Critical radiosensitizing parameters of dual-modality gold nanodandelion against cancer radioresistance in MV X-ray and proton therapies

Gold-based nanomaterials have emerged as promising nanoradioenhancers (NREs), with growing experimental and clinical evidence supporting their efficacy in both photon and proton therapies. However, the underlying mechanisms of nanoparticle-induced radioenhancement in clinical radiotherapy remain unclear. Here, we investigate the NRE mechanisms of multibranched gold nanodandelions (GNDs) under 10 MV photon and 160 MeV proton irradiation. Specifically, we evaluate irradiation-induced reactive oxygen species (ROS) generation under clinical photon and proton settings and demonstrate that nanoparticle composition and ambient oxygen concentration markedly influence ROS production. We further observe that aggregated-form GNDs (aGNDs) exhibit significantly enhanced ROS production, particularly singlet oxygen (¹O₂) and superoxide anion (O₂^•–), an effect that can be attributed to the increased probability of interaction between X-rays and aGNDs. In contrast, other parameters, including nanoparticle shape, surface functionalization, and accessible surface area, contributed minimally to ROS enhancement. In addition, combining matrix metalloproteinase (MMP) targeting capability with peroxidase-mimicking activity, we observe excessive ROS production immediately after irradiation, accompanied by more severe and sustained DNA damage and mitochondrial dysfunction. Furthermore, the strong therapeutic effect of GND-enhanced proton therapy was confirmed in vitro using radioresistant U87-MG RR and CT-2A RR cell lines, which showed significantly reduced colony formation compared to cells treated with proton therapy based on conventional spherical gold nanoparticles (AuNPs). Collectively, our physical and biological evidence suggest that aggregation of GNDs enhances ROS production following proton and MV photon irradiation, with the potential to improve the therapeutic efficacy of clinical radiotherapy.