Tests on lab critters must be more relevant to human trials, scientists say

Different types of rat respond to drugs in substantially different ways that can be tracked by metabolic analysis, according to scientists who say their finding has major implications for designing animal experiments.

Led by Julian Griffin of the University of Cambridge, UK, the researchers looked at the metabolic response of two different strains of rat exposed to the same liver toxin, orotic acid. This compound causes changes in the liver that mimic the symptoms of human fatty liver disease, thus providing a model for the disease.

A close look at the rats’ liver tissue, however, showed that things are not so straightforward. A comparison between Wistar and Kyoto rats, widely used experimentally for drugs testing, showed that resulting liver metabolites varied greatly between the two rat strains.1 

’This shows us that the strain and genetic profile of one animal can make its response to a drug, or susceptibility to a disease, very different,’ Griffin told Chemistry World. ’It’s important to understand how the strain interacts with a treatment or lesion in order to have an effective animal model.’

Flawed model

The study follows a recent report in the British Medical Journal that questioned the effectiveness of animal studies in developing medical treatments for humans. The report took a selection of established clinical interventions and compared the results of animal tests of the treatments with human clinical trails. In three out of the six interventions they looked at, animal and clinical results produced fundamentally different results. 

Peter Sandercock, one of the authors of the BMJ study told Chemistry World  that animal and clinical researchers needed to talk to each other when setting up animal studies. ’Experiments need to be designed to translate well into clinical study,’ he said. ’Clinical researchers need to engage in dialogue with researchers who design animal experiments for this disparity to be addressed.’

Dialogue could go some way to solving this problem. But Griffin believes that studying the metabolic responses of experimental animals can directly improve animal models, making them more relevant to their clinical counterparts - potentially by more closely mimicking the human metabolic response.

This kind of detailed metabolic profiling, which reveals the biochemical responses to drugs, environment and disease, has been dubbed metabonomics.

Donald Robertson, a toxicologist working for Pfizer in the US, agrees that metabonomics has a role in refining animal models and bringing them closer to the clinic but throws in a note of caution for drug development. ’Humanising animal models is certainly useful for efficacy testing but it’s far too early to say what effect this will have in drug safety testing,’ he told Chemistry World

’The very valuable thing that can be achieved through understanding these metabonomic differences is to reduce the number of animals needed in a trial,’ he emphasises. ’Often toxicologists use a large population of animals in order to cancel out any bias caused by these differences, but if we understood them, this would not be necessary.’ 

Victoria Gill