Transcriptomic cellular diversity of the early human developing brain

Abstract

The complexity of the mammalian brain is partly reflected in its cell type diversity which influences the function of neurons that encode the behavior of animals. Brain cell type diversity emerges during embryonic stages, a critical period when neurons start to become functionally active and establish their connectivity across the brain. Since the pioneering of single-cell RNA-sequencing (scRNA-seq), we can question when and how cellular diversity arises in the brain in a large-scale manner.

This thesis aims to study the human brain during the first trimester by using scRNA-seq to obtain a global view of the basic principles of the developing brain. First, I introduce human embryology from a historical perspective and summarize key concepts in central nervous system (CNS) development. I review few gaps in the field related to our findings, followed by current approaches and nomenclatures used in the field of single-cell genomics that applies to development. To put our work into perspective, I present an overview of the latest efforts to study human brain development at the single-cell level, both in the healthy and diseased brain.

Then I present the following two papers and a manuscript: In Paper I we used scRNA-seq to construct a cell taxonomy of the adult mouse nervous system. We describe two major groups: neuronal- and non-neuronal cells that were subdivided into distinct cell types. Overall, the neurons were transcriptionally similar across brain regions, whereas non-neuronal cells such as astrocytes, formed subgroups and were regionally distinct. The whole dataset revealed an organization that reflects the developmental origin of all cell types.

Paper II describes a method, RNA velocity, that infers temporal changes from static scRNAseq gene expression measurements. By realigning sequencing reads, this method detects and makes use of the unspliced and spliced mRNA, whose relative abundance is used to measure the change of rate in gene expression (the time derivative) in different tissues. This method is particularly suitable for developmental lineages, which was shown and validated both in vitro and in situ in this study.

Paper III presents a single-cell atlas of the human developing CNS across all major brain regions during postconceptional weeks (p.c.w.) 5 to 14. We observe that major cell classes emerge during this period, most of them being regionally diverse and to a surprisingly high degree among glial cells. We display the high resolution of this data by resolving several lineages in the forebrain and validated the spatial location of transcriptional cell types at 5 p.c.w. by using single-molecule FISH. As a whole, this study represents a reference of human brain development during the first critical period in life.

To tie these studies together, our findings on glial diversity were partially shared between the adult mouse and developing human CNS. We further showed that an RNA velocity-based method can be used to model the cell cycle dynamics in cortical tissue. To conclude, I discuss advantages and limitations of single-cell transcriptomics, its future challenges and how using this technology sheds light on the early human developing brain as is described in this thesis.