o., which corresponds to approximately 600 mg/day in patients if corrected with the interspecies dose scaling factor[15]) did not result in significant increases in liver enzyme activity in the plasma or liver pathology, even after repeated dosing (up to 4 weeks).[16-18] Similarly, rats exposed to INH (400 mg/kg/day) for 1 week produced mild hepatic steatosis (which is usually not seen in DILI patients), but not hepatocellular necrosis.[17] A

recent study that analyzed a large number of biomarkers of both hepatic injury and activation of the immune system in rats LDK378 receiving combined INH and rifampicin found no evidence of liver injury and concluded that the rat was not a suitable animal model to replicate the delayed type of INH hepatotoxicity.[19] Finally, rabbits developed minimal hepatic toxicity (liver enzyme leakage) after repeated administration of INH (50 mg/kg p.o., followed by three doses of 35 mg/kg every 3 h for

2 days). To generate more significant enzyme leakage, pretreatment with phenobarbital was required,[20] which may have multiple effects by itself. Similar difficulties in recapitulating cell injury were encountered in vitro; for example, high concentrations (> 26 mM) of INH were required to induce apoptosis in HepG2 cells and human and murine lymphoma find more cell lines,[21] but more relevant concentrations (1–3 mM) did not produce lethal cell injury in cultured mouse hepatocytes.[18] Because INH therapy is often combined with other antitubercular drugs, e.g. rifampicin, a number of studies have aimed at developing animal models for this particular cotreatment. However, in rats or mice, even INH/rifampicin cotreatment did not

model the toxicity seen in humans. Rather, the cotreatment mildly increased the leakage of liver enzymes and caused vacuolization of centrilobular hepatocytes, and increased the number of apoptotic cells in the liver, without causing overt liver damage.[22, 23] Only when using human hepatocytes in culture, rifampicin potentiated INH toxicity.[24] This makes sense, as rifampicin has been known to exert its effects through the human PXR, a 上海皓元医药股份有限公司 nuclear receptor that not only is involved in the regulation of drug-metabolizing enzymes (including CYP3A), but also in regulating porphyrin metabolism.[25] In view of the fact that INH hepatotoxicity in patients is idiosyncratic (host-dependent), it is plausible that toxicity cannot be reproduced in normal, healthy animals. Only an animal model with a specific underlying predisposing condition (genetic, acquired) can likely mimic the situation of patients carrying these susceptibility factors (see below). However, even in mice strains with genetically impaired immune tolerance (Cbl-b−/− and PD1−/− mice), the clinical pattern of liver failure could not be recapitulated.[26] INH (Fig. 2) is a prodrug. In M.