CROSS-SITE REPRODUCIBILITY AND FIDELITY OF HUMAN CEREBRAL ORGANOIDS

Human cortical organoids have been adapted to model many neurodevelopmental disorders and viral infection. However, a lack of reproducibility across cell lines, across studies of the same neurodevelopmental disorder, and the plethora of available organoid protocols has called into question how well human organoids model the developing brain and what the sources of variability are. For my dissertation, I studied the cross-site reproducibility of human cortical organoids in collaboration with two external research groups. We show that cortical organoids generate cortical progenitors and neuronal subtypes with moderate to high fidelity to the developing human cortex. The same induced pluripotent cell (iPSC) line grown across three sites generated reproducible cortical cell types in similar proportions with consistent cortical wall-like structure. Gene expression associated with metabolism and cell stress varied in day 84 organoids, and early stem cell states influenced cell type proportions at two weeks of differentiation. Known limitations of organoids models, like hypoxia induced cell stress, necrotic cores and non-cortical differentiation in ~15% of cells were also reproducible across sites. To test for how well the same organoid protocol models inter-individual variation in brain traits, I started a separate collaboration with the Infant Brain Imaging Study (IBIS) and the Human Pluripotent Stem Cell Core to generate iPSC lines from participants with longitudinal MRI measurements of brain growth. We presumed that if organoids are a faithful model of corticogenesis, organoid phenotypes measuring progenitor production or proliferation would correlate increases in cortical surface area as predicted by the radial unit hypothesis. I differentiated organoids from clonal iPSC lines derived from 18 unique IBIS participants. After removing non-cortical differentiations by using visible neuroepithelial budding as an inclusion criteria for the analysis, I identified a correlation between cross-section area growth rate and cortical surface at 24 months. Future analysis of 3D imaging and scRNAseq data collected from the same organoids will also attempt to identify in vitro correlates to brain traits. Identification of early changes in stem cell or organoid culture that predicts strong cortical induction later in the differentiation will make cortical organoids a better and more reliable model of human corticogenesis.