Pharmacogenomic and in silico identification of isoform-selective AKT inhibitors from Pithecellobium dulce for precision cancer therapy

AKT1 and AKT2 are central but functionally distinct kinases within the PI3K-AKT-mTOR pathway, and isoform-specific genomic alterations in these proteins have important implications for cancer prognosis and therapeutic responsiveness. This study aimed to integrate cancer pharmacogenomics with structure-based modeling to identify natural compounds capable of selectively targeting AKT1 or AKT2.

Public cancer genomics datasets from TCGA and the Kaplan-Meier Plotter were analyzed to characterize mutation patterns, copy number alterations, and survival associations of AKT1 and AKT2 across malignancies. Based on isoform-specific differences, twenty phytochemicals from Pithecellobium dulce were docked against the allosteric binding sites of AKT1 (PDB: 3QKL) and AKT2 (PDB: 2JDO). Lead compounds were evaluated using ADME prediction and density functional theory to assess pharmacokinetic suitability and electronic stability. The dynamic behavior of ligand-protein complexes was examined through 200-ns molecular dynamics simulations using the Desmond-Schrödinger platform, and binding free energies were estimated via MM-GBSA analysis. Regulatory interactions involving AKT-associated non-coding RNAs were also examined to support pharmacogenomic relevance.