AI-powered discovery and structure-based optimization of new 7-hydroxycoumarine-based AKR1C3 inhibitors able to reverse chemoresistance in aggressive and rare tumors

Aldo-Keto Reductase 1C3 (AKR1C3) is an overexpressed enzyme in prostate and various other cancers, playing a crucial role in driving cancer growth by regulating steroid hormone synthesis and prostaglandin signaling. Its involvement in drug resistance further contributes to tumor aggressiveness and invasiveness, making AKR1C3 an attractive target for drug design and development.[1] To expand our portfolio of AKR1C3-targeting scaffolds, we leveraged an AI-driven virtual screening and identified a potent, selective inhibitor, compound 1, featuring a 7- hydroxycoumarin core (Figure). This molecule not only demonstrated strong enzymatic inhibition but also exhibited significant antiproliferative effects in the 22RV1 prostate cancer cell model, as well as synergistic activity with doxorubicin (doxo) in osteosarcoma (OS) cell lines. Notably, in both models AKR1C3 was upregulated. As in Figure, the IC50 of doxorubicin on SAOS-2 was reduced from 80 to 2 nM when co-administered with compound 1.[2] In addition, in the in-house obtained doxo-resistant HOS, KHOS and MG63 lines, where AKR1C3 resulted overexpressed, compound 1 was observed restoring doxo sensitivity as well as ROS production, leading to cell apoptosis. Building on the crystallographic pose of compound 1 in the AKR1C3 active site, we initiated a SAR study, designing a series of analogs to enhance the potency of AKR1C3 inhibition. In this work, we will showcase the structure-driven in silico design, optimized synthesis, and in vitro evaluation on purified enzyme and tumor models of the novel 7-hydroxycoumarin scaffold-based inhibitors, highlighting their potential to advance AKR1C3-targeted chemotherapy.