Multimodal spatiotemporal phenotyping of human organoid development

Abstract

Organoids generated from human pluripotent stem cells (PSCs) provide experimental systems to study development and disease. However, we lack quantitative spatiotemporal descriptions of organoid development that incorporate measurements across different molecular modalities. Here we focus on the retina and use a single-cell multimodal approach to reconstruct human retinal organoid development. We establish an experimental and computational pipeline to generate multiplexed spatial protein maps over a retinal organoid time course and primary adult human retina, registering protein expression features at the population, cellular, and subcellular levels. We develop an analytical toolkit to segment nuclei, identify local and global tissue units, infer morphology trajectories, and analyze cell neighborhoods from multiplexed imaging data. We use this toolkit to visualize progenitor and neuron location, the spatial arrangements of extracellular and subcellular components, and global patterning in each organoid and primary tissue. In addition, we generate a single-cell transcriptome and chromatin accessibility time course dataset and infer a gene regulatory network underlying organoid development. We then integrate genomic data with spatially segmented nuclei into a multi-modal atlas enabling virtual exploration of retinal organoid development. We visualize molecular, cellular, and regulatory dynamics during organoid lamination, and identify regulons associated with neuronal differentiation and maintenance. We use the integrated atlas to explore retinal ganglion cell (RGC) spatial neighborhoods, highlighting pathways involved in RGC cell death. Finally, we show that mosaic CRISPR/Cas genetic perturbations in retinal organoids provide insight into cell fate regulation. Altogether, our work is a major advance toward a virtual human retinal organoid, and provides new directions for how to approach disorders of the visual system. More broadly, our approaches can be adapted to many organoid systems.

Comprehensive characterization of fetal and mature retinal cell identity to assess the fidelity of retinal organoids

“…One of the key limitations of the study is that the fidelity metrics do not address important features of the retina, such as the profiles of the other omics layers, spatial patterning, retinal organoid functionality, and domain-specific characteristics. Recently, a few studies have begun profiling the retinal organoid using multi-modal single-cell technologies ( Thomas et al., 2022 ; Wahle et al., 2022 ). The study by Thomas et al.…”

Identification of Evolutionarily Conserved VSX2 Enhancers in Retinal Development

“…Retinal organoids serve as a powerful tool for modeling human retinal development and disease (Foltz and Clegg, 2019, (Kruczek and Swaroop, 2020, (Wahle et al, 2023, (Ludwig et al, 2023, (Saha et al, 2022, (Phillips et al, 2014). To understand the role of the Vsx2 SE modules in human retinal organoids, we individually deleted Region 0 or Region 3 and assessed their growth and development over three established developmental stages (Capowski et al, 2019).…”

Robust phenotyping of highly multiplexed tissue imaging data using pixel-level clustering

“…In Gut et al, the authors also use a SOM to cluster pixels in iterative indirect immunofluorescence imaging (4i) images and relate these pixel features to cells, terming them Multiplexed Cell Units (MCUs). 6,44 In spatial transcriptomics, pixel-level features have been used to perform cell annotation and infer tissue substructures. [45][46][47][48] Furthermore, in the spectral imaging field (e.g.…”