Six-week-old adult male Sprague-Dawley rats (n = 31; weight, 195 - 210 g; Japan SLC Inc., Hamamatsu, Japan) were randomly assigned to undergo laminectomy only (laminectomy, n = 4), spinal cord injury only (injury, n = 10), or spinal cord injury with hind limb immobilization (injury + immobilization, n = 17). Histological study was performed for laminectomy (n = 2), injury (n = 6) and injury + immobilization (n = 6) group respectively. ELISA study was performed for laminectomy (n = 2), injury (n = 4) and injury + immobilization (n = 4) group respectively. Injection study was performed for injury + immobilization group (PBS injection, n = 3; GDNF injection, n = 4). The laminectomy group was a control group in this study. The protocol was approved by the Animal Committee of Aino University, Japan. All procedures were performed in accordance with the Guidelines for Animal Experiments of Aino University and in compliance with the Japanese Regulations for Animal Welfare.

Basic surgical procedures and postoperative care were as described previously [11] [14] -[16] . Laminectomy was performed at the mid-thoracic vertebrae 7/8. The spinal cord was crush-injured by dropping a 10-g metal rod from a height of 7.5 cm using a New York University (NYU) impactor [17] -[19] .

After surgery, the right and left hind limbs were immobilized with aluminum plates on soft sponges, with the ankle and knee joints fully extended. The gastrocnemius and tibialis anterior muscles were immobilized in the contracted and extended positions, respectively. The quadriceps and hamstring muscles were immobilized, but rats could move their hind limbs at the hip joints.

Recombinant GDNF protein (R & D Systems, Inc., Minneapolis, MN) or control phosphate-buffered saline (PBS) injection was administered to the injury + immobilization rats. GDNF at a concentration of 200 μg/mL or PBS alone was injected into the gastrocnemius muscles of both legs, using a 26G needle, at a depth of 1 mm and for a duration of 3 minutes; this was done twice weekly for 3 weeks following the crush injury. The gastrocnemius muscle was removed and, prior to transcardial paraformaldehyde perfusion, the wet weight was measured (PBS group, n = 3; GDNF injection, n = 4).

The spinal cords of the rats were cut into 2-mm lengths from the lesion site to the lumbar enlargement and were embedded in optimal cutting temperature (OCT) compound (laminectomy, n = 2; injury, n = 6; injury + immobilization, n = 6) (Tissue Tek, Sakura Finetechnical, Tokyo, Japan) at 3 weeks postintervention. The gastrocnemius muscle was removed prior to paraformaldehyde perfusion at 3 weeks postintervention. The muscles were then immediately frozen in acetone chilled with dry ice, and were embedded in OCT compound. Sections (thickness, 10 μm) were cut axially from the blocks by using a cryostat (CM1510S; Leica), and every fifth section was saved. A set of sections was mounted on 20 silane-coated glass slides (Matsunami, Tokyo, Japan) for each animal. Sections were used for hematoxylin and eosin (HE) staining and immunohistochemical analysis. The primary antibodies were anti-neuronal nuclei (NeuN) mouse polyclonal (1:100; Chemicon International, Inc., CA), anti-glial fibrillary acidic protein mouse monoclonal (1:300; Sigma, Tokyo, Japan) for astrocytes, and anti-GDNF goat polyclonal (1:10; R & D, Minneapolis, USA). Secondary antibodies were cyanine 3 (Cy3)- labeled anti-mouse IgG (1:1000; GE Healthcare Bio-Sciences Corp., NJ) and anti-goat Alexa Fluor 488 IgG (1:1000; Molecular Probes, OR).

The diameter of α-MN in lumbar spinal cord distribute from 30 to 50 μm, and peak at 40 μm [20] . Sections (thickness, 10 μm) were cut axially from the blocks by using a cryostat; every 5th section was saved. A set of sections was mounted on 20 silane-coated glass slides for each animal. Ten slides from 1 animal in each group were stained with anti-NeuN antibody, and the sections were examined by fluorescence microscopy (ECLIPSE E600, Nikon, Tokyo, Japan; DP71, Olympus, Tokyo, Japan). We scanned each section at 100× magnification, and counted the identified motor neurons at 40-μm intervals to prevent from biased counting. Mean α-MN counts were calculated for each group using the counts from each animal. The number of large (diameter, ≥40 μm) cells containing a single nucleolus located in the ventral horn was recorded and measured on both sides of each section with the use of an image analysis software (ImageJ, version 1.44; http://rsbweb.nih.gov/ij/index.html) [21] [22] .

We examined the HE sections to measure cavity volume [14] [15] . Images were obtained with the aid of a digital microscope (BZ-8000, Keyence). High-resolution images were used to trace the cavity areas. Areas identified in individual sections were measured with the use of an image analysis software (ImageJ). The areas of the cavities were measured on 6 serial axial sections with a section-to-section interval of 2 mm. These six measurements were added and averaged to produce the relative quantification data for each animal.

Fifty micrograms of spinal cord at L4/5 from each animal (injury, n = 4; injury + immobilization, n = 4) was homogenized in 1000 μL of ice-cold lysis buffer (137 mM NaCl, 1% NP40, 20 mM Tris-HCl [pH, 8.0], and a protease inhibitor (Halt Protease Inhibitor Cocktail, Thermo Fisher Scientific Inc., Kanagawa, Japan) using a mechanical homogenizer. Spinal cord homogenates were centrifuged at 1500 × g for 15 min at 4˚C. The resulting supernatant was divided into several portions and frozen at −80˚C.

For measurement of GDNF protein levels, spinal cord samples (laminectomy, n = 2; injury, n = 4; injury + immobilization, n = 4) were examined by ELISA with acid treatment procedure (GDNF kit, Novus Biologicals, CO). ELISA was performed twice for each sample.

Analysis of variance was used to characterize group differences. In the case of a significant F ratio by a one-way analysis of variance, a Tukey’s post-hoc test was performed. In all cases, the level of significance was set at P < 0.05 or P < 0.01. Data are presented as the mean ± standard deviation, unless otherwise indicated.

We previously reported that the injury + immobilization group had lower GDNF levels in muscles compared to the injury group. In this study, intramyocellular GDNF proteins were significantly decreased in the injury and injury + immobilization groups at 3 weeks postintervention (Figures 1(A)-(C)).

GDNF content at the L4/5 spinal cord segment was measured using ELISA. The injury + immobilization

group exhibited a significant decrease in GDNF proteins at 3 weeks postintervention compared to the injury and laminectomy group (Figure 2(A)). L4/5 spinal cord segments were not immunopositive for GDNF in injury + immobilization group, but slightly immunopositive in injury group (Figures 1(D)-(L)) compared to the laminectomy group.

We administered the recombinant GDNF proteins into the gastrocnemius muscle of rats in the injury + immobilization group. We measured the wet weight of muscle as the degree of disuse muscle atrophy. The average of wet weight of gastrocnemius muscle in PBS group (n = 3) and GDNF injection group (n = 4) was 1.13 ± 0.35 and 1.16 ± 0.25 respectively. The wet weight of gastrocnemius muscle in GDNF injection group did not differ from that of the PBS group at 3 weeks postintervention. The number of L4/5 α-MNs was 3.0 ± 1.41 in the control PBS injection group and 6.25 ± 1.25 in the GDNF injection group (Figure 3(A), Figure 3(B)). α-MN degeneration was partially rescued by exogenous GDNF proteins.

The mean volume of these cavities around the lesions did not differ significantly between the injury and injury + immobilization group (Figure 2(B)) or the PBS injection group and GDNF injection group at 3 weeks postintervention (Figure 3(C), Figure 3(D)).