Occurrence and In Vitro Bioactivity of Estrogen, Androgen, and Glucocorticoid Compounds in a Nationwide Screen of United States Stream Waters

Abstract

In vitro bioassays are sensitive, effect-based tools used to quantitatively screen for chemicals with nuclear receptor activity in environmental samples. We measured in vitro estrogen (ER), androgen (AR), and glucocorticoid receptor (GR) activity, along with a broad suite of chemical analytes, in streamwater from 35 well-characterized sites (3 reference and 32 impacted) across 24 states and Puerto Rico. ER agonism was the most frequently detected with nearly all sites (34/35) displaying activity (range, 0.054-116 ng E2Eq L^-1). There was a strong linear relationship (r² = 0.917) between in vitro ER activity and concentrations of steroidal estrogens after correcting for the in vitro potency of each compound. AR agonism was detected in 5/35 samples (range, 1.6-4.8 ng DHTEq L^-1) but concentrations of androgenic compounds were largely unable to account for the in vitro activity. Similarly, GR agonism was detected in 9/35 samples (range, 6.0-43 ng DexEq L^-1); however, none of the recognized GR-active compounds on the target-chemical analyte list were detected. The utility of in vitro assays in water quality monitoring was evident from both the quantitative agreement between ER activity and estrogen concentrations, as well as the detection of AR and GR activity for which there were limited or no corresponding target-chemical detections to explain the bioactivity. Incorporation of in vitro bioassays as complements to chemical analyses in standard water quality monitoring efforts would allow for more complete assessment of the chemical mixtures present in many surface waters.

Figure 1.

Concentrations of in vitro estrogenic activity (panels a, b) and estrogen compounds (panels c, d) across sampling sites. In vitro estrogen receptor transcriptional activation (T47D-KBluc, panel a) and bioluminescent yeast estrogen screen (BLYES, panel b) reported as 17β-estradiol equivalents (E2Eq). Independent chemical analyses for suites of analytes were conducted by USEPA (panel c) and USGS (panel d).

Figure 2.

Linear relationship between estimated and actual in vitro activity (log transformed 17β-estradiol equivalents (E2Eq)) in the T47D-KBluc estrogen receptor transcriptional activation assay. E2Eq values were estimated from concentrations of natural and synthetic estrogens from the USEPA chemical analysis (Figure 1) based on relative potency factors of the individual compounds in the T47D-KBluc assay. Dotted lines indicate 95% prediction interval.

Figure 3.

Concentrations of in vitro androgenic activity (panel a) and androgen compounds (panels b, c) across sampling sites. In vitro androgen receptor transcriptional activation (MDA-kb2, panel a) reported as dihydrotestosterone equivalents (DHTEq). Independent chemical analyses for suites of analytes were conducted by USGS (panel b) and USEPA (panel c).

Figure 4.

Concentrations of in vitro glucocorticoid activity across sampling sites. In vitro glucocorticoid receptor transcriptional activation (CV-1 cells transduced with human GR and luciferase reporter genes) reported as dexamethasone equivalents (DexEq).