Cancer is one of the major causes of death worldwide. The number of pathways and enzymes is involved in cancer proliferation and metastasis. One such enzyme is dihydrofolate reductase (DHFR). DHFR is an essential enzyme in DNA synthesis by catalyzing the reduction of dihydrofolate to tetrahydrofolate, which, in turn, is used in the biosynthesis of nucleotides, hence serves as an attractive target for the treatment of cancer. Therefore, this study aims to identify potential inhibitors through a combination of tandem structure‐based screening of an in‐house database of natural and synthesized compounds. This approach yielded compound Qth‐90 (2‐(3,4‐dihydroxyphenyl)chromane‐3,5,7‐triol) as a seed for the design of novel inhibitors. The inhibitory potential of the seed molecule was further evaluated through in vitro biochemical assays, demonstrating 90.64% inhibition with IC50 value of 11.10 ± 0.24 μM. To enhance the inhibitory potential of the seed molecule, the molecule was subjected to the deep learning web server WADDIACA for the generation of de novo molecules. A total of 100 novel molecules were generated using a trained artificial intelligence model employing deep learning–based de novo approaches, derived from the seed molecule. Among the 100 generated compounds, the top two were selected on the basis of their high synthetic accessibility (SA) score of 66.5578 and 99.00, respectively, along with optimum ADMET properties, lower docking scores, stronger binding interaction, and minimum binding free energy. These compounds exhibited relative activity similar to the seed molecule and are expected to serve as potential DHFR inhibitors, paving the way for further advanced studies.
Shareef et al. (Thu,) studied this question.