Biologically based pharmacodynamic models: tools for toxicological research and risk assessment.

The cancer model contains extensive biological detail, but is not fully validated. It is extremely useful for organizing what is known about malignant transformation, for linking measures of tissue dose with biological processes, and for proposing new experiments. It became very clear, as the cancer model was developed, that we lack an adequate understanding of the details of biochemical mechanisms of carcinogen action linking the target-tissue dose of carcinogen with effects on cell birth, death, and mutation rates as specified by the MVK model. This problem extends to noncarcinogenic toxicants such as methotrexate and the organophosphates. In these cases the biologically based descriptions of toxicity stop at the first interaction of the chemical with tissue. The linkage from the primary interaction to ultimate toxic effect is a "black box" that, at present, can only be described empirically rather than in biologically based terms. The models for dioxin and prednisolone do illustrate how biological events several steps removed from the initial tissue interaction of the toxicant can be described in quantitative, biologically based terms. As such they point the way for future development of more realistic descriptions in which the use of purely empirical linkages is reduced. Rigorous organization of current knowledge and specification of a logical sequence of new experiments encompass the major reasons for developing biologically based (structural) descriptions of any process. It bears emphasis that biologically based tissue response models, for cancer, for chloroform cytotoxicity, for functional impairment following organophosphate exposure, etc, are in reality quantitative formulations of an hypothesis or proposed mechanism incorporating the pharmacokinetics of the chemical and the linkage between target-tissue dose and toxic effect(s). The quantitative model allows us to see if the hypothesis is consistent with pertinent data and aids in designing experiments to define, refute, or bolster the general nature of the proposed linkages between tissue dose, tissue response and the whole organism's response to the chemical exposure. In this sense these biologically based models are quantitative descriptions of the current level of understanding of toxic effects in animals. We began this review article by looking at the literature on empirical tissue-response models to educate ourselves on the history of PD modeling. We rapidly found that the clear distinction between PK and PD models described in the literature blurs when one moves to biologically based modeling of tissue disposition and tissue response. Many articles on physiologically based pharmacokinetics turn out on close scrutiny to be as closely related to tissue response as they are to kinetics.(ABSTRACT TRUNCATED AT 400 WORDS)