Pharmaron poster on elucidating drug-induced hepatotoxicity mechanisms via untargeted lipidomics analysis in rat hepatocytes

Poster Authors:

Sicong Zhang, Xiaoxue Li, Jing Lai, Danxi Li, Chunyan Han, Mandy Xu

PK-ADME, Pharmaron Beijing Co., Ltd., Beijing 100176, China

Presented at the 2026 Society of Toxicology (SOT) Annual Meeting & ToxExpo

San Diego, CA | March 22–25, 2026

Why study drug-induced liver injury (DILI)?

Drug‑induced liver injury (DILI) is one of the main reasons why drug candidates fail during development or are withdrawn from the market. In order to reduce the risks and costs associated with bringing a new drug candidate to market it is important to understand mechanistic pathways early.

This poster presents an untargeted lipidomics workflow to investigate how liver cells respond to various toxic drugs. By looking at changes in lipids in rat liver cells, we can identify early signs of toxicity and learn about the biological processes involved, such as:

  • Mitochondrial damage
  • Oxidative stress
  • Apoptosis
  • Cholestasis
  • Steatosis

We used advanced data tools (OSC‑PLS‑DA, enrichment analysis, and machine‑learning) to identify lipid changes that point to different types of liver damage, including apoptosis, mitochondrial disruption, cholestasis, steatosis, and oxidative stress.

The results show that combining lab tests with advanced metabolomics can help predict liver toxicity early. This method is becoming increasingly useful for toxicology, ADME (absorption, distribution, metabolism, excretion), and safety teams focused on making better decisions during drug discovery and early development.

What’s inside the poster?

  • How untargeted lipidomics can highlight early changes linked to different toxicity mechanisms
  • Visual plots showing how toxic and non‑toxic treatments separate
  • Which metabolic pathways change under different toxic stresses
  • How machine learning helps select the most informative biomarker candidates
  • Why rat plated hepatocytes can be a cost‑effective and informative early‑stage model

Download the poster to review all figures, datasets and analysis.