A tumor-agnostic, topology-informed scoring framework for drug repurposing: application to CDK4/6 inhibitor resistance in HR+ breast cancer

Resistance to cyclin-dependent kinase 4/6 (CDK4/6) inhibitors poses a critical obstacle in managing hormone receptor-positive (HR⁺) breast cancer, underscoring an urgent need for new therapeutic strategies. Here, we developed a novel, tumor-agnostic, network-based scoring framework integrating transcriptomics, protein–protein interaction (PPI) networks, and drug-target affinity data to systematically prioritize repurposable therapeutic agents. Our dual-scoring system quantitatively evaluates each target’s network centrality and pharmacological relevance, demonstrating robust prediction of drug sensitivity across multiple validation cohorts (enrichment p = 0.00433). Application of this model specifically to CDK4/6-resistant breast cancer identified sorafenib as a highly-ranked therapeutic candidate, with FGFR3 emerging as a critical node within resistance networks. Subsequent molecular docking and dynamics simulations confirmed stable sorafenib binding to FGFR3. Biological validation further demonstrated that FGFR3 knockdown partially restored resistance to sorafenib in previously sensitized CDK4/6-resistant cells. Collectively, this scalable and biologically validated framework illustrates a powerful approach for discovering actionable resistance mechanisms and repurposing clinically available drugs, offering tangible pathways toward precision oncology.