Abstract
The present study investigated the ability of cotransplantation of Schwann cells (SCs) and olfactory ensheathing cells (OECs) combined with treadmill training in facilitating neuronal plasticity and promoting hindlimb function recovery of subacute moderate thoracic (T10) spinal cord contusion in rats. Two weeks postinjury, SCs were injected directly into the lesion, while OECs were injected into the adjacent tissues. The treadmill training with the rats began postinjury on day 7, with each session lasting 20 ± 10 min per day, 5 days per week, for 10 weeks. At the 11th week postinjury, OECs were found migrating longitudinally and laterally from the injection site to the injury site through the gray and white matter, while some traveled along the central canal or pia. The SCs remained densely packed and concentrated at the transplant site. The transplanted SCs supported ingrowth of numerous, densely populated neurofilament-positive (NF+), MBP+ axons. The OECs promoted elongation of moderate NF+, GAP-43+ axons and a few MBP+ axons in parallel with OEC processes. The GFAP immunoreactivity in the spared tissue surrounding the graft of SCs and OECs at the lesion site was less intense than that in the DMEM group. Treadmill training had no effect on GFAP immunoreactivity. Treadmill training increased the number of TH-immunoreactive neurons in the gray matter of L2 spinal cord. Moreover, cotransplantation of OECs and SCs significantly increased the BBB score during 5–8 weeks postinjury alongside treadmill training between 5 and 11 weeks. Cotransplantation of OECs and SCs combined with treadmill training resulted in the highest BBB score at 4 and 11 weeks. The study details the differential mechanisms of neuronal plasticity: (1) axon growth and remyelination induced by cotransplantation of OECs and SCs and (2) neuron plasticity below the lesion enhanced by treadmill training. The synergistic effects of the combined strategy enhance functional recovery. This manuscript is published as part of the International Association of Neurorestoratology (IANR) supplement issue of Cell Transplantation.
Keywords
Introduction
Contusive spinal cord injury (SCI) gives rise to a complex lesion that includes cellular and axonal loss, demyelination, involvement of growth inhibitory molecules, and astroglial scar or cavity formation (53). These changes limit axon regeneration and result in permanent neurological deficits in people with SCI. In order to optimize functional recovery of the spinal cord, some combination strategies have been tested (12,19,48). One such approach combines cell transplantation with rehabilitation therapy.
Transplantation of olfactory ensheathing cells (OECs) and Schwann cells (SCs) may be a very promising therapeutic strategy for spinal cord repair. It has previously been reported that OECs can express and secrete growth-promoting neurotrophic factors (40,55,62), adhesion molecules (4,6), and extracellular matrix proteins (9,56). Furthermore, OEC transplantation reduces nerve tissue loss, facilitates axon growth and regeneration, promotes remyelination, and improves motor function in various models of spinal cord injury (20,31,32,37,38,41,54). SCs have also been reported to produce multiple growth factors (45,61), cell adhesion molecules (23), and extracellular matrix molecules (52). They also support the elongation of axons (13) and remyelination (11,57,63). Compared with OECs, the SC graft is more effective in promoting axonal sparing, regeneration, and functional improvement than SCs/OECs combined. This is true for the acute (11), subacute (49,57), and chronic (2) moderately contused thoracic spinal cord in adult rats. However, it is the transplanted OECs (54,55) rather than the SCs (63) that show the long distance migratory ability in the spinal tissue. An OEC transplant demonstrates more favorable interactions with astrocytes and a potential advantage in promoting axonal growth through the astroglial scar over SCs into both the uninjured and intact spinal cord (1). Our preliminary investigations also demonstrated similar results.
Treadmill training has been found to induce spinal plasticity (16,18,22,29,33,50,59) and to promote recovery of stepping in spinal cord-transected animals (3,15,42) and also in severe, but incomplete, spinal cord injury patients (17,24,26,60). Treadmill training can further enhance hindlimb stepping after OEC transplantation in spinal cord-transected rats (33) and improve motor functions in complete spinal cord injury patients (39), while step training alone had no effect.
This study aims to examine the efficacy of OEC/SC cotransplantation combined with treadmill training in spinal cord-contused rats in terms of cell behavior and functional recovery.
Materials and Methods
Subjects and Grouping
A total of 40 adult female Sprague–Dawley (SD) rats were obtained from the Weitonglihua Company (Beijing, China). They were 75 ± 1 days of age and weighed between 220 and 250 g at the time of surgery for induction of SCI. They were assigned randomly from the start of the study to four groups and were later subject to an incomplete SCI in the laboratory. Each group contained 10 rats and received the following treatments, respectively. 1) Cotransplantation of OECs and SCs 2 weeks postinjury (OEC/SC group); 2) treadmill training 1 week postinjury combined with OEC and SC cotransplantation 2 weeks postinjury (OEC/SC training group); 3) treadmill training 1 week postinjury and Dulbecco's modified Eagle's medium (DMEM; vehicle) injection 2 weeks postinjury (training-DMEM group); and 4) DMEM injection 2 weeks postinjury (DMEM group). Hindlimb movement observation was conducted after the injury. Spinal cords were removed at the end of the planned observation, and immunostaining was performed. All procedures were carried out with researchers blinded to the subject groups.
All rats were individually housed in light- and temperature-controlled rooms and given ad libitum access to food and water. All experiments were performed according to the National Institute of Health standards and the guidelines of Beijing Army General Hospital for management of laboratory animals. This study was approved by the Institutional Ethics Committee of the Beijing Army General Hospital for the management of laboratory animals. We certified that all applicable institutional and governmental regulations concerning the ethical use of animals were followed during the course of this research.
Cell Culture
OECs. OECs were cultured from the olfactory bulbs of 7-day-old transgenic SD rats expressing green fluorescence protein (GFP) obtained from Dr. Hongyun Huang at Beijing Hongtianji Neuroscience Academy, China. Briefly, olfactory bulbs were removed and separated from the meninges. The tissues were dissociated and incubated in DMEM/nutrient mixture F12 (DMEM/F12, Gibco, Gaithersburg, MD, USA) with 0.25% trypsin (Worthington Biochemical, Lake Wood, NJ, USA) at 37°C for 60 min. Trypsinization was stopped by adding DMEM/F12 (1:1 mixture) supplemented with 10% fetal bovine serum (FBS, Gibco-Invitrogen, Carlsbad, CA, USA). The dissociated cells were washed thoroughly, resuspended, and then cultured with DMEM/F12 (1:1 mixture) supplemented with 10% FBS for 7 days. To characterize OECs, the cultured cells were fixed with 4% paraformaldehyde (PFA, Sinopharm Chemical Reagent Beijing Co., Ltd., Beijing, China) in phosphate buffer (PB, Sinopharm Chemical Reagent Beijing Co., Ltd.) and then treated with primary antibody of mouse anti-p75 (1:500, Santa Cruz Biotechology Inc., Santa Cruz, CA, USA) and secondary antibody of goat anti-mouse Alexa Fluor 546 (1:400, Molecular Probes Inc., Eugene, OR, USA) for fluorescence immunostaining. The cytoplasm of positive cells was stained red. Most of the cultured cells were expressing p75 and therefore identified as OECs.
SCs. SCs were obtained from the sciatic nerves of 7-day-old SD rats. The sciatic nerves were removed, separated from outer fibers, minced, and incubated in DMEM supplemented with 10% FBS. Two weeks later, after outward migration of fibroblasts, the explants were transferred to new dishes, separated enzymatically, and then incubated in DMEM/10% FBS supplemented with three mitogens: bovine pituitary extract (2 mg/ml; Invitrogen), forskolin (0.8 g/ml; Sigma, Ronkonkoma, NY, USA), and heregulin (2.5 nM; Genentech, San Francisco, CA, USA). Cells were grown to confluency and then were identified by immunostaining of S-100 expression with streptavidin–biotin complex/immunoperoxidase (SABC) method. Briefly, the cultured cells were fixed with 4% PFA in PB and then incubated with rabbit anti-S-100 IgG (1:200; Wuhan Boster Biological Technology Ltd., Wuhan, China). Experimental procedures of SABC immunochemistry followed the instructions provided by the reagent company (Wuhan Boster Biological Technology Ltd.). The cytoplasm of positive cells was stained as brown-yellow. Eighty-eight percent to 95% cultured cells expressed S-100 and so were identified as SCs.
Preparation of SCI Model
The contusion lesion was induced by the MASCIS impactor developed at W.M. Keck Center for Collaborative Neuroscience of Rutgers University, NJ, USA. According to Young's method (64), all the rats were anesthetized with intraperitoneal injection of 4% pentobarbital (35 mg kg−1; Sigma), and the spinal cord with the covering dura mater intact was exposed by laminectomy at thoracic vertebra T10. The exposed spinal cord at the T10 level was moderately injured by dropping the 10-g rod of the MASCIS impactor from a height of 25 mm. This was designed to produce only an incomplete injury. After injury, the muscles and skin were sutured (Sinopharm Chemical Reagent Beijing Co., Ltd.) in layers. The rats were allowed to recover in a warmed cage with easy access to water and food. Penicillin (20 mg kg−1, intramuscular; Beijing Shuanglu Medicine Company, Beijing, China) was administered immediately postsurgery and then daily for 7 days. The rats were maintained for a total of 11 weeks after injury.
Cell Transplantation
For transplantation, the cultured cells were harvested by trypsinization with 0.05% trypsin. Before trypsinization, SC nuclei were labeled by adding Hoechst 33342 (Sigma) to the media at 37°C for 10 min. Thirty minutes prior to implantation, GFP-OECs and SCs labeled by Hoechst 33342 were washed thoroughly and respectively harvested with 0.05% trypsin and 0.02% EDTA (Invitrogen), centrifuged, washed again, resuspended, and counted. After counting, OECs and SCs were resuspended in DMEM at a density of 105 cells/μl and were kept on ice.
Two weeks postinjury, the rats were reanesthetized with 4% pentobarbital, and the lesion site was reexposed. In the OEC/SC group and OEC/SC training group, 4 × 105 OECs in 4 μl of DMEM were injected into the midline at 1 mm rostral and caudal to the epicenter of the contused area at four depths (1.75, 1.25, 1, and 0.5 mm; 0.5 μl per injection depth), and 2 × 105 SCs in 2 μl of DMEM were injected into the lesion site. Injections were made using a 5-μl glass Hamilton syringe (Hamilton, Reno, N V, USA) with a pulled glass pipette, held in a micromanipulator with a microinjector at a rate of 2 μl/min (Quintessential Stereotaxic Injector, Stoelting Co., Wood Dale, IL, USA). The pipette was kept in place for an additional 5 min to minimize leakage upon withdrawal. After the injections, the muscle and the skin were sutured separately. The rats received postoperative care as described above. The treadmill training group and DMEM group received injections of DMEM in the same dosage at the same injection sites.
Treadmill Training
Rats in the treadmill training groups performed bipedal locomotion on a treadmill (Simplex II, Hangzhou Medical Instruments Company, Hangzhou, China). The rats were held by a harness around the belly and suspended over the treadmill in a semierect position to allow the hindlimbs to step on the treadmill. As each rat individually exercised, we moved them manually to keep the toes extended, footpad contacted on the treadmill during the stance phase, and interlimbs coordinated. Training began postinjury day 7 for the treadmill training groups, and the OEC/SC training group lasted for 20 ± 10 min per day, 5 days per week, for 10 weeks. Treadmill speeds were initially set at 3 m/min and increased on a daily basis up to 11 ± 13 m/min according to the tolerance of each rat.
Assessment of Functional Recovery From Spinal Cord Contusion
Functional recovery of 10 rats in each of the four groups was evaluated weekly from 1 to 11 weeks postinjury by Basso, Beattie, and Bresnahan (BBB) score (7) on a scale of 0 (complete paralysis) to 21 (normal mobility). The locomotor activities of the hindlimbs, trunk, and tail were observed weekly in an open field for 4 min. Observers were not aware of the treatment received by each rat, thus ensuring blind scoring.
Histology
On postinjury day 77, the rats were deeply anesthetized (45 mg kg−1 pentobarbital) and transcardially perfused with 100 ml of cold (4°C) 0.9% NaCl (Sinopharm Chemical Reagent Beijing Co., Ltd.) and then 200 ml of 4% PFA (0.1 M, pH 7.4). The spinal cords were removed and postfixed overnight in 4% PFA at 4°C and then transferred to phosphate-buffered 30% sucrose (Sinopharm Chemical Reagent Beijing Co., Ltd.) for 48 h at 4°C for tissue cryoprotection. Segments of the T8–T11 thoracic spinal cord (15 mm long) and segments of the L2–L3 lumbar spinal cord (5 mm long) of five rats from each group were dissected, removed, and embedded in tissue Tec OCT (Leica, Wetzlar, Germany). The long segments contained the entire lesion site. The embedded spinal cords of 15 and 5 mm were coronally and horizontally sectioned, respectively, at 15 μm on a freezing microtome (Leica) and mounted on a series of eight slides (Sinopharm Chemical Reagent Beijing Co., Ltd., China) so that each slide contained every eighth section. Six sections were mounted per slide and stored at 4°C until further processing.
Immunohistochemistry
Immunohistochemistry was performed for fluorescent microscopy as follows. Sections were washed three times with phosphate-buffered saline (PBS; Sinopharm Chemical Reagent Beijing Co., Ltd.) and incubated with 0.1% Triton X-100 (Sigma) with 1% normal goat serum (Invitrogen) in 0.1 M phosphate buffer, pH 7.4, for 30 min at room temperature. Incubations with the primary antibodies were performed overnight at 4°C. After repeated washing with PBS, sections were incubated with their respective secondary antibodies for 60 min at room temperature, washed three times with PBS, coverslipped, and examined in a confocal microscope (Zeiss 510 Meta, Jena, Germany) or a fluorescence microscope (Olympus BX 51, Tokyo, Japan) for analysis.
Every third, fourth, fifth, and sixth serial coronal section of the T8–T11 thoracic spinal cords was analyzed by fluorescent immunohistochemical staining for glial fibrillary acidic protein (GFAP)/neurofilament (NF), 43-kDa growth-associated protein-43 (GAP-43), myelin basic protein (MBP), and tyrosine hydroxylase (TH), respectively. Every fourth horizontal section of the L2–L3 lumbar spinal cord was stained with anti-TH. Primary antibodies used included either a mix of the polyclonal antibody anti-GFAP (Z0334, 1:400; Dako, Glostrup, Denmark) with one of the following monoclonal antibodies: anti-NF (N0142, 1:200; Sigma), anti-GAP-43 (1:200, Beijing Bofeikang Biotechnology Company, Beijing, China), and anti-MBP (SMI-94, 1:1,000, Sternberger Monoclonals Inc., Lutherville, MD, USA), or with monoclonal antibodies for anti-TH (1:200, Santa Cruz Biotechnology Inc.) only. The fluorescent secondary antibodies used included a mix of Alexa Fluor 546 goat anti-mouse and Alexa Fluor 633 goat anti-rabbit (1:400, Molecular Probes Inc.).
Quantitative Assessment of Immunostaining of TH Expression in the L2–L3 Lumbar Spinal Cord
A frame of rectangular shape of 600 μm2 was superimposed on the gray matter of the spinal cord at four alternate locations and on both the left and right side for each animal. All of the TH positively stained neurons in the rectangular frame were counted automatically by Image-Pro Plus V 6.0 (Media Cybernetics, Silver Spring, MD, USA) under fluorescence microscopy (Olympus BX 51, Tokyo, Japan). The number of TH+ neurons from each of the four rectangular frames per section was recorded and then summed for each section. Six sections per spinal cord were counted, summed, and then averaged for each rat. Five rats were inspected from each group.
Statistical Analysis
One-way ANOVA followed by the least significant difference (LSD) test was used for comparing counts of TH-positive neuron in the L2–L3 lumbar spinal cord. A mixed factorial (repeated measures) ANOVA followed by the Tukey–Kramer test was used for comparison of weekly functional recovery patterns after injury. Differences were accepted to be statistically significant at p < 0.05 compared with controls. Data are represented by mean and standard error.
Results
Cell Migration From the Site of Transplantation
On day 77 postinjury, OECs were found migrating longitudinally and laterally from the injection site to lesion site (Fig. 1A). Most of the OECs moved to the lesion site in both gray and white matter (Fig. 1A) and some along the central canal (Fig. 1B) or pia (Fig. 1C). The OECs exhibited two shapes. One was rounded or flattened astrocyte-like, while the other was a Schwann cell-like morphology. Transplanted SCs remained largely concentrated at the transplanted site and were densely packed together (Fig. 1D, E). The majority of OECs migrating longitudinally were Schwann cell-like OECs that aligned parallel to the longitudinal axis of the spinal cord. The SCs displayed parallel, swirling arrangements and were perpendicular to the longitudinal axis of the spinal cord.
Photomicrographs of OECs and SCs cotransplanted in spinal cord. Many green fluorescent protein-olfactory ensheathing cells (GFP-OECs; green) migrated longitudinally and laterally in gray matter and white matter (A, scale bar: 728 μm) and some by central canal (B, scale bar: 50 μm) or pia (C, scale bar: 50 μm) on day 77 after injury. While OECs migrate from the injection site to lesion site (D, scale bar: 100 μm), many Schwann cell (SC) nuclei stained with Hoechst (blue) were concentrated at the injection site and were still densely packed together on day 77 after injury (E, scale bar: 100 μm).
Axonal Growth Supported by Cotransplanted OECs and SCs
On day 77 postinjury, some NF-immunoreactive (NF+) axons were observed parallel to the OEC processes (Fig. 2). The NF+ axons were generally found in Schwann cell-like OECs, while GFAP was expressed mainly by astrocyte-like OECs (Fig. 2A). Substantial NF+ axons were observed entering SC grafts from the host cord and occupying the lesion site. A few NF+ axons were present along the rim of the SC grafts and followed a regular course along the longitudinal axis of the spinal cord. The majority of NF+ axons possessed swirling arrangements in the epicenter of the lesion site (Fig. 2D). Compared with the stout shape of axons in the SC graft, the NF+ axons in OECs displayed a circular shape and irregular arrangements. We also found that some OECs supported GAP-43+ elongation (Fig. 2B, C, F), and some of the axons supported by OECs coexpressed NF and GAP-43 (Fig. 2C, F).
Photomicrographs of axonal growth support by cotransplanted OECs and SCs. Some neurofilament-immunoreactive (NF+, red, A, B, C; scale bar: 20 μm) and many growth-associated protein-43-positive (GAP-43+; B, red; C, F, blue; scale bar: 20 μm) axons projected parallel to the OEC processes. Substantial NF+ axons entered the SC grafts from the host cord and occupied the lesion site. A few NF+ axons were present on the rim of the SC graft and followed a regular course along the longitudinal axis of the spinal cord, and the majority of NF+ axons formed a swirled arrangement in the epicenter of the lesion site (D, red; scale bar: 50 μm). In the treadmill training-only group, a few NF+ axons within the glial fibrillary acidic protein-positive (GFAP+) tissue around the lesion were observed, and no axon was spared in the center of the lesion site (E, red; scale bar: 75 μm). The GFAP immunoreactivity in the spared tissue surrounding the lesion cavity formed a distinct thick glial scar in the treadmill training-only group (E, blue). The density of GFAP immunoreactivity was reduced, and only a little GFAP-immunoreactive tissue invaded the NF+ axons supported by the grafts in OEC/SC group (D).
In the treadmill training-only group, a few NF+ axons within the GFAP+ tissue around the lesion were observed, while no axons were spared in the center of the lesion site (Fig. 2E). NF+ axons in DMEM control group demonstrated features similar to that in the treadmill training-only group.
Reduction of GFAP Immunoreactivity at Injury Site by Cotransplanted OECs and SCs
At the lesion site, GFAP immunoreactivity was observed in the spared tissue surrounding the lesion cavity, and a distinct thick glial scar was formed in the treadmill training-only group (Fig. 2D), similar to that of the DMEM group. However, in the OEC/SC group (Fig. 2C), the lesion site was filled with grafts of SCs and OECs, and no evident glial border between normal tissue and the grafts was found. The density of GFAP immunoreactivity in the spared tissue surrounding the graft of SCs and OECs was less than that in the treadmill training and DMEM group. A few GFAP-immunoreactive tissues invaded the NF+ axons supported by the grafts. When compared with the OEC/SC group, OEC/SC training had no effect on the GFAP immunoreactivity at the lesion site.
Myelination of Axons by Cotransplanted OECs and SCs
On day 77 postinjury, elongation of a few myelinated axons were observed with the OEC transplant (Fig. 3A). In the SC transplant, numerous myelinated axons were observed, rostral and caudal to the grafts. At the lesion site, most MBP+ axons were irregularly packed together in the center of the SC graft, and a few MBP+ axons were present along the rim of the SC graft and followed a regular course along the longitudinal axis of the spinal cord (Fig. 3B).
Photomicrographs of myelination by cotransplanted OECs and SCs. A few myelin-basic protein (MBP)-immunoreactive (MBP+, red) fibers projected along with OECs (A, scale bar: 20 μm). Numerous MBP+ fibers irregularly packed together in the center of SC graft and a few MBP+ axons were present at the rim of the SC graft and followed a regular course along the longitudinal axis of the spinal cord (B, scale bar: 50 μm). In the treadmill training group, a few MBP+ axons within the GFAP+ tissue around the lesion were observed, and no MBP+ axon was spared in the center of the lesion site (C, scale bar: 100 μm).
In the treadmill training-only group, a few MBP+ axons were observed within the GFAP+ tissue around the lesion, and no MBP+ axon was spared in the center of the lesion site (Fig. 3C). MBP+ axons in the DMEM control group demonstrated features similar to that in the treadmill training-only group.
TH Immunostaining
Neither OECs nor SCs supported elongation of TH+ axons. TH-immunoreactive neurons were observed in different distributions. They were mainly located in the Rexed laminae VII and X and in the dorsal horn of the spinal cord in all four groups. The soma of most TH+ neurons were 10–25 μm in diameter, with only a few being located at the dorsal horn margin in the OEC/SC training group (Fig. 4A). In the laminae IX of the ventral horn in the treadmill training-only group (Fig. 4B), the soma size was about 30–45 μm. However, they were also observed at laminae VIII and IX of the ventral horn in the treadmill training group (Fig. 4B, C). In addition, some bouton-like varicosities of TH-immunoreactive fibers surrounding the TH+ neuron were observed in the OEC/SC group (Fig. 4D). The number of TH+ neurons in the OEC/SC training group and the treadmill training-only group is greater than that in the OEC/SC group (p < 0.01, p < 0.05, respectively) and the DMEM group (p < 0.001, p < 0.01, respectively), while no difference was detected between the OEC/SC group and the DMEM group (p > 0.05) (Fig. 5).
Photomicrographs of TH immunostaining. Tyrosine hydroxylase (TH)-immunoreactive neurons with somata diameter of 30–45 μm were observed at the dorsal horn margin in the OEC/SC training group (A) and the laminae IX in the treadmill training-only group (B). TH+ neurons with a somata diameter of 15–25 μm were observed at VII and X in the treadmill training-only group (C), OEC/SC group (D), and Dulbecco's modified Eagle's medium (DMEM; vehicle) group (E). Some bouton-like varicosities of TH-immunoreactive fibers were present surrounding the TH+ neurons in the OEC/SC group (D). Scale bar: 100 μm.
The number of TH-positive neurons in L2 cord of rats in the four treatment groups. ***p < 0.001, OEC/SC training group, training-DMEM group versus spinal cord injury (SCI)-DMEM group. ⋆p < 0.05, ⋆⋆⋆p < 0.001, OEC/SC training group, training-DMEM group versus OEC/SC group.
Functional Recovery From Spinal Cord Contusion
Twelve hours after the SCI, the BBB score of all four groups was 0. The animals of the DMEM group exhibited a characteristic improvement in open-field locomotion that reached a plateau by 6 weeks postinjury. In contrast to the DMEM group, the OEC/SC training and treadmill training group exhibited a significantly greater improvement in BBB score from 4 weeks postinjury, and the OEC group showed a transient improvement in BBB score from 5 to 8 weeks postinjury. In addition, The OEC/SC training group exhibited a significantly greater improvement in BBB score from 9 to 10 weeks postinjury than the OECs and treadmill training-DMEM group, respectively (Fig. 6).
Basso, Beattie, and Bresnahan (BBB) score of rats in the four treatment groups. **p < 0.01, ***p < 0.001, OEC/SC training group versus SCI-DMEM group. #p < 0.05, ##p < 0.01, ###p < 0.001, OEC/SC group versus SCI-DMEM group. ⋆p < 0.05, ⋆⋆p < 0.01, ⋆⋆⋆p < 0.001, treadmill training group versus SCI-DMEM group.
Discussion
This study indicates that cotransplantation of OECs/SCs facilitates axonal growth, migration, and remyelination and reduces the density of GFAP immunoreactivity at the lesion site, while treadmill training increases neural plasticity below the lesion in spinal cord-contused rats. The combined strategy therefore exerts a synergistic effect on functional recovery. Both OECs and SCs showed a higher cell survival rate, but only OECs migrated from the injection site to the lesion site. Two different phenotypic patterns of axonal growth were observed. GAP-43+ and/or NF+ axons extended parallel to the processes of migrating OECs, while NF+ and MBP+ axons of the SCs located in the lesion site consisted of mostly swirling arrangements in the epicenter of the lesion site. This cotransplantation of OECs and SCs strategy led to a transient hindlimb functional recovery from 5 to 8 weeks postinjury. On the other hand, 30 min of daily treadmill training over 10 weeks did not promote axon growth but rather increased neurons in the L2 cord expressing TH. Treadmill training resulted in a more permanent functional recovery from 4 to 11 weeks postinjury.
OECs transplanted rostral and caudal to the injury exhibited migration over a significant distance from the injection site to the lesion site via the gray matter, white matter, central canal, or along the pia as previously reported (54,55). Pearse (49) ascribed the OEC migration in gray and white matter to the leakage of dyes used for cell labeling and dye transfer to migrating host cells in these former studies. They suggested that this could also be the result of the different injury environments of specific types of SCI. However, in this study, mainly the Schwann cell-like OECs rather than the astrocyte-like OECs exhibited significant migration parallel to the longitudinal axis in gray matter or white matter, suggesting that the migration of OECs depends on the nature of the OECs (27), rather than the injection pressure (43), leakage of dyes, or contamination by dyes. OEC support regulates not only the NF+ axons but also the GAP-43+ axonal growth. Some axons supported by OECs concomitantly expressed both NF and GAP-43. The circular shape and irregular arrangement of the majority of NF+ or GAP-43+ axons parallel to migrating OECs along the longitudinal axis of the spinal cord suggested that axons sprouted or regenerated. Since GAP-43 is a marker of axonal regeneration, its presence in growing axon terminals has an important role in axonal growth and neurotransmitter release (8,10). OECs also promoted remyelination in accordance with many other reports of the moderately contused adult rat thoracic spinal cord model (57) and other models (4,30–32,37). This is in contrast to the very low numbers of MBP+ axons that could be myelinated by the small numbers of exogenous or endogenous SCs as reported by Pearse in their subacute contusion model (49).
In line with other well-documented reports (13,63), our results showed that although the transplanted SCs were unable to migrate from the lesion site, they formed grafts bridging the contused cord and supporting substantial axonal growth and remyelination. Only a few of the NF+ and MBP+ axons resided at the rim of the SC graft and followed a regular course along the longitudinal axis of the spinal cord as Pearse previously described (49). Most of the NF+ and MBP+ axons created swirling arrangements in the epicenter of the lesion site. The swirling axons in the epicenter of the lesion seemed to be perpendicular to the longitudinal axis of the spinal cord in alignment with the cells. In contrast to the ring shape and arrangement parallel to the OEC processes in axons supported by OECs, the shape of the axons in SCs indicate that SCs support axonal sprouting. The prolonged axons may be from locally spared host supraspinal axons or regenerating axons induced by the OECs (51,57).
A few tissues with GFAP immunoreactivity were observed infiltrating into the graft of the injury site. The density of GFAP immunoreactivity was decreased, and no border was observed at the injury site in the cotransplantation of OEC/SC group compared to the DMEM group. This is in line with other reports that the OECs inhibit astrocyte proliferation and hypertrophy (36) and give rise to an opening for the GFAP immunoreactivity (1), although SC grafts induced more enhanced GFAP immunoreactivity with a distinct astroglial border between the normal and injured tissues in the normal, contused spinal cord (36).
When compared with the DMEM controls, cotransplantation of OECs/SCs not only supported axon growth and remyelination, and inhibited the GFAP immunoreactivity at the lesion site, but also promoted locomotor recovery at 5–8 weeks postinjury. Neither corticospinal tract fibers nor TH+ axons are supported by OECs and SCs, so the locomotor improvement is probably associated with the propriospinal and brain stem neuron projections such as serotonergic (5-HT; 5-hydroxytryptamine) (49) or noradrenergic (NA) axons (64), as well as peripherally derived NA axons (58). One possible explanation for this finding that the locomotor recovery occurs only at 5–8 weeks postinjury rather than persisting to 11 weeks postinjury is the death of the transplanted cells (49). These discrepancies about OEC migration, axon growth, remyelination, and phenotype of axon in different researches may also be due to differences in the SCI models employed, the cell culture procedures used, or the location of the implant site relative to the injury, transplantation time (49).
That 30 min of daily treadmill training over 10 weeks that displayed a significantly improved hindlimb locomotor function and changed the distribution and number of TH+ neurons in the L2 cord indicates the important role of the plasticity of TH+ interneurons and motoneurons in the L2 spinal cord toward functional recovery. Other reports also show that dopamine β-hydroxylase (DBH) expression was increased in the somata within the intermediate gray matter of spinal segments caudal to the transection in rats with the fourth thoracic cord transected (14), and the voluntary locomotor recovery was related to the plasticity of dopaminergic and serotoninergic systems in spinal cord-transected zebrafish (34). The greater number and wider distribution even at laminae IX of TH neuronal somata after treadmill training may reflect significant upregulation of the TH+ expression in intrinsic interneuronal neurons and a switch by other neurons to a TH phenotype, revealing an interneuron and motoneuron plastic response to treadmill training within the spinal cord. This result is in line with previous reports that step training helps to improve interneuron plasticity and to change motoneuron properties and synaptic inputs (50). The interneuron and motoneuron plasticity induced by treadmill training helps reorganization of sublesional neuronal networks (5,46) and assists in restoring normal patterns of the hindlimb (25). However, except for intrinsic changes below the lesion in the central pattern generator (CPG) and afferent inputs (5) and enhanced neural activity above the lesion site (21), some spared descending fibers from the brain were correlated to recognition of spinal circuitry and locomotor recovery after incomplete SCI.
Based on other reports (14,47,58), we speculate that the neurons with somata of 30–45 μm located in the marginal dorsal horn may be noradrenergic and the other 10–25 μm somata may be dopaminergic. However, the results reported by Mouchet et al. (47) and Takeoka et al. (58) stated that TH-positive dopaminergic interneurons with a somata of 20–25 μm in diameter were located in the cervical and sacral spinal cord, while DBH-positive somata (NA neurons) with a diameter of 30–35 μm were only in cervical levels, and this would differ from Cassam et al.'s results (14) that DBH-positive interneurons were present throughout the spinal cord of intact rats. The discrepancy between their reports may be due to differences in rat strains or even substrains (58). Therefore, noradrenergic and dopaminergic neurons may be intermingled at the gray matter of L2 spinal cord, and treadmill training may increase both noradrenergic and dopaminergic neuronal plasticity.
Similar to the effect of other combined therapy (28,33) on hindlimb functions in spinal cord-injured rats, the application of treadmill training 1 week postinjury combined with cotransplantation of OECs and SCs 2 weeks postinjury demonstrated a synergistic effect on motor function recovery. The results that cotransplantation of OECs/SCs promoted axon growth and remyelination with resultant transient locomotor recovery at 5–8 weeks postinjury, along with the combined treadmill training, led to further significant locomotor recovery and confirmed other reports of combined therapy that repair or formation of new networks is not sufficient to enhance function to a significant degree. Even if repair to the circuitry is optimal, training may be crucial for restoring a substantial level of motor function (28), and training-induced changes require either some spared descending fibers or need to be combined with other approaches (35). The above evidence suggests that the mechanism of this synergistic effect is probably associated with 1) the repair of damaged circuitries or recognition of spinal circuitry derived from both axon growth and remyelination promoted by OECs and SCs and 2) the enhancement of the plasticity of repaired or recognized circuitries induced by treadmill training. In contrast, another study suggests that the application of locomotor training, which enhances task-dependent plasticity, might have “replaced” the lost descending pathways, while antineurite outgrowth inhibitor (Nogo-A) antibodies increased regrowth and sprouting of descending fiber-promoting axon regeneration, but in combination, they did not lead to a synergistic, but rather an interference, effect (44). These discrepancies in combination strategies may be derived from the differences in the models, treatment measures, and the time of intervention in the various studies cited.
In summary, our study indicates that cotransplantation of OECs/ SCs combined with treadmill training has a synergistic effect on motor function recovery of spinal cord-contused rats.
Footnotes
Acknowledgments
This study was supported by Natural Science Foundations of China (grants 30872604 and 81171862) and Science Foundations of Beijing Hongtianji Neuroscience Academy (grant 2007-101). We would especially like to thank Dr. Dajue Wang (NSIC Stoke Mandeville Hospital, UK) for his kind support and help in preparing the manuscript. The authors declare no conflicts of interest.