Abstract EN:

CAMP occupies a strategic position to control responses to a large variety of developmental cues. However the role of individual adenylate cyclase subtypes is unclear. In the barrelless mouse strain (brl), déficient for the adenylate cyclase 1 (AC1) gene, the fine topography of the retinotectal map is altered, with a lack of confinement of retinal axon arbors in both the rostro-caudal and medio-lateral dimensions (Ravary et al. 2003). Two principal processes are known to be involved in the refinement of retinal axon topography: one depends on the repulsive guidance cues, ephrins, and the other depends on the spontaneous activity generated in the retina. We have identified the ACs involved in the development of retinal projections and investigated the interactions between cAMP, neuronal activity, and ephrin-As that are critical for the topography of retinal axons along the rostrocaudal axis of the superior colliculus (SC). To identify the ACs involved in the development of retinal maps, we have described the developmental expression of the ACs in the retina and in the brain. We have identified AC1, AC2, ACS, and ACS in the layer of cells projecting from the retina (the retinal ganglion cells (RGCs)). We found that AC1, AC2 and ACS are expressed in the targets of the retinal axons. To test whether these ACs are critical for the formation of retinal maps, we have traced the retinal axons in adult WT, AC1-/-, brl, AC5-/- and AC8-/- mice. Only AC1-/- and brl mice show altered retinal projections. Thus AC1 seems to be the only AC involved in the development of retinal maps. To study the role of AC1 we have developed a co-culture model that mimics the development of retinal projections in vivo. We have confronted temporal or nasal retinal explants with a mesencephalic slice containing the SC. In this model, temporal (and not nasal) axons retract from the caudal SC in an ephrin-A-dependent way. Temporal axons from brl temporal retina do not retract where confronted with a WT SC, whereas a brl mesencephalic slice does not alter the retraction of WT axons. Thus AC1 is required in the RGCs, not in the SC, for correct refinement of retinal projections. We have tested the interaction between AC1 and ephrin-As by inducing the retraction of retinal axons by soluble ephrin-A5. The rate of collapse of brl growth cones is normal but after collapsing their axons that to retract, showing that AC1 perturbs the normal targeting of retinal axons by altering thé ephrin-A response. As AC1 is stimulated by extracellular calcium entry and neuronal activity is required for thé formation of retinal projections, we hypothesized that thé stimulation of AC1 by activity is critical for the ephrin-A induced axonal retraction. In our co-culture model, blocking activity alters the retraction of temporal axons in a manner similar to thé absence of AC1 and the perturbation of ephrin-A signalling. The retraction defects of retinal axons that lack AC1 are reproduced by blocking activity with TTX. These defects are rescued by generating cAMP oscillations in TTX treated growth cones, but not by a continuous stimulation of the ACs. A manuscript describing thèse findings is in préparation. Thus AC1 is the only AC required for the refinement of retinal projections, and acts in the growth cones by generating cAMP oscillations in response to neuronal activity in order to allow the axonal retraction induced by ephrin-As.