Spinal cord injury: taking a detour to recovery (Commentary on Schnell et al.)


For years, the ‘holy grail’ of spinal cord injury research has been the attainment of long tract regeneration across the lesion site to provide reconnection of the brain with the motor and sensory centers of the completely severed cord. This goal has been more than elusive, but the few hints of success have usually involved the combination of several different therapeutic strategies aimed at the multiple barriers to regrowth and reconnection (e.g. Lu et al., 2004; Fouad et al., 2005; Bunge, 2008; Tom et al., 2009). The adult central nervous system expresses multiple inhibitory molecules that impede axonal elongation, including myelin proteins discovered by Martin Schwab and his group years ago (Schwab, 2004). Neurotrophic factors can be used to rev-up the latent growth capacities of adult neurons, and have been used with some success (e.g. Tom et al., 2009). In addition, even if axons can re-grow to reach their targets, they must form appropriate functional synapses in order to promote recovery.

In this issue, Schnell et al. (2011) employed a strategy aimed at each of these targets in a slightly less daunting paradigm, a complete lateral hemisection of the rat spinal cord. There is much clinical relevance in this model, as many human spinal cord injuries spare a portion of the descending tracts. Using a combination of anti-NogoA to overcome myelin inhibition, viral vector-mediated expression of the neurotrophin NT-3 to enhance outgrowth, and overexpression of a glutamate receptor subunit that enhances synaptic plasticity, they showed enhancement of recovery. And most impressively, they verified the establishment of new synaptic connections to motor neurons caudal to the lesion using intracellular recording. Did they attain the holy grail? That is, did the new connections regenerate across the lesion site on the same side of the spinal cord (see their Figure 4c)? To answer that question, they performed re-severing of the lesion site while recording from motor neurons exhibiting the new polysynaptic connections. The new synaptic connections remained intact, suggesting that the connections were not regrowth of the original descending inputs, but rather a novel circuit that must have detoured around the lesion onto the other side of the cord, and then back again. These new connections were reliably formed only when the full combination treatment was given.

An anatomical study labeling propriospinal axons from the cervical cord on the same side as the more caudally placed hemisection showed that in the full combination-treated rats, axons could be traced across the cord above the lesion, then back across below the lesion, and indeed a detour pathway for signals to reach the caudal cord. Finally, these new connections were associated with a somewhat better recovery of locomotor function, again only when the full combination was given.

These findings lend support to a growing understanding that the partially injured spinal cord can exhibit reorganization and can produce new connections that are important for partial recovery (e.g. Bareyre et al. 2004; Courtine et al., 2008; Rosenzweig et al., 2010). Indeed, even when all of the descending fibers on one side of the cord are severed, as was the case here, considerable recovery of both hindlimbs occurs (see their results on the BBB locomotor score; Basso et al., 1995). The enhanced recovery seen by Schnell et al. (2011) was above and beyond that normally seen, again suggesting that the new connections, verified by electrophysiology, may provide a substrate for additional recovery.

This new work and other recent studies give hope that judicious targeting of multiple aspects of axonal growth and circuit plasticity can be translated to human spinal cord injury, and that if there is any sparing of the cord at all, new connections can be formed by sprouting and re-routing without the absolute requirement for long tract regeneration. Yet, many questions and caveats remain. The enhancement of axonal sprouting and reconnection may not always yield positive outcomes. Indeed, the first real evidence for axonal sprouting after spinal cord injury (a hemisection!) showed dorsal root sprouting that was thought to provide the substrate for hyperreflexia and spasticity (McCouch et al., 1957). More recently, it has been recognized that neurotrophic treatments that enhance sprouting can lead to autonomic dysreflexia and neuropathic pain (e.g. Krenz & Weaver, 2000; Yajima et al., 2005), and that over-growth of descending corticospinal tract axons may contribute to motor dysfunction in cerebral palsy (e.g. Freil & Martin, 2007). Continued work on the basic biology of axonal growth and connections in both the adult and the developing spinal cord are needed to direct efforts to provide the most efficacious combination of treatments that will yield the best outcomes in humans.