Identifying And Evaluating Sources Of Variability In Rodent Studies

Environmental Health Sciences

Investigators

Linked publications & trials

Abstract

Short-term rodent bioassays, such as the uterotrophic bioassay, have been developed to assess estrogenic effects of chemicals. The outcomes of these assays may be affected by factors other than the chemicals under study. For example, many laboratory diets contain phytoestrogens that have estrogenic effects. Also, diets vary in the amount of metabolic energy that is available. High energy consumption may interfere with estrogenic responses. Furthermore, different rodent strains exhibit differing sensitivities to estrogens. Because bioassays for estrogenic effects are conducted at many laboratories around the world, using different diets and different rodent strains, conflicting results sometimes occur. For the endpoint of timing of vaginal opening, we showed that the commonly used Sprague-Dawley rat is less sensitive to estrogenic chemicals than the Fischer 344 rat or the CD-1 mouse. We also found that phytoestrogen level in the diet is more highly correlated with some estrogenic responses than is metabolic energy. We recommend that rodents in bioassays for assessing estrogenic effects should be fed a low metabolic energy diet that is free of phytoestrogens. Furthermore, sensitivity to estrogenic effects varies among strains of mice and rats as well as among endpoints, so strain selection should be made carefully.[unreadable] [unreadable] The NTP conducts rodent micronucleus tests as part of its battery of genetic toxicity testing. Normally, as red blood cells mature, they shed their nucleus. If chromosome damage has occurred in the red blood cells, small parts of the nucleus (micronuclei) may remain in the cell. In the micronucleus test, blood or bone marrow of mice or rats exposed to a chemical is examined under the microscope and numbers of mature red blood cells containing micronuclei are counted. Recently, a new technique using flow cytometry has been proposed for counting micronuclei in a less mature population of red blood cells. An advantage of this technique is that many more cells can be examined per animal than is feasible with microscopy. We contributed to two studies comparing microscopic and flow cytometry enumerations of micronuclei. 1) Blood and bone marrow samples from studies of nine chemicals in mice and rats were evaluated using both microscopic and flow cytometry techniques. We found that the two techniques produced very similar micronucleus counts and they resulted in the same conclusion about whether a chemical is genotoxic. 2) Using a range of typical values of micronucleus occurrences, we conducted computer simulations to study the effects of the number of cells examined per animal on the power and sensitivity of the micronucleus test. Based on both of these studies, the NTP has decided to adopt flow cytometry for the micronucleus test.[unreadable] [unreadable] In the mid-1990s, several NTP studies of chemicals were partially compromised by an infection of Heliobacter hepaticus. This bacterium causes hepatitis in mice which often leads to liver tumors. Because liver tumors were observed in these studies, it was not clear whether H. hepaticus or the chemical exposure was responsible. For most of these studies, the chemical caused cancer in other organs unaffected by the infection, so the carcinogenic effect of the chemical could be determined. For triethanolamine, however, tumors were observed only in the liver. Several years later, the NTP conducted a second, identical study of triethanolamine in mice that were free of H. hepaticus infection. However, at that time, the diet that the mice were fed had been changed and animals were larger in the second study. Because liver tumor rates increase with body weight, among other factors, our analyses include statistical modeling of liver tumor rates using historical control data from the each study. We are analyzing data from both studies to determine whether liver tumor rates can be attributed to triethanolamine, H. hepaticus, diet or body weight. [unreadable] [unreadable] Several pathologists examining kidneys of NTP rats have noticed a distinctive tumor morphology that they believe occurs spontaneously. The current diagnostic rules, however, do not distinguish this morphology from chemically-induced adenomas and carcinomas. We are analyzing the occurrences of this tumor morphology from among all NTP studies to determine whether the occurrences are, indeed, random. If they are random, the diagnostic rules may be refined to separate these tumors from chemically-induced tumors, potentially removing noise from the signal of chemical induction of kidney tumors. This would improve the accuracy of NTPs determinations regarding carcinogenicity.[unreadable] [unreadable] Because study design is critical to the collection of high quality data and may control factors involved in extraneous variability of data, we contributed to NTP study design teams for 5 chemicals, to the re-design of expanded Reproductive Assessment by Continuous Breeding (RACB) studies, and to the re-design of NTPs two-year rodent bioassays to include in utero exposure. We are also providing advice to the new NTP Host Susceptibility initiative that seeks to understand the role that genetic differences play in the variation of responses to environmental exposures.