Analysis of elimination half-lives in MamTKDB 1.0 related to bioaccumulation: Requirement of repeated administration and blood plasma values underrepresent tissues

When building the novel public mammalian toxicokinetic database (MamTKDB) we collected and included 3927 elimination half-lives (^elimt_1/2) for 1407 xenobiotics in various species (rat, human, mouse, dog, monkey, rabbit, cattle, pig, sheep, guinea pig, hamster, horse and goat) with specification of compartment (whole body, organ/tissue, cell type, medium) studied. Here we describe and analyse the collected data in MamTKDB 1.0. Most ^elimt_1/2 are for humans and rats and their data differ in some ways: whereas the rat data are mainly for pesticides, the human data are mainly for pharmaceuticals and environmental contaminants. There are also differences in types of compartments studied and in metabolites followed: human ^elimt_1/2 are mainly whole body based (i.e. based on blood plasma or excretion), animal data are additionally for various organs/tissues, cells or media. Contrary to human studies, animal studies regularly administrate radiolabeled (e.g. ¹⁴C) substances and distribution of both parent and eventual metabolites are followed, measuring the radioactivity. In rats, substances had been given through single, preconditioning or repeated administration. Single administration studies dominated, but repeated studies generally had longer ^elimt_1/2 than single or preconditioning studies for which ^elimt_1/2 were similar. Repeated administration studies should better ascertain steady state conditions throughout the body, a process involving time-dependent tissue loading, and the data show that for most substances, repeated studies are required to address bioaccumulation potential. About 65% of the substances in MamTKDB 1.0 fulfilled the octanol-water and octanol-air partitioning-based screening criteria (log K_ow > 2 and log K_oa > 5) for further bioaccumulation assessment and/or testing, and most of the substances with long ^elimt_1/2 in both humans and rats fulfill these criteria. Of note, however, there are also many chemicals with log K_ow > 2 with intermediate or short ^elimt_1/2. Per- and polyfluoroalkyl substances (PFAS) stand out in that they often have log K_oa < 5. Rats are poor toxicokinetic test models for perfluoroalkyl acids (PFAAs) for which pigs (and possibly mice) ^elimt_1/2 data resemble those of humans better. Perfluorinated carboxylic acids (PFCAs) and perfluorinated sulfonic acids (PFSAs) of similar molecular weight had similar ^elimt_1/2 in the species tested. For polychlorinated biphenyls (PCBs), ^elimt_1/2 increases with the degree of chlorination in humans. In relation to other compartments, blood plasma/serum had among the shortest ^elimt_1/2 in rats and often underrepresent ^elimt_1/2 in tissues. Rat data were divided into 38 compartment (tissue or media) types out of which 20 had sufficient data for correlational tests. In general, there was a strong degree of correlation of rat ^elimt_1/2 in-between most compartments, but there were also exceptions. Surprisingly, the correlation between brain and white fat was relatively weak. Interestingly, several substances or their metabolites bound to haemoglobin in red blood cells. MamTKDB 1.0 allows investigation on how certain chemical characteristics influence ^elimt_1/2 and is a promising database for assessment of bioaccumulation potential.