All experiments with adult Sprague-Dawley rats (230-250 g, female) were carried out in accordance with the guidelines issued by the institutional animal care and use committee at the Yeungnam University, Daegu, Korea. Animals were acclimatized to controlled laboratory environments (12 hr light/dark cycles) with free access to food and water. Contusive SCI was induced using a weight-drop device developed at New York University (NYU) (21, 22). Briefly, animals were anesthetized with intraperitoneal injections of Zoletil (Virbac Laboratories, France, 50 mg/kg) and laminectomy was performed at T9 to expose the dorsal surface of dura matter. The exposed spinal cord was compressed by dropping a 10 g rod (2.5 mm in diameter) from a height of 25 mm using a NYU impactor. All rats received a daily intramuscular injection of antibiotics (prophylactic kanamycin, 1 mg/kg) for 1 week after injury to prevent urinary tract infections. Manual bladder evacuation was carried out twice per a day until voiding reflexes were reestablished.

A Basso, Beattie, and Bresnahan (BBB) scores were measured as previously described (23). Rats with BBB score of 6 within 4 days after SCI were selected and randomly assigned to rTMS (n = 5) or sham (n = 5) group. The transcranial rTMS was delivered using a customized magnetic stimulator (Biocon-1000Pro, Mcube Technology, Seoul) (19). A round prototype coil (7 cm diameter) of biphasic current waveform was positioned around bregma at 1 cm from the skull. The stimulation parameters were used as the follow: biphasic current waveform, 370 µs (pulse width), 0.2 T (intensity at 100% stimulation), 25 Hz (repetition rate), and 3 sec per on/off cycle for 20 min (total stimulation time). Sham stimulation was applied using the same method, but to prevent magnetic stimulation of the brain, the coil was not activated and the active surface of the coil was placed perpendicularly to the skull. Real and sham stimulations were applied daily 5 days per week for 8 weeks from the 4th day after SCI.

For immunohistochemical staining for microglial and astrocytic markers in spinal cords, all experimental animals were sacrificed at 8 weeks after SCI. After the deep anesthesia with intraperitoneal injections of Zoletil, a catheter was inserted into the left ventricle, and whole blood stream was rinsed with 100 mL of saline, followed by 500 mL of 4% paraformaldehyde (PFA), as previously described (8, 10). Spinal cords at the L3-S1 level, that is, below the sites of injury were removed, post-fixed for 2 days in 4% PFA, stored in 20% sucrose for at least 1 day, and then, transversely sectioned at 40-50 µm using a cryostat (Leica, Germany). To enhance the penetration of antibody to tissue, spinal cord sections were reacted with 50% ethanol for 30 min, and then treated with 10% normal donkey serum (NDS, Jackson Immunoresearch, USA) to block non-specific secondary antibody reactions. Sections were then incubated overnight in a mixture of primary antibodies; anti-Iba1 (Chemicon; 1:1,000) and anti-GFAP (Chemicon, 1:1,000), and then incubated with Alexa 488-conjugated donkey anti-rabbit (Invitrogen, USA, 1:200) and Cy3-conjugated donkey anti-mouse (Jackson Immunoresearch, USA, 1:200) for 3 hr. At each step, sections were rinsed with 1 × PBS 3 times. Finally, sections were mounted with Vectashield (Vector Lab, USA) and immunofluorescent images were acquired using a cooled CCD camera (F-View II, Soft imaging system, Germany) attached to a light microscope (Leica DMR) for the quantitative analysis of Iba1 and GFAP immunoreactivities in the dorsal horn. Images obtained from 6 spinal cord sections (4 sections from L4 and L5 segments) per rat in rTMS and sham groups were analyzed using image analysis software (Analysis Pro, SIS, Germany).

Collected data were inputted into SPSS/PC version 14.0 and continuously analyzed. Results are expressed as mean ± standard deviations (SDs). Wilcoxon Signed Rank test was used to analyze baseline measures and the Mann Whitney U test was used for behavioral tests. P values of < 0.05 were considered statistically significant.

In the spinal cord of L4 and L5 segments, microglial cells, which were immunoreactive for Iba1, were evenly distributed in the gray matter and showed the wavy pattern in naïve control (Fig. 1A, a). At 8 weeks after SCI, Iba1 positive microglia numbers were higher, especially in the lamina II and III, than in naive control (Fig. 1B). In high power view of the dorsal horn at lamina II, activated microglia included round-shaped perikaryon were observed (Fig. 1b). Numbers of activated microglia increased after SCI was subsequently reduced in the spinal dorsal horn of rTMS rats (Fig. 1C, c). Quantitative analysis of Iba1-positive areas revealed a significant reduction (27.2%) in Iba1-positive areas in dorsal horn of rTMS group (average area = 26,283.32 µm²), as compared with sham group (average area = 36,104.2 µm²) (P = 0.028) (Fig. 1D).

In astrocytes, the immunoreactivity of GFAP increased in spinal dorsal horn after SCI, versus naive control (Fig. 2A, B). In particular, the density of GFAP positive cells in lamina II was notably increased, indicating that numbers of reactive astrocytes were elevated after the injury (Fig. 2b). In spinal dorsal horn, reactive astrocytes numbers subsequently decreased in the rTMS group, but not in sham group (Fig. 2C). Quantitative analysis of GFAP-positive areas confirmed a significant decrease of GFAP expression (29.2%) in the dorsal horn from rTMS group (average area = 48,192.7 µm²), as compared with sham group (average area = 68,058.6 µm²) (P = 0.009) (Fig. 2D).