Synergistic actions of olomoucine and bone morphogenetic protein-4 in axonal repair after acute spinal cord contusion.

To determine whether olomoucine acts synergistically with bone morphogenetic protein-4 in the treatment of spinal cord injury, we established a rat model of acute spinal cord contusion by impacting the spinal cord at the T8 vertebra. We injected a suspension of astrocytes derived from glial-restricted precursor cells exposed to bone morphogenetic protein-4 (GDAs(BMP)) into the spinal cord around the site of the injury, and/or olomoucine intraperitoneally. Olomoucine effectively inhibited astrocyte proliferation and the formation of scar tissue at the injury site, but did not prevent proliferation of GDAs(BMP) or inhibit their effects in reducing the spinal cord lesion cavity. Furthermore, while GDAs(BMP) and olomoucine independently resulted in small improvements in locomotor function in injured rats, combined administration of both treatments had a significantly greater effect on the restoration of motor function. These data indicate that the combined use of olomoucine and GDAs(BMP) creates a better environment for nerve regeneration than the use of either treatment alone, and contributes to spinal cord repair after injury.

Introduction

Following spinal cord injury, axons cannot regenerate effectively (Fry, 2001) owing to an inhibitory internal microenvironment (Tator, 1995; Wang et al., 2002; Tang et al., 2003; Kuerten et al., 2011). Numerous studies have been carried out with the aim of facilitating axonal regeneration following spinal cord injury (Akpek et al., 1999; Sharp et al., 2010; Park et al., 2011; Bazley et al., 2012; Han et al., 2012; Hill et al., 2012; Ozdemir et al., 2012; Wiliams and Bunge, 2012; Yazdani et al., 2012; Fan et al., 2013; Piao et al., 2013), with little success. Axonal regeneration is a complicated process controlled by various factors (Davies et al., 1997; Akpek et al., 1999; Bramlett and Dietrich, 2007). In normal spinal cord tissue, mature astrocytes maintain the morphology of neurons and provide vital nerve growth factors (Giménez y Ribotta et al., 2001). However, excessive proliferation of glial cells after spinal cord injury leads to the formation of glial scars (Shechter et al., 2011; Parry and Engh, 2012), which secrete harmful factors, such as chondroitin sulfate proteoglycans. Furthermore, after spinal cord injury, a cavity is formed (Kakulas, 1987; Tator, 1995). Glial fibers surrounding the cavity begin to grow in a disorderly manner (Davies et al., 2006). Together, the cavity and dense, disordered glia constitute a physical and chemical barrier to axonal regeneration (Hobohm et al., 2005; Wu et al., 2012). Therefore, to repair spinal cord damage, combination therapy is necessary to suppress glial cell proliferation and reduce the size of the cavity.

Olomoucine (2-(2-hydroxyethylamino)-6-benzylamino-9-methylpurine) is a purine derivative that competitively inhibits the activity of cyclin-dependent kinase (CDK) 1, CDK2 and CDK5 in vitro. The molecule binds to the enzymes’ adenosine triphosphate binding sites and inhibits interactions between CDK1 and cyclin B, CDK2 and cyclin E or A, as well as the activity of extracellular signal-regulated kinase 1 or mitogen-activated protein (Glab et al., 1994; Morgan, 1995). Inhibition of CDK interrupts the cell cycle at the G1/S and G2/M checkpoints, and blocks DNA synthesis and cell division (Fan et al., 1999; Arris et al., 2000). Olomoucine inhibits glial proliferation and glial scar formation in the spinal cord after injury, and reduces inhibitory factors; thus, it may improve the microenvironment of the injury zone for axonal regeneration (Tian et al., 2006).

Glial-restricted precursor cells are the precursors of oligodendrocytes and astrocytes. When induced by bone morphogenetic protein-4, they develop into a novel population of glial-restricted precursor cells-derived astrocytes (GDAsBMP) with unique properties (Gregori et al., 2002; Liu et al., 2002; Phillips et al., 2012). When transplanted into injured spinal cord, GDAsBMP fill the injury cavity, moderately inhibit glial scarring, realign host tissues, and delay the expression of inhibitory proteoglycans (Davies et al., 2006). Furthermore, GDAsBMP can also significantly reduce axonal atrophy in the central nervous system (Davies et al., 2006). Together, this evidence suggests that GDAsBMP may be highly effective in the repair of the damaged spinal cord (Davies et al., 2006).

Therefore, in the present study, we investigated the separate and combined effects of olomoucine and GDAsBMP in rats with T8 contused spinal cord, using behavioral and histochemical techniques.

Materials and Methods

Animals

A total of 72 healthy adult female specific-pathogen-free Sprague-Dawley rats, aged 6 weeks and weighing 200 ± 10 g, were provided by the Experimental Center of the Academy of Military Medical Sciences in China (License No. SCXK (army) 2007-004). All rats were housed at 23°C and relative humidity of 50%. The protocols were approved by the Rehabilitation Medical Ethics Committee, Capital Medical University, China. All rats were randomly divided into four groups: model (control), olomoucine, GDAsBMP, and olomoucine + GDAsBMP.

Preparation of spinal cord contusion model

Animals were anesthetized with an intraperitoneal injection of 1% sodium pentobarbital (35 mg/kg). Approximately 5 mm of spine was exposed at T8, and spinal cord contusion was performed using the New York University impactor device, with a weight of 10 g dropped onto the spinal cord from a height of 50 mm. The contusion model was considered successful when the hind legs and tail twitched spastically several times until flaccid paralysis occurred (Anderson et al., 2004) ().

Figure 1

Equipment and method for establishing a model of spinal cord injury.

Spinal cord contusion at T8 was carried out by dropping a weight of 10 g onto the spinal cord from a height of 50 mm.

Cell and drug treatment

GDAsBMP were transplanted into the injured spinal cord in the GDAsBMP and olomoucine + GDAsBMP groups. GDAsBMP cell suspension (30,000/μL; Sigma-Aldrich, St. Louis, MO, USA) was injected into the T8 segment of the spinal cord immediately after injury, using a micro-syringe fixed on a stereotaxic frame (Zhenghua Biological Equipment Co., Ltd., Huaibei, Anhui Province, China). Twelve injection sites were used, each receiving 0.5 μL: two points 0.5 mm either side of the posterior median sulcus, placed at three levels along the spinal cord (at the center of the injury, 1 mm rostral to the injury, and 1 mm caudal to the injury), and each at two depths (0.5 mm and 1.5 mm) (). The needle was inserted first to 1.5 mm depth; the cells were slowly injected, and the needle was maintained in place for 1 minute. The needle was then pulled out to 0.5 mm depth, and maintained in place for 10 minutes, before being removed completely. Finally, the wound was sutured. In the model and olomoucine groups, wounds were sutured immediately after the model had been established.

Figure 2

Cell transplantation in rats after T8 spinal cord contusion.

Glial-restricted precursor-derived astrocytes induced by bone morphogenetic protein-4 (GDAsBMP) were transplanted at 12 injection sites. Dorsal view: The six entry points (green) for cell transplantation. Sagittal view: the three rostrocaudal levels and two depths at which cells were injected (green).

Thirty minutes and 24 hours after injury, rats in the olomoucine and olomoucine + GDAsBMP groups were injected intraperitoneally with 4 mg/kg olomoucine (Alexis Biochemicals, San Diego, CA, USA) in physiological saline (80 mg/mL).

All rats were injected with penicillin (50,000 U/kg) for 3 days after surgery, and manual expression of the bladder was performed until autonomic micturition reflex appeared. Additionally, cyclosporin A (10 mg/kg intramuscularly) was administered to rats that had received GDAsBMP.

Evaluation of behavior using the Basso Beattie Bresnahan Locomotor Rating Scale

The behavior of rats in each group was observed at 3, 7, 14, 21 and 28 days after surgery. The Basso Beattie Bresnahan scale was used to evaluate hind limb motor function (Basso et al., 1995). Nine indexes were observed: body movement; trunk and abdomen position; trunk stability; front and hind limb coordination; gait; paw action; paw rotation; toe position; tail position. The hind limb motor functions of the rats were divided into 22 levels, in which a score of 0 indicated panplegia and a score of 21 indicated normal movement. Because a full bladder may affect locomotor activity, artificial micturition was carried out 1 hour before evaluation.

Anterograde tracing of motor and sensory axons with biotinylated dextran amine

Biotinylated dextran amine motor tracing was conducted 2 weeks before the rats were sacrificed. Rats were anesthetized with 1% pentobarbital sodium (35 mg/kg intraperitoneally) and secured on a stereotaxic frame. Bregma and the sagittal suture were used as reference points (Ersahin et al., 2011). Holes were drilled into the skull to allow 16 cortical injection points: 1.5 and 2.5 mm lateral to the sagittal suture, bilaterally, at four rostrocaudal levels (1 mm rostral to bregma; bregma; 1 mm caudal to bregma; 2 mm caudal to bregma) (). A 10% solution of biotinylated dextran amine -10,000 (Molecular Probes, Eugene, OR, USA) was injected slowly into the cortex (0.5 μL per injection point; 0.1 μL/min) at a depth of 1 mm from the brain surface. The total volume of biotinylated dextran amine injected was 8 μL ().

Figure 3

Anterograde axon tracing for biotinylated dextran amine (BDA).

(A) Anterograde motor axon tracing: 16 injection sites (green) in the motor cortex. (B) Anterograde sensory axon trace: two injection sites (green) on the dorsal spinal cord approximately 5 mm caudal to the injured area.

Biotinylated dextran amine sensory tracing was conducted 1 week before the rats were sacrificed. Animals were anesthetized as described above, and secured in the stereotaxic frame. The spinal dura mater was exposed at the T10 vertebra. Biotinylated dextran amine solution (10%; 0.5 μL) was slowly injected (0.1 μL/min) into the spinal cord at two points: approximately 5 mm away from the caudal margin of the injury area, bilaterally (0.5 mm either side of the posterior median sulcus), at a depth of 0.5 mm (). The needle was maintained in place for 10 minutes before being withdrawn slowly.

Double-labeling immunohistochemistry

At 3, 7 and 28 days after surgery, six rats from each group were deeply anesthetized and perfused transcardially with 4% paraformaldehyde solution. The injured section of the spinal cord, 15 mm in length, was removed and fixed in 4% paraformaldehyde solution for 6 hours, before immersion in 30% sucrose/PBS overnight at 4°C. The tissue was embedded in optimal cutting temperature tissue freezing medium (Sakura, Tokyo, Japan) and serial sections (50-μm thick) were cut using a freezing microtome. According to the direction of the biotinylated dextran amine trace, the center of the injury was designated “0”, and sections were taken at −1.5, −1, 0, +1, +1.5 and +5 mm (where “−” represents the proximal end of the biotinylated dextran amine injection and “+” represents the distal end). Antigen retrieval was carried out by incubating the sections in sodium citrate buffer with 0.1% Triton X-100 for 10 minutes. All samples were then blocked with 5% standard goat serum for 30 minutes at room temperature, and incubated with rabbit anti-glial fibrillary acidic protein polyclonal antibody (1:100; Sigma-Aldrich) overnight at 4°C, followed by Avidin-Cy3 (1:50; Sigma-Aldrich) and goat anti-rabbit IgG-FITC (1:50; Zymed Laboratories Co., San Diego, CA, USA) for 2–3 hours at room temperature, and washed in PBS with 0.1% Tween 20. Tissue sections were placed under a fluorescence microscope (DMLA 4000B and TCS SP5 confocal; Leica, Wetzlar, Germany) to observe biotinylated dextran amine expression (Avidin-Cy3-labeled, red) and glial fibrillary acidic protein (IgG-FITC-labeled, green). Image-Pro Plus 6.0 software (Media Cybernetics, Rockville, MD, USA) was used to calculate integrated optical density values expressed by glial fibrillary acidic protein at 28 days after spinal cord contusion. A high integrated optical density value represents a severe degree of glial scarring. Cavity area and the angle between glial fibrillary acidic protein-immunopositive fibers in the injury region were also calculated using Image-Pro Plus. The angle represents the linear arrangement of fibers, with a small angle indicating good linearity for axonal growth and transport (Davies et al., 2006).

Statistical analysis

All data were expressed as the mean ± SD, and analyzed using SPSS 16.0 software (SPSS, Chicago, IL, USA). Intergroup comparisons were performed using one-way analysis of variance, and t-tests were used for pairwise comparisons. P < 0.05 was considered statistically significant.

Results

GDAsBMP reduced lesion cavity size after spinal cord contusion

Double-immunohistochemistry revealed that a small cavity appeared shortly after spinal cord injury, and grew gradually, especially in the first 3–7 days. In the olomoucine group, the size of the cavity was not significantly different from that in the model group at any time point. However, in the GDAsBMP transplantation group, no cavity was observed at 3 days after surgery, and at 7 days the cavity size was notably smaller than in the model and olomoucine groups (P < 0.01). The difference remained significant until the end of the experiment (28 days; P < 0.05). This suggests that GDAsBMP prevent cavity formation in the early stages of spinal cord injury. In the olomoucine + GDAsBMP group, the size of the cavity was not significantly different from that in the GDAsBMP group, indicating that olomoucine does not influence the inhibitory action of GDAsBMP on the cavity ().

Figure 4

Astrocytes (AS) in rats with T8 contused spinal cord after injury (immunohistochemistry) and effect of glial-restricted precursor-derived astrocytes induced by bone morphogenetic protein-4 (GDAsBMP) and/or olomoucine on cavity formation.

(A) Model group: Large cavity (arrow) and dense scars (arrowhead); (B) olomoucine group: cavity size similar to that of the model group, but less glial scarring; (C) GDAsBMP group: smaller cavity than in model rats but dense surrounding scar; (D) olomoucine + GDAsBMP group: small cavity and little scarring; (E) longitudinal section of model group: dense scarring and disorganized fibers. Glial fibrillary acidic protein (GFAP), astrocyte marker; positive expression is labeled by IgG-FITC (green). Scale bars: (A–D) 2 mm; (E) 125 μm. (F) Effect of GDAsBMP and olomoucine on cavity area in the damage zone of rats with spinal cord contusion. Data are expressed as the mean ± SD; n = 6 rats per group. One-way analysis of variance was used for intergroup comparison; t-tests were used for pairwise comparisons. *P < 0.05, **P < 0.01, vs. model group; †P < 0.05, ††P < 0.01, vs. olomoucine group.

Olomoucine inhibited glial scar formation after spinal cord contusion

Double immunofluorescence did not reveal any significant differences in glial fibrillary acidic protein expression between the model and GDAsBMP groups (P > 0.05), suggesting that the application of GDAsBMP has no effect on glial scar formation after spinal cord injury. However, glial fibrillary acidic protein expression in both olomoucine-treated groups was significantly lower than in the model and GDAsBMP groups (P < 0.05), with no difference between the olomoucine and olomoucine + GDAsBMP groups (P > 0.05). This suggests that olomoucine effectively prevents astrocyte activation and gliosis, but that this effect is not strengthened by GDAsBMP (Figures ).

Effects of glial-restricted precursor-derived astrocytes induced by bone morphogenetic protein-4 (GDAsBMP) and olomoucine on glial fibrillary acidic protein (GFAP) expression in the damage zone of the spinal cord in rats after spinal cord contusion.

Data are expressed as the mean ± SD; n = 6 rats per group. One-way analysis of variance was used for intergroup comparison; t-tests were used for pairwise comparisons. *P < 0.05, vs. model group; #P < 0.05, vs. GDAsBMP group.

Effects of GDAsBMP and/or olomoucine on glial fibrillary acidic protein-immunopositive fibers and biotinylated dextran amine-positive axons after spinal cord contusion

In the model group, 28 days after spinal cord injury, astrocytes in the damage zone had proliferated. Glial fibrillary acidic protein-immunopositive fibers were thick and branched, typical of compact scar tissue. However, after olomoucine intervention, gliosis was inhibited, and glial fibrillary acidic protein-immunopositive fibers were relatively sparse and thin, although the angles between them were still large and the fibers formed a dense mass. Morphological differences in biotinylated dextran amine-positive axons were observed between the two groups at the proximal end of the injury and included thinner or thicker axons, formation of neuromas, branching and sprouting, and bending or winding growth. A few biotinylated dextran amine-positive axons were seen at the center of the injury site ().

Figure 6

Longitudinal sections showing effect of GDAsBMP and olomoucine treatment on GFAP-immunopositive fibers and BDA-positive axons in rats with T8 contused spinal cord 28 days (d) after injury (double-fluorescence immunohistochemistry).

(A) Model group: Disorganized fibers and dense scar tissue in the damage zone; BDA-positive axons were curved and winding at the injury margin (arrow). (B) Olomoucine group: Light scarring but large cavity; BDA-positive axons thick and clustered in some parts, forming a neuroma (arrow). (C) GDAsBMP group: Dense scarring at the injury border; some linearly-arranged BDA-positive axons (arrow) passing through cavity. (D) Olomoucine + GDAsBMP group: GFAP-immunopositive positive fibers and BDA-positive fibers in linear array with high correlation. GFAP (astrocyte marker) labeled with FITC (green). BDA (axonal regeneration marker) labeled with Avidin-Cy3 (red). GDAsBMP: Glial-restricted precursor-derived astrocytes induced by bone morphogenetic protein-4; GFAP: glial fibrillary acidic protein; BDA: biotinylated dextran amine.

In the GDAsBMP group, 28 days after injury, compact scar tissue was visible in the damage zone, accompanied by some small regions where glial fibrillary acidic protein-positive fibers lay along the vertical axis of the spinal cord. These fibers had obvious linear characteristics and kept in line with the direction of the regenerated biotinylated dextran amine-positive neural fibers. The mean angle between glial fibrillary acidic protein-immunopositive fibers in this region was 29.66°, notably smaller than the 62.34° of reticular fibers (P < 0.01; ). However, the linear characteristics of most of the fibers in this region were lost at the scar border, so that the majority of accompanying biotinylated dextran amine-positive axons lost continuity. In the olomoucine + GDAsBMP group, linear fibers and sparse reticular fibers coexisted in the damage zone. Glial fibrillary acidic protein-immunopositive fibers were more linear at the border, and biotinylated dextran amine-positive fibers passed through the cavity.

Figure 7

Effect of GDAsBMP and olomoucine treatment on angles between GFAP-immunopositive fibers in the damage zone in rats with T8 contused spinal cord at 28 days after injury.

The angles between GFAP-immunopositive fibers in the damage zone represents the linearity among the fibers. The smaller the angle, the more linear the arrangement. A linear arrangement is beneficial for axonal growth across the lesion cavity. Data are expressed as the mean ± SD; n = 6 rats per group. One-way analysis of variance was used for intergroup comparison; t-tests were used for pairwise comparisons. *P < 0.05, vs. model group; #P < 0.05, vs. GDAsBMP group. GDAsBMP: glial-restricted precursor-derived astrocytes induced by bone morphogenetic protein-4; GFAP: glial fibrillary acidic protein.

Effects of GDAsBMP and/or olomoucine on Basso Beattie Bresnahan scores after spinal cord contusion

One day after injury, the Basso Beattie Bresnahan score in all animals was 0. On days 3 and 7 post-injury, the scores in each group improved considerably. In the GDAsBMP and olomoucine + GDAsBMP groups, Basso Beattie Bresnahan scores were higher than those in the model and olomoucine groups (P < 0.05). By 14 days, the rate of recovery slowed in the model and GDAsBMP groups, so that the difference in Basso Beattie Bresnahan score between the olomoucine and GDAsBMP groups was lost (P > 0.05, day 14); however, Basso Beattie Bresnahan scores remained greater in the olomoucine and GDAsBMP groups than in the model group (P < 0.05), and in the olomoucine + GDAsBMP group than in the olomoucine and GDAsBMP groups (P < 0.05). After day 14, no further changes in Basso Beattie Bresnahan score were observed in any group except the olomoucine + GDAsBMP group, in which Basso Beattie Bresnahan score improved until the end of the 28-day observation period, and remained significantly different from the other groups (P < 0.05; ).

Figure 8

Effects of glial-restricted precursor-derived astrocytes induced by bone morphogenetic protein-4 (GDAsBMP) and olomoucine treatment on behavior in rats with T8 contused spinal cord at 28 days after injury.

Data are expressed as the mean ± SD; n = 18, 12, 6, 6, and 6 at 3, 7, 14, 21, and 28 days, respectively. One-way analysis of variance was used for intergroup comparison; t-tests were used for pairwise comparisons. *P < 0.05, vs. model group; #P < 0.05, vs. GDAsBMP group. BBB: Basso Beattie Bresnahan locomotor score.

Discussion

Following spinal cord injury, the earliest change in histology is bleeding, which may result from the rupture of the thin-walled vessels in the central gray matter (Gao et al., 2013; Liu et al., 2013a; Zhao et al., 2013). Later, along with increased activity of polymorphonuclear leukocytes and gitter cells around the bleeding vessels, the cells release large amounts of inflammatory mediators (Xu and Cheng, 2013; Zhang et al., 2013). The permeability of the vascular wall is increased, causing a secondary reaction. The injury area grows and stimulates glial activation (Huang et al., 2013; Liu et al., 2013b). Glia proliferate widely within the region of damage, and start to constitute the dense scar tissue together with the extracellular matrix (Bunge, 2001; Seki et al., 2002), in which the cell processes become arranged in a grid. These scars form physical barriers that prevent axons from following their original growth path. Additionally, glia secret harmful factors, further impeding axonal growth through the damage zone. This results in bending or winding of the axon terminal, or the formation of neuroma, and a loss of contact with distal processes. For these reasons, prevention of glial scarring is very important in the treatment of spinal cord injury.

Olomoucine regulates the cell cycle of activated glial cells in vitro (Tian et al., 2006). In rats with spinal cord injury, administration of olomoucine inhibits the proliferation of gitter cells, reduces secondary reactions in the spinal cord after injury, and decreases tissue edema and syringomyelia (Tian et al., 2007). We hypothesized that olomoucine would reduce the release of inflammatory mediators, thus impairing astrocyte activation, and regulate the astrocyte cell cycle, block proliferation, and ultimately inhibit the formation of scar tissue. However, we did not observe any notable differences between the olomoucine and model groups in glial fibrillary acidic protein expression, cavity area or Basso Beattie Bresnahan score 3 or 7 days after injury, suggesting that olomoucine has little effect in the early stages of injury.

A possible reason why olomoucine did not interfere in the glial cell cycle in the damaged spinal cord as expected may be because its action was delayed by the complex microenvironment in the damaged spinal cord, causing it to miss the microglial regulation window. Later, the size of the lesion depends on the degree of hemorrhagic necrosis and gliosis. Olomoucine-treated rats had larger cavities than rats in the other groups, because of the drug's gliosis-inhibiting function. However, the degree of inhibition was small, and not statistically significant.

From 14 days post-injury, Basso Beattie Bresnahan scores in the model group stabilized, whereas in the olomoucine group they continued to improve, suggesting that the drug somehow influences axonal repair at 7–14 days. Although immunohistochemistry was not performed at these time points, the results at 28 days show that olomoucine had significantly inhibited the formation of glial scars around the lesion cavity. Glial proliferation occurred in the damaged area from 7 days after injury. At this point, when a large number of astrocytes enter the proliferation cycle, stimulated by inflammatory factors, olomoucine may start to inhibit further activation. Thus, the growth of the glial scar, which secretes harmful factors, is inhibited, allowing more axons to grow through the damaged area and form synapses with the distal region, and facilitating the recovery of neural function. However, the extent of necrosis and size of the lesion cavity were not affected by olomoucine alone; various substances that inhibit production of gitter cells, myelin sheath formation, and axonal regeneration, may impair the restorative effect of olomoucine in spinal cord injury.

The lesion cavity is an important factor impeding neuroregeneration. It appears about 3 days after injury, initiated from coagulative necrosis in the center of the damaged spinal cord (Zhou and Li, 2006). Gitter cells start to proliferate, and digest and clear the necrotic tissues following the secondary reaction. A necrotic cavity is formed, increasing the area of spinal damage. When regenerative axons reach the proximal area of the cavity, they are unable to cross it because of a lack of supportive structures. However, in the present experiment, we observed that when GDAsBMP were transplanted into the injured spinal cord, the cavity area was notably smaller at every time point from 3 days after surgery, compared with the model group. It has been shown in vitro that GDAsBMP migrate to the injury site, proliferate, and fill the cavity (Kobayashi et al., 1997; Schucht et al., 2002). Furthermore, GDAsBMP delay the expression of inhibitory proteoglycans, promote axonal regeneration of the tractus rubrospinalis, and inhibit neuronal atrophy in the red nucleus (Kobayashi et al., 1997; Schucht et al., 2002). They also secrete brain-derived neurotrophic factor, which is neuroprotective and prevents axonal atrophy (Davies et al., 2006). Thus, transplantation of GDAsBMP into the damaged area provides not only a physical channel for axonal growth, but also a biochemical environment conducive to neuronal regeneration.

We also showed here that although GDAsBMP are astrocytes, when induced by bone morphogenetic protein 4, their processes are characterized by distinct linear features during growth and differentiation after injury. That is, these cells grow along the vertical axis of the spinal cord, with small angles between processes and a staggered conformation. Our double-labeling fluorescence study shows that this linear arrangement correlates with the growth of neural fibers. Together, this evidence demonstrates that GDAsBMP can bridge the lesion cavity and shorten the path of axonal growth, enhancing the efficiency of axons passing through the injury interface and thus promoting the repair and growth of nerves.

Previous studies demonstrated that GDAsBMP might be also capable of restricting the glial scar (Hill et al., 2004). However, in our experiment, we could not find evidence to support this. Although the rats that received GDAsBMP showed a trend in behavioral recovery earlier than other groups, no significant changes in Basso Beattie Bresnahan scores occurred after 2 weeks. The reason was that, although GDAsBMP promote the growth of motor and sensory axons into the area around the damage in the early phase after injury, when the axons reach the distal interface, a substantial glial scar has formed in the host tissue. Cell processes of these host astrocytes are disordered and irregular. GDAsBMP lose their linear properties in the border area. As a result, most of the regenerative nerve fibers cannot continue growing across this glial barrier, and axonal terminals twist or form neuromas.

When syringomyelia began to occur 3 days after injury, transplanted GDAsBMP gradually migrated into the corresponding area, and cell proliferation began. This migration filled the cavity effectively, and processes stretched out along the longitudinal axis of the spinal cord, forming the linear channel through which axons could grow. At the same time, GDAsBMP also delayed the expression of inhibitory proteoglycans in the local environment, secreted neuroprotective factors such as brain-derived neurotrophic factor, and stimulated axons growing into the injury center. Although olomoucine prevents proliferation of ordinary astrocytes, GDAsBMP did not lose their characteristics when combined with olomoucine. We observed no differences in the size of the spinal cord cavity between the GDAsBMP and olomoucine + GDAsBMP group at any time point after injury, indicating that the migration, proliferation and differentiation capabilities of GDAsBMP were not affected by the administration of olomoucine. This is probably because of the unique cellular characteristics of the GDAsBMP, and because the action of olomoucine was delayed, as described above. Thus, when olomoucine started to inhibit astrocytes, the GDAsBMP had already filled the injury space.

From 7 days after injury, olomoucine started to regulate and control the glial cell cycle in the damage zone, inhi539-548bit scar tissue formation, and decrease the secretion of harmful factors. However, the proliferation of GDAsBMP was not significantly influenced by olomoucine during this period. The GDAsBMP bridged the injury cavity, and the processes maintained their linear characteristics in the interface of the host tissues to provide axonal growth channels. We have shown in the present study that olomoucine, in combination with GDAsBMP transplantation, notably increases the number of nerve fibers that cross the region of damage and enter the distal interface of host tissues, because olomoucine restricted the formation of glial scar tissue around the cavity. Moreover, hind limb motor function in rats that received olomoucine and GDAsBMP improved consistently throughout the whole observation period, resulting in final Basso Beattie Bresnahan scores that were significantly better than those of rats in the other groups.