Model Predicts Drug-Induced Liver Toxicity

A novel model can determine the effect of certain drugs on human liver cells.

The authors of a new study used a new computational modeling process to observe and compare the toxic effects 15 drugs may have on the liver.

Prescription drugs and medication combinations that treat various conditions can result in liver damage and serious side effects. Laboratory experiments involving liver cells can determine the underlying mechanisms behind organ damage, but they cannot accurately determine how the drugs may impact human patients. Even preclinical animal studies cannot accurately depict the potential damage.

To improve the conversion from in vitro experiments to real-world data, researchers created a novel computational modeling approach to simulate how liver cells are impacted by different drugs, according to the study, which was published by PLOS Computational Biology.

The investigators compiled observations and clinical evidence about how drugs are dispersed and metabolized in the body to develop a more accurate view of how different drugs in different doses can affect the liver.

Previously, the team of investigators confirmed the accuracy of this approach in a proof-of-concept study. In the current study, they compared the potentially toxic effects of 15 drugs at clinically relevant doses through the model.

First, the researchers created whole-body models that mimic a specific drug after ingestion. Then, lab data were integrated into the model to more accurately predict the drug’s effect on liver cells, according to the study.

The investigators found that different drug classes caused similar responses, such as which specific genes would be involved with the response to a toxic dose, according to the study.

Although further studies are needed, this new computational approach may lead to faster diagnosis of toxic side effects. It could also determine which gene transcripts could become a marker of liver toxicity to develop a fast, minimally invasive test for liver toxicity, according to the study.

The model may also prove useful as a predictor of toxicity for investigational drugs to ensure their safety prior to regulatory approval, and could even be used to determine potentially dangerous drug-drug interactions that can lead to hospitalization or even death in some cases.

"Consistently applied to the design of clinical development programs, the approach presented has the potential to early identify medical and economic risks of new drugs," concluded study co-author Dr Lars Kuepfer.