Combination of Multifaceted Strategies to Maximize the Therapeutic Benefits of Neural Stem Cell Transplantation for Spinal Cord Repair

Abstract

Neural stem cells (NSCs) possess therapeutic potentials to reverse complex pathological processes following spinal cord injury (SCI), but many obstacles remain that could not be fully overcome by NSC transplantation alone. Combining complementary strategies might be required to advance NSC-based treatments to the clinical stage. The present study was undertaken to examine whether combination of NSCs, polymer scaffolds, neurotrophin-3 (NT3), and chondroitinase, which cleaves chondroitin sulfate proteoglycans at the interface between spinal cord and implanted scaffold, could provide additive therapeutic benefits. In a rat hemisection model, poly(e-caprolactone) (PCL) was used as a bridging scaffold and as a vehicle for NSC delivery. The PCL scaffolds seeded with F3 NSCs or NT3 overexpressing F3 cells (F3.NT3) were implanted into hemisected cavities. F3.NT3 showed better survival and migration, and more frequently differentiated into neurons and oligodendrocytes than F3 cells. Animals with PCL scaffold containing F3.NT3 cells showed the best locomotor recovery, and motor evoked potentials (MEPs) following transcranial magnetic stimulation were recorded only in PCL-F3.NT3 group in contralateral, but not ipsilateral, hindlimbs. Implantation of PCL scaffold with F3.NT3 cells increased NT3 levels, promoted neuroplasticity, and enhanced remyelination of contralateral white matter. Combining chondroitinase treatment after PCL-F3.NT3 implantation further enhanced cell migration and promoted axonal remodeling, and this was accompanied by augmented locomotor recovery and restoration of MEPs in ipsilateral hindlimbs. We demonstrate that combining multifaceted strategies can maximize the therapeutic benefits of NSC transplantation for SCI. Our results may have important clinical implications for the design of future NSC-based strategies.

Keywords

Stem cells Spinal cord injury (SCI)Tissue scaffold (Pclf polymer)Neurotrophin 3 Motor evoked potential Chondroitinase ABC

Introduction

Spinal cord injury (SCI) interrupts neural connections between the brain and the rest of the body, leading to a paralysis and loss of sensation below the level of injury. The pathological consequences of SCI are multitude: death of local segmental neurons and glial cells, severance of axons in white matter, formation of fluid-filled cystic cavities, and demyelination of spared white matter (6,33,52). All these pathological processes, at least to some degree, contribute together to the devastating functional outcomes after SCI, warranting the development of a multifaceted combinatorial approach to achieve clinically relevant functional recovery (19,50).

Preclinical studies of various animal models of SCI have shown that neural stem or progenitor cells (NSCs) possess multiple therapeutic potentials that might be exploited to reverse diverse pathological processes. For example, NSCs can modify the injured microenvironment and provide neuroprotection (27,41). Grafted NSCs can replace lost neuronal populations at segmental levels (15,60), and NSCs or lineage restricted glial progenitor cells can differentiate into myelinating oligodendrocytes and augment the remyelination of spared tissues (10,15, 30,31,46). Furthermore, NSC transplantation has been reported to increase axonal regeneration and/or plasticity (40,44). For these reasons, it is anticipated that the NSC transplantation will become a viable and effective therapeutic option to improve functional outcome after SCI (16,38).

However, many considerable obstacles remain to be overcome before NSC transplantation becomes available in clinic. The formation of fluid-filled cystic cavities severely impedes axonal growth in lesion areas, and it has been suggested that transplanted NSCs are not capable of bridging these cavitary lesions (51,59). Furthermore, survival and migration of grafted NSCs are often limited in the injured spinal cord (26,37,48), and NSCs transplanted into injured spinal cord tissue frequently either remain undifferentiated or differentiate predominantly into astrocytes, which can contribute to glial scar formation (11,12). In addition, glial scars and inhibitory extracellular matrices such as chondroitin sulfate proteoglycans (CSPGs) could interfere with migration and axonal growth of grafted NSCs, reducing the possibility of the proper integration of NSCs with the host neural circuit (26,32,35).

Therefore, we reasoned that a combination of complementary strategies aimed at removing specific obstacles would potentiate the efficacy of NSC-based treatments. Previous studies have shown that NSC transplantation in combination with a complementary strategy provided improved therapeutic benefits as compared with NSC alone. Teng et al. found that implanting polymer scaffold seeded with NSCs led to an improved functional recovery (57). Brain-derived neurotrophic factor (BDNF) or neurotrophin-3 (NT3) coadministered with NSCs or directly transduced to NSCs improved the survival and migration of grafted NSCs (7,20,29,49,62). Moreover, combining chondroitinase with NSC grafts to degrade inhibitory extracellular matrices enhanced NSC migration (26). What remains to be determined is whether a combination of multifaceted complementary strategies could provide therapeutic benefits of greater magnitude and to what extent a multifaceted combinatorial approach could enhance the therapeutic effects of NSC transplantation.

In an effort to identify an effective NSC-based treatment strategy for SCI, we tested whether a combinatorial strategy based on the use of polymer scaffold, NSCs, and NT3 could improve recovery of locomotor function, and whether the inclusion of a chondroitinase treatment to degrade CSPGs deposited at the interface between the spinal cord and implanted scaffold could augment the therapeutic benefits conferred by this combinatorial strategy.

Materials and Methods

Establishment of Human NSCs Modified to Secrete NT3

Telencephalon tissues from gestational week 15 human fetal brains were utilized to generate an immortalized human NSC line (F3) using a retroviral vector encoding v-myc, as described previously (18,34). The permission to use fetal tissues was granted by the Clinical Research Screening Committee involving Human Subjects of the University of British Columbia. The PG13 mouse packaging cell line was transfected with plasmid pLHCX-NT3 vector containing full-length human NT3 cDNA (ATCC, Manassas, VA) using lipofectamine. Retroviral supernatants obtained from the packaging cells were applied to F3 NSCs. Hygromycinresistant clones were screened and isolated, and one of the clones (F3.NT3) was expanded and used for the transplantation.

Pretreatment of PCL Scaffolds and the Cell Seeding Procedure

A biocompatible poly(e-caprolactone) (PCL) scaffold was used in this study. The scaffold was produced by reacting PCL diols (MW = 2,000) with acryloyl chloride and fabricated using a gas-foaming/salt-leaching method (36). The dimensions of PCL scaffold were determined to fit hemisected lesion cavities. Before cell seeding, the PCL scaffolds were sterilized with 70% ethanol, and then washed five times in distilled water, after which they were subjected to vacuum desiccation for 10 min to remove trapped air bubbles. Before seeding, F3 or F3.NT3 cells were grown attached to culture dishes, dissociated into single cells by a brief trypsin treatment, and then suspended at a density of 10⁵ cells/μl. A 2.5-μl aliquot of medium containing cells was then slowly injected using a pipette tip into both sides of each scaffold (a total of 5 × 10⁵ cells/scaffold). The PCL scaffolds seeded with cells were transferred into a 12-well plate and incubated on shaker at 37°C for at least 3 days before implantation.

Surgical Procedures

All animal procedures were approved by the Animal Experiment Review Committee at Ajou University School of Medicine. Adult female Sprague-Dawley rats (250–300 g) were used. After being anesthetized with 4% chloral hydrate (400 mg/kg), rats were subjected to dorsal laminectomy at the seventh thoracic vertebral level (T7–8). After dura was opened, a chunk of right side spinal cord was excised using iridectomy scissors. Remaining tissue was removed by vacuum suction until the vertebral body at the ventral side was visualized. Special care was taken to confirm that the right side hemicord was entirely removed and that the lesion did not extend into the left side spinal cord. A PCL scaffold seeded with human NSCs was implanted into the lesion site immediately after hemisection. Animals were randomly allocated to four experimental groups: 1) animals that received hemisection alone [control (CTL) group, N = 8), 2) animals implanted with PCL scaffold without cells (PCL group, N = 7), 3) animals implanted with PCL scaffold seeded with F3 NSCs (PCL-F3 group, N = 9), and 4) animals implanted with PCL scaffold with F3.NT3 cells (PCL-F3.NT3 group, N=9). The above animals were sacrificed at 9 weeks after all behavioral and electrophysiological tests had been completed. A different set of animals was sacrificed at 2 weeks post-SCI for ELISA and Western blot. All animals received daily intraperitoneal cyclosporine (Sandimmun; Novartis, Bern, Switzerland) at a dosage of 10 mg/kg, beginning from 1 day before surgery until sacrifice. Prophylactic antibiotics were injected intraperitoneally on the next day following each surgery. Bladder care was provided until spontaneous voiding resumed.

Infusion of Chondroitinase ABC

Protease-free chondroitinase ABC (C-ase) was purchased from the Seikagaku Corporation (Tokyo, Japan). One day prior to infusion, an enzyme aliquot (10 U/ml, 0.2 ml) was loaded into the reservoir of the Alzet 2004 osmotic pump (Durect, Cupertino, CA) and incubated overnight at 37°C. Immediately after implantation of PCL scaffold seeded with F3.NT3 cells, a PE-10 tubing (Intramedic, NJ) attached to the osmotic pump was inserted into the subarachnoid space from the intervertebral space one level rostral to the lesion and slowly advanced under the lamina until the tube opening was located on top of the implantation site. The osmotic minipumps delivered the enzyme solution at a rate of 0.25 μl/h continuously for 28 days. The weight equivalent of penicillinase (P-ase; Sigma, St. Louis, MO) was delivered as a control injection. Eighteen animals that received PCL implants seeded with F3.NT3 cells were randomly divided into C-ase (N = 9) and P-ase (N = 9) groups. All these animals were sacrificed at 9 weeks post-SCI after behavioral and electrophysiological assessments had been done. A different set of animals was sacrificed at 2 weeks post-SCI (N=3 per group), and the spinal cord sections of these animals were subjected to immunohistochemistry with CS56 and C4S antibody to detect intact and degraded CSPGs, respectively.

ELISA and Western Blot Analysis

The levels of NT3 in culture medium were measured using a human NT3 ELISA kit (R&D Systems, Minneapolis, MN). The detailed procedure was described in an instruction manual provided by the manufacturer. The experiment was replicated three times. NT3 levels in the spinal cord tissue implanted with PCL scaffolds seeded with F3.NT3 cells were also measured by ELISA. Briefly, 16 animals were sacrificed with an overdose anesthesia at 2 weeks post-SCI (N = 4 per group). Spinal cord tissue spanning ±5 mm from the epicenter was dissected and homogenized in lysis buffer (4 μl/mg tissue) containing protease inhibitor cocktail (Sigma). To measure NT3 levels only in the host spinal cord tissue, PCL scaffolds were manually detached from the spinal cord tissues and not included in the protein extraction. For Western blot analysis, spinal cord tissue blocks (spanning 2–3 mm) immediately caudal to the tissues used for ELISA were homogenized in the same lysis buffer. The blots were probed with growth associated protein 43 (GAP-43; 1:3000; rabbit; Chemicon; Temecula, CA) and synaptotagmin (1:3000; mouse; Chemicon) antibodies. All blots were also probed with mouse anti-β-actin to control for differences in loading amount. After incubation with peroxidase-conjugated appropriate secondary antibodies, antigen-antibody reactions were visualized using a chemiluminescence kit.

Assessment of Locomotor Recovery

Locomotor recovery was assessed using the Basso, Beattie, and Bresnahan (BBB) locomotor rating scale and the grid walk test. Animals were assigned with new identification codes after surgery to ensure that behavioral performances were rated in a blind manner. The BBB locomotor rating scale was used to measure the quality of hindlimb function during open field locomotion. Because animals were subjected to hemisection injury, right (ipsilateral) and left (contralateral) hindlimbs were assessed separately. For the grid walk test, animals were pretrained for 7 days before surgery to walk on grid runway (30 × 140 cm with 50 × 50-mm holes) for a water reward. The test was conducted at 4 and 7 weeks post-SCI, and animals were retrained for 5 days before testing. On the day of tests, four runs were recorded using a three-charge-coupled device (CCD) digital video camera, and later analyzed frame by frame in slow motion. The average number of limb placement errors per run was determined for each animal.

Electrophysiological Assessment

The integrity of the motor pathway through the spinal cord was assessed by the transcranial magnetic motor evoked potentials (39). After all the behavioral tests had been completed (at 8 weeks post-SCI), transcranial magnetic stimulation (Magstim model 2002, Magstim, Wales, UK) was delivered around the vertex of animals anesthetized with chloral hydrate (400 mg/kg). A custom-generated double coil (diameter 25 mm) was used to elicit focal stimulation. Recording needle electrodes were inserted into the gastrocnemius muscle. Stimulation was started from subthreshold intensities and the stimulus intensity was progressively increased until maximal evoked potentials were obtained. Final stimulation intensities ranged from 80 to 100 A/μs. Motor evoked potentials (MEPs) were recorded at both contralateral and ipsilateral gastrocnemius muscles. MEP amplitude in each animal was determined by averaging the results of 7–10 trials.

Tissue Processing and Immunohistochemistry

For histological analysis, animals were anesthetized with an overdose of chloral hydrate and perfused with heparinized saline (0.9%), followed by 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. Spinal cords were then dissected, postfixed in 4% paraformaldehyde at room temperature for 2 h, and then cryoprotected in a graded series of sucrose solutions. Longitudinal or transverse 20-μm sections were cut using a cryostat (Leica) in a 1:10 series and thaw-mounted onto Super Frost plus slides. For immunohistochemistry, spinal cord sections were incubated overnight at 4°C with anti-human mitochondria (1:300; mouse and rabbit; Chemicon), anti-APC-CC1 (adenomatous polyposis coli; 1:200; mouse; Calbiochem, La Jolla, CA), anti-GFAP (glial fibrillary acidic protein; 1:500; rabbit; Chemicon), anti-MAP2 (microtubule associated protein; 1:200; rabbit; Chemicon), anti-neurofilament-M (150 kDa) (1:200; rabbit; Chemicon), anti-CS56 (chondroitin sulfate; 1:200; mouse; Sigma), anti-C4S (degraded CSPGs; 1:200; mouse; Chemicon), and rabbit 5-HT antibodies (5-hydroxytryptamine; 1:5000; Immunostar, Hudson, WI). Spinal cord sections were then washed and incubated with fluorophore-tagged appropriate secondary antibodies. For fluorescence staining, cover slips were mounted onto glass slides using Gelvatol and examined under an Olympus confocal laser scanning microscope.

Quantitative Measurements of Cell Numbers and Axonal Growth

Human mitochondria-positive cells in the entire longitudinal spinal cord sections were counted using unbiased stereology. Cell counting was performed on an Olympus BX51 Microscope that was coupled with the Stereo Investigator 8 software (MBF Bioscience, Williston, VT). Human mitochondria-positive cells within optical dissectors randomly placed in regions of interest were counted using stereological criteria. Stereological estimation was performed systematically using the formula embedded in the software. To determine the percentage of grafted human NSCs that were colocalized with differentiation markers GFAP, APC-CC1, and MAP2, two longitudinal sections showing the highest graft survival were chosen for each animal. The numbers of human NSCs colocalized with these markers were counted in the entire sections, and these were then divided by the total numbers of human mitochondria-positive cells. Averages of these two sections were taken to represent final percentage values. To determine the maximal migration distance of grafted human NSCs, a montage image of the entire longitudinal spinal cord section showing the greatest migration distance was created using Photoshop software. Distances between the scaffold/spinal cord interfaces and the cells that migrated most were measured in each montage image and averaged for each group.

To measure the extent of neurofilament (NF)-positive axon growth within PCL scaffolds, two consecutive longitudinal spinal cord sections that contain the largest lesion areas were chosen. Three regions of interest (ROIs) measuring 1422 × 1249 μm were placed inside PCL scaffolds along the rostral, caudal, and medial borders in each section. NF-positive axons within these ROIs were manually traced at 200× magnifications using the Metamorph software 7.01 (Universal Imaging Corporation, Downingtown, PA). Numbers and lengths of axons in three ROIs were calculated using Metamorph software and averaged for the two sections in each animal. To quantify 5-HT axons in the caudal spinal cord, three transverse sections spaced 200 μm apart from each other were chosen from spinal cord blocks caudal to lesion sites. Left and right ventral horn regions in each section were imaged at 200×. ROIs were then drawn using a freehand selection tool to cover the entire ventral horn. The numbers of pixels occupied by the serotonergic (5-HT) fibers at the ventral horns were quantified. Serotonergic fiber densities of right side were normalized versus those of left side and percentage serotonergic innervations of lesioned versus nonlesioned sides were calculated and averaged for the three sections per animal.

Extent of Myelination in White Matter

Tissue sections were immersed for 8 min in a staining solution consisting of 240 ml of 0.2% eriochrome cyanine RC (Sigma) and 10 ml of 10% FeCl₃^*6H₂O (Sigma) in 3% HCl. Sections were then washed with running tap water followed by differentiation in 1% aqueous NH₄OH. In each animal, the eriochrome-stained spinal cord section containing the largest lesion area was chosen, and the two more sections 200 μm dorsal and ventral each to the chosen section were included in the analysis (three consecutive sections per animal). Stained sections were viewed on an optical microscope (Olympus BX51) and images were captured using a CCD camera (JVC digital camera KY-F1030). A single ROI (120 × 90 μm) per section was chosen in contralateral white matter. The areas of myelin staining above a predetermined threshold were quantified using Image J software (National Institutes Health). Percentages of ROIs stained by myelin were averaged for three sections per animal.

Statistical Methods

Statistical analysis was performed using SPSS version 12.0 (Chicago, IL) or GraphPad Prism software version 4.0 (San Diego, CA). The unpaired Student's t-test or one-way ANOVA followed by Tukey's post hoc test was used to compare group means at single time points. Repeated measures two-way ANOVA with Tukey's post hoc test was used to compare mean values at multiple time points (BBB scores) or different locations from the epicenter (cell numbers). Numerical values are presented as mean ± SEM.

Results

Implantation of PCL Scaffolds Seeded with Human NSCs Into Hemisected Spinal Cord

A total of 5 × 10⁵ parental human NSCs (F3) or NT3 overexpressing human NSCs (F3.NT3) were seeded into PCL scaffolds and cultured for at least 3 days before implantation. Scanning EM showed that seeded NSCs were attached to the inner walls of the PCL scaffold and packed into its micropores (Fig. 1A). To determine whether seeded F3.NT3 cells survive and function properly inside PCL scaffolds, levels of NT3 in culture media were measured. As expected, freely grown F3.NT3 (dissociate culture) secreted almost 10-fold more NT3 into culture media than parental F3 cells. Although NT3 levels were reduced by almost a half, F3.NT3 cells seeded into PCL scaffolds still secreted substantial amounts of NT3 (Fig. 1B).

Figure 1.

Seeding human neural stem cells (NSCs) into a poly(∊-caprolactone) (PCL) scaffold. (A) Scanning electron microscope image of a PCL scaffold seeded with human NSCs overexpressing NT3 (F3.NT3). NSCs were found attached to the walls of the PCL scaffold and packed into its micropores. (B) Comparison of NT3 levels in culture media between NSCs in dissociated culture and seeded in PCL scaffolds. The diagrams below the abscissa illustrate different culture conditions. Although NT3 levels in culture media were lower than those from freely grown F3.NT3 cells, F3.NT3 cells seeded in PCL scaffolds produced significant amounts of NT3 in media as determined by ELISA. Gray bar: F3, black bar: F3.NT3. ^***p < 0.001 by Student's t-test.

Right-sided spinal cord hemisection was performed at the T7 level and PCL scaffolds with or without seeded human NSCs were implanted immediately into hemisected cavities to bridge lesion sites. At 9 weeks post-SCI, PCL scaffolds were found to adhere well to residual spinal cord tissues and to form a rostrocaudal bridge across lesion cavities (Fig. 2A, B). Surviving human NSCs were identified by their immunoreactivities against human-specific mitochondrial antigen. Both F3 and F3.NT3 cells migrated into residual spinal cord tissues across scaffold–spinal cord interfaces (Fig. 2C–F). Numbers of human NSCs were comparable for F3 and F3.NT3 cells within scaffolds, but F3.NT3 cell numbers in residual spinal cord tissues were markedly greater than those of F3 cells. Stereological counting of grafted cells revealed that F3.NT3 cells were more widely distributed in a rostrocaudal direction, and the differences between the two cells were maintained at up to ±6 mm from the epicenter (Fig. 2G). The total number of human NSCs in the entire rostrocaudal extent was also significantly higher for F3.NT3 than F3 cells (29,256 ± 9,291 and 14,507 ± 3,696, respectively; p < 0.001), indicating that the overall survival was better in F3.NT3 cells. We also found that the mean migration distance was significantly greater in animals with PCL-F3.NT3 grafts (p < 0.01) (Fig. 2H).

Figure 2.

Survival and migration of human neural stem cells (NSCs) grafted inside PCL scaffolds. (A, B) Representative longitudinal spinal cord sections from PCL-F3 (A) and PCL-F3.NT3 groups (B) at 9 weeks post-spinal cord injury (SCI). Spinal cord sections were stained with glial fibrillary acidic protein (GFAP; green) and human-specific mitochondria antibody (red) to detect grafted NSCs of human origin. PCL scaffolds are nicely adhered to residual spinal cord tissue and formed a rostrocaudal bridge across the lesion cavities. Rostral direction is leftwards. Dotted lines demarcate scaffold borders. Scale bar: 250 μm. (C–F) Higher magnification images of boxed regions in the longitudinal section of PCL-F3 (C, D) or PCL-F3.NT3 (E, F) groups. The number of F3.NT3 cells that migrated into the spinal cord was markedly higher than that of F3 cells. DAPI-stained nuclei are shown in blue. Scale bar: 50 μm. (G) Stereological counting of grafted cells. ^*p < 0.05, ^**p < 0.01, and ^***p < 0.001 by repeated measures two-way ANOVA followed by Tukey's post hoc analysis. Error bars represent mean ± SEM. (H) Quantification of maximal migration distance. Each triangle or circle represents maximal migration distance in each animal, and each line represents group mean value. ^**p < 0.01 by Student's t-test. N = 9 per group.

F3 or F3.NT3 cells that migrated into remaining spinal cord tissues appeared to acquire mature neural cell phenotypes (Fig. 3). F3.NT3 cells observed in remaining spinal cord tissue were frequently colocalized with neuronal marker MAP-2 (Fig. 3D, G). On the other hand, only a small proportion of F3 cells differentiated into MAP-2-positive neurons (Fig. 3A, G). F3.NT3 cells also more frequently differentiated into oligodendrocytes than F3 cells (Fig. 3B, E, G). Approximately 10% of F3 and F3.NT3 cells that migrated into remaining spinal cord were found to be positive for GFAP, and there was no difference between the two cells (Fig. 3C, F, G). However, in contrast to F3 and F3.NT3 cells observed in remaining spinal cord tissues, human NSCs that remained within the scaffolds showed no evidence of differentiation into mature neural cells.

Figure 3.

Phenotypic differentiation of NSCs after implantation into hemisected spinal cord. Grafted NSCs were detected by immunoreactivity against human mitochondrial antigen (red) at 9 weeks post-SCI. Arrows indicate the colocalizations of cell type-specific markers (A–F). Merged cells are shown in yellow. Compared to F3 cells (A, D), F3.NT3 cells were frequently colocalized with neuronal (microtubule-associated protein 2; MAP-2) and oligodendrocytic (CC1) markers (B, E). (C, F) Small percentage of F3 or F3.NT3 cells differentiated into GFAP-positive astrocytes. (G) Quantification of percent NSCs that differentiated into neurons (MAP-2), oligodendrocytes (CC1), or astrocytes (GFAP). Scale bar: 25 μm. ^**p < 0.01 and ^***p < 0.001 by Student's t-test. N = 9 animals for each group. Error bars represent mean ± SEM.

Assessment of Functional Recovery

BBB locomotor and grid walk tests were performed to assess locomotor recovery. All groups showed similar BBB scores immediately after injury (Fig. 4A, B). Locomotor deficits rapidly recovered during the first two weeks in all animals. Animals in PCL-F3.NT3 group exhibited clearly better locomotor recovery than those in the other groups at 3 weeks and continued to show better improvement until 7 weeks post-SCI. Locomotor scores also improved in PCL-F3 group (mean BBB score at 7 weeks post-SCI, 12.7 ± 1.7 and 13.9 ± 1.8 for ipsi- and contralateral, respectively), but the mean BBB score at this time point was significantly less than that in PCL-F3.NT3 group (15.1 ± 1.5 for ipsilateral and 16.6 ± 1.3 for contralateral). At 7 weeks post-SCI, animals in PCL-F3.NT3 group fully regained coordinated plantar stepping with improved ankle positioning. Implantation of PCL scaffold alone provided slight functional benefits compared to the injury alone (CTL) group. Repeated measures two-way ANOVA revealed a highly significant treatment effect over time in both hindlimbs (p < 0.001 for both hindlimbs). Tukey's post hoc analysis showed significant differences between PCL-F3.NT3 and the other three groups (Fig. 4A, B). The difference between PCL-F3 and CTL groups was also statistically significant. In grid walk, animals in CTL group made significantly more frequent errors in both hindlimbs than those in the other groups (Fig. 4C, D). Implantation of PCL scaffold alone markedly reduced the number of errors (p < 0.001 compared to CTL). These errors were further reduced in PCL-F3 group. Animals in PCL-F3.NT3 groups showed the highest accuracy in hindlimb placement. One-way ANOVA followed by Tukey's post hoc analysis revealed a statistically significant difference between PCL-F3.NT3 and PCL-F3 groups. Taken together, these results suggest that PCL-F3.NT3 implantation provided the greatest behavioral benefits.

Figure 4.

Behavioral assessment of locomotor recovery. (A, B) Recovery of Basso, Beattie, and Bresnahan (BBB) locomotor scores over the 7-week period after injury for ipsilateral (right) (A) and contralateral (left) (B) hindlimbs. Animals in PCL-F3.NT3 group showed best recovery. ^*p < 0.05, ^**p < 0.01, and ^***p < 0.001 by repeated measures two-way ANOVA followed by Tukey's post hoc analysis. (C, D) Grid walk. Average numbers of ipsilateral (C) and contralateral (D) hindlimbs placement errors on grid per run were counted at 4 and 7 weeks post-SCI. Animals in PCL-F3.NT3 group showed highest accuracy on grid walk. ^*p < 0.05, ^**p < 0.01, and ^***p < 0.001 by one-way ANOVA followed by Tukey's post hoc analysis. N = 8, 7, 9, and 9 for control (CTL), PCL, PCL-F3, and PCL-F3.NT3 groups, respectively, for all graphs. Error bars represent mean ± SEM.

We further examined whether the improvements in hindlimb locomotor function were accompanied by electrophysiological restoration of motor evoked potentials (MEPs). In sham-operated animals, transcranial magnetic stimulation of the vertex readily elicited MEP responses at both hindlimbs of mean amplitudes 7.2 ± 1.6 and 7.3 ± 1.2 mV (ipsilateral and contralateral, respectively) (Fig. 5A, B). MEP responses were almost absent in CTL group in both hindlimbs. None of PCL, PCL-F3, and PCL-F3.NT3 groups showed ipsilateral MEP response with the amplitude of more than 1 mV (Fig. 5A, C). In the contralateral side, none of PCL alone group and only two of PCL-F3 group (22%, 2/9) exhibited appreciable MEP response (>1 mV). In PCL-F3.NT3 group, 67% (6/9) animals showed MEP responses with the amplitude of more than 1 mV (Fig. 5B). Mean amplitude of PCL-F3.NT3 groups in the contralateral side was 2.6 ± 1.7 mV and significantly higher than those of the other groups (p < 0.001) (Fig. 5D).

Figure 5.

Recovery of motor evoked potentials in response to transcranial magnetic stimulation. Motor evoked potentials (MEPs) were recorded at 8 weeks post-SCI in both ipsilateral and contralateral gastrocnemius muscles in response to transcranial magnetic stimulation over the vertex at intensities of 80–100 A/μs. (A, B) Representative MEP recordings of ipsilateral (A) and contralateral (B) hindlimbs. The MEP responses disappeared in both hindlimbs after injury. Significant MEP recovery was observed only in PCL-F3.NT3 group in the contralateral side. (C, D) Quantitation graphs of MEP amplitudes in ipsilateral (C) and contralateral (D) hindlimbs. ^*p < 0.001 by one-way ANOVA followed by Tukey's post hoc analysis. N = 4, 8, 7, 9, and 9 for sham, CTL, PCL, PCL-F3, and PCL-F3.NT3 groups, respectively. Error bars represent mean ± SEM.

Mechanism of Functional Recovery Mediated by PCL-F3.NT3 Implantation

First, we determined whether implantation of PCL scaffolds with F3.NT3 increased the concentration of NT3 in host spinal cord tissue at 2 weeks after injury (Fig. 6A). ELISA showed that levels of NT3 in the spinal cord of PCL-F3.NT3 group were significantly higher than in the other three groups (p < 0.05 vs. PCL-F3, and p < 0.01 vs. PCL and CTL groups). We also examined whether PCL-F3.NT3 implantation promoted neuroplasticity in the spinal cord caudal to lesions where the spinal locomotor center is located at the 2 weeks after injury when rapid behavioral improvement was observed. GAP-43 and synaptotagmin were chosen as a marker for axonal plasticity and synaptogenesis, respectively. Western blot showed that PCL-F3.NT3 implantation markedly increased the expression of GAP-43 and synaptotagmin at 2 weeks post-SCI (Fig. 6B). In fact, GAP-43 levels in PCL-F3.NT3 group were almost fourfold higher and synaptotagmin levels were threefold higher than in CTL group (Fig. 6C, D). GAP-43 expression was also significantly increased in PCL-F3 group compared to CTL and PCL groups (Fig. 6B, D), but levels of GAP-43 expression in PCL-F3.NT3 group was significantly higher than in PCL-F3 group (p < 0.01). Consistent with the Western blot data, GAP-43 immunoreactivity in neuropil was found to increase in the caudal gray matter (lumbar level), especially in the intermediate regions and ventral horns, of animals in PCL-F3.NT3 group (Fig. 6G, H). In contrast, only subtle immunoreactivity was observed in CTL group (Fig. 6E, F).

Figure 6.

Mechanism of functional recovery mediated by implantation of PCL scaffold with F3.NT3 NSCs. (A) ELISA of NT3 levels in host spinal cord tissue at 2 weeks post-SCI. The levels of NT3 in the spinal cord of PCL-F3.NT3 group were significantly higher than those in the other groups. N= 4 per group. (B–D) Western blot analysis of caudal spinal cord tissue at 2 weeks post-SCI. (B) Representative blots of growth-associated protein-43 (GAP-43) and synaptotagmin. PCL-F3.NT3 implantation markedly increased the expression of GAP-43 and synaptotagmin. (C, D) Quantification of GAP-43 (C) and synaptotagmin (D) expressions. GAP-43 levels in PCL-F3.NT3 group were almost fourfold higher and synaptotagmin levels were threefold higher than in CTL group (D). ^*p < 0.05, ^**p < 0.01, and ^***p < 0.001 by one-way ANOVA followed by Tukey's post hoc analysis. N = 4 per group. (E–H) Representative images of GAP-43 immunofluorescent staining in the transverse lumbar sections (lesioned side). GAP-43 immunoreactivity (green) was regionally increased in PCL-F3.NT3 group (G). Apparent GAP-43 immunoreactivity was observed only in dorsal horn region in CTL group (E), and there was very subtle neuropilic immunoreactivity in the ventral motor region (F). In contrast, GAP-43-positive axon terminals were readily observed in the ventral motor region of PCL-F3.NT3 group (H). The boxed regions in (E) and (G) are magnified in (F) and (H), respectively. DAPI-stained nuclei are shown in blue. Scale bars: 500 μm (E, G), 50 μm (F, H). (I–L) Representative eriochrome cyanine (EC)-stained longitudinal spinal cord sections from CTL (I), PCL (J), PCL-F3 (K), and PCL-F3.NT3 (L) groups. The images were taken at the white matter contralateral to spinal lesions. Scale bars: 100 μm. (M) Quantification of percent EC-stained areas in the contralateral white matter. ^*p < 0.05 and ^***p < 0.001 by one-way ANOVA followed by Tukey's post hoc analysis. N = 8, 7, 9, and 9 for CTL, PCL, PCL-F3, and PCL-F3.NT3 groups, respectively. All error bars represent mean ± SEM.

Electrophysiological assessment showed that electrical transmission through contralateral (left) spinal cord was also compromised after right side spinal cord hemisection (Fig. 5). This was consistent with the finding of a recent study in which a decline in electrical transmission through contralateral uninjured fibers was suggested to be due to chronic progressive demyelination (1). Because NT3 has been known as a strong signal to promote myelination (43,53), we reasoned that the recovery of contralateral MEP responses in PCL-F3. NT3 group was due, at least in part, to improved myelination in contralateral white matter. Eriochrome cyanine (EC) staining of myelin in contralateral white matter of CTL group showed a coarse arrangement of myelinating fibers with frequent unstained foci interspersed with myelin fragments (Fig. 6I), and the staining pattern was similar with PCL and PCL-F3 groups (Fig. 6J, K). In contrast, myelinating fibers were more compactly arranged in PCL-F3.NT3 group without apparent demyelinating foci (Fig. 6L). Quantification revealed significant increases of EC-stained areas in PCL-F3.NT3 group compared to the other groups (p < 0.001) (Fig. 6M).

Degradation of CSPGs Enhanced Migration of Transplanted Cells and Axonal Remodeling

It has been shown that CSPGs deposition around the injury sites inhibits axonal regeneration and/or cellular migration (17,26,45). Therefore, we asked if the combined use of CSPG degrading enzyme could provide additional therapeutic benefits in animals with implantation of PCL scaffold with F3.NT3 cells. Chondroitinase ABC (C-ase) or a control penicillinase (P-ase) was infused using osmotic minipumps in animals with PCL-F3.NT3 implantation. C-ase infusion almost completely abolished CS56 staining (intact CSPGs), which was densely observed around PCL scaffolds in P-ase control animals (Fig. 7). C-ase treatment also revealed immunoreactivity against C4S (degraded CSPGs), indicating that C-ase successfully degraded CSPGs at the interface between PCL scaffolds and spinal cord tissue. Degradation of CSPGs promoted the migration of F3.NT3 cells into residual spinal cord tissue (Fig. 8A, B). In P-ase group, F3.NT3 cells migrated across the interface between scaffolds and spinal cord, but a large proportion of migrating cells did not cross the GFAP-positive glial scars or remained adjacent to glial scar fronts (Fig. 8C, D). Combination of C-ase increased the extent of F3.NT3 cells that crossed glial scar fronts, especially in the caudal direction (Fig. 8E, F). The total number of cells was not significantly different between C-ase and P-ase groups (34,228 ± 17,706 and 25,695 ± 4,226, respectively; p = 0.18). However, numbers of F3.NT3 cells that migrated into rostrocaudal spinal cord were significantly higher at several sites in C-ase group (Fig. 8G), and the longest migration distance from the implant was also significantly greater in C-ase group (p < 0.01). The extent of differentiation of F3.NT3 cells into mature neural cell phenotypes was not different between P-ase and C-ase groups (data not shown).

Figure 7.

Chondroitinase ABC abolished chondroitin sulfate proteoglycans (CSPGs) at the interface between PCL scaffolds and residual spinal cord tissue. Animals with PCL-F3.NT3 implants were treated with penicillinase (P-ase, A–D) or chondroitinase ABC (C-ase, E–H), respectively. The effectiveness of C-ase to degrade CSPGs was examined using antibodies against intact (CS56, green) and digested (C4S, red) chondroitin sulfate. C-ase infusion almost completely abolished CS56 staining, which was densely observed around the PCL scaffold in P-ase control. Scale bar: 250 μm (A, E), 50 μm (B–D, F–H).

Figure 8.

Chondroitinase ABC treatment promoted migration of F3.NT3 NSCs. (A, B) Representative longitudinal spinal cord sections of penicillinase (P-ase, A) and chondroitinase ABC (C-ase, B) treated animals with PCL-F3.NT3 implants at 9 weeks post-SCI. Spinal cord sections were stained with GFAP (green) and human-specific mitochondria antibody (red) to detect grafted NSCs of human origin. F3.NT3 cells in C-ase group showed more frequent migration in a rostrocaudal direction. Rostral direction is leftwards. Dotted lines demarcate scaffold borders. Scale bar: 250 μm. (C–F) Higher magnification images of boxed regions in the longitudinal section of P-ase (C, D) or C-ase (E, F) groups. Note that F3.NT3 cells more readily cross glial scar fronts. DAPI-stained nuclei are shown in blue. Scale bar: 50 μm. (G) Stereological counting of grafted cells. ^*p < 0.05 and ^**p < 0.01 by repeated measures two-way ANOVA followed by Tukey's post hoc analysis. Error bars represent mean ± SEM. (H) Quantification of maximal migration distance. Each triangle or circle represents maximal migration distance in each animal, and each line represents group mean value. ^**p < 0.01 by Student's t-test. N = 9 per group.

The extent of axonal growth into PCL scaffolds (Fig. 9A-D) was also evaluated. In P-ase group, only a small number of short neurofilament-positive axons were observed inside PCL scaffolds, and these were was mostly limited to the very vicinity of the interface (Fig. 9A, B). In contrast, the number of axons observed inside PCL scaffolds was markedly higher in C-ase group (Fig. 9C, D, I), and the lengths of ingrowing axons were also greater in C-ase group (Fig. 9J). We also compared the extent of lumbar ventral horn innervation by serotonergic fibers (Fig. 9E-H), which are thought to play an important role in locomotor recovery after SCI (8). Although serotonergic (5-HT) axon areas in the contralateral ventral horn tended to be higher in C-ase group, the difference was not statistically significant (6.1 ± 4.9% and 9.6 ± 6.0%. in P-ase and C-ase groups, respectively; p = 0.19). In P-ase group, mean serotonergic axon density in the ipsilateral ventral horn was only 32.8 ± 8.8% of contralateral side (Fig. 9K). C-ase treatment substantially increased the extent of serotonergic innervation in the ipsilateral side, reaching 56.2 ± 11.5% of the contralateral side (Fig. 9H, K).

Figure 9.

Chondroitinase ABC treatment in combination with PCL-F3.NT3 implantation promoted axonal remodeling. (A–D) Longitudinal spinal cord sections of animals with PCL-F3.NT3 implants treated with penicillinase (P-ase) (A, B) or chondroitinase ABC (C-ase) (C, D). (B) and (D) are magnified images of the boxed regions in (A) and (C), respectively. Neurofilament (NF)-positive axons are shown in green. NF axons were found to grow into PCL scaffolds (arrows) in C-ase treated group. PCL scaffolds were often visualized by background staining (arrowheads, red) with anti-human mitochondria antibody. Dotted lines indicate scaffold borders. (E–H) Serotonin (5-HT) axons were visualized by immunofluorescence staining in transverse sections of the caudal spinal cord of animal treated with P-ase (E, F) or C-ase (G, H). 5-HT axon innervation in the ventral gray matter in the lesioned side (right) (F, H) was markedly reduced compared to that in the contralateral side (left) (E, G). In C-ase group, ipsilateral 5-HT axonal densities were higher than those in P-ase group (H). (I–K) Quantitation graphs of the numbers (I) and lengths (J) of NF axons growing into PCL, and the percent ipsilateral 5-HT axon density compared to contralateral side (K). ^***p < 0.001 by Student's t-test. Error bars represent mean ± SEM. Scale bars: 100 μm.

Degradation of CSPGs Augmented the Functional Recovery Mediated by PCL-F3.NT3 Implantation

BBB locomotor test was performed to examine whether C-ase treatment further improved the quality of hindlimb locomotion in animals implanted with PCL-F3.NT3. Animals treated with C-ase (N = 9) showed behavioral recovery superior to P-ase group (N = 9) in both hindlimbs (Fig. 10A, B). Repeated measures two-way ANOVA revealed a significant treatment effect over time (p < 0.05 for both hindlimbs). Locomotor performance was also evaluated at 4 and 7 weeks post-SCI using the grid walk test. C-ase treatment significantly improved the accuracy of hindlimb placement compared to P-ase treatment in both hindlimbs (Fig. 10C, D).

Figure 10.

Functional recovery following chondroitinase ABC treatment in combination with PCL-F3.NT3 implantation. (A, B) Recovery of BBB locomotor scores over the 7-week period after injury for ipsilateral (right) (A) and contralateral (left) (B) hindlimbs. Chondroitinase (C-ase, N = 9) treatment significantly improved recovery of BBB scores over time in animals with PCL-F3.NT3 implantation compared with penicillinase (P-ase, N = 9) group. ^*p < 0.05 by repeated measures two-way ANOVA. (C, D) Grid walk. Average numbers of hindlimb placement errors per run were counted at 4 and 7 weeks (4W and 7W) post-SCI for ipsilateral (C) and contralateral (D) hindlimbs. C-ase treatment significantly decreased the number of errors in both hindlimbs. ^*p < 0.05 and ^**p < 0.01 by Student's t-test. N = 9 per group. (E, F) Comparison of the mean amplitudes of motor evoked potentials (MEPs) recorded in ipsilateral (E) and contralateral (F) gastrocnemius muscles at 8 weeks post-SCI. C-ase treatment significantly improved the mean amplitude of MEPs in ipsilateral hindlimbs. ^*p < 0.05 by Student's t-test. N = 9 per group for all graphs. Error bars represent mean ± SEM.

MEP responses to transcranial magnetic stimulation were measured to compare electrophysiological recovery (Fig. 10E, F). C-ase treatment in animals implanted with PCL-F3.NT3 recovered recognizable MEP responses in ipsilateral hindlimbs, where no significant MEP recovery was observed by PCL-F3.NT3 implantation (Fig. 5A, C). Only 11% (1/9) of animals in P-ase group exhibited ipsilateral MEP response with the amplitude of more than 1 mV, whereas this percentage was elevated in C-ase group (67%, 6/9). The mean amplitude was also significantly different in C-ase group (1.4 ± 0.7 mV in C-ase and 0.6 ± 0.5 mV in P-ase group, p < 0.05) (Fig. 10E). However, MEP amplitude was not significantly changed in contralateral hindlimbs (Fig. 10F), where implantation of PCL scaffolds with F3.NT3 already significantly recovered MEP responses (Fig. 5B, D). Collectively, these data indicated that degradation of CSPGs augmented the behavioral and electrophysiological recovery mediated by PCL-F3.NT3 implantation.

Discussion

The present study demonstrated that the implantation of polymer scaffold seeded with NT3 overexpressing NSCs provided marked behavioral and electrophysiological recovery in a rat hemisection SCI model. Degradation of CSPGs deposited at the interface between scaffolds and spinal cord tissue further enhanced the therapeutic benefits of PCL-F3.NT3 implantation. The full combination enabled the injured rats to attain well-coordinated plantar stepping accompanied by improved ankle positioning and toe clearance (mean BBB score, 16.0 ± 1.5 and 17.4 ± 1.3 for ipsi- and contralateral hindlimbs, respectively) and reduced paw placement errors on grid to a minimum (3.5 ± 1.0 and 2.1 ± 1.0 per run for ipsi- and contralateral hindlimbs, respectively). Furthermore, animals with the full complement of combinatorial strategies regained bilateral MEP responses to transcranial magnetic stimulation. Therefore, our study highlights the additive benefits conferred by combining the following complementary strategies: 1) transplantation of NSCs with polymer scaffold, 2) incorporation of neurotrophic factor gene, and 3) modulation of inhibitory extracellular matrix. The combination of all these strategies together was feasible and led to greatest functional recovery, providing an applicable and promising framework for future NSC-based transplantation therapies targeting spinal cord repair.

The implantation of artificial scaffolds has been employed to bridge lesion cavities and to create a favorable tissue environment for axonal growth (9,21). Various scaffolding materials including biodegradable polymers have been tested in animal models of SCI (24,47,54). In the present study, we found that the polymer scaffold itself could provide some therapeutic benefits in terms of behavioral recovery. Animals with PCL scaffold alone exhibited significantly higher precision in hindlimb placement during the grid walk than those with injury alone. A previous study also reported a significant functional improvement by implantation of polymer scaffold (57). Implantation of polymer scaffolds may provide mechanical support and prevent instability-related functional impairments. Furthermore, implanted polymer scaffolds may deter the infiltration of inflammatory cells and fibroblasts into the lesion cavities with suppression of glial scar formation (57). In addition to their own positive contribution, bioengineered polymer scaffolds may find their versatile utility in integrating combinatorial strategies, especially cell-based interventions (47,61). As also shown in the present study, polymer scaffolds have proved to be compatible with a variety of cell types and successfully delivered cells of therapeutic interests to the lesioned spinal cord (23,47,54,56). The PCL scaffold in our study appeared to function like a reservoir supplying migratory NSCs to the host spinal cord. Polymer scaffolds are primarily expected to function as a bridge filling tissue defects (9). Although the PCL scaffolds physically filled the lesion cavity by providing structural support, they were not able to guide axonal growth without the addition of C-ase infusion. It has been suggested that glial scars forming on- and off-ramps of polymer scaffolds can impede their bridging function by preventing axonal growth in and out of polymer scaffolds (21).

Parental NSCs (F3 cells) grafted with PCL scaffolds survived the transplantation and migrated to host spinal cord. Although implantation of PCL scaffold seeded with F3 cells appeared to improve accuracy of hindlimb placement compared to PCL alone group, behavioral recovery assessed by BBB score was not significantly different between PCL-alone and PCL-F3 groups. The incorporation of NT3 gene into F3 cells substantially improved cell survival and migration. F3.NT3 cells were more numerous than F3 cells along the rostrocaudal extent from the implantation site. Furthermore, the maximal distance of F3.NT3 migration was also markedly greater than that of F3 cells. A previous study reported that NT3 treatment, by acting through its cognate receptor TrkC, enhanced the survival and migration of transplanted NSCs in the rat spinal cord (13). In addition, we found that F3.NT3 cells differentiated into mature neurons and oligodendrocytes more frequently than F3 cells. Therefore, the incorporation of NT3 gene could positively regulate the intrinsic properties of grafted NSCs such as survival, migration, and differentiation in an autocrine or paracrine fashion. As a result, a much larger number of new neurons and oligodendrocytes derived from NSCs were supplied to the host spinal cord by PCL-F3.NT3 implantation.

The animals implanted with PCL-F3.NT3 exhibited better locomotor recovery than those in the other three study groups. We speculate that several mechanisms may be responsible for PCL-F3.NT3-mediated functional recovery. Increased NT3 concentration in host spinal cord tissue (Fig. 6A) could promote neuroplasticity in the caudal spinal cord, which is directly responsible for hindlimb locomotion. It has been previously demonstrated that NT3 enhances the sprouting of corticospinal axons and that NT3 signaling plays a critical role in synaptogenesis (25,42,55). The increases in the expression of GAP-43 and synaptotagmin indicated that PCL-F3.NT3 implantation markedly promoted neuroplasticity in the caudal spinal cord. Intriguingly, the increase in GAP-43 immunoreactivity was found in the intermediate regions (where corticospinal fibers terminate) and ventral motor regions (where ventral motor neurons receive supraspinal inputs). This suggests that neuroplasticity was enhanced in a regionally specific manner and that synaptic connections were actively remodeled only in motor pathway, but not in behaviorally irrelevant regions. It is also possible that F3.NT3-derived neurons migrating to host spinal cord contributed to the observed locomotor recovery by participating in the formation of new intraspinal relay circuits (2). NT3 can directly enhance oligodendrocyte precursor proliferation and survival (3,4) and also promotes myelination by endogenous oligodendrocytes (43). In this regard, it was intriguing to find that MEP response was not elicited in the contralateral hindlimbs after hemisection injury. A recent study demonstrated that electrical transmission through contralateral spared white matter was impaired, and this was ascribed to a decrease in the number of myelinated axons (1). Our data strongly suggested that the absence of MEP response in contralateral hindlimb (Fig. 5B, D) was due to the demyelination of contralateral white matter and that MEP recovery could be explained by greater myelination in PCL-F3.NT3 group.

Although PCL-F3.NT3 implantation resulted in substantial recovery of locomotor function, we hypothesized that CSPGs surrounding PCL scaffolds would still limit the extent of functional benefits. The present study showed that degradation of CSPGs by intrathecal infusion of C-ase enhanced migration of F3.NT3 cells to the spinal cord and increased the extent of axonal growth into PCL scaffolds. This enhancement in migration induced by C-ase could increase the number of F3.NT3-derived neurons or oligodendrocytes located in host spinal cord, allowing more chances for intraspinal relay circuit formation or replacing lost oligodendrocytes. Because surrounding CSPGs inhibit axonal growth into scaffolds, PCL scaffolds would not function adequately as an actual bridge without C-ase (21). Therefore, degradation of CSPGs may be essentially required for the proper bridging functions either by polymer scaffold or other biological grafts (14,32). The increases in ipsilateral 5-HT innervation may explain our observation that MEP responses were significantly recovered by C-ase in ipsilateral hindlimbs (Fig. 10E), where PCL-F3.NT3 implantation failed to restore recognizable MEP responses (Fig. 5A, C). Combination of C-ase treatment could promote the establishment of new axonal connections to the ipsilateral lumbar motor neurons by supporting axon growth directly through implanted PCL scaffolds. Alternatively, C-ase treatment might promote the sprouting of contralateral spared axons to the ipsilateral lumbar regions (5,8). At any rate, these findings indicated that combination of C-ase treatment complemented the PCL-F3.NT3 implantation by strengthening ipsilateral supraspinal connections to caudal ventral horn.

One caveat in this study is that we use the hemisection SCI model, which has less clinical relevance compared to contusion or compression injury. We chose this model because it creates a defined cavity into which the solid PCL scaffold can be implanted without additional surgical damages. Because contusion or compression injury is accompanied by progressive cavitation that is characteristic of human SCI (58), these models would be more suited for a preclinical testing of a new therapeutic strategy. However, irregular and unpredictable shape and dimensions of lesion cavities would make it almost impossible to implant the solid PCL scaffold. Injectable hydrogels that can cast into irregular cavities would allow us to test the current combinatorial strategy in a more clinically relevant model (22,28). The timing of the combinatorial treatment also needs to be addressed. Because our acute treatment paradigm cannot be instituted for human patients, it would be desirable to test the treatment efficacy in a chronic injury setting. In this regard, a recent study that demonstrated the therapeutic effects of the combinatorial strategy involving hydrogel scaffold and mesenchymal stem cells appears to be highly encouraging (23).

Summarizing, this study demonstrated that combining multifaceted strategies could maximize the therapeutic benefits of NSC transplantation to repair the injured spinal cord. We believe that our results have important clinical implications regarding the future design of NSC-based therapeutic strategies. Developing an appropriate combinatorial strategy would accelerate the clinical translation of NSC-based transplantation approaches for human victims suffering from traumatic SCI.

Footnotes

Acknowledgments

This study was supported by Korea Research Foundation Grant (KRF-2006-E00439, B.G.K), Brain Science Fundamental Technology Development Project (2009-0081466, B.G.K), Ajou University School of Medicine (2007-2008, B.K.G.), and 21st Century Frontier Research Fund (SC-3111, B.W.Y). The authors declare no conflicts of interest.

References

Arvanian

V. L.

; Schnell

; Lou

; Golshani

; Hunanyan

; Ghosh

; Pearse

D. D.

; Robinson

J. K.

; Schwab

M. E.

; Fawcett

J. W.

; Mendell

L. M.

Chronic spinal hemisection in rats induces a progressive decline in transmission in uninjured fibers to motoneurons. Exp. Neurol. 216(2): 471–480; 2009.

Bareyre

F. M.

; Kerschensteiner

; Raineteau

; Mettenleiter

T. C.

; Weinmann

; Schwab

M. E.

The injured spinal cord spontaneously forms a new intraspinal circuit in adult rats. Nat. Neurosci. 7(3): 269–277; 2004.

Barres

B. A.

; Raff

M. C.

; Gaese

; Bartke

; Dechant

; Barde

Y. A.

A crucial role for neurotrophin-3 in oligodendrocyte development. Nature 367(6461): 371–375; 1994.

Barres

B. A.

; Schmid

; Sendnter

; Raff

M. C.

Multiple extracellular signals are required for long-term oligodendrocyte survival. Development 118(1): 283–295; 1993.

Barritt

A. W.

; Davies

; Marchand

; Hartley

; Grist

; Yip

; McMahon

S. B.

; Bradbury

E. J.

Chondroitinase ABC promotes sprouting of intact and injured spinal systems after spinal cord injury. J. Neurosci. 26(42): 10856–10867; 2006.

Basso

D. M.

; Beattie

M. S.

; Bresnahan

J. C.

Graded histological and locomotor outcomes after spinal cord contusion using the NYU weight-drop device versus transection. Exp. Neurol. 139(2): 244–256; 1996.

Blesch

; Lu

; Tuszynski

M. H.

Neurotrophic factors, gene therapy, and neural stem cells for spinal cord repair. Brain Res. Bull. 57(6): 833–838; 2002.

Bregman

B. S.

Development of serotonin immunoreactivity in the rat spinal cord and its plasticity after neonatal spinal cord lesions. Brain Res. 431(2): 245–263; 1987.

Bunge

M. B.

Bridging areas of injury in the spinal cord. Neuroscientist 7(4): 325–339; 2001.

10.

Cao

; Xu

X. M.

; Devries

W. H.

; Enzmann

G. U.

; Ping

; Tsoulfas

; Wood

P. M.

; Bunge

M. B.

; Whittemore

S. R.

Functional recovery in traumatic spinal cord injury after transplantation of multineurotrophin-expressing glial-restricted precursor cells. J. Neurosci. 25(30): 6947–6957; 2005.

11.

Cao

Q. L.

; Howard

R. M.

; Dennison

J. B.

; Whittemore

S. R.

Differentiation of engrafted neuronal-restricted precursor cells is inhibited in the traumatically injured spinal cord. Exp. Neurol. 177(2): 349–359; 2002.

12.

Cao

Q. L.

; Zhang

Y. P.

; Howard

R. M.

; Walters

W. M.

; Tsoulfas

; Whittemore

S. R.

Pluripotent stem cells engrafted into the normal or lesioned adult rat spinal cord are restricted to a glial lineage. Exp. Neurol. 167(1): 48–58; 2001.

13.

Castellanos

D. A.

; Tsoulfas

; Frydel

B. R.

; Gajavelli

; Bes

J. C.

; Sagen

TrkC overexpression enhances survival and migration of neural stem cell transplants in the rat spinal cord. Cell Transplant. 11(3): 297–307; 2002.

14.

Chau

C. H.

; Shum

D. K.

; Li

; Pei

; Lui

Y. Y.

; Wirthlin

; Chan

Y. S.

; Xu

X. M.

Chondroitinase ABC enhances axonal regrowth through Schwann cell-seeded guidance channels after spinal cord injury. FASEB J. 18(1): 194–196; 2004.

15.

Cummings

B. J.

; Uchida

; Tamaki

S. J.

; Salazar

D. L.

; Hooshmand

; Summers

; Gage

F. H.

; Anderson

A. J.

Human neural stem cells differentiate and promote locomotor recovery in spinal cord-injured mice. Proc. Natl. Acad. Sci. USA 102(39): 14069–14074; 2005.

16.

Eftekharpour

; Karimi-Abdolrezaee

; Fehlings

M. G.

Current status of experimental cell replacement approaches to spinal cord injury. Neurosurg. Focus 24(3-4): E19; 2008.

17.

Fitch

M. T.

; Silver

Glial cell extracellular matrix: Boundaries for axon growth in development and regeneration. Cell Tissue Res. 290(2): 379–384; 1997.

18.

Flax

J. D.

; Aurora

; Yang

; Simonin

; Wills

A. M.

; Billinghurst

L. L.

; Jendoubi

; Sidman

R. L.

; Wolfe

J. H.

; Kim

S. U.

; Snyder

E. Y.

Engraftable human neural stem cells respond to developmental cues, replace neurons, and express foreign genes. Nat. Biotechnol. 16(11): 1033–1039; 1998.

19.

Fouad

; Schnell

; Bunge

M. B.

; Schwab

M. E.

; Liebscher

; Pearse

D. D.

Combining Schwann cell bridges and olfactory-ensheathing glia grafts with chondroitinase promotes locomotor recovery after complete transection of the spinal cord. J. Neurosci. 25(5): 1169–1178; 2005.

20.

Fournier

A. E.

; GrandPre

; Strittmatter

S. M.

Identification of a receptor mediating Nogo-66 inhibition of axonal regeneration. Nature 409(6818): 341–346; 2001.

21.

Geller

H. M.

; Fawcett

J. W.

Building a bridge: Engineering spinal cord repair. Exp. Neurol. 174(2): 125–136; 2002.

22.

Hejcl

; Lesny

; Pradny

; Michalek

; Jendelova

; Stulik

; Sykova

Biocompatible hydrogels in spinal cord injury repair. Physiol. Res. 57(Suppl. 3): S121–132; 2008.

23.

Hejcl

; Sedy

; Kapcalova

; Toro

D. A.

; Amemori

; Lesny

; Likavcanova-Masinova

; Krumbholcova

; Pradny

; Michalek

; Burian

; Hajek

; Jendelova

; Sykova

HPMA-RGD hydrogels seeded with mesenchymal stem cells improve functional outcome in chronic spinal cord injury. Stem Cells Dev. 19(10): 1535–1546; 2010.

24.

Hejcl

; Urdzikova

; Sedy

; Lesny

; Pradny

; Michalek

; Burian

; Hajek

; Zamecnik

; Jendelova

; Sykova

Acute and delayed implantation of positively charged 2-hydroxyethyl methacrylate scaffolds in spinal cord injury in the rat. J. Neurosurg. Spine 8(1): 67–73; 2008.

25.

Iarikov

D. E.

; Kim

B. G.

; Dai

H. N.

; McAtee

; Kuhn

P. L.

; Bregman

B. S.

Delayed transplantation with exogenous neurotrophin administration enhances plasticity of corticofugal projections after spinal cord injury. J. Neurotrauma 24(4): 690–702; 2007.

26.

Ikegami

; Nakamura

; Yamane

; Katoh

; Okada

; Iwanami

; Watanabe

; Ishii

; Kato

; Fujita

; Takahashi

; Okano

H. J.

; Toyama

; Okano

Chondroitinase ABC combined with neural stem/progenitor cell transplantation enhances graft cell migration and outgrowth of growth-associated protein-43-positive fibers after rat spinal cord injury. Eur. J. Neurosci. 22(12): 3036–3046; 2005.

27.

Imitola

; Park

K. I.

; Teng

Y. D.

; Nisim

; Lachyankar

; Ourednik

; Mueller

F. J.

; Yiou

; Atala

; Sidman

R. L.

; Tuszynski

; Khoury

S. J.

; Snyder

E. Y.

Stem cells: Cross-talk and developmental programs. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359(1445): 823–837; 2004.

28.

Jain

; Kim

Y. T.

; McKeon

R. J.

; Bellamkonda

R. V.

In situ gelling hydrogels for conformal repair of spinal cord defects, and local delivery of BDNF after spinal cord injury. Biomaterials 27(3): 497–504; 2006.

29.

Johnson

P. J.

; Tatara

; Shiu

; Sakiyama-Elbert

S. E.

Controlled release of neurotrophin-3 and platelet derived growth factor from fibrin scaffolds containing neural progenitor cells enhances survival and differentiation into neurons in a subacute model of SCI. Cell Transplant. 19(1): 89–101; 2010.

30.

Karimi-Abdolrezaee

; Eftekharpour

; Wang

; Morshead

C. M.

; Fehlings

M. G.

Delayed transplantation of adult neural precursor cells promotes remyelination and functional neurological recovery after spinal cord injury. J. Neurosci. 26(13): 3377–3389; 2006.

31.

Keirstead

H. S.

; Nistor

; Bernal

; Totoiu

; Cloutier

; Sharp

; Steward

Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remyelinate and restore locomotion after spinal cord injury. J. Neurosci. 25(19): 4694–4705; 2005.

32.

Kim

B. G.

; Dai

H. N.

; Lynskey

J. V.

; McAtee

; Bregman

B. S.

Degradation of chondroitin sulfate proteoglycans potentiates transplant-mediated axonal remodeling and functional recovery after spinal cord injury in adult rats. J. Comp. Neurol. 497(2): 182–198; 2006.

33.

Kim

B. G.

; Hwang

D. H.

; Lee

S. I.

; Kim

E. J.

; Kim

S. U.

Stem cell-based cell therapy for spinal cord injury. Cell Transplant. 16(4): 357–366; 2007.

34.

Kim

S. U.

Human neural stem cells genetically modified for brain repair in neurological disorders. Neuropathology 24(3): 159–171; 2004.

35.

Kinouchi

; Takeda

; Yang

; Wilhelmsson

; Lundkvist

; Pekny

; Chen

D. F.

Robust neural integration from retinal transplants in mice deficient in GFAP and vimentin. Nat. Neurosci. 6(8): 863–868; 2003.

36.

Kweon

; Yoo

M. K.

; Park

I. K.

; Kim

T. H.

; Lee

H. C.

; Lee

H. S.

; Oh

J. S.

; Akaike

; Cho

C. S.

A novel degradable polycaprolactone networks for tissue engineering. Biomaterials 24(5): 801–808; 2003.

37.

Lee

S. I.

; Kim

B. G.

; Hwang

D. H.

; Kim

H. M.

; Kim

S. U.

Overexpression of Bcl-XL in human neural stem cells promotes graft survival and functional recovery following transplantation in spinal cord injury. J. Neurosci. Res. 87(14): 3186–3197; 2009.

38.

Louro

; Pearse

D. D.

Stem and progenitor cell therapies: Recent progress for spinal cord injury repair. Neurol. Res. 30(1): 5–16; 2008.

39.

Loy

D. N.

; Magnuson

D. S.

; Zhang

Y. P.

; Onifer

S. M.

; Mills

M. D.

; Cao

Q. L.

; Darnall

J. B.

; Fajardo

L. C.

; Burke

D. A.

; Whittemore

S. R.

Functional redundancy of ventral spinal locomotor pathways. J. Neurosci. 22(1): 315–323; 2002.

40.

; Jones

L. L.

; Snyder

E. Y.

; Tuszynski

M. H.

Neural stem cells constitutively secrete neurotrophic factors and promote extensive host axonal growth after spinal cord injury. Exp. Neurol. 181(2): 115–129; 2003.

41.

Madhavan

; Ourednik

Neural stem/progenitor cells initiate the formation of cellular networks that provide neuroprotection by growth factor-modulated antioxidant expression. Stem Cells 26(1): 254–265; 2008.

42.

Martinez

; Alcantara

; Borrell

; Del Rio

J. A.

; Blasi

; Otal

; Campos

; Boronat

; Barbacid

; Silos-Santiago

; Soriano

TrkB and TrkC signaling are required for maturation and synaptogenesis of hippocampal connections. J. Neurosci. 18(18): 7336–7350; 1998.

43.

McTigue

D. M.

; Horner

P. J.

; Stokes

B. T.

; Gage

F. H.

Neurotrophin-3 and brain-derived neurotrophic factor induce oligodendrocyte proliferation and myelination of regenerating axons in the contused adult rat spinal cord. J. Neurosci. 18(14): 5354–5365; 1998.

44.

Mitsui

; Shumsky

J. S.

; Lepore

A. C.

; Murray

; Fischer

Transplantation of neuronal and glial restricted precursors into contused spinal cord improves bladder and motor functions, decreases thermal hypersensitivity, and modifies intraspinal circuitry. J. Neurosci. 25(42): 9624–9636; 2005.

45.

Monnier

P. P.

; Sierra

; Schwab

J. M.

; Henke-Fahle

; Mueller

B. K.

The Rho/ROCK pathway mediates neurite growth-inhibitory activity associated with the chondroitin sulfate proteoglycans of the CNS glial scar. Mol. Cell. Neurosci. 22(3): 319–330; 2003.

46.

Mothe

A. J.

; Kulbatski

; Parr

; Mohareb

; Tator

C. H.

Adult spinal cord stem/progenitor cells transplanted as neurospheres preferentially differentiate into oligodendrocytes in the adult rat spinal cord. Cell Transplant. 17(7): 735–751; 2008.

47.

Nomura

; Tator

C. H.

; Shoichet

M. S.

Bioengineered strategies for spinal cord repair. J. Neurotrauma 23(3-4): 496–507; 2006.

48.

Okada

; Ishii

; Yamane

; Iwanami

; Ikegami

; Katoh

; Iwamoto

; Nakamura

; Miyoshi

; Okano

H. J.

; Contag

C. H.

; Toyama

; Okano

In vivo imaging of engrafted neural stem cells: Its application in evaluating the optimal timing of transplantation for spinal cord injury. FASEB J. 19(13): 1839–1841; 2005.

49.

Park

K. I.

; Himes

B. T.

; Stieg

P. E.

; Tessler

; Fischer

; Snyder

E. Y.

Neural stem cells may be uniquely suited for combined gene therapy and cell replacement: Evidence from engraftment of Neurotrophin-3-expressing stem cells in hypoxic-ischemic brain injury. Exp. Neurol. 199(1): 179–190; 2006.

50.

Pearse

D. D.

; Bunge

M. B.

Designing cell- and gene-based regeneration strategies to repair the injured spinal cord. J. Neurotrauma 23(3-4): 438–452; 2006.

51.

Pfeifer

; Vroemen

; Blesch

; Weidner

Adult neural progenitor cells provide a permissive guiding substrate for corticospinal axon growth following spinal cord injury. Eur. J. Neurosci. 20(7): 1695–1704; 2004.

52.

Rowland

J. W.

; Hawryluk

G. W.

; Kwon

; Fehlings

M. G.

Current status of acute spinal cord injury pathophysiology and emerging therapies: Promise on the horizon. Neurosurg. Focus 25(5): E2; 2008.

53.

Rubio

; Rodriguez

; Arevalo

M. A.

In vitro myelination by oligodendrocyte precursor cells transfected with the neurotrophin-3 gene. Glia 47(1): 78–87; 2004.

54.

Schmidt

C. E.

; Leach

J. B.

Neural tissue engineering: Strategies for repair and regeneration. Annu. Rev. Biomed. Eng. 5: 293–347; 2003.

55.

Schnell

; Schneider

; Kolbeck

; Barde

Y. A.

; Schwab

M. E.

Neurotrophin-3 enhances sprouting of corticospinal tract during development and after adult spinal cord lesion. Nature 367(6459): 170–173; 1994.

56.

Sykova

; Jendelova

; Urdzikova

; Lesny

; Hejcl

Bone marrow stem cells and polymer hydrogels—two strategies for spinal cord injury repair. Cell. Mol. Neurobiol. 26(7-8): 1113–1129; 2006.

57.

Teng

Y. D.

; Lavik

E. B.

; Qu

; Park

K. I.

; Ourednik

; Zurakowski

; Langer

; Snyder

E. Y.

Functional recovery following traumatic spinal cord injury mediated by a unique polymer scaffold seeded with neural stem cells. Proc. Natl. Acad. Sci. USA 99(5): 3024–3029; 2002.

58.

Tuszynski

M. H.

; Gabriel

; Gerhardt

; Szollar

Human spinal cord retains substantial structural mass in chronic stages after injury. J. Neurotrauma 16(6): 523–531; 1999.

59.

Vroemen

; Aigner

; Winkler

; Weidner

Adult neural progenitor cell grafts survive after acute spinal cord injury and integrate along axonal pathways. Eur. J. Neurosci. 18(4): 743–751; 2003.

60.

Yan

; Xu

; Welsh

A. M.

; Hatfield

; Hazel

; Johe

; Koliatsos

V. E.

Extensive neuronal differentiation of human neural stem cell grafts in adult rat spinal cord. PLoS Med. 4(2): e39; 2007.

61.

Zahir

; Nomura

; Guo

X. D.

; Kim

; Tator

; Morshead

; Shoichet

Bioengineering neural stem/progenitor cell-coated tubes for spinal cord injury repair. Cell Transplant. 17(3): 245–254; 2008.

62.

Zhang

; Gu

; Zhao

; Wen

Combined treatment of neurotrophin-3 gene and neural stem cells is propitious to functional recovery after spinal cord injury. Cell Transplant. 16(5): 475–481; 2007.