Tissue morphology influences the temporal program of human cerebral organoids

Organs are the ensemble of different cell types in a complex architectural milieu. It is well-known that progression through fate decisions sets up the complex cellular makeup and architecture of an organ, but how that same architecture may impact on cell fate is less clear. We sought to examine this by taking advantage of the unique capabilities of organoids as a tractable in vitro model to interrogate how fate and form interact during organ development. Screening methodological variations encountered in the literature revealed that common protocol adjustments, such as small molecule patterning and exposure to exogenous extracellular matrix, impacted various aspects of morphology, from macro structure to tissue architecture. We demonstrated that overall morphology is a predictor of tissue architecture and that perturbing morphology results in changes in cytoarchitecture. Vice-versa, perturbing cytoarchitecture by mechanically redistributing cells in a random spatial conformation resulted in a simplified overall morphology. We next examined the impact of these morphological perturbations on cell fate through integrated snRNA-seq and spatial transcriptomics within the phenotypic landscape. Regardless of the specific protocol, organoids with more complex morphology better mimicked in vivo human fetal brain development than organoids with simplified morphology. Further, organoids with perturbed tissue architecture profiled with scRNA-seq over a time course experiment displayed aberrant temporal progression in cell fate, with cells being intermingled in both space and time. Finally, imparting a simplified morphology through physical encapsulation led to disrupted tissue cytoarchitecture and a similar abnormal temporal progression. These data not only point to the importance of tissue morphology in organoid fidelity compared to in vivo, but also demonstrate that cells require proper spatial coordinates in order to undergo the proper temporal trajectory of events. Overall design: Single organoids were incubated in 1 mL Accumax solution (Merck, A7089) supplemented with DNase I (Sigma, 04716728001) 1.25 U/mL. Organoids were incubated in the solution at 37°C for 20 minutes on a shaker followed by gentle agitation and pipetting. The enzymatic activity of Accumax was stopped by adding IDM+A or FBS (fetal bovine serum Merck, F2442) 1:10. The solution was filtered through 70 µm cell strainer prior to cell counting. The solution was centrifuged at 200g for 10 minutes at 4°C and the pellet resuspended in cell prefixation buffer (Parse Biosciences). Sample fixation and split-seq were performed according to the Evercode Fixation v3 Parse Biosciences protocol. Control (COs) were collected at day 45, 55, 70 for sequencing. Dissociated (DISS) organoids were collected for sequencing at day 48, 55, 70, and Constrained (CONS) organoids were collected at day 55. Parse Biosciences Evercode whole transcriptome WT was used for COs, CONS, and DISS single cells. For nuclear split-seq, nuclei were isolated as described above, and sample fixation and split-seq were performed according to the Evercode Fixation v3 Parse Biosciences protocol for single nuclei. Nuclei were sequenced with Parse Biosciences Evercode whole transcriptome WT Mini. 20,000 total nuclei and 60,000 total cells were loaded for split-seq. Single cells were sequenced on a S2 Novaseq lane (4 billion reads depth). Single nuclei were sequenced on a SP Novaseq lane (400 million reads depth). N=2 organoids per batch (3 batches) per condition per time point exception made for day48 (2 batches), day55 COs (5 batches), day55 DISS (2 batches), day55 CONS (3 batches, 5 organoids). For further details see attached metadata excel spredsheats and Parse Biosciences computational pipe line output (submitted as suplmentary files).