In liver development, a number of growth factors (GFs) and components of the extracellular matrix (ECMs) lead to differentiation of liver parenchymal cells. As the liver contains many cell types, specifically investigating their functional effects on hepatic stem cell populations is difficult. Prospective isolation and clonal assays for hepatic stem cells enable the examination of direct effects of GFs and ECMs on this rare cell fraction. Using previously purified cells that fulfill the criteria for hepatic stem cells, we examined how GFs and ECMs regulate differentiation in the developing liver. We show here that hepatocyte growth factor (HGF) induced early transition of albumin(ALB)-negative stem cells to ALB-positive hepatic precursors resembling hepatoblasts and then oncostatin M (OSM) promoted their differentiation to tryptophan-2, 3-dioxygenase (TO)-positive mature hepatocytes. During this transition, ECMs were necessary for the differentiation of stem cells and precursors, but their effects were only supportive. In the first step of stem cell differentiation induced by HGF, the expression of CCAAT/enhancer binding protein (C/EBP), a basic leucine zipper transcription factor, changed dramatically. When C/EBP function was inhibited in stem cells, they stopped differentiating to hepatocyte-lineage cells and proliferated actively. These are the first findings to illustrate the mechanism of hepatic stem cell differentiation in liver development.

In mouse embryogenesis, the liver primordium initially develops from the ventral foregut endoderm at embryonic day (E) 8(Wilson et al., 1963; Zaret, 2000). Signals from the adjacent cardiac mesoderm and septum transversum, which are mainly mediated by fibroblast growth factors (FGFs) and bone morphogenetic proteins (BMPs),induce cells to express albumin (ALB) and α-fetoprotein (AFP), and to initiate liver bud formation (Jung et al.,1999; Rossi et al.,2001). After the specification of the liver, hematopoietic cells move into this organ and produce oncostatin M (OSM), which induces the maturation of hepatocytes (Kamiya et al.,1999; Kinoshita et al.,1999). At postnatal stages, hepatocyte growth factor (HGF)produced by non-parenchymal liver cells (sinusoidal, stellate and endothelial cells) is involved in hepatic maturation(Hu et al., 1993; Kamiya et al., 2001). These findings demonstrate that several signals mediated by mesoderm-derived cells steer cell fates towards hepatic lineages and induce differentiation into functional hepatocytes. It is not known, however, whether these signals act directly or indirectly to affect hepatic stem cell growth and differentiation. Thus, it is essential to examine the effects of such growth factors (GFs) in single-cell-based studies using purified stem cell populations.

In a previous study, using flow cytometry and single-cell-based assays, we prospectively identified hepatic stem cells with multilineage differentiation potential and self-renewing capability(Suzuki et al., 2002). These cells expressed the hepatocyte growth factor receptor Met and were low-positive for CD49f (α6 integrin subunit), but did not express Kit (stem cell factor receptor), CD45 (leukocyte common antigen) and TER119 (a molecule resembling glycophorin, exclusively expressed on immature erythroid cells). Sorted stem cells could be clonally propagated in culture for over 6 months, where they continuously produced hepatocytes and cholangiocytes as descendants, while maintaining primitive stem cells that expressed neither albumin nor cytokeratin 19 during the greater part of their expansion. Our studies with highly enriched populations with stem cell activity showed that HGF was a critical requirement for proliferation(Suzuki et al., 2000; Suzuki et al., 2002). It remained unknown, however, whether the Met/HGF interaction had a role in stem cell differentiation. Spagnoli et al.(Spagnoli et al., 1998)established a bi-potential hepatic precursor cell line from transgenic animals that constitutively expressed the activated form of Met. This and our previous findings suggest that, while the Met/HGF interaction is crucially responsible for maturation of differentiating hepatocytes in the postnatal liver(Hu et al., 1993; Kamiya et al., 2001) and division of mature hepatocytes in liver regeneration(Michalopoulos et al., 1984; Ishiki et al., 1992), it is also involved, through a separate mechanism, in stem cell growth and differentiation.

In this study, we investigated the direct effects of GFs and extracellular matrix components (ECMs) on proliferation and bi-potential differentiation of prospectively isolated and clonally cultured hepatic stem cells. To analyze primitive stem cells and stem cell-derived differentiating hepatocytes,respectively, cell type was determined by the expression of ALB enhancer/promoter-EGFP. We demonstrated that HGF, but not FGFs, induced early transition from ALB-negative (ALB) stem cells to ALB-positive (ALB+) hepatic precursors through signaling via the CCAAT/enhancer-binding protein (C/EBP), a basic leucine zipper transcription factor. Subsequently, HGF was an effective mitogen for differentiating cells,while OSM inhibited their proliferation and induced their maturation, as assessed by expression of glucose-6-phosphatase (G6P) and tryptophan-2,3-dioxygenase (TO). By contrast, these factors suppressed differentiation into cholangiocyte-lineage cells. Inactivation of C/EBPs, even in the presence of both HGF and OSM, strongly inhibited the differentiation of stem cells into hepatocyte-lineage cells and allowed cells to self-renew efficiently. Although several ECMs could also induce differentiation of stem cells by regulating C/EBPs, their effect was much weaker and may just work supportively for stem cell differentiation. Our present data show that gradual effects from HGF and OSM, mediated by the transcription factor C/EBP, lead stem cells to differentiate into hepatocytes rather than cholangiocytes through the efficient expansion of differentiating cells and permissive signals inducing the maturation of hepatocytes.

Flow-cytometric sorting and culture of hepatic stem cells

Stem cell clones were prospectively isolated from C57BL/6 embryonic day (E)13.5 fetal mice (CLEA, Tokyo, Japan) and were clonally cultured as described(Suzuki et al., 2002). After the initiation of culture, cells were maintained in culture by replating them every 7 days. Three clones were randomly selected and those cells were used for examination. For the GF test, we used HGF (generous gift from Dr T. Ishii,Mitsubishi Pharma Corp), OSM (Sigma, St Louis, MO), acidic FGF (aFGF)(Peprotech, London, UK), and basic FGF (bFGF) (Peprotech), and for the ECM test, we used non-coated dishes, and laminin-, type IV collagen-, type I collagen-and fibronectin-coated dishes (Becton Dickinson, San Jose, CA).

Isolation of total RNA

We prepared total RNA from test samples and, as RNA standards for real-time PCR, from fetal, neonatal and adult liver, using an RNeasy Mini Kit (QIAGEN,Tokyo, Japan) according to the manufacturer's instructions. Total RNA was diluted and used for quantitative analyses.

Semi-quantitative RT-PCR analysis

Sorted 1×105 cells were used to prepare total RNA. After various dilutions of template cDNA, we optimized their concentration for each primer. In these concentrations, amplification by PCR did not reach plateau but could be used semi-quantitative analysis. PCR was conducted using hepatocyte-specific primers for albumin (ALB), α-1-antitrypsin(αAT), glucose-6-phosphatase (G6P), and tryptophan-2,3-dioxygenase (TO),and for positive control hypoxanthine phosphoribosyltransferase (HPRT) as described (Suzuki et al.,2000; Suzuki et al.,2002). PCR cycles were as follows: initial denaturation at 95°C for 4 minutes followed by 40 cycles of 94°C for 1 minute,56°C for 1 minute, 72°C for 1 minute and final extension at 72°C for 10 minutes. PCR products were separated in 2% agarose gel.

PCR primers and TaqMan fluorogenic probes

PCR primers and TaqMan fluorogenic probes for real-time quantitative PCR were designed using the Primer Express software program (Version 1.0) (Applied Biosystems, Tokyo, Japan). The sequences were as follows:hepatocyte-differentiation primers for ALB (forward, 5′-TGT CCC CAA AGA GTT TAA AGC TG-3′; reverse, 5′-TCT TAA TCT GCT TCT CCT TCT CTG G-3′; and probe, 5′-ACC TTC ACC TTC CAC TCT GAT ATC TGC ACA CT-3′), α-fetoprotein (AFP) (forward, 5′-CCT GTC AAC TCT GGT ATC AGC CA-3′; reverse, 5′-CTC AGA AAACTG GTG ATG CAT AGC-3′; and probe, 5′-TGC TGC AAC TCT TCG TAT TCC AAC AGG A-3′), αAT (forward, 5′-TCG GAG GCT GAC ATC CAC AA-3′;reverse, 5′-TCA ACT GCA GCT CAC TGT CTG G-3′; and probe,5′-TTC CAA CAC CTC CTC CAA ACC CTC AA-3′), G6P (forward,5′-GTT CAA CCT CGT CTT CAA GTG GAT-3′; reverse, 5′-TGC TGT AGT AGT CGG T GT CCA GGA-3′; and probe, 5′-TTT GGA CAA CGC CCG TAT TGG TGG-3′) and TO (forward, 5′-CAA GGT GAT AGC TCG GAT GCA-3′; reverse, 5′-TCC AGA ACC GAG AAC TGC TGT-3′; and probe, 5′-TGT GGT GGT CAT CTT CAA GCT CCT GG-3′);cholangiocyte-differentiation primers for cytokeratin 19 (CK19) (forward,5′-TGA AGA TCC GCG ACT GGT-3′; reverse, 5′-TAA AGT AGT GGT TGT AAT CTC GGG A-3′; and probe, 5′-CCA GAA GCA GGG ACC CGG ACC-3′), and γ-glutamyltranspeptidase (GGT) (forward, 5′-TTT GCC TAT GCC AAG AGG AC-3′; reverse, 5′-TTG CGG ATC ACC TGA GAC A-3′; and probe, 5′-ATG CTC GGT GAC CCA AAG TTT GTC G-3′);or miscellaneous primers for hepatocyte nuclear factor 1 (HNF1) (forward,5′-GCT AGG CTC CAA CCT TGT CAC G-3′; reverse, 5′-TTG TGC CGG AAG GCT TCC T-3′; and probe, 5′-AGG TGC GTG TCT ACA ACT GGT TTG CCA-3′), hepatocyte nuclear factor 4 (HNF4) (forward, 5′-TGG TGT TTA AGG ACG TGC TGC-3′; reverse, 5′-ACG GCT CAT CTC CGC TAG CT-3′; and probe, 5′-CAA TGA CTA CAT CGT CCC TCG GCA CTG T-3′), hepatocyte nuclear factor 6 (HNF6) (forward, 5′-CCG GAG TTC CAG CGC AT-3′; reverse, 5′-TCT TGC TCT TTC CGT TTG CA-3′;and probe, 5′-TCG GCG CTC CGC TTA GCA GC-3′), Met (forward,5′-GAT CGT TCA ACC GGA TCA GAA-3′; reverse, 5′-GGA AGA GCC CGG ATA ATA ACA A-3′; and probe, 5′-TGC AGG ATT GAT CAT TGG TGC GGT C-3′), CCAAT/enhancer binding protein-alpha (C/EBPα) (forward,5′-AGC AAC GAG TAC CGG GTA CG-3′; reverse, 5′-TTA TCT CGG CTC TTG CGC A-3′; and probe, 5′-CGG GAA CGC AAC AAC ATC GCG-3′) and CCAAT/enhancer binding protein-β (C/EBPβ)(forward, 5′-CGG ATC AAA CGT GGC TGA G-3′; reverse, 5′-CGC AGG AAC ATC TTT AAG GTG A-3′; and probe, 5′-ACG TGT AAC TGT CTA GCC GGG CCC TG-3′). All TaqMan probes used in this experiment carried a 5′ FAM reporter dye (Applied Biosystems).

Real-time PCR conditions

The RT and the PCR were performed in one step by using TaqMan EZ RT-PCR Core Reagents (Applied Biosystems). The reaction mixture (25 μl final volume) includes 100 or 500 ng total RNA, 5×TaqMan EZ buffer (5 μl),Mn(OAc)2 (3 mM), rTth DNA polymerase (0.1 U/μl), uracil N-glycosylase (0.01 U/μl), dATP, dCTP, dGTP, dUTP (each 300 μM), and forward and reverse primers (200 nM), and probe (100 nM). Reverse transcription was performed at 60°C for 30 minutes. PCR was performed as follows: initial denaturation at 95°C for 5 minutes followed by 60 cycles of 95°C for 15 seconds and 60°C for 1 minutes. A template-free control was included in each experiment. All template-free controls, standard RNA dilutions and test samples were assayed in triplicate.

Analysis of real-time PCR data

The starting amount of mRNA in each test sample was calculated by preparing a standard curve using known dilutions of RNA standards. For each dilution,the ABI-PRISM 7700 software (Applied Biosystems) generated a real-time amplification curve constructed by relating the fluorescence signal intensity(DRn) to the cycle number. The ΔRn value corresponded to the variation in the reporter fluorescence intensity before and after PCR, normalized to the fluorescence of an internal passive reference present in the buffer solution. The standard curve was then generated on the basis of the linear relationship existing between the Ct value (cycle threshold; corresponding to the cycle number at which a significant increase in the fluorescence signal was first detected) and the logarithm of the starting quantity. Starting quantities of mRNA in samples were quantified by plotting the Ct on this standard curve.

Gene transfer into hepatic stem cell cultures

Stable transfection of hepatic stem cell cultures was carried out by lipofection. Briefly, 10 μg of a construct containing both the enhanced green fluorescence protein (EGFP) driven by the ALB promoter (–0.3 kb)and enhancer, a region located 8.5-10.4 kb upstream of the ALB promoter(Pinkert et al., 1987), and the Zeocin resistance gene was used to transfect 1×106 cells using Lipofectamine 2000 (Gibco BRL, Gaithersburg, MD). Stably transfected cells were selected by growth on laminin-coated dishes (Becton Dickinson) in our standard medium supplemented with 600 μg/ml Zeocin (Invitrogen,Groningen, Netherlands) and isolated by using cloning rings (Iwaki Glass,Tokyo, Japan). The frequency of EGFP-positive cells was assayed by FACS-Vantage (Becton Dickinson). In this paper, representative data from a transfected stem cell clone are shown because similar results were obtained from others.

Retrovirus production and FACS analysis of transduced cells

The retroviral vector pGCsam (MSCV) is described elsewhere(Kaneko et al., 2001). A dominant-negative form of C/EBP (A-C/EBP) followed by IRES EGFP or IRES nerve growth factor receptor truncated in the cytoplasmic domain (tNGFR) were subcloned into pGCsam (GCsam-A-C/EBP-IRES-EGFP, GCsam-A-C/EBP-IRES-NGFR,respectively). To produce recombinant retrovirus, plasmid DNA was transfected into 293gp cells (293 cells containing the gag and pol genes but lacking an envelope gene) along with 10A1 env expression plasmid(pCL-10A1) (Miller et al., 1996) by CaPO4 co-precipitation, and supernatant from the transfected cells was collected to infect cells. To detect the expression of tNGFR on the cell surface, cells were stained by mouse anti-human NGFR (Chemicon, Temecula, CA) followed by phycoerythrin(PE)-conjugated rabbit anti-mouse Igs (Dako, Carpinteria, CA) and analyzed by FACS-Vantage (Becton Dickinson).

Purification of ALB and ALB+ cells from heterogeneous stem cell cultures

As an experimental source, we used fluorescence-activated cell sorting(FACS) to sort cells from the Met+ CD49f+/lowKit CD45 TER119 cell fraction in E13.5 fetal mouse livers, in which self-renewing multipotent hepatic stem cells are highly enriched, and then expanded these cells in clonal cultures. After cell transplantation of expanding stem cells, even into mice with immunodeficiency, we have never found abnormal development and tumor formation by donor-derived cells. Furthermore, karyotype-analysis using FACS and propidium iodide showed that they maintained normal G0/G1 (2n) and G2/M(4n) pattern similar to primary sorted cells. These results strongly suggest that those cells are not the product of transformation and are diploid without the karyotypic rearrangements.

To examine direct roles for GFs and ECMs in stem cell differentiation, stem cell clones were placed in several culture conditions and the state of differentiation was analyzed by using FACS and quantitative PCR(Fig. 1). In our stem cell cultures, however, in addition to self-renewing stem cells, differentiating progeny such as hepatocytes and cholangiocytes were spontaneously produced from stem cells. Owing to this heterogeneity, the target cells for the effects of GFs and ECMs could not be determined. To elucidate which steps of stem cell differentiation were affected by differentiation-inducible factors, we separated the original cell population into ALB and ALB+ cells using FACS, following gene transfer of the ALB enhancer/promoter-EGFP construct into stem cell cultures(Fig. 1, Fig. 2A). After FACS-sorting,semi-quantitative RT-PCR analysis was conducted to compare the expression of hepatocyte markers in EGFP+ (ALB+) cells with that in EGFP (ALB) cells. As expected, the expression of hepatocyte markers in ALB+ cells was much higher than in ALB cells (Fig. 2B). These data clearly show that stem cell-derived differentiating heaptocyte-lineage cells can be visualized and specifically separated from other lineage cells. We next examined the expression of the liver-enriched transcription factors in ALB and ALB+ cells. Interestingly, although there was little difference in the expression of HNF1 and HMF4 between ALB and ALB+ cells in the result of real-time quantitative PCR analysis,the expression of HNF6, which is known to be a regulator of pancreatic endocrine cell differentiation (Jacquemin et al., 2000), was much higher in ALB+ cells(Fig. 2C). These results suggest that HNF6 could be involved in hepatocyte-differentiation from hepatic stem cells in the liver development.

Fig. 1.

Experimental procedure to elucidate the manner of hepatic stem cell differentiation. The Met+ CD49f+/lowKit CD45 TER119 cells isolated from E13.5 fetal mouse livers were cultured clonally in 96-well plates. Cells that expanded in culture and produced hepatocytes and cholangiocytes as descendants while maintaining primitive stem cells undergoing self-renewing divisions were used to analyze differentiation status. We chose three independent stem cell clones for examination. Following transfection of the ALB enhancer/promoter-EGFP construct, cells were examined by using FACS and quantitative PCR after they had been cultured with several GFs or ECMs. Scale bar: 100 μm.

Fig. 1.

Experimental procedure to elucidate the manner of hepatic stem cell differentiation. The Met+ CD49f+/lowKit CD45 TER119 cells isolated from E13.5 fetal mouse livers were cultured clonally in 96-well plates. Cells that expanded in culture and produced hepatocytes and cholangiocytes as descendants while maintaining primitive stem cells undergoing self-renewing divisions were used to analyze differentiation status. We chose three independent stem cell clones for examination. Following transfection of the ALB enhancer/promoter-EGFP construct, cells were examined by using FACS and quantitative PCR after they had been cultured with several GFs or ECMs. Scale bar: 100 μm.

Fig. 2.

Separation of differentiating ALB+ cells from the stem cell population. (A) After gene transfer of the ALB enhancer/promoter-EGFP construct into expanding stem cell populations, ALB+ and ALB cells were sorted restrictively using FACS. Immediate re-analysis of sorted cells showed purification of both ALB+ and ALB cells. After the culture of each sorted cell subpopulation for 16 days on type IV collagen-coated dishes, FACS analysis demonstrated that ALB cells emerged from ALB+cells, and that ALB+ cells emerged from ALBcells. Percentages of fractionated cells are shown at the top of each panel. Establishment of the gate was based on the profile of the negative control.(B) Semi-quantitative RT-PCR analysis of sorted EGFP+(ALB+) and EGFP (ALB) cells. Note that ALB+ sorted cells expressed hepatocyte-lineage markers,such as ALB, αAt, G6P and TO, at much higher levels than did ALB cells. (C) Quantitative analysis of sorted ALB+ cells using real-time quantitative PCR. All data were normalized to the value of ALB sorted cells and fold-differences are shown. Representative data from a transfected stem cell clone are shown; three samples were examined for each protein. Data are mean±s.d. (D) (a-f) Several sorted ALB cells could form clonal colonies including both ALB+ and ALBcells at day 20. (g-i) Moreover, ALB+ sorted cells gave rise to ALB cells even 1 day after the initiation of culture, and finally formed mosaic colonies similar to those from ALBcells (black arrowhead in g, an original EGFP+ cell; white arrowhead in g, a daughter cell). (a,d,g) Phase contrast. (b,e,h) Enhanced green fluorescence protein (EGFP) imaging. (c,f,i) Merge. d-f are magnifications of a-c, respectively. Scale bars: 100 μm.

Fig. 2.

Separation of differentiating ALB+ cells from the stem cell population. (A) After gene transfer of the ALB enhancer/promoter-EGFP construct into expanding stem cell populations, ALB+ and ALB cells were sorted restrictively using FACS. Immediate re-analysis of sorted cells showed purification of both ALB+ and ALB cells. After the culture of each sorted cell subpopulation for 16 days on type IV collagen-coated dishes, FACS analysis demonstrated that ALB cells emerged from ALB+cells, and that ALB+ cells emerged from ALBcells. Percentages of fractionated cells are shown at the top of each panel. Establishment of the gate was based on the profile of the negative control.(B) Semi-quantitative RT-PCR analysis of sorted EGFP+(ALB+) and EGFP (ALB) cells. Note that ALB+ sorted cells expressed hepatocyte-lineage markers,such as ALB, αAt, G6P and TO, at much higher levels than did ALB cells. (C) Quantitative analysis of sorted ALB+ cells using real-time quantitative PCR. All data were normalized to the value of ALB sorted cells and fold-differences are shown. Representative data from a transfected stem cell clone are shown; three samples were examined for each protein. Data are mean±s.d. (D) (a-f) Several sorted ALB cells could form clonal colonies including both ALB+ and ALBcells at day 20. (g-i) Moreover, ALB+ sorted cells gave rise to ALB cells even 1 day after the initiation of culture, and finally formed mosaic colonies similar to those from ALBcells (black arrowhead in g, an original EGFP+ cell; white arrowhead in g, a daughter cell). (a,d,g) Phase contrast. (b,e,h) Enhanced green fluorescence protein (EGFP) imaging. (c,f,i) Merge. d-f are magnifications of a-c, respectively. Scale bars: 100 μm.

Following independent culture, both ALB and ALB+ cells gave rise to ALB+ as well as ALB cells as assessed by FACS and a confocal microscopy(Fig. 2A,D). In addition, those sorted ALB and ALB+ cells capable of proliferation in culture, both possessed the capacity for differentiation into cholangiocytes (data not shown). These results demonstrated that ALB expression begins at a very early stage of stem cell differentiation, and that lineage specification into either hepatocytes or cholangiocytes cannot be determined by its expression. Therefore, early-differentiated ALB+cells may represent a bipotent hepatic precursor such as hepatoblasts, which express ALB and AFP, but are still capable of differentiating into cholangiocytes in the developing liver(Shiojiri et al., 1991).

Regulation of growth and differentiation of purified ALB and ALB+ cells by GFs and ECMs

To examine the effects of GFs and ECMs on ALB and ALB+ cells, sorted cells were independently cultured in six-well plates (1×104 cells/well) using several conditions and examined 10 days later. Although single cell cultures of sorted cells required ECMs and HGF similar to primary cultures, this high number of purified cells allowed slow growth even without GFs and ECMs. For ALBsorted cells, HGF strongly induced their proliferation. FGFs had a smaller positive effect, while OSM, by contrast, inhibited their proliferation(Fig. 3A, upper left graph). For ALB+ sorted cells, extensive proliferation was also found when cultured with HGF, but not with OSM and FGFs(Fig. 3A, lower left graph). The laminin, type IV collagen, and type I collagen were more effective on ALB+ cells than ALB cells(Fig. 3A, right graph). FACS analysis of cultured cells indicated that induction of ALB+ cells from ALB sorted cells was stimulated by HGF and OSM, and, to a lesser extent, by laminin, type IV collagen and type I collagen(Fig. 3B, upper graph). By contrast, HGF and OSM strongly inhibited the generation of ALB cells from ALB+ sorted cells. Laminin, type IV collagen and type I collagen-coated dishes had similar inhibitory effects on the generation of ALB cells(Fig. 3B, lower graph). Interestingly, aFGF and bFGF, which induce hepatogenesis in ventral endoderm at E8 (Jung et al., 1999), had little effect on stem cell growth and differentiation in our culture system.

Fig. 3.

Growth and differentiation of sorted ALB and ALB+ cells cultured with GFs and ECMs. FACS-sorted ALB and ALB+ cells were cultured separately. For the GF test, cells were cultured with or without HGF, OSM, aFGF, bFGF or a mixture of all GFs on type IV collagen-coated dishes. For the ECM test, cells were cultured on either non-coated, or laminin-, type IV collagen-, type I collagenor fibronectin-coated dishes. After 10 days in culture, proliferation(A) and differentiation (B) of sorted cells were examined by cell counting (A)or FACS (B). Representative data from a transfected stem cell clone are shown;three samples were examined for each GF or ECM. Data are mean±s.d. (A)HGF strongly promoted the proliferation of both ALB and ALB+ sorted cells, whereas lesser effects were noted with OSM, aFGF and bFGF. OSM specifically suppressed the proliferation of ALB cells. For the ECMs, laminin, type IV collagen and type I collagen induced the proliferation of ALB+ cells, but to a lesser degree than HGF. All data were normalized to the value of no-GF (GF test) or a non-coated dish (ECM test) and fold-differences are shown. (B) FACS-analysis revealed that HGF and OSM have the potential to induce ALB+ cells from ALB sorted cells, and inhibit generation of ALB cells from ALB+ sorted cells. Laminin, type IV collagen and type I collagen possessed similar, but lesser, effects than did HGF and OSM. All data are normalized to the value at the day of sorting(day 0) (GF test) or to the value of a non-coated dish (ECM test) and fold differences are shown.

Fig. 3.

Growth and differentiation of sorted ALB and ALB+ cells cultured with GFs and ECMs. FACS-sorted ALB and ALB+ cells were cultured separately. For the GF test, cells were cultured with or without HGF, OSM, aFGF, bFGF or a mixture of all GFs on type IV collagen-coated dishes. For the ECM test, cells were cultured on either non-coated, or laminin-, type IV collagen-, type I collagenor fibronectin-coated dishes. After 10 days in culture, proliferation(A) and differentiation (B) of sorted cells were examined by cell counting (A)or FACS (B). Representative data from a transfected stem cell clone are shown;three samples were examined for each GF or ECM. Data are mean±s.d. (A)HGF strongly promoted the proliferation of both ALB and ALB+ sorted cells, whereas lesser effects were noted with OSM, aFGF and bFGF. OSM specifically suppressed the proliferation of ALB cells. For the ECMs, laminin, type IV collagen and type I collagen induced the proliferation of ALB+ cells, but to a lesser degree than HGF. All data were normalized to the value of no-GF (GF test) or a non-coated dish (ECM test) and fold-differences are shown. (B) FACS-analysis revealed that HGF and OSM have the potential to induce ALB+ cells from ALB sorted cells, and inhibit generation of ALB cells from ALB+ sorted cells. Laminin, type IV collagen and type I collagen possessed similar, but lesser, effects than did HGF and OSM. All data are normalized to the value at the day of sorting(day 0) (GF test) or to the value of a non-coated dish (ECM test) and fold differences are shown.

In addition to FACS-analyses we further examined the effect of HGF and OSM on stem cell cultures to determine in detail the mechanism by which they induce differentiation of hepatocyte-lineage cells. After 10 days culture of either ALB or ALB+ sorted cells, quantitative PCR was performed to examine the expression patterns of hepatocyte or cholangiocyte differentiation markers. In ALB sorted cell cultures, HGF strongly induced ALB and αAT expression, and OSM induced G6P expression, which is normally activated at the mid-late gestational stage. TO, a marker for mature hepatocytes, expressed only in adult liver, could not be detected because of its much lower expression(Fig. 4A, left). Although HGF also induced expression of ALB and αAT in ALB+ sorted cell cultures, it did not induce latter markers of hepatocyte differentiation, such as G6P and TO. By contrast, OSM efficiently induced the expression of G6P and TO, and promoted the maturation of differentiating hepatocytes(Fig. 4A, right). In cultures with either HGF or OSM, AFP expression was decreased in both ALB and ALB+ sorted cell cultures. Cholangiocyte marker gene expression was also suppressed in cultures of both cell types by HGF and OSM (Fig. 4B). The laminin, type IV collagen and type I collagen promoted the expression of hepatocyte marker genes in both ALB and ALB+ cell cultures. Their effects, however, were not as strong as those of HGF and OSM,because relatively high expression of hepatocyte-differentiation marker genes was not detected until 20 days of culture and, furthermore, the expression of TO was too low to be determined by quantitative analysis (data not shown).

Fig. 4.

Quantitative analysis of the effect of HGF and OSM on sorted ALB and ALB+ cells. FACS-sorted ALB and ALB+ cells were cultured separately with or without HGF or OSM on type IV collagen-coated dishes for 10 days. Then,quantitative PCR was performed to determine the expression of several hepatocyte or cholangiocyte marker genes. Representative data from a transfected stem cell clone are shown; three samples were examined for each GF. Data are mean±s.d. (A) In the ALB cell culture,HGF strongly induced ALB and αAt expression. The mid-latter marker G6P was induced exclusively by OSM, but TO expression was still not detected. In the ALB+ cell culture, however, OSM strongly induced G6P and TO expression. The expression of ALB in the ALB+ sorted cell cultures was also promoted by HGF and OSM, but its effects were not greater than in ALB cells. (B) Both HGF and OSM suppressed the expression of cholangiocyte marker genes such as CK19 and GGT in ALB and ALB+ sorted cell cultures. All data were normalized to the values of no-GF and fold differences are shown. ND, not detected.

Fig. 4.

Quantitative analysis of the effect of HGF and OSM on sorted ALB and ALB+ cells. FACS-sorted ALB and ALB+ cells were cultured separately with or without HGF or OSM on type IV collagen-coated dishes for 10 days. Then,quantitative PCR was performed to determine the expression of several hepatocyte or cholangiocyte marker genes. Representative data from a transfected stem cell clone are shown; three samples were examined for each GF. Data are mean±s.d. (A) In the ALB cell culture,HGF strongly induced ALB and αAt expression. The mid-latter marker G6P was induced exclusively by OSM, but TO expression was still not detected. In the ALB+ cell culture, however, OSM strongly induced G6P and TO expression. The expression of ALB in the ALB+ sorted cell cultures was also promoted by HGF and OSM, but its effects were not greater than in ALB cells. (B) Both HGF and OSM suppressed the expression of cholangiocyte marker genes such as CK19 and GGT in ALB and ALB+ sorted cell cultures. All data were normalized to the values of no-GF and fold differences are shown. ND, not detected.

These results demonstrated that HGF can initiate the primary differentiation of stem cells into ALB+ hepatic precursors. HGF subsequently works as a mitogen to expand early-differentiating cells, while OSM inhibits their proliferation and promotes their differentiation into mature cells. Both HGF and OSM direct stem cells to the hepatocyte lineage,and also suppress development into the bile duct lineage. Several ECMs such as laminin, type IV collagen and type I collagen also induce differentiation of stem cells, in a similar manner to HGF and OSM, but their effect is much less dramatic than that of GFs. ECMs may function in the maintenance of stem cells and play a supportive role in their differentiation. Alternatively, ECMs may not affect stem cells directly, but instead could act on stem cell-derived precursors to promote their survival and differentiation.

The expression of C/EBPs changed dramatically during stem cell differentiation

To reveal the transcriptional control of hepatocyte-differentiation from stem cells, we first examined the effect of HGF on the expression of C/EBPs,which are candidate key factors. C/EBP proteins comprise a family of transcription factors that have a bZIP structure, consisting of a DNA binding basic region and a leucine zipper dimerization domain (Lekstrom-Himes et al.,1998). They directly control the expression of genes encoding hepatocyte-specific proteins such as ALB and αAT(Costa et al., 1989; Maire et al., 1989; Trautwein et al., 1996). The expression of C/EBPα is particularly upregulated when hepatocytes shift from proliferation to the differentiation state(Rana et al., 1994; Runge et al., 1997). In the fetal liver of Cebpa–/– mice, hepatocytes exhibited biliary epithelial cell characteristics and many pseudoglandular structures appeared, suggesting an involvement of C/EBPα in directing differentiation of bipotent hepatic stem cells along the hepatocyte-lineage(Tomizawa et al., 1998). As shown in Fig. 5A, the expression of C/EBPα in an original stem cell population was stimulated,in a dose-dependent manner, by HGF as well as laminin, type IV collagen and type I collagen. The differentiation of hepatocytes, represented by the expression of ALB, αAT and G6P, was also stimulated in stem cell cultures by high concentrations of HGF in the presence of ECMs (data not shown). By contrast, the expression of C/EBPβ was decreased by these culture conditions (Fig. 5A). HGF and ECMs were found to have little affect on the expression of C/EBPδ and C/EBPγ (data not shown).

Fig. 5.

HGF and several ECMs regulate C/EBP expression during stem cell differentiation. (A) When cells were cultured on type IV collagen-coated dishes for 8 days, C/EBPα expression was promoted in a dose-dependent manner by HGF, as well as by the presence of laminin, type IV collagen and type I collagen. By contrast, C/EBPβ expression was decreased in these culture conditions. All data were normalized to the expression values from 0 ng/ml (HGF) or non-coated dishes (ECM) and fold differences are shown. Representative data from a stem cell clone are shown; three samples were examined for each concentration of HGF or each ECM. Data are mean±s.d.(B) C/EBPα was highly expressed in ALB and ALB+ sorted cells cultured with HGF for 10 days. Its expression was promoted exclusively in the transition of ALB cells to ALB+ hepatic precursors induced by HGF. All data were normalized to the value of no-GF and fold differences are shown. Representative data from a transfected stem cell clone are shown; three samples were examined for each GF. Data are mean±s.d.

Fig. 5.

HGF and several ECMs regulate C/EBP expression during stem cell differentiation. (A) When cells were cultured on type IV collagen-coated dishes for 8 days, C/EBPα expression was promoted in a dose-dependent manner by HGF, as well as by the presence of laminin, type IV collagen and type I collagen. By contrast, C/EBPβ expression was decreased in these culture conditions. All data were normalized to the expression values from 0 ng/ml (HGF) or non-coated dishes (ECM) and fold differences are shown. Representative data from a stem cell clone are shown; three samples were examined for each concentration of HGF or each ECM. Data are mean±s.d.(B) C/EBPα was highly expressed in ALB and ALB+ sorted cells cultured with HGF for 10 days. Its expression was promoted exclusively in the transition of ALB cells to ALB+ hepatic precursors induced by HGF. All data were normalized to the value of no-GF and fold differences are shown. Representative data from a transfected stem cell clone are shown; three samples were examined for each GF. Data are mean±s.d.

After sorting and culture of ALB and ALB+cells, C/EBPα was highly expressed in culture of ALBcells with HGF, in which hepatic precursors actively differentiated from stem cells and proliferated, rather than ALB+ cells(Fig. 5B). As C/EBPαexpression was also synchronized with the expression of hepatocyte-differentiation genes such as ALB and αAT in culture of ALB cells with HGF, as described in Fig. 4A (left), C/EBPa probably has a role in the primary differentiation of stem cells into hepatocyte-lineage cells.

Lack of C/EBP proteins inhibits hepatocyte differentiation from stem cells

Using a retroviral gene transfer system, we expressed C/EBPα orβ in stem cell cultures to examine their roles in differentiation. Transduced cells, however, stopped proliferating and died within a few days(data not shown). This suggested that too much C/EBP disrupted the homeostasis of hepatic stem cells. Therefore, we transduced cells with the retroviral vector GCsam-A-C/EBP-IRES-EGFP, which drives expression of both A-C/EBP and EGFP (Iwama et al., 2002),which disrupts the function of C/EBP proteins. A-C/EBP, a dominant-negative C/EBP that has the potential to antagonize all C/EBP members, is a 102 amino acid protein consisting of an N-terminal 9 amino acid Flag epitope, a 13 amino acid linker, a 31 amino acid designated acidic amphipathic helix, and a 49 amino acid leucine zipper domain of C/EBPα(Olive et al., 1996). The leucine zipper from A-C/EBP specifically interacts with endogenous C/EBP leucine zippers, and the N-terminal acidic extension forms a coiled coil with endogenous C/EBP basic regions. This heterodimeric coiled coil structure is much more stable than C/EBPα bound to DNA, and thus, the dominant-negative protein abolishes DNA binding of all endogenous C/EBP family members.

After transduction, EGFP-positive cells (98.2±0.8%; n=3)were sorted by FACS and then subjected to in vitro and in vivo assays. Stem cells expressing A-C/EBP failed to express hepatocyte-differentiation markers such as ALB, AFP, αAT, G6P and TO, even when cultured with both HGF and OSM, which normally induce hepatocyte differentiation(Fig. 6A). By contrast, the expression of CK19 and GGT, markers of cholangiocyte-differentiation, was relatively enhanced in cells expressing A-C/EBP. The expression of the transcription factors HNF1, HNF4 and Met was also activated by blocking C/EBP function, but this change was not significant. Interestingly, the expression of HNF6 was decreased in transduced cells, in a similar manner to other hepatocyte-differentiation markers (Fig. 6B). Both ALB immunocytochemistry and PAS staining also revealed that stem cells expressing A-C/EBP did not give rise to functionally mature hepatocytes expressing ALB and containing abundant glycogen stores(Fig. 6C). Instead of differentiation along the hepatocyte-lineage, cultured cells expressing A-C/EBP grew actively and formed a lot of large colonies including more than 100 cells in comparison with mock controls, suggesting activation or maintenance of self-renewal status (Fig. 6D). Sorted ALB+ and ALB cells were also transduced with the retroviral vector GCsam-A-C/EBP-IRES-NGFR and analyzed for differentiation potential into hepatocyte-lineage cells. In ALB+ cultured cells that expressed A-C/EBP(ALB+/NGFR+ cells) after FACS sorting,ALB cells emerged efficiently. Transduction of ALB cells, by contrast, strongly inhibited the generation of ALB+ cells (Fig. 6E). Taken together, these data indicate that the expression of C/EBPs, induced by HGF, OSM and ECMs, is a key event for primary differentiation of stem cells into bipotent precursors and hepatocyte-lineage cells, and that disruption of this transcriptional regulation leads to the maintenance of stem cell status, including self-renewal activity.

Fig. 6.

Dominant-negative C/EBP (A-C/EBP) inhibits hepatocyte differentiation from stem cells. (A,B) After culturing FACS-sorted EGFP+ transduced cells for 10 or 20 days, quantitative PCR was performed. The expression of hepatocyte-differentiation markers ALB, AFP, αAt, G6P and TO in the stem cell population was significantly suppressed even when cultured with both HGF(H) and OSM (O), which induce hepatocyte differentiation. By contrast, the expression of CK19 and GGT, which are markers of cholangiocyte differentiation, was elevated in cells expressing A-C/EBP. The expression of HNF1, HNF4 and Met was also relatively enhanced, but not significantly. The expression of HNF6, however, was decreased in transduced cells, in a similar fashion to hepatocyte marker genes. All data were normalized to the value of mock controls (day 10; no GF) and fold differences are shown. Representative data from a stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. (C) Immunocytochemical staining of ALB and PAS-staining were performed on cells expressing EGFP (mock) or A-C/EBP at day 20. In cells expressing A-C/EBP, few cells were positive for ALB and stored glycogen. (counterstain: Hematoxylin). (D) After culture of sorted EGFP(mock) or A-C/EBP-expressing cells for 5 days in clonal density cultures(5×102 cells/well in six-well plates), the number of large(>100 cells in a colony) and small colonies was counted (n=6). The number of large colonies was increased when cells were transduced with A-C/EBP. Data are mean±s.d. (*P<0.005). (E) After transduction of FACS-sorted ALB+ or ALB cells by A-C/EBP-NGFR and culture of each cell population for 14 days,ALB+/NGFR+ cells or ALB/NGFR+ cells were sorted separately and cultured with or without HGF and OSM for 10 days (plating density is 1×103 cells/cm2). Then, the generation of ALB cells from ALB+/NGFR+ sorted cells and the generation of ALB+ cells from ALB/NGFR+ sorted cells was examined by using FACS. Compared with mock controls, cells expressing A-C/EBP failed to maintain ALB expression in ALB+ sorted cells and were unable to generate ALB+ cells from ALB sorted cells. Representative data from a transfected stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. Scale bars: 100 μm.

Fig. 6.

Dominant-negative C/EBP (A-C/EBP) inhibits hepatocyte differentiation from stem cells. (A,B) After culturing FACS-sorted EGFP+ transduced cells for 10 or 20 days, quantitative PCR was performed. The expression of hepatocyte-differentiation markers ALB, AFP, αAt, G6P and TO in the stem cell population was significantly suppressed even when cultured with both HGF(H) and OSM (O), which induce hepatocyte differentiation. By contrast, the expression of CK19 and GGT, which are markers of cholangiocyte differentiation, was elevated in cells expressing A-C/EBP. The expression of HNF1, HNF4 and Met was also relatively enhanced, but not significantly. The expression of HNF6, however, was decreased in transduced cells, in a similar fashion to hepatocyte marker genes. All data were normalized to the value of mock controls (day 10; no GF) and fold differences are shown. Representative data from a stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. (C) Immunocytochemical staining of ALB and PAS-staining were performed on cells expressing EGFP (mock) or A-C/EBP at day 20. In cells expressing A-C/EBP, few cells were positive for ALB and stored glycogen. (counterstain: Hematoxylin). (D) After culture of sorted EGFP(mock) or A-C/EBP-expressing cells for 5 days in clonal density cultures(5×102 cells/well in six-well plates), the number of large(>100 cells in a colony) and small colonies was counted (n=6). The number of large colonies was increased when cells were transduced with A-C/EBP. Data are mean±s.d. (*P<0.005). (E) After transduction of FACS-sorted ALB+ or ALB cells by A-C/EBP-NGFR and culture of each cell population for 14 days,ALB+/NGFR+ cells or ALB/NGFR+ cells were sorted separately and cultured with or without HGF and OSM for 10 days (plating density is 1×103 cells/cm2). Then, the generation of ALB cells from ALB+/NGFR+ sorted cells and the generation of ALB+ cells from ALB/NGFR+ sorted cells was examined by using FACS. Compared with mock controls, cells expressing A-C/EBP failed to maintain ALB expression in ALB+ sorted cells and were unable to generate ALB+ cells from ALB sorted cells. Representative data from a transfected stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. Scale bars: 100 μm.

Fig. 6.

Dominant-negative C/EBP (A-C/EBP) inhibits hepatocyte differentiation from stem cells. (A,B) After culturing FACS-sorted EGFP+ transduced cells for 10 or 20 days, quantitative PCR was performed. The expression of hepatocyte-differentiation markers ALB, AFP, αAt, G6P and TO in the stem cell population was significantly suppressed even when cultured with both HGF(H) and OSM (O), which induce hepatocyte differentiation. By contrast, the expression of CK19 and GGT, which are markers of cholangiocyte differentiation, was elevated in cells expressing A-C/EBP. The expression of HNF1, HNF4 and Met was also relatively enhanced, but not significantly. The expression of HNF6, however, was decreased in transduced cells, in a similar fashion to hepatocyte marker genes. All data were normalized to the value of mock controls (day 10; no GF) and fold differences are shown. Representative data from a stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. (C) Immunocytochemical staining of ALB and PAS-staining were performed on cells expressing EGFP (mock) or A-C/EBP at day 20. In cells expressing A-C/EBP, few cells were positive for ALB and stored glycogen. (counterstain: Hematoxylin). (D) After culture of sorted EGFP(mock) or A-C/EBP-expressing cells for 5 days in clonal density cultures(5×102 cells/well in six-well plates), the number of large(>100 cells in a colony) and small colonies was counted (n=6). The number of large colonies was increased when cells were transduced with A-C/EBP. Data are mean±s.d. (*P<0.005). (E) After transduction of FACS-sorted ALB+ or ALB cells by A-C/EBP-NGFR and culture of each cell population for 14 days,ALB+/NGFR+ cells or ALB/NGFR+ cells were sorted separately and cultured with or without HGF and OSM for 10 days (plating density is 1×103 cells/cm2). Then, the generation of ALB cells from ALB+/NGFR+ sorted cells and the generation of ALB+ cells from ALB/NGFR+ sorted cells was examined by using FACS. Compared with mock controls, cells expressing A-C/EBP failed to maintain ALB expression in ALB+ sorted cells and were unable to generate ALB+ cells from ALB sorted cells. Representative data from a transfected stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. Scale bars: 100 μm.

Fig. 6.

Dominant-negative C/EBP (A-C/EBP) inhibits hepatocyte differentiation from stem cells. (A,B) After culturing FACS-sorted EGFP+ transduced cells for 10 or 20 days, quantitative PCR was performed. The expression of hepatocyte-differentiation markers ALB, AFP, αAt, G6P and TO in the stem cell population was significantly suppressed even when cultured with both HGF(H) and OSM (O), which induce hepatocyte differentiation. By contrast, the expression of CK19 and GGT, which are markers of cholangiocyte differentiation, was elevated in cells expressing A-C/EBP. The expression of HNF1, HNF4 and Met was also relatively enhanced, but not significantly. The expression of HNF6, however, was decreased in transduced cells, in a similar fashion to hepatocyte marker genes. All data were normalized to the value of mock controls (day 10; no GF) and fold differences are shown. Representative data from a stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. (C) Immunocytochemical staining of ALB and PAS-staining were performed on cells expressing EGFP (mock) or A-C/EBP at day 20. In cells expressing A-C/EBP, few cells were positive for ALB and stored glycogen. (counterstain: Hematoxylin). (D) After culture of sorted EGFP(mock) or A-C/EBP-expressing cells for 5 days in clonal density cultures(5×102 cells/well in six-well plates), the number of large(>100 cells in a colony) and small colonies was counted (n=6). The number of large colonies was increased when cells were transduced with A-C/EBP. Data are mean±s.d. (*P<0.005). (E) After transduction of FACS-sorted ALB+ or ALB cells by A-C/EBP-NGFR and culture of each cell population for 14 days,ALB+/NGFR+ cells or ALB/NGFR+ cells were sorted separately and cultured with or without HGF and OSM for 10 days (plating density is 1×103 cells/cm2). Then, the generation of ALB cells from ALB+/NGFR+ sorted cells and the generation of ALB+ cells from ALB/NGFR+ sorted cells was examined by using FACS. Compared with mock controls, cells expressing A-C/EBP failed to maintain ALB expression in ALB+ sorted cells and were unable to generate ALB+ cells from ALB sorted cells. Representative data from a transfected stem cell clone are shown; three samples were examined for each condition. Data are mean±s.d. Scale bars: 100 μm.

To describe precisely sequential liver cell lineages derived from hepatic stem cells, we analyzed the differentiation status of purified stem cell populations in clonal experiments, after the exclusion of a number of differentiated cells and other cell lineages in the liver. Our previous report made it possible to isolate primary cells with stem cell activity from fetal mouse livers using FACS and culture them from single cells(Suzuki et al., 2002). Based on our present findings, we propose a possible lineage of stem cell differentiation in the developing liver(Fig. 7). Analysis of isolated stem cells revealed that initiation of the early differentiation of hepatocyte-lineage cells from stem cells was directly mediated by HGF. This identifies a novel function of HGF in liver development. In postnatal liver,HGF functions as an inducer of hepatocyte maturation(Hu et al., 1993; Kamiya et al., 2001), and in the regenerating liver, it stimulates proliferation of adult hepatocytes(Michalopoulos et al., 1984; Ishiki et al., 1992). HGF has been suggested to serve a role in the physiological activities of organ-specific stem cells owing to its ability to elicit repair and regeneration of many adult organs (Kawaida et al., 1994; Michalopoulos and DeFrances, 1997; Matsuda et al., 1997; Matsumoto et al., 1997; Stolz et al., 1999; Menke et al., 1999; Fausto, 2000; Xian et al., 2000; Sakamaki et al., 2002). Indeed, it stimulates differentiation and proliferation of human CD34-positive hematopoietic precursors and human embryonic stem cells(Galimi et al., 1994; Schuldiner et al., 2000). Thus, direct regulation by HGF may be a common mechanism among stem cells in multiple organs and critical for their differentiation and proliferation. We found that during the transition to ALB+ cells from ALB stem cells, HGF promotes production of ALB+cells and allows them to proliferate efficiently. The ALB+ cells produced still possess the capacity to differentiate into cholangiocytes expressing CK19 and GGT in clonal cultures. Therefore, differentiating and proliferating ALB+ cells freshly generated from ALB stem cells may be equivalent to bi-potent hepatoblasts,which express several lineage markers and largely occupy the developing liver. Our previous data have also shown that isolated hepatic stem cells reside in the developing liver without expressing both hepatocyte and cholangiocyte lineage markers (Suzuki et al.,2000; Suzuki et al.,2002). Furthermore, such cells were much fewer in number than previously hypothesized and appear to decrease as gestation advances. These consolidated findings support our present hypothesis described above.

Fig. 7.

Summarized possible mechanism for hepatic stem cell differentiation. Because early-generated ALB+ cells can proliferate exclusively and differentiate into cholangiocyte-lineage cells, we speculate that transitory hepatic precursors expressing ALB exist in the developing liver and they are equivalent to bi-potent hepatoblasts.

Fig. 7.

Summarized possible mechanism for hepatic stem cell differentiation. Because early-generated ALB+ cells can proliferate exclusively and differentiate into cholangiocyte-lineage cells, we speculate that transitory hepatic precursors expressing ALB exist in the developing liver and they are equivalent to bi-potent hepatoblasts.

After the initiation of hepatic stem cell differentiation by HGF, OSM,which is produced by hematopoietic lineage cells, gave a permissive signal for differentiation of hepatic precursors to mature hepatocytes expressing G6P and TO. Both HGF and OSM strongly stimulated differentiation of ALB+hepatic precursors into hepatocyte-lineage cells, and also prevented them from restoring the ALB phenotype and differentiating into cholangiocyte-lineage cells. Several ECMs, such as laminin, type IV collagen and type I collagen, also induced differentiation of stem cells. Although ECMs are necessary for proliferation and differentiation of stem cells, their effects are likely to be only supportive for the differentiation of hepatocytes and cholangiocytes, based on their weaker potential for stem cell differentiation. Factors regulating cholangiocyte differentiation from stem cells or hepatic precursors were not identified in this study, suggesting that stem cell to cholangiocyte differentiation may be a favored pathway in stem cell differentiation, that does not require specific signals. Further intensive studies will be necessary to elucidate the manner of cholangiocyte differentiation from stem cells.

In ALB cells generated from ALB+ precursors in clonal cultures, both CK19+ cholangiocyte-lineage cells and ALB CK19 cells that possess a stem cell phenotype were identified (data not shown). These results suggest that ALB expression early during the differentiation of stem cells is flexible and cells can flow between stem cells and ALB+ hepatic precursors before they obtain gradual signals for differentiation first from HGF and secondarily from OSM. In hepatogenesis from the endoderm layer starting at E8,FGFs produced by cardiac mesoderm play a key role in the generation of ALB+ cells (Jung et al.,1999). Our present data, however, showed that FGFs have much smaller effects on the transition from ALB stem cells to ALB+ cells. The E13.5 fetal mouse livers that we used for isolating stem cells had already been apart from cardiac mesoderm, indicating that FGFs should have finished their role in early liver development by this stage. A few cells receiving no signals from cardiac mesoderm or other lining mesenchymal cells in early hepatogenesis may be maintained in an undifferentiated state until the E13.5 mid-fetal stage. FGFs may directly induce the differentiation of hepatocytes and/or ALB+ hepatic precursors from foregut endoderm, but HGF stimulates the differentiation of dormant stem cells preserved in developing livers. A number of factors are likely to work mutually as inducers or repressors of liver organogenesis based on subtle timing.

C/EBP proteins regulate liver-specific gene expression and cell proliferation (Costa et al.,1989; Maire et al.,1989; Rana et al.,1994; Trautwein et al.,1996; Soriano et al.,1998; Greenbaum et al.,1998). In particular, C/EBPα is highly expressed in quiescent hepatocytes and positively regulates hepatocyte-specific gene expression, such as ALB and αAt(Costa et al., 1989; Maire et al., 1989; Rana et al., 1994; Runge et al., 1997; Soriano et al., 1998). In the developing liver of C/EBPα knockout mice, a number of pseudoglandular structures that co-express antigens specific for hepatocytes and cholangiocytes were found in the liver parenchyma, but the formation of bile ducts was not affected (Tomizawa et al.,1998). These data demonstrated that C/EBPα is an important regulator of hepatocyte differentiation, but not of cholangiocytes in either liver development or regeneration. In the results presented here, C/EBPαexpression was highly induced during the progression of hepatocyte differentiation from ALB stem cells to ALB+hepatic precursors by HGF. We suggest that HGF directly regulates the expression of C/EBPα, which plays a crucial role in the transition of stem cells to ALB+ hepatic precursors. C/EBPb is also a liver enriched transcription factor (Descombes et al., 1990), which, similar to C/EBPα, is involved in the regulation of liver-specific genes such as ALB(Trautwein et al., 1996). In the transition from proliferating to differentiated hepatocytes, the C/EBPα:C/EBPβ ratio was found to be increased(Runge et al., 1997). In the present study, during the induction of hepatocyte-differentiation by HGF, the C/EBPα:C/EBPβ ratio was also highly increased. These results demonstrate that the relative proportions of C/EBPα and C/EBPβ may be important for hepatocyte differentiation from stem cells in the developing liver.

Using knockout mice, several functions for C/EBPs in liver development have been identified (Wang et al.,1995; Soriano et al.,1998; Tomizawa et al.,1998). However, as redundancy exists among C/EBP family members,their intrinsic roles in liver development remain to be clarified. For example, liver cells in C/EBPα knockout mice could differentiate into hepatocytes when they were cultured on Matrigel(Soriano et al., 1998),suggesting that partial redundancy with other C/EBP proteins exists. To eliminate these complicated interpretations, we used the dominant-negative A-C/EBP to abolish endogenous DNA binding of all C/EBP family members. The expression of A-C/EBP in ALB stem cells resulted in semi-complete inhibition of the generation of hepatocyte-lineage cells, even in cultures including both HGF and OSM. In addition, A-C/EBP expression in ALB+ cells advanced the transition to ALB cells. These findings, collectively, show that the C/EBP family, especially C/EBPα and β, which are directly regulated by HGF, are required for the early steps in hepatic stem cell differentiation. In addition to inhibiting differentiation, lack of all C/EBP functions in stem cells enhances their self-renewal divisions in culture. In the developing liver, however, the number of stem cells is very low and their proliferation is restricted, even with the low expression of C/EBP proteins. Thus, other mechanisms may exist to maintain their quiescent status in liver development.

A number of molecular events in the differentiation of hepatic stem cells,such as the interaction of C/EBPs and other elements, are required for the determination of hepatocyte or cholangiocyte lineage. C/EBPα-mediated growth arrest is known to require interaction with p21, a cyclin-dependent kinase (CDK) inhibitor and CDK2 (Timchenko et al., 1997; Harris et al.,2001). This mechanism may be involved in OSM-mediated growth arrest of hepatic precursors, in order to then induce differentiation into mature hepatocytes. Because such growth inhibition is irrelevant to the transcriptional activity of C/EBPα(Harris et al., 2001), another mechanism should exist to control transcription of genes important for liver development. Actually, in the transcriptional control regions of ALB andαAt, liver-enriched transcription factors such as HNF1(Baumhueter et al., 1990), HNF3 and HNF4 (Costa et al., 1989),and widely distributed proteins such as activator protein 1 (AP1) (Hu et al.,1994) bind to regulate these genes along with C/EBP proteins. The currently proposed cascade of sequential transcriptional control of hepatic stem cell differentiation, however, is still unreliable. Our present results show that HNF6 was expressed more highly in ALB+ cells than ALB cells during the early differentiation of stem cells,and HNF6 expression was also suppressed when stem cell differentiation was inhibited by dominant-negative C/EBP proteins. The HNF6-binding sequence, in fact, is present in the promoter regions of several hepatocyte-enriched genes,such as αAt, AFP, cytochrome P450, GLUT2 and TO (Samadani et al., 1996; Tan et al., 2002). Thus, HNF6 may be one possible candidate involved in the early transition of stem cells to hepatic precursors, and its expression may be regulated by C/EBPs. In Hnf6–/– mice, abnormalities of the intrahepatic and extrahepatic bile ducts and of the gallbladder were observed(Clotman et al., 2002). These data suggest that HNF6 is essential for differentiation and maturation of biliary lineage cells rather than hepatocytes. However, in the E13.5 developing liver of these mice a number of cytokeratin-positive biliary lineage cells emerged compared to normal mice, suggesting that HNF6 regulates not only morphogenesis of biliary tract but turning point of the differentiation of primitive hepatic stem cells.

A precise description of stem cell differentiation would allow the control of hepatocyte differentiation and the induction of liver regeneration by manipulating the endogenous stem cell compartment. Our clonal culture assay with hepatic stem cells should reveal the mechanism that regulates their self-renewal potential and the signals that restrict their proliferation and differentiation in the developing liver. Exploring diverse gene programs activated in stem cells or differentiating cells should provide a molecular framework for future research into liver development. Key elements of stem cells, such as quiescent status and pluripotency, would be elucidated by comparing hepatic stem cells with other tissue-derived stem cells. The prospective isolation and characterization of stem cells is required for better understanding what exactly a stem cell is and what it does.

We thank N. Ukawa, M. Mori and Y. Jinzenji for technical support, and Y. Morita for FACS operation. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan (14207046, 12557096) and JSPS Research Fellowship for Young Scientists. This work was also supported by a grant from NISSAN Science Foundation.

Baumhueter, S., Mendel, D. B., Conley, P. B., Kuo, C. J., Turk,C., Graves, M. K., Edwards, C. A., Courtois, G. and Crabtree, G. R.(
1990
). HNF-1 shares three sequence motifs with the POU domain proteins and is identical to LF-B1 and APF.
Genes Dev.
4
,
372
-379.
Clotman, F., Lannoy, V. J., Reber, M., Cereghini, S., Cassiman,D., Jacquemin, P., Roskams, T., Rousseau, G. G. and Lemaigre, F. P.(
2002
). The onecut transcription factor HNF6 is required for normal development of the biliary tract.
Development
129
,
1819
-1828.
Costa, R. H., Grayson, D. R. and Darnell, J. E., Jr(
1989
). Multiple hepatocyte-enriched nuclear factors function in the regulation of transthyretin and alpha 1-antitrypsin genes.
Mol. Cell. Biol.
9
,
1415
-1425.
Descombes, P., Chojkier, M., Lichtsteiner, S., Falvey, E. and Schibler, U. (
1990
). LAP, a novel member of the C/EBP gene family, encodes a liver-enriched transcriptional activator protein.
Genes Dev.
4
,
1541
-1551.
Fausto, N. (
2000
). Liver regeneration.
J. Hepatol.
32
,
19
-31.
Galimi, F., Bagnara, G. P., Bonsi, L., Cottone, E., Follenzi,A., Simeone, A. and Comoglio, P. M. (
1994
). Hepatocyte growth factor induces proliferation and differentiation of multipotent and erythroid hemopoietic progenitors.
J. Cell Biol.
127
,
1743
-1754.
Greenbaum, L. E., Li, W., Cressman, D. E., Peng, Y., Ciliberto,G., Poli, V. and Taub, R. (
1998
). CCAAT enhancer-binding protein beta is required for normal hepatocyte proliferation in mice after partial hepatectomy.
J. Clin. Invest.
102
,
996
-1007.
Harris, T. E., Albrecht, J. H., Nakanishi, M. and Darlington, G. J. (
2001
). CCAAT/enhancer-binding protein-alpha cooperates with p21 to inhibit cyclin-dependent kinase-2 activity and induces growth arrest independent of DNA binding.
J. Biol. Chem.
276
,
29200
-29209.
Hu, J. and Isom, H. C. (
1994
). Suppression of albumin enhancer activity by H-ras and AP-1 in hepatocyte cell lines.
Mol. Cell. Biol.
14
,
1531
-1543.
Hu, Z., Evarts, R. P., Fujio, K., Marsden, E. R. and Thorgeirsson, S. S. (
1993
). Expression of hepatocyte growth factor and c-met genes during hepatic differentiation and liver development in the rat.
Am. J. Pathol.
142
,
1823
-1830.
Ishiki, Y., Ohnishi, H., Muto, Y., Matsumoto, K. and Nakamura,T. (
1992
). Direct evidence that hepatocyte growth factor is a hepatotrophic factor for liver regeneration and has a potent antihepatitis effect in vivo.
Hepatology
16
,
1227
-1235.
Iwama, A., Osawa, M., Hirasawa, R., Uchiyama, N., Kaneko, S.,Onodera, M., Shibuya, K., Shibuya, A., Vinson, C., Tenen, D. G. et al.(
2002
). Reciprocal roles for CCAAT/enhancer binding protein(C/EBP) and PU.1 transcription factors in Langerhans cell commitment.
J. Exp. Med.
195
,
547
-558.
Jacquemin, P., Durviaux, S. M., Jensen, J., Godfraind, C.,Gradwohl, G., Guillemot, F., Madsen, O. D., Carmeliet, P., Dewerchin, M.,Collen, D. et al. (
2000
). Transcription factor hepatocyte nuclear factor 6 regulates pancreatic endocrine cell differentiation and controls expression of the proendocrine gene ngn3.
Mol. Cell. Biol.
20
,
4445
-4454.
Jung, J., Zheng, M., Goldfarb, M. and Zaret, K. S.(
1999
). Initiation of mammalian liver development from endoderm by fibroblast growth factors.
Science
284
,
1998
-2003.
Kamiya, A., Kinoshita, T., Ito, Y., Matsui, T., Morikawa, Y.,Senba, E., Nakashima, K., Taga, T., Yoshida, K., Kishimoto, T. et al.(
1999
). Fetal liver development requires a paracrine action of oncostatin M through the gp130 signal transducer.
EMBO J.
18
,
2127
-2136.
Kamiya, A., Kinoshita, T. and Miyajima, A.(
2001
). Oncostatin M and hepatocyte growth factor induce hepatic maturation via distinct signaling pathways.
FEBS Lett.
492
,
90
-94.
Kaneko, S., Onodera, M., Fujiki, Y., Nagasawa, T. and Nakauchi,H. (
2001
). The simplified retroviral vector GCsap with murine stem cell virus long terminal repeat allows high and continued expression of enhanced green fluorescent protein by human hematopoietic progenitors engrafted in non-obese diabetic / severe combined immunodeficiency mice.
Hum. Gene. Ther.
12
,
35
-44.
Kawaida, K., Matsumoto, K., Shimazu, H. and Nakamura, T.(
1994
). Hepatocyte growth factor prevents acute renal failure and accelerates renal regeneration in mice.
Proc. Natl. Acad. Sci. USA
91
,
4357
-4361.
Kinoshita, T., Sekiguchi, T., Xu, M. J., Ito, Y., Kamiya, A.,Tsuji, K., Nakahata, T. and Miyajima, A. (
1999
). Hepatic differentiation induced by oncostatin M attenuates fetal liver hematopoiesis.
Proc. Natl. Acad. Sci. USA
96
,
7265
-7270.
Lekstrom-Himes, J. and Xanthopoulos, K. G.(
1998
). Biological role of the CCAAT/enhancer-binding protein family of transcription factors.
J. Biol. Chem.
273
,
28545
-28548.
Maire, P., Wuarin, J. and Schibler, U. (
1989
). The role of cis-acting promoter elements in tissue-specific albumin gene expression.
Science
244
,
343
-346.
Matsuda, Y., Matsumoto, K., Yamada, A., Ichida, T., Asakura, H.,Komoriya, Y., Nishiyama, E. and Nakamura, T. (
1997
). Preventive and therapeutic effects in rats of hepatocyte growth factor infusion on liver fibrosis/cirrhosis.
Hepatology
26
,
81
-89.
Matsumoto, K. and Nakamura, T. (
1997
). Hepatocyte growth factor (HGF) as a tissue organizer for organogenesis and regeneration.
Biochem. Biophys. Res. Commun.
239
,
639
-644.
Menke, A., Yamaguchi, H., Giehl, K. and Adler, G.(
1999
). Hepatocyte growth factor and fibroblast growth factor 2 are overexpressed after cerulein-induced acute pancreatitis.
Pancreas
18
,
28
-33.
Michalopoulos, G. K. and DeFrances, M. C.(
1997
). Liver regeneration.
Science
276
,
60
-66.
Michalopoulos, G., Houck, K. A., Dolan, M. L. and Leutteke, N. C. (
1984
). Control of hepatocyte replication by two serum factors.
Cancer Res.
44
,
4414
-4419.
Miller, A. D. (
1996
). Cell-surface receptors for retroviruses and implications for gene transfer.
Proc. Natl. Acad. Sci. USA
93
,
11407
-11413.
Olive, M., Williams, S. C., Dezan, C., Johnson, P. F. and Vinson, C. (
1996
). Design of a C/EBP-specific,dominant-negative bZIP protein with both inhibitory and gain-of-function properties.
J. Biol. Chem.
271
,
2040
-2047.
Pinkert, C. A., Ornitz, D. M., Brinster, R. L. and Palmiter, R. D. (
1987
). An albumin enhancer located 10 kb upstream functions along with its promoter to direct efficient, liver-specific expression in transgenic mice.
Genes Dev.
1
,
268
-276.
Rana, B., Mischoulon, D., Xie, Y., Bucher, N. L. and Farmer, S. R. (
1994
). Cell-extracellular matrix interactions can regulate the switch between growth and differentiation in rat hepatocytes:reciprocal expression of C/EBP alpha and immediate-early growth response transcription factors.
Mol. Cell. Biol.
14
,
5858
-5869.
Rossi, J. M., Dunn, N. R., Hogan, B. L. and Zaret, K. S.(
2001
). Distinct mesodermal signals, including BMPs from the septum transversum mesenchyme, are required in combination for hepatogenesis from the endoderm.
Genes Dev.
15
,
1998
-2009.
Runge, D., Runge, D. M., Bowen, W. C., Locker, J. and Michalopoulos, G. K. (
1997
). Matrix induced re-differentiation of cultured rat hepatocytes and changes of CCAAT/enhancer binding proteins.
J. Biol. Chem.
378
,
873
-881.
Sakamaki, Y., Matsumoto, K., Mizuno, S., Miyoshi, S., Matsuda,H. and Nakamura, T. (
2002
). Hepatocyte growth factor stimulates proliferation of respiratory epithelial cells during postpneumonectomy compensatory lung growth in mice.
Am. J. Respir. Cell Mol. Biol.
26
,
525
-533.
Samadani, U. and Costa, R. H. (
1996
). The transcriptional activator hepatocyte nuclear factor 6 regulates liver gene expression.
Mol. Cell. Biol.
16
,
6273
-6284.
Schuldiner, M., Yanuka, O., Itskovitz-Eldor, J., Melton, D. A. and Benvenisty, N. (
2000
). From the cover: effects of eight growth factors on the differentiation of cells derived from human embryonic stem cells.
Proc. Natl. Acad. Sci. USA
97
,
11307
-11312.
Shiojiri, N., Lemire, J. M. and Fausto, N.(
1991
). Cell lineages and oval cell progenitors in rat liver development.
Cancer Res.
51
,
2611
-2620.
Soriano, H. E., Kang, D. C., Finegold, M. J., Hicks, M. J.,Wang, N. D., Harrison, W. and Darlington, G. J. (
1998
). Lack of C/EBP alpha gene expression results in increased DNA synthesis and an increased frequency of immortalization of freshly isolated mice [correction of rat] hepatocytes.
Hepatology
27
,
392
-401.
Spagnoli, F. M., Amicone, L., Tripodi, M. and Weiss, M. C.(
1998
). Identification of a bipotential precursor cell in hepatic cell lines derived from transgenic mice expressing cyto-Met in the liver.
J. Cell Biol.
143
,
1101
-1112.
Stolz, D. B., Mars, W. M., Petersen, B. E., Kim, T. H. and Michalopoulos, G. K. (
1999
). Growth factor signal transduction immediately after two-thirds partial hepatectomy in the rat.
Cancer Res.
59
,
3954
-3960.
Suzuki, A., Zheng, Y. W., Kondo, R., Kusakabe, M., Takada, Y.,Fukao, K., Nakauchi, H. and Taniguchi, H. (
2000
). Flow cytometric separation and enrichment of hepatic progenitor cells in the developing mouse liver.
Hepatology
32
,
1230
-1239.
Suzuki, A., Zheng, Y. W., Kaneko, S., Onodera, M., Fukao, K.,Nakauchi, H. and Taniguchi, H. (
2002
). Clonal identification and characterization of self-renewing pluripotent stem cells in the developing liver.
J. Cell Biol.
156
,
173
-184.
Tan, Y., Adami, G. and Costa, R. H. (
2002
). Maintaining HNF6 expression prevents AdHNF3beta-mediated decrease in hepatic levels of Glut-2 and glycogen.
Hepatology
35
,
790
-798.
Timchenko, N. A., Harris, T. E., Wilde, M., Bilyeu, T. A.,Burgess-Beusse, B. L., Finegold, M. J. and Darlington, G. J.(
1997
). CCAAT/enhancer binding protein alpha regulates p21 protein and hepatocyte proliferation in newborn mice.
Mol. Cell. Biol.
17
,
7353
-7361.
Tomizawa, M., Garfield, S., Factor, V. and Xanthopoulos, K. G. (
1998
). Hepatocytes deficient in CCAAT/enhancer binding protein alpha (C/EBP alpha) exhibit both hepatocyte and biliary epithelial cell character.
Biochem. Biophys. Res. Commun.
249
,
1
-5.
Trautwein, C., Rakemann, T., Pietrangelo, A., Plumpe, J.,Montosi, G. and Manns, M. P. (
1996
). C/EBP-beta/LAP controls down-regulation of albumin gene transcription during liver regeneration.
J. Biol. Chem.
271
,
22262
-22270.
Wang, N. D., Finegold, M. J., Bradley, A., Ou, C. N.,Abdelsayed, S. V., Wilde, M. D., Taylor, L. R., Wilson, D. R. and Darlington,G. J. (
1995
). Impaired energy homeostasis in C/EBP alpha knockout mice.
Science
269
,
1108
-1112.
Wilson, J. W., Groat, C. S. and Leduc, E. H.(
1963
). Histogenesis of the liver.
Ann. New York Acad. Sci.
111
,
8
-24.
Xian, C. J., Couper, R., Howarth, G. S., Read, L. C. and Kallincos, N. C. (
2000
). Increased expression of HGF and c-met in rat small intestine during recovery from methotrexate-induced mucositis.
Br. J. Cancer
82
,
945
-952.
Zaret, K. S. (
2000
). Liver specification and early morphogenesis.
Mech. Dev.
92
,
83
-88.