DOI: https://doi.org/10.21203/rs.3.rs-1727116/v1
Treadmill exercise is beneficial for spinal cord injury (SCI) both in bench and in bedside, but the most effective treadmill exercise intensity and mechanisms underlying the treadmill exercise on skilled motor function recovery remain elusive. Here, the improved skilled motor function recovery, enhanced nerve conduction capability, neuroplasticity, and axonal sprouting were observed in the SCI mice after training for 4 weeks. However, high exercise intensity (HEI) leads to vulnerability and impaired exercise tolerance during training. We further found that in the moderate exercise intensity (MEI) and HEI groups showed elevated expression of brain-derived neurotrophic factor (BDNF) and insulin-like growth factor 1 (IGF-1). Meanwhile, elevated phosphorylated levels of ribosomal protein S6 (p-S6) and protein kinase B (p-AKT) in mouse motor cortex were also observed, indicating the cortical mechanistic target of rapamycin (mTOR) pathway activation. To investigate the role of the cortical mTOR activation, we performed a rapamycin assay. After using rapamycin, the exercise-induced activation of the cortical mTOR pathway and the exercise-enhanced effects were inhibited. Together, the expression of neurotrophic factors and the activation of the cortical mTOR pathway are in an intensity-dependent manner. And the MEI is safer and more beneficial than the LEI and HEI. Based on the rapamycin assay, the exercise-induced activation of mTOR pathway is necessary for the enhanced motor cortex and spine remodeling, all of which further contribute to better-skilled motor function recovery. These dates may provide a new window to further understand the mechanisms underlying exercise training effects on the skilled motor function recovery following SCI.
Spinal cord injury (SCI) is a devastating central nervous system (CNS) condition associated with permanent motor and sensory deficits. The overall global incidence of SCI was 10.5 cases per 100,000 persons, resulting in an estimated 800,000 new cases annually worldwide[1]. Therefore, developing effective methods to promote functional recovery and exploring its underlying mechanisms are imperative.
Treadmill exercise is a simple, easy, and common therapy in SCI rehabilitation. Accumulating evidence demonstrated that there is a neuroprotective and nerve repair effect of the treadmill exercise on CNS injuries such as SCI, stroke, and traumatic brain injury (TBI) both in animals and humans[2–8]. Previous studies found that the treadmill exercise potentiates overall functional recovery after SCI through reduced neuronal apoptosis and syringomyelia area, promoting secretion of neurotrophic factors and enhanced neuroplasticity[9–12]. However, systematic research on the treadmill exercise intensity and mechanisms underlying the treadmill exercise on neuroprotection and nerve repair are poorly understood.
The exploration of the treadmill exercise intensity hindered might be because most SCI animal models induced complete paralysis of the hindlimbs which prevents animals from running. However, evidence showed that corticospinal tract (CST) lesion is less important for locomotor function in rodent animals[13]. Work by Tetzlaff et al. and Kanagal et al. suggested that dorsolateral funiculus crush of mouse cervical spinal cord results in sustained forelimbs skilled behavior deficits but limited hindlimbs locomotor deficits[14, 15]. Therefore, exploring the effects of different treadmill exercise intensities on cervical SCI mice becomes reality.
Recently, Chen et al. found that the treadmill exercise at a speed of 12 m/min activates the mechanistic target of rapamycin (mTOR) pathway in mouse motor cortex, leading to enhanced neuronal activity of layer 5 pyramidal neurons and increased dendritic spine formation and motor learning in normal adult mice[16]. Therefore, in the current study, we plan to 1) explore what treadmill exercise intensity is best appropriate for the cervical SCI mice, and 2) investigate whether the effects of neuroprotection and nerve repair are related to the exercise-induced activation of cortical mTOR pathway.
Female C57BL/6J mice (6–8 weeks) were purchased from Hunan SJA Laboratory Animal Co., Ltd. They were performed cervical 5 crush injury and were used for the treadmill exercise, electrophysiology method, Golgi staining, Western blotting, immunofluorescence staining and some behavioral assays. Animals were group-housed under normal light-dark cycle with food and water ad libitum. All animal experimental procedures were approved by the Animal Experimental Ethics Committee of Chongqing Medical University.
Two experiments were involved in this study. For the experiment I, animals were assigned randomly into four groups including Control group (n = 15, sedentary ), Low Exercise Intensity group (LEI, n = 15), Moderate Exercise Intensity group (MEI, n = 15) and High Exercise Intensity group (HEI, n = 15) (30%, 50% and 70% of maximum speed, respectively[17]). For the experiment II, animals were divided into four groups including Spine Cord Injury group (SCI, n = 15), Spine Cord Injury combined with Treadmill Exercise group (SCI + E, n = 15), Spine Cord Injury combined with Rapamycin group (SCI + R, n = 15) and Spine Cord Injury combined with Treadmill Exercise and Rapamycin group (SCI + E + R, n = 15). The SCI + E and SCI + R + E groups were performed the treadmill exercise at moderate exercise intensity. And in experiment II, the rapamycin administration (3 mg/kg, Solarbio, China) was performed every 3 days during the treadmill exercise in the SCI + R and SCI + E + R groups (Fig. 7a).
A six channels treadmill apparatus (KW-PT, KEW, China) was adopted. Animals were initially adapted to the treadmill exercise for 3 days (9 m/min, 20 mins daily). After that, a maximum exercise speed test was performed with a 5-minute warm-up (9 m/min) followed by an increase in treadmill speed (1 m/min every 1 min) until animal exhaustion, that is, until they were not able to run even after 10 electric stimuli[18, 19]. To follow mouse body condition changes at various stages, the maximum exercise speed test was performed on the first day of every training session in the trained mice. All trained groups performed the treadmill exercise for 4 weeks (30 min/day, 5 days/week). The Control, SCI, and SCI + R groups were only put onto the treadmill device for the same time and same environment. After training for 4 weeks, all animals were sacrificed for further assays (Fig. 1a; Fig. 7a).
The cervical crush injury was performed as previously described[20].After 3 days pre-training, animals underwent cervical 5 (C5) crush injury (Fig. 1a). The spinal cord was exposed by laminectomy at C5 level. The dorsal root entry zone was pierced with a syringe needle (26-gauge needle). A modified fine tipped Dumont #5 forcep (tip width ~ 200 µm, length ~ 2 mm; Dumostar Biology) was used in the surgery. Two prongs of the forcep were inserted into the bilateral dorsal horn gray matter to a depth of ~ 1 mm[14]. The forcep was closed and held for 15 s, and this process was repeated once more. Next, one prong of the forceps was inserted into the left dorsal horn gray matter to a depth of ~ 0.8 mm, whereas the other prong remained outside[14]. The forceps was also closed and held for 15 s and repeated this process (Fig. 1b). After surgery, the muscle and skin were sutured. Finally, different from before, a retrograde trace assay was used to examine whether the surgery was successful or failed. Any animal that showed positive Mini Ruby staining in neurons in motor cortex was removed from the study.
5% Mini Ruby, an anterograde tracer (Thermo, USA), was injected to retrogradely label neurons in motor cortex[21]. After anesthetizing, the micropipette was used to inject Mini Ruby (0.2 µl/site×4 sites, 0.05 µl/min) into left and side (lateral: 1.1 mm; depth: 0.5 mm) of the midline of spinal cord between C6 and C7 level, respectively (Fig. 1c). If surgery was successful, little neurons stained by Mini Ruby could be observed in the motor cortex of the injured mice.
Some mice were injected with biotinylated dextran amines (BDA; Thermo, USA) two weeks before sacrifice. The mice were anesthetized by intraperitoneally injecting 0.5% sodium pentobarbital (50 mg/kg; Shanghai Pharmaceutical Factory). The skin overlying the skull was disinfected and shaved. The right motor cortex area was exposed by using a dental drill. A Hamilton syringe (Hamilton, Switzerland) connected with a glass-pulled micropipette was used to inject stereotaxically. 2.4 µl 10% BDA was slowly injected into the right motor cortex (bregma: -0.5, -1 mm; lateral: 0.8, 1.2 mm; depth: 0.7 ~ 0.9 mm [22, 16, 23]) 0.6 µl/site × 4 sites; 0.2 µl/min. After injecting, the micropipette was retained for 3 mins before retraction, and then retracted slowly (0.2µm/min). The mouse head skin was sutured and disinfected, and then the mouse was returned to home cage.
One day before sacrifice, mice were anesthetized by using 0.5% pentobarbital sodium (50 mg/kg), and the hair overlaying head skin and around the left forelimb was shaved. After local disinfection and incision of head skin, the primary motor cortex was exposed by using a micro-dental drill. After that, the mice were immobilized on a stereotaxic instrument. The pia meninge was lightly touched by a stimulating electrode, and the recording electrodes were inserted into mouse left forelimb muscle. Meanwhile, the reference electrode was clamped near the scalp incision, and the ground electrode was inserted into mouse tail. An evoked potential machine (Powelab/16SP, ADInstruments, Inc. USA) was used to record the motor evoke potential (MEP). The stimulation frequency was 4 Hz, the wave width was 0.2 ms, and the stimulating intensity was 5 mA. Each test was repeated at least 3 times to obtain steady waveforms (specially, N1 and P1 waves), and the latency of each wave peak was recorded.
Mice were killed by cervical dislocation and their brains were washed in 0.9% saline. The FD Rapid GolgiStain Kit was used according to the specifications of the manufacturer (FD NeuroTechnologies, USA). Briefly, whole brain was immersed in mixed solution (solution A and B), and stored at room temperature for 2 weeks in the dark. Then, the brain was transferred into solution C, and stored at room temperature in the dark for at least 72 hours, after that, sectioned at 100 µm on a freezing microtome. After sectioning and mounting on gelatin-coated slides, staining was performed. After staining, the complexity of neurons was analyzed by Sholl analysis[24, 25], the total dendrites length of neurons was calculated by using Simple Neurite Tracer (SNT) method from ImageJ. For analysis of dendritic spine density, 1 ~ 3 neurons in the right motor cortex per brain slices were randomly chosen, and then the dendrites of ~ 100µm length from these neurons were randomly chosen to measure. The number of dendritic spines was counted using a hand counter. The dendritic spine density was expressed as the number of dendritic spines/dendritic length (µm)[26].
Longitudinal sections through the lesion were stained for BDA and GFAP to identify lesion edges and observe the condition of axonal regeneration.
Transverse sections from the area of 0 ~ 500 µm rostral to lesion site were immunofluorescent stained by BDA to observe the axonal sprouting in ipsilateral side. Based on the previous methods of analyzing the axonal sprouting[27, 28], the spinal cord was divided into several circular regions. The dorsal corticospinal tract (dCST) region in spinal dorsal horn of the ipsilateral spinal cord was taken as the center of the circle, and then different concentric circles were drawn outwards with a constant radius distance difference (Δ radius = 100 µm). The number of intersections between the different concentric circles (r = 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 and 1100 µm) and the axonal sprouting were counted (Fig. 4c). The number of intersections at different radius distances was used to represent the elongation length of axonal sprouting.
Tissue lysates were extracted from mouse primary motor cortex in RAPA containing protease and phosphatase inhibitors (Beyotime, China). After quantification by using bicinchoninic acid kit (BCA, Thermo, USA), 30 ~ 50 µg of protein samples were separated by SDS–PAGE and were transferred to 0.22 µm polyvinylidene fluoride (PVDF) membrane (Absin, China). After washing in 1X tris-buffered saline and Tween-20 (TBST), and blocking in 5% skim milk, the membrane was incubated with the primary antibody at 4°C overnight. The primary antibodies used were rabbit anti-S6 (Cell Signaling, USA, 1:1000), rabbit anti-p-S6 (Cell Signaling, USA, 1:1000), rabbit anti-Akt (Cell Signaling, USA, 1:1000), rabbit anti-p-Akt (Cell Signaling, USA, 1:1000), rabbit anti-IGF-1 (Cell Signaling, USA, 1:1000), rabbit anti- IGF1-R (Cell Signaling, USA, 1:1000), rabbit anti-p-IGF1-R (Cell Signaling, USA, 1:1000), rabbit anti-BDNF (Cell Signaling, USA, 1:500), rabbit anti-p-Trkb (Cell Signaling, USA, 1:1000), mouse anti-β-actin (Proteintech, China, 1:2000). After that, the membrane was washed and incubated in secondary antibody for 1 hour. The secondary antibodies used were Goat anti-rabbit and Goat anti-mouse (Servicebio, China, 1:3000). Protein bands were visualized by using an imaging system and machine (Tanon, China). Integrated gray values of each band was measured by using ImageJ (National Institutes of Health, Bethesda, USA).
Mice were perfused with 0.01M phosphate-buffered saline (PBS) followed by 4% paraformaldehyde (PFA). The spinal cord and brain were dissected, then post-fixed in 4% PFA for 24 hours, after that, dehydrated gradually in different concentration sucrose solutions (18%; 24%; 30%) for 3 days. Finally, embedded in an optimal cutting temperature compound (OCT; Tissue-Tek Sakura) and frozen at -80℃. The spinal cord and brain were sectioned transversely into 12 µm and 18µm slices by using a freezing microtome (Thermo, USA), respectively. All tissue sections were then stored at -80℃.
Tissue sections were taken from − 80℃ and thawed for 2 hours at room temperature. After washing in PBS (0.01 M, pH 7.4), incubating in 0.03% Triton X-100 and blocking in QuickBlock immunostaining block solution (Beyotime, China), the slices were incubated with primary antibodies containing rabbit anti-p-S6 (Cell Signaling, USA, 1:200), Guinea pig anti-NeuN (Millipore, Germany, 1:800) and rabbit anti-GFAP (Thermo, USA, 1:2000) overnight at room temperature, followed by secondary antibodies for 45 min at 37℃. For BDA staining, Alexa-594 nm Streptavidin (Thermo, USA, 1:300) was used. Finally, sections were sealed in glycerol, followed by observation under Nikon A1R laser confocal microscopy (Nikon, Japan).
All behavioral analyses were conducted by personnel blind to the group inclusion.
Forelimb preference to support their weight against the wall was assessed by placing mice in a plexiglass cylinder (Fig. 2d; height: 20 cm; diameter: 15 cm) for 15 mins. Two mirrors were placed behind the cylinder at an angle so that any forepaw motion could be observed[14]. The whole process was videotaped by using a smartphone (iphone 6, USA). 10 initial placements were recorded. Initial placement was classified as either “left” or “right” when the left or right forepaw was used to support its own body weight > 0.25 s independently. If mice used both forepaws together to support themselves > 0.25 s, this situation should be classified as “both”. The ratios of initial left forepaw use (left/total initial placements) and both forepaws use (both/total initial placements) at different time points were analyzed.
Error (slip, miss or drag) percent for the left limbs was evaluated by using a horizontal ladder, where mice crossed a 5.2 cm wide, 30 cm high ladder with 31 rungs each spaced 1.3 cm apart (Fig. 2a). Each mouse repeated the task 5 times at one time point. Repetitions that mice turned around halfway and backed to starting line were excluded. To avoid the mice learning and q123456rd adapting, 5 rungs were removed randomly at every time point. The whole process was videotaped with a high-definition camera (ONTOP X2S, China). In the end, the number of errors and total steps of the left limbs during 5 separate runs were counted[14]. Finally, a ratio (errors/total steps) was used to represent the error percentage of left limbs.
All data were presented as mean ± SEM. One-way or two-way analysis of variance (ANOVA) with Bonferroni’s post-hoc analysis were used to compare differences among multiple groups, respectively. All statistical analysis was performed by GraphPad Prism 8.0 (La Jolla, CA, USA). p < 0.05 was defined as statistically significant.
Establishment of the C5 spinal cord crush injury model.
We first examined whether C5 crush surgery could specifically injure the dorsal corticospinal tract (dCST) in adult mice. As shown (Fig. 1d), compared with the normal mice, only little neurons labeled Mini Ruby was observed in the right motor cortex of the injured mice, suggesting that the dCST could be crushed specifically and completely by the C5 crush surgery, and indicating the successful establishment of the C5 spinal cord crush injury model.
Treadmill exercise promotes skilled motor function recovery and enhances nerve conduction capability after SCI.
We next investigated whether the recovery of skilled motor function was enhanced by treadmill exercise after injury. Date showed all mice performed worse in the behavioral tests at 3 days and 1 week after injury (Fig. 2a-f). After the treadmill exercise for 2 and 4 weeks, we found that the mice in the LEI, MEI and HEI groups showed lower error rate of left limbs in the horizontal ladder test (P < 0.05, Fig. 2b-c) and higher usage rate of left forelimb and both forelimbs in the cylinder rearing test (P < 0.05, Fig. 2e-f) when compared with the control group. More importantly, mice in the MEI and HEI groups performed slightly better than the LEI group, even though, there was no significant difference among three trained groups.
Meanwhile, we asked whether the nerve conduction capability was also enhanced by performing the electrophysiology method to test the motor evoked potential (MEP) in mouse right motor cortex (Fig. 2g-i). Obviously, results showed that the latency of N1 wave (LEI, MEI, HEI vs. control, 6.3 ± 0.2, 6.4 ± 0.2, 6.3 ± 0.2 vs. 6.7 ± 0.4, P < 0.0001) and P1 wave (LEI, MEI, HEI vs. control, 7.2 ± 0.3, 7.1 ± 0.5, 7.2 ± 0.3 vs. 8.0 ± 1.1, P < 0.0001) in the LEI, MEI and HEI groups was significantly shorter than that in the control group, suggesting that the nerve conduction capability was potentiated by the treadmill exercise.
Exercise-induced enhancement of neuroplasticity in motor cortex is necessary for skilled motor function recovery.
To verify effects of the treadmill exercise on neuroplasticity, we then measured the alterations of dendrite complexity, spine density and the morphology in the right motor cortex using Golgi staining. Results showed that the neurite arborization significantly increased after training (Fig. 3a-b). The total length of dendrites was longer (LEI, MEI, HEI vs. control, 2160.45 ± 399.65, 2210.34 ± 388.12, 2109.99 ± 559.94 vs. 690.80 ± 116.92, P < 0.0001; Fig. 3a-c), and the density of dendritic spines was higher (LEI, MEI, HEI vs. control, 0.13 ± 0.02, 0.11 ± 0.01, 0.17 ± 0.03 vs. 0.02 ± 0.01, P < 0.0001; Fig. 3d-e) in the mouse right motor cortex of the LEI, MEI and HEI groups compared with the control group, suggesting that the complexity of neurons was enhanced by the treadmill exercise. Besides, a high correlation was found between the exercise-enhanced cortical neuroplasticity and the mouse performance in the horizontal ladder test (Fig. 3f-i). The longer the total length of cortical dendrites and the higher the density of dendritic spines, the lower the error rate of left limbs in the horizontal ladder test.
Treadmill exercise alone fails to promote axonal regeneration beyond lesion site but enhances axonal sprouting in ipsilateral side after SCI.
We next explored the effects of the treadmill exercise on axonal regeneration and sprouting. As shown (Fig. 4a), the treadmill exercise alone was not enough to promote axonal regeneration beyond the lesion site. But, the enhancement of axonal sprouting in ipsilateral gray matter region 0 ~ 500 µm rostral to the lesion site was observed in the LEI, MEI and HEI groups after 4 weeks treadmill exercise (P < 0.05; Fig. 4b-d). Some of the new axonal sproutings could even elongate to the ipsilateral ventricornu, a dense region of motor neurons, about 1,000 µm far from the spinal dorsal horn region, where dCST converged. It is may be one of the structural basis for the exercise-induced promotion of functional recovery in the cervical SCI mice.
Treadmill exercise facilitates neurotrophic factors expression in mouse motor cortex after SCI.
We further investigated the changes of molecular pathway in the cervical SCI mice after the treadmill exercise. As shown (Fig. 5), BDNF, IGF-1, phosphorylated tyrosine kinase receptor B (p-TrkB) and insulin-like growth factor 1 receptor (p-IGF1-R) were up-regulated in the mouse right motor cortex after chronic treadmill exercise in the MEI and HEI groups compared with the control and LEI groups (P < 0.05). And there was no significant difference between the control group and the LEI group. These results suggested that the upregulated neurotrophic factors are intensity-dependent, and are related to the exercise-enhanced neuroprotective and nerve repair effect.
Treadmill exercise activates cortical mTOR pathway and this exercise-induced activation can be inhibited by using rapamycin.
As the co-downstream of TrkB and IGF1-R, mTOR pathway plays an important role in modulating axonal regeneration and sprouting, plasticity and so on. Therefore, we further explored the effects of treadmill exercise on the activation of cortical mTOR pathway. Consistent with the changes of neurotrophic factors (Fig. 5), elevated phosphorylated levels of ribosomal protein S6 (p-S6) and protein kinase B (p-AKT) were only observed in the MEI and HEI groups (P < 0.05; Fig. 6a-d), indicating the activation of cortical mTOR pathway by the moderate and high intensity treadmill exercise. And as further evidence, elevated neuronal p-S6 in the mouse right motor cortex (Fig. 6f) was also only found in the MEI and HEI groups by immunofluorescence staining (P < 0.05; Fig. 6e-g).
According to the facts that the activation of cortical mTOR pathway was not observed in the LEI group (Fig. 6) and the HEI group showed decreased exercise tolerance and underwent the most total electric shocks during the treadmill exercise (P < 0.05; Fig. 7b-c), all of which leads to vulnerability. We therefore adopted the moderate intensity treadmill exercise in following experiment. To investigate the role of mTOR activation, a mTOR pathway inhibitor, rapamycin, was used. As expected, compared with the SCI + E group, the levels of p-S6 and p-AKT decreased after intraperitoneally injecting rapamycin in mice (P < 0.05; Fig. 7d-g). And a similar result was also validated by the immunofluorescence staining (P < 0.05; Fig. 7h-i).
Exercise-induced promotion of skilled motor function recovery and enhancement of nerve conduction capability after SCI are associated with the activation of mTOR pathway.
More importantly, rapamycin assay further demonstrated that the exercise-induced activation of cortical mTOR pathway is necessary for the exercise-enhanced functional recovery (Fig. 8). Compared with the SCI + E group, the exercise-induced skilled motor function recovery was attenuated in the SCI + E + R group as indicated by lower percentage of performance improvements in the horizontal ladder test (P < 0.05; Fig. 8a-b) and the cylinder rearing test (P < 0.05; Fig. 8c-d). And the exercise-enhanced nerve conduction capability was also eliminated by rapamycin, as proved by the results that the latency of N1 (SCI + E + R vs. SCI + E, 7.2 ± 0.2 vs. 6.5 ± 0.4, P < 0.0001; Fig. 8f) and P1 (SCI + E + R vs. SCI + E, 8.8 ± 0.4 vs. 7.8 ± 0.2, P < 0.0001; Fig. 8g) waves became longer in the SCI + E + R group compared with the SCI + E group.
Enhanced neuroplasticity in motor cortex and potentiated axonal sprouting are associated with exercise-induced activation of mTOR pathway.
We also found that the changes of cortical neuroplasticity are associated with the exercise-induced activation of cortical mTOR pathway (Fig. 9). Compared with the SCI + E group, the exercised-enhanced formation of dendrites (SCI + E + R vs. SCI + E, 735.88 ± 24.43 vs. 1827.29 ± 457.98, P < 0.0001; Fig. 9a-c) and dendritic spines (SCI + E + R vs. SCI + E, 0.03 ± 0.01 vs. 0.12 ± 0.02, P < 0.0001; Fig. 9d-e) were also attenuated after injecting rapamycin in the SCI + E + R group. Consistent with those results, the exercise-induced recovery of skilled motor function was removed as well (Fig. 8). What’s more, compared with the SCI + E group, the exercise-induced enhancement of axonal sprouting was attenuated by the injection of rapamycin in the SCI + E + R group as well (P < 0.05; Fig. 10).
Together, the expression of neurotrophic factors and the activation of cortical mTOR pathway in an intensity dependent manner. Activation of the cortical mTOR pathway induced by the treadmill exercise plays an important role in motor cortex and spine remodeling, including the enhanced complexity of neurons, potentiated nerve conduction capability, and increased axonal sprouting in the cervical SCI mice, all of which further contribute to better motor function recovery.
The current study showed the effects of different intensities treadmill exercise on skilled motor function recovery, nerve conduction capability, brain and spine plasticity, expression of neurotrophic factors in the cervical SCI mice. We found the cortical mTOR pathway can be activated via chronic moderate or high intensity treadmill exercise (50% or 70% of maximum speed, respectively). Furthermore, the moderate intensity is more beneficial and more tolerable for SCI mice during treadmill exercise compared with the low and high intensity. And we further demonstrated that the exercise-induced activation of cortical mTOR pathway is necessary for the nerve repair and neuroprotective effect. To our knowledge, it is the first systematic study about the effects of different intensities treadmill exercise on functional recovery and the exploration of molecular mechanism after SCI.
Various exercise training approaches such as enriched environments, single pellet reaching task, swimming, bicycling, treadmill exercise and so on, have been shown to improve functional recovery after SCI[29–32]. In particular, originated from the discovery that spinalized cats could be trained on a treadmill to recover weight-supported stepping of the hindlimbs[33, 34], the treadmill exercise is one of the most common methods to provide motor function rehabilitation following SCI. Ward et al. found that as little as 30 minutes of step training 6 days per week enhanced overground locomotion in rats with contusive SCI[35]. Work by Wang and Jung et al. suggested that the treadmill exercise improved locomotor function in SCI rats, assessed by Basso-Beattie-Bresnahan (BBB) scale and performance in grid walking test[12, 36]. Shibata et al. found that approximate half of the treadmill exercise trained mice exhibited improved motor function based on the Basso Mouse Scale (BMS) scores, while none of the control group mice exhibited improved motor function[37]. Similarly, we also observed the error rate of mouse left limbs decreased, assessed by the horizontal ladder test. Moreover, the treadmill exercise improved the willing of participating in climbing cylinder, as proved by the increased usage rate of the injured forepaw. Furthermore, the enhanced dendritic complexity, increased dendritic spine density, potentiated axonal sprouting and motor nerve conductive function were also observed in the trained mice, compared with the control group, which may lay the structural basis for motor function recovery.
Considering the previous work, the effects of neuroprotection and nerve repair depend on the treadmill exercise are related to exercise intensity. Ploughman et al. reported that frequent low intensity exercise (as in voluntary running wheels) may be safer for rats with stroke and has a delayed but sustained effect on BDNF that may support brain remodeling[38]. And Lou et al. found one week of low (at a speed of 5 m/min for the first 5 mins, 8 m/min for the next 5 mins, and 11 m/min for the remaining 20 mins) or moderate (at a speed of 8 m/min for the first 5 mins, 11 m/min for the next 5 mins, and 14 m/min for the remaining 20 mins) intensity treadmill exercise enhanced neurogenesis of hippocampus in juvenile rats[39]. Rahmati et al. suggested that mild exercise intensity (30%~50% of maximum speed) is the optimal intensity for improving hippocampal neuroplasticity in adult rats[18]. But other studies indicated that higher intensity are more beneficial for increasing the expression of BDNF and IGF-1[10, 40], which is consistent with our results in Fig. 5. As far as we know, no studies have adequately adjusted and applied the intensity of treadmill exercise to the cervical SCI mice according to the change of weekly maximum exercise speed. From some studies, the exercise intensity is based on the running speed[41], the session duration[42] or the number of repetitions[43]. In view of the considerable variations in running speed and mouse condition[42, 44], it is important to adjust the exercise intensity, based on the weekly maximum exercise speed test, in accordance with the exercise tolerance of each group to provide further functional recovery benefits. Similar to Hager et al.[45], we observed the impaired exercise tolerance and the increased probability to be injured during the treadmill exercise in the HEI group as well. In addition, the activation of cortical mTOR pathway was observed in the MEI and HEI groups except the control and LEI groups. Together, compared with the low or high exercise intensities, the moderate exercise intensity is safer and more beneficial for the cervical SCI mice.
Accumulating evidence has demonstrated that the exercise-enhanced effects of neuroprotection and nerve repair may be related to the up-regulation of neurotrophic factors[46, 47]. Both human and animal studies have shown that BDNF and IGF-1 are involved in cortical reorganization by exercise training[48, 49]. Consistent with those studied, we found the expression of BDNF, IGF-1, p-TrkB and p-IGF1-R were up-regulated through the moderate or high intensity treadmill exercise. To be more specific, BDNF binds with TrkB, which results in the formation of p-TrkB, further inducing the activation of downstream pathways including PI3K-AKT-mTOR pathway, MAPK/Ras pathway and PLC-γ signaling cascade[50]. Similarly, IGF-1 binds with IGF1-R leading to increased p-IGF1-R, which leads to the activation of PI3K-AKT-mTOR and MAPK/ERK pathways[51]. These signaling pathways are critical for neuronal survival, apoptosis, brain and spine plasticity, axon regeneration and sprouting.
At the downstream of TrkB and IGF1-R, the mTOR pathway, which widely involves in mediating the exercise-associated health benefits [52], can be potentially activated by BDNF and IGF-1[53, 54]. Recently, Chen et al. found that chronic treadmill exercise (12 m/min, 1 hour daily, for 3 weeks) activates the mTOR pathway, which is necessary for spinogenesis, neuronal activation, and axonal myelination leading to improved motor learning in adult mice[16]. Here, we found that 4 weeks moderate or high intensity treadmill exercise also activates the cortical mTOR pathway in the cervical SCI mice, which is essential for the exercise-enhanced nerve conduction capability, neuroplasticity and motor function recovery, as proved by the further rapamycin assay.
There are some limitations to our study. First, the differentiation degree of exercise intensity is not enough. It leads to no significant difference in the exercise-enhanced effects among three trained groups, hindering further concretization of exercise intensity. Second, the effects induced by exercise are extensive and comprehensive. Many changes of other pathways are also involved; however, we only investigated the change of mTOR pathway here. Third, the inhibitor used in this study is systemic, which not only inhibits the cortical mTOR pathway but also affects the activity of mTOR pathway in the whole body.
In summary, our study suggests that the expression of neurotrophic factors and the activation of the cortical mTOR pathway are in an intensity-dependent manner. And the moderate intensity is safer and more beneficial than the low and high intensity. The treadmill exercise enhances the neuroplasticity, axonal sprouting, nerve conductive function and skilled motor function recovery in the cervical SCI mice, which can be attributed to the activation of mTOR pathway, as validated by the rapamycin assay. Our results collectively emphasize the key role of mTOR pathway in adaptations of neural reorganization and provide more evidence for clinical treatment in patients with SCI using treadmill exercise training.
Author Contribution Supervision, funding acquisition, and resources supporting, Ce Yang, Yuan Liu, Botao Tan, Lehua Yu, and Ying Yin; study design, manuscript preparation and writing, Zuxiong Zhan and Botao Tan; manuscript review, Botao Tan; animal experiments, statistical collection and analysis, Zuxiong Zhan, Ying Zhu, Lu Pan, Yunhang Wang, Qin Zhao; experimental technology supporting, Sen Li and Haiyan Wang. All authors read and approved the final manuscript.
Funding This work was supported by the National Natural Science Foundation of China (No.81702221 and No.82002377), the Natural Science Foundation of Chongqing (No. cstc2020jcyj-msxm0161, cstc2019jcyj-msxmX0195, and cstc2018jcyjAX0180).
Date Availability All data in the current study are available from the corresponding authors on reasonable request.
Ethics ApprovalAll animal experimental procedures were approved by the Animal Experimental Ethics Committee of Chongqing Medical University [(2020)161].
Consent to Participate Not applicable.
Consent for PublicationAll the authors have read and approved the final version of the manuscript.
Conflict of Interest The authors declare no competing interests.