Objective
The objective of this collaborative research project is to develop and validate a toolbox of AI-designed mini proteins called EpiBinders that can specifically target key epigenetic regulators involved in islet cell development and functional maturation.
Specifically, the overarching goal of our efforts is to use these newly designed EpiBinders alongside molecules known as "guide RNAs," which function as ZIP codes, in SCs to enable the precise activation of genes that will “turn-on” insulin production. As a result, the process of converting SCs into insulin-producing cells can be greatly accelerated.
Conduct rigorous in vitro validation to ensure the effectiveness and specificity of the EpiBinders in regulating the epigenetic landscape of SCs towards islet cell differentiation and functional maturation.
Apply and validate the toolbox of EpiBinders to regulate the activation of gene networks supporting the generation of more homogenous and islet tissue.
Characterize states of differentiation and functional maturation of epigenetically-manipulated SC-derived islets, and test their endurance to metabolic/inflammatory stressors, both in vitro and in vivo in cell transplantation models.
Background Rationale
A major hurdle in the production of islet tissue from SCs is the high variability between different SC lines for their propensity to efficiently produce functional islet tissue. This challenge is likely linked to the pattern of genome-wide activity, i.e. genes that are silenced and/or that are accessible for activation (a DNA state referred to as methylome) existing between different SC lines. This constitutes a barrier that current factor-based protocols of differentiation are unable to overcome across different SC lines. Based on our preliminary work, we postulate that our intervention with AI-designed EpiBinders at distinct stages of SC differentiation into islet tissue will eliminate or greatly reduce these bottlenecks and efficiently recruit larger pools of progenitors into more homogeneous downstream lineages, at each stage of differentiation, ultimately promoting a more robust production of functional islet cells. This AI-designed EpiBinders approach eliminates the “off-target” effects observed in methods involving overexpression of epigenetic enzymes, allows the enhancement of selected/targeted genes, and establishes epigenomic inheritance throughout cell division and subsequent differentiation into islet cells.
Our preliminary work provides proof of principle that a timed epigenetic de-repression of PDX1 and NGN3 genes at the stage of pancreatic progenitors and islet progenitors, respectively, increases the fraction of insulin-producing beta-cells over immature progenitors, and accelerate their development when compared to current protocols of differentiation. Based on these results, we also anticipate that this AI-designed EpiBinders intervention can be used to regulate the expression of other genes at later stages of islet cell differentiation to promote and accelerate the functional maturation of beta-cells, i.e., the acquisition of glucose-induced insulin secretion.
Collectively, these considerations provide a strong rationale for investigating the use of AI-designed EpiBinders as highly specific and efficient drivers of SC differentiation into functional islet tissue for transplantation in T1D.
Description of Project
Stem cell (SC)-based replacement therapy is emerging as promising cure for Type 1 diabetes. However, there are challenges with the current methods that include variable difficulties in controlling the development of SCs into the specific types needed for treating diabetes, as well as variations in how individual SC lines respond to signals that are supplied to cox them to become islet beta-cells. This results in a mixture of different types of cells, some of which may not be fully developed or functional. To address these challenges, we are working on a revolutionary new approach to improve the efficiency of deriving insulin-producing cells from stem cells. Our approach involves using Artificial Intelligence to design small proteins called EpiBinders. In our preliminary research, we found that the utilization of EpiBinders alongside molecules known as "guide RNAs," which function as ZIP codes, enables the precise targeting of EpiBinders to specific genomic regions. This targeted approach facilitates the activation of select genes of interest which will “turn-on” insulin production. As a result, the process of converting SCs into insulin-producing cells can be greatly accelerated. Accordingly, our initial experiments show that targeting certain genes important for the acquisition of a beta-cell identity can increase the production of insulin-producing cells and shorten almost by half the time required to turn stem cells into beta-cells. Our studies will also assess how well these EpiBinders-based intervention endows SC-derived beta-cells with the ability to resist and endure metabolic stressors related to diabetes, such as high glucose and inflammatory cues, both in lab settings and in cell transplantation experiments.
In summary, our project aims to improve the process of turning stem cells into insulin-producing cells by using Artificial Intelligence to design specialized proteins that target specific genes involved in cell development and functional maturation.
Anticipated Outcome
Work proposed under this collaborative research project will fill an important gap in knowledge by validating the efficacy of novel miniprotein binders (EpiBinders) in targeting and activating specific regulators of the epigenome, that in turn will allow for their use in efforts to accelerate the production of functional islet tissue for transplantation in T1D. Based on the combinatorial strategy of our experimental plan, we anticipate that the simultaneous targeting of genes regulating early commitment of SC toward the pancreatic islet cell lineage, and the activation of genes that control the acquisition and maintenance of islet cell metabolic performance and endocrine function, will greatly advance effort to widen the application of cell replacement therapy to a larger population of T1D patients. Ultimately, we also anticipate that our intervention will accelerate the production of islet tissue from SC, a result that would greatly reduce costs of production for transplantation.
This expected outcome is supported by our preliminary work that demonstrates a striking increase in the number of endocrine cells (~4 to 7.8-fold) following the EpiBinders(EBdCas9)-mediated activation of PDX1 and NGN3 expression in early pancreatic progenitors and islet progenitors.
Based on the reported function of genes that regulate critical mitochondrial functions, as well as insulin production and secretion, we expect that intervention with EBdCas9-and/or TET1/2-Epibinders to de-repress the expression of these gene will result in improved glucose/mitochondrial respiration coupling and glucose-regulated insulin secretion in SC-derived beta-cells.
On the other hand, the use of EpiBinders for DNMT3A will allow to repress the expression of genes whose activity has been shown to be detrimental for the acquisition and/or maintenance of beta-cell function.
Collectively, our approach is expected to demonstrate that this epigenetic intervention with EpiBinders will lead to much higher yields of SC-derived islets that can become functional sooner and long-term in vivo after transplantation.
Relevance to T1D
Stem cell (SC)-based replacement therapy is emerging as promising cure for Type 1 diabetes. Yet, several challenges remain to be overcome for this approach to be widely applied to a large demographic of T1D patients. These challenges include the heterogeneity of SC-derived islet cell preparations that often comprise different types of cells, some of which may not be fully developed or functional. To address these challenges, we are working on a revolutionary new approach to improve the efficiency of deriving insulin-producing cells from SCs. Our approach involves using Artificial Intelligence to design small proteins called EpiBinders. In preliminary work, we discovered that when these EpiBinders are used in conjunction with molecules called “guide RNAs” that work as ZIP codes for specific genes of interest, SCs turn into insulin-producing cells at much faster rate. Accordingly, our initial experiments show that targeting certain genes important for the acquisition of a beta-cell identity with the EpiBinders can increase the production of insulin-producing cells and shorten almost by half the time required to produce beta-cells. Our studies will also assess how well these “fast-track beta-cells” resist metabolic stressors related to diabetes, such as high glucose and inflammatory cues, both in lab settings and in cell transplantation experiments.
In summary, by exploiting the power of AI-designed regulators of gene expression we anticipate that the results from our collaborative research effort will dramatically improve and accelerate the process of producing functional islet tissue for transplantation in T1D patients.