From Structure to Impact: QSAR Assessment of Flame RetardantsInduced Lung Toxicity

Organophosphate flame retardants (OPFRs) are widely used in industries to reduce flammability and improve fire safety in plastic products without compromising performance. These applications are common in consumer goods and industrial products, such as electronics, building materials, furnishings, and textiles. Still, the extensive use of OPFRs in plastics has raised concerns about their potential environmental and health impacts [1] as their exposure has been linked to diseases such as asthma, pulmonary fibrosis, or acute respiratory distress syndrome. The study seeks to correlate the molecular structures of a set of OPFRs, characterized by their physicochemical molecular descriptors, with their biological effects on lung cells. This research aims to assess the potential implications of human exposure to OPFRs, enhancing our understanding of how structural variations in these compounds may affect their toxicity and impact on lung health. First, molecular descriptors were generated from 9X OPFRs compound names by converting them into SMILES representations. The, the RDKit package in Python compute 210 molecular descriptors from these SMILES structures, enabling a detailed analysis of the chemical and structural properties of the OPFRs under study. Next, the study experimentally investigated seven OPFRs¿tris(2-butoxyethyl) phosphate (TBOEP), triphenyl phosphate (TPhP), 2-ethylhexyl diphenyl phosphate (EHDPhP), tris(1,3-dichloro-2-propyl) phosphate (TDCPP), tris(2-ethylhexyl) phosphate (TEHP), tricresyl phosphate (TCP), and tris(2-chloroethyl) phosphate (TCEP)¿using two and three-dimensional (3D) cell cultures of the A549 lung cancer cell line. The research focused on evaluating the phenotypic effects and lipidomic alterations induced by these compounds. After exposing the cells to the OPFRs for 72 hours, the study evaluated cytotoxicity, reactive oxygen species (ROS) generation, and interleukin-8 (IL-8) secretion. Moreover, lipid extracts from the cells were analyzed using UHPLC/Q-TOF-MS in positive ionization mode to identify lipids most affected by exposure to OPFRs. Finally, the PCA exploration allowed the classification of OPFRs into different families based on their physicochemical properties, dividing them into various main groups. The seven compounds studied can be considered as models for these major groups related to aromatic rings, bromine, and alkyl constituents. Subsequently, various multivariate calibration methods (e.g., PLS-DA, ANNs, ...) were evaluated to correlate these molecular descriptors with experimentally measured phenotypic and lipidomic properties. The resulting models facilitated the prediction of the biological and lipidomic behavior of different OPFR families. These results contribute valuable insights into the safety assessment of OPFRs, highlighting the need for careful evaluation of their impact on human health.