A selenium-deficient Caco-2 cell model for assessing differential incorporation of chemical or food selenium into glutathione peroxidase.

Assessing the ability of a selenium (Se) sample to induce cellular glutathione peroxidase (GPx) activity in Se-deficient animals is the most commonly used method to determine Se bioavailability. Our goal is to establish a Se-deficient cell culture model with differential incorporation of Se chemical forms into GPx, which may complement the in vivo studies. In the present study, we developed a Se-deficient Caco-2 cell model with a serum gradual reduction method. It is well recognized that selenomethionine (SeMet) is the major nutritional source of Se; therefore, SeMet, selenite, or methylselenocysteine (SeMSC) was added to cell culture media with different concentrations and treatment time points. We found that selenite and SeMSC induced GPx more rapidly than SeMet. However, SeMet was better retained as it is incorporated into proteins in place of methionine; compared with 8-, 24-, or 48-h treatment, 72-h Se treatment was a more sensitive time point to measure the potential of GPx induction in all tested concentrations. Based on induction of GPx activity, the cellular bioavailability of Se from an extract of selenobroccoli after a simulated gastrointestinal digestion was comparable with that of SeMSC and SeMet. These in vitro data are, for the first time, consistent with previous published data regarding selenite and SeMet bioavailability in animal models and Se chemical speciation studies with broccoli. Thus, Se-deficient Caco-2 cell model with differential incorporation of chemical or food forms of Se into GPx provides a new tool to study the cellular mechanisms of Se bioavailability.