Role of spared pathways in locomotor recovery after body-weight-supported treadmill training in contused rats

Abstract

Body-weight-supported treadmill training (BWSTT)-related locomotor recovery has been shown in spinalized animals. Only a few animal studies have demonstrated locomotor recovery after BWSTT in an incomplete spinal cord injury (SCI) model, such as contusion injury. The contribution of spared descending pathways after BWSTT to behavioral recovery is unclear. Our goal was to evaluate locomotor recovery in contused rats after BWSTT, and to study the role of spared pathways in spinal plasticity after BWSTT. Forty-eight rats received a contusion, a transection, or a contusion followed at 9 weeks by a second transection injury. Half of the animals in the three injury groups were given BWSTT for up to 8 weeks. Kinematics and the Basso-Beattie-Bresnahan (BBB) test assessed behavioral improvements. Changes in Hoffmann-reflex (H-reflex) rate depression property, soleus muscle mass, and sprouting of primary afferent fibers were also evaluated. BWSTT-contused animals showed accelerated locomotor recovery, improved H-reflex properties, reduced muscle atrophy, and decreased sprouting of small caliber afferent fibers. BBB scores were not improved by BWSTT. Untrained contused rats that received a transection exhibited a decrease in kinematic parameters immediately after the transection; in contrast, trained contused rats did not show an immediate decrease in kinematic parameters after transection. This suggests that BWSTT with spared descending pathways leads to neuroplasticity at the lumbar spinal level that is capable of maintaining locomotor activity. Discontinuing training after the transection in the trained contused rats abolished the improved kinematics within 2 weeks and led to a reversal of the improved H-reflex response, increased muscle atrophy, and an increase in primary afferent fiber sprouting. Thus continued training may be required for maintenance of the recovery. Transected animals had no effect of BWSTT, indicating that in the absence of spared pathways this training paradigm did not improve function.

FIG. 1.

Study design. Electrophysiological testing and histological analyses were performed at the conclusion of the experiments (Cntl, control; Ctsn, contusion; Tx, transection; IHC, immunohistochemistry; CGRP, calcitonin gene-related peptide; BWSTT, body-weight-supported treadmill training).

FIG. 2.

Robotic device used for body-weight-supported treadmill training (BWSTT) and for locomotor assessment. The animals were trained and tested on the treadmill with 75% body weight support (BWS) at a speed of 7 cm/sec. The robotic arm was attached to the ankle using a neoprene cuff as shown in the photo. These arms moved the ankles in a semi-circular step cycle trajectory during training, and recorded the ankle trajectories during testing.

FIG. 3.

Open-field locomotion by Basso-Beattie-Bresnahan (BBB) testing. Transected animals had significantly lower BBB scores than contused rats, and no effect of BWSTT was observed for either injury group (Cntl, control; Ctsn, contusion; Tx, transection; BWSTT, body-weight-supported treadmill training).

FIG. 4.

Mean±standard error of mean of swing duration (A), step height (B), and step length (C; *=significant difference from controls; #=significant difference between the two groups [p<0.05]; ^=significant difference from week 9). Body-weight-supported treadmill training (BWSTT) resulted in a significant difference in the trained versus untrained contused animals at week 5, but not at week 9. After Tx injury, at week 10 trained animals retained their locomotor performance and were significantly different from untrained animals, but not at week 11, except for step height.

FIG. 5.

Mean±standard error of mean of the ratio of H-reflex (H_10Hz/H_0.3Hz×100; #=significant difference from untrained animals with similar injury, p<0.05 by analysis of variance). Body-weight-supported treadmill training (BWSTT) led to a significant improvement in the rate depression property of the H-reflex in animals with only contusion or transection injury (Cntl, control; Ctsn, contusion; Tx, transection).

FIG. 6.

Superimposed averages of stimulation frequencies of 0.3 Hz and 10 Hz for M-wave and H-wave. One representative animal from each of the seven groups is shown. Uninjured control animals showed 100% rate depression of the H-wave at 10 Hz (no H peak was observed). Injured animals exhibited decreased sensitivity to this rate depression property. Body-weight-supported treadmill training (BWSTT) helped restore this property in trained animals with contusion or transection injury (Cntl, control; Ctsn, contusion; Tx, transection).

FIG. 7.

Mean±standard error of mean of the soleus muscle weight/body weight (*=significant difference from control, p<0.05 by analysis of variance; #=significant differences between groups with similar injuries; Cntl, control; Ctsn, contusion; Tx, transection; BWSTT, body-weight-supported treadmill training).

FIG. 8.

Calcitonin gene-related peptide (CGRP) immunostaining in the dorsal horn of the L4–L6 lumbar spinal cord. CGRP-immunoreactive (IR) axons were localized to laminae I and II of the dorsal horn in control animals. In all injured animals, the CGRP-IR was significantly denser (p<0.05 by analysis of variance [ANOVA]) compared to controls. However, body-weight-supported treadmill training (BWSTT) in contused animals (Ctsn + BWSTT) appeared to diminish CGRP-IR (p<0.05 by ANOVA) when compared to untrained animals with similar injuries (Ctsn; Cntl, control; Tx, transection).

FIG. 9.

Mean±standard error of mean of the calcitonin gene-related peptide (CGRP) density (mean gray value) in the dorsal horn of the L4–L6 spinal cord (#=significant difference between groups with similar spinal cord injuries, p<0.05; Cntl, control; Ctsn, contusion; Tx, transection; BWSTT, body-weight-supported treadmill training).