Acute exercise prevents the development of neuropathic pain and the sprouting of non-peptidergic (GDNF- and artemin-responsive) c-fibers after spinal cord injury

Abstract

Spinal cord injury (SCI) impaired sensory fiber transmission leads to chronic, debilitating neuropathic pain. Sensory afferents are responsive to neurotrophic factors, molecules that are known to promote survival and maintenance of neurons, and regulate sensory neuron transduction of peripheral stimuli. A subset of primary afferent fibers responds only to the glial cell-line derived neurotrophic factor (GDNF) family of ligands (GFLs) and is non-peptidergic. In peripheral nerve injury models, restoration of GDNF or artemin (another GFL) to pre-injury levels within the spinal cord attenuates neuropathic pain. One non-invasive approach to increase the levels of GFLs in the spinal cord is through exercise (Ex), and to date exercise training is the only ameliorative, non-pharmacological treatment for SCI-induced neuropathic pain. The purpose of this study was 3-fold: 1) to determine whether exercise affects the onset of SCI-induced neuropathic pain; 2) to examine the temporal profile of GDNF and artemin in the dorsal root ganglia and spinal cord dorsal horn regions associated with forepaw dermatomes after SCI and Ex; and 3) to characterize GFL-responsive sensory fiber plasticity after SCI and Ex. Adult, female, Sprague-Dawley rats received a moderate, unilateral spinal cord contusion at C5. A subset of rats was exercised (SCI+Ex) on automated running wheels for 20min, 5days/week starting at 5days post-injury (dpi), continuing until 9 or 37dpi. Hargreaves' and von Frey testing was performed preoperatively and weekly post-SCI. Forty-two percent of rats in the unexercised group exhibited tactile allodynia of the forepaws while the other 58% retained normal sensation. The development of SCI-induced neuropathic pain correlated with a marked decrease in the levels of GDNF and artemin in the spinal cord and DRGs. Additionally, a dramatic increase in the density and the distribution throughout the dorsal horn of GFL-responsive afferents was observed in rats with SCI-induced allodynia. Importantly, in SCI rats that received Ex, the incidence of tactile allodynia decreased to 7% (1/17) and there was maintenance of GDNF and artemin at normal levels, with a normal distribution of GFL-responsive fibers. These data suggest that GFLs and/or their downstream effectors may be important modulators of pain fiber plasticity, representing effective targets for anti-allodynic therapeutics. Furthermore, we highlight the potent beneficial effects of acute exercise after SCI.

Keywords: Artemin; Central pain; GDNF; Mechanical allodynia; Spinal cord injury; Thermal hyperalgesia.

Figure 1. At-level tactile allodynia is prevented by acute exercise

(A) Five weeks after SCI, at-level allodynia occurred in ~42% of the rats with SCI (10 out of 24 rats). Exercise intervention reduced the incidence of at-level allodynia to 6% (1 out of 17 rats; χ²=14.529, p=0.0001). The paw withdrawal threshold to tactile stimuli for the ipsilesional (B) and contralesional (C) forepaw were recorded over a period of 35 days. Before SCI, rats withdraw their forepaws from a stimulus force of 60 g (indicated by dashed line). Two behavioral cohorts emerge after SCI, those that maintain tactile paw withdrawal thresholds at or near 60g force (SCI No Allodynia) and those that exhibit a significant reduction in paw withdrawal thresholds that plateau near 30 g force (SCI Allodynia). These distinct responses to SCI were apparent by 14 dpi and persisted to at least 35 dpi (**p<.05 vs all other groups). Rats that were exercised (SCI + Exercise) starting at 5 dpi did not develop allodynia in either the ipsilesional or contralesional forepaw. There was significant hyposensitivity at 7 dpi in the contralesional forepaw of the SCI No Allodynia and SCI + Exercise groups that resolved by 14 dpi (^##p<.05 vs. naïve).

Figure 2. Acute exercise does not prevent at-level thermal hyperalgesia

The latency to paw withdrawal to a noxious thermal stimulus for the ipsilesional (A) and contralesional (B) forepaw was recorded over a period of 35 days. Before SCI, rats withdraw their forepaws after 11 seconds (indicated by dashed line). After SCI, there was no difference in paw withdrawal latency between SCI No Allodynia, SCI Allodynia and SCI + Exercise groups at any time point for ipsilesional or contralesional forepaws, as all groups exhibited a significant decrease in paw withdrawal latency compared to baseline values (**p<.05 vs naïve control).

Figure 3. Anatomical assessments in SCI rats with and without tactile allodynia

Representative transverse sections of the lesion epicenter stained for myelin sacrificed at 9 days post injury in groups SCI No Allodynia (A), SCI Allodynia (B), and SCI + Exercise (C). Sections from SCI rats sacrificed at 37 days post injury in groups SCI No Allodynia (D) SCI Allodynia (E) and SCI + Exercise (F) are presented. The 200 kdyne impact produced a moderate, unilateral lesion with near complete degeneration of grey matter and a spared rim of white matter on the right side of the spinal cord. Importantly, the grey and white matter of the contralesional spinal cord appears normal (scale bar=0.5mm). There were no significant differences in tissue sparing through the rostrocaudal extent of the lesion between any groups at either time point (G).

Figure 4. Exercise prevents SCI-induced reduction of GDNF

In naïve rats, the concentration of GDNF was 29 pg/mg tissue in the C7 and C8 dorsal root ganglia (dashed line in A and B). At 9 and 37 dpi, the concentration of GDNF was significantly reduced in the ipsilesional (A) and contralesional (B) DRGs of the SCI Allodynia group (*p<.05 vs all other groups). Exercise starting at 5 dpi maintained naïve concentrations of GDNF in the DRG. The concentration of GDNF in the naïve dorsal half of the spinal cord was 60 pg/mg tissue (dashed line in C and D). (C) The concentration of GDNF decreased in the ipsilesional dorsal spinal cord of the SCI Allodynia group compared to all other groups (*p<.05) at 9 and 37 dpi. (D) GDNF levels were significantly decreased in the contralesional dorsal cord of the SCI No Allodynia at 9 but not 37 dpi (#p<0.05 vs naïve). Exercise starting at 5 dpi maintained the levels of GDNF in the dorsal spinal cord at both time points.

Figure 5. Exercise prevents SCI-induced reduction of artemin

In naïve rats, the concentration of artemin was 29 pg/mg tissue in the C7 and C8 dorsal root ganglia (dashed line in A and B). At 9 and 37 dpi, the concentration of artemin was significantly reduced in the ipsilesional and contralesional DRGs of the SCI Allodynia group (*p<.05 vs all other groups). Exercise starting at 5 dpi maintained naïve concentrations of artemin in the DRG. The concentration of artemin in the normal dorsal half of the spinal cord was 69 pg/mg tissue (dashed line in C and D). The concentration of artemin decreased in both the ipsilesional (C) and contralesional (D) dorsal spinal cord of the SCI Allodynia group compared to all other groups (*p<.05) at 9 dpi. Exercise starting at 5 dpi maintained the levels of artemin in the dorsal spinal cord. At 37 dpi, there were no significant differences between groups.

Figure 6. Exercise prevents SCI-induced redistribution of GFL-responsive afferents in the ipsilesional dorsal horn

Representative section of the ipsilesional C7 dorsal horn of Naïve (G),SCI No Allodynia with 9 (A) or 37 d (D) survival, SCI Allodynia with 9 (B) or 37 d (E) survival and SCI + Exercise with 9 (C) or 37 d (F) survival. Proportional area of positively labeled tissue within the dorsal horn was determined across the C7 and C8 spinal cord. The average proportional areas for these regions of interest are represented in H. Significantly greater proportional area was labeled for Isolectin-B4, a marker of non-peptidergic c-fibers at both 9 and 37 dpi of the SCI Allodynia group (*p<.05 vs all other groups; dashed line indicates the proportional area of Isolectin-B4+ afferents in the dorsal horn of naïve rats). (I)The degree of afferent labeling predicted at-level allodynia of the ipsilesional forepaw (r²=0.52; p<.05).

Figure 7. Exercise intervention prevents SCI-induced redistribution of GFL-responsive afferents in the contralesional dorsal horn

Representative section of the contralesional C7 dorsal horn of Naïve (G),SCI No Allodynia with 9 (A) or 37 d (D) survival, SCI Allodynia with 9 (B) or 37 d (E) survival and SCI + Exercise with 9 (C) or 37 d (F) survival. Proportional area of positively labeled tissue within the dorsal horn was determined across the C7 and C8 spinal cord. The average proportional areas for these regions of interest are represented in H. Significantly greater proportional area was labeled for Isolectin-B4, a marker of non-peptidergic c-fibers at both 9 and 37 dpi of the SCI Allodynia group (*p<.05 vs all other groups; dashed line indicates the proportional area of Isolectin-B4+ afferents in the dorsal horn of naïve rats). (I) The degree of afferent labeling predicted at-level allodynia of the contralesional forepaw (r²=0.64; p<.05).