Development of a translational murine model for antibody-drug conjugate-induced interstitial lung disease

, is transforming HER2+ breast cancer treatment but is linked to severe off-target toxicity, notably interstitial lung disease (ILD), a serious and mechanistically unresolved adverse event. The preclinical prediction of ILD, however, remains poor and largely dependent on non-human primates (NHP). Here, we establish a robust murine model that recapitulates delayed-onset, ILD-like pathology following repeated ADC administration. Mice treated with trastuzumab deruxtecan (T-Dxd) developed progressive pulmonary injury manifested by alveolar septal thickening, inflammatory infiltrates, pneumocyte hyperplasia, and subsequent fibrotic remodeling. These key histopathological changes closely parallel clinical observations and findings reported in non-human primates. Notably, the delayed nature of injury, emerging after repeated dosing rather than acutely, aligns with clinical ILD timing and further underscores the need for extended observation windows in preclinical studies. Using this model, we further demonstrate its ability to differentiate ADCs with distinct pulmonary safety profiles. Under identical dosing paradigms, Trastuzumab and an SN-38-based Trastuzumab-ADC showed minimal lung pathology, supporting the model's ability to distinguish ADCs with divergent pulmonary safety profiles. The model's discriminative capacity was further explored by studying NJB-T-PL119, a novel Trastuzumab-based ADC incorporating an alternative linker-payload. This novel ADC retained antitumor efficacy comparable to T-Dxd yet produced only low-grade pulmonary changes. This platform provides a scalable, ethically feasible, and mechanistically relevant system for early de-risking of ADC-associated lung injury, enabling informed molecule design and prioritization before advancing to NHP toxicology or clinical evaluation.