Manuel Haas

• Pyrrolizidine alkaloids (PAs) are secondary plant metabolites, which can occur as contaminants in predominantly plant-based foods. Only 1,2-unsatured PAs are bioactivated by cytochrome P450 (CYP450) enzymes, especially CYP3A4, into DNA reactive metabolites, which are known to be genotoxic and hepatoxic in vivo. The PA toxicity is strongly influenced by the chemical structure as demonstrated by several in vitro studies. However, quantitative genotoxicity data are needed, particularly in primary human hepatocytes (PHH). The organic cation transporter 1 (OCT1) has been identified as a key player in cellular uptake of cyclic diester, although little is known about a structure-dependent transport of PAs. A major objective of the thesis was the investigation of the relationship between PA structure and the toxicity of PAs with different degree of esterification in human liver models (HepG2-CYP3A4 and PHH). An in vitro genotoxicity battery, consistently demonstrated the DNA damage markers γH2AX and p53, as well as the alkaline comet assay, confirmed a structure-toxicity relationship in HepG2-CYP3A4 cells. The data were subject to benchmark dose (BMD) modeling to derive the genotoxic potential of each PA. BMD modeling yielded values in the range of 0.1–10 μM for most cyclic and open diesters, with monoesters displaying lower genotoxic potency. Cyclic diesters such as retrorsine and seneciphylline showed the highest genotoxic potential. Furthermore, cyclic and open diesters showed comparable cytotoxic potentials with effective concentrations at 50% cell viability (EC50) between 10 and 70 μM, with lasiocarpine at the top followed by seneciphylline. Besides, heliotrine and monocrotaline exhibited marked cytotoxic and genotoxic potentials, which were not comparable to other congeners within their degrees of esterification. Among the PA monoesters, only heliotrine was cytotoxic with an EC50 above 400 μM. Furthermore, heliotrine and monocrotaline exhibited significantly higher or lower toxic potentials, respectively, compared to PA congeners with the same esterification degree. Notably, the similar ranking was confirmed in PHH, with lasiocarpine exhibiting the highest genotoxic potential and approximately 3-4 times higher EC50 values for all PAs. The second objective of the thesis was to study the role of OCT1 in the transport of structurally different PAs, i. e. riddelliine (cyclic diester), lasiocarpine (open diester) and heliotrine (monoester), in metabolically competent human liver cell models (HepG2-CYP3A4 and PHH) and hamster fibroblasts. Using pharmacological inhibitors against OCT1, we observed that the OCT1-mediated uptake is crucial for PA-induced cytotoxicity and genotoxicity, which was independently on PA structure. Both cell models, HepG2-CYP3A4 and V79-CYP3A4 revealed strongly attenuated cytotoxicity upon OCT1 inhibition. Notable, the reduced OCT1 expression in V79-CYP3A4 cells correlated well with their reduced susceptibility to PA-induced cytotoxicity. PHH confirmed the results in combination with OCT1 inhibition. Furthermore, OCT1 inhibition reduced γH2AX and p53 levels, indicating less genotoxic stress. Consistently, OCT1 inhibition suppressed the DDR activation, as indicated by decreased checkpoint kinase phosphorylation during PA exposure. Our findings strongly support the concept of grouping PAs into potency classes for risk assessment. Furthermore, the metabolic activation by CYP3A4 as well as OCT1 as uptake transporter play a major role for PA toxicity.