(536c) Impact of Nutritional Stress and Liver Metabolic State On Xenobiotic Drug Transformation and Hepatotoxicity

Approximately half of all acute liver failures in the United States today are caused by drug-induced liver injury. Drug-induced hepatotoxicity is also cited as the most frequent cause of post-market withdrawal of an approved drug. A growing elderly population depends on multiple drugs for managing chronic illnesses, increasing the likelihood for harmful drug interactions. Purely empirical approaches for toxicity screening of multiple drug combinations will further raise the experimental cost burden of drug development and pre-clinical testing. Moreover, toxicity predictions based on animal models are limited in that there are large variances in the expression profiles of liver enzymes between species. Isolated human hepatocytes reflect the parenchymal fraction of the liver principally responsible for xenobiotic transformation and thus may be used as a surrogate for the organ. On the other hand, conventional in vitro toxicity screens focusing on a viability marker and a few xenobiotic transformation enzymes fail to take into account individual differences in nutritional status and metabolic enzyme profiles. There is substantial evidence that drug-induced liver injury can be traced to bioactivated electrophilic drug conjugates or other reactive intermediates that bring about covalent modification of cellular macromolecules . Reduced glutathione (GSH), a nucleophilic tripeptide, plays a key role in detoxifying these intermediates, thus coupling amino acid metabolism and xenobiotic transformation. Regenerating GSH from the oxidized form (GSSG) requires NADPH, a cofactor supplied by the pentose phosphate pathway and/or malate cycle. These substrate and cofactor dependencies suggest that the metabolic state of the liver can significantly impact the extent of drug biotransformation and the toxicity of the derived intermediates. The goal of our study is to investigate the hypothesis that nutritional stresses and deficiencies in metabolic enzymes (e.g. mutations that lower activity) are contributing factors in idiosyncratic adverse drug reactions.

In this study, we develop a framework to identify the metabolic burden of a drug based on in vitro flux data. The flux data were derived from metabolite profiles collected under different nutritional states generated by varying the concentrations of the major nutrients (glucose and amino acids) and hormones in the culture medium. The main test drug was Troglitazone (TGZ), an anti-diabetic agent withdrawn from clinical use due to reports of idiosyncratic hepatotoxicity. The drug-induced burden on each metabolic enzyme was quantified as the fractional change in flux relative to capacity flux. Capacity flux for each enzyme was estimated through linear programming with both stochiometric and thermodynamic constraints. Metabolic burdens estimated at various times after initial drug exposure were correlated with eventual cytotoxicity using a multiple regression algorithm to identify metabolic burdens that are significant predictors of toxicity. Results to date suggest that exposure to sub-toxic levels of TGZ directly impacts the metabolic flux distribution in cultured hepatocytes. The TGZ concentration threshold for cytotoxicity as well as the calculated metabolic burden varied with the nutrient composition of the culture medium, consistent with our hypothesis. Ongoing work examines the metabolic burden of drug combinations.