The Center for Alternatives to Animal Testing is an academic center affiliated with the Division of Toxicological Sciences in the Department of Environmental Health Sciences of the Johns Hopkins University Bloomberg School of Public Health.

 

Johns Hopkins School of Public Health

Abstract for TestSmart--A Humane and Efficient Approach to Screening Information Data Sets (SIDS) Data

Correlation between Daphnia Magna EC50 and Fathead Minnow LC50 values

Kurt Enslein
Health Designs, Inc.

Some years ago we examined relationships between daphnia magna and rat oral toxicity, and published 2 papers as a result of this effort (QSAR in Environmental Toxicology II, 91-106, 1987 KLE Kaiser, ed., and ACS Special Technical Publication 1007, 397-409, 1989). We found a strong relation between the toxicity endpoints of these two phylogenetically disparate species.

As a result of the interest in the HPV chemical program to minimize the number of vertebrate animals, we asked whether there existed a similar relationship between daphnia magna and fathead minnow toxicity. We had previously developed separate QSAR models for both of these species. The databases contained 252 and 617 chemicals, respectively. We therefore intersected the 2 databases, and were left with 93 chemicals for which data appeared in both sets.

We thereupon examined this subset to determine what correlations existed between the inverse molar daphnia magna EC50 and fathead minnow LC50 values.

All 93 chemicals: the correlation was 0.835, thus one species explaining approximately 70% of the variance in the other. We noted a number of outliers, i.e. chemicals that were either very toxic in daphnia magna or fathead minnow, or both. All 6 chemicals were agricultural chemicals: endosulfan, endrin, parathion, fenvalerate, methyl parathion, and malathion.

Acyclic chemicals: After removing the 6 outliers, the correlation had not changed substantially. We then determined whether certain chemical classes were predominant in the remaining 87 compounds. We found that the great majority were either acyclics (N=34) or substituted single benzenes (N=43). The correlation for the acyclics was 0.938, thus the variance explained was 88%. Most of the compounds in this subset probably acted by the narcosis mechanism.

Single benzenes: The correlation for these 43 compounds was 0.55, thus explaining only 30% of the variance.

Discussion

It is clear that the correlations could be considerably improved by the development of a set of QSAR models for each chemical class, as well as classes not substantially represented in the 93 compounds. It would be necessary to obtain additional daphnia magna data, probably requiring additional assays. We are experimenting with the use of our current daphnia magna model to determine whether using estimated values could be a substitute for assay data. However, we feel somewhat uncomfortable with deriving a QSAR model based on data derived from another QSAR model.