Deconstructing the Brain Serotonergic System: A Transcriptomic Analysis of 5-HT Neurons Across Space and Time

Abstract

Neurons that produce and release the monoamine serotonin (5-HT) are distributed across nine raphe nuclei of the brainstem and have a known role in a wide range of homeostatic and behavioral processes. While historically viewed as homogeneous, more recently discovered phenotypic heterogeneity within the serotonergic system is indicative of functionally distinct subtypes of 5-HT neurons. Consequently, a motivation in the field is to map the structure-function relationships of the system, contributing to both basic neuroscience and human health through informing functional separability. Cell phenotype is closely linked to gene expression; thus, determining transcriptomic differences between cells can be valuable to inform on functionally distinct neuron subtypes. Previous work from the Dymecki lab linked mature 5-HTergic neuron gene expression, progenitor cell subset of origin in the embryonic hindbrain (its rhombomeric (r) “lineage”), soma location anatomically, and subserved functions at the organismal level. Less well studied, though, is the heterogeneity contained within a given developmental lineage, despite evidence for regional variation in functional properties. One such example is the r0/r1 lineage of 5-HTergic neurons that comprise the dorsal raphe (DR) nucleus – the largest anatomically defined group of brain 5-HTergic neurons, linked with affective, cognitive, and sensorimotor functions. Through my dissertation work, I have sought to delineate DR 5-HTergic neuron heterogeneity mechanistically (as opposed to solely anatomically), using a novel analysis pipeline involving high-throughput single-cell RNA sequencing (scRNAseq) that informs anatomical and hodological mapping (soma and efferent boutons) using intersectional genetic strategies and histology. Cell subtype-specific scRNAseq then validates and expands the initial transcriptomic findings using an independent, orthogonal approach.

Through this iterative work we have built a new structure-function map of the 5-HTergic DR neuronal system, identifying fourteen subtypes of 5-HTergic neurons within the DR, embedded within a broader organization relating to their expression of other neurotransmitters in addition to 5-HT. Indeed, thirteen of the fourteen neuron subtypes express genes suggestive of co-transmission. Five of the identified neuron subtypes manifesting a GABAergic-serotonergic phenotype; six, a glutamatergic-serotonergic phenotype; and two, a predominantly glutamatergic phenotype with variable 5-HT levels. Within these broad divisions lie clusters with differential expression of genes spanning categories indicative of differences in functional identity. Such data may offer new ways to conceptualize serotonin-related disorders and ultimately provide novel circuit nodes and molecular pathways to consider for therapeutic development.

Towards determining if the discovered transcriptomic heterogeneity of the adult 5-HTergic DR is hard coded during embryonic development and thus present at birth, we extended our analyses to include early postnatal time points. We then found that while aspects of the identified transcriptomic structure of DR 5-HT neurons are present at P0, the full structure only becomes apparent at P5, and variability in the timing of expression changes in genes related to maturation and plasticity indicate that some populations may be more plastic, while others are hard coded.

Throughout these studies, the functional properties of the DR Met expressing population of 5-HT neurons, found in the caudal part of the DR, were characterized further. This population was found to be a predominant supra-ependymal projecting 5-HT population with fibers closely opposing proliferating cells in the adult neural stem cell niche of the subventricular zone. In addition, electrophysiological analysis showed they were hypoexcitable in comparison to other DR 5-HT populations. Furthermore, a population with similar anatomic distribution and gene expression profile was found in the human infant brainstem, highlighting the conservation of this unique population, and suggesting that the structure-function map characterized in mice will extend to the human brain more broadly.