Elucidating Key Characteristics of PFAS Binding to Human Peroxisome Proliferator-Activated Receptor Alpha: An Explainable Machine Learning Approach.

Per- and polyfluoroalkyl substances (PFAS) are widely employed anthropogenic fluorinated chemicals known to disrupt hepatic lipid metabolism by binding to human peroxisome proliferator-activated receptor alpha (PPARα). Therefore, screening for PFAS that bind to PPARα is of critical importance. Machine learning approaches are promising techniques for rapid screening of PFAS. However, traditional machine learning approaches lack interpretability, posing challenges in investigating the relationship between molecular descriptors and PPARα binding. In this study, we aimed to develop a novel, explainable machine learning approach to rapidly screen for PFAS that bind to PPARα. We calculated the PPARα-PFAS binding score and 206 molecular descriptors for PFAS. Through systematic and objective selection of important molecular descriptors, we developed a machine learning model with good predictive performance using only three descriptors. The molecular size (b_single) and electrostatic properties (BCUT_PEOE_3 and PEOE_VSA_PPOS) are important for PPARα-PFAS binding. Alternative PFAS are considered safer than their legacy predecessors. However, we found that alternative PFAS with many carbon atoms and ether groups exhibited a higher affinity for PPARα. Therefore, confirming the toxicity of these alternative PFAS compounds with such characteristics through biological experiments is important.

In vivo and in silico analyses of estrogenic potential of equine estrogens in medaka (Oryzias latipes).

Equine estrogens (EEs) are widely used in hormone replacement therapy pharmaceuticals for postmenopausal women. Previous studies have shown that EEs occur in the aquatic environment; however, the potential estrogenicity and risk of EEs in aquatic organisms, including fish, have yet to be studied in detail. Therefore, we evaluated the estrogenic potential of major EEs, namely equilin (Eq), 17α-dihydroequilin (17α-Eq), 17ß-dihydroequilin (17ß-Eq), equilenin (Eqn), 17α-dihydroequilenin (17α-Eqn), and 17ß-dihydroequilenin (17ß-Eqn), on medaka (Oryzias latipes) using in vivo and in silico assays. Quantitative real-time RT-PCR analyses revealed that expression levels of choriogenin L (ChgL) and choriogenin H (ChgH) in medaka embryos responded to various types and concentrations of EEs in a concentration-dependent manner, whereas transcription levels of vitellogenin 1 were not significantly affected by any of the EEs in the concentration range tested. The order of the in vivo estrogenic potencies of EEs was as follows: 17ß-Eq > Eq > 17ß-Eqn > Eqn > 17α-Eqn > 17α-Eq. Additionally, the 50% effective concentrations (EC50) of 17ß-Eq was lower than that of 17ß-estradiol. We also investigated the interaction potential of EEs with medaka estrogen receptor (ER) subtypes in silico using a three-dimensional model of the ligand-binding domain (LBD) for each ER and docking simulations. All six EEs were found to interact with the LBDs of ERα, ERß1, and ERß2. The order of the in silico interaction potentials of EEs with each ER LBD was as follows: 17ß-Eq > 17α-Eq > Eq > 17ß-Eqn > 17α-Eqn > Eqn. Furthermore, we identified the key amino acids that interact with EEs in each ER LBD; our findings suggest that amino acids and/or their hydrogen bonding may be responsible for the ligand-specific interactions with each ER. This study is the first to comprehensively analyze the estrogenic potential of EEs in medaka both in vivo and in silico.