Defining the Teratoma as a Model for Multi-Lineage Human Development

Studying early human developmental processes requires access to a model system that is easy to perturb, either genetically or chemically, genetically tractable and recapitulates key features of organogenesis. We propose that the teratoma, a recognized standardused for for validating pluripotency in stem cells, could be a promising platform for modeling human development. Teratomas differentiate to all germ layers, have regions of complex tissue-like organization with internal vasculature, and are relatively easy to generate and thus accessible, especially in comparison to human fetal tissue. Teratomas, however, have not been characterized rigorously to determine their suitability to model human developmental processes the diversity and identities of the cell types present, and their level of maturity in comparison to human developmental cell types. We thus systematically analyzed, perturbed, and engineered human pluripotent stem cell (PSC)-derived teratomas. Usingused histology, RNA in situ hybridization, and single cell RNA-seq (scRNA-seq) of 179,632 cells across 23 teratomas generated from 4 cell lines, we observed to find that teratomas reproducibly contain approximatelyat least 20 cell types across all three germ layers, and the cell type heterogeneity between teratomas was comparable to that found in organoid systems. We found that a significant fraction of injected PSCs engrafted, suggestingused lentiviral barcoding to determine there is minimalno bottlenecking during teratoma formation. Additionally, we mapped our teratoma data to published human fetal data and showed that the teratoma gut and brain cell types correspond well to existing fetal gut and brain scRNA-seq datasetscells. We then performed a CRISPR-Cas9 knockout screen in teratomas targeting key genes known to be embryonic lethal, and found that TWIST1, RUNX1, ASCL1, CDX2, and KLF6 knockouts resulted in assessed shifts in lineage abundance. using scRNA-seq. We found that TWIST1, RUNX1, ASCL1, CDX2, and KLF6 knockouts resulted in reproducible shifts in lineage abundance consistent with known biology. All of these These knockouts had effects on cell types from multiple germ layers, showing that the teratoma can serve as a platform to studymodel all major human lineages of human development simultaneously. Additionally, we ran a CRISPR-Cas9 screen targetingWe also knocked out the genes underlying Rett, Pitt-Hopkins, and L1 Syndromes and identified cell type specific shifts in gene expression. Finally, we engineered a novel miRNA-regulated suicide gene circuit that enriches for tissues in the teratoma that express a specific miRNA that selectively depletes cells that do not express a tissue-specific endogenous miRNA, allowing us to enrich for desired lineages in the teratoma, which we demonstrated using miR-124 to enrich for neuro-ectoderm. Specifically, we used miR-124 to successfully enrich for neuro-ectoderm. Taken together, we believe the teratoma is a promising platform for modeling multi-lineage human development, functional genetic screening, and cellular engineering Overall design: We first characterized 7 teratomas generated from H1 stem cells, with 3 of those teratomas containing lentiviral barcodes for additional lineage tracing. We characterized one teratoma each from H9, HUES62, and PGP1 stem cells to assess heterogeneity across cell lines. We ran 2 replicate CRISPR-KO screens targeting embryonic lethal genes. We generated 3 teratomas (with 2 x 10X runs for each teratoma) per screen. We then ran a CRISPR-KO screen targeting genes known to cause neural diseases using 2 teratomas (again with 2 x 10X runs per teratoma). Finally, we created a miRNA circuit that enriches for neuro-ectoderm cells expressing miR-124. We generated 4 GCV+ teratomas, 2 that used intratumoral injections of GCV, and 2 that used both intratumoral and intraperitoneal injections. We generated one GCV- miR-124 teratoma as a control.