Pushing For A Paradigm Shift In Cancer Risk Assessment

Recent studies are raising serious questions among toxicology researchers about the validity of cancer risk assessment methods as practiced today. Based on these findings, a growing number of scientists are calling for a thorough reevaluation of the criteria used to identify human carcinogens. "Twenty years ago, the assumptions that were made were appropriate, and the decisions made on those bases were appropriate because that's what we knew," says Samuel Cohen, a pathologist at the Epply Institute for Cancer Research at the University of Nebraska. "I think now that we've seen that there are examples that point a different way, that those have to be incorporated into the way we evaluate risk." But while many of the thousands of toxicologists, cell biologists, biochemists, and cancer researchers are struggling to establish a new paradigm for determining chemical risk, environmentalists are fighting to maintain the status quo. Being called into question are the standard bioassay protocols currently enlisted to evaluate cancer-causing chemicals in animal model systems. These protocols incorporate two basic assumptions: (1) a chemical that causes cancer in rats and mice has a high probability of causing cancer in humans (interspecies extrapolation); (2) chemicals that cause cancer when administered at high doses will also cause cancer when administered at low doses (dose extrapolation). Animal models long accepted as predictors of cancer in humans have, in some cases, been shown to be completely off the mark with results that are not transferable from species to species, toxicology researchers say. In other cases, it has been shown that threshold levels must be reached before a substance becomes tumorigenic, making dose extrapolation inappropriate. Moreover, some researchers believe that the danger from naturally occurring carcinogens far exceeds that from synthetic sources, leading them to conclude that estimates of human cancer risk from synthetic chemicals have been greatly exaggerated. On the basis of all the evidence, researchers believe that the time has come to reconsider public policy on cancer risk assessment. Cohen thrust the issue of risk assessment into the limelight with a 1990 paper (S.M. Cohen, et al., Science, 249:1007,1990) that called for the inclusion of mechanistic information in determining the relevance of animal test results for human exposure. Although a strong advocate of change, he insists that he is not against animal bioassays. Rather, he claims, he is "trying to encourage people to interpret them a little bit more carefully and specifically, when extrapolating to humans." In the case of saccharin, which causes bladder tumors in male rats when administered at high doses, Cohen says the tumors are linked to the presence of a protein in urine (au-globulin) to which saccharin binds. The tumors occur as a regenerative effect following erosion of the superficial bladder epithelial cells. The erosion is attributed to a precipitate in the urine generated when saccharin binds to the protein. However, a2u-globulin is specific to male rats and is not present in female rats, mice, or humans. In a recent paper, Cohen concludes, "On the basis of mechanistic considerations alone, humans are unlikely to develop bladder cancer as a consequence to exposure of saccharin, even if humans consumed levels as high as those consumed by the rat" (S.M. Cohen, et al., Chemical Research in Toxicology, 5:742, 1992). Nevertheless, because of the way that regulatory laws are written, saccharin cannot be used as a food additive, but by special legislative exemption it can be used as an artificial sweetener with an accompanying warning as to its carcinogenicity in animals. "There are still a lot of people who will not abandon the two assumptions of species extrapolation and dose extrapolation," says Cohen. "If you continue to [make these assumptions], we are going to get into some major difficulty, and vitamin C is a kind of point at issue. The sodium salt of vitamin C behaves very similarly to sodium saccharin in the male rat. If the same kind of extrapolation that the Environmental Protection Agency does currently for saccharin is done for vitamin C, you end up having a maximum allowable dose of vitamin C in the diet that is less than what you have to have to avoid scurvy." In addition, evidence suggests that the dose extrapolation assumption of the rodent bioassay may also be inappropriate for saccharin. When it is administered in the male rat, Cohen reports that "a threshold response appears to be involved. No effects are seen in the bladder epithelium at doses below 1 percent, while effects are observed when the chemical is administered at doses of 2.5 percent or higher." It is the effects of nongenotoxic agents such as saccharin that may be the most difficult to reconcile with the bioassay assumptions, he says. Genotoxic agents cause cancer by interacting with DNA and affecting the rate of genetic damage. Nongenotoxic agents cause cell proliferation, increasing the risk of cancer because with each cell DNA replication cycle there is a small but nonzero risk of genetic damage. Therefore, Cohen calls particularly for a new way of evaluating these nongenotoxics. Bruce Ames, the University of California, Berkeley, biochemist whose Ames test is used to evaluate numerous known human and animal carcinogens, is a well-known critic of current methods in cancer risk assessment. "In chronic tests at the maximum tolerated dose [MTD], more than half of all chemicals tested [both natural and synthetic] are carcinogenic in rats, and a high percentage of these carcinogens are not mutagens," he says. Since proliferation can be caused by toxicity of chemicals at high doses (cells are killed and then replaced), ignoring this contribution, Ames contends, "greatly exaggerates a risk factor for cancer that may be limited to high doses." In order to understand what causes tumors, Ames asserts, "The case is absolutely convincing that, in animal cancer tests, you need to measure two things. You want to measure both DNA damage and cell division. Animal cancer tests don't measure cell division, so they can't do risk assessment if they are leaving out one central thing. They do linear extrapolations from the high dose to the low dose. In the fine print EPA tells you that the risk could be zero, but they never say that in the press releases, and it never gets into the newspapers." Of even greater concern, perhaps, than the question of exaggerated risk is the possibility that, in some cases, linear extrapolation and interspecies extrapolation may lead to an understatement of actual risk to humans. In a major review of their own work and that of others, Miriam Poirier, a biochemist at the National Cancer Institute, and organic chemist Frederick Beland of the National Center for Toxicological Research examined the relationship between rodent tumor incidence induced by chronic carcinogen exposure and DNA adduct formation (M.C. Poirier, Chem. Res. Toxicol., 5:749, 1992). One study looks at the levels of dG-C8-ABP, the major DNA adduct produced by an aromatic amine found in cigarette smoke, 4-aminobiphenyl (4-ABP), and the concomitant bladder tumor incidences in male mice, compared with the DNA adducts measured in the bladders of human cigarette smokers and the increased risk of bladder cancer in male cigarette smokers. The comparison shows that a 50 percent bladder tumor incidence in mice is associated with levels of dG- C8-ABP that are nearly 200 times higher than those projected for a 50 percent bladder tumor incidence in humans. Poirier and Beland suggest that interspecies extrapolations for complex mixtures such as cigarette smoke may provide less accurate correlations than those for a single agent. "Low DNA adduct levels in tissues should not be considered to be evidence of low tumorigenic risk, as they sometimes are in rodents." Poirier says that the risk is only partially the result of the adducts from a suspected carcinogen. "If you think about it logically," she explains, "obviously there are lots of other compounds being taken in cigarette smoke and some of them are initiators and some of them are promoters. The risk is presumably a composite of the genotoxic effects of all these different things, plus whatever promoting influences happen to occur." In addition, "because humans have such a long life span as compared to rodents," says Poirier, "this allows for other tumorigenic events to occur, whether they be self-immune, genotoxic-immune, or promotional events. The application is not simple." For James Swenberg, a toxicologist at the University of North Carolina, part of the concern about current methods of risk assessment is their impact on the allocation of available resources. According to Swenberg, understanding how to do risk assessment better will allow for better strategies for cleaning up hazardous waste sites. "I'm familiar with research on one Superfund site, where the projected cleanup costs are $100 million," he says. "If the risk assessment is reduced by a factor of four through better science, the cost that is saved is $50 million, and the cost of doing that research is under $1 million. So it's a tremendous payback for the investment in research. Somehow that kind of information has not gotten through to the people who are making the funding decisions." And Ames believes that risk assessment that targets only synthetic or industrial chemicals diverts resources from important to unimportant things. "There is an imbalance in both the data and the perception of chemical carcinogens because little attention has been given to natural chemicals ... half of which have been shown to be carcinogenic at the MTD," he says. Poirier believes that "eventually what we are doing is going to change risk assessment, but it is a kind of grass-roots effort." This field, once the realm of theoreticians and epidemiologists, now includes a generation of scientists trained in carcinogenesis and toxicology who are bringing new insights into what really constitutes risk, she says. These are the people who will push the hardest for a reevaluation of the risk assessment status quo. Another sign of scientific schism is the rift that developed within a National Research Council committee over the use of MTD in cancer risk assessment. In Issues in Risk Assessment (National Academy Press, Washington, D.C., 1993), the majority of the committee continued to uphold the standard of using MTD testing as an initial trial test of possible carcinogens. "MTD testing should not be replaced with a dramatically new approach until there are more data to support such a change," says Bernard Goldstein, director of the Environmental and Occupational Health Sciences Institute at the University of Medicine and Dentistry of New Jersey in Piscataway. But about one-third of the committee dissented, advocating a testing approach based on under- standing the mechanisms by which moderate doses of a chemical affect animal physiology and health. "The scientific data from hundreds of tests have come in," says Richard Reitz, a committee member and an associate scientist with Dow Chemical Co. in Midland, Mich., adding that these data indicate a need to shift away from the use of MTD. "However, some [on the committee] are willing to be led there more rapidly than others." And, says Swenberg: "There are people who don't want to accept something. I think they have public health foremost in their minds, but they want absolute guarantees that that is the mechanism by which a chemical works. It's my feeling that if we require that degree of rigor, we will never move forward." This hotly debated research topic has also broken into the public arena, with EPA's release in early February of a list of 35 pesticides that will be banned or restricted because laboratory tests involving the aforementioned MTD methodology indicate they may cause cancer in humans. EPA had been technically violating federal statutes by allowing trace amounts of these pesticides to appear in food because it believed that such levels are safe. But the Ninth Circuit Court of Appeals in San Francisco ruled in 1992 in favor of the National Resources Defense Council (NRDC) and ordered EPA to come into compliance with the letter of the law. In this case, the law is the Delaney Clause, a 35-year-old section of the Federal Food, Drug and Cosmetic Act, which prohibits even the slightest trace of pesiticides or additives in food if laboratory tests indicate potential carcinogenicity. But even the very environmental group that is forcing the EPA to strictly enforce the Delaney clause is apparently softening its position on carcinogenic risk. "The philosophy of the Delaney clause was we don't know enough about carcinogens that we should simply allow them in the food supply," Al Meyerhoff, a senior attorney with NRDC, told the New York Times (K. Schneider, Feb. 1, 1993, page E6). "We are prepared to work toward legislation that replaces Delaney with a comprehensive reform that accomplishes real protections. Delaney is not sacrosanct." Sara Brudnoy teaches organic chemistry at Sage Junior College, Albany, N.Y.

Pushing For A Paradigm Shift In Cancer Risk Assessment

Interested in reading more?

Become a Member of