examined, with 6.8 ± 1.2 graft-derived axons quantified per 40μm-thick section of ventral root; Suppl. Fig. 2E–F). 56.7 ± 5.3% of these ventral root axons expressed the appropriate phenotypic motor neuronal marker choline acetyltransferase (ChAT; Suppl. Fig. 3F), likely derived from ChAT-expressing motor neurons inside grafts (Suppl. Fig. 2C). GFP-labeled axons were rarely detected in dorsal roots.

Open in a separate window

Figure 3

Myelination, Synapse Formation and Expression of Neurotransmitters

(A–B) Graft-derived, GFP-labeled axons are myelinated in many cases, (red, myelin-associated glycoprotein, MAG, confirmed by electron microscopy (T, transplanted, GFP-labeled axon). (C) GFP-expressing axon terminals are closely associated with host MAP-2-expressing neurons and dendrites, and (D) host Tuj1-expressing neuronal somata. (E) A z-stack image triple labeled for GFP, synaptophysin (Syn, inset), and ChAT, indicating co-association of graft-derived axons with a synaptic marker in direct association with host motor neurons (arrowhead indicates one of several examples). (F) Electron microscopy confirms that DAB-labeled GFP-expressing axon terminals form synapses (arrows) with host dendrites. Arrowhead indicates a separate, host-host synapse. (G–H) Expression of vGlut1/2 or GAD65 by GFP-labeled axons (arrows and insets) in close association with host motor (ChAT-labeled) neurons. Scale bar: A, 3μm; B, 200nm; C–E, 8μm; F, 200nm; G, 7μm; H, 6μm.

GFP-labeled, graft-derived axons in the host spinal cord formed dense bouton-like terminals around host dendrites and cell bodies, identified by labeling for the neuronal somal and dendritic markers MAP-2, the neuronal somal marker Tuj1 (βIII tubulin), and the motor neuron marker ChAT (Fig. 3C–E). Double immunofluorescent labeling for GFP and synaptophysin suggested synapse formation between grafted and host neurons in both rostral and caudal directions (Fig. 3E) at all distances from the lesion site, including sites sampled 25mm from the injury. Immunoelectron microscopy confirmed synapse formation between GFP-labeled graft-derived axons and host neurons and dendrites (Fig. 3F). Further, GFP-labeled terminal boutons expressed either the excitatory transmitter marker vesicular glutamate transporter 1/2 (vGlut1/2) or the inhibitory marker glutamic acid decarboxylase 65 (GAD65) (Fig. 3G–H). Many instances of co-localization of these terminal region markers with ChAT were also found (Fig. 3G–H). Collectively, these findings demonstrate exuberant projections of neural stem cell-derived axons from sites of spinal cord injury into the adult spinal cord together with synapse formation.

Neurons of the developing spinal cord express receptors to myelin-associated inhibitors such as Nogo (Josephson et al., 2002), a finding we confirmed in the present study in grafted tissue (Suppl. Fig. 5A). Thus, insensitivity to myelin-based inhibitors is an unlikely mechanism to account for the robust growth of axons observed in this study. Recently, the tumor suppressor gene “phosphatase and tensin homolog” (PTEN) has been identified as a major neuron-intrinsic regulator of axonal regeneration in the adult CNS (Park et al., 2008). PTEN inhibits the activity of “mammalian target of rapamycin” (mTOR), and neuronal knockout of PTEN has been reported to enhance central axonal regeneration (Park et al., 2008; Liu et al., 2010; Sun et al., 2011). To determine whether mTOR signaling is a mechanism contributing to the extensive growth of neural stem cell-derived axons through the lesioned adult CNS, we injected rapamycin, an mTOR inhibitor, into rats beginning two days after placement of grafts into the lesion site (6 mg/kg intraperitoneally, n=6 rats; 7 controls underwent vehicle injections). When examined two weeks later, rats subjected to mTOR blockade exhibited a significant, 50% reduction in graft-derived axonal outgrowth from the lesion compared to vehicle-injected controls (P<0.05; Fig. 4A; Suppl. Fig. 5B–E). Labeling for phospho-S6 (p-S6), an indicator of mTOR activity (Park et al., 2008), demonstrated abundant expression in graft-derived neurons and substantial reduction following rapamycin treatment (Suppl. Fig. 5H–J). Rapamycin injection did not alter neural stem cell graft survival or differentiation (Suppl. Fig. 5F–G, K). These findings demonstrate mTOR signaling as a mechanism contributing to the ability of early stage neurons to extend axons in the adult lesioned CNS.

Open in a separate window

Figure 4

Rapamycin Reduces Axon Growth; Host Axons Innervate Grafts

(A) Treatment with the mTOR inhibitor rapamycin significantly reduced outgrowth of graft-derived axons (*p<0.05, **p<0.01; see also Suppl. Fig 5). Veh, vehicle; Rap, rapamycin treated. Data are represented as mean ± SEM. (B–C) Host reticulospinal axons labeled with BDA regenerate into GFP-expressing neural stem cell grafts in site of T3 complete transection (C, from boxed area of panel B; arrows indicate GFP-labeled grafted cells with neuronal morphology). (D–E) Host serotonergic axons immunolabeled for 5-HT regenerate into GFP expressing E14 neural stem cell grafts in lesion site. (F) Bouton-like structures on host reticulospinal axons regenerating into graft co-localize with synaptophysin in a Z-stack image (arrows). (G) Host synaptophysin terminals (red) are closely apposed to GFP-labeled grafted cells exhibiting neuronal morphology. Scale bar: B, 64μm; C, 12μm; D, 70 μm; E, 10 μm; F–G, 2μm. .

Importantly, host axons also penetrated neural stem cell grafts in spinal cord lesion sites, including BDA-labeled reticulospinal axons and 5-HT labeled serotonergic (raphespinal) axons (Fig. 4B–G). Host inputs into grafts were both locally and supraspinally derived. While host supraspinal axons regenerated into grafts, in no case did they regenerate beyond the graft into host spinal cord parenchyma caudal to the lesion site. Like graft-derived axons, host axons penetrating grafts co-localized with the synaptic marker synaptophysin (Fig. 4F), establishing a mechanism for host-to-graft connectivity.

Functional and behavioral outcomes were measured in subjects that underwent T3 complete spinal cord transections and, two weeks later, neural stem cell implants in the lesion site (Fig. 5A). Hindlimb locomotion was severely impaired in both lesion control and grafted subjects for the first four weeks post-injury; in the 5^th week (three weeks post-grafting), recipients of neural stem cell grafts exhibited significant improvement on the BBB scale (Basso et al., 1996), reaching a functional plateau by the 8^th week post-lesion (p<0.01). The mean BBB score reached a level of 7, indicating movement about each joint of the hindlimb, in contrast to minimal movement if any in lesioned controls. To determine whether functional improvement depended on regeneration of host axons into the graft/lesion site, half the subjects underwent re-transection of the spinal cord just rostral to the T3 graft, resulting in a complete abolition of behavioral recovery when measured one week later (Fig. 5A and Suppl. Fig. 6).

Open in a separate window

Figure 5

Functional and Electrophysiological Improvement after T3 Complete Transection

(A) Hindlimb locomotion: BBB scores after T3 complete transection show significant improvement in subjects that received neural stem cell grafts (NSC, n=6) compared to lesioned controls (n=6). Re-transection (arrow, Re-T) at rostral interface of graft with host abolishes functional improvements when assessed one week later (**p<0.01, ***p<0.001). Data are represented as mean ± SEM. (B) Electrophysiological transmission across the T3 complete lesion site: (i) In intact animals, stimulation at C7 evoked a short latency (~3.0 msec), large amplitude response at T6. (ii) Transection of the cord at T3 completely abolished this response. (iii) In 4 of 6 lesion/grafted animals, recovery of an evoked response of prolonged latency (~5.5 msec) was observed. (iv) Re-transection of the spinal cord at T3, just rostral to the graft (green arrow), abolished the recovered evoked response. (C) Kynurenic acid (50 mM), a blocker of excitatory synapses, was applied to the graft to determine whether excitatory transmission across synapses in the graft was required to detect responses at T6. Indeed, kynurenic acid application substantially reduced the amplitude but not the latency of the evoked response. Also see Figure S6.

To confirm the dependence of functional recovery on host axonal regeneration into the lesion site, electrophysiological measurements were made in 6 subjects. A stimulating electrode was placed in the dorsal C7 spinal cord (3 spinal segments above the lesion) and recordings were made at T6 (3 spinal segments below the lesion). In intact animals, stimulation at C7 evoked a short latency (3.42± 0.32 msec) response (Fig. 5B-i) that was entirely abolished by T3 complete transection (Fig. 5B-ii). In two-thirds of animals with neural stem cell grafts in the lesion site, evoked responses were restored (Fig. 5B-iii). Response latencies in the grafted group were prolonged (5.47 ± 0.24 msec compared to 3.42± 0.32 msec in intact subjects, P<0.05), suggesting formation of poly-synaptic relays through the graft. Response amplitudes in the grafted group (0.09 ± 0.03 mV) were smaller than intact subjects (0.27 ± 0.12 mV) (P<0.05; Fig. 5B-iii). Notably, re-transection of the spinal cord just rostral to the T3 graft completely abolished detectable activity upon C7 stimulation even when using high intensity stimuli (Fig. 5B-iv). To determine whether responses detected at T6 depended on the presence of synapses in the graft, the NMDA receptor blocker kynurenic acid (50 mM) was applied to the graft at T3, 7 weeks post-lesion. Kynurenic acid resulted in a significant, 50% reduction in the amplitude of the response recorded at T6 but did not affect latency of the response, indicating that excitatory synaptic transmission across the graft was required for detection of caudal transmission (Fig. 5C).

Thus, early stage neurons grafted in a fibrin matrix containing a growth factor cocktail extend large numbers of axons over long distances in the lesioned spinal cord, and form neuronal relays that significantly improve electrophysiological and functional outcomes. The magnitude of functional effect substantially exceeds those previously reported in studies of fetal or stem cell grafts to the injured spinal cord, possibly due to enhanced graft survival generated by co-implantation with fibrin matrices containing growth factors. These results suggest strong translational possibilities. Accordingly, we repeated these studies using cultured human neural stem cells.

6 adult female athymic nude rats (T-cell deficient) underwent T3 complete transections. One week later, a suspension of neural stem cells derived from human fetal spinal cord (566RSCUBQT-GFP, see Methods; gift of NeuralStem, Inc.) was embedded into fibrin matrices containing a growth factor cocktail, as above. These cells are currently being used in a human clinical trial in amytrophic lateral sclerosis (Boulis et al., 2011). Function was assessed over the next 7 weeks in grafted animals and compared to six lesioned controls, then rats were sacrificed for anatomical studies. Grafts completely and consistently filled lesion cavities (Fig. 6A), and a majority of grafted cells (57±1.7%) expressed the mature neuronal marker NeuN (Fig. 6B). As in rat donor studies, remarkable numbers of axons emerged from the grafts and extended for very long distances, over a total of 17 spinal segments (Fig. 6C–E and Fig. 7). In the rostral direction, axons extended 25 mm above the lesion site to the C4 level (Fig. 7A–C), and in the caudal direction axons reached the upper lumbar spinal cord (Fig. 7D–F). GFP-labeled, graft-derived axons formed bouton-like terminals in the host gray matter and expressed human-specific synaptophysin (Fig. 6F), suggesting synaptic connectivity between grafted human and host rat neurons. Tumor formation by implanted human neurons was not observed within the lesion site or remotely in the host spinal cord. As observed in rat donor studies, grafts of human neural stem cells induced significant locomotor recovery after T3 complete transection (Fig. 6G), with movement about all joints of the lower extremities. Once again, re-transecting the spinal cord immediately rostral to the lesion site abolished hindlimb movement, indicating that functional recovery was dependent on inputs into the graft from the host.

Open in a separate window

Figure 6

Human Neural Stem Cell Grafts in Fibrin Matrices after T3 Transection: Long-Distance Axonal Growth, Connectivity and Functional Improvement

(A) GFP-expressing 566RSC human neural stem cells (NSCs) survive and completely fill T3 complete transection site, 7 weeks post-grafting (GFP, green; GFAP, red). (B) GFP and NeuN labeling reveals that many grafted human neural stem cells express mature neuronal markers. Field is from graft center. (C–E)GFP-expressing human neurons robustly extend axons into the host spinal cord caudal and rostral to the complete T3 transection site (caudal shown). Axons extend through both white matter (WM) and gray matter (GM). (F) GFP-labeled human axon terminals express human-specific synaptophysin (hSyn) in host gray matter. Site shown is 7 mm caudal to lesion. Scale bar: A, 500 μm; B, 10 μm; C, 600 μm; D–E, 60 μm; F 3 μm. (G) There is significant functional improvement after human NSC grafts to sites of T3 complete transection (**p<0.01). Re-transection (arrow, Re-T) at rostral interface of graft completely abolishes functional recovery assessed one week later, indicating that host inputs are required to improve function. Data are represented as mean ± SEM.

Open in a separate window

Figure 7

Dense Axonal Outgrowth from Implants of Human 566RSC Neural Stem Cells

(A) Very large numbers of GFP-labeled axons extend into host spinal cord rostral to T3 transection site, 7 weeks post-grafting. C8 segment shown. (B–C) GFP-labeled axons are present in C6 and C4 white matter, five and seven spinal segments above the lesion, respectively. Asterisks indicate location of higher magnification views. (D–F) GFP-labeled axons also extend in high density caudal to T6 and T9, and are present in both white matter (WM) and gray matter (GM, dashed lines indicate interface). (F) Axons continue to extend caudally to L1. Overall, axons extend at least 25 mm in each direction in all subjects. Scale bar: 15 μm.

To determine whether the ability to extend axons in vivo is a property expressed by stem cells derived from different human sources, we also tested the Harvard University Embryonic Stem cell 7 (HUES7) line (Cowan et al., 2004). HUES7 cells were neurally induced (Yuan et al., 2011) and embedded in the same growth factor cocktail-containing matrix, then grafted to sites of spinal cord injury (Suppl. Fig. 7); because this component of the study was purely anatomical, we used a C5 lateral spinal cord lesion (hemisection) in which animal care is less difficult. Once again, remarkable numbers of axons extended from the lesion over very long distances, reaching the brainstem in the rostral direction and the lumbar spinal cord caudally (Suppl. Fig. 7A–G). In gray matter, these human terminals abundantly expressed the pre-synaptic marker synaptophysin (Suppl. Fig. 7H).

DISCUSSION

Findings of the present study indicate that neural stem cells can specify sufficient information to permit extensive, long-distance central nervous system axonal regeneration; additional experimental manipulation of the inhibitory environment of the adult CNS is not required to achieve this long-distance regeneration. Moreover, the axon growth arising from neural stem cells can support functional improvement in the most severe model of spinal cord injury, complete spinal transection, when co-grafted in a supportive fibrin matrix containing a growth factor cocktail. Axons extend in large numbers and over remarkably long distances to form connections with host axons. Indeed, axons grew through white matter, gray matter, and into ventral roots, grew in high density, and elongated at a rapid rate of 1–2mm per day even through inhibitory white matter. Moreover, host supraspinal axons regenerated into the neural stem cell grafts. This reciprocal growth resulted in a five-point improvement on the BBB scale that reflects recovery of the ability to move all joints of the hindlimbs. The mechanism of recovery, in contrast to previous studies with neural stem cells, is functional relay formation, indicated by the abolition of recovery by spinal re-transection immediately rostral to the lesion. These findings may accordingly represent the most comprehensive demonstration to date of the ability of novel neural relays to support functional recovery even after the most severe spinal cord injury (Steward et al., 2003).

Wictorin and colleagues reported in 1990 that early stage human fetal tissue grafted into partial striatal lesion sites could extend axons for long distances in the mature CNS (Wictorin et al., 1990). In 1997, Silver and colleagues also supported the concept that neurons could extend axons through adult CNS tissue over long distances, by grafting adult dorsal root ganglion neurons to white matter (Davies et al., 1997). However, these neurons did not extend axons across lesion sites and functional effects were not studied (Davies et al., 1999). More recently, Gaillard and colleagues reported long-distance extension of GFP-labeled axons from embryonic grafts placed into cortical lesion sites; this study was anatomical and lacked functional or circuit analysis (Gaillard et al., 2007). Fischer's group grafted mixed neuronal and glial restricted precursors to sites of partial spinal cord injury but reported only modest axonal outgrowth (Lepore and Fischer, 2005), and subsequent co-administration of growth factors enabled further axonal growth over only short distances of a few millimeters (Bonner et al., 2011). Previous studies of embryonic stem cell implants to sites of spinal cord injury used moderate and incomplete lesions, and reported only modest functional improvements (Cummings et al., 2005) (1–2 points on the BBB scale), with relatively modest graft survival in the lesion site. The classic embryonic spinal cord grafting literature (Jakeman and Reier, 1991; Theele et al., 1996; Giovanni et al 1997) used cells of a similar developmental stage to the rodent donor cells in the present study, but did not identify the extent to which early stage neurons were capable of extending very large numbers of axons over extraordinarily long distances in the lesioned adult nervous system; the improved survival, growth and outcomes in the present study are likely a result of the use of fibrin matrix and a growth factor cocktail, together with clear identification of cell fate and projection using the GFP reporter. Indeed, distances traversed by graft-derived axons in this study exceeded the length of the entire embryonic rat nervous system (Miller et al., 1998). Findings were not attributable to graft cell fusion with the host, as axons progressively elongated outward from the lesion over several days, an observation inconsistent with cell fusion.

Grafts of human neural stem cells exhibited growth that was equally extensive. Two different human cell sources, including 566RSC cells derived from human embryonic spinal cord (Johe et al., 1996) and HUES7 cells derived from human ES cells (Cowan et al., 2004; Yuan et al., 2011), exhibited identical patterns of growth in the injured adult spinal cord after neural induction. Thus, this growth potential is not cell-specific or species-specific, suggesting translational potential. Indeed, functional testing of 566RSC cells also revealed an ability to support extensive functional improvement in a severe SCI lesion.

EXPERIMENTAL PROCEDURES

Supplementary Material

ACKNOWLEDGMENTS

Footnotes

REFERENCES