Zebrafish behavioral assessment as a first-tier whole-organism NAM for developmental neurotoxicity: A multi-laboratory evaluation.

The development of New Approach Methodologies (NAMs) for developmental neurotoxicity (DNT) assessment requires demonstrating reproducibility across testing facilities and clearly defining their role in tiered testing strategies. This multi-laboratory study evaluated the zebrafish Light-Dark Transition Test (LDTT) across four independent laboratories testing 33 compounds (28 suspected DNT toxicants and 5 putative DNT-negatives), including organophosphate and pyrethroid insecticides (deltamethrin, parathion), flame retardants (TDCIPP, TMPP, BPDP), and neonicotinoids (nicotine, acetamiprid) using agreed-upon protocol elements. A novel three-tiered quality control framework removed 10.9% of plates passing basic viability criteria, providing a systematic approach applicable to multi-well plate-based behavioral assays. Developmental toxicity (DT) endpoints showed high inter-laboratory concordance (84-100%), while DNT activity calls were more variable (68-84%); however, potency estimates for active compounds showed similar reproducibility (mean 3.2-fold for DNT, 3.5-fold for DT). Zebrafish LDTT detected 93% of DNT-positive compounds and 100% of DNT-negative compounds as a single whole-organism assay, demonstrating complementary performance to the 13-assay DNT in vitro battery (DNT-IVB; 86% sensitivity, 60% specificity). Integration with DNT IATA case studies demonstrated that multi-laboratory zebrafish data strengthen weight-of-evidence evaluations, with independent validation establishing high confidence in the flame retardants TMPP and BPDP. Together, these findings support positioning zebrafish larval behavioral assays as an efficient first-tier whole-organism NAM that bridges mechanistic in vitro screening and mammalian guideline studies, providing organism-level points of departure and prioritizing compounds for further DNT-IVB and in vivo testing within tiered regulatory DNT assessment frameworks.