Investigating the Mechanisms of Action and the Identification of Breast Carcinogens by Computational Analysis of Female Rodent Carcinogena

Abstract

Public reporting burden for this collection of inforrmtion Is estimated to average I hour per response, induding the time for reviewing Instructions, searching existing data sources, gathering and mnaintalning the data needed, and completing and reviewing this collection of information. Send comnents regarding this burden estimate or any other aspect of this collection of information, indclng suggestions for reducing this burden to Washington Headquarters Services, Directorate for Infornation Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, AUTHOR(S)Albert R. Cunningham, Ph.D. PERFORMING ORGANIZA TION NAME(S) AND ADDRESS(ES)8 ABSTRACT (Maximum 200 Words)This project is investigating the potential that environmental estrogens may be involved in the etiology of breast cancer.We hypothesize that specific features of chemicals can be identified that are significantly associated with female and breast carcinogens and that these features are related to mechanisms of chemical carcinogenesis.Our overall scientific objective is to investigate the hypothesized relationship between environmental chemicals, xenoestrogens, and the development of breast cancer. With the success of the rat and mouse mammary carcinogen models we are preparing two manuscripts for publication. We are also pursuing work on a general chemical carcinogen manuscript and a one describing female-specific carcinogens.Also of importance, we are working on several xenoestrogen models that, although not detailed in the project proposal, will be of great importance for understanding the endocrine disruptor link to breast cancer.We have also developed a new structure-activity relationship program called cat-SAR that is producing predictive and mechanistically insightful models of mammary carcinogens.Looking forward I see no obstacles to the successful completion of. this project in a timely manner. During the early part of the project it was becoming evident that MCASE was not developing models that were of stellar predictivity. On account of successful modeling with SIMCA (soft independent modeling of class analogy) of aromatic amine Salmonella mutagens and skin sensitizing agents for a project supported by Proctor & Gamble, we spent some time investigating whether SIMCA models could be employed to produce adequate models relating to this project.We originally thought that SIMCA combined with HQSAR (hologram quantitative SAR) models appeared to be superior to MCASE models. The HQSAR-SIMCA approach utilized categorical biological data (i.e., carcinogen vs. non-carcinogen) and molecular fragments as SAR descriptors. Therefore, this seemed a reasonable substitute SAR approach for MCASE. However, upon consultation with the makers of Sybyl HQSAR-SIMCA, we learned that there was a large degree of random assignment of SAR descriptors. Basically, as it turned out, although the modeling software was able to produce models that could predict the activity of unknown clemicals-the...