Human plasma concentration-time profiles of troglitazone and troglitazone sulfate simulated by in vivo experiments with chimeric mice with humanized livers and semi-physiological pharmacokinetic modeling.

Troglitazone and its major metabolite troglitazone sulfate were intravenously administered to chimeric mice with different ratios of liver replacement by human hepatocytes. Total clearances were converted to hepatic intrinsic clearances normalized to their liver weight, with the assumption that extra-hepatic elimination of these compounds was negligible. These values were plotted against the replacement indices, and postulated values for virtual 100% chimeric mice were assumed to be equivalent to those in humans. Metabolic formation ratio was estimated by comparing AUCs of troglitazone sulfate after separate administration of troglitazone and troglitazone sulfate. Liver to plasma concentration ratios were obtained from direct measurement. These parameters were extrapolated to 100% chimeric mice and subjected to semi-physiological pharmacokinetic modeling using pharmacokinetic parameters for oral administration taken from literature. Our simulated plasma concentration-time profile of troglitazone agreed well with observed values obtained in clinical study. However, the profile of troglitazone sulfate was far below the reported values. Although the possible reasons for this discrepancy remains unsolved, the combination of chimeric mice with semi-physiological PK modeling proved to be a useful tool in understanding the function of each PK parameter in human pharmacokinetics of troglitazone and its conjugated metabolite.

Simulation of human plasma concentration-time profiles of the partial glucokinase activator PF-04937319 and its disproportionate N-demethylated metabolite using humanized chimeric mice and semi-physiological pharmacokinetic modeling.

1. The partial glucokinase activator N,N-dimethyl-5-((2-methyl-6-((5-methylpyrazin-2-yl)carbamoyl)benzofuran-4-yl)oxy)pyrimidine-2-carboxamide (PF-04937319) is biotransformed in humans to N-methyl-5-((2-methyl-6-((5-methylpyrazin-2-yl)carbamoyl)benzofuran-4-yl)oxy)pyrimidine-2-carboxamide (M1), accounting for ∼65% of total exposure at steady state. 2. As the disproportionately abundant nature of M1 could not be reliably predicted from in vitro metabolism studies, we evaluated a chimeric mouse model with humanized liver on TK-NOG background for its ability to retrospectively predict human disposition of PF-04937319. Since livers of chimeric mice were enlarged by hyperplasia and contained remnant mouse hepatocytes, hepatic intrinsic clearances normalized for liver weight, metabolite formation and liver to plasma concentration ratios were plotted against the replacement index by human hepatocytes and extrapolated to those in the virtual chimeric mouse with 100% humanized liver. 3. Semi-physiological pharmacokinetic analyses using the above parameters revealed that simulated concentration curves of PF-04937319 and M1 were approximately superimposed with the observed clinical data in humans. 4. Finally, qualitative profiling of circulating metabolites in humanized chimeric mice dosed with PF-04937319 or M1 also revealed the presence of a carbinolamide metabolite, identified in the clinical study as a human-specific metabolite. The case study demonstrates that humanized chimeric mice may be potentially useful in preclinical discovery towards studying disproportionate or human-specific metabolism of drug candidates.

Formation of the accumulative human metabolite and human-specific glutathione conjugate of diclofenac in TK-NOG chimeric mice with humanized livers.

3'-Hydroxy-4'-methoxydiclofenac (VI) is a human-specific metabolite known to accumulate in the plasma of patients after repeated administration of diclofenac sodium. Diclofenac also produces glutathione-conjugated metabolites, some of which are human-specific. In the present study, we investigated whether these metabolites could be generated in humanized chimeric mice produced from TK-NOG mice. After a single oral administration of diclofenac to humanized mice, the unchanged drug in plasma peaked at 0.25 hour and then declined with a half-life (t1/2) of 2.4 hours. 4'-Hydroxydiclofenac (II) and 3'-hydroxydiclofenac also peaked at 0.25 hour and were undetectable within 24 hours. However, VI peaked at 8 hours and declined with a t1/2 of 13 hours. When diclofenac was given once per day, peak and trough levels of VI reached plateau within 3 days. Studies with administration of II suggested VI was generated via II as an intermediate. Among six reported glutathione-conjugated metabolites of diclofenac, M1 (5-hydroxy-4-(glutathion-S-yl)diclofenac) to M6 (2'-(glutathion-S-yl)monoclofenac), we found three dichlorinated conjugates [M1, M2 (4'-hydroxy-3'-(glutathion-S-yl)diclofenac), and M3 (5-hydroxy-6-(glutathion-S-yl)diclofenac)], and a single monochlorinated conjugate [M4 (2'-hydroxy-3'-(glutathion-S-yl)monoclofenac) or M5 (4'-hydroxy-2'-(glutathion-S-yl)monoclofenac)], in the bile of humanized chimeric mice. M4 and M5 are positional isomers and have been previously reported as human-specific in vitro metabolites likely generated via arene oxide and quinone imine-type intermediates, respectively. The biliary monochlorinated metabolite exhibited the same mass spectrum as those of M4 and M5, and we discuss whether this conjugate corresponded to M4 or M5. Overall, humanized TK-NOG chimeric mice were considered to be a functional tool for the study of drug metabolism of diclofenac in humans.