Group 3: Ethics 1 - Animals in Experiments

A Summary of the Main Arguments

Here we summarise some of the common arguments against animal research as identified by a House of Lords report (The House of Lords, 2002) into the use of animals in scientific research.

Beyond the reasoning that animal research is unethical because it violates the principle of not harming other beings (as discussed in the sections on pain, and in the discussion on whether animals have moral standing) unless for their own benefit there is the simple matter that animal research is ineffective.

This is perhaps the biggest argument for moving away from animal research today. Since diseases function differently between different animals and humans it is hard to translate animal research into a clinical benefit for us. New drugs may be beneficial to some animals but ineffective or even dangerous when used on humans. For example, a drug called Rofecoxib (branded as Vioxx), an anti-inflammatory was deemed safe and effective when tested in monkeys but was later shown to cause heart attacks and strokes in humans (Jüni, et al., 2004). The inverse is also true, drugs that are harmful to animals can be beneficial to humans. For instance, the popular painkillers Aspirin, Paracetamol and Ibuprofen are toxic to cats and dogs (Villar, et al., 1998) but not to humans. Indeed, Aspirin causes birth defects, it is a teratogen, in animals but not humans (Kimmel, et al., 1971). This has led to potentially beneficial drugs to be discarded because they produced negative results in animals. There are differences even amongst (non-human) animals. The most commonly used species of monkey used in drug trials (the crab-eating macaque) is resistant to doses of Paracetamol that would kill both dogs, cats and humans (Tamai, et al., 2017).

As well as being ineffective some animal research is a complete waste, of time, money, and lives. The animal testing for Thalidomide, a drug used against nausea and to alleviate morning sickness in pregnant women which caused thousands of babies to be born with birth defects did not even include tests on pregnant animals and so this defect in the drug was missed (Hendrickx, et al., 1966). That animal research is a waste of money in many cases, makes it harmful to human health since it takes money away from non-animal research, as is often pointed out by HIV researchers where the breakthroughs have often occurred due to in vitro studies (Greek, 2002).

Another example of wasted research is the fact that much animal research is trivial, simply used to validate the large number of “me too” drugs being pushed out by drug companies. These are versions of drugs already on the market which are made simply to prolong the intellectual property rights of drug companies and make them more profits (Goldacre, 2014). If animals are to be used at all there is a case that they should only be used for drugs which are medically necessary, not to benefit drug manufacturers. Even when not performed on “me too” drugs, some of the research is of little relevance, with some animal research never being cited in the literature. There is also the possibility of duplicate testing, since negative findings are not likely to be published (so-called publication bias) many animal experiments are repeated even though the data has already been collected.

Part of the reason why animal research is ineffective is that animal models often fail to capture the way a disease behaves in humans. Since animals simply do not get many of the diseases we do, like heart disease, some cancers, HIV, Parkinson’s and schizophrenia to name but a few, they have to be artificially induced in animals where they just don’t always mimic the disease exactly. For example, MPTP a drug used to produce nearly all the effects of Parkinson’s by destroying dopamine neurons. But, it behaves differently depending on the animal. Whilst it mimics Parkinson’s disease in monkeys, administering it to rats does not produce a valid model since rodents seem immune to its effects (Emborg, 2007).

Then there the matter of drugs trials being carried out later, or even concurrently, in humans. This raises the question, if we are going to do human studies anyway why do them in animals first? Safety is of course a factor, but new techniques such as micro dosing which is discussed later offer a way around this. Somewhat more cynically it could be suggested that the sole purpose of the animal tests is to give legal protection to drug companies if their products are later shown to be toxic.

One of the main reasons that animal research continues is through habit, scientists becoming accustomed to particular ways of working. This is not helped by a regulatory environment that requires animal testing as a box ticking exercise in order to get approval to sell new drugs. Nor is it helped by the lack of funding into alternatives by drug companies and research councils. It has been argued that had the same amount of money and effort been invested in non-animal methods as animal methods then similarly effective drugs would have been discovered. This is changing now with the advent of the “three R’s” (Russel & Burch, 1959) policy of replacement, reduction and refinement which encourages the improvement of non-animal methods.

In an ideal world we would use the most scientifically valid method for our toxicology assessments, which would be to test directly in humans. Currently, this would involve perhaps too high a level or risk and would be unlikely to result in a sufficient number of volunteers. The rise of new technologies however could lead to this being the approach used in the future. However, before we start looking at alternatives to animal experimentation it is work discussing if animals are even worthy of special treatment in the first place that would warrant their replacement, in other words do they have "moral standing".

Consequently, an important question in deciding whether or not animals should be used in experiments, and thereby suffer “harm”, is whether they have a moral standing to begin with. A typical argument for the use of animal experiments reasons as follows:

1. Persons have moral standing.
2. Normal human beings have moral standing.
3. Other beings may nor may not be persons.

We then pose the question, “what is it to be a person?” to-wit the answer usually involves rationality, or self-consciousness, or some other uniquely (as far as we’re aware) human characteristic. On the basis that nonhuman animals do not display the same characteristics as we do, it is quickly concluded that we and the other animals are different to people in some way. That nonhuman animals have a lesser status than we do. Because of this, whilst humans enjoy full moral rights nonhuman animals do not.

This argument can be countered on the basis that, none of these human-like characteristics are of any importance to determining what moral rights something has. Arguably what is more important is for the thing to be sentient, that is capable of feeling pain and suffering. If something is capable of experiencing harm then surely it has a right not to be subjected needlessly to such harm? There is of course another “gotcha” here, when we say needlessly we imply that there are some situations where the harm is justified. This is true, there are certainly some situations today in which we make such judgements. The standout example being the use of euthanasia on severely sick animals, where death is used as a means to prevent any further harm. The point is not to argue whether all suffering of animals is wrong, it is instead to argue that the benefits to humans should not stand as a reason for harming animals.

Ultimately the similarities between humans and nonhumans, particularly for higher mammals means they should be treated similarly (Rachels, 1999).

A somewhat related line of argument for treating animals as we treat ourselves comes from the founder of the utilitarian philosophy, Jeremy Bentham. He argues (Bentham, 1876) that “a full grown horse or dog” is a more rational animal than an infant “a day, a week, or even a month old”. Yet we would take a moral issue with experimenting on infants. This he says shows that it is not “reason, nor […] speech” that should be used to determine if it is moral to experiment on something, but rather the question, “can [it] suffer?”. Consequently, all creatures capable of suffering are equal to humans and should be treated as such.

The question then becomes which animals are capable of suffering? This is discussed in the section on pain.

Since there is no argument that these methods could inflict pain, the only possible justification is that the animals do not suffer, or if they do their suffering is necessary.

The problem with justifying animal experiments on the basis that the animals do not suffer because they’re incapable of feeling pain is that pain is a personal experience. We cannot measure pain directly in other human beings, yet alone other animals. As mentioned already, the degree to which it is “right” to use an animal in experiments is often linked to the ability of the animal to suffer. This is called the argument-by-analogy approach. This states that if an animal responds to a stimulus in the same way as we do they it is likely to have had an analogous experience.

This allows us to distinguish between the responses of different animals. For example, even Protozoa (eukaryotic microorganisms), the simplest of invertebrates have a “pain response” in that, where present, they move their hair-like organelles called cilia faster if subjected to a harmful stimulus, like being poked with a needle (Naitoh & Eckert, 1974). And they do this without having a nervous system.

You would probably say that Protozoa can’t feel pain, requiring perhaps that the response is more like the human pain response which often persists beyond a simple reflex action (Morton & Griffiths, 1985). For example, many mammals also become immobile on experiencing pain, become aggressive and stop feeding or drinking, among other things.

This “moving of the goal posts” is a dangerous approach, and it has lead us to inflict suffering upon animals which, after further research have been considered to be more capable of feeling pain than we originally thought. The stand out example here being the Cephalopods, the group of marine animals including the octopus and squid. These animals were not added to EU animal protection legislation until 2010 (European Parliament and Council, 2010). We must recognise that our knowledge of other animals’ physiology and functioning is changing all the time with the rise of new technologies. For example, it wasn’t until rapid genetic screening could be performed that the fruit fly gene for nociception (named painless, due to mutants being unable to feel ‘pain’) was discovered (Tracey, et al., 2003). We cannot go back and undo the suffering we have caused to these animals after the fact. This at the very least suggests we should be more cautious in choosing which animals to experiment on, if indeed we must experiment on them at all. We certainly should be collecting more data on this front.

Indeed, to suggest that feeling pain is dependent in some sense on self-awareness simply because that is how we suffer does not mean that the same has to be true of other species (Smith, 1991). They could suffer in their own way and it would be just as bad as how we perceive our own suffering.

The Purpose of Pain

It seems that by default, the majority of people link capacity to feel pain with reasoning ability, or degree of consciousness. They would say that there is a positive correlation between mental ability and ability to feel pain. As evidence by our willingness to use other animals over non-human primates for example, or to favour invertebrates rather than vertebrates. Even though as mentioned above there is increasing evidence that these animals too are capable of experiencing some sort of pain.

Dawkins (2011) argues that there is no reason why the correlation must be positive, he suggests that pain is a primal feeling that one shouldn’t require intellect to experience. He goes on to suggest that the correlation may even me negative. The argument being that the purpose of pain, at least from an evolutionary perspective, is to prevent animals from repeating actions that lead to bodily harm. By extension he reasons, is it not then plausible that a species such as ourselves might need less to experience less pain? Since, we are capable of reasoning we should be able to figure out what is good or bad for us whereas a less intelligent species might need more pain, to reinforce a lesson that we can learn without the need for such a powerful motivation.

There is no general reason to believe that non-human animals feel pain any less than we do, and consequently we should treat the harm we inflict on animals as the equivalent, morally, of doing the same thing to another human being. This would undoubtable render many of the experiments we perform on animals as morally wrong, particularly when as we discuss in the section on humans in phase I trials there is often just as little justification for using humans as there are animals.

Even when animal studies are conducted “well” the results will not necessarily translate to clinical practice because the models lack external validity. Many animal constructs have both poor construct validity and predictive validity.Whilst transgenic mice can be engineered to exhibit some of the symptoms of Alzhiemer's disease the changes neither completely nor exclusively reflect what happens to humans when they get Alzheimer's disease (Langley, 2014). We can not assume that things like gene function are conserved between animals and humans, despite the fact that 99% of mouse genes have a corresponding gene in the human genome (Mouse Sequencing Consortium, 2000), until functional equivalence has actually been shown (Lynch, 2009).

A model has construct validity if the correctly reflects the underlying physiological processes in humans. It is clear that in many cases animal models are not construct valid. Predictive validity refers to whether a model accurately predicts what will happen in humans, since 92% of drugs fail clinical trials, mostly because of unpredicted effects in humans it could be suggested that animal models are not construct valid either.

A large part of this are the differences between animals and human disease pathophysiology but other reasons include:

• Differences in comorbidities between animals and humans.
• The use of co-medication in humans.
• Animal studies usually consist of animals which are young and healthy whereas, in patients, disease mainly occurs in elderly people with co-morbidities.
• Animal studies are performed in rather homogenous populations (e.g. lab rats) whereas groups of patients are far more heterogenous.
• Animal studies often use only either male or female animals but the disease affects both male and female patients.
• The doses given to the animals are toxic or would not be tolerated by patients.
• Differences in outcome measures and the timing of outcome assessment between animal studies and clinical trials.
• Animal studies use short onset-to-treatment times that are unobtainable clinically.

Many animal models have not been evaluated systematically for reproducibiliy, specificity or sensitivity (Hartung, 2008) yet costly clinical trials, risking participants, are conducted on the basis of such models.

We mentioned previously that some potentially beneficial drugs are discarded because they produce a negative result in animal studies. A similar risk is that animal research leads to the development of many of the same compounds (Markou, et al., 2009), the so-called "me too" compounds. Because we are always testing functionality in the same way it makes sense that most of the drugs would also function the same way.

The widespread perception that animal studies are useful may perhaps be driven by the fact that neutral or negative animal studies are more likely to remain unpublished. Even this assumption is starting to be questioned by many researchers who feel that the value of animal experiments is limited, and that the only reason to carry out animal research is because they are legally compelled to do so (Watts, 2007).

We are also seeing improvements to basic biological techniques, transforming them from complex, time consuming experiments into high throughput operations that allow for massive data collection whilst remaining precise. For example, quantitative proteomics has been used to identify and quantify proteins involved in neurodegenerative diseases.

There have also been improvements to in vivo cellular models. Previously cell models often consisted of cancer cell lines in monolayer cultures (Langley, 2014), which as rightly pointed out by proponents of animal experimentation fail to replicate the complex architecture and cellular interactions that take place in a complete living organism. New tissue engineering approaches allow us to create controllable 3D models of human tissue which can replicate the spatial structure of real tissue. These models have better viability than traditional cultures and many of the tissue processes more closely match what occurs in vivo (Langley, 2014). Taking this technology further, we are now starting to see “organ on a chip” devices which combine microfluidics with 3D culture and can reproduce the structural and biochemical features of entire organs in vitro. Lung and gut models have been developed and brain-on-a-chip systems are currently in development. See, for example, Park, et al. (2015).

Another recent advance, in the form of microdosing, uses humans for in vivo studies, rather than merely human cells. A microdose is a very small dose of a compound, usually below 100µg, which is too small to create either a pharmacological effect or an adverse reaction in the person to which it is administered (Rowland, 2010). Microdosing has been made possible by new analytical chemistry methods that make it possible to detect molecular concentrations as small as picograms per millilitre in blood. Even newer methods, like accelerator mass spectrometry, allow the detection of individual molecules tagged with radioactive carbon-14. In the words of Malcolm Roland (2010), this technology has “the ability to detect a liquid compound even after one litre of it has been dissolved in the entire oceans of the world”.

Unfortunately, it is not thought that microdosing will be able to replace animal studies, since it means using humans for untested medicines when people want to know that a compound is likely to be safe before administering it. This means some sort of animal testing will always be necessary. However, it does mean that the unnecessary exposure to of animals to high doses of unwanted compounds will be reduced, since the same data can be collected with microdosing as a traditional pharmacokinetics study.

Additionally, by being able to collect more pharmacokinetic data from animals in a non-invasive way microdosing will allow us to compare animal and human pharmacokinetics. This will allow us to make a better assessment of which animals we use to predict human pharmacokinetics by identifying which species are similar to us from a drug action perspective. This is important because currently we tend to use larger animals, such as the non-human primates based on the principal of allometry (Rowland, 2010). This is the assumption that the difference between animals and humans is mainly due to size. As we become able to make better predictions with smaller animals we will need to use higher order animals less. The advantage being, that as discussed earlier these animals suffer in a way closer to how we suffer so it is better to eliminate their use before the use of other animals. This is a common approach in animal experiments, to work up the scale tree, for example neural tissue transplant experiments were first done in rodents to show that the cells survived before progressing to monkeys (Redmond, 2012), by reducing the scale we need we can ultimately reduce the number of animal experiments we have to perform. Although, as we have seen in the case of MPTP, rodent studies do not necessarily produce what will happen in humans or monkeys, so some technology improvements are still needed here.

Emborg, M. E., 2007. Nonhuman Primate Models of Parkinson's Disease. ILAR Journal, 48(4), pp. 339-355.

European Parliament and Council, 2010. Directive 2010/63/EU of the European Parliament and of the Council on the protection of animals used for scientific purposes.. Official Journal of the European Union, pp. 276/33-276/79.

Goldacre, B., 2014. Bad pharma: how drug companies mislead doctors and harm patients. s.l.:Macmillan.

Greek, R., 2002. Animal studies and HIV research. BMJ, Volume 324, p. 236.

Hackam, D. & Redelmeier, D., 2006. Translation of research evidence from animals to humans.. JAMA, 296(14), pp. 1731-2.

Hartung, T., 2008. Thoughts on the limitations of animal models. Parkinsonism & related disorders, Volume 14, pp. S81-S83.

Hendrickx, A. G., Axelrod, L. R. & Clayborn, L. D., 1966. 'Thalidomide’ syndrome in baboons.. Nature, 210(5039), pp. 958-959.

Jüni, P. et al., 2004. Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. The Lancet, 364(9450), pp. 2021-2029.

Kimmel, C. A., Wilson, J. G. & Schumacher, H. J., 1971. Studies on metabolism and identification of the causative agent in aspirin teratogenesis in rats.. Teratology, 4(1), pp. 15-24.

Langley, G. R., 2014. Considering a new paradigm for Alzheimer's disease research. Drug Discovery Today, 19(8), pp. 1114-1124.

Lynch, V. J., 2009. Use with caution: Developmental systems divergence and potential pitfalls of animal models.. The Yale journal of biology and medicine, 82(2), p. 53.

Markou, A. et al., 2009. Removing Obstacles in Neuroscience Drug Discovery: The Future Path for Animal Models. Neuropsychopharmacology, Volume 34, pp. 74-89.

Morton, D. B. & Griffiths, P. H. M., 1985. Guidelines on the recognition of pain, distress and discomfort in experimental animals and an hypothesis for assessment.. Vet Rec, 116(16), pp. 431-436.

Mouse Sequencing Consortium, 2000. Why Mouse Matters. [Online]
Available at: https://www.genome.gov/10001345/importance-of-mouse-genome/
[Accessed 3 7 2017].

Naitoh, Y. & Eckert, R., 1974. The control of ciliary activity in protozoa.. Cilia and flagella, Volume 12, pp. 305-352.

Park, J. et al., 2015. Three-dimensional brain-on-a-chip with an interstitial level of flow and its application as an in vitro model of Alzheimer's disease. Lab on a Chip, 15(1), pp. 141-150.

Rachels, J., 1999. Do Animals Have Moral Standing?. [Online]
Available at: http://www.jamesrachels.org/MoralStanding.pdf
[Accessed 22 06 2017].

Redmond, E., 2012. Using Monkeys to Understand and Cure Parkinson Disease. The Hastings Centre Report, 42(6).

Rowland, M., 2010. Interview: Interview with Professor Malcolm Rowland.. Bioanalysis, 2(3), pp. 385-91.

Russel, W. & Burch, R., 1959. The Principles of Humane Experimental Technique. London: Methuen.

Smith, J. A., 1991. A Question of Pain in Invertebrates. ILAR Journal, 33(1-2), pp. 25-31.

Tamai, S. et al., 2017. A monkey model of acetaminophen-induced hepatotoxicity; phenotypic similarity to human.. The Journal of toxicological sciences, 42(1), pp. 73-84.

The House of Lords, 2002. SELECT COMMITTEE ON ANIMALS IN SCIENTIFIC PROCEDURES, London: The Stationary Office.

The Royal Society Animals in Research Comittee, 2004. The use of non-human animals in research: a guide for scientists, London: The Royal Society.

Tracey, D. W., Wilson, R. I., Laurent, G. & Benzen, S., 2003. painless, a Drosophila Gene Essential for Nociception. Cell, 113(2), pp. 261-273.

Trevan, J. W., 1927. The error of determination of toxicity.. Proceedings of the Royal Society of London. Series B, Containing Papers of a Biological Character, 101(712), pp. 483-514.

van der Worp, H. B. et al., 2017. Can Animal Models of Disease Reliably Inform Human Studies?. PLoS Medicine, 7(3), p. e1000245.

Villar, D., Buck, W. & Gonzalez, J., 1998. Ibuprofen, aspirin and acetaminophen toxicosis and treatment in dogs and cats.. Veterinary and Human Toxicology, 40(3), pp. 156-162.

Watts, G., 2007. Alternatives to animal experimentation. British Medical Journal, 334(7586), pp. 182-184.