Towards the Delivery of Genetically-Encodable Protein Degraders to Undruggable Targets via Lipid Nanoparticles

Introduction: Targeted protein degradation (TPD) methods, such as PROTACs and molecular glues, induce close proximity between E3 ubiquitin ligases and proteins of interest (POI), sending the POI to degradation via the proteasome. These technologies enable the degradation of pathogenic proteins driving disease, but are only able to target a small percentage of the proteome due to the limited accessibility of small molecule binders, which require the target protein to be both structurally ordered and contain a binding pocket. To overcome these challenges, we have developed ubiquibodies (uAbs) which fuse computationally-derived, target-specific peptides guides to the catalytic domains of E3 ubiquitin ligases, creating a genetically-encodable and programmable TPD platform for any target of interest. Still, for this technology, delivery of genetically-encodable payloads remains a challenge. Recently, spurred on by the COVID-19 pandemic, lipid nanoparticles (LNPs) have emerged as a potent delivery system of mRNA. In this work, we develop an LNP platform for the treatment of pediatric sarcomas, most notably Ewing sarcoma and fusion-positive alveolar rhabdomyosarcoma (FP-ARMS), by encoding uAbs targeting undruggable, driver fusion oncoproteins as mRNA into LNP formulations and demonstrating potent and functional target degradation in clinically-relevant cell lines. Our work motivates delivery of our uAbs in vivo for therapeutic development.

Materials and

Methods: We utilize our lab’s generative language models, namely PepMLM, to generate peptide guides to EWS-FLI1 (the driver fusion oncoprotein of Ewing sarcoma) and PAX3-FOXO1 (the driver of FP-ARMS) using the target sequence alone, and subsequently fuse these peptides upstream of the CHIP E3 ubiquitin ligase catalytic domain. PepMLM has been validated on multiple undruggable, disordered targets prior to this work, providing confidence in this approach. We further utilize standard Western blotting techniques with verified antibodies to the target proteins, and an Annexin V staining kit to evaluate apoptosis induction in these cells via flow cytometry.

We further synthesized LNPs using a four-lipid cocktail consisting of ionizable lipid (DLin-MC3-DMA), phospholipid (DSPC), cholesterol, and PEGylated lipid (DMG-PEG2000), mixed in optimized molar ratios. LNPs were produced by microfluidic mixing of lipid ethanol solutions with aqueous mRNA solutions, facilitating spontaneous encapsulation. The mRNA, encoding uAbs against specific oncoproteins and a control GFP, was transcribed in vitro using a T7 RNA polymerase system and purified through oligo(dT)-cellulose chromatography. The finalized LNPs were characterized by dynamic light scattering for size and zeta potential, and their mRNA encapsulation efficiency was assessed via RiboGreen assay. This setup ensures the targeted delivery of functional mRNA, enabling effective protein degradation in pediatric sarcoma models.

Results, Conclusions, and Discussions: We first identify candidate uAb fusions that systematically demonstrate functional degradation of these proteins when transfected as plasmids in A673 Ewing sarcoma cells and RH4 ARMS cells, respectively, via Western blotting and Annexin V staining. As a result of these experiments, we downselect the top three uAbs for each target. Next, we encode these top uAbs as mRNA (alongside control GFP mRNA) into the aforementioned LNP formulation to deliver our uAbs as mRNA into FP-ARMS and RH4 cells, demonstrating high uptake efficiency (via flow cytometry of GFP) and functionally potent degradation of fusion oncoproteins, EWS-FLI1 and PAX3-FOXO1 via Western blotting and apoptosis induction via Annexin V staining. Here, we establish a potential, all-in-one therapeutic platform for degradation of undruggable pediatric sarcoma targets, via our CRISPR-analogous uAb platform. Notably, our system can design a peptide (and therefore a uAb) to any target protein, enabling encapsulation of potent degraders via LNPs and subsequent in vivo delivery. Our next steps will be to deliver our uAbs into both cell-derived xenograft (CDX) and patient-derived xenograft (PDX) models of various rare diseases, such as Ewing sarcoma and FP-ARMS, to demonstrate reduction of disease phenotype. In total, we envision that our algorithm-to-in vivo pipeline can serve a general treatment platform for diseases considered undruggable by small molecule means.

Acknowledgements (Optional):