University of Washington has received a USD 1.71 million NIDDK RC2 grant to develop AI-designed mini-proteins that epigenetically direct pluripotent stem cells into functional pancreatic islets, addressing a persistent bottleneck in stem cell-based diabetes therapies.
University of Washington has received a USD 1.71 million NIDDK RC2 grant to develop AI-designed mini-proteins that direct pluripotent stem cells into functional pancreatic islets through targeted epigenetic editing, aiming to address a major manufacturing bottleneck in stem cell therapies for diabetes.
Current protocols for differentiating stem cells into islet tissue produce heterogeneous cell preparations with variable proportions of functional beta-cells across different stem cell lines — a problem that limits the scalability of cell replacement approaches for diabetes. The Washington team, led by Vincenzino Cirulli, proposes to overcome this using AI-designed mini-proteins called EpiBinders, which are fused to catalytically inactive Cas9 (dCas9) and targeted to promoters of key islet developmental genes, including PDX1 and NGN3, to enforce epigenetic reprogramming via the Polycomb Repressive Complex 2 (PRC2). Preliminary data from reporter stem cell lines — engineered to allow live-monitoring of cell fate decisions — indicated that transient EpiBinder targeting accelerated differentiation into glucose-responsive beta-cells and improved yield and homogeneity across multiple stem cell clones.
The award reflects growing NIH investment at the intersection of AI-driven protein design and regenerative medicine, a space where computational tools are increasingly being applied to overcome biological variability in cell therapy manufacturing.