Reg-2 is a 16-KDa secreted protein belonging to a family of proteins structurally related to c-type lectins, which are upregulated in pancreatitis and have been shown to possess antiapoptotic actions on pancreatic cell lines (Iovanna et al.,1991; Orelle et al.,1992; Ortiz et al.,1998). Reg-2 normally is not expressed in the adult rat central (CNS) or peripheral nervous system (PNS), but it is expressed in a cytokine-dependent manner in both sensory and motor neurons during development (Livesey et al.,1997). Interestingly, Reg-2 is strongly upregulated in the PNS after nerve axotomy or crushing in adult rats, and it is also reexpressed in the CNS during Alzheimer's disease, spinal cord injury (SCI), or inflammation of peripheral motor and sensory nerves (De la Monte et al.,1990; Nishimune et al.,2000; Averill et al.,2002,2008; Fang et al.,2010). The expression of Reg-2 is driven by cytokines including the interleukin-6 family, ciliary neurotrophic factor (CNTF), leukemia inhibitory factor, and cardiotrophin (CT-1; Livesey et al.,1997). Previous studies demonstrated that expression of Reg-2 is also required for the survival-promoting actions of CNTF on motoneurons (Livesey et al.,1997; Nishimune et al.,2000). Indeed, blocking Reg-2 expression specifically abrogated the survival effect of CNTF, suggesting that it could be an important intermediate involved in the survival signaling pathway of CNTF-related cytokines (Nishimune et al.,2000).
The rapid increase in expression of CNTF after CNS injury has been hypothesized to play an intimate role in neuronal survival and functional recovery (Albrecht et al.,2002; Yokota et al.,2005; Müller et al.,2007). Previous investigation has shown that CNTF can rescue motoneurons from apoptosis induced by sciatic nerve axotomy (Oliveira et al.,2002), and intrathecal infusion of exogenous CNTF following SCI can reduce tissue damage, protect the rubrospinal descending tracks, and enhance the functional recovery of adult rats (Ye et al.,2004).
In our own studies, blocking endogenous Reg-2 expression in cultured embryonic neurons led to a significant reduction of neuronal survival and neurite outgrowth, whereas astrocyte survival was unaffected. The addition of the Reg-2 into the culture media after peroxide treatment or cellular hypoxia insult induced by mitochondrial poisoning reduced lactate dehydrogenase levels and cell death. Our data suggest that Reg-2 is essential for survival and neurite outgrowth of developing spinal cord neurons, and it plays a protective effect on spinal cord neurons against following in vitro injury (Fang et al.,2010).
As a continuation of our investigations, the present studies were aimed at determining whether Reg-2 has a neuroprotective effects in vivo using an adult rat spinal cord transection model, because the increase expression of Reg-2 has been detected in both motor and sensory neurons following transection injury of the spinal cord (Fang et al.,2010). The agents were delivered intrathecally into the site of injury over a period of 7 days, and the in vivo effects of Reg-2 at two doses (based on in vitro tested concentrations; Ye et al.,2004; Fang et al.,2010) were compared with CNTF and saline vehicle. Histological and immunohistochemical assessment, FluoroGold (FG) retrograde tracing, and locomotor testing were used to assess the effect of Reg-2 on white matter sparing, myelination, inflammation, long tract axonal protection, and functional recovery in rats with SCI.
MATERIALS AND METHODS
A total of 144 adult female Sprague-Dawley rats, weighing 200–250 g (obtained from the Experimental Animal Center of Zhejiang University), were used. Eight rats were used as a sham group, and the other 136 rats were randomly assigned into four SCI groups: saline vehicle (200 μL, i.t.), low-dose Reg-2 (genway, CA, 100 μg, i.t.), high-dose Reg-2 (500 μg, i.t.), and CNTF (recombinant rat ciliary neurotrophic factor, PeproTech, NJ, 10 μg, i.t.). The animals in each group were as follows: survival for 24 hr (N = 3), 7 days (N = 11), and 42 days (N = 20) post-SCI. Drugs were delivered chronically via implantation of Alzet osmotic minipumps over 24 hr (N = 12, pump model 2001D; 8.0 μL/hr; Alza Corp., Palo Alto, CA) and 7 days (N = 124, pump model 2001; 1.0 μL/hr). All animal procedures used in this study were carried out in accordance with the Guide for the care and use of laboratory animals in Zhejiang University.
Preparation of Osmotic Pumps
Before surgery, osmotic minipumps were prepared under sterile conditions and filled with 0.9% saline vehicle, Reg-2, or CNTF in saline. Cannulae consisting of polyurethane tubing were sterilized overnight in 100% ethanol before being attached to the flow moderator of the pump. All cannulae were examined carefully before sterilization for patency at the free end and for fluid leakage at the adapter site. The pumps were incubated overnight at room temperature in sterile saline for priming.
Surgical Procedures and Drug Delivery
The animals were anesthetized by an intraperitoneal injection of 1% Nembutal (40 mg/kg). A laminectomy was then performed at the T9–T10 level. The spinal cord of the sham-operated group was exposed but left intact. In the other surgery groups, the rat spinal cord was cut transversely against the inner wall of the vertebral canal between T9 and T10, and a second cut was performed 1 mm rostral to the incision. Blood and cord tissues in this gap were removed by gently aspiration with a fine suction pipette. Animals showed a clear spasm of lower limbs and tail upon the injury. The cord stumps were then slightly lifted by microforceps to check the completeness of lesion, and a visible separation of stump was observed under operating microscope. Catheters were then introduced into the subarachnoid space through small holes in the dura at T12, and the tips advanced to T9, with infusion starting immediately. The catheter was sutured to the muscles, and the pump was placed into a subcutaneous pocket. The dura was sealed and the wound was closed in layers.
After surgery, the rats were placed in a temperature- and humidity-controlled chambers overnight; penicillin was injected intramuscularly for 7 days (25,000 UI per rat, b.i.d.). Manual bladder emptying was performed at least three times daily until reflex bladder emptying was established.
The Basso Beattie Bresnahan locomotor rating scale.
The Basso Beattie Bresnahan (BBB) locomotor rating scale was used to test the behavioral consequence of the surgical procedures (Basso et al.,1995,1996; N = 5 per group). The BBB 22-point scale (0–21) can differentiate hindlimb locomotor activity over a wide range of injury, from a score of 0, indicative of no observed hindlimb movements, to a score of 21, representative of a normal ambulating rodent. Before surgery, naive rats were individually placed in the center of a circular enclosure made of molded plastic with a smooth, nonslip floor (90 cm diameter, 7 cm wall height) for 4 min to evaluate the locomotor activities of the trunk, tail, and hindlimbs and to ensure that all subjects consistently obtained a maximum score of 21. Subsequent testing was then initiated 3 days after surgery and then twice a week for 6 weeks. Typically, the rats were placed in the open field for 4 min and scored by observers blinded to the experimental treatments (see Basso et al.,1995,1996).
Grid-walk analysis was also used to assess hindlimb locomotor deficits (Behrmann et al.,1992; N = 5 per group), and this was performed from 3 to 6 weeks postinjury on control and treated rats, as none of the animals with insufficient hindlimb recovery achieving a minimum score of 11 on the BBB scale at first 2 weeks following surgery (Reynolds et al.,2004). The grid runway is a 3 × 3 square foot plastic mesh with the hole size of 2.5 × 2.5 cm2. Each animal was placed on the grid and allowed to perform active grid walking task for a period of 3 min. The assessment was performed blind by two investigators, with each investigator observing only the right or left hindlimb. During this time period, the number of footfalls (fall of the hindlimb, including at least the ankle joint, through the grid surface) was determined individually for each hindlimb. The number of footfalls of left and right hindlimbs occurring over the 3 min active grid walking period was then counted. For accuracy, the number of errors per active grip walking period was averaged from the four trials at each time point. If an animal was not able to move its hindlimbs, a maximum of 20 errors was given (Metz et al.,2000).
Fluorogold Retrograde Tracing
Fluorogold (Fluorochrome, Denver, CO) retrograde tracing was used to determine the extent to which spared descending axons reached the rostral lumbar enlargement (N = 4 per group). Briefly, 5 weeks postinjury, 4% FG was injected into the lumbar enlargement at 1 mm from the midline and 1.5 mm from the dorsal cord surface on both sides (0.5 μL per injection) using a glass micropipette attached to a pneumatic picopump which was ∼12 mm distal to the lesion epicenter. One week later, rats were sacrificed in preparation for histological analysis. Transverse sections of 25 μm thickness from the C6 and T5 segments were chosen as representative of cervical and thoracic segments. The number of FG-labeled neurons was counted bilaterally within these two cord segments (3 mm in length). For transverse sections (of 40 μm thickness, every fourth section) from the entire brainstem and sensorimotor cortex, the total number of FG-labeled neurons in the spinal nucleus of trigeminal nerve (SPT), dorsal thalamus (DT), the caudal pontine reticular nucleus (Pnc), the red nucleus (RN), and the hindlimb area of motor cortex (Ctx-HL) was counted bilaterally. Only FG-labeled neurons with nuclei were mapped and counted to avoid the duplication of single cells in the count.
Perfusion and Tissue Processing
Animals were sacrificed at 1 (N = 8) and 6 weeks (N = 8) following injury. The rats were given a lethal dose of Nembutal and then perfused intracardially with saline containing 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The spinal cord and brain tissues were then carefully harvested. The implanted osmotic pumps and cannulas were removed and examined for their contents. All osmotic pumps were empty of their contents at the time of sacrifice. Spinal cords were carefully dissected, and a 1.5-cm spinal cord segments containing the injury site were fixed in the same fixative for 4 hr and then transferred into 30% sucrose in PBS until the tissue dropped into bottom of the container. Twenty-micron-thick sections were cut on the freezing microtome horizontally (N = 3) and transversely (N = 5) using a Leica cryostat. The sections were then mounted onto 0.02% poly-L-lysine-coated slides. All sections were collected. One of every 10 of the sections was taken as a set, and then the 10 sets of sections were processed for histological or immunohistological purposes: hematoxylin and eosin (H&E), cresyl violet, Luxol fast blue (LFB) staining, silver degenerating axons staining described by Eager (1970), immunohistochemistry staining for neurofilament (NF)-M, TNF-alpha, caspase 3, and immunofluorescence labeling for growth-associated protein-43 (GAP43), CD45, and CD68/ED1.
To measure the area of the cystic cavity within spinal cord, sections H&E staining was used. Five sections from the injury epicenter of each animal were analyzed; the images were captured by a Nikon CCD camera, and the cross-sectional area of the cavity was measured using NIH Image software. The maximal value of cross-sectional area of each cavity was selected at each section. Two independent observers made measurements over the same images, and the final maximum cross-sectional area of cavity in each animal was the mean value of both observers.
Five transverse sections from each animal were selected for cresyl violet staining at random. The sections were photographed under 400× magnification using a Nikon TE-300 microscope in three vision fields per section. Neuronal counts were performed on both anterior horns, with counting restricted to the neurons having a well-defined nucleolus and a cell body rich in endoplasmic reticulum.
Transverse sections from the injury epicenter were also stained for myelin using LFB. Digital photomicrographs were obtained at 40× magnification and analyzed for spare white matter area using NIH image software. The injury epicenter was determined for each animal at the rostrocaudal level along the spinal cord axis that contained the least amount of spared myelin per transverse section. The degenerating axons of each section were also identified using Eager's silver staining method.
The sections were permeabilized and blocked with 0.3% Triton X-100/10% normal goat serum in 0.01 M PBS for 30 min and then incubated with polyclonal rabbit antibodies: anti-160KD NF-M (1:1,000, Neuromics, MN), antitumor necrosis factor alpha (TNF-α 1:1,000, ProSci Incorporated, CA), and anti-caspase-3 (1:500; Cayman Chemical, MI) overnight at 4°C. Sections were incubated with secondary biotinylated goat anti-rabbit IgG antibody (1:400; Vector Laboratories, CA) for 1 hr at room temperature, followed by avidin–biotin peroxidase complex (ABC kit, Thermo Fisher Scientific, MA). After incubation for 5 min with 0.02% DAB and 0.003% H2O2 in 0.005 Tris-HCl, the sections were counterstained with hematoxylin. Primary antibody omission controls were used to further confirm the specificity of the immunohistochemical labeling. Five transverse sections from each animal were selected at random, and images were photographed under 400× magnification in three visual fields per section. TNF-α and caspase-3 immunoreactive cells were counted on preserved anterior horns of each spinal cord sections.
The sections were per-treated according to the method described above and incubated with primary polyclonal rabbit anti-160KD NF-M primary antibodies (1:1,000, Neuromics, MN), anti-GFAP (1:200, Thermo Fisher Scientific, MA), and monoclonal antibodies including mouse anti-GAP43, CD45, and CD68/ED1 (1:100; Santa Cruz Biotechnology, CA) overnight at 4°C. The sections were then washed with PBS and incubated with 1:200 TRITC- or FITC-conjugated goat anti-rabbit/mouse IgG secondary antibodies for 1 hr at 37°C (Invitrogen, CA). The sections were finally coverslipped with antifade Gel/Mount aqueous mounting media (SouthernBiotech, AL). All control sections were incubated in PBS without primary antibodies. Photographs of GAP-43 immunoreactive cells were counted, and immunoreactive area of CD45 and CD-68 was analyzed with NIH image software.
Rats were sacrificed by decapitation at 24 hr post-SCI (N = 3 per group), and one 10-mm spinal cord segment containing the injury epicenter was prepared for Western blot. Total proteins were extracted using 2 mM PMSF in a EDTA-free 1 mL ice-cold RIPA buffer added protease inhibitor cocktail. The protein concentrations were determined using the Bradford protein assay. SDS-PAGE was performed on a 10% polyacrylamide slab gel and separated proteins then electrophoretically transferred to PVDF membrane at 70 V for 1.5 hr at 4°C in a Bio-Rad TransBlot apparatus. After blocking nonspecific binding sites with bovine serum albumin, the membrane was incubated for 12 hr at room temperature with primary rabbit polyclonal anti-MPO (1:200; Santa Cruz Biotechnology, CA), then washed three times for 5 min with TBST at room temperature, incubated with a secondary horseradish peroxidase (HRP)-conjugated goat anti-rabbit (1:5,000, Santa Cruz, CA) antibody for 1 hr, and washed with TBST again. The Western blot was visualized using the ECL Plus detection system as described in the technical manual provided by Amersham Pharmacia Biotech, followed by imaging and quantification of protein bands using Bio-Rad Quantity One 1D software. To normalize the protein bands to a gel loading control, membranes were washed in TBST and reprobed with rabbit anti-β-actin (1:5,000, AbCam, MA) followed by incubation with HRP-conjugated goat anti-rabbit antibody (1:5,000, Santa Cruz, CA) and ECL detection. For the negative control, the primary antibody was omitted.
Data are presented as mean ± SD. One-way analysis of variance with post hoc Tukey t-tests were used to determine statistical significance. A P value of <0.05 was considered statistically significant. All statistical tests and graphs were performed or generated, respectively, using GraphPad Prism Version 4.0 (GraphPad Prism Software, CA).
Reg-2 and CNTF Attenuated Neuron Loss in Injured Spinal Cord
There was a remarkable loss of neurons in the anterior horn of rats following SCI. The total neuron count reduced to 48% of the sham group at 7 days after injury and further decline to 28% at Day 42 (Fig. 1A,B). Compared with the vehicle control, Reg-2 protein and CNTF treatment significantly reduced neuron loss induced by injury at both 7 and 42 days (P < 0.01, Fig. 1). Although the ratio of neuronal survival was not obviously different among the three SCI drug-delivery groups in 1 week (P > 0.05, Fig. 1A), there were 25%–30% more survival neurons in the CNTF and high-dose Reg-2-treated groups versus the low-dose Reg-2-treated group at 6 weeks following injury (P < 0.05, Fig. 1B).
Reg-2 and CNTF Inhibited Inflammatory Reaction and Apoptosis After SCI
MPO was used as a marker for neutrophil infiltration, CD45 as a marker for extravasated leukocytes, and CD68 as a marker for activation of resident microglia and extravasated macrophages. Posttransection, the expression level of MPO greatly increased in vehicle-treated group at 24 hr postinjury, whereas infusion of CNTF and Reg-2 protein resulted in a significant decrease in MPO activity and neutrophil infiltration (P < 0.01, Fig. 2). Moreover, the number of CD45+ cells was found to be greatly increased in vehicle-treated rats over most of the cross section of the spinal cord at the injury epicenter (Fig. 3A) and over several millimeters rostral and caudal to it. This infiltration was markedly attenuated by CNTF and Reg-2 treatment (Fig. 3B–D). Similar results were seen with CD68 staining (Fig. 3E–H). Sham-operated rats had no marked immunostaining for CD45 and CD68. This confirmed that CNTF and Reg-2 treatment inhibited extravasation of inflammatory cells into the spinal cord. Quantification of the area of immunostaining at 10 mm distances in a spinal cord segments from 5 mm rostral to 5 mm caudal from the epicenter showed that CNTF and Reg-2 treatments reduced inflammation at all postinjury times from 1 to 6 weeks (P < 0.01, Fig. 4A,B,E,F). Generally, CNTF and high-dose Reg-2 had better effects than the low-dose of Reg-2 (P < 0.05, Fig. 4A,B,E,F). Regression analysis showed that the extent of inflammation at 7 or 42 days postinjury was correlated with white matter loss (P < 0.01, Fig. 4C,D,G,H).
The expression of active caspase-3, an enzyme critically involved in the execution of the mammalian apoptotic cell death program, was also detected. The number of caspase-3 IR neural cells was reduced in Reg-2 and CNTF treatment groups following SCI (Fig. 5). Significant differences were found in active caspase-3 expression between the low-dose Reg-2 group and other two treatment groups at 7 days following SCI, but not at 42 days (Fig. 6A,B).
Reg-2 and CNTF Reduced White Matter Loss and Demyelination
The lesion area, cavitation, white matter sparing area, and spared axons at the injury epicenter were measured. In the vehicle group, 7 days following SCI, only a portion of the ventral and dorsolateral white matter remained present (about 42%) compared with sham-operated rats (Fig. 7A,B). Following Reg-2 protein and CNTF treatment, much more of the ventral and lateral white matter were spared (Fig. 7C,D). Rats in CNTF delivery group could keep the myelination area up to 77% of the sham group, implying that much more sparing white matter was present in the CNTF group compared with the vehicle control and Reg-2 treatment groups (P < 0.01, Fig. 8A).
With the formation of the cystic cavity upon 6 weeks following SCI, the sparing of the white matter area within the spinal cord of rats with SCI dropped further (Fig. 8C). More sparing of white matter in the spinal cord of rats among the three drugs-infused groups was still visible compared with vehicle controls (P < 0.01, Fig. 8D). More sparing of white matter could be found in rats treated with CNTF and high-dose Reg-2 compared with the lower dose of Reg-2 at this time point, while no significant differences between CNTF- and Reg-2-treated group were observed (P > 0.05, Fig. 8D).
Numerous black color degenerating axons as well as the loss of axons were demonstrated in vehicle control with Eager's sliver staining (Fig. 7F), while within the Reg-2- and CNTF-treated groups, a larger area with tan color normal stained axons were observed at 7 and 42 days following SCI (Fig. 7E,G,H). Regression analysis showed that sparing of white matter area at 7 and 42 days postinjury correlated with BBB scores (P < 0.01, Fig. 8B,F).
Six weeks following SCI, large cystic cavities appeared in the injury epicenter of spinal cord in the saline vehicle group (Fig. 9A); the majority of the area in spinal parenchyma was destroyed and only a thin rim of tissue remained (Fig. 9E). The cavities within the spinal cord of the drug-infused rats were much smaller than that of vehicle group (P < 0.01, Figs. 8E, 9B–D,F–H). The average volume of the cavity was found to reduce to 51% in low-dose Reg-2-treated group and to 40% and 39% in the CNTF- and high-dose Reg-2-treated groups, respectively. A lower lesion volume was confirmed in CNTF- and the high-dose Reg-2-treated group compared with the low-dose Reg-2-treated group (P < 0.05, Fig. 8E).
NF-IR stained axons appeared in the area of cavity in each groups. Indeed, the cystic cavity was encircled by glial tissues (Fig. 10F). In vehicle control rats, unchained NF-IR axons surrounded the cavity, and there were only a few NF-IR fibers pass through along the margin of cord (denoted by arrows); no intact fibers passed continuously through the injury epicenter (Fig. 10A,B). However, numerous NF-IR axon profiles were found to penetrate heavily into the spared white matter area in CNTF- and Reg-2-treated groups, implying the presence of more intact axons within this area. NF-IR axons were also seen directly adjacent to the lesion as well as within the lesion cavity in all three groups (Fig. 10C–E).
Reg-2 and CNTF Increased Axonal Sparing of Proprio- and Supraspinal Origins
Intrathecal infusion of either Reg-2 or CNTF resulted in a significant increase in the mean number of FG-labeled propriospinal neurons located at both T5 and C6 spinal levels (Fig. 11). However, no statistically significant difference was found in the number of FG-labeled cells among the three different treated groups at both cord levels (P > 0.05, Fig. 12A).
FG-labeled neurons in selected supraspinal nuclei were also counted to assess whether the Reg-2 and CNTF could induce a reduction of lesion volume and an increase in axon sparing of neurons in these areas (Figs. 13, 14). Significant statistical differences in the number of FG-labeled neurons between the saline- and drug-treated groups were detected in the caudal pontine reticular nucleus, the red nucleus, the hindlimb area of motor cortex (Figs. 12B, 13, P < 0.01), the spinal nucleus of trigeminal nerve, and the dorsal thalamus (Figs. 12B, 14, P < 0.01). However, no statistically remarkable differences were found in the number of FG-labeled neurons between the CNTF- and high-dose Reg-2-treated groups (P > 0.05), but significant differences occurred compared with the low-dose Reg-2-treated group (P < 0.05, Fig. 12B).
Reg-2 and CNTF Treatments Promoted Functional Recovery After SCI
All injured animals developed obvious bilateral hindlimb paralysis immediately following SCI and presented partial functional recovery over the 6-week observation period. Only the sham-operated animals showed stable scores. During the first 7 days after injury, none of the rats in the vehicle control group had obvious locomotive activity, but the Reg-2- and CNTF-treated rats began to display higher BBB scores (Fig. 15A), implying either Reg-2 or CNTF treatments improved the functional recovery. Six weeks after injury, all injured rats showed progressively increasing behavioral recovery. Moreover, the rats treated with Reg-2 and CNTF continued to display better behavioral scores than the controls (P < 0.05, Fig. 15A). Four weeks after surgery, all rats from these treated groups also exhibited a remarkable improvement in BBB subscore, and the CNTF- and high dose Reg-2-treated group also gained higher subscores than low-dose Reg-2-treated group at 5 weeks postinjury (P < 0.05, Fig. 15B). At the same time point, the average score of the footfalls in grid-walk analysis of drug-treated groups was significantly lower than vehicle controls (P < 0.01, Fig. 15C); regression analysis also showed that the BBB score was highly correlated to the error score of grid-walk analysis at 6 weeks postinjury (P < 0.0001, Fig. 15D).
Reg-2 and CNTF Treatments Enhanced Expression of GAP-43 and Inhibited Expression of TNF-α in Neurons Following SCI
Following transections, an extensive expression of TNF-α was found in neurons and other neural cells of spinal cord in the vehicle treatment group (Fig. 16A). Comparatively, TNF-α IR neural cells were evidently reduced in CNTF- and Reg-2-treated groups at 7 and 42 days following SCI (Figs. 16B–D, 17A,B, P < 0.05). The low-dose Reg-2-treated group shows significant differences when compared with the other two treated groups (P < 0.05, Fig. 17A,B).
Except for a few GAP-43-labeled cells locating at the dorsal column, there was no other GAP-43 expression in sham group. Following SCI, the expression of GAP-43 increased in cell bodies and fibers in each group (Fig. 16E–H). The CNTF- and Reg-2-treated rats had much more GAP-43-labeled cells when compared with vehicle control at 7 and 42 days following SCI (P < 0.05, Fig. 18A,B). Significant differences were found between the low-dose Reg-2 group and the other two treated groups at 7 days following SCI, but not at 42 days (Fig. 18A,B).
The neuroprotective effects of CNTF following SCI and other CNS disease have long been confirmed by previous researchers (Linker et al.,2002; Oliveira et al.,2002; Ye et al.,2004; Lu et al.,2009). Studies have shown that intracerebroventricular or intrathecal infusion, but not intravenous or local injection, is more effective at reversing neuronal deficits (Meazza et al.,1997). In this investigation, CNTF was administrated as a positive control to evaluate the effects of its downstream signaling molecule, Reg-2. To maintain a continuous administration with a consistent dose, we used implantable Alzet osmotic pumps to deliver drugs intrathecally.
SCI is a major cause of disability (Profyris et al.,2004). The functional decline following SCI is caused by both direct injury and secondary pathophysiological mechanisms induced by the initial trauma (Carlson and Gorden,2002; Houle and Tessler,2003). The initial trauma (mechanical, chemical, poisonous, etc.) rapidly and directly ruins neurons and glia and a delayed secondary pathology follows (Donnelly and Popovich,2008; Fitch and Silver,2008). The latter is characterized by neuronal and glial apoptosis, axonal degeneration, increased blood–CNS barrier permeability, and a complex neuroinflammatory response that can lead to extensive demyelination, tissue damage, which further causes exacerbation of neuron loss (Norenberg et al.,2004; Profyris et al.,2004; Fleming et al.,2006). At later stages, the cord is observed to display a larger cavity formation, and the injury epicenter usually has reactive gliosis surrounding the cystic cavity. This creates a highly nonpermissive environment to axonal regrowth (Fitch and Silver,2008).
Treatments targeting the neuroinflammatory reaction may be reasonably assumed to alleviate the secondary injury (Donnelly and Popovich,2008). Our investigation has shown that Reg-2 and CNTF intrathecal infusion attenuated leukocytes infiltration and macrophages extravasation, the latter being a major contributor to demyelination (Shuman et al.,1997; Popovich et al.,1999,2002; Gris et al.,2004). The anti-inflammatory mechanism of action of CNTF involves an inhibition of the production of TNF-α (Meazza et al.,1997; Linker et al.,2002). We also detected the expression of TNF-α in the CNTF and Reg-2 treatment groups. The results revealed that high dose of Reg-2 exerted similar effects compared with CNTF to decrease TNF-α expression. This is logical because TNF-α has long been known to contribute toward the demyelination of axon and as a gliosis trigger (Linker et al.,2002; Inukai et al.,2009; Lu et al.,2009). Indeed, microinjection of TNF-α into the brain elicited macrophage and leukocyte recruitment, whereas injections into the spinal cord elicited neutrophil and macrophage infiltration and activation (Donnelly and Popovich,2008). The inhibition of TNF-α, and perhaps also other inflammatory-related cytokines, may also contribute to the beneficial anti-inflammatory effects of CNTF and Reg-2.
Apoptosis has been demonstrated as a mechanism of cell death after SCI (Crowe et al.,1997; Liu et al.,1997). Following such an insult, the apoptotic cells that were mainly affected were neurons and oligodendrocytes. The cell death obviously contributes to locomotor deficits and widespread demyelination damage during the secondary injury (Crowe et al.,1997; Liu et al.,1997; Warden et al.,2001). Caspase-3 is one of the key executors of apoptosis and is responsible for the cleavage of proteins such as the nuclear enzyme poly(ADP-ribose) polymerase. Increased expression of caspase-3 has been shown after SCI (Springer et al.,1999). Our results showed either Reg-2 protein or CNTF reduced the expression of active caspase-3 after SCI, indicating that CNTF and Reg-2 have antiapoptotic profiles. The total neuron count also suggested that Reg-2 and CNTF can reduce neuron loss in the injured spinal cord, which is consistent with its proposed antiapoptotic effects.
It is also evident that the expression of GAP-43 increases initially after injury (Kobayashi et al.,1997), while the application of Reg-2 protein and CNTF could further improve and maintain this upregulation. Thus, it is suggested that in these treated animals, the prevention of neuronal loss and the stimulation of GAP-43 expression were correlated with an increased regenerative capacity of injured neural cells (Xu et al.,2009).
After acute spinal cord contusive injury, the main neuronal pathophysiology is hemorrhage, edema, and axonal demyelination at the injury site (Ye et al.,2004). In later stage of secondary injury, obvious cavitation and reactive gliosis surrounding the cystic cavity will present in spinal cord parenchyma at lesion epicenters and adjacent rostral/caudal regions. In the control group, the severe tissue loss and cavity formation left only a little spare region of white matter in the ventrolateral margins at each level. The amount of spared tissue closely correlated with the residual neurological function (Basso et al.,1996; Rosenberg and Wrathall,1997). Our observations have demonstrated that Reg-2 infusion at low and high doses can cause a 49% and 61% reductions in the total lesion volume, respectively. The reduction in lesion size led to an increase in white matter sparing surrounding the lesion epicenter, as shown by LFB staining. Injury-induced demyelination was also significantly decreased after drug delivery. Regression analysis showed a decrease of inflammation area in each treatment group, which was correlated with the white matter sparing area. This is reasonable, because inflammation contributes to tissue loss after SCI (Donnelly and Popovich,2008) and may partly explain the neuroprotective effects of CNTF and Reg-2. Moreover, the area of the spared white matter also correlated well with BBB scores, implying that inhibition of demyelination and attenuated tissue damage following Reg-2 and CNTF treatments probably contributed to better outcomes in functional recovery.
It is important to appreciate that an increase in white matter sparing is accompanied by an increased sparing of axons. Because the cavitation and reactive gliosis surrounding cavity are barriers to prevent the growth and regeneration of axons, the reduction of cavity volume indicates that axons are going through the injury epicenter, especially the descending and ascending long distance axons, as demonstrated by FG-retrograde tracing.
The C6 and T5 segments were chosen as representative cervical and thoracic segments as axons from these regions form propriospinal projections descending to the spinal cord caudal to the injury. Counting the number of FG-labeling neurons in selected propriospinal and supraspinal regions where axons traversing through the SCI site retrogradely transporting of FG allowed us to compare axonal sparing amongst different treatment groups (Iannotti et al.,2004; Liu et al.,2007). Reg-2 and CNTF treatment provided a two- to threefold increase in the number of FG-labeled propriospinal neurons in the T5 and C6 cord segments. An increase in the number of labeled neurons in SPT and DT indicated an increase of sparing ascending axons to supraspinal regions, whereas in Pnc, RN, and Ctx-HL, it indicated an enhancement to protected descending axons from supraspinal regions. These data suggest that Reg-2 and CNTF infusion not only protected spinal cord tissues from secondary damage but also enhanced the sparing of selective proprio- and supraspinal axons that traverse the lesion site. All the ascending and descending pathways above- mentioned play a role in the hindlimb function (Kuypers et al.,1982; Walberg,1982). Thus, Reg-2- and CNTF-treated groups all gained much better BBB scores compared with vehicle controls. It should be noted that, although no statistically significant differences were found in the number of FG-labeled propriospinal neurons among all these treated groups, CNTF- and high-dose Reg-2-treated animals had a greater number of FG-labeled supraspinal neurons and a lower cavity volume than the low-dose Reg-2-treated group. Thus, the BBB score analysis revealed that more animals from these two treated groups had a higher hindlimb function, consistent with a coordinated stepping. It seems the long distance supraspinal axons are more susceptible to the volume change of lesion cavity, sparing white matter, and myelination area.
In this study, extraordinary BBB scores of more than 16 at 6 weeks postinjury in the CNTF and high-dose Reg-2 treatment groups were recorded; this degree of protection is much higher than that of previous reports from complete spinal cord transection models (Basso et al.,1996; Chen et al.,2004; Zhang et al.,2007).One possibility is that the transection injury of the spinal cord in our mode was incomplete, and then some intact fibers could have been left in the ventral and ventrolateral funiculi. Indeed, such incomplete spinal cord incision may often been encountered because of the rough inner surface and curved lateral recesses of the vertebral canal (You et al.,2003). Even 4.82% of remaining fibers in unilateral ventrolateral funiculus were able to sustain a certain spontaneous recovery of locomotion, with the BBB score reaching 7 in untreated rats with spinal cord transection injury performed by blade within a certain segment, as quadruped animals contain much larger capacity of spontaneous recovery than primates and humans (Rossignol et al.,1996; Field-Fote,2000). In this study, a little sparing of white matter occurred at the ventrolateral area of spinal cords in vehicle control group with averaged BBB score of 7.7 at 6 weeks postinjury. Further, a few NF-IR fibers passed through the cord along its margin in longitudinal sections of control group (Fig. 10A,B), which is similar to the previous description (You et al.,2003). Moreover, the magnitude of locomotor recovery was correlated with the percentage of remaining nerve fibers along the periphery of the ventrolateral spinal cord, exclusive of most of the descending motor tracts, implying an essential role of propriospinal connections in the initiation of spontaneous locomotor recovery (You et al.,2003). FG-labeled neural cells were found at T5 and C6 segments in our control group, but rarely in the brainstem and cortex, which is consistent with the suggested essential role of propriospinal connections. Clearly, therefore, although we were very careful in the process of transection surgery, some fibers in ventral and ventrolateral funiculi may have remained in each group.
Animals with transected spinal cords have been shown to be capable of executing a variety of complex innate behaviors (i.e., walking, scratching, and paw-shaking) in the absence of supraspinal and proprioceptive sensory input (Field-Fote,2000). These spontaneous improvements in hindlimb locomotor movements were attributed to the formation of functional neuronal connections within the locomotor central pattern generator (CPG; Yakovenko et al.,2007). Moreover, the lower thoracic segments and the lumbar segments of rats were thought to contain the CPG (Chen et al.,2004; Yakovenko et al.,2007), and a resection of all the segments below the T10, or a resection of 2–3 mm spinal cord including the whole segment of T10, could produce a complete injury with BBB scores lower than 3 at 6 weeks postoperation (Basso et al.,1996; Chen et al.,2004). Thus, the remaining fibers and the spontaneous improvements of hindlimb locomotor movements generated by the CPG could all contribute to the extraordinary high BBB score in our vehicle-treated control animals of 7–8.
Although the possible incomplete incision may leave remaining fibers in each animal, only the vehicle-treated animals acquired an average score below 8 at 6 weeks postinjury. With the consistency of our surgical procedure, the outcomes from double-blind BBB scoring, tract tracing, and histological assessment by two individuals “blinded” to rat treatment status; improved locomotor function was only associated with drug treatment groups. Moreover, the grid-walk locomotor test also confirmed the improved motor function recovery of hindlimbs in the drug-treated groups as revealed by BBB scoring and subscore analysis. Therefore, the evident protective effect of CNTF and Reg-2 was showed in our study.
In fact, the secondary damage interrupting the remaining spinal pathways, including hemorrhage, edema, subsequent axonal demyelination, and neuronal death, was more severe to the remaining ventral fibers than the primary blade spinal cord transection itself. Although similar volume of remaining fibers may be left in all animals after transection surgery, most of the remaining fibers in the control animals were prone to destruction under the unfavorable microenvironment induced by secondary injury, that is, more severe inflammatory reaction, gliosis barriers, destructive factors, and less neurotrophins in the injury area. Nevertheless, Reg-2 and CNTF treatment exerted neurotrophic effects on the improvement of the microenvironment, and then more fibers were remained that presumably contributed to the higher scores in functional recovery, compared with vehicle control with the same extent injury. In our study, the vehicle control rats got higher scores of 7–8 compared with 2–3 from other reports in complete transection models. Then, the extraordinary scores of CNTF and Reg-2 treatment groups of 15–17 may correspond to the scores of 11–13 in complete transection models. Although treatment was only administered for 7 days, the sustained neuroprotective effects could be observed at 6 weeks after injury, suggesting that neurotrophic treatments had long-lasting effects even when terminated after 1 week.
On the basis of these studies, we conclude that Reg-2 expression not only exists as an obligatory intermediate in the survival signaling pathway of CNTF-related cytokines but can also play a role as a neuroprotective agent by reducing spinal cord secondary injury, protecting the remaining fibers, promoting axonal growth, decreasing cell death and demyelination area, and finally improving functional recovery following SCI. Reg-2 acted in a dose-dependent manner, and its neuroprotective spectrum resembles the effects of CNTF, the neurotrophic factor upstream to induce Reg-2 expression. Although the exact mechanisms are still unclear, the reupregulation of Reg-2 expression in injured neurons following CNS injury and disease was likely to be regarded as a protective reaction to harmful insults. Future studies of Reg-2 effects in vivo with the aim of further optimizing the application dose and combination with other neuroprotective factors targeting different pathways from the neurotrophic cytokines family may achieve an even better protection and functional increased recovery outcomes after SCI and possibly in other CNS pathological conditions.