In silico postscreening study of antisickling agents to treat sickle cell disease

Sickle cell disease (SCD) affects about 100,000 people in the U.S. and millions worldwide. Although curative therapies exist, their application is limited by cost and donor availability, making drug therapy the most accessible option for most patients. High-throughput in vitro screenings have identified potential antisickling drugs, but these assays cannot replicate organ-specific oxygen dynamics or account for pharmacokinetics (PK) and pharmacodynamics (PD) of the drug, which are essential to determine their therapeutic efficacy in vivo. To address these limitations, we develop a computational platform that complements experimental screening by integrating PK/PD models with a kinetic model of erythrocyte sickling, enabling the identification of drug candidates most likely to succeed in clinical trials. We first demonstrate the platform's effectiveness by comparing the predicted efficacies of voxelotor and osivelotor. Our predictions show that short-term treatment interruption of voxelotor could result in a rapid elevation of the erythrocyte sickling ratio, potentially leading to the increased vaso-occlusion incidence reported in its clinical trials. In contrast, osivelotor achieves antisickling effects comparable to those of voxelotor at lower doses and enables greater tolerance to treatment interruption due to its improved PK properties. Then, we examine bitopertin, a repurposed drug, and project its antisickling efficacy to be markedly lower than hydroxyurea or voxelotor, indicating limited potential as an antisickling drug. Furthermore, our platform can predict the potential efficacy of multiagent therapies. Altogether, this in silico framework provides an efficient postscreening tool for predicting in vivo drug efficacy, thereby accelerating the discovery of new antisickling therapies for SCD.