Systematic comparison of temporal hepatotoxicant-induced gene network responses across three liver test systems.

Abstract

Drug-induced liver injury (DILI) arises from dynamic and time-dependent cellular stress responses that remain insufficiently captured by conventional single-timepoint toxicogenomic assessments. We systematically characterized temporal and concentration-dependent transcriptomic responses to the clinically relevant hepatotoxicants ketoconazole, diclofenac and nitrofurantoin across three human liver in vitro models: primary human hepatocytes (PHH), hiPSC-derived hepatocyte-like cells (HLC) and HepG2 cells. Time-resolved RNA sequencing (0-48 hours) combined with likelihood ratio testing identified time-responsive genes (TRGs), which were subsequently integrated into TXG-MAPr gene co-expression modules to enable mechanistic interpretation at the network level. Across all models and compounds, a conserved core stress response was observed, characterized by activation of ER stress (ATF4), oxidative stress (NRF2) and heat shock (HSF1) pathways, while distinct model-specific adaptive programs reflected differences in metabolic competence and differentiation status. Mapping TRGs onto co-expression networks revealed coordinated temporal activation patterns and highlighted both shared and system-specific transcriptional programs. Concentration-response analysis at 24 hours demonstrated that module-level transcriptomic points of departure (tPODs) were highly reproducible across models for a subset of functionally annotated networks, particularly ER stress modules associated with hepatocellular injury in vivo. Notably, these modules showed substantial gene-level concordance across systems, supporting their biological robustness and translational relevance. These findings establish that time-resolved, network-based transcriptomics provides mechanistically grounded, reproducible and quantitative endpoints that enhance cross-system comparability and offer a scalable framework for regulatory toxicology and next-generation chemical risk assessment.

Key Findings

• A conserved core stress response involving activation of ER stress (ATF4), oxidative stress (NRF2), and heat shock (HSF1) pathways was observed across three human liver in vitro models exposed to hepatotoxicants.
• Time-resolved RNA sequencing combined with gene co-expression network analysis revealed coordinated temporal activation patterns and both shared and model-specific transcriptional programs.
• Module-level transcriptomic points of departure (tPODs) for ER stress modules were reproducible across models and showed substantial gene-level concordance, supporting biological robustness and translational relevance.

Clinical Significance

Citation

Danilyuk Tamara Y, Schouten Marou, Burgers Elsje Jet al.. Systematic comparison of temporal hepatotoxicant-induced gene network responses across three liver test systems. Toxicological sciences : an official journal of the Society of Toxicology. 2026-Jul-24.