OUP user menu

NT-3 expression from engineered olfactory ensheathing glia promotes spinal sparing and regeneration

Marc J. Ruitenberg, Dane B. Levison, Seok Voon Lee, Joost Verhaagen, Alan R. Harvey, Giles W. Plant
DOI: http://dx.doi.org/10.1093/brain/awh424 839-853 First published online: 16 February 2005


Adenoviral (AdV) vectors encoding neurotrophin-3 (AdV-NT-3) or the bacterial marker enzyme β-galactosidase (LacZ gene) were used to transduce olfactory ensheathing glia (OEG) cultures. AdV vector-transduced OEG expressed high levels of recombinant neurotrophin as shown by in situ hybridization and enzyme-linked immunosorbent assay techniques. The biological activity of vector-derived NT-3 was determined in a dorsal root ganglia neurite outgrowth assay. Engineered cell suspensions were then injected into adult Fischer 344 rat spinal cord immediately after unilateral cervical (C4) corticospinal tract (CST) transection. Transplanted animals received a total of 200 000 cells; either non-transduced OEG or OEG transduced with AdV vectors encoding NT-3 or LacZ, respectively. At 3 months after injury, lesion volumes were significantly smaller in all OEG-transplanted rats when compared with control (medium-injected) rats. Anterograde tracing of the lesioned CST projection, originating from the contralateral sensorimotor cortex, showed a significantly greater number of distal CST axons only in OEG-NT-3-transplanted rats. Behavioural analysis was performed on all rats using open field locomotion scoring, and a forelimb reaching task with Eshkol–Wachman movement notation. Analysis of behavioural tests revealed no significant differences in recovery between experimental groups, although movement analysis indicated that possible compensatory mechanisms were occurring after OEG implantation. The results demonstrate that OEG transplantation per se can promote tissue sparing after injury, but, after appropriate genetic modification, these olfactory-derived cells become far more effective in promoting long-distance maintenance/regeneration of lesioned adult CST axons.

  • corticospinal tract
  • gene therapy
  • neurotrophin-3
  • olfactory ensheathing glia
  • regeneration
  • AdV = adenoviral
  • BDA = biotinylated dextran amine
  • BrdU = bromodeoxyuridine
  • CST = corticospinal tract
  • DRG = dorsal root ganglion
  • GFAP = glial fibrillary acidic protein
  • NT-3 = neurotrophin-3
  • OEG = olfactory ensheathing glia


The adult mammalian spinal cord is normally unable to regenerate axons after injury, in part due to the growth-inhibiting properties of the mature CNS. Axonal regrowth within the spinal cord can be induced experimentally to a certain extent by altering the inhibitory microenvironment of the CNS, e.g. via neural cell transplantation such as olfactory-ensheathing glia (OEG), which can be extracted and purified from the olfactory nerve layer of the olfactory bulb (e.g. Ramon-Cueto et al., 1998; Yan et al., 2001; Plant et al., 2003). Previous studies have demonstrated that a proportion of severed spinal cord axons can regrow and elongate in the presence of OEG (Ramon-Cueto and Nieto-Sampedro, 1994; Li et al., 1997, 1998, 2003; Ramon-Cueto et al., 1998, 2000; Lu et al., 2001, 2002; Takami et al., 2002; Keyvan-Fouladi et al., 2003; Plant et al., 2003; Chuah et al., 2004; Ramer et al., 2004).

Primary OEG are known to produce several trophic factors (Ramon-Cueto and Avila, 1998; Boruch et al., 2001; Woodhall et al., 2001) that could contribute to their growth-permissive properties. Trophic factors secreted by OEG include nerve growth factor, neurotrophin-4/5 and neuregulin (Thompson et al., 2000; Boruch et al., 2001). Brain-derived neurotrophic factor is also produced in small amounts, but neurotrophin-3 (NT-3) is not expressed in detectable quantities (Boruch et al., 2001; Woodhall et al., 2001). Importantly, NT-3 has been reported to stimulate the regrowth of important motor pathways after spinal cord injuries, in particular the corticospinal tract (CST), either in combination with other factors or when given as a single injection (Schnell et al., 1994; Grill et al., 1997; Blits et al., 2000). Adult CST neurons express the high affinity receptor for NT-3, TrkC (Giehl and Tetzlaff, 1996), which continues to be expressed after injury (Liebl et al., 2001). Thus, this neurotrophin appears to be a critical factor in CST growth and plasticity.

It was shown recently by us (Ruitenberg et al., 2003) and others (Cao et al., 2004) that genetic engineering of OEG can result in a cell type that is more effective in promoting spinal cord repair. Our previous study focused on injury and regeneration of the rubrospinal tract (Ruitenberg et al., 2003). In the present study, with the specific aim of providing an environment more conducive to CST regrowth after injury, we engineered adult OEG to secrete NT-3 using an ex vivo adenoviral (AdV) vector transduction technique (Ruitenberg et al., 2002a). These modified OEG were then transplanted to the injured rat cervical spinal cord after a unilateral CST lesion. We used the novel approach of visualizing both left and right CST projections using different biotin dextran labels, thus enabling us to analyse and quantitate the interplay of lesioned versus intact descending axonal tracts in control and OEG-transplanted rats. Behavioural analysis was assessed over a 3 month time period using open field locomotion (Basso et al., 1995), pellet retrieval (Thallmair et al., 1998) and movement analysis (Whishaw et al., 1993).

Material and methods


Adult female Fischer 344 rats (120–150 g; Animal Resource Center, Australia) were used for the in vivo study of spinal cord injury and analysis of the behaviour resulting from such an injury. All experimental procedures conformed to National Health and Medical Research Council guidelines (Australia) and were approved by The University of Western Australia Animal Ethics Committee. Rats were maintained at 90% of normal body weight by administering measured amounts of food each day. This was to ensure a constant appetite so that the animals were motivated to perform the food pellet retrieval task (Whishaw and Pellis, 1990). Initially, 29 rats were trained for behaviour analysis for 2 weeks prior to surgery; two were unable to learn the task and were excluded from the experiment. The remaining 27 animals underwent surgery at the end of the pre-training period and were allowed 2 weeks recovery before post-injury behavioural testing. Analysis of forelimb function using the forelimb reaching task was then performed three times a week, continuing for 10 weeks, before the animals were sacrificed for histological analysis.

OEG cultures from adult rat

OEG were extracted from the olfactory bulb nerve layer of adult female Fischer 344 rats (8 weeks old, 200 g; n = 5) as described in Plant et al. (2003). In brief, the olfactory nerve layer was dissected away from the ventral part of the olfactory bulbs, diced into small pieces and incubated with 0.25% trypsin (Gibco-BRL, USA) and 50 µg/ml DNase (Boehringer Mannheim, Germany) in Hank's buffered salt solution (HBSS; Sigma, USA) for 60 min at 37°C. The tissue was then triturated until a single cell suspension was obtained. Cells were harvested by low-speed centrifugation, re-suspended and plated onto poly-l-lysine- (PLL; Sigma) coated dishes. The medium was replaced after 3–4 days and OEG were left for another 4 days before being immunopurified from contaminating cells using the low affinity nerve growth factor receptor p75 antibody (Plant et al., 2002).

AdV vector infection and transgene expression analysis

The conditions for optimal transduction of primary OEG cultures have been described elsewhere (Ruitenberg et al., 2002a). OEG were infected with AdV vectors, encoding either LacZ or NT-3, at a multiplicity of infection (m.o.i.) of 100. The calculated dose of recombinant virus was added to the culture medium and incubated with the cells overnight in a 37°C tissue culture incubator. Next day, virus-containing medium was replaced with fresh medium and the cells left for 48 h to allow initiation of transgene expression. AdV vector-transduced cells were either stained for expression of the LacZ gene using X-gal histochemistry or examined for NT-3 mRNA expression using standard non-radioactive in situ hybridization techniques.

Secretion and biological activity of AdV vector-derived NT-3 was confirmed in vitro prior to transplantation. Conditioned media from non-modified OEG-, AdV-LacZ- or AdV-NT-3- (OEG-LacZ and OEG-NT-3, respectively) transduced cell cultures were analysed for the presence of AdV vector-derived NT-3 using an enzyme-linked immunosorbent assay (ELISA; Ruitenberg et al., 2003). The Emax™ immunoassay system (Promega, USA) was used to determine levels of NT-3 in OEG-conditioned medium according to the manufacturer's instructions. The amount of secreted NT-3 was expressed as ng of NT-3 derived from a transduced culture per day. The biological activity of AdV vector-derived NT-3 was determined by assaying neurite outgrowth from dorsal root ganglion (DRG) explants, similar to the methods of Dijkhuizen et al. (1997). Embryonic day 15 (E15) DRG explants were grown for 48 h in the presence of conditioned medium from OEG cultures infected with AdV-NT-3, AdV-LacZ or no virus, respectively. Neurite outgrowth was visualized with the mouse monoclonal antibody 2H3 against neurofilament (1 : 1000; Developmental Hybridoma Bank, University of Iowa, IA, USA). Photo images were taken using an IX70 inverted microscope (Olympus, Melbourne, Australia) with a Magnafire Digital camera (Optronics, USA).

In vivo analysis of transgene expression was confirmed in three rats at 7 days post-OEG implantation using sagittal sections through the cervical spinal cord that were subjected to standard in situ hybridization using digoxygenin (DIG)-labelled cRNA antisense probe against NT-3 mRNA (Ruitenberg et al., 2003).

In vitro characterization of transduced OEG

The purity of OEG cultures used for transplantation was confirmed by immunocytochemistry for antigenic OEG markers. Glial fibrillary acidic protein (GFAP) and S100 protein were labelled with rabbit anti-cow GFAP (DAKO Corp., USA) or rabbit anti-S100, respectively, and then visualized with Alexa Fluor®-488 conjugated goat anti-rabbit IgG (Vector Laboratories Inc., USA). Immunoreactivity for low-affinity nerve growth factor receptor p75 was visualized with a mouse monoclonal antibody and Cy™3-conjugated goat anti-mouse secondary antibodies (Jackson ImmunoResearch Laboratories Inc., USA). OEG purity was calculated from the ratio of p75-positive cells to the total cell number (nuclear stain; Hoechst 33342 dye) in the culture.

To examine if transgenic NT-3 had a mitogenic effect on adult OEG derived from the olfactory bulb nerve layer, cell proliferation was assayed in vitro by bromodeoxyuridine (BrdU; Roche, Germany) incorporation. Purified OEG were prepared as described above and plated onto PLL-coated coverslips (15 000 cells/12 mm coverslip) in a 1 : 1 mixture of Dulbecco's modified Eagle's medium (DMEM; Sigma) and Ham's F-12 nutrient mix (DF medium), supplemented with 10% fetal bovine serum (DF-10S) plus mitogens pituitary extract (2 µg/ml; Sigma) and forskolin (0.2 µM; Sigma). The following day, the medium was replaced with either serum-free chemically defined (CD) medium (Yan et al., 2001) or fresh, mitogen-containing DF-10S, and the cells were left for 48 h. Twelve OEG cultures growing in CD medium were then transduced with AdV-LacZ or AdV-NT-3 (2 × n = 6; m.o.i. 100) as described above. Additional cultures were grown in triplicate per experimental condition; either in CD medium or in DF-10S plus mitogens with the addition of increasing amounts of recombinant NT-3 (0, 1, 10, 25, 50 and 100 ng/ml; Peprotech Inc., USA). The latter experimental groups were fed daily with new medium containing freshly thawed NT-3. One day after the first exposure to NT-3, all cultures were pulsed with 10 µM BrdU and incubated for 24 h prior to fixation. Uptake of BrdU by dividing cells under the different experimental conditions was detected via immunocytochemistry using a polyclonal sheep anti-BrdU antibody (1 : 1600; Fitzgerald, USA) and monoclonal S100 antibody (1 : 1000; Sigma) to identify OEG. Per experimental condition (n = 3), photo images were taken of five random fields across each coverslip for cell counts. The fraction of proliferating cells was calculated from the ratio of BrdU-positive cells to the total cell number (Hoechst). The whole experiment was repeated twice. A total of just over 72 000 cells were counted in this experiment, with an average number of 824 ± 34 and 3006 ± 75 cells counted per experimental condition in serum-free and mitogen-containing medium, respectively.

The biological activity of recombinant NT-3 (50 ng/ml) was confirmed in culture using E15 DRG explants (prepared as described above). Explants (n = 6) were grown in Neurobasal™ medium (Gibco) for 72 h. Explant cultures were then triple stained with nuclear Hoechst dye and antibodies against pan-neurofilament (1 : 500; Zymed, USA) and S100 protein using standard immunohistochemical techniques.

Spinal cord surgery

The methods for spinal cord surgery and cell suspension injections were described previously (Ruitenberg et al., 2002a,b). Of the 27 operated rats, seven received injections of OEG-NT-3 in a 1 : 1 vehicle mix of DF medium, seven were injected with OEG-LacZ, seven received non-modified OEG and another six were injected with vehicle alone. Two rats did not recover from the surgery and another animal continued to lose weight, reacting adversely to the injury, and consequently was sacrificed during the first postoperative week. Thus 24 animals were used for post-lesion analysis of behaviour. An additional three rats were used to test for NT-3 mRNA expression in OEG transplants in vivo (see above).

For surgery, animals were anaesthetised with halothane [1.5% halothane; Rhone-Poulenc Chemicals Pty Ltd, Australia; (v/v) in 60% N2O/38.5% O2]. A rostral to caudal incision was made in the skin covering the cervical neck region, and the underlying shoulder and back muscles were separated using forceps. Laminectomy of the C4 vertebra was performed and the dura was carefully cut to expose the spinal cord. A specially designed micro-knife was then used to make a 1.5 mm deep medio-lateral incision across the right half of the spinal cord, which aimed to transect the dorsal hemisphere completely including the CST.

A Hamilton syringe (5 µl; Bonaduz, Switzerland) with an attached capillary needle (tip diameter: 60 µm) was used to give each animal two injections (2× 1 µl; each containing 100 000 cells) of OEG-NT-3, OEG-LacZ, non-transduced OEG or medium alone. OEG suspensions for implantation purposes were prepared as originally described in Ruitenberg et al. (2002a). The researcher performing spinal surgery and OEG implantation was unaware of transplant conditions. Injections were made 0.5 mm lateral to the dorso-medial sulcus on the right side of the spinal cord at a depth of 0.8 mm in an attempt to place the OEG within the dorsal CST projection area. The first injection was made 1 mm proximal to the lesion and the second 1 mm distal to the lesion. Shoulder muscles were then sutured together and the skin was closed with Michel surgical clips (Fine Science Tools, Canada). During the first week, animals were given subcutaneous injections of benacillin (150 mg/ml procaine penicillin, 150 mg/ml benzathine penicillin, 20 mg/ml procaine hydrochloride; Troy Laboratories Pty Ltd, Australia) every second day and daily injections (0.3 ml) of the analgesic buprenorphine hydrochloride (0.3 ml, Reckitt and Colman, UK).

Behavioural analysis

BBB open field scoring

Functional performance of impaired fore- to hindlimb coordination was investigated by two independent investigators, blinded to the experimental treatment, using the Basso–Beattie–Bresnahan (BBB) locomotor rating scale (Basso et al., 1995).

Forelimb reaching task and movement analysis

The test apparatus was a clear Perspex test box with an opening in the right side of the anterior wall and an external shelf on which to place food pellets (see Fig. 2). The floor of the box consisted of a metal grating, thus any dropped food pellets could not be retrieved. Dustless precision pellets (45 mg; BioServ, USA) were used for this task, as they were small enough for the rats to grasp easily and appeared an attractive reward. Food pellets were placed 2.5 cm from the inner edge and towards the left side of the shelf to encourage the animals to reach with the right forelimb. Each pellet was immediately replaced with another irrespective of whether it was eaten or dropped. A successful reach was scored if the rat reached through the opening, grabbed the food pellet with its right paw and transferred it to its mouth. Successful reaches with the left forelimb were not scored. An unsuccessful reach was made if the food pellet was dropped inside the test box. If the food pellet was incidentally knocked off the shelf outside the test box, this was not scored and the pellet was replaced. Averaged weekly success ratings, i.e. the number of successful reaches performed in 2 min, were calculated from three independent measurements over a 7 day period.

For the analysis of movements involved in the forelimb reaching task, animals were filmed once a week from three different angles: (i) a dorsal view looking down on the animal; (ii) an anterior view of the whole animal; and (iii) an anterior close-up of the right forelimb and head. Each of the movements involved in forelimb reaching was rated with the Eshkol–Wachman movement notation (Whishaw et al., 1993).

Anterograde CST tracing

Following the conclusion of behavioural experiments, the 24 rats were anaesthetised with halothane (Rhone-Poulenc Chemicals Pty Ltd, Australia) and an incision was made through the skin covering the skull. Two elongated craniotomies were made at 5 mm either side of the midline from −7.1 mm bregma to +2.9 mm lambda. Anterograde tracer biotinylated dextran amine (BDA, 10% in 2 µl of dH2O; Molecular Probes, USA) was injected through these holes in three oblique injections into the sensorimotor cortex in each hemisphere using a Hamilton syringe. BDA conjugated with Texas red (BDA-R) was injected into the right cerebral cortex, ipsilateral to the spinal cord lesion, whereas BDA conjugated with fluorescein (BDA-G) was injected into the left cortex, contralateral to the lesion site.

Perfusion and tissue processing

Ten days after cortical tracer injections, animals were deeply anaesthetised with 0.8 ml of Nembutal (intraperitoneal, pentobarbitone sodium, 60 mg/ml; Rhone Merieux, Australia) and transcardially perfused with 100 ml of heparinized Dulbecco's phosphate-buffered saline wash, followed by 100 ml of 4% paraformaldehyde in Sorenson's buffer, pH 7.4. The head and vertebral column were dissected from each animal and post-fixed for 24 h. The brain and spinal cord were extracted from the skull and vertebra and stored intact in 0.1 M phosphate buffer (PB, pH 7.4). The position of the injury in the spinal cord was measured from the caudal edge of the cerebellum to confirm that all animals were lesioned at the same level. A 2 cm segment was cut from the spinal cord, with the lesion at the mid-point of this segment, and embedded in gelatine. The embedded spinal cords were sagittally sectioned (40 µm thickness) on a CO2-freezing microtome (Polycut, Reichert-Jung, Australia). Consecutive series of sections were transferred to 24-well plates containing 0.1 M PB with 0.01% sodium azide (Sigma), and stored at 4°C until processed for immunohistochemistry.


Twenty-one of the 24 cortically injected animals were successfully labelled with BDA tracer. Out of these 21 animals, one (OEG-NT-3-transplanted) rat was categorized as receiving an incomplete lesion because the relatively large numbers of remaining BDA-G-labelled fibres present in and beyond the damaged area were obviously concentrated in one area and projected in a straight parallel bundle reminiscent of the intact anatomical projection. Furthermore, axon terminations were seen throughout the distal grey matter of the spinal cord segment; this animal was excluded from further histological analysis.

After sectioning, the first complete spinal cord section was assigned the number 1; this being the lateral right-most part of the spinal cord. Approximately 50 sections were collected per spinal cord. Every eighth section on the lesioned side was immunostained for GFAP, visualizing reactive astrocytes in the lesion area. Sections adjacent to these were immunostained for p75, a surface receptor that is persistently expressed on the vast majority of OEG after their transplantation into injured spinal cord (Ruitenberg et al., 2002a). Fluorophore-conjugated secondary antibodies were used to visualize labelled cells. The immunohistochemical procedures for GFAP and p75 staining on lesioned spinal cord tissue were described previously (Ruitenberg et al., 2002a, 2003).

Quantification of BDA-labelled CST axons

All sagittal sections that contained BDA-G-labelled axons were optically sectioned at low magnification (×4 objective) using a BioRad MRC-1000 confocal laser-scanning microscope attached to a Nikon Optiphot-2 microscope (Centre for Microscopy and Microanalysis, UWA). Quantitative comparisons of BDA-G-labelled axons were based on the level of fluorescence in each section. Using a computerized image analysis program (ImageSpace, Molecular Dynamics), background and non-specific staining in each image was excluded by altering the intensity thresholds. The length and density of BDA-G-labelled axons across the lesion were determined by taking measurements at four distances (Fig. 5A and B): 1 mm proximal to the middle of the lesion site (−1 mm), the middle of the lesion (0 mm), 0.5 mm distal to the middle of the lesion (+0.5 mm) and 1 mm distal to the middle (+1.0 mm). At each distance, a measure of fluorescence was obtained (number of pixels within the altered thresholds × average pixel intensity). The initial adjustment of intensity thresholds could not entirely exclude all background signal from each image. To control for this, measurements from three areas of ventral white matter (representative of background staining) were averaged to give a value for background fluorescence, which was then used to correct signal fluorescence at the four distances across the lesion site.

In addition to the quantitative analysis in sagittal sections, additional analysis of signal fluorescence in the lesioned CST projection was performed in transverse sections (40 µm) both at 1 cm proximal and at 1 cm distal from the lesion epicentre. Digital images were collected and analysed using a free form profile measurement tool that defined the labelled and non-labelled areas. The intensity of fluorescence was compared between 1 cm proximal and 1 cm distal in order to obtain a percentage label. Background fluorescence was subtracted from the final value as described earlier.

Assessment of spinal tissue sparing

Every fourth sagittal section from the spinal cord was used to determine the volume of spared tissue (Plant et al., 2003). The analysis was performed in five rats from each of the medium-injected or OEG-, OEG-LacZ- and OEG-NT-3-implanted groups. In each section, the pixel number in the damaged area and the total number of pixels in a 2.5 mm long spinal cord segment, with the lesion epicentre in the middle, was measured. The border of the damaged area was defined as an obvious discontinuity in the density of small cells, i.e. transplanted cells and inflammatory cells, and the absence of healthy looking spinal neurons. This area contained the grafted cells and usually some small cysts. Measurements from each section were summed per rat and corrected for the total number of sections. Spared spinal tissue was considered to be the difference between the number of pixels in the area of damaged tissue (cyst area and degenerate area) and the number of pixels in the whole segment. Similar methods were applied to estimate graft size, using p75-stained sections, as there is a concordance between p75 staining and OEG viability for up to 4 months after transplantation (Ruitenberg et al., 2002a). The relative amount of degenerate tissue could be calculated by subtracting the number of pixels in the p75-positive area from the total pixel number of the damaged area.

Statistical analysis

All statistical analysis was performed on standard statistics computer software (SPSS for Windows, Release 10). Data sets were analysed for significance using one-way and repeated measures ANOVA (analysis of variance) or simple contrast tests. All data were represented as means = SEM except in Fig. 5C, which shows means ± SD to better visualize intra-group variation.


Analysis of transgene expression

Primary p75-purified OEG cultures were efficiently transduced with first generation E1-deleted AdV vectors encoding LacZ or NT-3 (m.o.i. 100; Fig. 1A and B). High levels of transgene expression were seen in vitro; >95% of the cells were labelled 3 days after infection. Transduced cells displayed similar morphology to control OEG cultures, cytotoxicity appeared absent and control cultures showed no expression of either LacZ or NT-3 (not shown).

Fig. 1

In vitro and in vivo analysis of AdV vector-mediated transgene expression. (A) X-Gal cytochemistry of primary OEG, 72 h after transduction with AdV-LacZ vector. Virtually all cultured cells expressed the LacZ gene product. (B) Using a cRNA probe specific for NT-3 mRNA, cells transduced with AdV-NT-3 vector expressed high levels of this neurotrophin after 72 h in vitro. (C) Conditioned medium from AdV-NT-3 vector-transduced OEG cells was added to E15 dorsal root ganglion (DRG) cultures and neurite outgrowth was visually inspected after 48 h. AdV vector-derived NT-3 in OEG conditioned medium was biologically active, as demonstrated by the induction of robust neurite growth from cultured DRG explants. (D) Higher magnification of the outlined area in C showing neurite extensions with growth cone-like structures. (E) Conditioned medium taken from control (OEG-LacZ) cultures induced very little outgrowth from DRG explants. (F) In vivo confirmation of AdV vector-mediated NT-3 expression at 7 days after injection of AdV-NT-3 vector-transduced OEG suspensions into the lesioned spinal cord. Scale bars: A, B = 83 µm; C, E = 1 mm; D = 275 µm; F = 170 µm.

Conditioned medium from OEG cultures was analysed for the presence of NT-3 using standard ELISA, 4 days after AdV vector transduction. AdV-NT-3 vector-transduced cultures produced net NT-3 levels of 40 ± 7.4 ng/105 cells/24 h. NT-3 protein could not be detected in non-transduced and OEG-LacZ cultures.

Biological activity of AdV vector-derived NT-3 was confirmed by analysing neurite outgrowth from embryonic DRG explants in the presence of conditioned medium from OEG cultures. After 2 days, robust neurite outgrowth was observed only from DRG explants grown in supernatant from OEG-NT-3 cultures (Fig. 1C and D). Virtually no neuritic growth was observed from DRG explants that were grown in conditioned medium from either non-transduced OEG or OEG-LacZ cultures (Fig. 1E).

In vivo analysis of transgene expression showed that, apart from endogenous levels, there was robust NT-3 mRNA expression in the damaged area at 7 days post-implantation (Fig. 1F). Numerous labelled cells of bipolar OEG morphology were detected in the transplantation area. No such staining was present in animals that received injections with OEG-LacZ or ‘medium only’. As documented previously (Ruitenberg et al., 2003), focused p75 labelling in adjacent immunoreacted spinal cord sections was co-extensive with regions containing intense NT-3 mRNA expression.

Behavioural analysis

Open field testing

Behavioural analysis of OEG-grafted rats during over-ground locomotion (BBB score) did not reveal detectable changes from control (medium-injected) animals over the analysed period (10 weeks). Gait was frequently to constantly weight supported with plantar steps but there was only occasional coordination between fore- and hindlimbs. A steady rating of 12 was scored for all animals from week 4 onwards.

Movement analysis and pellet retrieval success ratings

Body rotation after surgery was abnormal in most experimental animals as they had difficulties in raising the right forelimb. The head and shoulders would turn in anti-clockwise motion almost 90°, such that the right forelimb was raised in the air. The forepaw was then pronated over the food pellet and the limb retracted back to drag it towards the inner edge of the shelf as illustrated in Fig. 2A. Components such as grasp, supination and release were virtually absent.

Fig. 2

Behavioural assessment of forepaw function in experimental animals. (A) Pre-surgery skilled forepaw movements were normal in experimental rats. All animals were capable of targeted reaching through the opening with their right forepaw; they grabbed the food pellet and, subsequently, transferred it to their mouth. At 3 weeks after the surgery, pellet retrieval was severely impaired in all experimental animals, i.e. the lesioned forepaw was paralysed and the rats were not capable of making the skilled reaches that are required to grab the food pellet reward. At 12 weeks after surgery, success ratings in pellet retrieval had significantly improved among all experimental groups, but further analysis showed that there was no concurrent improvement in Eshkol–Wachman movement notation for the impaired forepaw. The animals had adapted to partial forepaw paralysis by turning the head and shoulders in an anti-clockwise motion almost 90°, such that the lesioned forelimb was raised in the air. This forepaw was then pronated over the food pellet and the limb retracted back to drag it into the animals mouth as illustrated. Components such as grasp, supination and release were virtually absent. (B) Weekly averages of success ratings, i.e. the number of pellets retrieved per 2 min time interval, for each of the four groups before lesioning and during the testing period. Note the (spontaneous) improvement that was observed over time in all groups. (C) Weekly averages of Eshkol–Wachman movement notation, i.e. rating of skilled forelimb use, for each of the four groups analysed. (D) ‘Success versus movement’ analysis of an individual animal from the medium-injected (left) and OEG-NT-3-implanted (right) animal groups. Note that animal 11 demonstrated a simultaneous improvement of pellet retrieval success and forelimb movement rating from week 7 to 9, whereas no such parallel recovery was observed in animal 18.

Pre-surgery baseline level of success (18–23 pellets eaten in 2 min) was not significantly different (P > 0.05) between the experimental groups. Movement rating re-started 2 weeks following surgery (week 3), when pellet retrieval was markedly impaired across all groups (0–6 pellets eaten in 2 min; see Fig. 2B). All experimental animals were tested until 12 weeks after surgery, at which point success ratings had improved across all groups (9–25 pellets/2 min). Repeated measures ANOVA showed a statistically significant improvement in animals' success ratings over time (P < 0.001), but not between the experimental groups (P = 0.110).

Pre-surgery movement ratings for all experimental animals were scored between 0 and 2, indicating normal forelimb performance. Movement ratings were dramatically impaired in all animals after surgery, indicating abnormalities in forelimb function, which remained relatively unchanged over the test period (11.3–15.2; Fig. 2C). Repeated measures ANOVA showed no statistically significant changes in post-surgery movement rating over time (P = 0.503) nor differences between the experimental groups (P = 0.506).

Pellet retrieval versus movement ratings

Behaviour of all animals, regardless of experimental treatment, centred around a pattern of improvement in pellet retrieval success rating without corresponding restoration of skilled forelimb use (exemplified by animal no. 18; Fig. 2D, left panel). An exception to this pattern was given by OEG-NT-3-transplated rat 11, which demonstrated a concurrent loss of both success and movement ratings over the first 7 weeks and, subsequently, a simultaneous improvement of these parameters from week 7 to 9 (Fig. 2D, right panel). After week 9, however, the movement rating remained relatively constant while pellet retrieval continued to improve.

Histological analysis

Anterograde CST tracing

Both dorso-medial CST projections were labelled with either BDA-G or BDA-R as illustrated (see diagram, Fig. 3A). Normal, uninjured Fischer 344 rats (n = 2) were used to optimize the tracing procedure (Fig. 3B and C). The border between the two projections was very distinct and only very small numbers of axons (<5%) were observed ipsilateral to the injected cortical hemisphere.

Fig. 3

Graphic representation and confocal photomicrographs of dual-coloured anterograde CST in normal and spinal cord-injured animals. (A) Schematic diagram of the tracer injection sites and subsequent labelling of the different sensorimotor cortical hemispheres and CST projections. (B) Representative horizontal spinal cord section showing the two individually labelled CST projections in an uninjured control animal. (C) Transverse section through the cervical (C4) spinal cord, showing the dorsal column region and labelled CST projections. Note the distinct labelling of axons within both CST projections. Only occasional BDA-G-positive fibres were seen in the BDA-R-labelled CST, and vice versa. (D) Representative confocal photomicrograph of an OEG-LacZ transplanted rat, 12 weeks post-surgery, showing successfully lesioned (BDA-G-labelled) CST axons while the BDA-R-labelled CST was still intact. Lesioned CST axons had developed enlarged, bulbous ends, and large fibre masses were seen to collect at the rostral edge of the lesion site. Virtually no regrowth of lesioned CST axons was induced in animals receiving control (OEG and OEG-LacZ) implants. (E) Consecutive section showing GFAP-positive hyperactive astrocytes surrounding the lesion and implantation site. cc = central canal; gm = grey matter; wm = white matter. Scale bars: B = 374 µm; C = 178 µm; D, E = 227 µm.

CST projections in experimental animals were labelled as described above. After processing, ∼50 sagittal sections (40 µm) were obtained per rat, 8–10 of which contained BDA-G-labelled fibres. In all operated rats, CST axons had developed enlarged, dystrophic ends and sprouts just proximal to the lesion site, which caused the BDA-G-labelled CST to appear wider here than in more proximal undamaged parts. Virtually no BDA-G-labelled fibres were seen in the lesion area itself or in distal spinal cord in vehicle-injected controls, or OEG- and OEG-LacZ-implanted rats (Fig. 3D). Immunohistochemistry for GFAP showed the presence of hypertrophic astrocytes within a 200–300 µm radius of the damaged area (Fig. 3E), with lesioned BDA-G-labelled CST axons abutting the proximal border of the glial scar.

In all transplanted animals, the phenotypic appearance and purity of primary OEG preparations, extracted from the olfactory bulbs of adult Fischer 344 rats, were examined prior to transplantation. Cultured OEG were immunoreactive for S100 (not shown), GFAP and p75 (Fig. 4A). The purity of all cultures used for transplantation purposes was between 95 and 98%, as calculated from the ratio of p75-stained cells to the total cell number. After 12 weeks, numerous p75-immunoreactive cells were seen within the damaged area of all OEG-transplanted animals (Fig. 4B). Many fewer cells, most probably host Schwann cells that had entered the injured cord (Ruitenberg et al., 2003), were found in vehicle-injected controls.

Fig. 4

Transplant purity and distal CST axons in OEG-NT-3-transplanted rats. (A) After immunopanning, primary OEG cultures for transplantation purposes were enriched in p75-positive (red) cells and between 95 and 98% pure as determined from the ratio of p75-positive cells and the total number of Hoechst dye-positive (blue) nuclei. Cultured OEG were also immunoreactive for GFAP (green). (B) Representative image of intense p75 immunoreactivity (asterisk) at the lesion and implantation site, 12 weeks after surgery, as observed in OEG-transplanted animals only. (C) Low-power micrograph of a typical OEG-NT-3-transplanted rat at 12 weeks after surgery. Distal BDA-G-labelled CST axons (arrows) extended from the proximal mass (white line), bypassing the lesion area through the more ventral portions, and were detected at least up to 1 cm distally. (D) Higher magnification of the lesion and implantation area, (outlined by the dashed line) in another OEG-NT-3-transplanted rat showing a significant number of BDA-G-labelled CST axons (arrows) entering the ventral portions of the damaged area and extending into the opposing part of the spinal cord. (E) Transverse spinal cord section, 1 cm distal to the C4 lesion site, taken from an OEG-LacZ-transplanted rat at 12 weeks after surgery. Note the presence of BDA-R-positive axons in the intact CST projection (left) whereas BDA-G-labelled axons are absent in the degenerated projection area on the lesioned side of the spinal cord (dashed line), demonstrating the completeness of the injury. (F) Scattered BDA-G-labelled fibres were seen at 1 cm distal to the injury site in OEG-NT-3-transplanted animals only. CST = corticospinal tract; gm = grey matter; wm = white matter. Scale bars: A = 41 µm; B = 91 µm; C = 870 µm; D = 279 µm; E, F = 71 µm.

Strikingly, four out of the five animals analysed receiving OEG-NT-3 implants showed BDA-G-labelled fibres extending from the proximal mass of lesioned CST axons through the lesion site and into more distal parts of the spinal cord as far as at least 1 cm (Fig. 4C). These distal CST axons were scattered in distribution, had a beaded appearance and uneven trajectory (Fig. 4D), features regarded as being characteristic of regenerating CST fibres (Steward et al., 2003).

In the transverse sections, no BDA-G fibres were seen 1 cm distal to the lesion site in any of the vehicle-injected, OEG- or OEG-LacZ-implanted rats (Fig. 4E). Distal BDA-G-labelled CST axons were, however, seen at this level in four of the five OEG-NT-3 transplanted animals (Fig. 4F). On average, ∼9% of proximally labelled fibres were seen at least 1 cm distal to the lesion site (see later), which was the maximum distance measured. These labelled axons were scattered and not obviously concentrated in a localized area of white matter, the latter organization being more indicative of an incomplete CST lesion.

Quantification of dorsal CST axons

Labelled CST fibres in sagittal sections were quantified by measuring signal fluorescence at four distances in a rostral to caudal direction across the lesion site (Fig. 5A and B). These values were then averaged for each experimental group, i.e. vehicle-injected, OEG-, OEG-LacZ- and OEG-NT-3-transplanted rats (Fig. 5C). The fluorescence intensity of CST fibres was very high proximal to the lesion area (−1.0 mm), decreasing sharply in the centre (0 mm) and more distal parts of the lesion (+0.5 and +1.0 mm). There was a trend towards increased numbers of distal CST axons in the OEG-NT-3-implanted animal group, showing consistently higher signal fluorescence in the lesion centre and more caudal parts of the spinal cord than all other experimental groups (Fig. 5C). Simple contrast tests showed that statistical significance in increased fluorescence measures in the OEG-NT-3 group was approximated at both +0.5 mm (P = 0.085) and +1.0 mm (P = 0.066) when compared with its control, the OEG-LacZ group.

Fig. 5

Quantitative analysis of distal CST axons at 12 weeks after injury and OEG implantation. (A) Typical example of a confocal microscope image of an OEG-implanted animal, showing a sagittal section through the spinal cord with the lesion epicentre in the middle. (B) Converted image of A that was used by the computer to quantify fibre fluorescence using image analysis software. The position of the vertical bars signifies the regions of analysis. (C) Measurements of BDA-G-labelled axons at the lesion site, as obtained from sagittal spinal cord sections from all experimental groups and presented as pixels of fluorescence (mean ± SD). There was a trend towards increased numbers of axons in the more distal parts of the lesion and implantation site in OEG-NT-3-implanted animals. (D) Quantification of BDA-G-traced CST axons in transverse spinal cord sections, 1 cm distal to the injury site. The obtained measurements are expressed as the percentage of total fluorescence at 1 cm proximal to the injury site to compensate for variability in efficiency of CST tracing (relative percentage of fluorescence intensity ± SEM). Significantly more distal BDA-G-labelled fibres were detected in OEG-NT-3-implanted rats compared with all other experimental animal groups (P ≤ 0.0005).

In transverse sections, the density of CST fibres within both BDA-G (lesioned) and BDA-R (uninjured) labelled projections at 1 cm distal to the injury was expressed as a percentage of total label (measured 1 cm proximal to the injury). In vehicle-injected control animals, signal fluorescence 1 cm distal to the injury site was reduced to only 0.66 ± 0.2% (mean ± SEM; n = 5) of the proximal signal. Similar values were found for the OEG- (1.5 ± 0.7%; n = 5) and OEG-LacZ- (1.1 ± 0.2%; n = 5) transplanted groups. Strikingly, in OEG-NT-3-transplanted animals, much higher values of ∼9% of the proximal label were seen 1 cm below the injury site (9.1 ± 1.43%; n = 5; Fig. 5D). ANOVA and post hoc analysis revealed a significant overall treatment effect between the experimental groups [F(3,20) = 15.45; P = 0.0001]. Further analysis showed no statistical differences between groups that received vehicle injections, OEG or OEG-LacZ transplants (P > 0.57), but the distal CST label in OEG-NT-3-transplanted animals was significantly greater compared with other groups (P < 0.0005). There was no evidence of progressive changes in BDA-R-labelled fibre densities or compensatory growth into the BDA-G projection and termination areas distal to the injury site.

OEG transplants promote tissue sparing

Spinal tissue sparing was calculated from the damaged area (including cysts, degenerate tissue and transplanted OEG) within a 2.5 mm long spinal cord segment (Fig. 6A). In the vehicle-injected group, 75.2 ± 4.8% (mean ± SEM, n = 5) of the spinal cord segment examined was spared. In contrast, significantly more (P ≤ 0.006) spinal tissue was spared in OEG- (87.7 ± 1.0%; n = 5), OEG-LacZ- (87.5 ± 0.9%; n = 5) and OEG-NT-3- (87.7 ± 0.8%; n = 5) implanted groups. Mean cyst area and number were similar between experimental groups (P > 0.05; Fig. 6B). The amount of degenerate tissue was significantly decreased in all transplanted animals (P < 0.004; Fig. 6C). As expected, all OEG-transplanted animals had greater p75 immunoreactivity in the grafted area than medium-injected control rats (P < 0.0001; Fig. 6D). While there was no significant difference between OEG and OEG-LacZ groups (P = 0.78), the p75-positive area in the OEG-NT-3 group was significantly larger than all other OEG-transplanted groups (P < 0.004). This increase might indicate improved survival and/or proliferation of engineered donor cells (Bianco et al., 2004) and/or perhaps increased host Schwann cell invasion in this group (Ramer et al., 2004).

Fig. 6

OEG implants reduce the spinal cord lesion size and preserve tissue. (A) Statistical analysis showed that the mean percentage of spared spinal tissue was significantly greater in all OEG-transplanted animal groups compared with vehicle-injected controls (P ≤ 0.006). (B) No significant changes in total cyst number and cystic (pixel) area were seen between the different experimental groups. (C) The area of degenerate tissue was significantly smaller in all OEG-implanted animals and even further reduced in the OEG-NT-3-transplanted animal group (P < 0.004) (D). The p75-positive (pixel) area in the grafted region was significantly larger in OEG- and OEG-LacZ-transplanted rats but even further increased following implantation of AdV-NT-3-transduced OEG (P < 0.004). A single asterisk indicates statistical significance from vehicle-injected lesioned control animals, whereas double asterisks signify significance of OEG-NT-3-transplanted rats compared with all other experimental groups.

In an additional series of in vitro experiments, we studied the mitogenic effects of (virally encoded) NT-3 on adult OEG cultures via BrdU incorporation. Cells were grown in either serum-free (CD) or mitogen-containing (DF-10S) medium. We found no evidence that transduction of cells with AdV-NT-3 resulted in increased proliferation of purified adult OEG that were obtained from the olfactory bulb nerve layer. Additional experiments, adding recombinant NT-3 to OEG cultures, also showed no mitogenic effects of this neurotrophin at any concentration (Fig. 7A–C). Further analysis of total cell numbers after plating failed to reveal any effects of NT-3 on OEG viability, at least over the 2 day time period in vitro (not shown). Interestingly, however, DRGs that were grown in the presence of NT-3, to confirm biological activity of this recombinant neurotrophin, showed that in addition to robust neurite outgrowth, there was also increased Schwann cell dispersal from the explants (Fig. 7C and D). The latter observation is consistent with an effect of NT-3 on Schwann cell migration and chemokinesis (Maniwa et al., 2003).

Fig. 7

In vitro analysis of mitogenic effects of NT-3 on purified adult OEG. (A and B) Proliferation of purified adult OEG, obtained from the rat olfactory bulb nerve layer, in DF-10S medium in the absence (A) and presence of NT-3 (B; 50 ng/ml). Cultures are triple-stained with S100 (green), BrdU (red) and Hoechst dye (blue). In proliferating cells, the nuclear co-localization of BrdU signal and Hoechst dye results in a purple nuclear stain. (C) Quantitative analysis of OEG proliferation using cells that were either grown in CD- or DF-10S medium, containing various amounts of NT-3, or transduced with AdV-NT-3. As expected, the percentage of proliferating cells was much lower in medium, containing no serum, pituitary extract and forskolin. No evidence was found that NT-3, either derived from transduced cells or added as recombinant protein to the culture medium, had a mitogenic effect on OEG (P > 0.05). (D and E) Confirmation of the biological activity of recombinant NT-3. E15 DRG explants were (immuno-) stained for pan-neurofilament (green), S100 protein (red) and Hoechst dye (blue nuclear stain). (D) In the absence of NT-3, virtually no neurite extension from DRG explants was observed. (E) As expected, addition of NT-3 (50 ng/ml) to the culture medium induced robust, radial fibre outgrowth from the explant. Note the increased dispersal of Schwann cells in the presence of NT-3. Scale bars: A, B = 50 µm; D, E = 100 µm.


We report the effects of genetically modified OEG transplants on spinal cord tissue sparing and CST sprouting/regeneration after cervical injury in rats. Significantly more spinal cord tissue was spared in all OEG-transplanted animals compared with vehicle-injected controls. Qualitative and quantitative histological analysis of BDA-G-labelled CST axons showed that OEG implants engineered to secrete NT-3 induced long-distance maintenance/regrowth of these axons for up to at least 1 cm distal to the injury site. Behavioural analysis revealed some post-lesion recovery but there were no significant differences between experimental groups. As no changes were seen in the pattern and distribution of uninjured BDA-R-labelled CST axons distal from the lesion site, our results point to compensatory involvement of intraspinal circuitry and/or other descending supraspinal pathways in post-lesion recovery.

OEG-implanted rats had significantly more spared spinal tissue (10–15%) at 12 weeks after injury compared with vehicle-injected controls. This suggests the preservation of white matter tracts and intraspinal neurons in the area neighbouring the primary injury site, which could contribute to functional adaptation after the injury. These findings are consistent with previous reports showing that OEG are capable of promoting spinal cord tissue sparing (Takami et al., 2002; Plant et al., 2003; Ruitenberg et al., 2003; Verdu et al., 2003; Ramer et al., 2004).

In transplanted animals, many of the cells present at the site of engraftment were p75 positive. This population bridged the lesion from the rostral to caudal extents of intact spinal cord, providing a cellular substrate for growing axons towards the opposing end of the spinal cord. Although a proportion of these p75-positive cells may be host (non-myelinating) Schwann cells (Takami et al., 2002; Ruitenberg et al., 2003; Boyd et al., 2004; Ramer et al., 2004), previous work, tracking the fate of green fluorescent protein (GFP)-labelled OEG in similar transection models, showed that grafted OEG survived (Ruitenberg et al., 2002; Sasaki et al., 2004) and made up a significant proportion of these p75-immunoreactive cells (Ruitenberg et al., 2002a).

In both vehicle-injected and OEG-implanted animal groups, the majority of lesioned CST axons formed enlarged bulbous ends along the proximal edge of the lesion. In rats that received vehicle injections or control implants (OEG and OEG-LacZ groups), CST axons did not elongate beyond the injury site, which is in agreement with Takami et al. (2002) but in contrast to data reported by other groups (Li et al., 1998; Ramon-Cueto et al., 2000; Nash et al., 2002; Keyvan-Fouladi et al., 2003). It is difficult to assess if these discrepancies represent true differences in outcomes or whether they are more related to experimental differences in lesion procedures, OEG preparation protocols and/or transplant purity. Only in OEG-NT-3 transplanted rats was there evidence of CST sprouting 0.5 and 1 mm distal to the lesion epicentre; others have described similar effects of NT-3 on CST axons (Schnell et al., 1994; Senut et al., 1995; Grill et al., 1997; Blits et al., 2000).

Significantly more BDA-G-labelled CST fibres were detected up to 1 cm distal to the lesion in the OEG-NT-3 group. These axons had a beaded appearance (Li et al., 1997) and traversed the damaged area through the more ventral portions of the lesion and implantation site. Further distally, the majority of BDA-G-labelled CST axons were arranged in a more scattered, defasciculated pattern with an uneven trajectory along the rostro-caudal axis of the denervated spinal cord. Such features are thought to distinguish axonal regrowth from normal CST organization (Steward et al., 2003). However, morphological criteria that are used to discriminate true axonal regeneration from sparing, in particular in partial spinal cord transection models, remain an important area of debate. In our study, there was a comparable amount of preserved spinal tissue in all OEG transplant paradigms, suggesting that the increased number of distal CST axons in the OEG-NT-3 group was not due to a difference in overall tissue sparing. Regeneration of at least some CST fibres thus seems likely, although it is also possible that NT-3 released from transduced OEG had a selective and specific effect on protecting compromised CST axons from subsequent secondary degeneration.

All experimental animals were tested for skilled forelimb reaching. They showed severe impairment immediately after injury followed by partial recovery. Success ratings in transplanted animals did not change significantly from controls over the period of analysis, which contrasts with the results of Nash et al. (2002), who combined methylprednisolone administration with OEG transplantation. Our results, however, support the original findings of Schrimsher and Reier (1993), who described a recovery in the pellet retrieval task performance over time after a C4 cervical lesion. Movement ratings remained relatively constant over the testing period and did not differ between the experimental groups. This indicates increasing success in the forelimb reaching task over time, but no parallel improvement in animals' movements to reach for the food pellets. Hence, our findings suggest that the improvements in pellet retrieval are due to spinal or supraspinal compensatory mechanisms and learning adaptation, rather than recovery of skilled movements due to anatomical CST regeneration.

Closer inspection of movement ratings (Whishaw et al., 1993) revealed that specific task components were impaired by unilateral CST lesions. These included supination, grasping and release of the food pellet. Impairments of distal forelimb muscle were previously reported following similar cervical lesions of the dorsal columns that include the CST (Schrimsher and Reier, 1993; Whishaw et al., 1993; McKenna and Whishaw, 1999; Raineteau et al., 2001). It is therefore apparent that future studies involving behavioural analysis should consider the exact movements that are affected from damage to specific spinal motor pathways. In the present study, a task that may have been more sensitive to a CST lesion in rats would be one requiring supination, grasping and release of the forepaw (e.g. Galea and Darian-Smith, 1997; Ballermann et al., 2001).

Several mechanisms, other than axonal regeneration, have been suggested to contribute to functional recovery after spinal cord injury (e.g. Houweling et al., 1998; Jakeman et al., 1998; von Meyenburg et al., 1998; Tuszynski et al., 1999; Corbetta et al., 2002; Gomez-Pinilla et al., 2002; Blits et al., 2003; Ruitenberg et al., 2003; Bareyre et al., 2004). Proposed mechanisms of action include functional compensation from related, spared pathways, the stimulation of ‘central pattern generators’ (simple motor programs) within the distal spinal cord, and induced plasticity locally in the spinal cord or higher brain centres. Conceptually, increased plasticity may allow for remaining tissue, surrounding a partial lesion, to take over the functional roles of the damaged motor tracts and thereby lead to recovery of movement. In our study, the rubrospinal tract, which has many functional similarities to the CST (Kennedy, 1990) and is involved in fine motor control (Whishaw et al., 1990), is unlikely to play a role due to it also being lesioned along with the CST. Interestingly, a study by Weidner et al. (2001) reported spontaneous sprouting of the ventral CST after ablation of the dorsal CST, which correlated with improvements in forelimb function. Although we did not study the ventral portion of the CST, a similar mechanism could have operated here. Taken together, we argue that functional improvements do not necessarily reflect the extent of anatomical regeneration and reconnection of descending nerve tract circuitry within the injured rodent spinal cord.

Gene therapy combined with OEG transplantation had a significant effect on the maintenance/regeneration of CST fibres, and sparing of spinal cord tissue in general. These results give credence to further study of OEG transplantation in combination with long-term secretion of neurotrophic factors, e.g. by using lentiviral vectors (Ruitenberg et al., 2002a). There is, however, an important need to discriminate between functional axonal regeneration, compensation and axonal plasticity as potential outcomes after spinal cord injuries. Development of more appropriate measures of sensorimotor performance in rodents (e.g. Metz and Whishaw, 2002) will allow better, more realistic and hopefully more clinically relevant interpretation of functional outcomes after spinal cord injury in these animals.


We wish to thank C.L. Christensen for assistance with histology, Dr P. Wood (The University of Miami) for providing the p75 antibody, and Dr B. Blits for the AdV vector stock encoding NT-3. The 2H3 monoclonal antibody, developed by Drs T.M. Jessell and J. Dodd, was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242, USA. Work in the laboratory of G.W.P. was supported by the National Health and Medical Research Council (NHMRC research grant No. 9935975 and RD Wright fellowship No. 303265), The Western Australian Institute for Medical Research, The Neurotrauma Research Program of Western Australia, The Australasian Spinal Research Trust and The Ramaciotti Foundation. The work performed in the laboratory of J.V. was funded by the Netherlands Organization for Scientific Research (NWO-GMW Pioneer grant No. 030-94-142) and the New Drug Research Foundation (NDRF; research grant No. 014-80-010).


View Abstract