Cytotoxicity-driven prediction of developmental toxicity is complemented by transcriptional biomarker signatures of perturbed neuroectodermal differentiation in vitro

Abstract Developmental toxicity testing requires human-relevant approaches that provide both predictive information and mechanistic insight. Here, we investigated the relative contributions of cytotoxicity and targeted transcriptomics to developmental toxicity assessment in a human induced pluripotent stem cell-based model of early neuroectodermal differentiation. A biologically anchored 190-gene transcript biomarker panel was derived from a previously published transcriptome dataset using complementary biomarker-selection strategies. An independent set of 18 developmental toxicants and non-toxicants was then evaluated based on concentration-dependent cytotoxicity normalized to maternal plasma C max values and biomarker expression. We found that cytotoxicity provided substantial discriminatory power in random forest-based classification of developmental toxicants and non-toxicants, yielding an AUC of 0.852, demonstrating robust discrimination for a complex and heterogeneous toxicological endpoint. In contrast, transcriptional responses were frequently detected at concentrations substantially below those causing cytotoxicity, but this increased sensitivity did not improve random forest classification of the two groups. Instead, targeted transcriptional profiling revealed compound-specific perturbation patterns that were not captured by cytotoxicity. Principal component analysis and Differentiation Pattern Clustering (DiPaC) distinguished impaired progression towards neuroectodermal identity from activation of divergent developmental transcriptional programs, exemplified by methoxyacetic acid and dimethadione, respectively. These findings demonstrate that cytotoxicity assessment normalized to physiological maternal plasma concentrations can provide substantial discriminatory information in an in vitro model of early embryonic development, while targeted transcriptional profiling provides complementary mechanistic insight into how developmental differentiation is perturbed. Distinguishing predictive from mechanistic value may therefore be important for the interpretation and integration of transcriptomic endpoints in developmental toxicity assessment.

Authors

Institutions

Publication Details

Journal
Archives of Toxicology
Published
2026-11-17
DOI
https://doi.org/10.1007/s00204-026-04643-1
Primary Topic
Animal testing and alternatives
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Cytotoxicity-driven prediction of developmental toxicity is complemented by transcriptional biomarker signatures of perturbed neuroectodermal differentiation in vitro

Patrick Nell, Antonia Thomitzek, Jan Georg Hengstler, David Feuerborn et al.
Archives of Toxicology
Animal testing and alternatives
article

Cytotoxicity-driven prediction of developmental toxicity is complemented by transcriptional biomarker signatures of perturbed neuroectodermal differentiation in vitro

Patrick Nell, Antonia Thomitzek, Jan Georg Hengstler, David Feuerborn, Andreas Scholtz-Illigens, Franziska Kappenberg, Karolina Edlund, Jörg Rahnenführer, Sarah M. Seidel, Katharina Derksen, Daniel Hösle, Annalena Weissert, Franziska Schneider
article en

Abstract

Abstract Developmental toxicity testing requires human-relevant approaches that provide both predictive information and mechanistic insight. Here, we investigated the relative contributions of cytotoxicity and targeted transcriptomics to developmental toxicity assessment in a human induced pluripotent stem cell-based model of early neuroectodermal differentiation. A biologically anchored 190-gene transcript biomarker panel was derived from a previously published transcriptome dataset using complementary biomarker-selection strategies. An independent set of 18 developmental toxicants and non-toxicants was then evaluated based on concentration-dependent cytotoxicity normalized to maternal plasma C max values and biomarker expression. We found that cytotoxicity provided substantial discriminatory power in random forest-based classification of developmental toxicants and non-toxicants, yielding an AUC of 0.852, demonstrating robust discrimination for a complex and heterogeneous toxicological endpoint. In contrast, transcriptional responses were frequently detected at concentrations substantially below those causing cytotoxicity, but this increased sensitivity did not improve random forest classification of the two groups. Instead, targeted transcriptional profiling revealed compound-specific perturbation patterns that were not captured by cytotoxicity. Principal component analysis and Differentiation Pattern Clustering (DiPaC) distinguished impaired progression towards neuroectodermal identity from activation of divergent developmental transcriptional programs, exemplified by methoxyacetic acid and dimethadione, respectively. These findings demonstrate that cytotoxicity assessment normalized to physiological maternal plasma concentrations can provide substantial discriminatory information in an in vitro model of early embryonic development, while targeted transcriptional profiling provides complementary mechanistic insight into how developmental differentiation is perturbed. Distinguishing predictive from mechanistic value may therefore be important for the interpretation and integration of transcriptomic endpoints in developmental toxicity assessment.

Archives of Toxicology
TU Dortmund University (DE), University Hospital Bonn (DE), Max Planck Institute of Molecular Physiology (DE), Leibniz Research Centre for Working Environment and Human Factors (DE)
Animal testing and alternatives
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon