Sensory experience (“nurture”) acts upon a genetically hardwired blueprint (“nature”) to sculpt the brain during early developmental stages known as “critical periods” (CPs). In the process, immature neurons acquire diverse identities and assemble into circuits defined by specific synaptic connections. Disrupted organization of neural circuits during CPs underlies a myriad of neural disorders, making the study of experience-dependent neural development fundamentally and clinically important. Recent advances in single-cell transcriptomics have enabled the “bottom-up” characterization of neural diversity at molecular resolution. By applying unsupervised machine learning approaches to large-scale gene expression measurements, catalogs of cell types have been generated and used to study development, function, dysfunction, and evolution across the nervous system. However, the influence of sensory experience during CPs on the development and maturation of diverse neuronal types remains unknown. Furthermore, due to the destructive nature of single-cell transcriptomic assays, temporal changes cannot be directly observed. They must be inferred from high-dimensional gene expression “snapshots” in distinct samples of single cells spanning developmental stages, conditions, and animals. In this thesis, I present developmental transcriptomic cell type atlases in three regions of the mouse brain: the retina, the whisker somatosensory cortex, and the primary visual cortex. I then use these atlases to understand the role of sensory experience in the maturation of cell types within each system.
In the first part of this thesis, I investigate the role of vision in the development of 45 retinal ganglion cell (RGC) types, the retina’s sole output neuron. The development and connectivity of RGCs are patterned by activity-independent transcriptional programs and activity-dependent remodeling. To inventory the molecular correlates of these influences, high-throughput single-cell RNA sequencing (scRNA-seq) was applied to mouse RGCs at six embryonic and postnatal ages. I identified temporally regulated modules of genes that correlate with, and likely regulate, multiple phases of RGC development, ranging from differentiation and axon guidance to synaptic recognition and refinement. Some of these genes are expressed broadly, while others, including key transcription factors and recognition molecules, are selectively expressed by one or a few of the 45 transcriptomically distinct types defined previously in adult mice. Next, I used these results as a foundation to analyze the transcriptomes of RGCs in mice lacking visual experience due to dark rearing from birth or due to mutations that ablate either bipolar or photoreceptor cells. 98.5% of visually deprived (VD) RGCs could be unequivocally assigned to a single RGC type based on their transcriptional profiles, demonstrating that visual activity is dispensable for acquisition and maintenance of RGC type identity. However, visual deprivation significantly reduced the transcriptomic distinctions among RGC types, implying that activity is required for complete RGC maturation or maintenance. Consistent with this notion, transcriptomic alterations in VD RGCs significantly overlapped with gene modules found in developing RGCs. These results provide a resource for mechanistic analyses of RGC differentiation and maturation and for investigating the role of activity in these processes.
In the second part of this thesis, I investigate the influence of vision on the development of cell types in the mouse primary visual cortex (V1), the area of the brain that processes visual information. By combining unsupervised and supervised machine learning approaches, single-nucleus RNA sequencing, visual deprivation, genetics, and functional imaging, my colleagues and I discovered that vision selectively drives the specification of glutamatergic neuronal types in upper layers (L) (L2/3/4), while deep-layer glutamatergic neuronal types, GABAergic neuronal types, and non-neuronal cell types are established before eye opening. Furthermore, I found that L2/3 neuronal types form an experience-dependent spatial continuum defined by the graded expression of ~200 genes, including regulators of cell adhesion and synapse formation. One of these genes, Igsf9b, a vision-dependent gene encoding an inhibitory synaptic cell adhesion molecule, is required for the normal development of binocular neurons in L2/3. These results raise the intriguing possibility that sensory experience acts differentially in regulating the plasticity of individual cell types and provide a blueprint for future studies into the experience-dependent development of other brain regions.
In the third part of this thesis, I investigate whether experience-dependent regulation of cell types exists in another cortical region known as the primary whisker somatosensory cortex (wS1). wS1 is a major model system to study the experience-dependent plasticity of cortical neuron physiology, morphology, and sensory coding. However, the role of sensory experience in regulating neuronal cell type development and gene expression in wS1 remains poorly understood. I assembled and annotated a transcriptomic atlas of wS1 during postnatal development comprising 45 molecularly distinct neuronal types that can be grouped into eight excitatory and four inhibitory neuron subclasses. Using this atlas, I examined the influence of whisker experience from postnatal day (P) 12, the onset of active whisking, to P22, on the maturation of molecularly distinct cell types. During this developmental period, when whisker experience was normal, ~250 genes were regulated in a neuronal subclass-specific fashion. At the resolution of neuronal types, I found that only the composition of layer (L) 2/3 glutamatergic neuronal types, but not other neuronal types, changed substantially between P12 and P22. These compositional changes resemble those observed previously in the primary visual cortex (V1), and the temporal gene expression changes were also highly conserved between the two regions. In contrast to V1, however, cell type maturation in wS1 is not substantially dependent on sensory experience, as 10-day full-face whisker deprivation did not influence the transcriptomic identity and composition of L2/3 neuronal types. A one-day competitive whisker deprivation protocol also did not affect cell type identity but induced moderate changes in plasticity-related gene expression. Thus, developmental maturation of cell types is similar in V1 and wS1, but sensory deprivation minimally affects cell type development in wS1.
Finally, I highlight the major conclusions of each chapter, discuss their historical context, and suggest future work on the experience-dependent development of neuronal cell types and circuits in the mammalian nervous system.
