Computer model catches harmful chemicals

Wednesday 15 Aug 12

A computer model developed at the National Food Institute, Technical University of Denmark can contribute to the evaluation of chemical safety in food, drugs and consumer products – and thus help to identify and catch substances before they do any harm. The model has identified 8,000 potential endocrine-disrupting substances so far.

Chemical pollution of the environment and food is strongly suspected of having harmful effects on human health. But now the National Food Institute has developed a computer model that can help identify and catch substances before they do any harm.

Chemicals must not activate receptor

The computer model catches the harmful chemicals by predicting whether a given substance can bind to and activate a certain receptor in the cells of the body, PXR (pregnane X receptor). PXR plays a key role in regulating the balance of our hormones, both sex hormones and thyroid hormones, which are important for brain development. This means that PXR is involved in controlling many processes in the body.  

If humans are exposed to chemicals that activate PXR, it can affect our natural hormones and the balance between them, leading to adverse health effects. A large number of known endocrine disrupters are capable of activating PXR, which indicates that the receptor plays an important role in endocrine disruption.

8,000 substances can disturb important functions

The computer model was used to screen around 50,000 chemicals and indicated that more than 8,000 can bind to the receptor and potentially disturb important functions of the body. These are substances that can be present in our food, environment and cosmetic products.  

The computer model is a so-called QSAR computer model. QSAR stands for "quantitative structure-activity relationship". At the National Food Institute, QSAR models are used to calculate and predict the harmful effects of chemicals on human health.

The National Food Institute has set up a QSAR database with predictions for nearly 185,000 substances from more than 150 different models. The database is continuously expanding as new models are developed, offering unique possibilities for conducting statistical analyses, which can help uncover new and unknown biological correlations.

A detailed analysis of data from the combined database shows that foetal defects, hormone disruption and genetic damage occur more frequently among the chemicals predicted to bind to PXR. Overall the work performed at the National Food Institute indicates that PXR binding is linked to an increased risk of serious health effects.

Model can also be used in drug development

In general the QSAR models increase the information about a given substance, minimise animal testing and save both companies and the authorities time and money. They can also be used to prioritise testing and regulatory activities of chemicals.  

The PXR computer model could also provide important information, not only when assessing health effects, but also in the development of new chemicals and drugs. Knowledge about PXR activation can be an important piece of information for identifying problematic substances and promote their substitution with less harmful alternatives.

Find out more

Find out more about the PXR computer model in an article published in "Toxicology and Applied Pharmacology": QSAR model for human pregnane X receptor (PXR) binding: Screening of environmental chemicals and correlations with genotoxicity, endocrine disruption and teratogenicity.