SAT-240 In Silico To In Vivo Analyses of Tert-Butylphenols Used in Food Contact Materials Identify Retinoid-X Receptor Activators

Abstract Disclosure: B.J. Mengeling: None. A. Ettayapuram Ramaprasad: None. M. Smith: None. K.A. Durkin: None. M. La Merrill: None. J.D. Furlow: None. The potential for food contact chemicals to disrupt genetic programs in development and metabolism raises concerns. The nuclear receptor (NR) superfamily of transcription factors controls many of these genetic programs. Within the NR family, the retinoid-X receptor (RXR) is a DNA-binding partner for one-third of the other NRs. Disruption of its activity, therefore, has the potential to generate adverse outcomes in several NR pathways. We used machine learning and other in silico methods to identify RXR-interacting candidates from a list of over 57,000 chemicals. Butylphenols comprised the largest, high probability, structural group with 58 compounds; several are food contact chemicals with widespread commercial use as antioxidants. Prior ToxCast data from in vitro reporter assays suggested that bulky, aliphatic substitution at the C4 position of 2,6-di-tert-butylphenol facilitated RXR activation. We tested this hypothesis using six butylphenols with the following C4 substitutions: H, methyl, ethyl, sec-butyl, tert-butyl, and methyl(dimethylamino). We used a Xenopus laevis precocious metamorphosis assay that analyzes disruption of thyroid hormone receptor (TR) signaling in vivo, where TR-RXR heterodimers are the predominant DNA-binding form. Using an integrated luciferase reporter driven by TR-RXR binding, we assayed TR action and quantified morphological changes. Thyroid hormone (TH) was required for activation, but half of the tested set of tert-butylphenols potentiated TH action at nanomolar concentrations. Molecular modeling studies showed that the three positive butyl-phenols formed more frequent, stable interactions with RXRα, and the bulkiness of their moiety at the C4 position increased steric complementarity with the RXR ligand binding pocket. Our findings significantly add to rising concerns about the safety of these well-known antioxidants in food packaging and establishes a paradigm for machine learning coupled with a convenient, small animal in vivo validation approach to identify new chemicals of concern that can interact with RXR-NR controlled genetic pathways. Presentation: Saturday, July 12, 2025