Abstract
BACKGROUND: The vision of a National Research Council (NRC) committee (the Committee on Toxicity Testing and Assessment of Environmental Agents) for future toxicity testing involves the testing of human cells in in vitro assays for "toxicity pathways"--normal signaling pathways that when perturbed can lead to adverse effects. Risk assessments would eventually be conducted using mathematical models of toxicity pathways (TP models) to estimate exposures that will not cause biologically significant perturbations in these pathways. OBJECTIVES: In this commentary we present our vision of how risk assessment to support exposure standards will be developed once a suitable suite of in vitro assays becomes available. DISCUSSION: Issues to be faced basing risk assessments on in vitro data are more complex than, but conceptually similar to, those faced currently when applying in vivo data. Absent some unforeseen technical breakthrough, in vitro data will be used in ways similar to current practices that involve applying uncertainty or safety factors to no observed adverse effect levels or benchmark doses. TP models are unlikely to contribute quantitatively to risk assessments for several reasons, including that the statistical variability inherent in such complex models severely limits their usefulness in estimating small changes in response, and that such models will likely continue to involve empirical modeling of dose responses. CONCLUSION: The vision of the committee predicts that chemicals will be tested more quickly and cheaply and that animal testing will be reduced or eliminated. Progress toward achieving these goals will be expedited if the issues raised herein are given careful consideration.