Tale of Three <i>N</i>-Nitrosamines and the Variables Needed to Assess Their Carcinogenicity In Silico Incorporated into a Single Workflow

<i>N</i>-Nitrosamine impurities in pharmaceuticals present a considerable challenge for regulators and industry alike, where the absence of carcinogenic-potency studies has left a gap that must be adequately filled to protect public health. In the interim, this means balancing risk assessment with the necessity to continue research, development, and supply of pharmaceuticals. In the long term, we need a cost-effective solution that optimizes both. As if beholden to Newton's Third Law, every crisis breeds an opportunity of equal magnitude. Consequently, cross-industry consortia have been racing to find a solution by advancing our current science. Recent spotlight has been on in silico tools, as a fast and increasingly reliable alternative to in vivo and in vitro testing. Because <i>N</i>-nitrosamine bioactivation lends itself uniquely to quantum mechanics (QM) approaches, the integration of electronic-structure considerations has emerged as the dominant in silico approach. This signifies a considerable leap in predictive toxicology, which has, for much of its existence, relied on atomistic (quantitative) structure-activity relationships, i.e., (Q)SARs. Here we present a validation of an integrated docking-QM approach within the CADRE program and demonstrate its utility on three different impurities, <i>N</i>-nitroso-7-monomethylamino-6-deoxytetracycline, <i>N</i>-nitroso-dabigatran etexilate, and 1-methyl-4-nitrosopiperazine. We show that a combined in silico strategy, which considers bioavailability, transport, cytochrome P450 binding, and reactivity, can be leveraged to supplement the overly conservative Carcinogenic Potency Categorization Approach (CPCA) in setting the daily acceptable intake (AI) using defensible, highly mechanistic, and quantitative drivers of <i>N</i>-nitrosamine metabolism. To that end, we argue that while <i>N</i>-nitroso-7-monomethylamino-6-deoxytetracycline and 1-methyl-4-nitrosopiperazine are cohort-of-concern impurities, <i>N</i>-nitroso-dabigatran etexilate is not a potent carcinogen (TD₅₀ > 1.5 mg/kg/day), contrasting the CPCA-derived AI. Lastly, we discuss how the CADRE tool can be integrated with the broader landscape of QM methods and the CPCA into a single harmonized in silico strategy for carcinogenicity assessment of <i>N</i>-nitrosamine impurities.