RESUMEN
Mn is an essential element that causes neurotoxicity in humans when inhaled at high concentrations. This metal has well-recognized route-dependent differences in absorption, with greater proportionate uptake for inhalation versus dietary exposure. Physiologically-based pharmacokinetic (PBPK) models for Mn have included these route specific differences in uptake and their effect on delivery of Mn to target tissues via systemic circulation. These PBPK models include components describing ingestion and inhalation, homeostatic control (concentration dependent biliary elimination and gastrointestinal absorption), and delivery to target sites within the brain. The objective of this study was to combine PBPK modeling of target tissue Mn concentration and categorical regression analysis to identify Mn intake levels (both by food and air) that are expected to cause minimal toxicity. We first used the human PBPK model to describe blood Mn data from three occupational exposure studies, demonstrating consistency between model predictions and measured data. The PBPK model was then used to predict concentrations of Mn in the globus pallidus (the presumed target tissue for motor function disruption in humans) for various epidemiological studies. With the predicted globus pallidus concentration of Mn, we conducted categorical regression modeling between globus pallidus Mn and severity-scored neurological outcome data from the human cohorts. This structured tissue dose - response analysis led to an estimated 10% extra risk concentration (ERC10) of 0.55µg/g Mn in the globus pallidus, which is comparable to similar values estimated by the Agency of Toxic Substances and Disease Registry and Health Canada (after translation from external exposure to tissue dose). The steep dose-response curve below this ERC10 value may be used to inform the choice of adjustment factor to translate the ERC10 as a point of departure to a reference concentration for occupational or environmental exposure to Mn. Because these results are based on human epidemiological data and a human PBPK model, adjustment or translation of results from animals to humans is not required.