Carcinogenic risks of aromatic amines and their radical oxidation products: Construction of adverse outcome pathways, development of priority control lists, and mechanistic analysis of carcinogenesis

Aromatic amines (AAs), widely distributed in the environment, have raised global concerns due to their carcinogenic risks. However, existing studies mainly focus on individual compounds or cancer types, with limited exploration of the carcinogenic potential of their oxidative products. To comprehensively evaluate human carcinogenic risks, this study selected 100 representative AAs and constructed adverse outcome pathways (AOPs) for bladder, colorectal, and pancreatic cancers. Additionally, 400 oxidative metabolites were generated through reactions with reactive oxygen species (ROO·, HO·, O₂·⁻, and O₃). A multi-indicator evaluation framework was established by integrating molecular docking, ADMET prediction, entropy weighting, coefficient of variation, CRITIC, and subjective weighting methods to quantify the overall carcinogenic risk. Target-specific analyses revealed that cytochrome P450 (CYP450) plays a critical role across AOPs, while O₃-induced metabolites show higher carcinogenic potential. A prioritization list was developed using the TOPSIS method combined with standard deviation classification, identifying 5 AAs of “special concern,” 21 of “high concern,” 51 of “moderate concern,” and 23 of “low concern.” Structural feature and residue-level interaction analyses indicated that AAs with complex architectures, multiple aromatic rings, bulky substituents, or the ability to form electrophilic metabolites exhibited higher risks. Electrostatic interactions were identified as the primary driving force for AA–CYP450 binding, and a greater number of amino acid binding sites enhanced binding affinity and carcinogenic risk. Overall, this study proposes a systematic risk grading framework, establishes a prioritization strategy for AAs control, and elucidates structural and mechanistic determinants contributing to carcinogenic risk heterogeneity.