Carcinogen Test Process Challenged


In a direct challenge to the federal government’s primary method of determining the safety of chemicals, two new papers published today argue that feeding high doses of toxic chemicals to rats and mice does not accurately predict their potential to cause cancer in humans.

The results seem likely to trigger a re-evaluation of the animal testing process. The researchers argue in the journal Science that it is the unrealistically high doses of the chemicals rather than their inherent cancer-causing ability that leads to tumor formation.

The high chemical doses themselves cause most of the cancers simply by increasing the rate of cell proliferation, a process already known to cause tumors, according to independent reports by two pairs of researchers: molecular biologists Bruce Ames of UC Berkeley and Lois Swirsky Gold of the Lawrence Berkeley Laboratory and pathologists Samuel M. Cohen and Leon B. Ellwein of the University of Nebraska Medical Center.


In the low doses at which humans would normally be exposed to such chemicals--which range from pesticide residues on fruits and vegetables to food additives such as the artificial sweetener saccharin--most of the rodent carcinogens would be harmless, the researchers said.

“Clearly, we can no longer just accept that information from animal bioassays is sacrosanct, as proof that the chemicals will cause cancer in humans,” Cohen said in a telephone interview.

Ames was even more blunt: “We think the current approach to cancer risk assessment is bankrupt.”

Although the new reports reflect a growing feeling among researchers that rodent tests for carcinogenicity need to be re-evaluated, the authorities responsible for conducting them remain convinced of their validity.

“There are only two definitive ways to tell whether chemicals have potential to cause cancer,” said Richard A. Griesemer, head of the Division of Toxicological Research and Testing at the National Institute of Environmental Health Sciences. “One is through epidemiological studies in humans. . . . The second way is to produce cancers in animals.”

Ames thinks that regulatory agencies will eventually swing around to his way of thinking, but he admits that the process will be slow. “The regulatory bureaucracy is like a supertanker--it isn’t easy to change its direction.”

Ames is a controversial researcher who is often perceived by environmentalists as a proponent of the use of synthetic chemicals. He frequently argues that humans are exposed to much larger amounts of naturally occurring carcinogens in food than they are to synthetic chemicals, and that neither exposure represents a major risk.

In today’s paper and three others scheduled to appear during September in the prestigious Proceedings of the National Academy of Sciences, Ames and Gold expand this argument.

They review the data on the 427 chemicals that have so far been tested for carcinogenicity in both mice and rats. That data shows that more than half (212 of 350) of the synthetic chemicals tested and half (37 of 77) of the naturally occurring chemicals are “rodent carcinogens” that cause cancer in both species.

Humans are exposed to relatively high doses, about 1,500 milligrams per day, of the naturally occurring chemicals. In contrast, humans are exposed to an average of only about 0.09 milligrams of pesticides and chemical residues per day.

Most of the naturally occurring chemicals are toxins used by plants to protect themselves from insect predators. The chemicals are common in a broad variety of fruits and vegetables, as well as in coffee and other beverages. Other chemicals that cause cancer in rodents are produced by intense heat during the cooking of food. Humans consume about 2,000 milligrams per day of such chemicals.

But there is no evidence that any of these naturally occurring chemicals cause cancer in humans at normal doses, say the researchers. Over the course of millennia, humans have evolved protective enzymes to detoxify these chemicals and to correct any genetic and cellular damage they may inflict. These broad-based defenses should be able to cope equally well with synthetic chemicals, which frequently are similar to natural chemicals.

“Cabbage and broccoli contain a chemical whose breakdown products behave in the body in much the same way as does dioxin, one of the most feared industrial chemicals,” Ames and Gold said. If the natural chemical does not cause cancer, then it is unlikely that dioxin will, they said. Dioxin has been shown to cause cancer in rodents but thus far not in humans.

But more important, both the Ames and Cohen groups argue, the high incidence of carcinogenesis observed in the rodent tests suggests that something unusual is going on. As many as 60% of the rodent carcinogens do not interact directly with deoxyribonucleic acid--DNA, the blueprint of life--and such interaction is generally assumed to be necessary to produce cancer.

A variety of non-animal laboratory tests, the most famous of which was developed by Ames, can be used to identify such genetic damage.

Hence, said Cohen, “there has to be something else to it,” some indirect activity of the chemicals. That something else, both groups contend, is the initiation of cell proliferation caused by the unusually high doses but not from direct DNA damage.

Very high doses of most of the rodent carcinogens kill some of the cells they contact, thereby inducing the growth of replacement cells. And many different studies have demonstrated that such proliferation induces tumor formation.

But the rodent carcinogens, by and large, do not provoke cellular division in doses below those used for testing. Hence, they do not produce cancer except in the very highest doses, such as might be produced by industrial exposure.

Both groups support controls to limit chemical pollution, but they argue that trying to rid the world of pesticide residues and water pollutants found only at very low levels is counterproductive. It is not likely to improve public health and will divert time and money away from important health problems, they say.

In particular, Ames adds, trying to reduce pesticide residues on fruits and vegetables will raise their cost and make them less widely available. “The pesticide residues don’t matter at all,” he said. “There’s just too little of them. The thing we should be doing is to eat more fruits and vegetables because they protect against cancer.”

To test the chemicals at lower doses in rodents is possible but would call for much larger numbers of animals in order to observe a significant number of tumors, said Griesemer. For example, testing saccharin would require 85,000 animals. “That would be prohibitively expensive and . . . would tie up most of the labs in the country,” he said. “It’s really not feasible.”