Hepatocyte culture systems are being developed for applications in bioartificial liver development and in vitro toxicology. However, when cells are removed from their host and put into culture, the output of critical functional markers is quite lower than those observed in vivo or predicted based on theoretical models of cellular capabilities. We seek to develop models of hepatocyte metabolism that will allow us to optimize the functionality of hepatocytes in these cell-based devices.

A major focus of our efforts has been on the application of optimization and pathway analysis techniques to elucidate what reactions and pathways are being used in culture and how we might intervene to redirect the metabolism in a more desired direction. Our initial work focused on central metabolism in hepatocyte cultures. First, linear programming was applied to determine the optimal flux distribution of the hepatic metabolism given bounds encompassing a range of physiological states. Further analysis of the optimal solution using a mixed integer linear programming framework revealed the existence of eight different flux distributions that each produces the same optimal value of urea. The flexibility of the network to utilize alternative pathways as well as the importance of different reactions was analyzed using a bi-level programming model, where at the upper level the optimization is performed to minimize the changes from the native flux distribution. Pathway analysis and logic-based programming were used to elucidate the relative importance of individual pathways and reactions within the overall network by clarifying the dependencies among the various pathways and metabolites.

We have recently begun to explore the interactions between central metabolism and xenobiotic metabolism, which has been modeled less extensively than the major anabolic and catabolic pathways. We reconstructed a hepatic detoxification network by adding to previous descriptions of central metabolism reactions corresponding to acetaminophen and ethanol detoxification pathways. The combined network consists of 106 metabolites and 121 reactions, and pathway analysis was performed using the FluxAnalyzer software. The reaction network included hepatic, APAP, and ethanol metabolism and the synthesis of conjugative species. Pathway analysis revealed the impact of APAP on the overall hepatocyte metabolic network including its effects on amino acid uptake, cysteine biotransformation, ATP synthesis and ethanol metabolism. We were able to show the importance of metabolite (cysteine) and ATP availability in preventing the formation of the oxidative toxic intermediate NAPQI by the synthesis of PAPS, UDP-glucuronic acid and glutathione. The metabolism of ethanol was also shown to be ATP dependent; this additional cellular demand contributes to the depletion of detoxifying species, thus leading to increased hepatotoxicity. Lastly, our results identified cysteine to be a critical amino acid in detoxification, which coincides with the current treatment of acetaminophen overdose, administration of N-acetylcysteine as a source of cysteine to aid in glutathione synthesis.

A kinetic model of APAP elimination was developed by compiling experimental data on molecular rate processes from animal, liver microsomes, and cellular studies. Dynamic modeling of APAP metabolism illustrated the formation of toxic metabolites in a dose-dependent fashion and the requirement for conjugation species. The time scale for onset of hepatotoxic effects was elucidated from the dynamics of depletion of conjugation species. In addition, we were able to show how fasting increases an individual's susceptibility to acute liver failure when taking equivalent drug doses.

Extended Abstract Status: Not Uploaded