Interaction of extracellular signals during the polarization and orientation of retinal ganglion cells in vivo

Abstract

Neurons are highly polarized cells with a single, long process responsible for signal output, the axon, and a series of shorter and often ramified neurites capable of receiving and integrating signals, the dendrites. During nervous system development, these cells arise from a neuroepithelium, whose polarity is given, in part, by the presence of apical junction complexes and by the association with a basal lamina, enriched in Laminin1. Hence, a polarity transition, from an epithelial to a neuronal polarity, is an inherent step in neuronal differentiation. During this process, neuronal orientation within the tissue, essential for the correct transmission of information, is established. In this work, we aim to dissect some of the molecular mechanisms governing retinal ganglion cell (RGC) orientation in the zebrafish, with an emphasis on proteins polarized within the neuroepithelium: Laminin1 and the Slit-Robo signaling pathway, which act as a positive and negative signal for axon extension, respectively. We hypothesize that it is the functional interaction of these two signals that guides RGC orientation in vivo. Firstly, we decided to better characterize the role of the main candidate from the Slit-Robo pathway, Slit2, in RGC differentiation. We generated a null mutant line for slit2, in which we found a disorganization of retinal axons at the optic chiasm, but no neuronal orientation defects. This phenotype was similar to the one obtained upon the injection of a morpholino oligomer directed against slit2, and was consistent with mRNA expression, as evidenced through fluorescent in situ hybridization. Moreover, slit3 knock-down in slit2 mutant embryos resulted in aberrant crossing of retinal axons at the midline. On the other hand, injection of slit2 morpholino in bashful/laminin1 embryos caused orientation defects in a significant proportion of RGCs, not evident in non-injected bashful/laminin1 embryos. Parallel to this, we set out to analyze RGC differentiation in re-aggregated retinal cultures. In these organoids, cells self-organize into layers reminiscent of retinal lamination in vivo, with RGCs located at the outer-most region. These cells were able to grow axons over the surface of the aggregates, albeit with difficulty, as evidenced through time-lapse imaging. This could possibly be due to the lack of superficial Laminin, which we found accumulated at the center of the organoid and associated with retinal pigment epithelium cells. When analyzing the function of the Slit-Robo pathway in RGC differentiation in these conditions, we found both RGC axons and cell bodies inside the astray/robo2-derived organoids. While axon guidance defects are a hallmark of astray/robo2 embryos, they do not present mispositioned RGCs, pointing to the possibility of compensatory mechanisms operating in vivo and absent in re-aggregated cultures. We thus conclude that Slit2 is essential for the organization of retinal axons in the optic chiasm, and that it functions together with Slit3 to guide axon crossing at the midline. Moreover, RGC orientation in vivo depends on the cooperation between Lamininæ1 and Slit2, while cell positioning in vitro depends, at least partially, on Robo2.