Mechanism-based risk assessment strategy for drug-induced cholestasis using the transcriptional benchmark dose derived by toxicogenomics.

Cholestasis is one of the major causes of drug-induced liver injury (DILI), which can result in withdrawal of approved drugs from the market. Early identification of cholestatic drugs is difficult due to the complex mechanisms involved. In order to develop a strategy for mechanism-based risk assessment of cholestatic drugs, we analyzed gene expression data obtained from the livers of rats that had been orally administered with 12 known cholestatic compounds repeatedly for 28 days at three dose levels. Qualitative analyses were performed using two statistical approaches (hierarchical clustering and principle component analysis), in addition to pathway analysis. The transcriptional benchmark dose (tBMD) and tBMD 95% lower limit (tBMDL) were used for quantitative analyses, which revealed three compound sub-groups that produced different types of differential gene expression; these groups of genes were mainly involved in inflammation, cholesterol biosynthesis, and oxidative stress. Furthermore, the tBMDL values for each test compound were in good agreement with the relevant no observed adverse effect level. These results indicate that our novel strategy for drug safety evaluation using mechanism-based classification and tBMDL would facilitate the application of toxicogenomics for risk assessment of cholestatic DILI.

Assessment of hepatic function in rabbits with steroid-induced cholestatic liver injury.

Methyltestosterone (MT) or ethinyl estradiol (EE) was administered to adult rabbits for 20 weeks beginning with initial daily doses of 0.4 mg/kg MT and 0.015 mg/kg EE for three weeks, then these dosages were doubled at 3-week intervals to a maximum dosages 6.4 mg/kg and 0.24 mg/kg, respectively. Within 2 weeks, the serum gamma-glutamyltransferase activity of MT and EE treated rabbits was significantly greater than controls and increased progressively throughout the treatment period. Aspartate aminotransferase activity was also increased at 2 weeks and remained so for 17 weeks. Serum alkaline phosphatase was elevated at 2 weeks but thereafter was normal indicating that this enzyme is of no value in detecting steroid-induced hepatic dysfunction. Elevated serum bile acid concentration and prolonged BSP clearance indicated marked hepatic excretory dysfunction at higher dose levels. Histologic abnormalities were observed in the livers of both MT and EE treated rabbits. These lesions were more severe in the EE group in which there was marked bile duct proliferation, mononuclear cell infiltration of portal areas, and perilobular fibrosis. The studies indicate that the rabbit is susceptible to development of hepatic injury when receiving 17 alpha-alkyl substituted steroids and may be a useful animal model for investigations of the pathogenesis of steroid-induced cholestatic liver injury.