Hemoglobinopathies, including sickle-cell disease (SCD) and β-thalassemia (TDT), represent
severe genetic disorders impacting a significant population worldwide, with few effective
therapeutic options available. Recent research has identified CHD4, a chromatin-
remodeling ATPase, as a promising target for these diseases due to its role in repressing γ-
globin—a fetal hemoglobin isoform that ameliorates disease pathology. However, no
specific inhibitors of CHD4 are known. Auxiliary domains such as the HMG domain, which is
essential for CHD4 activity, represent possible targets for inhibitor design, but the small size
and convex surface of this domain makes inhibitor design challenging via conventional drug
discovery methods.
To address this challenge, we have combined RoseTTAFold Diffusion (RFDiffusion), an
advanced deep learning-based protein design algorithm, with a high-throughput plate-
based screening platform to accelerate the discovery and validation of candidate CHD4
inhibitors. RFDiffusion was utilized to rapidly generate 60 de novo protein scaffolds
predicted to bind to the CHD4 HMG domain. The high-throughput assay enabled systematic
assessment of binding affinities and inhibitory effects, facilitating the prioritization of lead
candidates for further development.
Here we describe our progress to date in this endeavor, which holds promise as a general
approach for the development of tool compounds that could inform future therapeutic
design.
Keywords: RFDiffusion, CHD4, Hemoglobinopathies, Protein Design, High-Throughput
Screening, Chromatin Remodeling.