AI驱动发现IRP2-EGFR双靶向策略揭示在结直肠癌中的协同抗肿瘤疗效

Background: IRP2 is a key regulator of intracellular iron homeostasis and is known to be overexpressed in colorectal cancer (CRC). As a relatively novel therapeutic target, IRP2 has been implicated in tumor progression, and our previous study demonstrated the antitumor effects of IRP2 inhibitors. To broaden the therapeutic potential of IRP2 inhibition and establish effective combination strategies, we applied a multifaceted AI-based predictive approach to identify compounds capable of synergizing with the IRP2 inhibitor. Methods: We predicted compounds capable of producing a synergistic effect with IRP2 inhibition using an AI-based model on the AIGEN InSight platform. Cell viability was assessed using CCK-8, and synergy scores were calculated using the SynergyFinder 3.0. To investigate the mechanism underlying the combination effect at the RNA level, we performed RNA sequencing (RNA-seq) and identified differentially expressed genes (DEGs) using the DESeq2 R package. Pathway enrichment analysis was conducted, and key findings were validated in CRC cell lines using qRT-PCR and western blotting. Results: Based on the AI-driven prediction model, Osimertinib was identified as a top-ranked candidate, achieving an AI-score of 0.73, which places it within the top 0.03% of all evaluated compounds. IRP2 inhibitor was subsequently optimized to generate KS-20260 with improved potency. The combination of KS-20260 and Osimertinib exhibited differential synergy scores across seven colorectal cancer cell lines. Notably, LoVo and DLD-1 cells showed the highest synergy scores of 10.845 and 7.843, respectively, and the combination treatment markedly reduced the protein levels of IRP2 and phosphorylated EGFR compared with single-agent treatments. Using RNA-seq data from LoVo cells treated with KS-20260, Osimertinib, or their combination, we analyzed pathways that were more affected by the combination treatment than the additive effect, defined as the average response of individual treatments. The Gene Set Enrichment Analysis (GSEA) revealed a significant downregulation of cell-cycle-related gene sets (e.g., CELL_CYCLE and CELL_CYCLE_PHASE_TRANSITION). Supporting these transcriptomic findings, the mRNA expression of key cell-cycle regulators, including CDK1, AURKB, CENPF, and E2F8, was decreased in the combination-treated cells. Furthermore, in the preliminary in vivo experiment, the combination-treated group (TGI 57.01%) exhibited a greater reduction in both tumor volume and tumor weight compared with the single-agent treatment groups (KS-20260, 38.37%; Osimertinib, 20.57%). Conclusion: This study demonstrates that KS-20260 and Osimertinib combination therapy disrupts cell-cycle progression and enhances antitumor efficacy in colorectal cancer, supporting dual IRP2-EGFR targeting as a promising therapeutic strategy.