Multiomic profiling defines cell fate plasticity of in vitro-derived islets

Abstract

Transplantation of insulin-secreting {beta}-cells differentiated from human pluripotent stem cells holds great potential as a cell therapy for treating insulin-dependent diabetes. While these stem cell-derived islets (SC-islets) are able to reverse diabetes in animal models, they are not fully equivalent to their in vivo counterparts. To better define the state of the cell types generated within these SC-islets and provide a resource for identifying deficiencies in lineage specification, we used single-cell multiomic sequencing to simultaneously measure the chromatin accessibility and transcriptional profiles of SC-islets at multiple time points as well as primary human islets. The integrated analysis of both the transcriptional and chromatin landscape for each cell provided greater resolution for defining cell identity, allowing us to derive novel gene lists for identifying each islet cell type. Furthermore, this multiomic analysis revealed that the difference between SC-{beta} cells and enterochromaffin-like cells, which are a major off-target from in vitro differentiation, is a gradient of progressive cell states rather than a stark difference in identity. The chromatin landscape of primary human islets was much more restricted, suggesting that stem cell-derived cells are not fully locked into their cell fate. While long term culture of SC-islets both in vitro and in vivo does close overall chromatin state, only in vivo transplantation directs cells toward their correct identities. Collectively, our multiomic analysis demonstrates that both the chromatin and transcriptional landscapes play significant roles in islet cell identity, and these data can be used as a resource to identify specific deficiencies in the chromatin and transcriptional state of SC-islet cell types.