Class I histone deacetylases (HDACs) are key regulators of cell proliferation and they are frequently dysregulated in cancer cells. We report here the synthesis of a novel series of class-I selective HDAC inhibitors (HDACi) containing a 2-aminobenzamide moiety as a zinc-binding group connected with a central (piperazin-1-yl)pyrazine or (piperazin-1-yl)pyrimidine moiety. Some of the compounds were additionally substituted with an aromatic capping group. Compounds were tested in vitro against human HDAC1, 2, 3, and 8 enzymes and compared to reference class I HDACi (Entinostat (MS-275), Mocetinostat, CI994 and RGFP-966). The most promising compounds were found to be highly selective against HDAC1, 2 and 3 over the remaining HDAC subtypes from other classes. Molecular docking studies and MD simulations were performed to rationalize the in vitro data and to deduce a complete structure activity relationship (SAR) analysis of this novel series of class-I HDACi. The most potent compounds, including 19f, which blocks HDAC1, HDAC2, and HDAC3, as well as the selective HDAC1/HDAC2 inhibitors 21a and 29b, were selected for further cellular testing against human acute myeloid leukemia (AML) and erythroleukemic cancer (HEL) cells, taking into consideration their low toxicity against human embryonic HEK293 cells. We found that 19f is superior to the clinically tested class-I HDACi Entinostat (MS-275). Thus, 19f is a new and specific HDACi with the potential to eliminate blood cancer cells of various origins.

Class I histone deacetylases, HDAC1, HDAC2, and HDAC3, represent potential targets for cancer treatment. However, the development of isoform-selective drugs for these enzymes remains challenging due to their high sequence and structural similarity. In the current study, we applied a computational approach to predict the selectivity profile of developed inhibitors. Molecular docking followed by MD simulation and calculation of binding free energy was performed for a dataset of 2-aminobenzamides comprising 30 previously developed inhibitors. For each HDAC isoform, a significant correlation was found between the binding free energy values and in vitro inhibitory activities. The predictive accuracy and reliability of the best preforming models were assessed on an external test set of newly designed and synthesized inhibitors. The developed binding free-energy models are cost-effective methods and help to reduce the time required to prioritize compounds for further studies.

This review outlines integrated platforms combining genome mining, computational design, synthetic biology, and high-throughput screening to enable the programmable biosynthesis of tailor-made natural products.