Cancer and Birth Defects
More on tests for cancer and birth defects
This is quite a long and detailed document. For a shorter version, go to:
For information on other kinds of safety tests, go to:
Toxicity tests, including LD50
Carcinogen tests
Chemicals that can cause cancer are called carcinogens. Tests for such substances involve lifetime feeding studies (usually 2 years) in rats and mice of both sexes. Some of the animals are fed the maximum tolerated dose (MTD) over their entire lifespan. This is the highest dose the animals can tolerate without dying (1), a dose which is much higher than humans would be exposed to. Another group is fed half the MTD, while another group is given no chemicals at all.
Teratology tests
Teratology is the study of defects which occur in foetuses during pregnancy. Rabbits and rats or mice are fed various doses of the test substance during the first part of their pregnancy, when organs of the foetuses are being formed. The day before the babies are due to be born they are surgically removed and examined for abnormalities in the skeleton or soft tissues (1).
Problems with carcinogen tests
First of all, these tests use large numbers of animals since equal numbers of males and females from 2 species (rats and mice) are included. Secondly, many animals develop painful tumours.
Animal tests are expensive and time consuming, which means that many chemicals can't be tested. According to the American Chemical Society's registry:
- there are 4 million known chemicals,
- the number is increasing by 6000 a week,
- about 63,000 chemicals are in commercial use,
- about 1000 new chemicals enter the market each year.
There are also scientific problems associated with animal tests. Over half the chemicals tested in rats and mice produce tumours. It is very unlikely that all these chemicals cause cancer in humans. In one study, 20 chemicals were tested on animals. According to the Agency for Research on Cancer, these chemicals don't cause cancer in humans. However, 19 of the 20 substances caused cancer in animals (3).
One reason for this inconsistency may be differences between species. Researchers have noted: " Some tumorigenic mechanisms in rats and mice are now known to be absent in humans, and this casts doubt on the relevance of rodents as models for carcinogenesis in humans" (4).
Another reason for the inconsistency is probably the very large doses (MTD) of chemicals fed to animals. Some researchers have concluded that chronic dosing at the MTD causes a high rate of cell division, and increased chances of mutations leading to cancer. The abnormal cell division doesn't occur at the low doses humans are likely to be exposed to, so the rodent tests don't show what will happen under normal circumstances to humans (5).
There isn't complete agreement in results even between the closely related rodent species of rats and mice. One review looked at all National Cancer Institute Bioassay Technology Reports that involved oral feeding in rats and mice of both sexes, a total of 171 studies.
For the chemicals that were carcinogens (caused cancer):
- 21% of these chemicals caused cancer in both sexes and both species,
- 32% caused cancer in one species only,
- 46% were not consistent across species or sex.
A number of studies have examined the percentage of chemicals for which results from rats are consistent, that is, either positive in both species or negative in both species. Here are some examples:
-
Results were either positive or negative for both species for 74% of 266 chemicals. That leaves 68 chemicals for which there was no agreement. (7)
-
Agreement was 76% for 392 chemicals. Of these chemicals, 76% that produced cancer in rats also did so in mice, while 70% of chemicals that produced cancer in mice also did so in rats. (8)
-
For 244 chemicals 71% were either positive in both species or negative in both species. Of the chemicals that caused cancer, only 46% did so in both species (9).
- For 181 pharmaceuticals submitted for registration in Germany and The Netherlands, results were positive or negative in both rats and mice in 70% of cases (10).
Rats and mice are more similar to each other than they are to humans, so the agreement between animal tests and human risk of cancer is likely to be less than the 70% in these studies. Some researchers have concluded:
" For almost all of the chemicals tested to date, rodent bioassays have not been cost-effective. They give limited and uncertain information on carcinogenicity, generally give no indication of mechanism of action, and require years to complete " (11).
Another problem concerns naturally occurring cancer in the control groups. In each study, animals that are fed chemicals are compared to a control group, that is, animals that are not fed chemicals. The trouble is that cancer is quite variable in these groups, and sometimes is quite common.
These were the rates of tumours in 30 groups of animals not given chemicals (1):
| male B6C3F1 mice | liver adenoma/carcinoma | 10%-68% |
| male F344 rats | pituitary gland adenoma/carcinoma | 12%-60% |
| female F344 rats | pituitary gland adenoma/carcinoma | 3%-76% |
| female B6C3F1 mice | pituitary gland adenoma/carcinoma | 0%-36% |
In other words, in some groups of the untreated male mice the rate of cancer was as low as 10% whereas in others it was as high as 68%. This variable and sometimes high rate of naturally occurring cancer makes it more difficult to explain the role of the test chemical in causing tumours, and to apply the findings to humans.
Alternative tests for carcinogens
There is a range of alternative tests to assess whether a chemical is likely to cause cancer. Most carcinogens (cancer-causing) chemicals have their effect by damaging genetic material and altering the way cells reproduce. The alternative tests look for such genetic damage, either to genes or to chromosomes.
Genetic mutations in bacteria
The most widely used test is the Ames Test, based on special strains of Salmonella bacteria. These strains have a defective gene, which means they can't produce histidine, an amino acid needed for growth. If a test chemical causes gene mutations it will change the defective gene back to normal, and the bacteria can grow into visible colonies.About 1 billion bacteria are added to a petri dish, together with liver enzymes to activate test chemicals. This is to test both the chemical itself and the products it is broken down into by the liver. After 2 days researchers count the number of colonies of bacteria with normal histidine genes (12).
Sometimes other micro-organisms are used to look for gene mutations, including E.coli bacteria or yeast. However, the Ames Salmonella Test is the best known and is widely used around the world. These bacteria are very sensitive to chemicals causing genetic damage
Genetic mutations in mammal cells
In these tests, lack of particular enzymes in cells is a sign that a mutation has taken place. Human white blood cells (lymphocytes) can be used. Researchers have mostly studied the gene that controls the HGPRT enzyme. Lack of the enzyme makes cells resistant to substances which limit the growth of normal cells. By examining the growth in cell colonies, researchers can calculate the extent of gene mutations (13).
DNA repair in mammal cells
Mammals cells have the ability to repair some of the damage to DNA, so looking for evidence of DNA repair is another way of measuring genetic damage. Human liver cells (hepatocytes) from transplant donors very effectively separate chemicals that cause genetic damage from those that do not (14). Genetic damage can also be studied in human hepatocytes by looking for DNA strand breaks and DNA migration (15).
Chromosome abnormalities in mammal cells
The chromosomes in a dividing cell have two strands, called sister chromatids. One type of genetic damage involves material from one chromatid breaking off and joining to the other in what is called sister-chromatid exchange. Human white blood cells (lymphocytes) are often used in this kind of study. For example, lymphocytes were used to assess whether calcium silicate would be a safe replacement for asbestos. However, this material produced gaps and breaks in chromatids, as well as sister-chromatid exchanges (16).
Usefulness of alternative tests
Some cancer-causing chemicals were first found by alternative tests. Furylfuramide (AF-2) was tested on both rats and mice, but was considered safe enough to use as a food additive in Japan. However, it caused mutations in bacteria, and chromosome damage in cell cultures, and was later banned. Other chemicals detected by alternative tests and later shown to be carcinogenic include ethylene dichloride, ethylene dibromide, some hair dye ingredients, and tris-BP, a flame retardant used in children's pyjamas from 1972-1977 (12).
Tests with human tissue can be used to understand more about how chemicals cause damage to genetic material, and which tissues are most likely to be affected. For example, the effect of components of cigarette smoke on different organs has been compared, using tissues from hospitals. Compounds found in faeces and linked to colon cancer have been shown to cause chromosome damage in colon tissue and cells. Cell cultures have confirmed that asbestos is particularly damaging to mesothelial cells compared to other cells-asbestos causes mesothelioma in humans (17).
Each test has limitations, so the best approach is to use at least 3 tests to measure gene mutation, DNA repair and chromosome abnormalities. If all 3 are positive the chemical is almost certainly a carcinogen. If 2 tests are positive it is probably a carcinogen, but if only 1 test is positive the result should be treated with caution (18).
Computer databases can help to interpret the results of tests. There are now many computer programs that look for patterns in chemicals known to cause cancer. These programs try to link particular molecular structures to particular effects in the body. When results from two such programs were combined, systems called Hazardexpert and COMPACT agreed with animal tests in 86% of cases, while DEREK and COMPACT agreed in 87% of cases. Ideally, of course, the computer predictions would be compared to human data since animal tests are not perfect (19).
A more recent model was developed on the basis of information about 306 chemicals. The model could predict 96% of the carcinogens in this group. It could predict 94% of carcinogens in a new group of 301 chemicals (20).
Plant tests for environmental carcinogens
Plants are used in several tests for gene mutations and chromosome. They are particularly useful for detecting environmental pollution. For example, the Tradescantia stamen hair assay has been used for many years by Japanese scientists to detect radioactivity around nuclear power stations. Plants can also be used to test waste water, factory effluents, exhaust fumes, as well as pollutants in soil and around oil refineries and factories. The following are some examples (21):
Gene mutations in plants
- Tradescantia stamen hair assay. The flower of Tradescantia is usually blue. However, if a gene mutation has occurred stamen hair cells will be pink, so the number of pink cells are counted after the flower has been exposed to a chemical.
- Zea mays (corn) waxy locus in pollen assay.
- Arabidopsis thaliana white embryo assay. Seeds are exposed to a chemical and are then grown until the plant produces seed pods. Immature seeds which contain mutant genes will appear white rather than green (22).
Chromosome abnormalities in plants
- Vicia faba (broad bean) root tip assay. Broad beans are germinated and the roots of seedlings are exposed to the test chemical. The cells of the root tips are then examined under the microscope for chromosome abnormalities.
- Allium cepa (onion) root tip assay.
- Tradescantia micronucleus assay.
Plants are more sensitive to chemicals that damage DNA than mammals (23), so some chemicals that are damaging to plants will not be damaging to humans. However, the plants are a very good warning system. It is fairly certain that a chemical which doesn't cause damage in plants is safe for mammals.
Carcinogens that don't cause genetic damage
All the tests described so far detect chemicals that cause genetic damage. However, there are chemicals that cause cancer without any genetic damage. Cell transformation tests using Syrian hamster embryo cells have been successfully used to detect such carcinogens (24). This test showed that the hormone diethylstilbestrol was a carcinogen even though it did not cause gene mutations (25). In another study, Chinese hamster lung cells were used to show that chemicals like diethylstilbestrol alter normal cell division (26).
Human cells can be used to test chemicals suspected of causing peroxisomal proliferation, that is, an abnormal increase in liver enzymes and cells. In fact, human cells should be used because animal cells don't respond in the same way to these chemicals (27).
Some chemicals don't cause cancer themselves, but create changes that help the cancer cells to thrive. A cell transformation test using a human leukemia cell line has been devised to identify such chemicals. Round cells suspended in fluid attach to plastic surfaces and spread out when exposed to a tumour promoting chemical (28).
As in all other areas, researchers are working on computer programs to predict the effect of chemicals. One program focusing on carcinogens that don't cause genetic damage was able to predict the results of animal tests in 70%-80% of cases (29). As noted before, ideally such programs should be comparing their results with what is known about human cancers.
Problems with tests for teratogens
Chemicals which cause birth defects are called teratogens. The thalidomide tragedy in 1961 showed how damaging drugs can be to foetuses. Many children were born with deformed limbs after their mothers took thalidomide as a sedative or for morning sickness.
This tragedy also highlighted the large differences between species in response to drugs. Thalidomide was not tested as a teratogen before going on the market. It was tested on pregnant rats, mice and guinea pigs after deformed babies were born, but only in New Zealand White rabbits did it produce birth defects anything like those seen in humans. A writer on the subject summed up the results of animal tests as follows:
" In approximately 10 strains of rats, 15 strains of mice, 11 breeds of rabbits, 2 breeds of dog, 3 strains of hamsters, 8 species of primates and in other such varied species as cats, armadillos, guinea pigs, swine and ferrets in which thalidomide has been tested, teratogenic effects have been induced only occasionally " (30).
Humans are more sensitive to thalidomide than other species. The opposite is often true -- around 800 substances are teratogens in some species of mammal, but only about 25 are known to affect humans. For example, aspirin causes birth defects in rats, mice, guinea pigs, cats, dogs and monkeys, but not when taken by humans (30).
In addition to species differences between adult animals, there are also species differences in the structure and biochemistry of the placenta. A review of these differences concluded:
" Drug distribution varies with species so that penetration or non-penetration of a drug in a test animal is no guarantee that this will also be true of the human placenta. Susceptibility to teratogenic effects of a given drug varies with species, strain, and stage of foetal development " (31).
Testing human placentas
There is a ready supply of human placentas in hospitals. They can be kept viable after birth by circulating nutrient fluid through them. Various effects of drugs can be tested on these placentas, and the results will be directly relevant to humans. For example, the drug AZT reduces HIV transfer from mother to foetus, but may also have damaging effects. An in vitro study showed that AZT produces major biochemical changes in the placenta (32).
High blood pressure can be a problem in pregnancy, and drugs used to treat it may also affect circulation in the placenta. An in vitro study showed that one commonly used drug dilated blood vessels in the placenta, whereas another drug did not (33).
Women continue to take drugs during pregnancy for serious conditions such as epilepsy. An in vitro study showed that two commonly used drugs both cross the placenta. However, one drug accumulates in placental tissue, whereas the other does not (34).
Alternative tests for teratogens
There are alternatives to feeding chemicals to pregnant animals and examining the effect on full term foetuses. Not only do such tests destroy thousands of lives, but they are also inefficient. As one researcher has commented:
" Current drug-testing procedures, mostly animal reproduction (in vivo) studies, are time consuming, expensive, and yield responses that vary from species to species. With 50,000 to 70,000 different chemicals already in the market place and some 200 to 400 new drugs being produced every year, alternatives to whole animal, or in vivo studies are needed " (35).
Some tests use rat or mouse embryos removed from the mother after about 10 days of pregnancy. The embryos are kept alive and tested for 1-4 days in vitro. Adult and young animals are still killed, so such tests don't have many advantages over standard teratogen tests.
Cell cultures
Cells from embryos or newborn animals are used in some tests. These cells continue to grow, group together, and develop into specific kinds of tissue as they would in the developing animal. Chemicals that cause birth defects will interfere with these processes in the cell culture. Tests have used brain cells from young mice (36), neural retina cells from chick embryos (37), and cells from the limb buds of mouse foetuses (38).Animals are still being killed in these tests, although fewer animals are involved. An alternative is to collect human cells from operations or autopsies. For example, human cartilage cells (chondrocytes) from babies have been used to test the effect of chemicals on the developing skeleton (39).
Permanent cell lines
The advantage of permanent cell lines is that they were collected from humans or animals many years ago and are now grown permanently in culture. The MOT (mouse ovarian tumour) cell line and the HEPM (human embryonic palatal mesenchymal) cell line have been used to test possible teratogens (40, 41). Many chemicals that cause birth defects affect the growth of cells. However, these cells don't develop into specific kinds of tissue like the fresh cell cultures, so they have not been as accurate a model for the developmental problems of embryos.A recent test has combined the advantage of a permanent cell line (no further animals killed) with the advantage of changing embryonic cells. The development of the mouse ES (embryonic stem) cell line can be stopped by a particular chemical until researchers are ready to carry out a test. When the chemical is removed, the cells develop into specific kinds of tissue. Chemicals that cause birth defects will affect this development. So far this test has been very accurate (42).
Hydra test
The hydra is a tiny, simple animal consisting of a feeding tube with tentacles at the end. If the mid-section is cut into pieces, each piece will grow into a new hydra. These developing "embryos" are exposed to the test chemical for 3 days, and the effect on their normal development is scored. Results have been quite accurate. For example, the hydra test predicts the danger of thalidomide, whereas rat and mouse tests do not (43, 44)
Drosophila eggs
Fruit fly eggs are collected and homogenised when they are only a few hours old. The eggs contain cells which will develop into different kinds of tissue, such as muscle and nerve fibres. During normal development the muscle cells (myocytes) group together into myotubes and the nerve cells (neurons) group together into small ganglia. The egg cells are exposed to a test chemical for 24 hours. If the chemical causes birth defects, it will interfere with the normal development of myotubes and ganglia. This test correctly predicted 94% of 100 test chemicals (45).
Commercial in vitro testing
Alternative test are not merely a pipedream, a hope for the future -- they are being used in many laboratories now. There are companies which use the alternative tests described in this document. They carry out sets of tests on behalf of other companies or government agencies. For example, visit the web sites of:
I would like to see References for this document on carcinogen and teratogen tests.
If you would like detailed information on other types of safety tests, go to:
Toxicity tests, including LD50