Hepatic precursors derived from murine embryonic stem cells contribute to regeneration of injured liver

Authors

  • Jeonghoon Heo,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute (NCI), National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Valentina M. Factor,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute (NCI), National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Tania Uren,

    1. Laboratory of Hepatitis Viruses, Division of Viral Products, CBER/Food and Drug Administration, Bethesda, MD
    Search for more papers by this author
  • Yasushi Takahama,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute (NCI), National Institutes of Health, Bethesda, MD
    Current affiliation:
    1. 1st Department of Surgery, Nara Medical University, Nara, Japan; Ju-Seog Lee is currently affiliated with the Department of Molecular Therapeutics, University of Texas, M.D. Anderson Cancer Center, Houston, TX
    Search for more papers by this author
  • Ju-Seog Lee,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute (NCI), National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Marian Major,

    1. Laboratory of Hepatitis Viruses, Division of Viral Products, CBER/Food and Drug Administration, Bethesda, MD
    Search for more papers by this author
  • Stephen M. Feinstone,

    1. Laboratory of Hepatitis Viruses, Division of Viral Products, CBER/Food and Drug Administration, Bethesda, MD
    Search for more papers by this author
  • Snorri S. Thorgeirsson

    Corresponding author
    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute (NCI), National Institutes of Health, Bethesda, MD
    • Laboratory of Experimental Carcinogenesis, National Cancer Institute, 37 Convent Dr., Bldg. 37/Room 4146, Bethesda, MD 20892
    Search for more papers by this author
    • fax: 301-496-0734


  • Potential conflict of interest: Nothing to report.

Abstract

We established an efficient system for differentiation, expansion and isolation of hepatic progenitor cells from mouse embryonic stem (ES) cells and evaluated their capacity to repopulate injured liver. Using mouse ES cells transfected with the green fluorescent protein (GFP) reporter gene regulated by albumin (ALB) enhancer/promoter, we found that a serum-free chemically defined medium supports formation of embryoid bodies (EBs) and differentiation of hepatic lineage cells in the absence of exogenous growth factors or feeder cell layers. The first GFP+ cells expressing ALB were detected in close proximity to “beating” myocytes after 7 days of EB cultures. GFP+ cells increased in number, acquired hepatocyte-like morphology and hepatocyte-specific markers (i.e., ALB, AAT, TO, and G6P), and by 28 days represented more than 30% of cells isolated from EB outgrowths. The FACS-purified GFP+ cells developed into functional hepatocytes without evidence of cell fusion and participated in the repairing of diseased liver when transplanted into MUP-uPA/SCID mice. The ES cell-derived hepatocytes were responsive to normal growth regulation and proliferated at the same rate as the host hepatocytes after an additional growth stimulus from CCl4-induced liver injury. The transplanted GFP+ cells also differentiated into biliary epithelial cells. In conclusion, a highly enriched population of committed hepatocyte precursors can be generated from ES cells in vitro for effective cell replacement therapy. (HEPATOLOGY 2006;44:1478–1486.)

Hepatocyte transplantation is an effective treatment for liver failure and/or end-stage liver disease. However, the shortage of donor organs and the difficulties of cyropreservation and long-term culturing of mature hepatocytes have limited the clinical application of cell-based therapy. Recently, the use of embryonic stem (ES) cells has attracted considerable interest for cell replacement therapy because of their capacity to proliferate indefinitely in vitro while retaining the potential to differentiate into all types of cells including hepatocytes.1–5 Clinical application of ES cells requires a simple and efficient protocol for directing their differentiation into a specific cell type. Several studies have reported the induction of ES cell differentiation into hepatocytes both in vitro6–9 and in vivo.10 The in vitro approaches involve the formation of embryoid bodies (EBs) to re-create the inductive microenvironment required for liver organogenesis11, 12 or treatment with specific growth factors and cytokines critical for hepatocyte differentiation such as hepatocyte growth factor, fibroblast growth factor, and oncostatin.1, 13 Also, the introduction of genes promoting endodermal differentiation into ES cells,14 modification of culture microenvironment by supplementation with extracellular matrix proteins or coculture with other cell types,15, 16 is used to direct ES cells toward a hepatocytic lineage. Nevertheless, these strategies generally have been inefficient, producing small numbers of hepatocyte precursors within heterogeneous cell populations. Furthermore, ES-derived hepatocytes have been found to have little utility for physiological and pathological liver remodeling, and their transplantation into liver is frequently associated with formation of teratomas.12, 17

In the present study we adapted the ES/EB methodology and defined the conditions that promote efficient differentiation of mouse ES cells toward the hepatocytic lineage both in vivo and in vitro. First, we show that a serum- and growth-factor-free chemically defined medium favors propagation of hepatocyte precursors in EB-derived cultures. Second, we find that highly enriched populations of early-stage hepatocyte precursors can be isolated from EB outgrowths by fluorescence-activated cell sorting (FACS) using green fluorescent protein (GFP) expression driven by a lineage-specific (albumin) promoter. Finally, we demonstrate that in a mouse model of liver injury ES-derived GFP+ cells give rise to functional hepatocytes that integrate into and replace diseased parenchyma without formation of teratomas following transplantation.

Abbreviations

ES, embryonic stem; EBs, embryoid bodies; FACS, fluorescence-activated cell sorting; GFP, green fluorescent protein; ALB, albumin; AG, albumin promoter/enhancer-green fluorescent protein; EGFP, enhanced green fluorescent protein; PCR, polymerase chain reaction; MUP, major urinary protein; uPA, urokinase-type plasminogen activator; SCID, severe combined immunodeficiency; BrdU, bromodeoxyuridine; GS, glutamine synthetase.

Materials and Methods

Establishment of ES Cell Lines Expressing GFP Under the Control of ALB Promoter/Enhancer (AG-ES Cells).

A 2.3-kb mouse albumin (ALB) promoter/enhancer18 was cloned into the promoterless enhanced green fluorescent protein (EGFP) vector, pEGFP-1 (BD Bioscience Clontech, Palo Alto, CA), after digestion with SacI and KpnI restriction enzymes. The resulting construct, pALB-GFP, was electroporated into the HM-1 ES cell line19 and the Hepa 1-6 HCC cell line, which was used as a positive control for GFP expression. Clones transfected with pALB-GFP were referred to as AG-ES or AG-Hepa 1-6 cells. Several independent clones were used to confirm the stable genomic integration of pALB-GFP through more than 10 passages in culture.

Culture and Differentiation of AG-ES Cells.

Undifferentiated AG-ES cells were maintained as described.20 To generate embryoid bodies, the AG-ES cells were dispersed into a single-cell suspension in Iscove's modified Dulbecco's medium (IMDM; Invitrogen, Carlsbad, CA) containing 20% fetal bovine serum (FBS; HyClone, Logan, UT), 2 mmol/L L-glutamine (Invitrogen), 300 μmol/L monothioglycerol (Sigma, St. Louis, MO), and antibiotics and cultured by the hanging drop method (1 × 103 ES cells/30 μL).21 After 5 days, EBs were replated on collagen IV-coated plates and cultured for an additional 26-28 days. To induce differentiation into hepatocytes, EBs were grown in the following media: (1) IMDM supplemented with 20% FBS, 2 mmol/L L-glutamine, and 300 μmol/L monothioglycerol, and antibiotics; (2) William E serum-free medium (Invitrogen) supplemented with 1× ITS (BD Bioscience), 10 μmol/L hydrocortisone-21-hemisuccinate (StemCell Technologies Inc., Vancouver, BC, Canada), 0.05% bovine serum albumin (Invitrogen), 2 mmol/L ascorbic acid, 10 mmol/L nicotinamide (Sigma), 1 μmol/L dexamethasone (Sigma), 2 mmol/L L-glutamine, and antibiotics; and (3) HepatoZYME-SFM (Invitrogen) serum-free medium designed for primary hepatocyte cultures. The media were changed every 2 days.

RNA Isolation and RT-PCR Analysis.

Total RNA was extracted from either cultured cells or FACS-sorted GFP+ and GFP cell fractions using TRIzol reagent (Invitrogen). Polymerase chain reaction (PCR) products were separated by electrophoresis on 1.2% agarose gel containing ethidium bromide. Primers and optimal annealing temperatures are listed in Supplementary Table 1 (Supplementary material for this article can be found on the Hepatology website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html).

Quantitative PCR was performed using a SYBR Green PCR Core Reagents kit (Applied Biosystems, Foster City, CA) as previously described.20 mRNA expression was normalized by the internal control of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and then represented as the values relative to the control.

Immunohistochemistry.

EB cultures were fixed in 4% paraformaldehyde and 0.15% picric acid in phosphate-buffered saline (PBS) at room temperature for 20 minutes, washed in PBS, and treated with permeabilization buffer containing 0.1% Triton X-100, 10% donkey serum, and 1% BSA in PBS for 45 minutes. Cells were incubated with mouse ALB antibody (1:20; R&D systems, Minneapolis, MN) at 4°C overnight, followed by goat anti-mouse Cy3-conjugated secondary antibody (1:250; Jackson Immunoresearch Labs, West Grove, PA) at room temperature for 45 minutes. Slides were mounted using mounting medium with DAPI (Vector Labs, Burlingame, CA) and examined with a Zeiss 510 confocal laser scanning microscope (Carl Zeiss, Germany).

Fluorescence-Activated Cell Sorting.

To isolate GFP+ cells, outgrowths from EBs were digested in a buffer containing 0.1% collagenase type II (GIBCO) and 1.2 U/mL dispase II (Roche Applied Science, Indianapolis, IN) for 20 minutes at 37°C. A single-cell fraction was separated from the nondigested EB clumps using a 40-μm cell strainer. Cells were resuspended in PBS containing 10% FBS and 2 mmol/L EDTA and sorted on a FACS Vantage (BD Bioscience) equipped with Turbo Sort for high-speed sorting. After sorting, viability, determined for each cell population by trypan blue exclusion, was typically more than 90%.

Transplantation.

Homozygous major urinary protein-urokinase-type plasminogen activator/severe combined immunodeficiency (MUP-uPA/SCID) mice were generated by crossbreeding MUP-uPA transgenic mice, described by Weglarz et al.,22 with SCID-BEIGE mice purchased from Taconic. Cells (106 in 50 μL of HBSS) were injected into the inferior poles of the spleens of 3-week-old mice of both sexes anesthetized with isoflurane. Cellular engraftment in the liver was examined 5, 32, and 82 days after transplantation. Each treatment group consisted of 2-4 animals. For acute CCl4injury, animals that had received transplanted cells 82 days previously were subjected to CCl4(Sigma) by inhalation for 15 minutes, using a 2-L inhalation chamber containing 100 μL of CCl4 (Sigma) as previously described23 and sacrificed 2 days later. Bromodeoxyuridine (BrdU; Boehringer Mannheim, Indianapolis, IN; 150 mg/kg body weight) was injected intraperitoneally 2 hours before sacrifice. All studies were done according to National Institutes of Health guidelines for animal care.

Histology and Immunohistochemistry.

Liver tissues were fixed by perfusion with 4% paraformaldehyde. After additional fixation in 4% paraformaldehyde for 2 hours at 4°C, the tissues were either cryoprotected in 30% sucrose overnight and embedded in OCT compound (Sakura, Torrance, CA) or embedded in paraffin. Cryostat sections (8 μm) were washed with PBS and permeabilized in PBS with 0.1% BSA, 0.3% Triton X-100, and 10% normal goat serum for 45 minutes at room temperature. Antimouse GFP (1:1,000; Molecular Probes, Eugene, OR), anti-human/mouse serum ALB (1:20; R&D system), anti-mouse Cnx32 (1:30; Santa Cruz Biotechnology Inc., Santa Cruz, CA), anti-mouse ATP-binding cassette, subfamily G, member 2 (Abcg2; 1:200; Kamiya Biomedical Company, Seattle, WA), anti-mouse glutamine synthetase (GS; 1:300; BD Transduction Laboratories, Franklin Lakes, NJ), anti-mouse cytochrome P450, family 2, subfamily E (Cyp2E; 1:200; Stressgene, Ann Arbor, MI), and anti-rat A6 (1:10) were diluted in PBS containing 1% BSA and 0.6% Triton X-100 and incubated overnight at 4°C. Expression of the proteins was visualized with appropriate Alexa-594- (1:1,000; Molecular Probes), Cy3- (1:200; Jackson ImmunoResearch Inc.), or Texas Red (1:200; Jackson ImmunoResearch Inc.)-conjugated secondary antibodies. Slides were mounted with medium containing DAPI (Vector labs). BrdU-DNA immunostaining was performed as previously described.24 Images were captured using a Zeiss LSM 510 NLO confocal microscope (Carl Zeiss, Inc., Thornwood, NY) or a fluorescence microscope. Quantification of GFP-positive areas and BrdU-positive cells was performed using Photoshop 7.0 (Adobe System, Mountain View, CA) and Image-Pro Plus V 5.12 (Media Cyberneties, Inc., Silver Spring, MD) software. To determine the percentage of BrdU-positive cells in ES-derived or host parenchyma, at least 1,200 cells from each group of mice were counted.

Statistical Analysis.

The Student t test was used to determine the differences between groups. P values of less than .05 were considered significant.

Results

Establishment of ES Cell Lines Expressing GFP Under Control of ALB Promoter (AG-ES Cells).

To track the cells committed to the hepatic lineage, we used a system in which the expression of GFP was controlled by ALB promoter/enhancer. A 2.3-kb ALB promoter/enhancer construct18 was cloned into a promoterless EGFP vector referred to as pALB-GFP (Fig. 1A). When pALB-GFP was introduced into the Hepa 1-6 HCC cell line expressing ALB protein (AG-Hepa 1-6), bright GFP fluorescence was detected, demonstrating the specificity of pALB-GFP expression (Fig. 1B). After transfection of pALB-GFP into HM-1 ES cells, several AG-ES cell lines were established. The AG-ES cells displayed the same morphology, growth rate, and capacity of EB formation (data not shown) as the wild-type ES cells (HM1-ES) and did not express either ALB or GFP mRNA (Fig. 1C). However, EBs generated from AG-ES cells expressed both GFP and ALB mRNA (Fig. 1C), again indicating the specificity of the pALB-GFP construct. The intensity of the GFP fluorescence signal was weaker in the AG-EB derived cells than in the AG-Hepa 1-6 cells yet sufficient for FACS analysis (Fig. 1D).These results show that AG-ES cell lines are useful for detection and isolation of cells committed to the hepatocytic lineage.

Figure 1.

Establishment of ES cell lines expressing GFP under the control of ALB promoter/enhancer. (A) Construction of pALB-GFP vector. A 2.3 kb-ALB promoter/enhancer was inserted into the multiple cloning site of the promoterless EGFP vector after digestion with SacI and KpnI restriction enzymes. (B) Confocal image of AG-Hepa 1-6 cells transfected with pALB-GFP. (C) RT-PCR analysis of GFP and ALB in AG-Hepa 1-6 cells and AG-EBs cultured for 17 days. (D) Flow cytometry analysis of GFP expression in differentiated AG-ES cells (AG-EB) at 24 days. Wild-type EB (WT-EB) and AG-Hepa 1-6 cells were used as negative and positive controls, respectively (ES, embryonic stem cell; GFP, green fluorescent protein; ALB, albumin; EGFP, enhanced green fluorescent protein; AG, albumin promoter/enhancer–green fluorescent protein; EBs, embryoid bodies).

In Vitro Differentiation of AG-ES Cells Into Hepatocytic Lineage.

Various media were tested for the capacity to induce hepatocytic differentiation, as judged by morphology and expression of hepatocyte-specific markers. EBs displayed remarkably different growth patterns depending on medium composition (Fig. 2A). In serum-rich media, EBs grew overconfluent and produced a multilayered culture with a mixed-cell phenotype. EBs maintained in a commercial serum-free medium (HepatoZYME designed for primary hepatocyte cultures) failed to form cell outgrowths and rapidly disintegrated. In contrast, EBs cultured in chemically defined serum-free medium generated distinct outgrowths that expanded as monolayer cultures. By 7-10 days, the first GFP-expressing cells were detectable by fluorescence microscopy as clusters surrounded by colonies of rhythmically contracting cells (Fig. 2C). Cells in these clusters expanded into sheets of tightly packed cells that displayed a typical hepatocyte morphology, including binucleated cells (Fig. 2D-E) and coexpressed GFP and ALB (Fig. 2F-K).

Figure 2.

In vitro differentiation of AG-ES cells into hepatocytic lineage. (A) Effect of serum and serum-free culture environments on EB differentiation. EBs were cultured on collagen IV–coated plates in either 20% FBS, serum-free defined, or HepatoZYME-SFM growth medium for 2 and 7 days (original magnification, ×40). (B) Expression of ALB, AFP, GFP, and G6P was analyzed by RT-PCR in AG-EBs cultured in 20% FBS- or serum-free defined medium for 29 days. (C) Confocal image of GFP-expressing cells in AG-EBs cultured in serum-free defined medium for 10 days. GFP+ cells, detected near areas of beating cardiomyocytes, are indicated by red lines (original magnification, ×400). (D) EB outgrowth displaying hepatocyte morphology after 13 days in serum-free defined medium (original magnification, ×100). (E) EB outgrowth displaying hepatocyte morphology after 13 days in serum-free defined medium at higher magnification (×800). A red arrow points to binucleated cells characteristic for hepatic differentiation. (F-H) Confocal images of GFP (×400) in outgrowths of EBs cultured in serum-free defined medium (ALB detected by immunohistochemistry with anti-ALB). (I-K) Confocal images of colocalization of GFP with ALB (×400) in outgrowths of EBs cultured in serum-free defined medium (ALB detected by immunohistochemistry with anti-ALB). (L-M) Real-time PCR analysis of GFP and ALB in AG-EBs cultured in serum-free defined medium. The mRNA level of each gene was normalized to GAPDH levels and is expressed as a ratio of that found in AG-EBs cultured in serum-rich medium for 37 days (ALB, albumin; AFP, alpha-fetoprotein; GFP, green fluorescent protein; G6P, glucose-6-phosphatase).

Reverse-transcription (RT)-PCR analysis confirmed that the expression of ALB and hepatocyte lineage-specific markers was much higher in EBs grown in the defined serum-free medium than in those grown in serum-rich medium (Fig. 2B). Furthermore, expression of glucose-6-phosphatase (G6P), a mature hepatocyte marker, was detected only in EBs cultured in the serum-free medium. Accumulation of ALB and GFP mRNA transcripts was highly correlated and reached the maximum by 26-28 days, as shown by quantitative real-time PCR analysis (Fig. 2L-M). Similarly, ALB level in serum-free medium was more than 100-fold higher than that detected in serum-rich medium. Thus, our serum-free culture conditions favored the differentiation of ES cells toward hepatocytic lineage cells.

Isolation and Characterization of GFP+ Cells.

To enrich GFP+ cells, EB cultures were partially dispersed into a single-cell suspension by collagenase/dispase digestion targeted to EB outgrowths. This approach produced a high yield of GFP+ cells. Eighteen percent of the isolated cells were GFP-positive by 14 days (Fig. 3A) and had almost doubled by 28 days of expansion in culture. When isolated ES cells were separated into GFP+ and GFP fractions using FACS, the purity of GFP+ cells was 95% (Fig. 3B). The expression of hepatocyte-specific genes, including ALB, alpha-1 antitrypsin, G6P, TO, and hepatocyte nuclear factor 4, was higher in GFP+ cells but was also detected at a lower level in GFP cells (Fig. 3C). Microarray analysis of FACS-sorted GFP+ and GFP cells confirmed that GFP+ but not GFP cells displayed a gene expression pattern similar to that of hepatic lineage cells (Supplementary Data 1). Thus, a highly enriched population of hepatic precursors can be isolated from mouse ES cells using a GFP reporter gene regulated by a hepatocyte-specific promoter.

Figure 3.

Isolation and characterization of GFP+ cells derived from AG-EBs. (A) Time-dependent increase in percentage of GFP+ cells. AG-EBs were dissociated into single cell suspensions with 0.1% collagenase type II and 1.2 U/mL dispase II, and GFP-positive (GFP+) cells were separated from GFP-negative (GFP) cells by FACS. Data are expressed as mean ± SD; the number of experiments is indicated in parentheses. (B) FACS analysis of the GFP+cell population isolated from AG-EBs cultured in serum-free defined medium for 26 days. From left to right are dot plots representing the background signal derived from undifferentiated AG-ES cells (AG-ES), and the GFP fluorescence of cells derived from AG-EBs before and after sorting. (C) Comparison of hepatocyte-specific gene expression by RT-PCR in GFP+ and GFP cell populations isolated from AG-EBs cultured for 26 days and in adult liver tissue (AG, albumin promoter/enhancer-green fluorescent protein; ES, embryonic stem cell; EBs, embryoid bodies; GFP, green fluorescent protein; ALB, albumin; AAT, alpha 1-antitrypsin; G6P, glucose-6-phosphatase; TO, tryptophan 2,3-dioxygenase; HNF4, hepatocyte nuclear factor 4; GAPDH, glyceraldehyde-3-phosphate dehydrogenase).

ES Cell-Derived GFP+ Cells Differentiate Into Functional Hepatocytes and Repopulate Diseased Liver of MUP-uPA/SCID Mice.

To examine whether in vitro-generated hepatic precursor cells can differentiate into functional hepatocytes, we performed cell transfer experiments using chimeric MUP-uPA/SCID mice. This model facilitates cell engraftment because of the absence of B- and T-cell-mediated immune response and also provides transplanted cells with a competitive growth advantage from the persistent damage of the host hepatocytes caused by the uPA transgene driven by a hepatocyte-specific MUP promoter.22 ES cell-derived GFP+ cells progressively replaced the diseased parenchyma of MUP-uPA/SCID mice (Fig. 4A -C). Five days after intrasplenic injection, small clusters of GFP+ cells (2-3 cells) were detected in close proximity to portal veins. The size of the GFP+ clusters in a liver cross-section increased progressively to 30-40 cells by 32 days and to 100-400 cells by 82 days after transplantation. The distribution of hepatocyte clusters within the liver sections was variable. Quantitative analysis of GPF-positive areas using confocal images (Fig. 4D) showed that liver repopulation varied from 0% to 20% in randomly selected liver sections, averaging 1.94% ± 5.81% by 82 days after transplantation (Table 1). In the MUP-uPA/SCID mice restoration of injured parenchyma is normally achieved via proliferation of endogenous hepatocytes that lose the MUP-uPA transgene. Therefore, these data show that ES-derived hepatocytes can effectively compete with the endogenous source of liver regeneration. When the GFP cell fraction was transplanted, only one small GFP+ cluster of hepatocytes was found. Significantly, no teratomas were observed up to 82 days after transplantation of either GFP+ or GFP cell fractions (Table 1; Fig. 4F). In striking contrast, multiple teratomas developed in the recipient livers 10 days after transplantation of undifferentiated ES cells (Fig. 4G-H).

Figure 4.

Liver repopulation by AG-ES cell–derived GFP+ cells in MUP-uPA/SCID mice. GFP+ cells (1 × 106 cells/50 μL HBSS per mouse) isolated from AG-EBs after 26–28 days in culture were injected into the inferior poles of the spleens of 3 week-old MUP-uPA/SCID mice. As controls, undifferentiated AG-ES cells (1 × 106 cells/50 μL HBSS per mouse) or 50 μL of saline was injected. (A-C) GFP+ cluster size had increased from 2-3 to 100 cells 82 days after injection (original magnification, ×100; insets show higher magnification, ×400) of GFP+ hepatocyte clusters. (D) A representative liver section with 4.5% liver repopulation. Twelve adjacent confocal images after GFP immunostaining were tiled together to make the composite. White arrows indicate GFP+ hepatocyte clusters (original magnification, ×100). (E) Injection of saline did not result in formation of teratomas. (F) Injection of GFP+ cells did not result in formation of teratomas. (G) Undifferentiated ES cells developed multiple teratomas 10 days after injection. (H) Morphology of teratomas that developed in undifferentiated ES cells 10 days after injection (H&E staining; original magnification, ×100; AG, albumin promoter/enhancer–green fluorescent protein; ES, embryonic stem cell; EBs, embryoid bodies; GFP, green fluorescent protein; MUP-uPA/SCID, major urinary protein-urokinase plasminogen activator/severe combined immunodeficiency; H&E, hematoxylin and eosin).

Table 1. Quantification of Liver Repopulation by ES Cell-Derived GFP+ Cells in MUP-uPA Mice
 Cell Types
GFP+GFPES cellsSaline
  • NOTE. Data are presented as mean ± SD (n = 23 ∼ 35 liver tissue sections per group).

  • *

    In additional 2 mice, no teratoma formation in liver, lung, heart, or spleen were found 104 days after transplantation of ES cell-derived GFP+ cells.

Number of animals4322
Days after injection82821082
No of hepatocyte cluster/cm210.9 ± 8.60.4 ± 0.600
No of duct structures/cm21.6 ± 1.30.4 ± 0.700
% of repopulation1.94 ± 5.810.02 ± 0.0800
Teratoma formationNo*NoYesNo

To address whether liver repopulation can be the result of fusion between injected mouse ES cell-derived GFP+ cells and resident hepatocytes, we performed a PCR-based microsatellite analysis using D15Mit175 primers to detect polymorphisms between donor (129/Sv) and recipient (C57BL/6) cell genomes (Supplementary Data 2) and quantitative fluorescence imaging to analyze hepatocyte size and binuclearity (Supplementary Data 3). No clear evidence of cell fusion was observed in our experiments, suggesting that GFP+ hepatocytes in the repopulated liver resulted mostly from the integration and differentiation of ES cell-derived GFP+ cells. Thus, the population of ES cell-derived GFP+ cells contains hepatic precursors that have the capacity to home in on and functionally repopulate injured liver without teratoma formation.

The GFP+ cells displayed normal morphology and histological staining characteristic of differentiated hepatocytes (Fig. 5). They expressed high levels of ALB (Fig. 5A-C) and established typical junctional contacts (Fig. 5D-F) and common biliary capillaries (Fig. 5G-I) with each other and with neighboring host hepatocytes. In addition, the architecture of GFP+ parenchyma was similar to that of the surrounding liver tissue, as illustrated by zonal heterogeneity in the expression of Cyp2E1 (Fig. 5J-L) and GS (Fig. 5M-O) along the hepatic plate.25 Occasionally, ES-derived hepatic progenitor cells gave rise to bile ducts (Fig. 5P-R), suggesting their bipotential properties.

Figure 5.

Differentiation of AG-ES cell–derived GFP+ cells in the livers of MUP-uPA/SCID mice. Individual and merged images of liver sections 82 days after transplantation are shown. (A-C) Liver sections stained with anti-ALB (original magnification, ×200). (D-F) Sections stained with anti-Cnx 32 in media containing DAPI (white arrow in F inset, establishment of junctional connections between GFP+ cells and neighboring hepatocytes; original magnification, ×400). (G-I) Sections stained with anti-Abcg2 in media containing DAPI (white arrow in J inset, establishment of common biliary capillaries between GFP+ cells and neighboring hepatocytes; original magnification, ×400). (J-L) GFP+ hepatocytes expressed Cyp2E when located pericentrally in sections stained with anti-Cyp2E in media containing DAPI (original magnification, ×200). (M-O) Expression of GS in GFP+ hepatocytes was detected only when cells were in close contact with central but not portal veins in sections stained with anti-GS in media containing DAPI (original magnification, ×200). (P-R) Both green and red signals were intensified for visualization in sections stained with anti-A6 (original magnification, ×200). AG, albumin promoter/enhancer-green fluorescent protein; ES, embryonic stem cell; GFP, green fluorescent protein; MUP-uPA/SCID, major urinary protein-urokinase plasminogen activator/severe combined immunodeficiency; H&E, hematoxylin and eosin; ALB, albumin; Cnx 32, connexin 32; Abcg2, ATP-binding cassette, subfamily G, member 2; Cyp2E, cytochrome P450, family 2, subfamily e; GS, glutamine synthetase; cv, central vein; pv, periportal vein.

ES Cell-Derived Hepatocytes Show Normal Proliferative Response to Additional Growth Stimuli.

Finally, we measured the proliferative activity of GFP+ cells versus the surrounding parenchyma using GFP and BrdU immunohistochemistry. ES cell-derived hepatocytes proliferated at the same rate as recipient hepatocytes (2.53% ± 0.77% vs. 2.11% ± 0.49%, P = .18) and displayed a comparable proliferative response to additional growth stimulation from CCl4-induced liver damage (9.81% ± 2.19% vs. 7.30% ± 1.28%, P = .17; Fig. 6A-H). These findings show that ES-derived hepatocytes are responsive to normal growth regulation in the liver.

Figure 6.

GFP+ hepatocytes subjected to normal growth regulation. MUP-uPA mice were treated with a single dose of CCl4 82 days after transplantation of GFP+ cells. Each animal received an i.p. injection of BrdU (150 mg/kg of body weight) 2 hours before sacrifice. Staining was performed on serial liver sections. Black or white lines outline the areas occupied by GFP+ hepatocytes. Black arrows indicate necrotic areas and white arrows indicate BrdU-positive hepatocytes (GFP, green fluorescent protein; MUP-uPA/SCID, major urinary protein-urokinase plasminogen activator/severe combined immunodeficiency; CCl4, carbon tetrachloride; BrdU, bromodeoxyuridine; H&E, hematoxylin and eosin). (A) H&E-stained serial liver tissue section from untreated mice (original magnification, ×100). (B) H&E-stained serial liver tissue section from mice exposed to CCl4 for 48 hours (original magnification, ×100). (C) GFP-immunostained serial liver tissue section from untreated mice (original magnification, ×100). (D) GFP-immunostained serial liver tissue section from mice exposed to CCl4 for 48 hours (original magnification, ×100). (E) BrdU-immunostained serial liver tissue section from untreated mice (original magnification, ×100). (F) BrdU-immunostained serial liver tissue section from mice exposed to CCl4 for 48 hours (original magnification, ×100). (G) BrdU-immunostained serial liver tissue section from untreated mice at a higher magnification of ×200. (H) BrdU-immunostained serial liver tissue section from mice exposed to CCl4 for 48 hours at a higher magnification of ×200.

Discussion

In the present study we demonstrate that a highly enriched population of early-stage hepatocyte precursors can be generated from ES cells in vitro. Also we show that these precursors differentiate into functional hepatocytes that integrate into and replace diseased liver parenchyma without evidence of fusion with resident hepatocytes.

To detect ES-derived hepatic precursors both in vitro and in vivo, we introduced the GFP transgene driven by the mouse ALB promoter18 into ES cells. ALB promoter has an obvious advantage over the alpha-fetoprotein promoter used in previous studies16, 26 because it can trace mature hepatocytes as well as hepatocyte precursors.27 In our study, the expression of the GFP transgene was stable and did not decrease during ES cell growth, as reported previously.28, 29

To induce hepatic differentiation, we used the developmental potential of the ES/EB system to mimic hepatogenesis in the early embryo.1 Consistent with the inductive role of cardiac mesoderm in hepatic differentiation,30 the first GFP+ cells that displayed morphologic characteristics and expressed hepatocyte-specific markers were found in close proximity to contracting cardiomyocytes. The differentiation of EB-derived cells along hepatocytic lineage depended on the culture environment. The GFP+ cells increased in number in serum- and growth factor-free media, suggesting a convergence of signals from neighboring cells of different types was sufficient to support the hepatocytic potential of ES cells and/or to facilitate the survival of early-stage hepatocyte precursors. In contrast, in serum-rich medium the growth of hepatocyte-like cells was surpassed by other cell types. Thus, the culture of ES cells in a serum-free chemically defined medium was more effective in directing ES cells toward the hepatocytic lineage than was culture in a medium containing 20% FBS.

The capacity of ES cell-derived hepatocytes to engraft and repopulate the recipient liver has been questioned. Although several studies have shown that ES cell-derived hepatocytes incorporate into recipient liver upon transplantation7, 10, 12, 26 the repopulation efficiency was very low, not exceeding 0.1%.12, 26 In our study, ES cell-derived hepatocytes replaced from 1% to 20% of the parenchyma in randomly selected areas of the liver in MUP-uPA/SCID mice. The repopulating ES cell-derived cells displayed characteristics of normal differentiated hepatocytes and established structural connections with each other and with neighboring host hepatocytes, indicating their functional integration. A similar initial repopulation was found after transplantation of bipotential mouse embryonic liver stem cell lines in Alb-uPA/SCID transgenic mice.31

It has been recently reported that transplanted monkey EB cells repopulate the diseased parenchyma of ALB-uPA/SCID mice via fusion with host hepatocytes.32 In the present study we used a highly purified fraction of mouse ES cell-derived cells committed to hepatocytic lineage and did not find evidence of cell fusion based on the analysis of donor DNA level and nucleus size. Our data are consistent with a report that mouse embryonic liver stem cell lines integrate and repopulate damaged livers of ALB-uPA/SCID mice without cell fusion.31 It seems possible that transplantation of unfractionated EB cells and/or differences in donor cell origin and state of maturation/differentiation32 may account for the discrepancies in the results. Consistent with this, transplantation of adult hepatocytes into fumarylacetoacetate hydrolase-deficient mouse liver33 resulted in liver repopulation without cell fusion, whereas bone-marrow-derived cells contributed to liver regeneration in the same model via cell fusion and formation of hybrid cells.34 Nevertheless, our results cannot rule out a low level of cell fusion of the transplanted ES cell-derived hepatocyte progenitors with the resident hepatocytes.

Hepatic cell lines derived from fetal liver are able to differentiate into both hepatocytes and biliary epithelial cells.31, 35 In accordance with these findings, we found the presence of ES cell-derived bile duct structures expressing A6 antigen, a marker of hepatic oval and biliary epithelial cells.36 These data suggest that a population of ES cell-derived GFP+ cells may contain bipotential hepatic precursors that can differentiate into both hepatocytes and bile duct-like cells.

It is well established that ES cells are tumorigenic when transplanted into adult mice,37 making their use for cell replacement therapy problematic. Teratoma formation has been found after transplantation of cells derived from 15-day-old EB cultures7 or ES cells treated with acidic FGF for 7 days.17 Chinzei et al.12reported an inverse relationship between the duration of ES cells in culture and the likelihood of teratoma formation after transplantation. In the present study, ES cell–derived GFP+ cells isolated with high purity using FACS were subjected to normal growth regulation and did not produce teratomas within 104 days after transplantation into the diseased liver of MUP-uPA/SCID mice, whereas undifferentiated ES cells gave rise to numerous tumors only 10 days after transplantation. Similar results were observed after transplantation of GFP-positive12, 26 or PECAM1-negative cells isolated from EB-derived cells.7 Thus, differentiation of ES cells in culture and purification using either positive or negative selection may significantly decrease the tumorigenic potential of ES cells.

In conclusion, this is the first study to report a robust serum- and feeder-free culture system for generation of significant numbers of bipotential hepatic precursors from mouse ES cells. Upon transplantation into the diseased livers of MUP-uPA/SCID mice, in vitro-differentiated ES cells acquire the differentiated properties and growth pattern of adult hepatocytes without development of teratomas and thus provide a valuable source of hepatocyte precursors for tissue repair.

Acknowledgements

We thank Joe W. Grisham for valuable advice and critical reading of the manuscript, Stephen Wincovitch Sr. for assistance with confocal imaging, and Barbara J. Taylor for FACS sorting.

Ancillary