Disentangling the roles of molecular guidance cues and spontaneous activity in the emergence of visual topographic maps

Abstract

Prior to the onset of sensory experience neural circuits develop through a combination of genetic instructions and activity-dependent mechanisms. Classic studies supported a model in which axon guidance molecules and spontaneous activity act independently and sequentially to first guide axons to form a rough circuit at the target tissues which is then refined by clustering correlated inputs. Recent experiments have questioned this model proposing that spontaneous activity has a direct effect on axon guidance mechanisms in different developmental stages such as differentiation, axon guidance or map topography. Using the retinotopic map of mice as a model we have developed an experimental approach in vivo to block activity in a sustained and reliable way in retinal ganglion cells (RGCs) to definitively clarify the role of spontaneous activity during development and unravel the controversial question of whether or not it is required to activate the repulsive signaling mediated by EphA/ephrinAs during the formation of visual topographic maps. In this study, we first performed calcium recordings in the developing retina of alive animals in order to analyze the patterns of retinal spontaneous activity in vivo during the first postnatal week and found calcium transients with a clear correlation between neighboring cells. Then we blocked spontaneous activity in embryonic RGCs in vivo (by ectopic expression of the inward rectifier potassium channel Kir2.1) and showed that differentiation, axonal growth and axon pathfinding are activity-independent processes in these neurons. In addition, contrary to previous studies in vitro, our results show that, although spontaneous activity is essential for local arborization of retinal axons in the termination zone (TZ) at the targets, it is dispensable for the activation of EphA/ephrinA signaling and therefore for topography, as predicted by the classic model.