Development of a curated Hershberger database

Abstract

A systematic literature review was conducted to identify Hershberger bioassays for ∼3200 chemicals including those used to validate the OECD/US EPA guideline assay, US EPA's chemicals screened for endocrine activity, and the library of chemicals run in US EPA 's ToxCast in vitro assays. For 134 chemicals that met pre-defined criteria, experimental results were extracted into a database used to characterize uncertainty in results and evaluate the concordance of the Hershberger assay with other in vivo rodent studies that measure androgen-responsive endpoints. Of 25 chemicals tested in >1 Hershberger study, 28% had disagreements between studies (i.e. ≥1 positive and ≥1 negative study), and of the 65 chemicals tested in Hershberger studies and other in vivo studies with androgen-responsive endpoints, 43% indicated disagreements, though in some cases these may be explained by differences in study designs or physiology of the animal model. Ultimately, 49 chemicals were identified with reproducible androgen pathway responses confirmed in ≥2 in vivo rodent studies that could be considered reference chemicals useful for validating alternative methods.

CoMPARA: Collaborative Modeling Project for Androgen Receptor Activity

“…First, a multiassay AR pathway model was developed based on the results of 11 assays covering the androgen signaling pathway and combining the in vitro results into an AUC score representing the whole AR activity to mimic the in vivo results (Kleinstreuer et al 2017). These assays were run on the same initial library of 1,855 ToxCast™ chemicals that the ER assays were run on, and the developed pathway model was validated using reference chemicals with known in vitro results from the literature (Kleinstreuer et al 2017) and further compared with a set of chemicals with reproducible results in vivo (Browne et al 2018;Kleinstreuer et al 2018a). Note that the goal of this project is to predict in vitro AR activity.…”

Evaluation of androgen assay results using a curated Hershberger database

“…The current EDSP Tier 1 screening assays for AR pathway activity are the AR in vitro binding assay and the in vivo rodent Hershberger, a short-term bioassay for identifying (anti)androgenic effects of chemicals on accessory sex tissue (AST) weights [12,13] (US EPA 2009a, OECD 2009). The database described in the companion paper [1] (Browne et al in prep) includes 602 Hershberger studies on 216 unique chemicals, and details on test chemical, experimental design, and experimental results were recorded and assessed using a priori criteria to ensure quality of results. Reference chemicals were identified based on in vivo effects on the androgen pathway that were reproducible in more than one Hershberger study, or in a Hershberger study and another in vivo study design with androgen responsive endpoints.…”

“…

A set of 39 reference chemicals with reproducible androgen pathway effects in vivo, identified in the companion manuscript [1], were used to interrogate the performance of the ToxCast/Tox 21 androgen receptor (AR) model based on 11 high throughput assays. Cytotoxicity data and specificity confirmation assays were used to distinguish assay loss-of-function from true antagonistic signaling suppression.

…”

Workshop on the validation and regulatory acceptance of innovative 3R approaches in regulatory toxicology – Evolution versus revolution

“…The cells are simultaneously treated in test compound and differentiation is phenotypically assessed in a single well using three distinct phenotypic endpoints: adipogenic gene activation (FABP4 protein expression), lipid accumulation (AdipoRed neutral lipid staining), and adipokine secretion (adiponectin secretion). (Kleinstreuer et al, 2016), (Hoffmann et al, 2018), (Browne et al, 2018), it might be better not to compare DA data directly with animal data to calculate values of predictivity. Rather, the ultimate decision after DA-based IATA evaluations should be compared with current, accepted WoE decision matrixes to calculate the statistical performance.…”