Impact of Endurance Training on Regeneration of Axons, Glial Cells, and Inhibitory Neurons after Spinal Cord Injury: A Link between Functional Outcome and Regeneration Potential within the Lesion Site and in Adjacent Spinal Cord Tissue

Abstract

Endurance training prior to spinal cord injury (SCI) has a beneficial effect on the activation of signaling pathways responsible for survival, neuroplasticity, and neuroregeneration. It is, however, unclear which training-induced cell populations are essential for the functional outcome after SCI. Adult Wistar rats were divided into four groups: control, six weeks of endurance training, Th9 compression (40 g/15 min), and pretraining + Th9 compression. The animals survived six weeks. Training alone increased the gene expression and protein level of immature CNP-ase oligodendrocytes (~16%) at Th10, and caused rearrangements in neurotrophic regulation of inhibitory GABA/glycinergic neurons at the Th10 and L2 levels, known to contain the interneurons with rhythmogenic potential. Training + SCI upregulated markers for immature and mature (CNP-ase, PLP1) oligodendrocytes by ~13% at the lesion site and caudally, and increased the number of GABA/glycinergic neurons in specific spinal cord regions. In the pretrained SCI group, the functional outcome of hindlimbs positively correlated with the protein levels of CNP-ase, PLP1, and neurofilaments (NF-l), but not with the outgrowing axons (Gap-43) at the lesion site and caudally. These results indicate that endurance training applied before SCI potentiates the repair in damaged spinal cord, and creates a suitable environment for neurological outcome.

Keywords: endurance training; neurological outcome; regeneration of cell populations; spinal cord compression.

Figure 1

Physiological intersegmental differences between the number of GABA/glycinergic neurons in Th10 and L2 segments in dorsal horns (DHs), intermediate zone (IZ), and ventral horns (VHs). Results are presented as mean ± SD (n = 3). Data were statistically evaluated using Student’s parametric t-test; ** p < 0.01; **** p < 0.0001.

Figure 2

Immunohistochemical analysis of GABA (B,C,E,F; green) and Glyt2 (H,I; green) neurons with BDNF (A,C,G,I; red) and GFRα (D,F; red) in the intermediate zone and ventral horns of Th10 spinal segments in naïve control group. White arrows show the overlapping of the inhibitory neurons with BDNF or GFRα and DAPI, heads show Glyt2 neurons (n = 3). Scale bars: (A–I), 100 µm.

Figure 3

Fluorescent labeling of BDNF growth factor (A,C,J,L; red) and GFRα receptor (D,F,G,I; red) after six weeks of endurance training in specific areas of Th10 and L2 segments of spinal cord. Microphotographs also show the overlapping of these factors with GABA (C,D,I; green) and Glyt2 (L; green) inhibitory neurons (white arrows) (n = 3). Scale bars: (A–L), 100 µm.

Figure 4

Immunoexpression of BDNF (A,C,D,F; red), GFRα (G,I,J,L; red), GABA (B,C,E,F; green) and Glyt2 (H,I,K,L; green) neurons in nontrained (SCI) and pretrained (T + SCI) groups with Th9 compression. Differences in the experimental groups are shown in the intermediate zone of Th10 segment (just caudally from the lesion site). White arrows indicate the overlapping of BDNF with GABA, GFRα with Glyt2 and DAPI, heads show GABA- and GFRα-immunoreactive puncta (n = 3). Scale bars: (A–L), 100 µm.

Figure 5

Immunohistochemical, relative gene expression, and protein level analyses of oligodendrocytes in thoracic spinal cord. Fluorescent APC staining of mature oligodendrocytes colocalized with DAPI in SCI and T + SCI groups in caudal segment (−1) (n = 3) (A). Levels of ΔΔCt for immature CNP-ase oligodendrocytes and the ratio of CNP-ase/βactin indicate the effect of endurance training on intact vs. traumatized tissue (B). Levels of ΔΔCt for mature PLP1 oligodendrocytes and the ratio of PLP1/βactin indicate the effect of endurance training on intact vs. traumatized tissue (C). White arrows show colocalization of APC with DAPI. Results are presented as mean ± SD (RT-PCR: n = 4, WB: n = 5). Data were statistically evaluated using one-way ANOVA and post hoc Tukey’s HSD test; * p < 0.05; ** p < 0.01; *** p < 0.001. Scale bar: 100 µm.

Figure 6

Gene expression, protein level analysis, and fluorescent labeling of light chains of neurofilaments (NF-l) and newly outgrowing fibers (Gap-43) in experimental groups. Levels of ΔΔCt for NF-l, Gap-43, and the ratio of NF-l/βactin and Gap-43/βactin show the effect of endurance training on intact vs. traumatized spinal cord tissue (A,B) (RT-PCR: n = 4, WB: n = 5). Microphotographs show the staining of NF-l (green) and Gap-43 (red) in SCI and T + SCI groups in caudal segment (−1) (C) (n = 3). Red arrows indicate the magnification of representative dashed white squares, showing colocalization of NF-1 with Gap-43. Results are presented as mean ± SD. Data were statistically evaluated using one-way ANOVA and post hoc Tukey’s HSD test; * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Scale bar: 100 µm.

Figure 7

Neurological Basso–Beatie–Bresnahan (BBB) score showing locomotor activity of experimental animals in intact control (IC), SCI and T + SCI groups. Scoring ranges from 0 points (complete paraplegia) to 21 points (normal hindlimb movements). Results are presented as mean ± SD (n = 9). Data were statistically evaluated using repeated measures of variance (repeated one-way ANOVA multiple comparisons); experimental groups vs IC (^# p < 0.05); SCI vs T + SCI (* p < 0.05).

Figure 8

Values of peak latency (ms) recorded from musculus gastrocnemius prior to spinal cord injury (0 weeks) and 2, 4, and 6 weeks after trauma in the SCI and T + SCI groups. Results are presented as mean ± SD (n = 4). Data were statistically evaluated using Student’s parametric t-test; ** p < 0.01.

Figure 9

Linear regression analysis of correlation between protein levels for CNP-ase (A,B), PLP1 (C,D), and BBB neurological score (A–D) in the epicenter of injury and adjacent segments in SCI and T + SCI groups. Scatterplots of individual values with regression line correlation coefficient (r) and coefficients of determination (R²) calculated using regression analysis and p-Value (n = 5).

Figure 10

Linear regression analysis of correlation between gene expression for CNP-ase (A,B), PLP1 (C,D), and BBB neurological score (A–D) in the epicenter of injury and adjacent segments in SCI and T + SCI groups. Scatterplots of individual values with regression line correlation coefficient (r) and coefficients of determination (R²) calculated using regression analysis and p-Value (n = 4).

Figure 11

Linear regression analysis of correlation between protein levels for NF-l (A,B), Gap-43 (C,D), and BBB neurological score (A–D) at the lesion site, cranial, and caudal segments in Th9 compression groups (SCI and T + SCI). Scatterplots of individual values with regression line correlation coefficient (r) and coefficients of determination (R²) calculated using regression analysis and p-Value (n = 5).

Figure 12

Linear regression analysis of correlation between gene expression for NF-l (A,B), Gap-43 (C,D), and BBB neurological score (A–D) at the lesion site, cranial, and caudal segments in Th9 compression groups (SCI and T + SCI). Scatterplots of individual values with regression line correlation coefficient (r) and coefficients of determination (R²) calculated using regression analysis and p-Value (n = 4).

Figure 13

Schematic illustration of the spinal cord showing localization of three areas of interest (dorsal horns, intermediate zone, and ventral horns) for quantitative analysis of GAD 65- and Glyt2-positive cells. Measuring frame size in each area was 1.1 × 1 mm.