A novel ERβ high throughput microscopy platform for testing endocrine disrupting chemicals

Abstract

In this study we present an inducible biosensor model for the Estrogen Receptor Beta (ERβ), GFP-ERβ:PRL-HeLa, a single-cell-based high throughput (HT) in vitro assay that allows direct visualization and measurement of GFP-tagged ERβ binding to ER-specific DNA response elements (EREs), ERβ-induced chromatin remodeling, and monitor transcriptional alterations via mRNA fluorescence in situ hybridization for a prolactin (PRL)-dsRED2 reporter gene. The model was used to accurately (Z' = 0.58-0.8) differentiate ERβ-selective ligands from ERα ligands when treated with a panel of selective agonists and antagonists. Next, we tested an Environmental Protection Agency (EPA)-provided set of 45 estrogenic reference chemicals with known ERα in vivo activity and identified several that activated ERβ as well, with varying sensitivity, including a subset that is completely novel. We then used an orthogonal ERE-containing transgenic zebrafish (ZF) model to cross validate ERβ and ERα selective activities at the organism level. Using this environmentally relevant ZF assay, some compounds were confirmed to have ERβ activity, validating the GFP-ERβ:PRL-HeLa assay as a screening tool for potential ERβ active endocrine disruptors (EDCs). These data demonstrate the value of sensitive multiplex mechanistic data gathered by the GFP-ERβ:PRL-HeLa assay coupled with an orthogonal zebrafish model to rapidly identify environmentally relevant ERβ EDCs and improve upon currently available screening tools for this understudied nuclear receptor.

Keywords: Endocrine disruptors; Estrogen receptor alpha; Estrogen receptor beta; High content analysis; High throughput microscopy; Zebrafish.

Fig. 1

Characterization of GFP-ERβ:HeLa array cells A) Representative images of GFP-ERβ:PRL-HeLa cells taken following 2-h of the indicated treatments (DMSO, E2 100 nM, 4HT 100 nM). Images are maximum intensity projections acquired at 60x/1.42 and deconvolved. Scale bar:10 μm. B–C) array size (pixels, B) and nuclear GFP intensity (C) metrics after time course analysis of GFP-ERβ:PRL-HeLa cells treated with E2 (10 nM) and 4HT (100 nM). D-E) six-point dose response analysis at the 2-h time-point with E2 and 4HT (1pM to 1 μM) measuring array size (D) and nuclear GFP-ERβ levels (E). In panels B–E the dashed line represents the DMSO control. *p < 0.05 as compared to DMSO control. Error bars are from eight technical replicates.

Fig. 2

ERβ elicits a strong and reproducible transcriptional response from the PRL array (A) Representative images of dsRED2 smFISH after DMSO, E2 (10 nM), or 4HT (100 nM) treatment for 2-h; (B) quantitation of dsRED2 intensity at the PRL array after a time course treatment with E2 (10 nM) or 4HT (100 nM); C) E2 and 4HT dose-response analysis of the dsRED2 response at 2-h; and D) effect of the indicated antagonists on basal and E2 (10 nM)-stimulated dsRED2 transcription. Cells were treated with 4HT (1 μM), ICI182,780 (ICI, 1 μM), PHTTP (1 μM), actinomycin D (ActD, 1 μg/ml), and flavopiridol (Flavo, 100 nM) ± E2 10 nM *p < 0.05 as compared to DMSO control; **p < 0.05 as compared to E2 treated cells. All experiments are represented as mean and standard deviation from eight technical replicates.

Fig. 3

Comparison of ERα/ERβ PRL-HeLa biosensor models shows model sensitivity in differentiating ER selective compounds (A) Heatmap showing logEC₅₀ values for GFP-ERβ:PRL-HeLa and GFP-ERα:PRL-HeLa cells treated with the indicated chemicals (10pM to 1 μM) for 2-h as measured by transcriptional reporter and chromatin remodeling features; (B) Transcriptional response (dsRED2) of ERα (red) and ERβ (green) PRL-array cells treated with multiple doses of a ERβ selective agonist (DPN), a ERβ selective antagonist (PHTTP), a ERα selective agonist (PPT), or a ERα selective antagonist (MPP). Results are shown as mean and standard deviation from 8 technical replicates. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

EPA45 Compound library comparison to ToxCast ERβ dimerization data A) Heatmap showing the effect of EPA45 chemicals on the GFP-ERβ:PRL-HeLa cell line represented as maximal fold change over DMSO control. Data is shown as average of three independent screens. B–C) Comparison of determined logEC₅₀ values of GFP-ERβ:PRL-HeLa array size (B) and dsRED2 intensity (C) vs. Odyssey Thera BiFC ERβ-ERβ dimerization assay. D-E) dose-response analysis for corticosterone (D) and phenobarbital (E) using the array size and dsRed2 intensity metrics. Data is average ± standard deviation of three independent experiments.

Fig. 5

Comparison of ERs using PRL-HeLa models identifies ERβ interacting compounds in the EPA45 library of control chemicals Scatter plots comparing logEC₅₀ values for the array size (A) or the dsRED2 transcriptional output (B) metrics in ERα and ERβ-containing cells treated for 2-h with a six-point dose-response of the EPA45 chemical library.

Fig. 6

Transgenic ERE:GFP reporter zebrafish as an in vivo orthogonal assay to define ERα vs. ERβ activities Representative images of the ERE:GFP Zebrafish at 3dfp treated with: (A) DMSO; (B); 10 nM E2; (C) 1 μM DPN; and, D) 1 μM genistein. Red arrows highlight GFP signal in the liver while white arrows indicate the heart valve. Whole-well images were captured at 4x/0.16. Scale bar is 100 μm. A minimum of ten Zebrafish per treatment were assayed. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Supplementary Figure 1. Validation of the PRL-ERβ model through immunofluorescence, Western blot, and doxycycline titration of ERβ expression. A-B) Immunofluorescence validation. PRL-ERβ (A) and PRL-ERα (B) cells were immunolabeled with three different ERβ primary antibodies from Abcam (top panels), GeneTex (middle panels), and Upstate (now available from Millipore-sigma) (bottom panels) and compared with the GFP-ER localization. Images are maximum intensity projections after being captured at 20x/0.75. Scale bar is 20 μm. C) Titration experiment of doxycycline treatment for 24 h in the PRL-ERβ cell line to determine GFP-ERβ nuclear expression. D) Western blot showing ERβ expression selectively in PRL-ERβ (lane 2) vs. PRL-ERα (lane 3) cells using the Millipore/Sigma antibody. Lane 1 is the molecular weight marker.