Molecular design of hypothalamus development

Abstract

A wealth of specialized neuroendocrine command systems intercalated within the hypothalamus control the most fundamental physiological needs 1 , 2 . Nevertheless, a developmental blueprint integrating molecular determinants of neuronal and glial diversity along temporal and spatial scales of hypothalamus development remains unresolved 3 . Here, we combine single-cell RNA-seq on 51,199 cells of ectodermal origin, gene regulatory network (GRN) screens in conjunction with GWAS-based disease phenotyping and genetic lineage reconstruction to show that 9 glial and 33 neuronal subtypes are generated by mid-gestation under the control of distinct GRNs. Combinatorial molecular codes arising from neurotransmitters, neuropeptides and transcription factors are minimally required to decode the taxonomical hierarchy of hypothalamic neurons. Differentiation of GABA and dopamine but not glutamate neurons relies on quasi-stable intermediate states with a pool of GABA progenitors giving rise to dopamine cells 4 . An unexpected abundance of chemotropic proliferation and guidance cues commonly implicated in dorsal (cortical) patterning 5 was found in the hypothalamus. Particularly, Slit / Robo loss-of-function impacted both the production and positioning of periventricular dopamine neurons. Overall, we uncover molecular principles shaping the developmental architecture of the hypothalamus and show how neuronal heterogeneity is transformed into a multimodal neural unit to endow a virtually infinite adaptive potential throughout life.

The cellular and molecular landscape of hypothalamic patterning and differentiation from embryonic to late postnatal development

“…To identify region-specific differences between mature oligodendrocytes or astrocytes of the hypothalamus and other brain regions, previously published cortical scRNA-seq 12,13 were used to identify differential gene expression, using age as the variance with default parameters (first-pass gene lists) 14 . Published hypothalamic datasets 13,[15][16][17] were then used to compare to these cortical scRNAseq datasets using the identified genes (second-pass gene lists). Identified differential genes were then validated by matching the Allen Brain in situ atlas data using co-coframer 11 to validate spatial expression of these differential genes, as some observed differences could reflect batch effects resulting from scRNA-seq library preparations from different laboratories (third-pass gene lists).…”

“…To identify the developmental origins of ependymal cells and subtypes of tanycytes, transcription factors that are highly expressed near the midpoint of the pseudotime branch-where cells are not full mature tanycytes but no longer gliogenic progenitors-were extracted and clustered to scRNA-seq data obtained from adult tanycytes 16,18 using Garnett v0.2.9 19 .…”

“…This process was necessary to identify cells comprising specific hypothalamic nuclei, since at older ages, genes that are highly informative at providing information about spatial localization within the hypothalamus are often no longer highly expressed. This molecular stepping stone approach was validated by identification of glutamatergic VMH neurons using the public droplet-based scRNA-seq dataset 16,18 . Spatial information of the identified clusters were validated by comparison to the Allen Brain in situ atlas data using co-coframer 11 , as well as by matching to the published SMART-seq data obtained from the adult VMH 24 .…”

“…Progress in this area has been hampered, however, by the fact that hypothalamic cell types thus far have remained quite poorly characterized, despite recent efforts aimed at using scRNA-seq to classify cells in different regions of the adult hypothalamus [11][12][13][14][15] , and more recently in limited embryonic and early postnatal periods 16,17 . Still less is known about how hypothalamic cell types acquire their identities during development.…”

FGF21 signaling in glutamatergic neurons is required for weight loss associated with dietary protein dilution

“…In silico analysis of deposited single-cell RNA sequencing. Single cell RNA sequencing data of the whole hypothalamus was downloaded from GEO (accession number: GSE132730) 33 and loaded into the R package Seurat (v3.1). Klb expressing cells in the "Neurons" sub-class of data were then isolated.…”