Animal research figures – categorising severity
Every year the Home Office publishes figures on the use of animals in scientific procedures in the UK. This year, it has changed its methods to record the maximum severity of the procedures every animal has experienced in its lifetime. This way of counting will produce a much more accurate picture of animal research in the UK. But what exactly is a procedure? And what’s the difference between a mild procedure and a severe one? Here Dr Sara Wells, Director of the MRC Mary Lyon Centre, explains.
Let’s start with the basics. Why exactly does the Home Office publish numbers every year?
Animal research in the UK is regulated under the Animals (Scientific Procedures) Act 1986. This requires that every year the Home Office publishes information about protected animals used in scientific research.
To carry out research with these animals, researchers must apply for a project licence from the Home Office, and every licence-holder must then complete ‘annual returns’ about the animals used. All these returns are consolidated and form the report published today.
Yes. This means any living vertebrate other than man and any living cephalopod (eg octopus, cuttlefish, squid).
The use of other animals in research, such as fruitflies or nematode worms, is not subject to the same kind of scrutiny.
What kind of information is published?
Until this year, the reports included information about the number of regulated procedures started on animals in a calendar year.
From this year, the reports count every animal at the end of the experiment, as well as the maximum severity of procedures it has experienced in its lifetime. This is much more of an administrative burden, but also a far more accurate and transparent way of assessing the cost of research in terms of what the animals actually experience.
In 2014 3.87 million procedures were performed in UK laboratories. This is split almost exactly between experimental procedures (1.93 million) and animals that have been used for breeding purposes alone (1.94 million).
I guess that means we can’t compare this year’s report with last year’s then?
You’re right – they’re not comparable. And because the old method counted when a procedure began and the new method counts when it finished, there is potential for double counting if people try to compare them.
This year’s report is really starting with a clean slate. But it does mean that numbers should now be comparable across Europe. The change in reporting has been brought in because an EU Directive on animal research has been transposed into UK law. However many of the scientific community have supported the idea of retrospective severity assessment for a while.
What’s the benefit of measuring animal use in this way?
It’s a really transparent way of looking at science. In the past we had to say “what category might our animals fall into?” and apply for licences based on our estimates. Now we’re looking at every animal and saying “what has this animal actually experienced?”
So it’s really giving a representation of the level of suffering of laboratory animals and directly tying this to the scientific output of the experiments they’ve been used in — I think that’s really important because we can look at the cost of animal research, including to the animals, against its benefits.
Back to the report – what exactly is a ‘procedure’?
A procedure is a whole programme of work with a specific scientific endpoint. For example, if you were researching diabetes. you could breed a mouse with an alteration in a gene involved in glucose metabolism, maintain that mouse until it was an adult, take a blood sample and a urine sample, and measure the fat content of the animal using imaging equipment while it was under anaesthetic. These protocols all form part of a single procedure — where the scientific endpoint is finding out about biological processes involved.
The number and order of protocols in one procedure is tightly controlled within each project licence.
And are only animals used in scientific experiments counted?
All animals which undergo regulated procedures are included. However, breeding genetically altered animals is also a procedure. Therefore the numbers on the annual returns represent both experimentation and breeding where the animal is carrying a genetic change but has been used for breeding alone.
Until recently, one of the most common methods of creating a genetically altered (GA) animal – such as a mouse for example – was to make the genetic change in embryonic stem cells, inject these into an early-stage mouse embryo and then implant the embryo into a female mouse.
It then takes a few generations of breeding to establish a colony of mice. This mouse line might then be bred with mice carrying a different genetic alteration to assess the genetic impact of the combination. During these breeding steps it is unavoidable that mice are generated with the wrong combination of genes for the scientific studies. These mice also have to be included in the statistics although they won’t be used any further and will be culled.
‘Wild type’ mice – those without genetic alterations – which are only used in or produced by breeding and then humanely culled are not counted in the returns because only regulated procedures are counted.
And the procedures are graded by severity?
Yes, the definition of a regulated procedure is one that may cause an animal pain, suffering, distress or lasting harm, but within that definition procedures are classified as mild, moderate and severe. The severity is classified under the maximum an animal has reached in the whole of its lifetime — if an animal lives for a year and experiences some mild pain but also, say, 10 minutes of moderate pain, it is classed as moderate severity.
This year the severity classification of the vast majority of animals used for breeding only is mild or under (94 per cent), whereas experimental procedures in animals have been categorised as higher severity: mild or below is 67 per cent, moderate 25 per cent, and 8 per cent of animals reached a severe classification.
Mild procedures include an animal experiencing short-term mild pain such as having an injection. A GA animal being born, never showing any effects of the altered gene they are carrying, never being used in an experiment and then being killed humanely would be counted as a mild procedure.
To avoid double-counting of GA animals, animals are only returned once – under their breeding if they’re not used in an experiment, and under the severity of the experiment if they are.
A moderate procedure involves an animal suffering short-term moderate pain, or longer-term mild pain. This includes procedures like surgery under general anaesthetic.
A severe procedure may cause severe pain, suffering or distress or long-term moderate pain, suffering or distress. This would include experiments where an animal suffers from a debilitating disease and gets sick very quickly, before there is the opportunity for intervening (eg sudden paralysis in stroke models). In these cases it is usual for the animal to be put down as soon as these symptoms are seen. Also, if an animal is found dead and one can’t be absolutely sure that its death wasn’t to do with the experimental conditions, it is classed as severe.
In the last decade, the numbers of procedures – a proxy for the numbers of animals used – has increased. Why is that?
The figures increase largely because research using GA animals is a growing area. With GA mice, for example, we are generating much more sophisticated models of disease with multiple genetic changes rather than just one. Breeding in all these combinations of genes together inevitably needs more mice that just one genetic change.
Most researchers are very conscious of their breeding schemes and only breed animals they can use. But there will always be animals that can’t be used – you can’t research the female reproductive system in male mice for example.
There’s also an element of fluctuation in the number of animals used due to things like strategic investments in particular areas of research, and changes in regulatory requirements.
This year, the numbers show a slight decrease, however we should be cautious about comparing to previous years as the method for collecting the figures has changed dramatically.
And what about long term? Aren’t there new techniques that don’t use as many animals, for example?
Gene editing with techniques such as CRISPR-Cas9 make genetically altering animals much easier and cheaper, and in themselves require fewer animals to generate a GA animal. This is because the technique involves injecting the genetic change of choice straight into an embryo and implanting it into a mouse. The number of offspring carrying the change is much higher than for traditional methods, so it’s much more efficient.
But even as these techniques are used more, that doesn’t mean using fewer animals. The general trend I expect to see is not necessarily an increase in the number of animals but an increase in the number of different varieties that are generated. Because CRISPR-Cas9 is easier and cheaper, it’s likely that researchers will be able to study more mutations, especially those found in human populations, than they would have done previously.
This will lead to better science because the model animals will much more closely reflect the human diseases. As well as creating a mouse with a missing protein to discover the function of that protein, if you think a particular point mutation causes a change to the protein, you can now make the point mutation itself in the animal. So it’s a great system that’s much more reflective of human genetic disorders.
Remember that we’re in a sequencing age – as we learn more about the genetics of human disease from sequencing projects, we’re going to want to generate many more models of the genes involved.