Quantifying the effect of human interindividual kinetic differences on the relative potency value for riddelliine N-oxide at low dose levels by a new approach methodology.

Pyrrolizidine alkaloid N-oxides (PA-N-oxides) are predominant in plants and herbal foods, and are converted to pyrrolizidine alkaloids (PAs) upon consumption, leading to toxicity. The effect of interindividual kinetic differences on the relative potency values of PA-N-oxides compared to their PAs (REP_{PANO to PA}) was studied, with riddelliine N-oxide (RIDO) and riddelliine (RID) as model compounds. In vitro kinetic data measured in incubations with 30 fecal and 25 liver S9 donor samples showed high variation across individuals, where the interindividual variability was captured with Bayesian multilevel regression. The distributions of influential PBK model parameters were used as input for physiologically based kinetic (PBK) modeling combined with Monte Carlo (MC) simulations to calculate the probability distribution of REP_{RIDO to RID} values. At low dose levels, interindividual differences were shown to be a factor that influences the REP_{RIDO to RID} value while neither dose nor endpoint used plays a role. The distribution of the REP_{RIDO to RID} value ranged from 0.71 to 0.97 (95th percentile) with a mean value of 0.87. The approach described enables determination of interindividual REP_{PANO to PA} values at low dose levels, which are not accessible in in vivo experiments quantifying the REP_{PANO to PA}value.