A novel humanized mouse lacking murine P450 oxidoreductase for studying human drug metabolism

Abstract

Only one out of 10 drugs in development passes clinical trials. Many fail because experimental animal models poorly predict human xenobiotic metabolism. Human liver chimeric mice are a step forward in this regard, as the human hepatocytes in chimeric livers generate human metabolites, but the remaining murine hepatocytes contain an expanded set of P450 cytochromes that form the major class of drug-metabolizing enzymes. We therefore generated a conditional knock-out of the NADPH-P450 oxidoreductase (Por) gene combined with Il2rg ^{- /-} /Rag2 ^{- /-} /Fah ^{- /-} (PIRF) mice. Here we show that homozygous PIRF mouse livers are readily repopulated with human hepatocytes, and when the murine Por gene is deleted (<5%), they predominantly use human cytochrome metabolism. When given the anticancer drug gefitinib or the retroviral drug atazanavir, the Por-deleted humanized PIRF mice develop higher levels of the major human metabolites than current models. Humanized, murine Por-deficient PIRF mice can thus predict human drug metabolism and should be useful for preclinical drug development.Human liver chimeric mice are increasingly used for drug testing in preclinical development, but express residual murine p450 cytochromes. Here the authors generate mice lacking the Por gene in the liver, and show that human cytochrome metabolism is used following repopulation with human hepatocytes.

Fig. 1

Generation of the PIRF strain and deletion of the murine P450 (Por) oxidoreductase. a Schematic representation of deleted and transgenic loci in the PIRF strain. b qPCR of Por mRNA 7 days after intravenous injection of adenovirus expressing the CRE recombinase (Adeno-Cre). c Immunostaining for Por protein demonstrating a gradient across the hepatic acinus with higher pericentral (cv) and lower periportal (pv) expression. Seven days after injection with Adeno-Cre, Por is barely detectable. d Western blotting confirms almost complete disappearance of Por protein 7 days after injection. Results are expressed in mean values ± s.e.m. of triplicates (N = 3). Scale bar 50 μm. PIRF, Por ^c/c /Il2rg ^−/− /Rag2 ^−/− /Fah ^−/−

Fig. 2

Deletion of Por with an adenoviral vector expressing CRE recombinase in PIRF mice. a 2 weeks after injection of the adenovirus (2.2 × 10¹⁰pfu per mouse) hepatocytes start to accumulate lipids. b H&E stain of liver lobe 4 weeks after transduction with adenovirus showing macro and microvesicular steatosis. Right picture is a higher magnification of boxed area on left. c Clonal expansion of Por expressing hepatocytes in PIRF mice. Mice were injected intravenously with adenovirus expressing Cre recombinase to delete the floxed Por gene. Scale bar 50 μm. PIRF, Por ^c/c /Il2rg ^−/− /Rag2 ^−/− /Fah ^−/−

Fig. 3

Gene expression profiling of humanized PIRF mice following deletion of Por. a Immunostaining of humanized PIRF and FRG mice for murine Por (mPor) and human nuclei (hNuc) after injection of Adeno-Cre (2.2 × 10¹⁰ pfu/mouse) once (1x) or twice (2x). Counterstaining in merged picture using DAPI. b Experimental outline for murine and human transcriptomics from chimeric livers with or without Por deletion. Chimeric mice (N = 7) injected with either Adeno-Cre or Adeno-GFP (2.2 × 10¹⁰ pfu/mouse) were euthanized after 1 week and liver tissue from all seven lobes was analyzed. Comparison of murine c and human d cytochromes originating from the same chimeric livers. e Gene expression of the main drug metabolizing human cytochromes in humanized, Por-deleted PIRF (Hu-PIRF 2x) mice and originating isogenic human hepatocytes (N = 2 donor hepatocytes and 4 isogenic humanized mice). Gene expression has been normalized to four human housekeeping genes and their murine counterparts (PSMB2, PSMB4, RAB7A, and VPS29; Psmb2, Psmb4, Rab7, and Vps29). Mean with range are given in c–e. FRG, Fah ^−/− /Rag2 ^−/− /Il2rg ^−/−; PIRF, Por ^c/c /Il2rg ^−/− /Rag2 ^−/− /Fah ^−/−

Fig. 4

Xenobiotic metabolism in humanized PIRF mice. Mass spectrometry analysis of murine feces within 16 h after intravenous injection of gefitinib a cancer therapeutic. a Selection of abundant and decreased gefitinib metabolites upon murine P450 oxidoreductase (Por) deletion in non-humanized PIRF mice (N = 3). b Gefitinib metabolites and known biotransformation. c Experimental setup: not-humanized PIRF mice and Humanized (Hu) FIRF and FRG mice were injected once (1x, after transplantation of human hepatocytes) or twice (2x, before transplantation and after reaching high human chimerism) before doing drug studies. All experimental groups contain three animals, if humanized, with high (>70%) and similar human chimerism (see Methods). Murine Por-deleted, human liver chimeric PIRF (Hu-PIRF 2x) mice show the most abundant human metabolite, M4 d, and the human-specific metabolite M28 e of gefitinib (N = 3). f Mass spectrometry analysis of PIRF liver homogenates 30 min after injection of atazanavir, a retroviral therapeutic. A major human metabolite (M15) is shown. The overall abundance of metabolites from atazanavir or gefitinib was set as 100% in each sample. The data are expressed as mean ± s.e.m. *P < 0.05 using non-parametric Mann–Whitney test. FRG, Fah ^−/− /Rag2 ^−/− /Il2rg ^−/−; PIRF, Por ^c/c /Il2rg ^−/− /Rag2 ^−/− /Fah ^−/−