Research Grants 1997-1998

A Human Model for ACD

Jurij J. Hostynek, PhD
Euroamerican Technology Resources, Inc., Lafayette, California

Proposed overall is the development of a quantitative structure-activity relationship (QSAR) model which will predict the potential and potency for delayed type contact hypersensitivity (ACD) reactions upon skin (or systemic) contact with new or untested small molecular weight nonelectrolytes in man.

Prior to large scale development, production and commercialization of new chemicals, their potential to cause hypersensitivity disease upon dermal or systemic exposure requires investigation as part of the overall assessment of their potential for adverse health effects. In order to properly assess such risks to humans, identification of potential sensitizers currently involves the use of laboratory animals, mostly the guinea pig. However, conventional animal test methods used for risk assessment purposes are time consuming and costly, requiring a relatively large number of animals, and rely on a subjective evaluation of clinical results.

The discriminant technique of two-valued multiple regression analysis of molecular descriptors is known to successfully distinguish between allergens and non-allergens. Its application to analysis of clinical human sensitization data in particular will lead to a similar algorithm predictive of human sensitization potential that, if proven successful, will make traditional animal testing superfluous, ultimately taking animal models out of the testing loop.

In Phase II the predictive value of the model will be tested by analyzing randomly selected structures for their sensitizing ability, which had been set aside from the learning set for the purpose of validation. In Phase III, the predictive value of the QSAR model will be raised from a nes/no classification acheived in Phase I and confirmed in Phase II to that of a quantitative tool for the assessment of allergenic potency of small molecular weight chemicals. Through better resolution of clinical data on record, the relative potency of known allergens in the model will be categorized in distinct levels of ranking (non, weak, moderate, strong) and analyzed by Rank Transform Regression Analysis.