<i>N</i>-nitrosamines: in silico modelling of DNA reactivity and identification of metabolic precursors

The discovery of N-nitrosamines (NNAs) as impurities in several pharmaceuticals has renewed activities in assessing their mutagenic and carcinogenic potential. In the current investigation, the binary mutagenic potential of NNAs is re-investigated using the mechanism-based structure-activity approach of the TIMES models. Emphasis is placed on meeting the OECD (Q)SAR principles for model validation and the organization's (Q)SAR prediction principles. A curated data set of 41 small and complex NNA-containing substances tested in a standard battery of Salmonella typhimurium strains with and without rat microsomal activation was assessed for these tasks. Structural boundaries are initially derived from activating mechanisms for interactions of parent NNAs with DNA described in the literature. These activating mechanisms include direct-acting mutagenicity (denitrosation of parent molecules) or DNA interactions after S9 metabolic activation (alpha-hydroxylation). After analysis of the 41 NNAs, structural features that mitigate or 'mask' the covalent binding of NNAs to DNA expanded the original alert definition. The structural fragments' predictive capabilities (performance) for the activating and negating mechanisms of these 41 chemicals are excellent. Three false positives and no false negatives are reported. Moreover, the role of metabolism in the N-nitrosation of secondary amines and tertiary amines after conversion to secondary amines under in vivo conditions is explained with descriptions of new metabolic transformations. These transformation boundaries are applied to different inventories to search for parent structures that are potential in vivo metabolic precursors of NNAs.