SOT’s read-across workshop, “Read-Across: Building Scientific Confidence in the Development and Evaluation of Read-Across for Regulatory Purposes Using Tox21 Approaches,” was presented to a standing-room only crowd hungry for insight into perfecting the art of read-across. In the read-across approach, endpoint information for one chemical (source chemical) is used to predict the same endpoint (such as cancer or aquatic toxicity) for a similar chemical (target chemical). Usually, this is based on structural similarity or sharing mode or mechanisms of action. For data-poor chemicals, read-across shows great promise to assess physicochemical properties, toxicity, environmental fate, and ecotoxicity endpoints.
The use of read-across continues to expand around the world, propelled largely by REACH data requirements. Thomas Hartung said that more than 75 percent of REACH dossiers include read-across, while the use of other data gap filling approaches, such as QSAR, trail well behind. Despite the widespread use of read-across, expertise among rank and file toxicologists is relatively limited.
Currently, there is not agreement and alignment on what constitutes a good read-across prediction. What does success really look like in a read-across? According to Grace Patlewicz, read-across beauty is in the “eye of the beholder,” which underscores the need to strengthen methods to perform read-across, as well as to characterize uncertainty to communicate strength in the read-across prediction.
Workshop attendees learned about OECD’s recent evaluation of four case studies from around the world that evaluated differences in read-across evaluations, including a Canadian read-across of 14 substituted diphenylamines and a Japanese read-across of 19 allyl esters. Many read-across practitioners are familiar with OECD’s guidelines on grouping chemicals for read-across (2nd edition, 2014), which can be downloaded from OECD at no charge. OECD’s 141-page guidance provides a good framework for read-across, but has limited practical examples for newer practitioners to follow. OECD’s new case study evaluation demonstrates that OECD is focused on moving guidance into practice. This approach is proven to build expertise, and the US Environmental Protection Agency took a similar approach years ago when rolling out benchmark dose modeling to the masses. Hopefully, more read-across case studies will be published by OECD and others.
Nicholas Ball warned of the dangers of negative read-across, which occurs when reading across to confirm the absence of toxicity for one endpoint (such as liver toxicity), but failing to recognize that the surrogate chemical may cause toxicity for another endpoint (such as testicular toxicity).
A treasure trove of upcoming meetings and publications mentioned during the workshop should contribute to increased rigor and body of knowledge that goes into read-across.