Chapter 4
Cosmetics testing on animals – a scientific critique
print

This chapter describes how and why each animal test is undertaken and the various endpoints which are assessed in the evaluation of cosmetic ingredients. Major scientific criticisms are provided and the tests’ effects on the well-being of animals are described. Many animal tests are unvalidated and there is little evidence of their reliability, particularly in predicting long-term adverse outcomes in humans, such as cancer or reproductive effects.
There are many instances of species differences in response to test chemicals – even members of the same species may respond in a number of ways that depends upon their health, breed, gender and age. This variation makes it difficult to extrapolate to humans, and these issues are explored in this chapter for each toxicity test.
It has been stated, in connection with animal tests of potential human chemical hazard exposure that “the main conclusions to be drawn from this discussion are that the use of animals in toxicological studies does not provide a reliable basis for extrapolation to human health” (12).

The animal tests

1. Acute toxicity

Acute toxicity testing is undertaken in order to determine the possible hazard of a single exposure to a chemical or product through the oral, skin or respiratory routes. The assessment of the potential to kill (lethality), which has traditionally been part of acute systemic (the whole body) toxicity tests, has relied on the measurement of that dose of a test chemical which kills half of those animals tested – this is the LD50 test. The LD50 value has been used for decades to classify chemicals on the basis of acute toxic hazard, and in addition to define doses for other in vivo tests.

During acute toxicity testing it is also possible to use endpoints which do not involve the death of the animal. Here, target organs such as the liver, kidneys or respiratory system, can be tested for the specific toxicities of chemicals.

The oral LD50 test method [OECD test guideline 401, EC guideline B1] which used three to five dose groups of animals, each comprising up to ten individuals, was deleted in 2001 and replaced by the following refinement and reduction methods:

a. The fixed dose method [OECD 420, EC B1 bis] no longer uses death as the endpoint and reduces levels of pain and distress in the smaller number of animals used. It is therefore a refinement of the LD50 test.

b. The acute toxic class method [OECD 423, EC B1 tris] does not involve a precise LD50 value but rather seeks to find a range of dosages of a test chemical where death is expected. The test involves a stepwise series of doses and takes more time than the original LD50, or the fixed dose method. It uses smaller numbers of animals but stress and pain will still be experienced by those individuals undergoing the test.

c. The up-and-down procedure [OECD 425] permits an estimation of the LD50 value, confidence intervals and the observation of toxic effects. This method reduces the number of animals used in comparison with the original LD50 test, but again still produces adverse and harmful effects and death in animals undergoing the test.

The LD50 test is still used for assessment of chemical toxicity using the skin application or inhalation routes (as opposed to oral dosing), also relevant to the testing of cosmetics ingredients.

Rats are usually used, although rabbits are also subjected to acute skin toxicity tests. One dose is applied per group, and at least five animals per test and control (who do not receive the test substance but instead a harmless material) groups. Fourteen days of observation is the norm and animals are fasted prior to the oral ingestion of the test substance. All animals are autopsied at the end of the test period in order to find if there has been any damage caused by the test. Any sex-specific response is noted. In inhalation experiments, 10 animals (five from each sex) are used for each concentration of the test substance.

Key criticisms:

– the immune, physiological, genetic, sex and other health indexes are factors influencing the validity of the test results. In addition, exposure to other chemicals encountered by the animals in the laboratory introduces variability in the outcome of the test.
– there are several significant species differences in the role of detoxifying organs and sites within the body, where toxic chemicals accumulate. Species show differences in the types and amounts of P450 enzymes (in the liver). The kidney also removes substances from the body in different ways and at different rates in test animals, compared to humans (13).
– the time course of the clearance and accumulation of any given substance varies both between species, and from one breed to another of the same species.
The environmental conditions in which animals are kept are often poor. Anxious and stressed animals make unreliable test subjects.

2. Skin absorption

Skin absorption is defined as the passage through the skin into the blood
circulation of a substance which has been applied to the skin; for example, an ingredient of a moisturiser. The way in which chemicals pass through the skin depends upon a number of factors, such as the size and chemical characteristics of the ingredient of the cosmetic being assessed, as well as the anatomical structure of the skin.
The SCCNFP in principle accepts any scientifically validated method, including human studies and in vitro skin methods as well as in vivo tests, to assess the skin absorption of a new cosmetic ingredient. All novel cosmetic ingredients submitted to the SCCNFP for approval and subsequent inclusion into the Annexes of the Directive 76/768/EEC must have skin absorption data relevant to humans.

Many species of animal have been used in the past to provide information about the passage of chemicals through the skin and their subsequent distribution throughout the body. Rats are most often used (14). After administration of the test material, the rats are killed and the amount of test substance absorbed is estimated. However, now that the OECD has accepted a test-tube method (see Chapter 5), the rat experiment should no longer be conducted within the EU.

Key criticisms:

– in a number of significant ways, rat skin differs from human skin including variations in skin and hair structure. Tests with rats invariably over-estimate skin penetration in humans.

3. Skin corrosivity and irritancy

Skin corrosion tests on animals have been used for many years to assess the potential of a substance to cause deep and irreversible damage to the skin. This damage involves visible death of tissues (necrosis) through the various layers of the skin, down to the dermis and even the muscle layers. Signs of corrosion can occur after up to four hours of test substance application. Such corrosive damage is not expected with cosmetic products, but some substances which are corrosive – such as alkalis –are used in cosmetics, especially soaps and cleansers. Their corrosive potential depends upon their final concentration, other ‘neutralising’ chemical ingredients, exposure routes and similar factors.
Animal tests of corrosivity cause high levels of pain, suffering and distress and are no longer permitted in Europe, as OECD-approved in vitro methods are available. In vitro corrosivity tests are now in both EU and OECD test guidelines [B40, TG430 and TG431 respectively].
Skin irritation tests were developed to determine the potential of substances to cause reversible skin damage such as swelling and inflammation, usually after a single application.
Albino rabbits are most often used. A single dose is applied to an area where the fur is removed, and a shaved but untreated region close by is used as a control. The irritancy is scored by checking treated skin against the appearance of untreated skin.
Up to three rabbits are used for each test substance, and exposure is usually for up to four hours.

Key criticisms:

– animals differ in their immune, physiological and genetic status, causing variations in response to potentially damaging chemicals. This makes it very difficult to predict how the same substance might affect humans.
– the anatomy and structure of the skin vary between test species and humans, so a simple extrapolation to likely human responses is rather dubious.
– animals tend to show different responses to the same test chemical depending upon their age.
– the rabbit, commonly used for irritancy tests, is a notoriously poor predictor of human skin irritation (15).
Substances which elicit irritation will cause varying amounts of pain, discomfort and anxiety in those animals tested.

4. Eye irritancy

Eye irritation tests have been developed to measure the potential of substances to cause redness, swelling and distension in the eye. Signs include discharge, swelling of the iris and lid, or ulceration or opacity of the cornea after a single test application.
OECD guidelines now specify prior non-animal screening of cosmetic ingredients, such as those known to be strongly alkaline or acidic, before any animal eye irritation tests are conducted. When such tests using animals are performed, adult albino rabbits are used, up to three adults for each test.
Single dose effects are monitored for up to 21 days. The other, non-tested eye serves as a control. In most cases no anaesthetic is used, and is only resorted to when it is thought that considerable pain may ensue. An illustrated standard guide is used to score levels of irritancy. Clearly pain, discomfort and stress are still part of the response to damaging cosmetic ingredients.

Key criticisms:

– there are marked differences in the structure of the cornea of the eye of rabbits and humans. Rabbit corneal mean thickness is 0.37 mm as against around 0.51 mm in humans. Also the rabbit cornea comprises 25 per cent of the surface area of the eye in comparison to only 7 per cent in humans.
– rabbits produce a smaller volume of tears than do humans. This means that any material placed in the eye will remain longer in rabbits, and hence possibly have a more marked effect than may be found in humans.
– the ranking and scoring of any damage which results from the administration of test substances is highly subjective, significant variation being found both within and across laboratories. Dr D Swanston of Porton Down, a major toxicological research establishment in the UK, has said: “…no single animal species has been found to model exactly for the human eye either in anatomical terms or in response to irritation”(16) .

5. Skin sensitisation (allergic reaction) and photosensitisation

The sensitisation test is intended to assess the ability of a test substance to elicit an allergic reaction in the skin – allergic contact dermatitis.
For most chemicals, a less severe method than the traditional guinea pig test has recently been accepted by the EU and the OECD. This is the local lymph node assay (LLNA), a refinement and reduction method using mice to assess sensitisation [OECD 429]. The method is based on measurements of specific responses (cell proliferation) induced in the lymph nodes which drain an area where a test chemical is applied (17). Chemicals which sensitise the skin lead to a marked increase in the activity of adjacent lymph nodes.
Mice are used and the assay measures the triggering of an immune response in the animal to three daily applications of a test chemical to the surface of the ear. Standard mouse strains are used and three different doses of the chemical are applied. Mice are killed after five days and their lymph nodes and blood cells examined. The LLNA method uses fewer animals and is likely to cause less suffering than the former widely-used methods, described below.
Until recently, and still for some chemicals, the traditional tests for skin sensitisation used guinea pigs, there being two common methods [OECD 406, EC B6].
a. The Magnusson Kligman Guinea Pig Maximisation test (GPMT), which uses an adjuvant – a substance that boosts the immune reaction but also causes more pain.
b. The Buehler test, a non-adjuvant (and less sensitive) test.
Albino guinea pigs are most often used in these two tests. The substance is delivered to the shaved skin of the shoulder by either surface application or by injection. A minimum of 10 animals are used in each test group and at least five in the control group. Further testing is carried out in larger group sizes (up to 20) if the initial response is equivocal. Multiple doses are applied in order to evoke local allergic reaction.

Key criticisms:

– the guinea pig tests are highly subjective and have poor reproducibility.
– the test material is applied to an area of shaved skin, which is not the normal way in which humans use a range of cosmetics (except shaving products like aftershave and various balms).
– the structure of the animals’ skin (mouse or guinea pig) differs from that of humans.
– the immune, physiological and genetic characteristics of the test animals influence sensitisation reactions.
– large doses of a test chemical, especially when injected, do not mimic the ways in which allergies are triggered in humans.
– there is little evidence that potency of a chemical effect in the guinea pig predicts potency effects in the human (18).

6. Subchronic toxicity

Chronic toxicity results from persistent or progressive administration of a substance which results in damage to cells, tissues or organs. Of particular relevance to cosmetic ingredients are the oral and skin routes in subacute (28 days) and subchronic (90 days) repeat-dose studies.
The 28- and 90-day oral toxicity tests in rodents are the most commonly used long-term tests for whole-body toxicity. The highest dose administered is intended to cause some toxicity, with possible pain and suffering but not death. Once the test is completed the animals are killed and examined for organ damage and other unwanted effects.
In the development of cosmetic ingredients having specific biological effects and which come into contact with human skin for long periods of time, the whole-body effect is a major focus for toxicity tests.
In 28-day tests, the most commonly used species is the rat, although for repeat-dose skin tests guinea pigs, rats or rabbits are used. Ten animals receive each dose, plus 10 animals in the control group. The test chemical is administered daily via the skin, by inhalation or by mouth and, after killing, the animals are examined both pathologically and biochemically.
The 90-day test also uses rats. Forty animals comprise each test and control group. Three doses of test chemical are used in each group. The main routes of administration are oral and inhalation.

Key criticisms:

– different species absorb, metabolise and excrete chemicals in varying ways and at different rates, making extrapolation from rodents to humans problematic.
– there are species differences in pharmacological and biochemical responsiveness, and even discrepancies between different breeds of rats (19).
– the physiological, immune and dietary status of the animals used is likely to affect the interpretation of the test outcome.
– scaling-up from small animals with short lifespans, such as rats, to larger, long-lived humans, is always difficult. It is often attempted by calculations relating dose to body size, but these are only ‘guesstimates’.
– the ways in which exposure normally occurs is crucial in deciding the long-term toxic dose. For example, humans rarely receive repeated, same-size doses of a cosmetic ingredient orally over a long period of time, and yet many animal tests are conducted in this way.
– in long-term studies there can be problems in defining the relevant endpoints and how should they be evaluated (20).
The impact on test animals includes many of the effects listed under acute toxicity – these include pain and distress, low-level discomfort and feelings of nausea, agitation and other possibly subtle clinical signs. Death may occur; and can be preceded by fits, behavioural changes, vomiting, discharges from mouth or anus, or loss of consciousness.

7. Mutagenicity and genotoxicity

Mutagens are substances which increase the rate of genetic change (mutations) in the genes or chromosomes. Mutations occur spontaneously within any population, and mutagens increase the number of such alterations. Mutations are, in some instances, the first steps in the formation of a cancer.
Genotoxicity is a wider term which refers to the ability of a chemical to interact in various ways with DNA and/or other parts of the cell nucleus.
Several in vitro genotoxicity tests are available. The SCCNFP regards the combination of two in vitro tests to be, in general, sufficient to discover mutagenic and/or genotoxic potential of a specific chemical. Depending upon the results of such tests, SCCNFP may require other in vitro or in vivo tests to also be undertaken.
Animal tests for mutagenicity/genotoxicity:
In principle, tests for mutagenicity/genotoxicity involve an indirect assessment of toxic effects on the blood cells and their genes, within the bone marrow. Two techniques are commonly used:
a. The in vivo mammalian bone marrow cytogenetic test
This test assumes that actively dividing blood cells in the bone marrow are especially sensitive to the effects of mutagens/genotoxins. Test animals are usually rodents such as rats, mice or Chinese hamsters. Ten animals (five of each sex) are used and have the test material administered either by mouth or by injection into the abdominal cavity. Two control groups are used for each test substance and dose regime. Single or multiple doses are used and bone marrow sampling is performed up to 48 hours after dosing. Preparations of the blood cells are examined with a microscope in order to find effects on the cell nucleus.
b. The micronucleus test
In this test for mutagenic/genotoxic effects, the overall technique and sample sizes are similar to [a] above, but the effect of the chemical is estimated by looking at a different aspect of the blood cell. Increasing numbers of cells showing damage to the nucleus (seen as the formation of micronuclei) indicate a gene-damaging chemical. Mice are usually used.

Key criticisms:

– many general criticisms of toxicity testing are relevant to these tests, especially the role of the immune system and the genetic susceptibility of the species and strains of animals.
– it is often unclear whether the test chemical actually reaches the bone marrow occurs – thus false negatives are likely (21).
– because of possible difficulties of penetration of the test material to the
bone marrow, large doses are used. This makes the relevance to humans dubious,
since humans are not often exposed to large doses of cosmetic ingredients.
– injection of test substances into the abdominal cavity is not an appropriate
route of administration for cosmetic ingredients.
– the tests are limited to a few tissues and a narrow range of endpoints (22).
Many of the criticisms of the tests for carcinogenicity (see [13], below) also apply to assessing mutagenic/genotoxic effects in animals. Repair enzymes and the role of ‘scavenging’ molecules in animals and humans are very different and this will have an impact.

8. Photoirritancy and phototoxicity

There is no validated animal tests for this endpoint. The in vitro method is the method of choice: the 3T3 neutral red uptake phototoxicity test (see Chapter 5).

9. Photomutagenicity and photogenotoxicity

A number of in vitro methods are available to assess light-induced mutagenic and genetic effects. These include: bacterial and yeast mutation assays, tests for detecting clastogenicity or tests for gene mutations in in vitro mammalian cell systems.

10. Human data

Ethical studies on fully-informed human volunteers are fairly common in product testing, but less so in cosmetic ingredient testing. Examples include skin irritation or skin sensitisation tests involving human volunteers. In such cases, earlier animal or in vitro data on toxicity may be used to ensure the safety of the volunteers.
Ethical and practical aspects of using human volunteers in safety evaluation studies are discussed in the SCCNFP guidance notes (23).

11. Toxicokinetics

Toxicokinetic tests are designed to follow the time course of chemical absorption, distribution, metabolism and excretion, and toxic effects. Doses are either single or multiple.
Rodents are often the species used. Each dose group comprises four animals, and routes of administration are oral, inhalation or via the skin. Distribution, excretion and metabolism time courses are followed. After killing, animals are examined for the accumulation of test substance in target organs at different times from the start of the administration.
Key criticisms:
– there are the common species- and strain-related difficulties of interpreting the relevance of animal data for humans.
– the elimination of toxic substances may follow different pathways in the humans than in rodents and other test species.
– similarly, the rates of detoxification and subsequent elimination are species- and strain-dependent and there are marked problems of trying to extrapolate to other species.

12. Metabolism

After absorption into the bloodstream the fate of a substance in the body is determined to a large extent by its metabolism, mainly in the liver. Some chemicals are inactivated by metabolism, while others may be metabolised to toxic compounds and either stored in various parts of the body or excreted. Only in specific cases does the SCCNFP require animal or in vitro studies in order to assess potential adverse metabolic effects.
Metabolism tests are undertaken in ways similar to those described for toxicokinetics (see [11] above), and the same criticisms of relevance and extrapolation also apply.

13. Teratogenicity, reproductive toxicity and carcinogenicity

Teratogenicity tests
A number of chemicals are known to influence the development of the fetus whilst in the womb, the end result being either bodily malformations or the death of the fetus. This process is known as teratogenesis.
The test for potential teratogens involves dosing pregnant female animals during the period of organ formation in the developing embryo. Attempts are made to assess whether the test substance causes malformations in the embryos.
Rats, mice, hamsters and rabbits have been used – the animals being most often rabbits and rats. Tests are normally conducted on 20 female rodents or 12 female rabbits per dose, with similar numbers in control groups.
Three dose levels are given, where the highest is sufficient to evoke minor changes in the mother (for example loss of weight). Test substances are usually given by mouth and the embryos are killed and examined for any evidence of toxicity-induced anatomical change.

Key criticisms:

– the tests are designed to identify gross effects of chemical administration; more subtle damage is likely to be overlooked.
– there are significant differences in the extent to which chemicals cross the placenta, depending on species.

– the test is costly in time and funds.
– problems of relevance arise, especially when inbred genetically-constant strains of animals are used – since humans are genetically mixed.
– test animals have much shorter lifespans in comparison with humans, causing difficulties in scaling up.

Reproductive toxicity tests
Reproductive toxicology tests seek to evaluate cosmetic ingredients which may have an effect upon the reproductive capacity. Doses are administered to male and female animals during their reproductive cycles (that is, sperm formation and absorption in the male and during two oestrous cycles in the female).
The oral route is most often used and doses are given daily, usually to groups of rats or mice. Assessment is made of effects on fertility, pregnancy and maternal behaviours (such as feeding and nest building).

Key criticisms:

– the reproductive systems and cycles of rodents and humans are very different, and the responses of organs such as the testes and ovaries to chemicals may vary in humans and other animal species.
– the influence of immune, physiological and dietary status on the results cause problems discussed in earlier sections.
– genetic constitution profoundly affects the reproductive toxicity of many
chemicals and this varies between humans and other animals.

Carcinogenicity tests
Tests use young rats and mice, with dosing starting very soon after weaning. The usual route is by mouth, but substances can be delivered by skin painting and by inhalation. Routes are chosen which attempt to mimic the ways in which humans might be exposed to the cosmetic ingredient. Three doses are usually tested with at least 100 animals for each dose, together with a further 100 animals being used as controls.
The outcome is assessed by blood sampling, body weight, pathological appearance and tissue and organ examination in order to detect cancers. These tests are expensive and take up to five years to complete.
Tumours in humans and animals normally occur as a result of a number of events, only one of which may be a single chemical exposure – this is not taken into account in the animal tests.

Key criticisms:

– animal carcinogenicity tests have very poor reproducibility, and there are species differences in results even between mice and rats.
– metabolic rates vary with the size of the animal – thus rats and mice are high metabolic rate species in comparison to humans (24) and their response to cancer-causing chemicals may differ from that of humans.
– there are many natural ‘scavenger’ chemicals present in animals – these act to clear away any potentially harmful molecules. The rates at which depletion occurs of one such ‘scavenger’ (glutathione) varies between different species, making comparisons difficult (25).
– mutagenesis (genetic damage) and carcinogenesis are known to be related to the effects of a highly active (26) form of oxygen which can damage DNA. Many animals used to assess carcinogenic and mutagenic potential have far less sophisticated DNA-repair mechanisms than are found in humans (27). It is not therefore surprising that small rodents are more prone to cancer than are humans (28), making extrapolation a matter of guesswork (29).
– variations exist between humans and rodents in the major drug-metabolising enzymes of the liver, such as the cytochrome P450s, which influences test outcomes. Furthermore, the kinds of P450s found in humans depend upon diet, genetic constitution, and smoking and alcohol consumption, as well as environmental exposure (30). Such complex effects cannot be modelled in laboratory animals.

OneVoice condemns the testing of cosmetic ingredients on animals because of the suffering caused. Moreover, the results of animal tests are often irrelevant to humans, because of species differences and the poor design of the tests. This means that the safety of consumers may be put at risk by some cosmetic ingredients which appear to be safe in animals such as rats, mice and rabbits.

 

12 - Millstone E (1989). Methods and practices of animal experimentation, in: Animal experimentation: the consensus changes, G Langley (Editor). Macmillan: London p. 72-87.
13 - Parke DV & Ioannides C (1990). Role of P450 in mouse liver tumor promotion. Progress in Clinical & Biological Research 331:215-230.
14 - Howes D et al (1996). Methods for assessing percutaneous absorption. The Report and recommendations of an ECVAM Workshop 13. ATLA 24: 81-106.
15 - Stitzel KA (2000). The use of clinical data to confirm the safety of consumer products for use on the skin. In: Progress in reduction, refinement and replacement of animal experimentation. Balls M, van Zeller A-M & M E Halder ME (Editors). Elsevier: Amsterdam. p. 587-600.
16 - Swanston DW (1991). The current status of non-animal alternative methods in skin and eye irritancy testing. In: Animals and alternatives in toxicology, Balls M et al (Editors). Macmillan: London. p. 23-51.
17 - Basketter DA et al (2002). Local lymph node assay - validations, conduct and use in practice. Food & Chemical Toxicology, 40:593-598.
18 - Basketter DA et al (1997). Skin sensitisation thresholds: determination in predictive models. Food & Chemical Toxicology 35:417-425.
19 - Monro A (1993). The paradoxical lack of interspecies correlation between plasma concentrations and chemical carcinogenicity. Regulatory Toxicology & Pharmacology 18:115-135.
20 - Fox JG et al (1979). Toxicology studies II: The laboratory animal. Food & Chemical Toxicology 17:661-675.
21 - Garner R C (1991). Genotoxicity testing. In: Animals and alternatives in toxicology. Balls M et al (Editors). Macmillan: London.
22 - ECVAM (2002). Alternative (non-animal) methods for chemicals testing: current status and future prospects. ATLA 30(suppl. 1):84-85.
23 - SCCNFP (2000). Notes of Guidance for Testing of Cosmetic Ingredients for their Safety Evaluation.
24 - Lewis DFV & Langley GR (1996). A validation study of the COMPACT and HazardExpert techniques with 40 chemicals. Mutation Research 369:157-174.
25 - Parke DV & Ioannides C (1990). Role of cytochromes P450 in mouse liver tumour production. In: Mouse liver carcinogenesis; mechanisms & species comparison Stevenson C et al (Editors). Alan Liss: New York. p. 215-230.
26 - Oberley LW & Oberley TD (1995). Reactive oxygen species in the aetiology of cancer. In: Drugs, diet and disease, Volume 1: Mechanistic approaches to cancer. C Ioannides and D F V Lewis (Editors). Ellis Horwood: London. p. 47-63.
27 - Hoeijmakers JHJ & Lehman AR (1994). Nucleotide excision-repair among species. In: Methods to assess DNA damage and repair: interspecies comparison, V Tardiff et al (Editors). John Wiley: Chichester. p. 57-82.
28 - Parke DV (1994). The cytochromes P450 and mechanisms of chemical carcinogenesis. Environmental Health Perspectives 102: 852-853.
29 - Parke DV (1990). ibid.
30 - Parke DV & Ioannides C (1990). ibid.