Nucleic acid drugs may cause toxicity through several mechanisms, including on-target effects, narrow off-target effects, and broad off-target effects. On-target toxicity refers to exaggerated pharmacological activity at the intended target, while broad off-target toxicity is mainly related to the molecule's structure, physicochemical properties, immune interactions, plasma protein binding, distribution, and metabolism. Because narrow off-target effects are often species-specific and difficult to predict reliably in animal models, toxicology study design usually focuses more on evaluating on-target and non-specific broad off-target risks through scientifically selected animal species.
Similar to small-molecule drugs, nucleic acid therapies generally require nonclinical toxicity studies in two animal species: one rodent species, such as mouse or rat, and one non-rodent species, such as dog or monkey. This two-species approach is important because nucleic acid drugs may show different toxicity sensitivity across species and may also have species-dependent pharmacokinetic profiles due to artificial nucleotide modifications. Even when short-term toxicity appears similar between species, long-term toxicity studies in two species can improve the ability to detect potential adverse effects. This approach is more consistent with the ICH M3(R2) framework for small molecules than with the simplified single-species approach sometimes allowed for biologics under ICH S6(R1).
For on-target toxicity assessment, animal species should be selected based on whether they can demonstrate the intended pharmacological activity of the nucleic acid drug. If multiple species show pharmacological effects, both one rodent and one non-rodent species should be used. If only one species is pharmacologically relevant, that species may be selected for on-target toxicity evaluation. If no conventional animal species can adequately model the pharmacological effect, a surrogate molecule may be considered, although surrogate-based studies can provide limited information because pharmacodynamic and toxicological responses may differ from those of the clinical candidate. When expected on-target effects are minimal, developers may instead rely on careful target biology assessment and stepwise clinical progression to manage remaining uncertainties.
For broad off-target toxicity assessment, the clinical candidate itself should be tested, and two animal species are generally recommended to maximize the detection of potential adverse reactions. If both selected species demonstrate pharmacological activity, on-target and off-target toxicities can be assessed together. If only one species is pharmacologically relevant, an additional non-relevant species may still be needed to evaluate non-specific off-target effects. Species selection should also consider known class effects, structural similarities to previously studied nucleic acid drugs, and interspecies differences in metabolism and PK. When a candidate's metabolic or pharmacokinetic profile differs substantially across species, models that more closely resemble humans should be prioritized; for example, non-human primates may better reflect human PK for some phosphorothioate-modified ASOs than mice.
Overall, animal species selection for nucleic acid drug toxicology studies should address two major dimensions: pharmacologically mediated on-target toxicity and structure-related broad off-target toxicity. For drugs targeting human-specific sequences, developers may design sequences shared by humans and animals to enable pharmacological assessment in conventional models. As the field advances, species selection strategies are expected to become more precise through the use of toxicology databases, structural analogy analysis, and emerging humanized models. A scientifically justified and translational animal model strategy is therefore essential for generating reliable nonclinical safety data and supporting the development of safe nucleic acid therapeutics.
