Provisional relative potency factors for pyrrolizidine alkaloids – scientifically justified?

Background:

Pyrrolizidine alkaloids (PAs) may occur in phytomedicinal preparations, food and feed, either as natural constituents of the plants used or as cross-contaminants from weeds etc. Some 1,2-unsaturated PAs exert toxic and carcinogenic effects, primarily in the liver, which depend on metabolic activation steps catalyzed by cytochrome P450 (CYP) monooxygenases. Currently, due to a lack of detailed data, all 1,2-unsaturated PA congeners are considered equally potent to the presumably most toxic congeners, riddelliine (or lasiocarpine). This approach is over-protective because it is well established that many PA congeners are less or considerably less toxic than these 'reference compounds'.

Objectives:

In order to quantify potential risks and translate the connection between structure and toxicity into practice, we generated data on cytotoxicity and genotoxicity of relevant PAs in relevant cell culture models. In particular, we wanted to expand the database for the evaluation of recently published in vivo-based interim Relative Potency (iREP) Factors [1] designed to overcome the default assumption of uniform toxicity of all PAs.

Methods:

Cytotoxicity of PAs was assessed in primary rat hepatocytes and in CYP3A4-overexpressing human HepG2 (HepG2-CYP3A4) cells. Micronuclei formation was measured in HepG2-CYP3A4 cells only.

Results:

It was found that all 1,2-unsaturated PAs tested are cytotoxic in the cell models used, although the relative toxicities vary widely in a congener-specific manner [2]. The suggested iREP factors provide a good basis for describing the relative toxicities while the in vitro potencies of a few congeners were overestimated. This discrepancy is possibly due to toxicokinetic factors not fully reflected in our models. Micronuclei formation in HepG2-CYP3A4 cells was in good accordance to cytotoxicity pointing to a common key event in both adverse outcomes. Further studies are aimed at replacing additional animal experiments but allowing a congener-specific risk assessment of human exposure to PAs.

Acknowledgement:

We would like to thank the Deutsche Forschungsgemeinschaft and the Kooperation Phytopharmaka for their support of these studies.

References:

[1] Merz KH, Schrenk D. Toxicol Lett 2016, 263: 44 – 57

[2] Gao et al. Food Chem Toxicol 2019, in press