Profiling Acute Oral and Inhalation Toxicity Data Using a Computational Workflow to Screen for Facile Chemical Reactivity

Abstract

Abstract

We are developing computational approaches for mechanisms of acute toxicity. Facile chemical reactivity involves formation of covalent adducts between electrophiles and nucleophiles and can be molecular initiating events for adverse outcome pathways. Using an open-source Konstanz Information Miner workflow, we developed a profiler to identify reactive moieties that covers five chemically facile-reactive classes: (1) Michael acceptors, (2) Schiff base formation, (3) acylation, (4) nucleophilic aromatic substitution, and (5) bimolecular nucleophilic substitution. We hypothesized that nonspecific chemical reactivity may be a more prominent molecular initiating event for high acute toxicity for inhalation compared to oral exposure routes because of ensuing respiratory irritation with pulmonary edema resulting in lethality. We tested this hypothesis by stratifying rodent acute oral and 4-hour inhalation toxicity databases (DBs) according to the globally harmonized system (GHS) classification and our profiler to screen the DBs for the incidence of reactive alerts in each class. Before screening for chemical reactivity, the distribution among GHS classes 1–5 was similar for compounds in the inhalation DB and highly weighted toward less toxic GHS classes 4 and 5 for the oral DB, likely due to oral first-pass metabolism. However, after screening, there was no trend in facile reactivity relative to GHS distribution for the oral DB, but enrichment of more severe GHS classes 1–3 for the inhalation DB, indicating a greater contribution of chemical reactivity to acute toxicity after inhalation exposure. These results demonstrate implementation of a reactivity profiler for the endpoint of acute toxicity and highlight its potential future utility for other toxicity endpoints.