3D Forebrain-like Organoids Mimic Fetal Brain Development and Provide a Platform of Studying the Effect Autism-Associated Mutations on Corticogenesis and Chromatin Dynamics

Background: Key regulatory networks control critical aspects of human cerebral cortex formation. Cellular differentiation, migration, and cortical lamination are controlled by precise dynamic shifts in epigenetic states. Genetic risk data from de novo mutations in simplex cases of autism spectrum disorder (ASD) converge during fetal brain development and risk genes form key regulatory networks. However, our understanding of the network characteristics and how specific mutations impact brain development remains limited. This is largely due to lack of available human biological samples of relevant genotypes, cell types, and developmental stages. Induced pluripotent stem cells (iPSCs) have emerged as a powerful tool for modeling human development. Moreover, the in vitro differentiation of 3D cortical organoids from iPSCs has been shown to recapitulate the same step-wise neurodevelopmental processes that occur during in vivo corticogenesis.

Objectives: To implement a scalable 3D forebrain-like organoid model system and characterize chromatin dynamics during corticogenesis at single-cell resolution.

Methods: We implemented a scalable 3D forebrain-like organoid model system, which has been reported to mimic the same critical fetal developmental time period implicated in ASD risk. Using a modified SpinΩ bioreactor protocol (Qian et al. 2016), we sampled dozens of individual human organoids derived from control iPSCs and matured for 30, 60 and 90 days in vitro (DIV). To assay dynamic changes in epigenetic states during organoid maturation, we measured chromatin accessibility using a single-cell combinatorial indexing ATAC-seq assay (Cusanovich et al. 2015) that allows for the generation for thousands of single cells in one experiment. Further, we multiplexed many organoids per experiment, allowing for valuable characterization of inter-sample heterogeneity.

Results: We generated over 10,000 single-cell ATAC-seq chromatin accessibility profiles from forebrain-like organoids. Clustering of these single-cell profiles revealed three clear cellular populations corresponding to neuroepithelial stem cells, neuroprogenitors, and post-mitotic neurons. To relate chromatin accessibility to putative transcription factor activity, we analyzed accessibility peaks for enrichment of specific transcription factor binding motifs. A pseudotemporal ordering analysis on the single-cell profiles revealed a succession of key brain transcription factor activities involved in cortical development. For example, cells putatively expressing transcription factor EMX1 increased between 30 and 60 DIV, reflecting a cell population transition from early neuroepithelial cells to neuroprogenitors, and then to early glutamatergic neurons. We see a strong correlation between the epigenomic shifts with previously cataloged transcriptional changes and transcription factor abundance validating the use of this model. Additionally, our genome-wide assessment of chromatin accessibility revealed previously uncharacterized putative enhancer regions. We are now extending these approaches to patient-specific iPSCs with de novo mutations in key transcriptional regulators, such as TBR1.

Conclusions: We present the first single-cell resolution characterization of chromatin dynamics during corticogenesis of forebrain-like organoids. These data support the use of organoids for modeling the effect of ASD-associated mutations on corticogenesis and discovering the critical changes in brain development that lead to ASD.