Advancing NAMs for Risk Assessment: Perspectives from SOT 2025
By Breanne Kincaid | April 17th, 2025
In March, I was fortunate to attend the Society of Toxicology’s annual meeting in Orlando, Florida, where experts in toxicology for risk assessment presented updates and overviews about the use of New Approach Methodologies (NAMs) in human health risk assessment.
NAMs are scientific models and test strategies that reduce reliance on animal data and enable human-relevant investigations of toxicity. They can include a variety of technologies: in vitro models (ranging from single cell types or complex microphysiological systems (MPS) that recapitulate tissue structure and functionality), computational models (activity and hazard prediction), transcriptomics methodologies (assessing all RNA molecules (gene transcripts) produced in a cell or tissue at a given time to understand which genes are active and how they respond to different conditions, such as a chemical exposure).
As regulatory agencies, in particular the U.S. Environmental Protection Agency (EPA), increasingly recognize the value of integrating NAMs, researchers, and stakeholders are beginning to integrate these technologies into their risk assessments efforts. These efforts were especially apparent in the conference session, “Evolution of Human Health Risk Assessment: New Science and Approaches That Are Ready for Prime Time.”
Matthew Gagne provided an international perspective of Health Canada’s approach to applying in vitro and in silico methods for chemical prioritization and screening through the agency’s computational toolkit, HAWPr. This platform integrates diverse data streams—including international regulatory records, in vitro assay results from EPA ToxCast, and predictive inputs from the OECD QSAR Toolbox—to automate risk assessment efforts. By leveraging expert rules and establishing a data hierarchy (in vivo > in vitro > in silico), the toolkit not only flags hazard indicators but also assigns confidence levels to predictions. Notably, the incorporation of IRAP (Identification of Risk Assessment Priorities) enables a systematic approach to hazard classification across endpoints such as carcinogenicity, mutagenicity, developmental toxicity, reproductive toxicity, repeat dose toxicity, and endocrine disruption. By instituting strict filtering criteria—such as adopting the U.S. National Toxicology Program (NTP) Integrated Chemical Environment (ICE) platform flags to eliminate spurious response curves and aligning assay provider, cell type, and cytotoxicity thresholds—the approach ensures that only robust in vitro signals are carried forward. This enhancement ultimately strengthens the predictive capacity of risk assessments.
In parallel, developments in in vitro transcriptomics were highlighted as a promising strategy for establishing transcriptomic points of departure (tPODs) – the lowest dose of a chemical that causes an observable adverse effect, based on gene transcript data – by mapping gene transcripts to known adverse outcome pathways (AOPs) – the cellular and molecular events that culminate in an observable negative health effect. The tPOD is used as a starting point to determine safe exposure levels for humans. Impressively, results from the tPOD approach closely replicated the existing POD for numerous chemicals that had been set using combinations of animal and in vitro data. Such methods are not only invaluable for reducing animal use in toxicity testing but also offer insights into mechanistic underpinnings of toxicity.
Alison Harrill introduced EPA’s Transcriptomic Assessment Product (ETAP), further highlighting the value of transcriptomics approaches. ETAP provides a workflow for using transcriptomic data from short-term animal studies to generate a tPOD and reference dose (RfD), which is an estimate of daily exposure that is unlikely to introduce an appreciable risk of adverse effects over a lifetime, for chemicals lacking toxicity or human epidemiological data for health risk assessments.
Kamin Johnson of Corteva Agriscience further solidified the practical application of these methodologies by discussing how gene expression data from medium-term exposure studies can be used to derive molecular PODs that inform chronic risk assessments for agrochemicals. Through a detailed case study on the pesticide halauxifen-methyl, Johnson illustrated how gene expression endpoints tied to a known AOP can serve as sensitive biomarkers to define a protective POD. This approach not only streamlines regulatory decision-making for new agrochemicals but also underscores the importance of mechanistic data in hazard evaluation.
Together, these presentations illustrate a clear shift towards integrated, mechanistically informed risk assessment strategies that reduce the number of animals required to generate toxicity data—paving the way for regulatory frameworks that are both scientifically robust and economically sustainable.
The views expressed do not necessarily reflect the official policy or position of Johns Hopkins University or Johns Hopkins Bloomberg School of Public Health.
