Unveiling the TP5 dual binding mechanism and structure-based search for potential dual inhibitors of tubulin and...

The dual inhibition of microtubule dynamics and immune checkpoint signaling represents a promising strategy in cancer therapy by simultaneously disrupting cell proliferation and immune evasion. Both αβ-tubulin and dimeric PD-L1 (dPD-L1) are critical targets in oncology due to their central roles in tumor growth, metastasis, and immune suppression. In this study, we employed an integrative computational approach encompassing molecular docking, molecular dynamics (MD) simulations, MM-GBSA binding energy calculations, ADMET and chemical space analysis, to identify potential novel dual inhibitors of αβ-tubulin and dPD-L. A library of ∼14 000 compounds, derived from ChEMBL and generative neural networks, was screened using key structural fragments associated with both targets. Among the candidates, the compound TP5, previously reported as a dual inhibitor with in vitro and in vivo efficacy, was included as a reference compound, despite lacking structural information about its binding modes. Our simulations provide detailed structural insights into TP5's interactions with both tubulin and PD-L1, revealing its colchicine-site mimicry and moderate stabilization of PD-L1 dimers. Additionally, several newly identified compounds displayed more favorable binding energies than TP5, particularly frag2_571, frag2_3214, frag2_3522, frag2_8390, frag2_8592, and bf_1, which exhibited stable dual-target interactions and similar structural effects to TP5. ADMET predictions supported their pharmacological potential, highlighting them as promising scaffolds for further optimization and experimental evaluation. These findings support the utility of fragment-based and AI-guided design in the discovery of multitarget anticancer agents.