(333d) On the Molecular Diversification and Maturation of Neurons in the Visual System

A central question in neurobiology is how diverse neuronal types in the mammalian brain arise through an interplay of genetic regulation and experience-dependent mechanisms. I will present our recent collaborative efforts to address this question through a combination of high-throughput gene expression measurements in mouse models, machine learning and statistical inference of large scale datasets, as well as histological assays. Our focus will be two areas of the mammalian visual system - the retina, which is the part of the eye that initiates vision, and the primary visual cortex, where the first stage of higher-order visual processing occurs.

In the retina, we have studied how retinal ganglion cells (RGCs), the primary output neurons of the retina, differentiate into 45 distinct types during embryonic and postnatal life. We use a statistical inference approach rooted in the mathematical theory of optimal transport to identify fate associations from "snapshot" measurements of RGCs at different stages of development in mice. We find that RGCs diversify through a gradual process, and predict that precursors are likely to be multipotential. Importantly, by analyzing three distinct models where mice are deprived of vision, we also show that RGC diversification appears to be genetically hardwired.

In contrast, we find a much more significant role for vision in the molecular maturation of neurons in the primary visual cortex. It has been long known that the cortex subdivided into five anatomically "layers" (L2, 3, .. 6)containing distinct populations of cells. By analyzing the postnatal maturation of these cells based on single-nucleus gene expression measurements in mice, we find that neurons in the superficial layers (2/3/4) require visual input to acquire their molecular identities, while neurons in the deeper layer (5/6) do not. Unexpectedly, we also discovered that vision plays a role in "patterning" these neurons into distinct sublayers which correlate with projection specificities to higher visual regions. Through rigorous perturbation assays we show that visual deprivation during early postnatal life completely abrogates neuronal maturation in these superficial layers, and their associated anatomical patterns. Thus, we identify role for experience mediated gene expression programs in the anatomical patterning of this brain region.

These two interconnected but contrasting studies suggest a tempting hypothesis that neural diversity sensory systems is likely shaped largely by hardwired genetic mechanisms, while those in higher-order processing centers are likely to be shaped significantly through experience dependent plasticity.