Endocrine Abstracts

Thyroid hormone (TH) signaling is essential for nervous system development. The clinical relevance of this simple statement is underscored by the severe neurological phenotypes resulting from global TH deficiency or perturbed local TH signaling during fetal development. A major aim of our research is to better understand the local action of TH during early human cortex development using cerebral organoids derived from human induced pluripotent stem cells (hiPSC) as a model system. We show that these organoids recapitulate major cytoarchitectural and transcriptional aspects of early cortex morphogenesis. Using single cell RNA-seq, we characterized TH-responsive gene expression programs with temporal and cell type-specific resolution and identified TH-induced changes in the relative abundance of neuronal cell types. Exploiting the laminar organization of progenitor and neuronal cell types in organoids, we next aimed to register TH-dependent gene expression programs in a spatial dimension. We use single molecule fluorescent in situ hybridization (smFISH) in combination with informative cell layer markers to characterize spatial expression profiles. Cryosections of organoids cultured in presence of a range of T3 media levels (0 to 20 nM T3) were stained by immunofluorescence or smFISH and high-resolution images were acquired by confocal microscopy. We next developed a quantitative image analysis pipeline based on tools implemented in ImageJ and QuPath software permitting (I) detection of changes in the relative abundance of neuronal cells types in specific cell layers and (II) quantification of spatial gene expression levels. One of the most critical steps in quantitative image analysis is a precise cell segmentation. Here, densely packed neuronal progenitor populations proved a formidable challenge that was successfully overcome by application of deep learning approaches. We will report results from a first series of proof-of-concept experiments where we document changes in neuronal subtype abundance in response to different T3 media levels. In addition, we devised an analytic pipeline to quantify changes in regional gene expression for T3-responsive genes (i.e. DIO3, SEMA3C, EPHA4) in developing organoids. Registering the spatial patterns of T3-induced gene expression will complement the analysis of cell type-specific gene expression changes in order to reveal the potential role of layer-specific niche characteristics in shaping coordinated responses of cortical tissue to varying TH levels.