Pharmaceutical agents, such as antiepileptic medications, can cross fetal barriers and affect the developing brain. Prenatal exposure to the antiepileptic drug valproate (VPA) is associated with an increased risk of neurodevelopmental disorders, including congenital malformations and autism spectrum disorder. In animal models and neural organoids, VPA has been shown to alter signaling pathways, such as Wnt pathway, providing insights into VPA-induced neurodevelopmental defects. Here, we exposed dorsal forebrain organoids to VPA for 30 days and examined effects at the tissue, cellular, and molecular level. VPA treatment disrupted ventricular-like regions, indicating defects in cell-cell and cell-matrix interactions. Transcriptomics analysis confirmed altered expression of extracellular matrix (ECM) genes and single cell RNA sequencing analysis identified genes involved in microenvironment sensing, such as cellular mechanosensing and Hippo-YAP/TAZ signaling pathway. Finally, proteomics analysis corroborated that VPA alters the microenvironment of the human dorsal forebrain organoids by disrupting the secretion of ECM proteins. Altogether, our study suggests that VPA-treated dorsal forebrain organoids serve as a model to investigate the role of extracellular processes in brain development and to understand how their disruptions might contribute to neurodevelopmental disorders.
Summary: Neural organoids are three-dimensional stem cell-derived models widely used to study brain development and model disease. Brightfield imaging offers a non-invasive readout of their growth and morphology, and recent advances have made automated organoid segmentation from such images increasingly attainable. Yet existing tools largely leave morphological characterization, quality assessment, and comparisons between conditions to manual or ad hoc analysis. Here we present Organotchi, an open-source tool that integrates automated segmentation with longitudinal morphological analysis of neural organoids. Organotchi quantifies ten morphological features, including area, perimeter, and eccentricity, and tracks cystic structures over time, reports variability within and between batches, and flags outlier organoids and segmentation failures, supporting objective organoid characterization and enhancing downstream analysis. As a proof of concept, we apply Organotchi to disease modelling, pharmacological perturbation, and protocol optimization in neural organoids. One use case assesses drug response in breast cancer spheroids, demonstrating its use beyond its development domain.
Pontocerebellar hypoplasia (PCH) is a group of ultrarare, neurodegenerative disorders characterized by cerebellar and pontine hypoplasia. Genetic analysis over the last two decades has revealed an increasing number of pathogenic variants in a wide range of broadly expressed genes functioning in RNA processing, tRNA metabolism, and translation. However, the mechanisms linking these ubiquitous processes to brain region-specific vulnerability are unknown. Here, we established a multi-level variant-to-function in silico framework to predict the molecular consequences of PCH-associated variants in TSEN complex genes. These variants were predicted to have heterogeneous effects on diverse protein properties, including stability, subcellular localization, and degradation, supporting variant-specific rather than uniform disease mechanisms. Complementary transcriptomic analyses showed that PCH-associated genes were not globally enriched in the prenatal cerebellum. Instead, their expression was coordinated in a stage- and cell type-specific manner during cerebellar development. We therefore hypothesize that multiple PCH-associated genes are regulated by a common set of transcription factors, providing an explanation of the selective vulnerability of the cerebellum and to the phenotypic convergence of genetically diverse PCH subtypes. In summary, this study prioritizes candidate variants for biochemical, cellular, and in vivo validation, and identifies regulatory programs, cell lineages, and developmental windows for targeted, mechanistically informed disease modelling.