Related Experiment Video
Updated: Mar 9, 2026
11:06
Network Pharmacology Prediction and Metabolomics Validation of the Mechanism of Fructus Phyllanthi against Hyperlipidemia
Published on: April 7, 2023
2.9K
Combining UHPLC profiling and random walk network-based in vitro analysis to predict herb-induced liver injury
Kwanyong Choi1, Jun Young Park2, Sunyong Yoo2
1Department of Food Science and Biotechnology, Seoul National University of Science and Technology, 232, Gongneung-ro, Nowon-gu, Seoul, 01811, Republic of Korea.
Summary
This study introduces a novel computational framework to predict herbal medicine hepatotoxicity. The method successfully identified potentially harmful herbs, guiding safer preclinical screening of botanical products.
Area of Science:
- Toxicology
- Pharmacology
- Computational Biology
Background:
- Herbal medicines are widely used globally.
- Their potential for liver toxicity (hepatotoxicity) is often underestimated.
- Predictive toxicology methods for herbal products are underdeveloped.
Purpose of the Study:
- To develop and validate a predictive toxicology framework for assessing herbal medicine hepatotoxicity.
- To identify specific pathways involved in herb-induced liver injury.
- To enable scalable preclinical screening of herbal products for safety.
Main Methods:
- Combined ultra-high-performance liquid chromatography (UHPLC) compound profiling with a Random Walk with Restart (RWR) network approach.
- Utilized herb-compound-target associations filtered by statistical thresholds.
- Validated predictions using in vitro assays (cell microscopy, mitochondrial membrane potential, ALT/AST release) and transcriptomic profiling (RNA sequencing, qRT-PCR) in HepG2 cells.
Main Results:
- RWR prioritized apoptosis, oxidative stress, and inflammatory pathways for six of seven tested herbs.
- The six prioritized herbal extracts induced significant hepatotoxicity in vitro, consistent with predictions.
- Astragalus membranaceus showed limited pathway enrichment and minimal toxic effects.
- RNA sequencing revealed broad transcriptomic perturbations and clustering of hepatotoxic extracts.
Conclusions:
- The integrated computational and experimental framework effectively predicts herbal medicine hepatotoxicity.
- This approach supports scalable preclinical safety assessment of botanical products.
- Expanding herb-compound-target databases can further enhance predictive accuracy for toxicological safety.
