The effort is sponsored by the National Institute of Environmental Health Sciences, one of the federal National Institutes of Health, to stimulate a search for cheaper and faster tests that could help it catch up with some of the 100,000 chemicals estimated to be in use in the world. While it has been called "Beat the Mouse," its formal title is the Predictive Toxicology Evaluation Experiment.
In the current supplement of the NIEHS journal, Environmental Health Perspectives, the participating scientists are making predictions of carcinogenicity using methods they hope will prove to be useful alternatives or supplements to the standard tests of the federal National Toxicology Program, which rely on dosing both mice and rats for as long as two years. The scientists are predicting the results of 30 of these NTP rodent tests for carcinogenesis by using computer machine-learning, comparisons of the test substances' molecular structures with those of known carcinogens, tissue or cell-culture tests, including the Ames test for mutagenicity, and logic and educated guesses made by experts.
The scientists in this evaluation experiment may help answer the question: Can chemicals
be tested faster, cheaper and accurately, using fewer or no rats and mice?
"The answer is important because we need information on more chemicals in our environment than we can get using the standard tests," Kenneth Olden, Ph.D., director of both the NIEHS and the National Toxicology Program, said. "A two-year rodent exposure study actually can require five years to reach final conclusions when you add in the planning time and the extensive analysis of results. It can cost $2-4 million. So we can't currently do nearly enough tests -- when you consider that there are thousands of chemicals which we swallow as drugs, use as pesticides and cleaners, and breathe, drink or eat. We need accurate methods that are shorter, cheaper and quicker."
As a result, NIEHS capitalized on the testing being performed by NTP and invited scientists to look at the same chemicals using whatever method they thought promising -- and to publish their predictions of their potential to cause cancer before the final results from standard testing are in. The resulting predictive papers fill more than a hundred pages of the NIEHS journal supplement volume 104.
Final results from some of the standard, two-year bioassays are beginning to come out, but not enough to indicate which predictive approaches worked best. But the purpose is not winning so much as "focusing the intellectual resources of different research groups on a common problem," NIEHS' Douglas W. Bristol, Joseph T. Wachman and Arnold Greenwell write in an introduction to the journal. They, along with NIEHS carcinogenesis and mutagenesis lab chief Ray Tennant, shepherded the effort. (In another effort, Tennant and others are determining if two lines of transgenic mice -- mice which have been given a human gene implicated in many human cancers, will provide accurate results in six months of tests, rather than the
The organizers acknowledge that when the results of the conventional two-year tests are in, researchers may initially check to see how accurate their predictions were -- "Hey, we were right on for everything but vanadium pentoxide and Yellow No. 7!"
But Bristol said the more important questions will follow: Can we adjust our methods to make our predictions 100 percent? Did we get the right answers for correct reasons? What can we learn by comparing our results, and reasons, with the different approaches of Ashby, Benigni or Moriguchi?
When the results of the 30 NTP bioassays are completed, a final tally will be prepared for a subsequent journal, Bristol, Wachman and Greenwell said.
A copy of the NIEHS' journal, Environmental Health Perspectives supplement volume 104 may
be obtained by calling Ruth McFarland, 919/541-3665.