Structure-activity model of chemicals that cause human respiratory sensitization

Abstract

We report a structure-activity model of chemicals with the potential to cause respiratory allergy developed through the CASE/MultiCASE systems. Chemicals documented to elicit a decrease in FEV1 of > or = 20% within 24 h of inhalation provocation challenge were used to form a learning set. Additional requirements for inclusion in the learning set were that chemicals had at least two contiguous nonhydrogen atoms and were nonmetallic. Forty chemicals met these criteria. The model identified several "structural alerts" including the isocyanate functionality (N = C = O), primary and secondary amines, substituted aromatic moieties, and distance descriptors. An external data-withholding exercise used to validate the model yielded a sensitivity of 0.95 and a specificity of 0.95. This model is applicable to initial prediction of the sensitizing ability of untested chemicals and may provide mechanistic insight into the process(es) of respiratory sensitization.