Quantitative Structure-Nanoparticle Assembly Prediction of Protein Degraders

Proteolysis-targeting chimeras (PROTACs) are catalytic protein degraders with promising preclinical activity. The clinical translation of PROTACs has been limited by poor pharmacologic properties and toxicities, in part due to their “non-druglike” characteristics, including large molecular weights. We found that the vast majority of PROTACs can self-assemble into nanoparticles, yielding nanoparticle PROTACs (nanoPROTACs) with ultra-high drug loadings. While PROTAC molecular features can be deleterious to their pharmacokinetic properties, we found that they can drive nanoencapsulation more efficiently than FDA-approved small molecule drugs. Using structure-based prediction algorithms, we identified spatial autocorrelation molecular descriptors that defined nanoPROTAC formation with 96% sensitivity at 100% specificity. NanoPROTACs, targeted to the tumor microenvironment via P-selectin, led to significantly enhanced tumor drug uptake, target degradation, tumor growth inhibition, and overall survival in solid tumor xenografts. These findings offer a broad strategy to improve the pharmacologic properties and therapeutic index of PROTACs and potentially other non-druglike experimental therapeutics.