Verification of In Vivo Estrogenic Activity for Four Per- and Polyfluoroalkyl Substances (PFAS) Identified as Estrogen Receptor Agonists via New Approach Methodologies

Abstract

Given concerns about potential toxicological hazards of the thousands of data-poor per- and polyfluorinated alkyl substances (PFAS) currently in commerce and detected in the environment, tiered testing strategies that employ high-throughput in vitro screening as an initial testing tier have been implemented. The present study evaluated the effectiveness of previous in vitro screening for identifying PFAS capable, or incapable, of inducing estrogenic responses in fish exposed in vivo. Fathead minnows (Pimephales promelas) were exposed for 96 h to five PFAS (perfluorooctanoic acid [PFOA]; 1H,1H,8H,8H-perfluorooctane-1,8-diol [FC8-diol]; 1H,1H,10H,10H-perfluorodecane-1,10-diol [FC10-diol]; 1H,1H,8H,8H-perfluoro-3,6-dioxaoctane-1,8-diol [FC8-DOD]; and perfluoro-2-methyl-3-oxahexanoic acid [HFPO-DA]) that showed varying levels of in vitro estrogenic potency. In agreement with in vitro screening results, exposure to FC8-diol, FC10-diol, and FC8-DOD caused concentration-dependent increases in the expression of transcript coding for vitellogenin and estrogen receptor alpha and reduced expression of insulin-like growth factor and apolipoprotein eb. Once differences in bioconcentration were accounted for, the rank order of potency in vivo matched that determined in vitro. These results provide a screening level benchmark for worst-case estimates of potential estrogenic hazards of PFAS and a basis for identifying structurally similar PFAS to scrutinize for putative estrogenic activity.

Keywords: PFAS; adverse outcome pathway; ecotoxicology; endocrine disruption; new approach methodologies.

Figure 1.

Relative abundance of vitellogenin (vtg) and estrogen receptor alpha (esr1) mRNA transcripts in the liver tissue of male fathead minnows exposed to FC8-diol (A, B), FC8-DOD (C, D), or FC10-diol (E, F) for 96 h. See Supplementary Table S.1 for full chemical names. Only results from experiment 6 are shown here for FC10-diol. Error bars indicate standard error. Common letters indicate treatments that were not significantly different from one another based on a post-hoc test. Concentrations are mean measured concentration from each study. n=11–12. For vtg, one extreme outlier was removed from the control group from the FC8-diol exposure.

Figure 2.

Relative abundance of insulin-like growth factor (igf1) and apolipoprotein eb (apoeb) mRNA transcripts in the liver tissue of male fathead minnows exposed to FC8-diol (A, B), FC8-DOD (C, D), or FC10-diol (E, F) for 96 h. See Supplementary Table S.1 for full chemical names. Only results from experiment 6 are shown here for FC10-diol. Error bars indicate standard error. Common letters indicate treatments that were not significantly different from one another based on a post-hoc test. Concentrations are mean measured concentration from each study. n=11–12.

Figure 3.

Comparison of estimated points of departure for FC8-diol and FC10-diol. In vitro AC50 refers to the in vitro 50% activity concentration based on the average of the two Attagene assays. Unadjusted in vivo benchmark concentration (BMC) represents the mean (across four gene expression endpoints) based on measured water concentrations. Bioconcentration (BC) adjusted BMCs = (in vivo unadjusted BMC * mean BC [Supplementary Table S.3]). BC and relative potency (REP) adjusted BMCs = (in vivo unadjusted BMC * mean BC * in vitro REP [Supplementary Table S.2]).