Assessment of endocrine disruptors: approaches, issues, and uncertainties

This paper focuses on the quantitative risk assessment of environmental endocrine-disrupting chemicals (EDCs) to human health. An EDC can be defined as an exogenous agent that interferes with the normal endocrine signaling and communication mechanisms. The normal feedback control system of natural hormones is responsible for regulatory mechanisms that maintain homeostasis. Hormones deliver their message to target cells by interacting with receptors, initiating signal transduction, gene transcription, and mRNA translation, and ultimately leading to cellular response. Because effects of EDCs include diverse disease endpoints such as cancer, reproductive toxicity, developmental toxicity, immune system effects, acute toxicity, and neurotoxicity, risk assessment of EDCs is necessarily endpoint-specific. From the quantitative viewpoint, it is best to model the normal endocrinology and then extend the model to include impacts attributable to a particular exogenous agent. A practical approach to such a complex process is to break the spectrum of biochemical and biological events into modular components: e.g., pharmacokinetics, biochemical/molecular (including cellular signaling), and cellular response/dynamics. A flexible mathematical procedure that is capable of modeling each of these components is suggested. However, a real biologically based model is not yet feasible because of a lack of necessary biological information. A challenge to risk assessors is how to develop a hybrid risk assessment approach that can use the limited biological information available for a specific agent and avoid relying on a default approach that incorporates no biological information. The USEPA's default approach is to derive benchmark dose (BMD) or benchmark concentration (BMC) on the basis of a predetermined empirical dose-response model. BMD (or BMC) is the highest dose (or concentration and duration) of exposure that is considered unlikely to cause adverse effects in a human population, including sensitive subgroups. Data from two studies are used to stimulate discussion of issues and the needs for new quantitative approaches and data for assessing endocrine disruptors. Statistical concepts about threshold effect and the U-shaped dose-response relationship are also discussed. This report is a condensed version of the one to be published in the monograph of the NATO Advanced Research Workshop, Endocrine Disrupters and Carcinogenesis Risk Assessment" held May 8-12, 2001, in Bialystok, Poland.