To the editor

The article by McDonald et al. published in Nature Medicine in Dec 1999 deserves comment1. First, related not only to the paper of McDonald et al. but also to many published reports on recovery after spinal cord lesion, is the definition of a functional recovery. BBB scores must be interpreted very carefully for hindlimb motor function. This function can be defined more specifically as locomotion, but authentic locomotion requires both functional and electrophysiological definition2,3. This was not true in the recent works4,5,6. This issue is especially essential when dealing with substitutive restoration, such as that done by McDonald et al. It is now well known that animals lesioned but not transplanted can show various hindlimb movements, such as paw shaking, that have nothing in common with locomotion except that the hindlimbs are moving7.

Moreover, the interpretations of McDonald et al. for the mechanisms of recovery are very unlikely. It is now widely agreed that the prerequisite for locomotion is the existence of a serotonergic and/or noradrenergic drive of the central pattern generator2,3. Monoaminergic axons are unmyelinated. Thus, it is very unlikely that remyelination by grafted cells can account for such a mechanism. Alternatively, monoaminergic neurons could have differentiated from grafted stem cells. To explore this, it would have been necessary to analyze the transmitter phenotype of the grafted cells.