Biologically Based Dose-Response Modeling in Developmental Toxicology: Biochemical and Cellular Sequelae of 5-Fluorouracil Exposure in the Developing Rat

Abstract

Mechanistically based dose-response models for developmental toxicity require elucidation of critical biological events that intervene between maternal exposure and adverse developmental outcome. We have examined some of the major events in the rat embryo/fetus following a subcutaneous injection of 5-fluorouracil (5-FU; 0-40 mg/kg) to the dam on Day 14 of gestation. This treatment resulted in reduced fetal weight that was significant at doses of 20 mg/kg and higher, generalized reduced ossification at doses above 25 mg/kg, and wavy ribs at doses of 30 mg/kg and higher. Numerous malformations including cleft palate and hindlimb defects were substantially increased at doses of 35 and 40 mg/kg. 5-FU inhibits thymidylate synthetase (TS), resulting in inhibited growth of rapidly proliferating tissues. To identify early events in the pathogenesis of hindlimb defects, we examined the effects of 5-FU on TS activity, cell cycle, growth, and morphology in the developing hindlimb as a function of dose and time. The rate of decline of TS activity following 5-FU exposure was dose related, although maximal inhibition and recovery were similar at doses within (20 and 40 mg/kg) and below (10 mg/kg) the range of detectable developmental toxicity. Flow cytometric analysis of nuclei from embryonic hindlimbs revealed a transient increase in the percentage of cells in S phase and decrease in G 0/G 1 phase 8 hr after maternal injection of 20-40 mg 5-FU/kg, but not at lower doses. Reduction in growth and morphometric changes of hindlimbs were observed only after maternal exposure to 40 mg/kg. The tissue specificity of these effects was examined by comparing the hindlimb with other embryonic tissues. There was also a dose-related decline of TS activity in the embryonic liver. However, the pattern of recovery of TS activity and cell cycle alterations were different in the liver than in the hindlimb, probably reflecting the higher cell proliferative rate in the liver at this stage. We have derived a quantitative, empirical model for induction of hindlimb defects based on TS inhibition and subsequent cellular events following 5-FU exposure. The model predicted a dose response similar to that of the observed data although the predicted curve was shifted toward lower doses. These results suggest that while this model may not capture all of the critical events involved in the induction of hindlimb defects following maternal exposure to 5-FU, it does reflect a central mechanism of its developmental toxicity. Biologically based dose-response modeling provides a framework for testing mechanistic hypotheses, and developing such models should ultimately improve our ability to perform risk assessments.