Skip to content ↓

Environmental chemicals may not mutate people

William G. Thilly
William G. Thilly

CAMBRIDGE, Mass.--Environmental chemicals are probably not producing human cancers by causing the mutations found in those cancers, an MIT toxicologist and epidemiologist says in the July 1 issue of Nature Genetics.

His arguments include the relationship between cigarette smoke and lung cancer. An exception is skin cancer and sunlight because there is a direct connection between exposure to sun and mutations in skin cells.

In work that has ramifications for cancer research and environmental and pharmaceutical regulatory policy, William G. Thilly, professor of biological engineering in MIT's Biological Engineering Division, and his students mathematically analyzed years of data on lung cancer, leukemias, lymphomas and cancers of the nervous system, pancreas and intestines. They found that the historical death rates for many common cancers that have risen steadily for Americans born in the 19th century could be explained without attributing the changing rates of human mutation to environmental causes.

Most environmental cancer research is based on the assumption that environmental agents in food, air and water enter our bodies and cause the mutations known to cause human cancers. Thilly found that instead of causing mutations in the first place, environmental carcinogens may accelerate the growth of cells in which a mutation has already occurred spontaneously as our DNA replicates and repairs itself over a lifetime.

Thilly bases his arguments on perspectives from epidemiology, experimental cancer studies and the results of technology developed in his laboratory that, for the first time, permits measurement of the kind and number of mutations in any gene in human tissues from which tumors arise.


Because four generations is too short a time to blame for inherited genetic changes, epidemiologists and cancer researchers generally agree that the cancer rate increases for Americans born in the 19th century had to be caused by changes in the environment. In searching for environmental factors, most research focused on the identity of mutagens and the mechanisms by which they cause mutations in experimental systems using bacteria through rodents as models.

For lung cancer, the environmental cause was identified 50 years ago as cigarette smoking; for most other cancers the environmental factors are not yet known.

Hold everything, Thilly argues in his Perspectives article in Nature Genetics. As a student in a class taught by MIT Nobel laureate Salvador E. Luria, he became aware of the crucial difference between inducing and selecting mutations. His present challenge depends on precise measurement techniques that pinpoint the kind and number of mutations in human genes in human tissues. His lab provided the data for his challenge and a general means to test the idea that environmental agents such as cigarette smoke cause cancer by causing mutations in the tissue at risk.

After his graduate student, Hilary Coller (now at the Fred Hutchinson Cancer Research Center in Seattle), found that mitochondrial DNA in smokers and nonsmokers had the same number and kind of mutations, Thilly said, "The old idea kind of went up in smoke."

Weighing various forms of evidence, he concluded, "The widely held hypothesis that environmental chemicals induce a significant fraction of human point mutations has not been supported by observation. Mutagens (except for sunlight) may have little, if any, effect on human cancer risk."


There is no doubt that genetic mutations--inherited or not--accumulate in cells during stages of cancer. Thilly found that these mutations occur just as regularly in cultured human cells in the laboratory that were not exposed to any environmental mutagens. It was only in the case of sun-exposed skin cells, which develop thousands of times more mutants than other cells, that an environmental factor was clearly and unambiguously shown to be mutagenic to humans.

Even cigarette smoking didn't increase "point mutations," or mutations in the DNA sequence. The number and kinds of point mutations were essentially identical in the bronchial cells of nonsmokers and smokers, even for identical twins where one sibling smoked and the other didn't.

However, the rate of growth of hypothetical precancerous lesions was calculated to be highly sensitive to exposure to cigarette smoke. Similarly, as deduced by another epidemiological group, uranium miners' exposure to radon did not appear to influence the initiation of lung cancer but did appear to accelerate the growth of precancerous tissue.


Tiny clusters of atypical cells are always present in our bodies. They can be found only through painstaking microscopic observation, and most of the time, they don't become problematic. Current methods can't predict which, if any, of these minutely different cells are precancerous. These precancerous cells, Thilly says, could be most vulnerable to environmental carcinogens.

In separate research, scientists recently proposed that cells unnecessarily "repairing" undamaged DNA might cause most genetic mutations that lead to cancer, with spontaneous DNA replication mistakes leading to the rest.

Thilly warns that, his arguments notwithstanding, environmental factors could be acting in mutagens in any organ for which it has not yet been demonstrated otherwise.

"The direct testing of the 'environmental mutagen' hypothesis in every human tissue from which the more common tumors arise would seem to be an important research priority and necessary if public policy is to be based on scientific evidence," he writes. "The failure to test this hypothesis may have already misdirected the search for the causes of both genetic change and cancer in humans."

Thilly suggested that more closely studying the potentially precancerous tissues of people who have a genetic history of early-onset colon, kidney or nervous system cancer might be most fruitful.

For more information, see

Related Topics

More MIT News