Abstract
BACKGROUND: The aim of this study was to investigate the mechanism of the Poly ADP-ribose polymerase (PARP) inhibitor pamiparib (PAM) and the tyrosine kinase inhibitor anlotinib (ANL), utilizing a Biological Factor Regulatory Neural Network (BFReg-NN) and metabolomics approach. METHODS: Potential drug combinations were identified by integrating bioinformatics and machine learning algorithmic strategies. In vitro, their effects on ovarian cancer cells were detected by MTT assay, clone formation and annexin V/PI double staining, scratch assay and Transwell assay. The effect of PAM in combination with ANL was investigated in a nude mouse ovarian cancer model. The mechanism of action was investigated using an untargeted metabolomics approach. The inhibitory effect of the combination of the two drugs on stem cell activity was detected using the tumorsphere assay, limiting dilution assay and RT-qPCR, and the changes in signaling pathway protein expression after treatment with the two drugs were detected using Western blotting. RESULTS: Predictive results confirmed the synergistic effect of the potential drug combinations, revealing that the potential mechanism of PAM combined with ANL in ovarian cancer is related to tumor stem cells. Overexpression of the PI3K/Akt signaling pathway is commonly associated with cancer recurrence and drug resistance. In vitro, the combination of PAM and ANL inhibited clone formation, proliferation, migration, and stemness of A2780 ovarian cells through the PI3K/Akt signaling pathway. In vivo, significant downregulation of p-PI3K, p-Akt, Bcl-2, and HIF1-α, and upregulation of BAX protein expression confirmed that the mechanism of action of combination therapy is related to PI3K/Akt pathway. CONCLUSION: The combination of PAM and ANL was more effective than monotherapy for treating ovarian cancer and holds potential to become a new therapeutic approach for ovarian cancer.