Locomotor recovery after lumbar spinal cord injury: fact or fancy?

A recent global survey reported that the prevalence of spinal cord injury (SCI) ranges from 250 to 900 per million (Singh et al. 2014). Therefore, up to 6.5 million people world-wide live with the tragic consequences of SCI. Trauma centred on the lumbar segments occurs in only about 20% of cases but is of particular concern. This region contains critical rhythm-generating circuitry for the production of stepping, as was first reported by Graham Brown over 100 years ago (Graham Brown, 1911). It is now generally accepted that the human spinal cord also contains the neural substrate for motor rhythm production. The intrinsic capacity of the spinal cord to generate patterned output gives rise to the enticing prospect that function might be restored if dormant circuits located below the site of injury can be re-activated. Thus, considerable research has focused on regeneration of damaged pathways in the cervical and thoracic regions. Epidural electrical stimulation of the lumbar cord, which bypasses cervical and thoracic lesions, can enable voluntary movements not otherwise possible in paraplegic humans (Angeli et al. 2014). But what if instead of cervical or thoracic cord lesions that disrupt descending neural projections the injury is centred on lumbar segments, thereby destroying rhythm-generating circuitry? Our present understanding of spinal cord function predicts a substantially worse prognosis in this situation. Indeed, adult rodents subjected to neurochemical ablation or compression injuries of upper lumbar segments do not recover locomotor function (Magnuson et al. 1999; Moonen et al. 2016). In contrast, in this issue of The Journal of Physiology, Züchner and colleagues report a strikingly different result in neonatal mice (Züchner et al. 2018). One-day-old mice received lumbar cord compression injuries resulting in up to 90% neuronal loss at the lesion epicentre. The rostral lumbar region (T13–L2) was targeted because previous studies suggest these segments have the greatest rhythmogenic capacity. Kinematic analysis was performed during air-stepping on postnatal days 2, 5 and 9, as well as during swimming tasks on postnatal days 15 and 25. Other mice, similarly injured on postnatal day 1, were tested in vitro; the complete spinal cord was removed on postnatal days 2, 3, or 4, placed in an oxygenated bath, and tested for its ability to generate locomotor activity in response to bath-applied neurochemicals. The results showed that locomotor output appeared by postnatal day 4 in vitro and virtually normal hindlimb locomotor patterns were observed by day 25 in vivo. How does hindlimb stepping recover so promptly and remarkably well in neonatal mice after destruction of rhythmogenic circuitry? It has long been known that neonatal animals have much greater potential for recovery than older animals after SCI (Stelzner et al. 1975). This ‘infant lesion effect’ dictates that those neural elements least mature at the time of insult are the ones most likely to successfully reorganize post-injury (Bergman & Goldberger, 1982). In the present study, Züchner and colleagues demonstrate sprouting of serotonergic fibres as well as altered neurochemical sensitivity of the network. Plasticity also occurs in the mature nervous system after trauma, although in the case of lumbar cord injury in adult animals it appears that plasticity alone is insufficient. Impressive recovery in neonatal mice may also be related, at least in part, to the longitudinally distributed nature of the locomotor network. There is evidence that multiple segments above and below the most rhythmogenic zone (T13–L2) also contain rhythmogenic elements that support hindlimb stepping as part of a distributed system (Cowley & Schmidt, 1997). Redundancy and/or other back-up mechanisms that help ensure locomotor output despite local failures in the network are features well-suited for the preservation of a basic survival function such as locomotion. Possibly these properties are more readily called into play in the immature spinal cord, especially under in vitro conditions, and are inaccessible in adult animals. Nevertheless, the distributed network concept is quite compatible with the results of adult human studies wherein the higher the level of cord injury (the greater the number of segments in continuity with the lumbar cord) the better the locomotor performance (Dietz et al. 1999). Despite the substantial difference in neonatal versus adult recovery, investigation of the immature nervous system, with its greater capacity for repair, is essential. Future studies may determine which specific neurotrophic factors or other mechanisms underlie successful recovery of locomotor function in neonatal but not mature animals, with a view to manipulating such factors in adult animals and eventually in humans. Indeed, therapeutically oriented studies bridging the biology of the mature and immature spinal cord have already begun. Injection of human fetal spinal cord-derived stem cells into the site of lumbar cord injury in adult rats improves hindlimb function (van Gorp et al. 2013). Also worth noting is that human spinal cord injury is often incomplete; some mechanisms of repair in the neonatal lumbar cord may prove more readily recruited in mature animals with less devastating injury. The ultimate goal in spinal cord research is to restore function. However, major therapeutic advances, whether using electrical stimulation, neuropharmacological modulation, training, robotics, molecular manipulation, or regenerative strategies, demand a more comprehensive understanding of the fundamental cellular and network mechanisms underlying normal behaviour, as well as the response of these systems to injury. Carefully designed experiments using animal models, of the type reported by Züchner and colleagues in this issue, not only help build the required knowledge base but can also prompt the spinal cord to reveal unanticipated, but valuable, secrets. None declared.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.