Human Hepatocytes Can Repopulate Mouse Liver: Histopathology of the Liver in Human Hepatocyte-Transplanted Chimeric Mice and Toxicologic Responses to Acetaminophen

Abstract

A human hepatocyte-transplanted chimeric mouse has been established by transplantation of human hepatocytes to urokinase-type plasminogen activator transgenic/severe combined immunodeficiency (uPA^+/+ /SCID) mice. These chimeric mice have various amounts of human hepatocytes that proliferate extensively and progressively replace mouse hepatocytes. In the chimeric liver, hepatic cords and sinusoid-like structures were observed. The human hepatocytes expressed human albumin, human cytochrome P450 enzymes, and human transporter proteins. Furthermore, electron microscopic analysis demonstrated bile canaliculi associated with human hepatocytes in the chimeric mouse livers. These results indicate that the chimeric mouse livers contain functionally intact and differentiated human hepatocytes. Additionally, the toxicologic response of hepatocytes to acetaminophen (APAP) administration was compared in normal and chimeric mouse livers. Following 1,400 mg/kg APAP, mild hepatocellular degeneration was observed in the human hepatocyte areas in the chimeric mice, compared with severe centrilobular hepatocellular necrosis in the ICR mouse livers. In conclusion, these chimeric livers contain functionally differentiated human hepatocytes, and are less susceptible to APAP toxicity, compared to ICR mice.

Introduction

The main functions of the liver include metabolism of nutrients, synthesis and secretion of bile, synthesis of the plasma proteins, destruction of spent blood cells, and detoxification of metabolic waste products and various toxins. Owing to the functional diversity and highly developed three-dimensional architectures of hepatocytes, vasculature, and associated nonparenchymal cells, investigations of human liver functions have been limited to in vitro and ex vivo assays.

Recently, a human hepatocyte chimeric mouse was established by transplantation of human hepatocytes into the liver of homozygous urokinase-type plasminogen activator transgenic/ severe combined immunodeficiency (uPA^+/+/SCID) mice (Tateno et al., 2004). Heterozygous uPA-transgenic mice (albumin pro-motor) were crossed with SCID mice to produce uPA^+/−/SCID mice that were then crossed to produce uPA^+/+/SCID mice. The uPA^+/+/SCID mouse undergoes continuous hepatocellular damage due to expression of the albumin-uPA transgene (Tateno et al., 2004), and also has immunologic tolerance to human hepatocytes as a result of the SCID mutation. Consequently, human hepatocytes can be transplanted into these mice and establishment of these hepatocytes can compensate for damaged endogenous murine hepatocyte functions. Periodic intraperi-toneal nafamostat mesilate injections are needed to maintain immunotolerance to the human hepatocytes.

To date, a few animal models containing human hepatocytes in the liver have been developed (Aurich et al., 2007; Dandri et al., 2001; Ho et al., 2005; Mercer et al., 2001; Meuleman et al., 2005; Ouyang et al., 2001). Most of these humanized liver models are not sufficiently populated with transplanted human hepatocytes and the normal hepatic architecture is not maintained or restored. In this study we characterized the morphologic and functional differentiation of transplanted human hepatocytes in the chimeric liver of uPA^+/+/SCID mice and evaluated their response to acetaminophen, a classical hepatotoxin.

Materials and Methods

Chemical

Acetaminophen (APAP) was purchased from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in 0.5% methylcellulose (Metolose SM-4000; Shin-Etsu Chemical, Tokyo, Japan) aqueous solution.

Treatment of Animals

To characterize the morphologic features of human hepato-cytes in the chimeric mouse liver, 23 uPA^+/+/SCID mice with humanized livers (chimeric mice), 5 uPA^WT/WT/SCID mice, and 2 uPA^+/−/SCID mice were examined (Table 1). The toxicological responses to APAP were evaluated in 9 chimeric and 15 ICR (SLC Japan, Atsugi, Japan) mice. All experiments were conducted in accordance with the Hiroshima Prefectural Institute of Industrial Science and Technology Ethics Board, the Ethics Committees of Kanazawa University, and the Animal Ethics Committee of Pfizer Global Research and Development Nagoya Laboratories.

Chimeric mice with humanized livers were generated by the method described previously using uPA^+/+/SCID mice (Figure 1; Tateno et al., 2004). uPA^+/− transgenic mice (B6SJL-TgN [Alb1Plau] 144Bri; The Jackson Laboratories, Bar Harbor, ME, USA) were crossed with SCID mice (Fox Chase SCID C.B-17/lcr-scid Jcl; Crea Japan, Tokyo, Japan) to generate uPA^+/+/ SCID, uPA^+/−/SCID, and uPA^WT/WT/SCID mice. Human hepato-cytes were purchased from In Vitro Technologies (Baltimore, MD, USA) or BD Gentest (Woburn, MA, USA). At 20–30 days of age, uPA^+/+/SCID mice were inoculated with human hepato-cytes through a small left-flank incision into the inferior splenic pole. Following inoculation, human albumin (hAlb) was measured weekly in 2-μL tail vein blood samples using an enzyme-linked immunoassay kit (Bethyl Laboratories, Montgomery, TX, USA). To maintain immunotolerance to the transplanted human hepatocytes, uPA^+/+/SCID mice were injected intraperi-toneally with 0.3 mg/200 μL/animal nafamostat mesilate (6-amidino-2-naphthyl p-guanidinobenzoate dimethanesulfonate; Torii Pharmaceutical, Tokyo, Japan) once every 2 days, once per day or twice per day for hAlb levels of up to 4 mg/mL, 6 mg/mL, or >6 mg/mL, respectively. Chimeric mice were generated by PhoenixBio (Hiroshima, Japan). All mice used in the studies were aged 11–14 weeks. Mice were housed in cages individually (175 × 245 × 125 mm) with free access to tap water and a pellet diet (CRF-1; CREA Japan, Tokyo, Japan). The animal room was maintained at 21–25°C with 40–70% relative humidity and a 12 h light-dark cycle.

For the APAP toxicity study, three groups of chimeric mice (12–14 weeks; n = 3/group, hAlb = 6.6–15.1 mg/mL) received a single dose of APAP by oral gavage (0 mg/kg, 1,400 mg/kg, and 1,400 mg/kg) and were killed at 24, 4, and 24 h postdose, respectively. To eliminate any potential effects on APAP toxicity, chimeric mice were not administered nafamostat in the 3 days prior to the study. For comparison with the chimeric mice, five groups of ICR mice (12 weeks; n = 3/group) were administered APAP by oral gavage (0 mg/kg, 400 mg/kg [two groups], and 1,400 mg/kg [two groups]) and killed at 4 or 24 h postdose. ICR mice were used in this experiment because sufficient uPA^WT/WT/SCID mice were not available from the bleeding scheme; in addition, the ICR strain is commonly used in toxicology study in Japan.

Necropsy and Tissue Preparations

Animals were killed by cutting the abdominal aorta after anesthesia with isoflurane (inhalation) or pentobarbital (70 mg/kg, intraperitoneal injection). Tissues were fixed in 4% paraformalde-hyde (PFA) or snap-frozen in liquid nitrogen. PFA-fixed tissues were analyzed by light or electron microscopy, and frozen-tissue sections were used for specific immunohistochemical assays. For light microscopy, PFA-fixed samples were trimmed, embedded in paraffin, sectioned to a thickness of 4 μm, and stained with hematoxylin and eosin (H&E) or used in immuno-histochemistry assays.

Immunohistochemistry

Immunohistochemistry (IHC) was used to demonstrate the expression of human proteins in transplanted hepatocytes, sinusoidal endothelial cells, and basement membranes. Markers of cell proliferation and apoptosis were also detected via IHC. Specific reagents and experimental conditions are summarized in Table 2. Frozen tissues were sectioned at 6 μm and mounted on glass slides, and postfixed in cold acetone or 4% PFA for 3 minutes and incubated with antibody. Additional PFA-fixed tissues were processed routinely, paraffin embedded, and sectioned at 6 μm. Pretreatment of paraffin tissue sections consisted of 0.01 M citric acid (pH 6.0) in distilled water at 100°C for 10 minutes (heat pretreatment, Table 2) or proteinase K (DakoCytomation, Tokyo, Japan) for 10 minutes. Murine mono-clonal antibodies were detected with avidin-biotin (Vectastain Elite ABC; Vector Laboratories, Burlingame, CA, USA) while peroxidase-conjugated rabbit anti-goat immunoglobulin G (IgG; Chemicon, Temecula, CA, USA) or goat anti-rabbit IgG (Vectastain) for the rabbit polyclonal antibodies were used as the secondary antibodies. IHC stains were visualized with the chro-mogen 3,3′-diaminobenzidine tetrachloride (DakoCytomation) and counterstained with hematoxylin.

Detection of Apoptosis

The In Situ Apoptosis Detection Kit (TACS2 TdT-DAB; Trevigen, Gaithersburg, MD, USA) was used to label apoptotic cells in paraffin sections (TdT-dUTP nick end labeling or TUNEL). TUNEL labeling required pretreatment of the tissue sections with proteinase K, and Mn²⁺ was used as the cation in the labeling reaction mixture. Paraffin-embedded small intestine was used as the positive tissue control for the proliferating cell nuclear antigen (PCNA) and TUNEL staining.

Image Analyses

Quantitative image analysis was carried with the Image Processor for Analytical Pathology (IPAP-WIN; Sumika Techno-service, Osaka, Japan). The replacement index (RI), which is the extent of human hepatocyte population in the mouse liver, was expressed as the percentage of CK8,18- or OCH1E5-positive staining area in the total liver section. Immunolocalization of CYP expression (CYP% = percentage of CYP-positive area in the total liver section) and PCNA-labeling indices (PCNA-LI% = percentage of PCNA-positive cells in the liver section) were calculated. For each animal, four fields (4× objective for CK8,18 or OCH1E5, 10× for PCNA) were measured to calculate the average score for each IHC marker.

Electron Microscopic Examinations

For transmission electron microscopy, small pieces of PFA-fixed liver from three mice (one uPA^WT/WT/SCID and two chimeric mice with RI > 90%) were postfixed in 2.5% glutaraldehyde, contrasted with 1% osmium tetroxide, and embedded in epoxy resin. Sections were cut to a thickness of 1 μm, stained with toluidine blue, and examined microscopically. Appropriate areas were selected in the semithin sections, and ultrathin sections were made from the trimmed blocks. Ultrathin sections mounted on the grid mesh were stained with uranyl acetate and lead citrate, and observed with a Hitachi H-7600 electron microscope (Hitachi, Tokyo, Japan).

Results

Histopathology of the Chimeric Mouse Livers

H&E and IHC analysis of the human hepatocyte markers demonstrated proliferation of human hepatocytes with varying degrees of replacement of mouse hepatocytes in the chimeric livers (Figure 2).

The livers from uPA^+/+/SCID chimeric mice exhibited multi-focal clear eosinophilic areas interspersed with or surrounded by basophilic hepatocytes (Figure 2A, 2B, 2G, and 2H). In contrast, livers from uPA^WT/WT/SCID mice consisted of normal appearing hepatocytes (Figure 2C and 2I). The clear eosinophilic areas in the chimeric mice were positive for OCH1E5 (Figure 2F), CK8,18 (Figure 2D and 2E), and hAlb (Figure 2J and 2K) by IHC and were interpreted to represent viable, transplanted human hepatocytes. As shown in Table 1, chimeric mice had varying amounts of human hepatocytes in the liver (RI up to 95%), and the RI generally correlated with the hAlb levels in the blood (data not shown). In areas negative for hAlb, CK8,18, and OCH1E5 staining, basophilic small hepatocytes (replacing irregular areas of single-cell hepatocyte necrosis) and brown pigment-laden macrophages were observed. These histological features were distinct from those of uPA^WT/WT/SCID mice (Figure 2). In chimeric mice with lower RI, normal mouse hepatocytes were observed in the liver; furthermore, these normal murine hepato-cytes were negative for human hepatocyte markers (Figure 2A). Thus, the chimeric mouse livers had three components: foci of repopulating human hepatocytes, degenerating donor mouse hepatocytes, and histologically normal mouse hepatocytes. Human hepatocytes in the chimeric livers were distinguished from murine hepatocytes by their clear cytoplasm and smaller nuclei on H&E staining (Figure 2). Furthermore, PCNA IHC demonstrated proliferation of human hepatocytes in the mouse liver. The human hepatocyte foci exhibited varying amounts of proliferation (PCNA labeling index [PCNA-LI], 22–68%) and apoptotic cells were observed in the murine hepatocyte areas (Figure 3). In well-differentiated foci of human hepatocytes (CK8,18 or OCH1E5 positive with low PCNA-LI), hepatic cords and sinusoid-like structures were observed (Figures 4E and 5A).

Laminin and CD31 IHC were used to evaluate the architecture of the human hepatocyte foci and the relationship with degenerating murine hepatocytes (Figure 4). In areas of degenerating mouse hepatocytes, the sinusoid structure was characterized by the presence of increased laminin positive filaments (Figure 4C). In the human hepatocyte areas, poorly vascularized hepatocyte aggregates (few CD31-positive sinusoid endothelial cells) were observed in the areas of proliferating (higher PCNA LI) human hepatocytes (Figure 4C and 4D). Developing hepatic cords were lined by a laminin–positive basement membrane and by CD31-positive sinusoid endothelial cells (Figure 4E and 4F).

Electron microscopy demonstrated developed bile canaliculi between human hepatocytes (Figure 5C and 5E). Furthermore, transplanted hepatocytes expressed CYP2E1 (predominantly diffuse, >70% hepatocytes, Figure 2M), CYP3A4 (faint and diffuse in >40% cells), and CYP3A5 (faint staining in <5% cells) proteins. In addition, IHC localized expression of the MRP2 transporter to the apical hepatocyte membrane (Figure 2N), with faint expression of MRP3 (Figure 2O), and MRP6 and P-glycoprotein (PGP; data not shown), in the basolateral membranes of the human hepatocytes.

Toxicological Responses to APAP

The toxicological response in the chimeric liver following APAP administration was compared with that of the ICR mouse liver (Figure 6), a standard toxicology strain in Japan. At 24 h postdose, all chimeric mice survived the oral administration of APAP at 1,400 mg/kg, while ICR mice were dead. In the livers of chimeric mice receiving 1,400 mg/kg APAP, mild vacuolation of hepatocytes (Figure 6E and 6F) and mild hepatocellular degeneration (Figure 6F) were observed in the human hepato-cyte areas. A few TUNEL-positive cells were observed in the human hepatocyte area at 4 h after administration of 1,400 mg/kg APAP, but these were not apparent at the 24 h time point. CYP IHC revealed that APAP–related changes were limited to a 63% decrease (relative to vehicle control) in CYP2E1 expression at 24 h (Figure 6J–L). APAP hepatotoxicity was mild in the chimeric mouse livers, in contrast to the severe centrilobular hepatocellular necrosis observed in the ICR mouse livers (Figure 6N and 6O). These results suggest that the chimeric mice are less susceptible to APAP toxicity than ICR mice. This may be due to differences between human and mouse hepatocytes, genetic differences between the uPA^+/+/SCID and ICR mice strains, or some combination of these two factors.

Discussion

In the present study, mouse-human chimeric livers contained variable numbers of human hepatocytes as demonstrated by IHC with human-specific hepatocellular markers and circulating hAlb. Human hepatocytes proliferated extensively (PCNA-LI, 22–68%) in the mouse liver, replacing damaged and degenerating murine hepatocytes and restoring some of the architectural features (bile caniliculi, sinusoidal endothelial cells, basal lamina). The chimeric mouse with the highest hAlb levels appeared healthy, except for small body size, as previously described (Tateno et al., 2004). In addition, the livers did not exhibit any inflammatory reactions associated with the presence of the human hepatocytes (Tateno et al., 2004).

The results described here suggest that the transplanted human hepatocytes can function effectively in terms of nutrition, bile acid secretion, and synthesis of hAlb or other proteins without adversely affecting the donor mouse. Furthermore, in the murine environment, these human hepatocytes retained expression of human CYPs and transporter proteins (Katoh and Yokoi, 2007; Nishimura et al., 2005; Tateno et al., 2004; Tsuge et al., 2005) and can be expected to retain human liver function. Although the light and electron microscopic features of the humanized hepatocyte areas showed incompletely developed architecture, sinusoid-like CD31-positive vascular channels, hepatic cord-like structures, and bile canaliculi were observed in the well-differentiated human hepatocyte areas.

Mouse hepatocytes undergoing replacement by transplanted human hepatocytes appeared shrunken and degenerate, similar to those reported in Alb-uPA^+/+ transgenic mice (Sandgren et al., 1991). Meuleman et al. (2005) morphologically and biochemically characterized chimeric livers from uPA^+/+/SCID mice repopulated with human hepatocytes and successfully infected them with HBV and HCV. Similar to our study, the transplanted human hepatocytes were swollen with clear cytoplasm, and the plasma of the chimeric mice contained hAlb as well as 21 additional human proteins (Meuleman et al., 2005). Prior mRNA-expression analysis of our uPA^+/+/SCID chimeric livers demonstrated expression of 20 human CYPs, 26 human phase II metabolic enzymes, and 21 human transporters (Nishimura et al., 2005). The reported mRNA expression profiles of CYP2E1, CYP3A4, and CYP3A5 mRNAs are consistent with protein localization of these CYPs to the human hepatocyte areas in our IHC study. Other animal models with humanized livers, including immunodeficient (Pfp/Rag2) mice (Aurich et al., 2007) and rats (Ho et al., 2005; Ouyang et al., 2001), have shown that it may be possible to transplant human hepatocytes into the livers; however, extensive morphologic, immunohistochemical, and ultrastructural characterization of the humanized livers has not been previously described.

In the present study, we demonstrate that transplanted human hepatocytes express human CYPs, MRPs, and PGP in the murine liver following repopulation; in addition, the toxicologic response to APAP in these chimeric mice is less dramatic than that in nonchimeric (ICR) mice.

High doses of APAP produce centrilobular hepatic necrosis, mediated by CYP-dependent metabolism (CYP2E1, CYP3A4, and CYP1A2) to N-acetyl-p-benzoquinone imine (NAPQI; Dahlin et al., 1984). APAP-induced hepatic injury is increased by depletion of glutathione and formation of covalent adducts with hepatic proteins, and these subsequent changes are mediated by NAPQI. Furthermore, NAPQI induces oxidant stress (peroxynitrate) and mitochondria-mediated cell death (Latchoumycandane et al., 2007). In addition, an in vitro examination using human hepatoma cells (Bel-7402) revealed that the cytotoxic effects of APAP were increased in the presence of S9 mixture (Zhang et al., 2007). These results indicate that APAP is a metabolism-mediated cytotoxicant for both rodents and humans.

Some transgenic mouse studies indicate that metabolism-related factors and transporter expression are necessary for the hepatotoxic effects of APAP. CYP2E1 knockout (KO; El-Hassan et al., 2003; Lee et al., 1996), Mrp3 KO (Manautou et al., 2005), CAR (constitutive androstane receptor, a key regulator of APAP metabolism) KO (Zhang et al., 2002), and SOD1 (Cu,Zn-superoxide dismutase) KO (Lei et al., 2006 ) mice have been shown to be more resistant to APAP toxicity. In the SOD1 KO mice, a reduction in CYP2E1 activity (as opposed to CYP2E1 protein) attenuates APAP toxicity. Although all these KO mice seem to have normal phenotypes, survival rates and tolerance to APAP are generally higher than those of the corresponding wild-type mice. The toxic effects in the wild-type mice with 400–600 mg/kg APAP were similar to those in ICR mice in our study. By comparison, our uPA^+/+/SCID chimera livers contained more differentiated hepatocytes than in vitro human hepatoma cells, and the hepatocytes in these chimeric liver expressed human CYP, MRP, and PGP proteins. Although CYP and MRP activities were not measured, the immunoexpression level of CYP2E1 was reduced in the human hepatocyte area 24 h after APAP administration. Our findings that uPA^+/+/SCID chimeric mice better tolerated APAP toxicity might suggest that the metabolizing proteins were not fully activated in the liver. Furthermore, immature hepa-tocytes such as oval cells are reported to be resistant to APAP injury (Kofman et al., 2005). Thus, the reduced toxicity to APAP in uPA^+/+/SCID chimeric mice may be due to functional immaturity of the repopulating human hepatocytes. Quantification of NAPQI and exposure level of APAP in treated ICR mouse liver, mouse and human hepatocyte areas of chimeric mouse liver, and uPA^WT/WT/SCID mouse liver might identify the cause of the differences in APAP toxicity among these mice. There may also be other factors contributing to this difference. Michael et al. (1999) showed the importance of hepatic macrophages in APAP toxicity. In addition, Liu et al. (2006) used intracellular adhesion molecule-1-deficient mice to demonstrate that neutrophil accumulation contributes to the progression of APAP-induced hepatotoxicity. In the present study, histopathological examination of the uPA^+/+/ SCID chimeric mice showed that nonparenchymal cells (such as Kupffer cells and sinusoid endothelial cells) were not clearly observed in the areas of proliferating human hepatocytes and that there were no inflammatory reactions in the liver associated with APAP administration. The absence of inflammatory cell or Kupffer cell activity might also attenuate the APAP hepatotoxic-ity in the uPA^+/+/SCID chimeric mice. Vacuolation and degeneration of human hepatocytes were induced by APAP, but severe necrosis was not observed in the chimeric mouse livers. The influence of nafamostat administration and the genetic background of the SCID mouse on the inflammatory cell reactions will be addressed in subsequent investigations.

Thus, the mechanisms by which APAP exerts its toxic effects are incompletely understood, even in rodents. At present, there is no evidence of species differences in the mechanisms underlying APAP toxicity or APAP sensitivity. The present study revealed clear differences in APAP toxicity levels between the uPA^+/+/SCID chimeric and ICR mice, but the results were obtained in partially humanized livers (RI < 95%), not a “completely” humanized liver. We believe this animal model will continue to be a useful tool to investigate drug metabolism, hepatotoxicity, and idiosyncratic liver diseases of humans.

In conclusion, human hepatocyte-transplanted chimeric mice contain viable, differentiating, and proliferating human hepato-cytes in the liver. The transplanted human hepatocytes express human proteins, develop complex architectural features (bile canaliculi and hepatic cords), providing an “in vivo” murine approach to investigate human liver function.