Bespoke Micellar Nanoparticles (MNPs) for In Vivo Delivery of Digital Base Editor Treatments for Primary Immunodeficiency Disease - ABSTRACT Digital base editors hold tremendous promise for precision gene therapies for monogenic diseases, such as severe combined immunodeficiency A (SCID-A), a primary immunodeficiency (PID). PIDs can be treated with allogeneic hematopoietic stem cell transplant (HSCT), however, >70% of patients lack suitable matched donors and of those transplanted, many encounter severe complications, such as graft-versus-host-disease and even death. An alternative and ideal approach would be genetic correction of hematopoietic stem/progenitor cells (HSPCs) in vivo, yet base editing therapies require tandem delivery of two component payloads (mRNA encoding base editor protein along with gRNA). Major clinical challenges include safe and effective development of delivery systems for two-component digital base editors, which involves packaging uniformity/stability, formulation optimization, safe/effective editing performance in vivo, and collective elucidation/prediction of structure-activity relationships. The objective of this application is to develop a build-test-learn-predict-refine closed loop machine learning (ML)-driven workflow of synthesis, formulation generation, and screening that will optimize micellar nanoparticles (MNPs) for binary base editor (gRNA and mRNA expressing adenine base editor, ABE) delivery in vitro with a novel engineered cell model and in vivo to HSPCs using a novel genetically humanized SCID-A mouse model (all developed by our team). To accomplish this, we will synthesize a novel uniform library of amphiphilic polymers for blended assembly into MNP vehicles. The proposed work is innovative and unique as a high throughput workflow will be combined with a novel multi-fidelity ML model to drive our synthesis/formulation strategy that will uniquely correlate chemistry/composition to size/stability (Aim 1), in vitro (Aim 2), and in vivo (Aim 3) performance relationships. The Specific Aims of this proposal are: Aim 1. Driven by ML, a novel library of precise amphiphilic block polymers will be synthesized, assembled into a library of compositions, formulated into MNPs with gRNA and mRNA expressing ABE, quantitatively characterized, and optimized for physicochemical properties. Aim 2. A novel engineered SCID-A cell model will enable quantitation and correlation of MNP cellular internalization, ABE editing, and toxicity to physicochemical properties via ML to identify and predict structural drivers of performance and safety in vitro. Aim 3. A novel genetically humanized SCID-A mouse model will enable quantitation and ML prediction of MNP physicochemical properties to functional in vivo ABE editing, immune system recovery, toxicity, and in vitro-in vivo performance trends. Our world-leading team is particularly well suited to undertake the proposed studies providing expertise in polymer synthesis and physicochemical characterization (Reineke), ML and computation (Sarupria), along with in vitro and in vivo model development, ABE therapies, and translational research (Moriarty and McIvor). Our pioneering and versatile MNP platform, ML, and biological discovery tools enable our long-term goals of achieving well-defined, safe, stable, and effective in vivo base editing therapy for PIDs currently treated by HSCT.
