AI-designed mini-proteins as drivers of pancreatic islet production from pluripotent stem cells

About This Grant

The ability to differentiate Stem Cells (SC) into functional pancreatic islets highlights new opportunities to address the tissue shortage for cell replacement therapies in diabetes. However, a barrier to the broad applicability of this approach across the human population is the variable efficiency of current protocols in controlling lineage biases and functional maturation of different SC lines, thereby yielding heterogeneous islet cell preparations containing variable proportions of endocrine and immature cell types. In this collaborative project, we propose to broaden the therapeutic potential of SC-based treatments by developing a new epigenetic approach to enhance the efficiency of islet tissue derivation from SCs. Our innovative approach will generate SC-derived pancreatic islets by activating, in a stepwise manner, key islet cell developmental and survival programs using Artificial Intelligence (AI)-designed mini-proteins (EpiBinders) capable of enforcing epigenetic regulation of select islets’ gene networks. In preliminary proof-of-principle studies, using newly developed reporter SC lines allowing live-monitoring of fate choices during differentiation, we provide evidence that transient targeting of an (AI)-designed mini-binder of the Polycomb Repressive Complex 2 (PRC2) fused to dCas9 (EBdCas9) to PDX1 and NGN3 promoters significantly accelerates the differentiation of SC lines into glucose-responsive islet -cells, increasing the yield and homogeneity of islet tissue from multiple SC clones. Building on these findings, our goals are to develop a pipeline of functionally validated AI-designed EpiBinders for epigenetic programming of SCs into islet tissue, establish protocols for EpiBinders’ delivery at select stages of SC differentiation, and validate their specificity and efficiency as drivers of islet cell development and function in multiple SC lines, both in vitro and in vivo in cell transplantation models. We anticipate that the new tools, knowledge, and resources developed through this interdisciplinary effort will significantly broaden the applicability of SC-based therapies for diabetes and for a wide range of degenerative diseases, while establishing impactful platforms that advance discovery and enable future innovation.