An alternative to in-vivo/in-vitro drug toxicity testing: Environmentally relevant approaches for toxicity assessment of pharmaceuticals towards rats

Excessive and inappropriate usage of human and veterinary pharmaceuticals alongside their improper disposal give rise to the contaminants of emerging concern (CEC), potentially harming the non-target living organisms. To meet European Medicines Agency (EMA) guidelines, comprehensive toxicological information is required for each chemical before market release. In this regard, quantitative structure-toxicity relationships (QSTRs) offer effective alternatives. This study aims to develop QSTR models based on the oral median lethal dose (LD₅₀) values of 702 pharmaceuticals in rats, adhering to the OECD principles. Several globally accepted validation parameters were analyzed to establish the reliability, robustness, etc., of the developed QSTR models. Intelligent consensus prediction (ICP) was applied to the individual models to boost predictivity of the models. This study elucidated that electronegativity, lipophilicity, the presence of pyrroles, and the number of tertiary amine groups are responsible for the toxicity toward rats. Further, the DrugBank database (investigational and experimental groups) was screened by the optimal model and the results were validated using real-world toxicity data. The developed models will greatly reduce the cost, time, and animal experimentation in the context of pharmaceuticals-induced toxicity in rat models, enabling toxicity data gap bridging and the development of safe pharmaceuticals strictly supporting RRR (reduction, refinement, and replacement) guidelines.