By screening the intestines of experimental mice, researchers can weed out those mice that do not respond well to the substance being examined. This can in some cases mean a halving of the number of mice required for an experiment, and perhaps even more. (Photo: CALAR)
By screening the intestines of experimental mice, researchers can weed out those mice that do not respond well to the substance being examined. This can in some cases mean a halving of the number of mice required for an experiment, and perhaps even more. (Photo: CALAR)

Can we avoid animal testing entirely?

Scientists are working flat out to find alternatives to animal testing. QSAR computer models are looking promising.

Published

A research team at the Danish National Veterinary Institute has spent years feeding computer models with information about toxins. These models help them study unknown substances for health hazards, and they can also in certain situations reduce or eliminate the need for animal testing.

The so-called QSAR models (Quantitative Structure-Activity Relationships) are computer models with a database filled with information about the chemical structures of known substances, and whether they can e.g. cause allergies or disrupt hormones. A number of mathematical models are then added to the database.

When researchers want to use a new substance in the development of a new drug, or perhaps when they need to use a softener in a plastic product, the models can compare the chemical structure of the new substance with existing substances in the database. If the new substance is similar to a known allergenic substance, can cause cancer or is harmful in some other way, the database will say so.

This is a clever way to eliminate potentially toxic substances and focus on those that appear not to be harmful. This way you can avoid spending time on animal testing of the substances which have been predicted to be harmful.

Models help eliminate poor drug candidates

If the models indicate that a substance may have unwanted side effects, then you can accelerate relevant experimental studies of them to ‘eliminate’ poor drug candidates before investing large sums in them. Such targeting could potentially remove the need for animal testing.

Eva Bay Wedebye

”These models can be used early on in e.g. the development of a new drug. You only need to have the chemical structure of your substance described on a piece of paper, and then you can run it through the models,” says Eva Bay Wedebye, a chief adviser at the Division of Toxicology and Risk Assessment at the National Veterinary Institute, the Technical University of Denmark (DTU VET).

“If the models indicate that a substance may have unwanted side effects, then you can accelerate relevant experimental studies of them to ‘eliminate’ poor drug candidates before investing large sums in them. Such targeting could potentially in certain situations remove the need for animal testing.”

Prior to the arrival of the QSAR models, this kind of information could only be arrived at through animal testing, so in this way the models are an important step towards reducing the number of animal experiments.

However, the method is not perfect. It can only be as good as the data in the database and will always be associated with some uncertainty. Nevertheless, the certainty increases as the database continues to grow.

Fewer animal trials – a lucky side effect

I am convinced that in ten years from now this will be entirely incorporated into our experimental methods.

Eva Bay Wedebye

The initial purpose of using these models was not to reduce the number of animal trials; this was just a lucky coincidence. The models were initially developed to make it easier to figure out if chemicals can be harmful to health.

This is also why the QSAR research group at DTU is working together with the Danish Ministry of the Environment. But the QSAR models remain a good alternative to animal testing, which would otherwise be necessary to determine the toxicity of unknown substances.

Unfortunately, many test animals are being used in other types of research, for instance in basic research. Here, too, scientists have come up with alternatives, one of which looks at the type of bacteria the animals have in their intestines.

Intestinal bacteria in mice explain test results

If you know the composition of bacteria in the intestines of your experimental mice, you can make do with only half as many mice. This may sound a bit odd, but the explanation is that the intestinal bacteria in the mice play a big role.

QSAR (Quantitative Structure-Activity Relationships) is a collective term for a number of models that compare the chemical structure of known organic substances with unknown substances. If, for example, we know that a certain substance causes hormonal imbalance, there is a great risk that similar substances do the same.

“If you screen the intestinal flora in experimental mice and pick only those that respond best, you can do with only those in your experiment. This can in some cases mean a halving of the number of mice, and in other cases even more,” says Axel Kornerup Hansen, a professor at the Department of Veterinary Disease Biology at the University of Copenhagen, who is also on the board of a newly-created centre that promotes alternatives to animal testing.

In one experiment, Hansen’s research group could explain up to 70-80 percent of the variation in how mice responded to having an allergenic product put on their ear – based only on their intestinal flora.

This means that the bacterial mix that the mice have in their stomachs greatly determines how their ear responds to contact with an allergenic substance.

A lot of money to be saved

Today, researchers use inbred mice to avoid noise in the experimental data. For example, it does not make much sense to use mice in a diabetes experiment if the mice do not get diabetes as a result of being different from the other mice. Those mice are then rendered redundant.

This can be sorted out by comparing the intestinal bacteria. If the intestinal flora gives widely different results in the experiment, this can muddy the results. If, however, it is determined beforehand that the mice have the same kind of bacteria – or at least if this is taken into account in the subsequent analysis – then a lot of noise can be removed from the experimental data. Without this noise, it is possible to reach the same conclusions using fewer animals.

The research into the importance of bacteria can be widely used to cut down on the number of test animals – even in basic research, where a vast majority of test animals meet their fate.

”I am convinced that in ten years from now this will be entirely incorporated into our experimental methods,” says Hansen, whose research group has tested its methods with many of the largest pharmaceutical companies in Denmark.

There is a smouldering interest in these possibilities, not least because there is a lot of money to be saved in cutting down the number of test animals.

-------------------------

Read the Danish version of this story at videnskab.dk

External links