Carcinogenicity Tests: How We Identify Cancer-Causing Chemicals

Every day, we come into contact with thousands of chemicals-through the food we eat, the air we breathe, and the products we use. But how do scientists figure out which of these substances might cause cancer? The answer lies in carcinogenicity testing: a multi-layered scientific process that combines molecular analysis, bacterial experiments, long-term animal studies, and human population data. These tests form the backbone of how regulatory agencies worldwide decide which chemicals are safe and which ones need to be restricted or banned.

Overview of carcinogenicity testing methods 🔗

Identifying whether a chemical causes cancer is not a single-step process. Scientists rely on a tiered approach that uses several complementary testing methods. Each tier offers a different level of evidence, and together they build a comprehensive picture of a chemical’s cancer-causing potential.

The main categories of carcinogenicity testing include:

Molecular structure analysis (SAR/QSAR): Computational methods that predict whether a chemical might be carcinogenic based on its molecular features. Short-term tests: Rapid laboratory assays, like the Ames test, that screen for genetic mutations linked to cancer. Long-term bioassays: Animal studies lasting up to two years that directly observe tumour development. Epidemiological studies: Research on human populations to identify cancer patterns linked to chemical exposures.

This layered system exists because no single method is perfect. Short-term tests provide rapid screening but may miss non-mutagenic carcinogens, while animal bioassays offer controlled conditions but may not perfectly predict human responses . Epidemiological studies examine real human populations but struggle to isolate the effects of individual chemicals. By combining all of these approaches, scientists arrive at more reliable conclusions.

Molecular structure analysis: predicting cancer risk from chemical features 🔗

Before any wet-lab experiment begins, scientists can get an initial read on a chemical’s cancer-causing potential by examining its molecular structure. This approach is known as Structure-Activity Relationship (SAR) analysis or, in its more advanced computational form, Quantitative Structure-Activity Relationship (QSAR) modelling.

How SAR and QSAR work 🔗

The fundamental idea behind SAR is straightforward: a chemical’s biological activity is closely linked to its molecular structure. Certain structural features-like electrophilic groups that can react with DNA-are known to be associated with carcinogenicity. As early as 1978, researchers Elisabeth and James Miller established that electrophilic molecules are predicted to be carcinogens, because DNA is rich in nucleophilic centres that can bind covalently to such substances .

QSAR takes this a step further by using computational models to mathematically relate molecular properties to biological effects. QSAR analysis uses numerical representations of a molecule’s structure, properties, and features-known as molecular descriptors-to predict various toxicity endpoints, including acute toxicity, genotoxicity, and carcinogenicity . These descriptors include things like molecular surface area, electronic charge distribution, and lipophilicity.

The practical value of computational screening 🔗

QSAR models are particularly useful because they can screen large numbers of untested chemicals quickly and cheaply. The U.S. EPA has supported the development of QSAR models to predict cancer slope factors, helping to prioritize and screen potential carcinogenic chemicals at hazardous sites . Tools like the OECD QSAR Toolbox allow toxicologists to assess chemicals using a combination of structural profiling and read-across from tested analogues.

However, these computational methods have limitations. Current research has found substantial variability between different QSAR models, indicating there is still progress needed before these methods can function as reliable stand-alone tools for carcinogenicity prediction . For this reason, QSAR results are typically used as a starting point-flagging chemicals for further testing rather than serving as a final verdict.

The Ames test and other short-term tests 🔗

When scientists need a fast, affordable way to screen chemicals for cancer-causing potential, they turn to short-term mutagenicity tests. The most well-known of these is the Ames test, developed by biochemist Bruce Ames in the early 1970s.

How the Ames test works 🔗

The Ames test is based on a key principle in cancer biology: most carcinogens are also mutagens-they damage DNA in ways that can lead to uncontrolled cell growth. The test uses specially engineered strains of Salmonella typhimurium bacteria that carry mutations in genes involved in histidine biosynthesis, making them unable to produce this essential amino acid on their own .

Here is how the process works in practice. The mutant bacteria are placed on a growth medium containing only a trace amount of histidine, along with the chemical being tested. If the chemical is mutagenic, it causes reverse mutations in the bacteria’s DNA, restoring their ability to synthesize histidine. These bacteria then form visible colonies on the plate, and more potent mutagens produce greater numbers of colonies on the agar media . Scientists compare colony counts from treated plates to untreated controls to determine whether the test chemical is mutagenic.

Accounting for mammalian metabolism 🔗

One important feature of the Ames test is the inclusion of a liver enzyme extract (called S9 mix) from rats. This addition is critical because many chemicals are not directly mutagenic-they only become harmful after the body’s metabolic processes transform them. Benzo[a]pyrene, for example, is not mutagenic itself, but its metabolic products are, which is why rat liver extract is optionally added to simulate metabolic effects .

Strengths and limitations 🔗

The Ames test became a game-changer in chemical safety screening. It is one of the eight tests required under the U.S. Pesticide Act and one of the six tests required under the Toxic Substances Control Act . Early studies by Ames showed that approximately 90% of known carcinogens could be detected by this test, although later studies showed the identification rate to be between 50-70% of known carcinogens .

The test has real-world impact. It was used to demonstrate that the flame retardant tris(2,3-dibromopropyl)phosphate and the food additive furylfuramide were mutagenic, ultimately leading to both being withdrawn from consumer products .

However, the Ames test is not foolproof. It can sometimes fail to detect genotoxic compounds, particularly those that cause large DNA deletions or those that operate through non-genotoxic mechanisms . Substances that are toxic to bacteria themselves (like antibiotics) also cannot be adequately tested. This is why a positive Ames test does not confirm a chemical is carcinogenic-it flags it for further investigation.

Beyond the Ames test: other short-term screening methods 🔗

The Ames test is just one tool in a broader toolkit of short-term genotoxicity assays. Other important tests include:

The micronucleus assay: This test examines whether substances cause chromosome damage in developing red blood cells inside bone marrow, where a micronucleus forms when a chromosome is broken or fails to migrate properly during cell division . Detection of such chromosomal damage matters because it has been linked to birth defects, infertility, and cancer.

The Comet assay: This test measures DNA strand breaks in individual cells by examining the pattern of DNA migration during electrophoresis-damaged DNA forms a distinctive “comet tail” shape.

Unscheduled DNA Synthesis (UDS) assay: This short-term in vivo assay, which measures DNA repair activity in rat liver, is recommended by most regulatory authorities when in vitro tests give positive results but in vivo cytogenetics assays are negative .

Together, these short-term tests provide a rapid, cost-effective screening layer that identifies chemicals warranting more expensive, long-term investigation.

The role of bioassays in carcinogenicity assessment 🔗

When short-term tests flag a chemical as potentially dangerous, scientists often move to the most rigorous experimental method available: the long-term animal bioassay. These studies are considered the gold standard for experimental evidence of carcinogenicity.

How bioassays are conducted 🔗

The U.S. National Toxicology Program (NTP) conducts long-term carcinogenicity studies that typically involve exposing both sexes of rats and mice to a substance for a period of two years -nearly the entire lifespan of these animals. Scientists use multiple dose groups along with untreated control animals to see whether the chemical increases tumour incidence.

These studies are conducted at high doses to maximize the potential for detecting carcinogenic effects, with the ultimate aim of extrapolating risk estimates to the much lower exposure levels that humans typically experience . This high-dose approach is necessary because testing at realistic human exposure levels would require enormous numbers of animals to detect statistically meaningful differences.

Interpreting bioassay results 🔗

Reading bioassay results is not always clear-cut. Ambiguous evidence may arise from statistically marginal results or from increases only in commonly occurring benign tumours . If tumours develop in only one species and are linked to species-specific toxicity, that pattern is more typical of non-genotoxic carcinogens.

When bioassay data are applied to human risk assessment, additional factors must be considered-including evidence from short-term genotoxicity tests, data on how the chemical is metabolized, and potential levels of human exposure. Mathematical models may be used for low-dose extrapolation, although no general agreement exists on which method is most appropriate, and calculated risk estimates can vary considerably between methods .

Moving beyond traditional bioassays 🔗

The traditional two-year rodent bioassay is expensive, time-consuming, and raises ethical concerns about animal welfare. In response, the scientific community is developing alternative approaches. The International Council for Harmonisation (ICH) has evaluated a Weight-of-Evidence approach that can substitute for the two-year rat study for certain pharmaceuticals . This approach combines data from shorter toxicology studies, genetic toxicology results, and information about hormonal perturbation and immune modulation to predict carcinogenic risk without requiring a full-length bioassay.

Emerging high-throughput and multiplexed in vitro approaches are also being developed to distinguish between genotoxic carcinogens, non-genotoxic carcinogens, and non-carcinogens, using human cell lines rather than animal models .

Epidemiological studies: learning from human populations 🔗

While laboratory tests and animal studies provide controlled experimental data, epidemiological studies offer something no other method can: evidence from actual human populations. Epidemiology examines disease patterns across groups of people to identify links between chemical exposures and cancer risk.

Types of epidemiological evidence 🔗

Cancer epidemiology typically relies on two main study designs. Cohort studies follow groups of people with known chemical exposures (often in occupational settings) over time, tracking whether they develop cancer at higher rates than unexposed populations. Case-control studies compare people who already have cancer with matched controls, looking backward to identify differences in past chemical exposures.

Occupational cohort studies have been particularly valuable, as workers in certain industries experience higher and more measurable exposures than the general population . For instance, studies of workers exposed to polychlorinated biphenyls (PCBs) provided key evidence that led the International Agency for Research on Cancer (IARC) to classify PCBs as carcinogenic to humans.

The IARC Monographs programme 🔗

The most influential system for evaluating carcinogenic evidence is the IARC Monographs programme, run by the World Health Organization’s cancer research agency. Since 1971, more than 1,000 agents have been evaluated, of which more than 500 have been identified as carcinogenic, probably carcinogenic, or possibly carcinogenic to humans .

IARC classifies agents into groups based on the strength of evidence. Group 1 agents are confirmed human carcinogens (like asbestos and tobacco smoke). Group 2A agents are probable carcinogens, Group 2B are possible carcinogens, and Group 3 agents are not classifiable based on current evidence. These classifications rely heavily on epidemiological data combined with animal and mechanistic evidence.

Challenges of epidemiological research 🔗

Epidemiology is powerful but comes with inherent difficulties. The human diet, for example, is a complex mixture of nutrients and chemicals that are notoriously difficult to measure in observational studies, and many of these could serve as plausible confounders of the effect under study .

Other challenges include long latency periods (cancer may take decades to develop after exposure), difficulty in accurately measuring past exposures, and the fact that people are typically exposed to multiple chemicals simultaneously. For environmental chemicals, cancer risk among general populations is generally considered lower than in occupational settings, but chronic exposure to relatively low levels may still contribute to cancer risk over a lifetime .

Despite these challenges, epidemiological studies remain essential. They are the only testing method that directly demonstrates cancer risk in humans. Strong epidemiological evidence, supported by animal and mechanistic data, provides the most convincing case for classifying a substance as a human carcinogen.

How these methods work together 🔗

No single carcinogenicity test can provide a complete answer. The strength of the assessment system lies in how these different methods complement one another. Molecular structure analysis and QSAR models offer rapid initial screening at minimal cost. Short-term tests like the Ames test quickly identify chemicals that damage DNA. Long-term bioassays provide controlled experimental evidence of tumour development. And epidemiological studies confirm whether chemicals actually cause cancer in human populations.

Regulatory agencies like the U.S. EPA, the European Chemicals Agency (ECHA), and IARC use data from all these sources to make decisions about chemical safety. The field continues to evolve-with advances in genomics, computational modelling, and organ-on-chip technology promising faster, more accurate, and more ethical methods of identifying cancer-causing chemicals in the future.

What do you think? With thousands of chemicals still untested for carcinogenicity, should regulatory frameworks require more pre-market testing before chemicals enter consumer products? And as computational tools like QSAR modelling improve, could they eventually reduce our reliance on animal bioassays-or will real-world human and animal data always remain irreplaceable?

References

1. https://assessments.epa.gov/risk/document/&deid=238403
2. https://qsartoolbox.org/
3. https://en.wikipedia.org/wiki/Ames_test
4. https://www.sciencedirect.com/topics/pharmacology-toxicology-and-pharmaceutical-science/ames-test
5. https://www.sciencehistory.org/stories/distillations-pod/the-ames-test/
6. https://ntp.niehs.nih.gov/research/apptoxres/cartox
7. https://www.frontiersin.org/journals/toxicology/articles/10.3389/ftox.2024.1353783/full
8. https://academic.oup.com/mutage/article/39/2/69/7596158
9. https://monographs.iarc.who.int/