Controlling Cell Fate During Directed Differentiation of Human Beta Cells

Abstract

Human beta cells derived from pluripotent stem cells have the potential to cure type 1 diabetes. Considerable progress has been made in differentiating insulin-secreting beta cells that can rescue animal models of diabetes, but these develop as a subset of cells within heterogeneous stem cell-derived (SC-) islets. Defining and controlling the cell composition of SC-islets is key to their future therapeutic application. In this dissertation, we apply first single cell RNA sequencing (scRNA-seq) to characterize cell populations within SC-islets, and then CRISPR screening to identify genes controlling their development.

First, we present a comprehensive scRNA-seq mapping of cell identities during directed differentiation of stem cell-derived beta cells. SC-beta cells, once formed, retain their identity in the absence of signaling modulators and gain expression of genes associated with a more functional, mature state. Cells previously referred to as polyhormonal are immature SC-alpha cells, and we show that SC-alpha cells are as similar to in vivo alpha cells as are SC-beta cells to their own counterparts. We provide the first identification of SC-enterochromaffin (EC) cells – serotonin-secreting enteroendocrine cells – as a population of comparable abundance to SC-beta cells. Finally, we develop a strategy for SC-beta cell enrichment using magnetic sorting on the marker CD49a.

Next, we present a genome-wide CRISPR knockout screen for genes controlling cell fate during SC-beta differentiation. Through bulk, flow-cytometry-based screens, we identify 244 genes whose loss-of-function alters the likelihood of differentiating into SC-beta, SC-alpha or SC-EC cells. Using Perturb-Seq, we expand the characterization of these perturbations by measuring their impact across all populations at several timepoints in differentiation. We generate pluripotent stem cell lines with increased SC-beta differentiation propensity by engineering homozygous deletions in the coding sequence of FBXL14. Finally, we compare loss of function observations with overexpression for key transcription factors.

Overall, this work resolves open questions about the nature of key cell types produced during SC-beta differentiation and provides a systematic exploration of the genes underlying their specification. Leveraging this map to guide genetic engineering of pluripotent stem cells provides a new layer of control in the creation of stem cell-derived islets optimized for therapeutic applications.