Lukas Rutz

• Pyrrolizidine alkaloids (PA) are secondary plant metabolites occurring in a great many plant species worldwide, known to exhibit hepatotoxic, genotoxic and carcinogenic properties after metabolic activation. In recent years, contamination of food, feed and herbal medicines with PA has become an increasing problem. The concept of interim relative potency factors (iREP) proposed by Merz and Schrenk in 2016 was a new approach for risk assessment of PA. While existing approaches of risk assessment assumed equivalent toxic potency for all PA congeners, the approach of Merz and Schrenk considered the structural features of individual PA congeners based on existing data from the literature. In order to generate further data on the structure-specific toxicity of PA, congeners of different structural classes were investigated in different in vitro test systems. In vitro cytotoxicity was investigated in primary rat hepatocytes, HepG2 C9 cells (overespressing human CYP3A4) and naïve HepG2 cells. Overall, it could be observed that lasiocarpine and the cyclic di-esters (except monocrotaline) showed much stronger cytotoxic effects in comparison to the tested mono-esters in both, primary rat hepatocytes and in HepG2 C9 cells. Primary hepatocytes were the most sensitive cells investigating cytotoxicity of different PA congeners, followed by the HepG2 C9 cells. This is confirmed by markedly higher metabolism rates for all investigated PA in primary rat hepatocytes determined in the metabolism experiments. In naïve HepG2 cells no cytotoxic effects could be observed. The influence of cytochrome P450 (CYP) on the formation of toxic metabolites seem to play a crucial role. This assumption could be beared using ketoconazole as CYP inhibitor and testing various pre-incubation times in primary rat hepatocytes. CYP activity was measured using 7-Benzoxyresorufin-O-Dealkylase (BROD) assay in primary rat hepatocytes and in HepG2 C9 cells. Glutahione (GSH) depletion using buthionine sulfoximine (BSO) showed slight stronger cytotoxic effects for several PA, but not for all tested. In contrast to the negative results of mutagenicity in ames fluctuation assay using Salmonella strains TA98 and TA100 with and without metabolic activation by S9 mix, all tested PA congeners showed micronuclei induction in the HepG2 C9 cell line. Again, laisocarpine and the cyclic di-esters (except monocrotaline) were the most potent ones. In conclusion, the data from cytotoxicity and genototoxicity experiments from the tested PA congeners confirm published iREP factors with a few exceptions, in particular for monocrotaline or echimidine. Additionally, metabolism of six selected PA was studied in primary rat hepatocytes and HepG2 C9 cells. Genrally, it was found that almost all tested cyclic and open-chained di-esters (except retrorsine) showed much higher metabolism rates in both cell types, in comparison to the mono-esters, for which only low metabolism rates could be measured. The same was observed for the quantified amounts of reactive metabolites in the supernatants of both cell types. In general, also these data bear the results from cytotoxicity and genotoxicity experiments and help to better understand the complex metabolism and the structure-specific toxicity of different PA congeners.