Introduction: Histone deacetylase 6 (HDAC6) and enhancer of zeste homolog 2 (EZH2) are crucial epigenetic regulators in cancer, with their synergistic roles playing a significant part in tumorigenesis and progression. However, dual-target peptide inhibitors that target both HDAC6 and EZH2 have not yet been developed. In this work, we present an AI‐driven de novo design strategy to develop novel peptide inhibitors that target HDAC6 and EZH2. Methods: Our approach used RFDiffusion to generate diverse peptide conformations, followed by ProteinMPNN to predict the most probable amino acid sequences. The HDAC6 and EZH2- binding peptides were then fused using linker sequences, and AlphaFold3 was employed to predict the resulting structures, thereby validating the feasibility of the fusion design and identifying high-confidence peptide candidates. Results: Through 100 ns molecular dynamics simulations, three peptide inhibitors were identified, and DTP-1 showed superior stability and binding free energies of －88.96 kcal/mol (HDAC6) and －62.53 kcal/mol (EZH2). Discussion: These findings substantiate the feasibility of an AI-guided workflow for designing dual-target peptide inhibitors that establish a foundation for enhancing the efficacy of epigenetic combination therapies and accelerating the development of dual-target inhibitors. Conclusion: This study introduces novel peptide inhibitors targeting HDAC6 and EZH2, along with a generalizable strategy for designing dual-target peptide inhibitors. DTP-1 emerges as a promising lead inhibitor for further preclinical development.
Zhang et al. (2026) studied this question.