Laboratory Mice Shown to be Influenced By Lab Environment

06/03/99    Portland, Ore.

Genetically identical mice in exact same studies react differently, study shows.

Research conducted at three laboratories in the U.S. and Canada shows that genetically identical mice react differently to controlled behavioral tests. The study, to be published in the June 4 issue of Science, indicates that even controlled research conducted at different laboratories can yield significantly different results.

"We would have come to three different conclusions if we had done the experiments on our own," said lead author John Crabbe, Ph.D., Portland Veterans Affairs Medical Center and professor, Oregon Health Sciences University Department of Behavioral Neurosciences.

"This study confirms a suspicion among scientists about laboratory influence and reaffirms the importance of carefully controlling and repeating experiments before coming to any conclusions," said Crabbe.

The study involved simple behavioral tests of several particular genetic strains of mice and one "knockout" strain. The behavioral characteristics of the genetic strains are well documented, but the mutant strain was missing a neurotransmitter receptor gene that made its behavioral characteristics less predictable. While results for some of the tests were consistent at the three labs, other tests had significant differences.

"We saw that some behaviors were more susceptible to differences in environment," said Crabbe. For example, an anxiety test that subjected the mice to exposed and elevated platforms--akin to walking out on a high-dive board--yielded different results at the three labs even among the same strains. The knockout strain was found to have the greatest degree of difference at each site. On the other hand, a test for preference between alcohol and water produced consistent results at the three labs, indicating strong genetic disposition to this behavior.

"The conclusion that unknown, subtle environmental features have profound effects on the behaviors of isogenic (identical) animals reinforces the idea that for behaviors like alcoholism, genes will define risk, not destiny," said Enoch Gordis, M.D., director of the National Institute on Alcohol Abuse and Alcoholism, one of the study's sponsors.

The study has direct implications for behavioral researchers, but Crabbe says it also carries a warning message for other genetic researchers. "The best evidence would come from multiple tests in multiple labs, which isn't always convenient and efficient. Our hope is to get some discussion going on this issue."

Other laboratories participating in the study were at the University of Alberta in Edmonton and State University of New York at Albany.