Lineages and molecular heterogeneity in the developing nervous system

Abstract

Information in the genome unfolds through a dynamic process leading to the molecular and anatomical organization of a physiologically functional organism. The nervous system is the most diverse and intricate architecture generated by this process. It is composed of hundreds of millions of cells of hundreds of different cell types, whose connectivity and interactions are the physiological underpinnings of our capacity to respond to stimuli, our ability to learn and our cognitive capabilities.

In this thesis, I explore the formation of tissues in the nervous system during embryonic development. In particular, I focus on changes in molecular composition that lead progenitor cells to generate a complex mix of cell types. The specific aim of this work is to address the lack of complete and systematic knowledge of the heterogeneity of neural tissues and to describe the progression of a cell through different molecular states. To achieve this, I took advantage of the new opportunities offered by single-cell expression profiling technologies to gain a holistic view of a developing tissue. To contextualize the work, I review the relevant literature and conceptual framework. Starting with a historical perspective, I discuss the concept of cell type and how it relates to developmental dynamics and evolution. I then review different aspects of developmental neuroscience, starting with general principles and then focusing on the main areas of interest: the ventral midbrain, the sympathetic nervous system, and postnatal development. Then the technological advances instrumental for this thesis are reviewed, with a focus on analysis methods for single-cell RNA sequencing. Finally, I discuss the relationship between lineages and gene regulation, and I introduce the reader to the idea of a global time derivative of gene expression through traditional systems biology modeling.

Then I present the results of three different studies. In paper I, we used single-cell RNA sequencing to describe the cell-type heterogeneity of sympathetic ganglia. We found seven distinct kinds of neurons, where only two had been previously described. Using lineage tracing, we shed light on the developmental origin of the new types. We linked their molecular profile to function and described how they innervate the erector muscles. Paper II describes the embryonic development of the ventral midbrain at the single-cell level. We characterized human and mouse embryonic tissues, identifying cell types and their homologies. We found an uncharacterized heterogeneity among radial glial cells and gained new insight into the timing of dopaminergic neurons specification. Finally, we presented a data-driven strategy to assess the quality of in vitro differentiation protocols. In paper III we addressed the major limitation of studying development with single-cell RNA sequencing: the absence of a temporal dimension. We described an analysis framework that uses the ratio of spliced to unspliced RNA abundance to estimate the time derivative of gene expression. The method was used to predict the future molecular states of cells and to determine their fate bias.

In these studies, we produced a rich description of tissue heterogeneity and answered different biological questions. The results were achieved by harnessing the information contained in the data through analysis approaches inspired by developmental or physical principles. In summary, this thesis provides new insight into several aspects of mammalian nervous-system development, and it presents analytical approaches that I predict will inspire future investigation of the developmental dynamics of single-cells.