Cluster-based read-across of ToxCast data identifies buprofezin as a partial PPARγ agonist with obesogenic potential

The identification of environmental obesogens has become increasingly urgent amid rising rates of metabolic disorders linked to chemical exposures. Here, we employed a cluster-based read-across framework that integrates structural descriptors with high-throughput screening (HTS) data from the ToxCast database to systematically identify potential obesogens. A total of 8971 chemicals were represented in a 2,217-dimensional structure-activity matrix, combining 1905 chemical fingerprints and 312 bioactivity endpoints, which were reduced and clustered into 135 distinct chemical groups. Among these, Cluster 14 was notably enriched with compounds affecting lipid metabolism pathways and included buprofezin, a thiadiazine insecticide not previously linked to obesogenic activity. Functional assays in human bone marrow-derived mesenchymal stem cells (hBM-MSCs) demonstrated that buprofezin promotes adipogenesis. Despite its classification as inactive in PPARγ-specific ToxCast assays, mechanistic studies revealed that buprofezin acts as a partial agonist of PPARγ (K_i = 14.8 µM), enhancing SRC1 coactivator recruitment. These findings underscore the limitations of current HTS data in capturing partial agonist activities and demonstrate the value of integrated read-across approaches in uncovering hidden chemical hazards. Given buprofezin's environmental persistence and widespread agricultural use, its potential metabolic health risks require further investigation.