Astrocytes have been considered over decades as accessory cells in the physiology and pathophysiology of the central nervous system. More specifically, their possible role in the pathophysiology of neurotrauma has never been considered until the beginning of the eighties, when Aguayo and colleagues demonstrated that the absence of axonal regeneration in the CNS of adult mammals was not an intrinsic property of neurons, but rather depended upon their cellular environment. Using oxysterols, which had been shown to block the reactivity of astrocytes, we found that the two major proteins of the astrocyte cytoskeleton, GFAP and Vimentin, were major actors of this reactivity in spinal cord injury. We managed then to obtain mice knock-out for the genes coding for these two proteins. Using co-cultures of newborn spinal cord KO astrocytes, with embryonic cortical neurons, we found that these astrocytes were more permissive for neuron survival and neuritic extensions than wild type astrocytes. Moreover, these KO astrocytes presented many biochemical characteristics of the so-called radial glia which operate as guides for neuronal migration and neuritic extension during embryogenesis. KO mice underwent lateral hemi-section of the spinal cord, and we found that at variance with wild type they did not develop astrocytic scars, sprouting of corticospinal and raphespinal axons was extensive, and they regained function of the paralyzed hindlimb. We have recently replicated these results both on in vitro and in vivo models with lentiviral vectors carrying siRNA for GFAP and Vimentin, on wild type mice. It appears then that reactive astrocytes, when adequately manipulated can be converted into permissive substrates for axonal regeneration and functional recovery after spinal cord injury.

Supported by INSERM, IRME, Verticale and Ikerbasque