In silico QSAR-guided design of Dammarane-type triterpenoids as potential PI3Kα-targeted anticancer agents

Cancer remains a major global health challenge, driven in part by dysregulation of key oncogenic signaling pathways such as phosphatidylinositol 3-kinase alpha (PI3Kα), which plays a central role in tumor growth and therapeutic resistance. In this study, an integrative computational strategy combining quantitative structure-activity relationship (QSAR) modeling, drug-likeness and ADMET prediction, and molecular docking was applied to investigate dammarane-type triterpenoid derivatives as potential PI3Kα inhibitors. A dataset of 22 reported compounds was analyzed using multiple linear regression (MLR), partial least squares (PLS), and principal component regression (PCR) models, all of which were rigorously validated by external test sets, Y-randomization, and applicability domain analysis, with the PCR model showing the highest predictive performance (R² = 0.833; R²_test= 0.79). Descriptor analysis identified lipophilicity, electronic distribution, and polar surface properties as key determinants of anticancer activity, while excessive molecular size negatively influenced potency. Guided by these insights, four new derivatives (D1-D4) were rationally designed and evaluated in silico, exhibiting favorable drug-likeness, high predicted oral absorption (89-100%), absence of AMES toxicity, and moderate synthetic accessibility. Molecular docking against PI3Kα (PDB ID: 8TSB) revealed stable binding for all designed compounds, with D1 emerging as the most promising lead, combining strong binding affinity (-5.70 kcal/mol) and favorable interaction patterns within the active site. Overall, this work demonstrates the potential of dammarane-type triterpenoids as PI3Kα-targeted anticancer agents and highlights the value of an integrated, cost-effective computational framework for rational lead identification and optimization, supporting future experimental development in line with global health priorities.