PLoS Comput. Biol. 8, e1002750 (2012)

Metabolism is often considered only in terms of intracellular biochemical networks or across a complex organism, but a full understanding of human metabolism requires integration of these extremes. Toward this end, Krauss et al. explore how organismally focused, physiologically based pharmacokinetic (PBPK) models and genome-scale metabolic network models at the cellular scale could best be merged. Considering that PBPK models are often parsed into vascular, interstitial and intracellular spaces, the authors identified the intracellular space as a natural point of contact between the two models. The authors defined two ways in which a compound could affect the models: In indirect coupling, drugs or other inhibitors act as regulatory elements, influencing cellular metabolism without changing the availability of the compound in the PBPK model. In direct coupling, molecules from the PBPK model serve as substrates in the metabolic network model, causing changes to both. The integrated model was successfully benchmarked against multiple data sets. As one example, the authors modeled the consequences of the purine analog allopurinol, used to treat high concentrations of uric acid; their conclusions demonstrated good agreement with long-term allopurinol dosing, even though the data used as input was limited to a single dose. These results highlight the ability of the merged models to provide mechanistic insights into complex disease states.