This presentation focuses on the history and status of the preclinical toxicologic assessment of recombinant proteins. Cytokines and growth factors are used as examples. There has been an evolution of thought over the past dozen years on the testing of these substances that has ranged from their being considered nontoxic, human-specific proteins for which no predictive testing could be done in animals, to the present view that they can be toxic, relevant testing may be possible in animals, and assessment approaches should be science-based and case-by-case. The challenge of appropriate testing of recombinant proteins has caused toxicologists in both industry and in regulatory authorities to reconsider not only testing procedures but the purpose of preclinical assessment. The form of regulatory guidelines (guidance vs. inflexible testing protocols) has been questioned, and the need for interaction between scientists in industry and regulatory agencies has been strengthened. This process has advanced the science of toxicology.