Machine Learning-Aided\n3D Dynamic SERS Strategy for\nPhysiological Mapping: Biotoxicity of Environmentally Dimensional\nAged Nanoplastics and Corresponding Protein Corona Complexes

Nanoplastics (NPs) are emerging pollutants that undergo\ninevitable\naging in the environment, raising concerns about human exposure and\nhealth hazards. Research on the cytotoxicity of various polymer types\nof NPs, aged nanoplastics (aNPs), and their interactions with proteins\n(aNPs-protein corona) is still nascent. Traditional cytotoxicity detection\nmethods often rely on end point assays with restricted temporal resolution\nand analysis of single or multiple biomarkers. Here, we propose a\nnovel approach integrating the 3D dynamic SERS strategy (DSS) with\nmachine learning to rapidly analyze the cell fate and death modes\ninduced by NPs, aNPs, and aNPs-protein corona complexes at the molecular\nlevel. PS, PVC, PMMA, and PC products from the water environment were\nused to prepare the corresponding NPs, and the impact of UV irradiation\non their physicochemical properties was examined. DSS systematically\nmaps the molecular changes in the cellular secretome caused by these\nNPs. Machine learning effectively extracts information from complex\nspectra, differentiating between biological samples. Results show\nprolonged UV exposure increases cell sensitivity to ferroptosis and\ncytotoxicity in various aNPs, while the protein corona on aNPs significantly\nmitigates toxicity associated with surface oxygen-containing functional\ngroups, resulting in a reduced similarity to ferroptosis signatures.\n3D DSS with machine learning technique analyzes the overall metabolite\nprofile at the molecular level rather than individual biomarkers.\nThis study is the first attempt to compare the biotoxicity of diverse\npolymer NPs, aNPs, and aNPs-protein coronas at cellular and molecular\nlevels in human hepatocytes, enhancing our understanding of the complex\nbiological impacts of NPs.