Yoshizato Project, Cooperative Link of Unique Science and Technology for Economy Revitalization (CLUSTER), Hiroshima Prefectural Institute of Industrial Science and Technology, Hiroshima, Japan
Hiroshima University Liver Project Research Center, Hiroshima, Japan
Developmental Biology Laboratory and Hiroshima University 21ststCentury COE Program for Advanced Radiation Casualty Medicine, Department of Biological Science, Graduate School of Science, Hiroshima University, Hiroshima, Japan
PhoenixBio Co., Ltd., Hiroshima, Japan
PhoenixBio Co., Ltd., 3-4-1 Kagamiyama, Higashihiroshima, Hiroshima 739-0046, Japan
Potential conflict of interest: Nothing to report.
We previously identified a small population of replicative hepatocytes in long-term cultures of human adult parenchymal hepatocytes (PHs) at a frequency of 0.01%-0.09%. These hepatocytes were able to grow continuously through serial subcultures as colony-forming parenchymal hepatocytes (CFPHs). In the present study, we generated gene expression profiles for cultured CFPHs and found that they expressed cytokeratin 19, CD90 (Thy-1), and CD44, but not mature hepatocyte markers such as tryptophan-2,3-dioxygenase (TO) and glucose-6-phosphatase (G6P), confirming that these cells are hepatic progenitor-like cells. The cultured CFPHs were resistant to infection with human hepatitis B virus (HBV). To examine the growth and differentiation capacity of the cells in vivo, serially subcultured CFPHs were transplanted into the progeny of a cross between albumin promoter/enhancer-driven urokinase plasminogen activator-transgenic mice and severe combined immunodeficient (SCID) mice. The cells were engrafted into the liver and were able to grow for at least 10 weeks, ultimately reaching a maximum occupancy rate of 27%. The CFPHs in the host liver expressed differentiation markers such as TO, G6P, and cytochrome P450 subtypes and could be infected with HBV. CFPH-chimeric mice with a relatively high replacement rate exhibited viremia and had high serum levels of hepatitis B surface antigen. Conclusion: Serially subcultured human hepatic progenitor-like cells from postnatal livers successfully repopulated injured livers and exhibited several phenotypes of mature hepatocytes, including susceptibility to HBV. In vitro–expanded CFPHs can be used to characterize the differentiation state of human hepatic progenitor-like cells. (HEPATOLOGY 2008.)
Studies using rodents with damaged livers have shown that parenchymal hepatocytes (PHs) have great growth potential. When mouse (m) hepatocytes were transplanted into the livers of albumin promoter/enhancer-driven urokinase plasminogen activator (uPA)-transgenic mice,1 they engrafted and repopulated the host liver. Serial transplantation experiments using m-hepatocytes in mice with tyrosinemia showed their enormous growth capacity.2 The replicative potential of rat hepatocytes has also been demonstrated by transplanting them into the partially hepatectomized liver of a retrorsine-treated rat,3 and uPA-transgenic mice crossed with severely immunodeficient mice, such as severe combined immunodeficient (SCID)/beige mice,4 SCID mice,5, 6 or recombination activation gene 2 knockout mice7 have been used to show the growth potential of human (h)-hepatocytes. When transplanted into uPA/SCID mice, PHs from a human juvenile male grew in the host liver to a level at which the proportion (replacement index) of the area of repopulated h-hepatocytes to the total number (host and donor) of hepatocytes reached 96% at 64 days posttransplantation.5 Such h-hepatocyte–chimeric mice have been used to study the pharmacological responses of h-hepatocytes5 and to investigate h-hepatitis viral infections.4, 6–8
In contrast, normal hepatocytes have limited replicative capacity in vitro and acquire an abnormal phenotype if they are cultured for extended periods.9, 10 Studies on hepatocytes cultured in a newly devised medium (hepatocyte clonal growth medium11, 12) revealed a subpopulation of highly replicative PHs, known as small hepatocytes (SHs), in both rats12 and humans.13 Their occupancy rate in h-liver ranged from 0.01% to 0.09% and was dependent on donor age.13 The h-SHs formed colonies and grew continuously through several subcultures, which led us to name them colony-forming PHs (CFPHs).13 Replication of the CFPHs was donor age–dependent up to passage 7 (p = 7),13 and the cells did not exhibit a normal hepatocytic phenotype. Instead, they exhibited the traits of hepatocytes or biliary cells depending on the culture conditions. In addition, the CFPHs were not susceptible to infection with hepatitis B virus (HBV) (unpublished data).
In this study, we generated gene expression profiles of CFPHs and transplanted serially subcultured CFPHs into homozygous uPA/SCID mice to examine their growth and differentiation capacity. Our results indicate that the cells were engrafted onto the liver parenchyma and repopulated the tissue, ultimately differentiating into mature hepatocytes. Importantly, the in vitro–propagated CFPHs became susceptible to infection with HBV. This study supports our previous suggestion that CFPHs from h-postnatal liver are hepatic progenitor-like cells with the potential to assume a normal hepatocytic phenotype.13
Materials and Methods
This study was performed with the approval of the Hiroshima Prefectural Institute of Industrial Science and Technology Ethics Board. PHs were isolated as described13, 14 from livers donated by a 12-year-old Asian male (12YM) and a 16-year-old Asian female (16YF) according to the guidelines of the 1975 Declaration of Helsinki. Cryopreserved PHs from a 9-month-old Caucasian male (9MM) and a 10-year-old Caucasian female (10YF) were obtained from In Vitro Technologies (Baltimore, MD) and BD Biosciences (San Jose, CA), respectively.
Culture of CFPHs.
Cryopreserved PHs from the 9MM, 12YM, and 16YF were thawed5 and serially subcultured to obtain in vitro–expanded CFPHs.13 Commercial 9MM PHs and freshly isolated 12YM and 16YF PHs were each subcultured to p = 3. The expanded cells were then cryopreserved, thawed upon use, and cultured on collagen-coated plates for 14-20 days as described.13
We detached 12YM CFPHs (p = 4 or 5) from culture plates by treatment with 0.25% Trypsin-EDTA (Invitrogen, Carlsbad, CA), suspended, incubated on ice for 30 minutes with m-monoclonal antibodies against hThy-1 (clone F15-42-1; Chemicon, Temecula, CA), and incubated with antibodies against m-immunoglobulin G Alexa-488 (Molecular Probes, Eugene, OR). We used m-immunoglobulin G1 as a negative control. The cells were then analyzed and separated using a fluorescence-activated cell sorter (Becton Dickinson, Franklin Lakes, NJ) as reported.12
Transplantation of PHs and CFPHs.
We detached 9MM and 12YM CFPHs (p = 4) from their culture plates and treated for 1 hour with DMEM containing 10% fetal bovine serum and 3 μg/mL anti–h-integrin α1 monoclonal antibodies (clone FB12, Chemicon).15 This procedure improved engraftment of the CFPHs in uPA/SCID m-liver and reduced host mortality.
Transplantation of PHs and CFPHs was performed as described previously.5 Homozygous uPA/SCID mice were injected with 0.75 × 106 9MM and 12YM PHs or 0.75-1.0 × 106in vitro–expanded 9MM and 12YM CFPHs into the inferior splenic pole. When necessary, 10 mM 5-bromo-2′-deoxyuridine (BrdU) (Sigma, St. Louis, MO) and 1.2 mM 5-fluoro-2′-deoxyuridine (Wako, Osaka, Japan) in saline were injected intraperitoneally into the mice at 10 μL/g body weight 1 hour prior to death. The animals were treated according to the guidelines of our local committee on animal experiments.
We transplanted 9MM and 12YM CFPHs into 6 and 40 uPA/SCID mice, respectively. The mice were then killed 3, 9, or 10 weeks later, depending on the experimental purpose. In a previous report, we used 9MM and 12YM PHs as donor cells.5 In this study, we used some of the preserved livers from these mice for histological examinations and as sources of RNA for reverse-transcription polymerase chain reaction (RT-PCR) analysis. The mice used in our transplantation experiments were separated into 7 groups (A-G) as shown in Table 1, which includes the rates of engraftment and replacement indices (RIs) of the chimeric mice.
Table 1. Summary of CFPH and PH Transplantation Experiments in uPA/SCID Mice
Blood samples (5 μL) were collected periodically after transplantation from the tail veins of the hosts, and the level of h-albumin (ALB) in each was determined using a Human Albumin ELISA Quantitation Kit (Bethyl Laboratories, Montgomery, TX) to monitor the growth of the transplanted CFPHs.
An RNeasy Tissue Kit (Qiagen, Valencia, CA) was used to isolate total RNA from freeze-thawed 9MM and 10YF PHs, cells of the h-hepatoma cell line HepG2, and 12YM and 16YF CFPHs (p = 4). RNA was also isolated with Isogen (Nippon Gene, Tokyo, Japan) from the livers of homozygous uPA/SCID mice and mice chimeric for 12YM PHs or 12YM CFPHs. Each RNA sample was treated with deoxyribonuclease (Takara Bio, Kyoto, Japan) and used as the template for RT-PCR. The RNA (1 μg) was reverse-transcribed with random hexamers using PowerScript Reverse Transcriptase (Clontech, Kyoto, Japan). All reactions were performed with Ex Taq (Takara Bio). Semiquantitative PCR was performed to allow linear amplification of the targets. The following h-specific or m and h cross-reactive genes were subjected to RT-PCR under the conditions shown in Supplementary Table 1: ALB, α1-antitrypsin (AAT), tryptophan-2,3-dioxygenase (TO), glucose-6-phosphatase (G6P), α-fetoprotein (AFP), cytokeratin 19 (CK19), biliary glycoprotein (BGP), Thy-1, CD44, multidrug resistance protein 1 (MDR1), multidrug resistance-associated protein 1 (MRP1), MRP2, and glyceraldehyde-3-phosphate dehydrogenase.
In Situ Hybridization.
Cryosections (7 μm thick) were fixed with 4% paraformaldehyde, then incubated with 100 ng/mL proteinase K for 10 minutes at 37°C. The sections were then treated at 90°C for 6 minutes and hybridized for 2 hours at 37°C with biotinylated h-DNA probes (Dako, Glostrup, Denmark). The sections were also used to detect whole h-genomic DNA using the GenPoint System (Dako) according to the manufacturer's instructions. Finally, they were stained with hematoxylin-eosin.
Immunohistochemistry and Histochemistry.
Formalin-fixed livers were embedded in paraffin and sectioned 5 μm thick. The sections were heated in a microwave oven for 5 minutes in Target Retrieval Solution (Dako), then placed at room temperature for 20 minutes. The livers used to generate frozen sections were embedded in OCT compound (Sakura Finechemicals, Tokyo, Japan), frozen in liquid nitrogen, and sectioned 5 μm thick. The cultured cells were fixed in cold ethanol for 10 minutes. The primary antibodies and conditions used for immunohistochemistry are listed in Supplementary Table 2. For bright-field immunohistochemistry, the antibodies were visualized using a Vectastain ABC Kit (Vector Laboratories, Burlingame, CA) using DAB substrates. Fluorescense immunohistochemistry was performed using Alexa 488–conjugated or Alexa 594–conjugated secondary antibodies (Molecular Probes). The nuclei were stained with Hoechst 33258. Glycogens were visualized using a periodic acid-Schiff (PAS) staining kit (Muto Pure Chemicals, Tokyo, Japan). RIs were determined using hALB-immunostained sections of chimeric m-livers as reported previously.5
We obtained h-serum containing high-titer HBV DNA (8.1 log10 genome equivalents/mL serum) from an HBV genotype C carrier after obtaining informed consent. The serum was kept at −80°C until use. Four CFPH-chimeric mice were intravenously injected with 100 μL of the HBV-positive serum 9-12 weeks after transplantation.
HBV Marker Analysis.
Hepatitis B surface antigen (HBsAg) was measured using an Architect Analyzer (Abbott, Osaka, Japan). Serum DNA was extracted using a SMITEST EX-R&D Nucleic Acid Extraction Kit (Genome Science Laboratories, Fukushima, Japan). Small amounts of HBV DNA (<300 copies/mL) were detected via nested PCR.8 If HBV DNA was detected during the initial round of PCR, the copy number was determined via real-time PCR as reported.8
Phenotypes of CFPHs In Vitro.
We seeded 9MM and 12YM PHs on culture dishes and confirmed that the CFPHs from the 2 donors were similar in morphology and replicative capacity. A small number of the CFPHs (0.01%-0.09% of the seeded PHs) began to replicate after 5 days, and the number of replicating cells gradually increased until colonies appeared at 17 days (Fig. 1A); after 21 days, the cells covered the surface of the dish (Fig. 1B). Most of the seeded PHs were not replicative, and they gradually flattened, acquiring a senescent morphology within 20 days of seeding (Fig. 1A). The CFPHs showed an epithelial cell–like morphology with scant cytoplasm (Fig. 1B), and they retained this appearance during subculture (Fig. 1C). The population doubling time (PDT) of the CFPHs gradually increased as the passage number increased. Up to p = 4, the CFPHs from the young donors replicated with a population doubling time of 170-220 hours; subsequently, the population doubling time increased until the cells finally became senescent.13
The expression of several marker genes was compared among PHs, HepG2 cells, and CFPHs (Fig. 1D). In our experience, no significant differences exist in the marker gene expression profiles of PHs among different donors, and the same trend applies to subcultured CFPHs.13 At p = 4, the CFPHs expressed less ALB and AAT messenger RNA compared with the PHs. The PHs expressed TO and G6P, both of which are markers of mature hepatocytes, whereas the CFPHs did not. CK19, a hepatic progenitor/biliary cell marker, was expressed in both the CFPHs and HepG2 cells, but not in the PHs. BGP, a cell–cell adhesion molecule in epithelium, endothelium, and myeloid cells,16 was expressed in the PHs and HepG2 cells, but only faintly in the CFPHs. The CFPHs, but not the PHs or HepG2 cells, expressed Thy-1, a hematopoietic/hepatic progenitor cell marker. AFP, a hepatic progenitor/carcinoma cell marker, was only detectable in HepG2 cells. CD44, an SH17 or oval cell marker,18 was strongly expressed in CFPHs, but only faintly in PHs and HepG2 cells. PHs and CFPHs faintly expressed MDR1. PHs expressed MRP2, but not MRP1. In contrast, CFPHs expressed MRP1, but not MRP2. A change from MRP2 to MRP1 expression during culture has been reported in rat hepatocytes.19
Thy-1 and CD44 expression in CFPHs was assessed via immunocytochemistry (Fig. 1E-F). A few CFPHs were positive for Thy-1 (Fig. 1E), whereas the majority was strongly positive for CD44 (Fig. 1F). Fluorescence-activated cell sorting indicated that a minor population of the CFPHs expressed Thy-1 (Fig. 1G-H), with an occupancy rate of 1%-3% (Fig. 1H). The CFPHs expressed CK7, CK8, CK18, and CK19 in the preconfluent state and became CK7- and CK19-negative in condensed regions postconfluence (data not shown), which is in agreement with our previous findings.13 Other hepatic stem cell markers such as CD34 and c-kit were undetectable in our CFPHs (data not shown).
Repopulation of CFPHs in uPA/SCID Mouse Liver.
We transplanted 12YM CFPHs (p = 4) into 27 homozygous uPA/SCID mice. The serum concentration of hALB was monitored posttransplantation as a measure of the RI of CFPHs (Fig. 2A). Approximately half of the hosts had no or only a small increase in the level of hALB throughout the experimental period. The remaining mice showed a continuous increase in the concentration of hALB, which reached >10 μg/mL after 9 to 10 weeks. Animal 27 showed the greatest increase, reaching 0.7 mg/mL after 10 weeks. The RI of each of the 14 mice in which blood hALB concentration was >8 μg/mL after 9 to 10 weeks was determined by dividing the hALB-positive areas by the entire area measured,5 and the data were plotted against the corresponding blood hALB concentrations (Fig. 2B). RIs between 0.2% and 27.0% were well correlated with blood hALB concentrations in the 9-728 μg/mL range.
Livers of mice engrafted with the CFPHs were subjected to immunohistochemical staining for hALB (Fig. 3A -D,H) and in situ hybridization using h-genomic DNA probes (Fig. 3I). hALB-positive cells were visible within 3 weeks posttransplantation as single cells or small clusters consisting of up to 25 cells (Fig. 3A-B). Larger clusters containing 20-450 hALB-positive cells appeared after 9 to 10 weeks (Fig. 3C for animal 2 and Fig. 3D for animals 17 and 27). To detect replicating CFPHs, the mice were given BrdU after 9 weeks. BrdU-positive CFPHs were observed at the edges of the colonies (Fig. 3E-G). Serial liver sections were prepared from CFPH-chimeric mice 9 to 10 weeks after transplantation for hALB immuohistochemistry (Fig. 3H) and for in situ hybridization with an h-DNA probe (Fig. 3I). The regions identified as containing h-hepatocytes by the 2 methods were identical.
Comparison of Repopulation by CFPHs and PHs.
PHs and CFPHs (p = 4) were prepared from the livers of 9MM and 12YM donors and transplanted into uPA/SCID mice, and the mice were killed 3 and 10 weeks posttransplantation. The transplanted cells were identified as hALB-positive from histological sections. The number of PH- and CFPH-derived clusters was 125.0 ± 28.2 (n = 3) and 3.3 ± 7.5 (n = 7), respectively, per cross-section of the left lobe of the livers 3 weeks after transplantation, suggesting that the rate of engraftment of the CFPHs was much lower than that of the PHs.
The CFPHs were smaller in size compared with the PHs after 3 weeks (Fig. 4A -B). The cytoplasm of the former was less abundant and more strongly stained for hALB than that of the latter. We observed hCD44 in the plasma membrane of the CFPH-derived cells (Fig. 4E), but not in that of the PH-derived cells (data not shown). At 10 weeks posttransplantation, the CFPHs had increased in size to match those of the PHs, whose sizes were unchanged (Fig. 4C-D), and hCD44 expression disappeared from the CFPH-derived cells (Fig. 4F). The diameter of each CFPH and PH was quantified as follows: 18.3 ± 5.1 μm (mean ± SD, n = 65) versus 25.8 ± 6.4 μm (n = 124) at 3 weeks and 27.0 ± 5.5 μm (n = 185) versus 25.8 ± 4.8 μm (n = 187) at 10 weeks. We found no significant differences in this parameter between the 12YM and 9MM samples. Thus, it appears that the CFPHs replicated without changing their original small size until 3 weeks posttransplantation, when they became larger.
Liver sections from the chimeric mice were stained with hematoxylin-eosin to compare the morphological features of PHs and CFPHs at 10 weeks. The repopulated CFPHs (Fig. 4G) showed no significant difference in morphology compared with the repopulated PHs (Fig. 4H). As reported previously,5, 6 the PHs in the chimeric livers were enlarged and had less eosinophilic cytoplasm than the PHs in h-livers. The livers of the mice that had low hALB levels at 10 weeks posttransplantation were mostly occupied by red nodules, which have been reported to be formed by the transgene-deleted hepatocytes of the host.20
Gene and Protein Expression Profiles of CFPHs in Chimeric Mice Compared with Those of PHs.
Three 12YM CFPH-chimeric mice (11, 15, and 17) were randomly selected from the mice in Fig. 2A and killed 10 weeks after transplantation. RNA was extracted from each liver to generate gene expression profiles via RT-PCR. RT-PCR was also performed on 2 12YM PH-chimeric mice that were included in a previous study.5 The CFPH livers expressed hALB, hAAT, hTO, hG6P, and hMRP2, but not hCK19, hThy-1, or hMRP1, just as in the PH-livers (Fig. 5). Previously, we showed that the PHs in chimeric mice expressed various h-cytochrome P450 (hCYP) subtypes in a manner similar to the donor liver.5 In this study, we found that the expression of hCYPs 1A2, 2C8, 2C9, 2D6, and 2E1, but not 3A4, in the CFPH-chimeric mice was similar to that in the PH-chimeric mice (data not shown). Expression of hCYP3A4 was very low (less than one-fifth) in CFPHs compared with that in PHs.
Protein expression was investigated immunohistochemically for the CFPH-chimeric livers at 3, 9, and 10 weeks posttransplantation. All of the examined CFPHs were Thy-1–negative, CK7-negative, CK19-negative, and AFP-negative (data not shown). The hALB-positive cells were coincident with the hCK18-positive cells at both 3 (data not shown) and 9 weeks posttransplantation (Fig. 6A -C). MRP2-positive signals were present on the bile canalicular membranes of the transplanted CFPHs at 10 weeks (Fig. 6D-F). CYP3A4-expressing CFPHs were localized in the pericentral zone (Fig. 6G-I) as reported previously,21 but their distributions were unique. Although some of the CFPHs were positive for CYP3A4, approximately 70% of them were negative. In contrast, all of the CFPHs in the pericentral zone strongly expressed CYP1A2 (Fig. 6J-L), which is known to be expressed in postnatal liver.22 The CFPHs in the chimeric mice were strongly PAS-positive (Fig. 6N), whereas the in vitro CFPHs were faintly PAS-positive (data not shown). From these results, we conclude that the transplanted CFPHs differentiated into functionally mature hepatocytes. No h-cell tumors were formed during any of our experiments in the uPA/SCID mice.
Infection of CFPH-Chimeric Mice with HBV.
To further examine whether CFPHs had exhibited normal differentiated phenotypes in chimeric mice, we tested their susceptibility to HBV infection. Four CFPH-chimeric mice with various serum hALB levels (0.2, 1.6, 7.3, and 222.0 μg/mL) were inoculated with 100 μL of HBV-positive h-serum at 9-12 weeks posttransplantation. The animals were then tested every 2 weeks for HBV viremia and serum hALB levels (Fig. 7A). The amount of HBV DNA in the animals increased between 2 and 8 weeks after inoculation, and all 4 mice developed measurable viremia within 8 weeks. However, a correlation was observed between the HBV DNA and/or HBsAg level and the hALB level: the former appeared to be high when the latter was high (Fig. 7A). HBsAg was detectable in the serum of the chimeric mice when they showed elevated virus titers: the HBsAg levels of chimeric mice with HBV DNA levels of 2 × 103, 5.2 × 105, 5.9 × 107, and 7.7 × 108 copies/mL 8 weeks after inoculation were <0.05, <0.05, 3.2, and 124.0 IU/mL, respectively. HBV was infectious to CFPH-chimeric mice with very low levels of hALB (<104 ng/mL), and all mice showed quantitatively measurable viremia (>103 copies/mL) up to 8 weeks after inoculation. In contrast, most PH-chimeric mice with <104 ng/mL hALB did not show quantitatively measurable levels of viremia up to 12 weeks after inoculation (data not shown) as reported previously.8 In this study, we confirmed that CFPHs were not susceptible to infection with HBV prior to transplantation. The presence of hepatitis B core antigen and HBsAg in the CFPHs from HBV-infected chimeric livers was examined immunohistochemically (Fig. 7C,E). CFPHs were positive for both antigens that were sporadically distributed in the same regions among the CFPH colonies. Hepatitis B core antigen–positive cells accounted for 18.7 ± 8.3% of the total number of CFPHs (n = 3; total cell count = 1,215) (Fig. 7C), and both the nucleus and cytoplasm of the cells showed signals (Fig. 7D).
This study supports our previous conclusion that CFPHs are h-hepatic progenitor-like cells.13 Cultured CFPHs expressed such hepatic progenitor cell markers as CK19, Thy-1, and CD44, but not mature hepatocyte markers such as TO and G6P. We also found that in vitro–expanded CFPHs in uPA/SCID mice were able to repopulate the parenchyma, in which they differentiated into mature hepatocytes. FISH (fluorescence in situ hybridization) using mouse X chromosome probes showed that the engrafted and propagated CFPHs did not fuse to the mouse cells (data not shown). Thus, replicative CFPHs isolated from postnatal liver are normal, functional hepatocyte progenitor-like cells.
The existence of stem/progenitor cells in the adult liver is controversial.23–25 In the present study, we showed that the CFPHs expressed CK19, Thy-1, and CD44, but not AFP, in serial culture. Thy-1 antigens are expressed in h-hepatic progenitor cells in fetal liver26 and in rat oval cells,27 but not in normal adult hepatocytes. We showed that Thy-1-expressing cells were present among the CFPHs at an occupancy of 1%-3%. SHs show greater growth potential than PHs in rats.12 Other studies have reported that CD44 is a specific marker for rat SHs in vitro and in vivo, and that its expression level decreases with SH maturation in vitro.17 Moreover, a recent study demonstrated that CD44 was strongly expressed by oval cells in a 2-acetylaminofluorene/partial hepatectomy, a D-galactosamine, and a retrorsine/partial hepatectomy rat model, but not by small hepatocyte-like progenitor cells (SHPCs)18 that appeared in a retrorsine/partial hepatectomy model.28 We detected CD44 expression in CFPHs at the plasma membrane. These results suggest that Thy-1 and CD44 may be common markers for both rat and h-hepatic progenitor cells.
Mouse embryonic liver stem cell lines differentiate into both hepatocytes and bile ducts in uPA/SCID mice.29 Like PHs, our CFPHs differentiated into mature hepatocytes, but not into biliary epithelial cells, in uPA/SCID mice. CFPHs are considered to be hepatic progenitor-like cells, like rat SHs12, 30–33 and SHPCs.28, 34 SHPCs are closely related to SHs; they are small and similar in size,28, 30 and both express CYP3A1 and 2E1 at a low level.28, 32 At 3 weeks posttransplantation, the CFPHs were small in size, had a large nucleus-to-cytoplasm ratio, and expressed hCD44, but not hCK19. At 10 weeks, the cells became bigger, assumed a morphology similar to that of PH-derived cells, and lost their expression of hCD44. The expression of hCYP3A4 was quite low (0.15-fold) among CFPHs compared with that of PHs (data not shown). In addition, the distribution of hCYP3A4-expressing CFPHs in the pericentral zone was unique: more than two-thirds of CFPHs did not express CYP3A4. In the case of the h-PH–chimeric mice, all PHs in the pericentral zone expressed CYP3A4 (data not shown).
Presently, we lack experimental data to explain the expression of hCYP3A4 in CFPH-chimeric liver, but CFPHs may require some specific environmental factor(s) for differentiation, which might be absent from mouse liver. Alternatively, some factors that specifically inhibit the differentiation of CFPHs might be present there. CK7-positive h-hepatic progenitor cells are present in the livers of uPA/SCID mice transplanted with h-postnatal liver-derived PHs,6 and these small cells are strongly immunoreactive to pan-cytokeratin with scant cytoplasm. The CFPHs were morphologically similar to these cells at 3 weeks posttransplantation, although we were unable to detect CK7-positive cells in either the PH- or CFPH-transplanted chimeric livers. However, CFPHs were hCK7-, hCK19-, and hCD44-positive, at least until 1 day posttransplantation (data not shown).
We reported previously that uPA/SCID livers were nearly completely replaced with young donor PHs at 10 weeks posttransplantation.5 In contrast, the RIs of our CFPH-chimeric mice were <30% at 9 to 10 weeks. CFPHs were rare in the host liver at 3 weeks posttransplantation, whereas several PHs were observed. The lower RIs of the CFPHs might be attributable to their lower engraftment efficiency.
In conclusion, h-hepatocytes in immunodeficient, and liver-injured mice are useful for the study of viral hepatitis. Repopulated h-hepatocytes are susceptible to infection with HBV6–8 and HCV.4, 6 Additionally, h-hepatocyte–chimeric mice are usually produced by transplanting fresh6, 7 or cryopreserved hepatocytes,4, 5 but sources of h-hepatocytes are limited. Several studies have reported on liver repopulation by in vitro–propagated cells from adult and fetal livers, such as immortalized mouse hepatic stem cells,29 rat SHPCs,34 immortalized h-hepatocytes transfected with full-length HBV,35 and fetal h-epithelial/hepatic progenitor cells.36, 37 However, the RIs in these studies were extremely low (less than a few percent). In the present study, we were able to produce CFPH-chimeric mice with RIs as high as 27%. Thus, CFPHs could be an alternative to h-hepatocytes as a source of hepatocytes for transplantation. Moreover, the CFPH-chimeric mice were susceptible to infection with HBV, even though their serum hALB levels were extremely low (102-103 ng/mL). CFPH-chimeric mice will be useful for studying h-HBV and for characterizing h-hepatic progenitor cells.
We thank H. Kohno, Y. Matsumoto, S. Nagai, A. Tachibana, Y. Yoshizane, and Y. Seo for providing technical assistance. We also thank Dr. K. Ohashi (Tokyo Women's Medical University) for helpful discussion and comments during the preparation of this manuscript.