Critical evaluation of ToxCast-Reactome predicted toxicity pathway correspondence of the human liver HepG2 activity profile with observed PFOA and PFOS hazards

Abstract

Per- and polyfluoroalkyl substances (PFAS) are ubiquitous in the environment due to their persistence and widespread use in consumer products. The environmental fate and human health effects of two commercially successful PFAS, perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS), have been extensively studied. There is strong evidence that hepatotoxicity, immunotoxicity, and developmental toxicity observed in PFOA/PFOS-exposed rodents involve the activation of peroxisome proliferator-activated receptor-α (PPAR α); however, studies in PPAR α -null mice suggest that PPAR α -independent mechanisms also contribute to hepatotoxicity and immunotoxicity. During the prior decade, manufacturing uses of PFOA and PFOS were phased out as part of global stewardship programs; however, thousands of newer, less characterized PFAS are currently used as replacements. Consequently, there is a need for methods to efficiently identify the potential hazards of these alternative PFAS. This study critically evaluated an in silico hazard screening framework for PFAS focusing on PFOA and PFOS, whose toxicity has been well characterized. Data for human liver HepG2 cells were obtained from ToxCast to identify genes activated by PFOA or PFOS representing a primary target organ, and biological pathways were inferred using the Reactome database. Several biological pathways identified through the ToxCast-Reactome framework corresponded to potential human or animal hazards suggested in epidemiology or laboratory studies, increasing confidence in the proposed screening methodology. Thus, this in silico framework should be useful for assessing the hazard of the newer, less characterized PFAS, and may be adapted to other classes of toxicants as well.