Clonal, cultured, murine fetal liver hepatoblasts maintain liver specification in chimeric mice


  • Charles E. Rogler,

    1. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Hong Chou Zhou,

    1. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Lauretta LeVoci,

    1. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Leslie E. Rogler

    Corresponding author
    1. Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    • The Marion Bessin Liver Center, Departments of Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461
    Search for more papers by this author
    • fax: 718-430-8975

  • Potential conflict of interest: Nothing to report.


Recent studies have shown a pluripotential nature of stem cells that were previously thought to be committed to specific lineages. HBC-3 cells are a clonal fetal murine hepatoblast cell line derived from an e9.5 murine embryo, and these cells can be induced to form hepatocytes and bile ducts in vitro and when transplanted into adult mouse livers. To determine whether HBC-3 cells can exhibit a pluripotential phenotype, we created chimeric mice by injection of enhanced green fluorescent protein (EGFP)–marked HBC-3 cells into wild-type or dipeptidyl dipeptidase IV (DPPIV) knockout blastocysts. Genetically labeled HBC-3 cells were identified by EGFP polymerase chain reaction (PCR) in all the major organs of many chimeric mice and visualized in chimeras as bright red DPPIV-positive cells in the DPPIV knockout chimeric mice. Strikingly, the HBC-3 cells maintained phenotypic and biochemical features of liver specification in every case in which they were identified in nonliver organs, such as brain, mesenchyme, and bone. In adult liver they were present as small foci of hepatocytes and bile ducts in the chimeras. Additional major histocompatibility complex (MHC) marker analysis and X and Y chromosome content analysis further demonstrated that HBC-3 cells did not acquire the phenotype of the organs in which they resided and that they were not present because of fusion with host cells. Conclusion: In contrast to other stem cell types, these data demonstrate that cultured murine fetal liver stem cells appear to maintain their liver specification in the context of nonliver organs in chimeric mice. (HEPATOLOGY 2007.)

Stem cell populations have been classified based on their capacity to self-renew and give rise to multiple different cell types.1 The different types of stem cells represent a continuum of developmental potential from the totipotent cells of the early preimplantation embryo to the highly restricted stem cells of tumors.1 Classic transplantation experiments suggested that the generation of different cell types from totipotent cells of the early vertebrate embryo occurs by the sequential restriction of potential. Programming of cells for specific functions occurs through cell–cell interaction, morphogen gradients, chromatin modification, and transcription factor activation.1–3

Classic transplantation experiments by LeDourain2 showed that liver specification followed interaction of the foregut endoderm with the cardiac mesenchyme. Recent molecular analysis has shown that specification of the foregut endoderm in mammals requires the development of competency because of the expression of Fox1a and Fox2a within the ventral foregut endoderm.3 Further liver specification is accomplished by sequential interactions of the competent region of foregut endoderm through fibroblast growth factor signaling from the cardiac mesenchyme followed by bone morphogenetic protein signaling from the septum transversum.4, 5 These interactions lead to the specification of a committed liver progenitor population within the hepatic diverticulum from which the epithelial compartment of the liver is derived.

Several lines of evidence have led investigators to question the irreversibility of cellular commitment. The pathologic observation of metaplasia suggested that under certain conditions cells could lose their commitment to a lineage and “transdifferentiate” into a type of cell that would normally be found elsewhere in the animal.6. Early work by Briggs and King7 and Fishburg et al8 in amphibians and more recent work in mammals have demonstrated that the process of commitment is epigenetic.9, 10 The results of these experiments led to the question of whether commitment was truly an irreversible process as had been suggested by the results of tissue transplantation. This also raises the question of whether tissue-resident stem cell populations are subject to the process of commitment or are generic cells whose lineage is determined by the nature of the niche in which they reside.

One approach to determine the potential of stem cell populations is to use them to generate chimeric mice by blastocyst injection or aggregation with morula stage mouse embryos. The generation of chimeric mice has been useful for fate mapping the mammalian embryo.11 This approach was also used to demonstrate that the stem cells of teratomas (embryonal carcinoma cells) could be reprogrammed to participate in normal development by injection into the blastocyst.12–14 In an analogous experiment, neural stem cells or bone marrow–derived mesenchymal stem cells have been injected into blastocysts, and the resultant chimeras showed contribution from these stem cell populations in a wide variety of tissues.15–17 More recent experiments have suggested that adult neural stem cells were incapable of forming chimeras after blastocyst injection because in the absence of fusion the cells rapidly differentiated into glial-like cells within the blastocyst.18

We have isolated a fetal hepatoblast cell line (HBC-3) by microdissection of the hepatic diverticulum from murine embryos on day 9.5 post coitus.19 These cells can be induced to differentiate along either the hepatocyte or bile ductular lineages in vitro.19–21 On transplantation into murine liver, they become integrated into the liver structure and show limited replication, as would be expected of a committed liver progenitor cell line.20 To investigate the developmental potential of these cells, we have injected enhanced green fluorescent protein (EGFP)–labeled HBC-3 cells into wild-type and dipeptidyl dipeptidase IV (DPPIV) knockout murine blastocyts and followed their fate in the resultant chimeric embryos, pups, and adult animals. Labeled cells were found in a wide variety of tissues in chimeric animals, and the analysis shows that HBC-3 cells retain their hepatic specification even when present in nonliver tissues in the chimeric mice.


DPPIV, dipeptidyl dipeptidase IV; EGFP, enhanced green fluorescent protein; FISH, fluorescent in situ hybridization; p.c., post coitum; PCR, polymerase chain reaction.

Materials and Methods

DPPIV knockout mice were graciously provided by Dr. Shafritz (Albert Einstein College of Medicine, Bronx, NY) Blastocyst injections were conducted by injecting 8 to 10 HBC-3 cells (clone 8) into 3.5-day-old blastocysts of C57Bl/6 mice or DPPIV knockout mice followed by transfer to foster C57Bl/6 pseudopregnant females. Chimeric animals were isolated on days 13.5 to 18.5 p.c, 1 week, 2 weeks postnatal, and 8 and 10 months of age. Animal protocols were approved by the Animal Care Use Committee of the Albert Einstein College of Medicine.

Polymerase Chain Reaction Analysis of EGFP and Neo.

Organ genomic DNA was prepared by the Trizol method. Polymerase chain reaction (PCR) of EGFP and Neo genes used primers that amplified 417-bp and 492-bp fragments, respectively. Primer sequences and reaction conditions are in Supplementary Methods available on the HEPATOLOGY website (

Estimation of Chimerism.

Quantitative real-time PCR was used to measure the relative ratio of the Y chromosome marker, Sry, and X chromosome marker, hypoxanthine-guanine phosphoribosyltransferase, in DNA from control and chimeric tissues. Artificial mixes of male and female control DNA were used as references to estimate the percentage of chimerism of tissues from male chimeras. PCR primers, amplification conditions, and Syber green labeling protocol are in Supplementary Methods.

Detection of DPPIV Activity.

DPPIV enzyme expression was determined on 5-micron-thick cryostat sections as described in Piazza et al22 and in the Supplementary Methods.

Detection of H2Kk.

Five-micron frozen sections of chimeric mouse adult or fetal tissues were fixed with ice-cold acetone for 5 minutes. The sections were stained using biotinylated anti-mouse H2Kk monoclonal antibody, avidin-conjugated horseradish peroxidase (Vector), and counterstained with Gill's hematoxylin, Photomicrographs were made using a Nikon Labophot microscope. Sections of wild-type C57Bl/6 mouse tissues were similarly stained with H2Kk and H2Kb as controls.

XY Fluorescent In Situ Hybridization.

Fluorescent in situ hybridization (FISH) for X and Y chromosomes was performed using standard protocols by the cytogenetics laboratory at Albert Einstein College of Medicine. Livers were microdissected from chimeric or wild-type mouse fetuses on day 14.5 p.c., dissociated by trituration acid methanol, fixed, and deposited onto slides by cytospin. Slides were hybridized with single paint probes for mouse X and Y chromosomes from Applied Spectral Imaging according to the manufacturer's specifications. Y probes were labeled with Cy5 and X probes were labeled with fluorescein isothiocyanate. Stained slides were observed using a Leica AOBS Laser Scanning Confocal Microscope. Fifty cells from each preparation were examined.


To investigate the pluripotential nature of whether HBC-3 cells can be induced to differentiate along additional cellular lineages, we tested their ability to form tissues in a chimeric mouse environment. Our hypothesis was that, if HBC-3 cells are capable of exhibiting plasticity, they will receive the needed signals and will contribute to multiple tissues in the developing embryo.

We developed an HBC-3 EGFP-expressing clone to follow the fate of cells in early blastocysts and embryos. This cell line did not exhibit any phenotypic changes from the parental line.20 Cells were injected into the blastocoel cavity of embryos at 3.5 days p.c. and the blastocysts placed in outgrowth culture (Fig. 1). The transplanted cells survived injection and persisted during blastocyst outgrowth formation. In some of the outgrowths, the cells remained as a cluster of green fluorescent cells associated with the developing epiblast, whereas in others, they were scattered throughout the outgrowth in both epiblast and trophoblast (Fig. 1), suggesting that they might contribute to both embryonic and extraembryonic tissues in chimeras.

Figure 1.

EGFP-labeled HBC-3 detected in cultured blastocyst outgrowths. EGFP-positive HBC-3 cells were detected in (A) inner cell mass and trophoblast cells of hatching blastocysts, 6 hours postinjection (original magnification ×150). (B) The inner cell mass 1 day postinjection (original magnification ×300). (C) The trophoblast 1 day postinjection (original magnification ×300). (D) The inner cell mass of blastocyst outgrowth at 4 days postinjection (original magnification ×300).

HBC-3 Genetic Markers Were Widespread in Chimeric Mice.

Similarly injected blastocysts were implanted into pseudopregnant females, and we isolated e14.5-day embryos. DNA was prepared from isolated fetal liver, yolk sac, placenta, and the remaining embryo. PCR for EGFP was positive for all the extraembryonic-derived placenta and yolk sac samples from 10 embryos (Fig. 2, Table 1). EGFP was detected in only 4 of 10 of the embryos and in only 2 of 10 of the livers. The presence of the cells in embryonic and extraembryonic tissues revealed a widespread distribution of HBC-3 cells in the chimeric embryos.

Figure 2.

EGFP DNA was detected in the liver, fetus, placenta, and yolk sac of e14.5-day chimeric fetuses. Chimeric fetuses were isolated, and the liver, fetus, placenta, and yolk sac were dissected. DNA was prepared and assayed for EGFP by PCR. (A) Analysis of EGFP in (E)-fetus minus the liver; (P)- placenta; (Y)- yolk sac of 4 chimeric fetuses (#7–10). (B) Analysis of EGFP in liver DNA samples from 10 e14.5 chimeras (#s 1–10) (N), negative control; (P), positive control; (Nt), no template.

Table 1. Detection Of EGFP DNA In E14.5-Day Chimeric Fetus by PCR
E14.5 fetusPCR for GFP
  1. L, liver; Em, embryo; Pl, placenta; Ys, yolk sac.

  2. Chimeric conceptuses were isolated on day 14.5 of gestation and dissected into 4 fractions. Genomic DNA was prepared from the 4 fractions as described in Materials and Methods. The presence of EGFP DNA−positive HBC-3 cells was detected using PCR as described in Materials and Methods. +, EGFP DNA was detected. −, EGFP DNA was not detected.


Analysis of 1-week, 4-week, and 10-month-old chimeric mice showed 2 patterns of HBC-3 chimerism (Fig. 3 and Table 2). In 1 group (Set A, Table 2), there was a limited contribution of HBC-3 cells, because EGFP was present in only selective lobes of the liver and in a few other tissues, including small intestine, bone, and brain. In the second group, (Set B, Table 2), there was a widespread, fine-grained chimerism in which all the tissues tested, including all the lobes of the liver, were EGFP positive. The difference in the degree of chimerism did not appear to be attributable to sexual dimorphism because male and female chimeras occurred at approximately equal numbers of set A and set B chimeras. We also checked our DNA samples for the Neo gene that was also present in HBC-3 clone 8 cells. In all cases, the Neo gene was present in the identical distribution as the EGFP gene (Supplementary Fig. 1).

Figure 3.

EGFP DNA was detected in multiple tissues of 1-week-old and 4-week-old chimeras. Two distinct patterns of chimerism were detected. (A) Chimeras with sparse chimerism in the liver and other tissues. (B) Chimeras with widespread chimerism in which EGFP was detected in all tissues examined. The data shown here are representative results. Organ code: 1, brain; 2, lung; 3, heart; 4, kidney; 5, spleen; 6, small intestine; 7, colon; 8, skin; 9, bone; 10; gonad, 11, bladder; 12 muscle; 13, pancreas; 14, stomach; 15, fat; 16–20; liver lobes; P, Positive control; N, negative control; Nt, no template.

Table 2. Summary of Animals Analyzed
Genotype of AnimalsTotal Number AnalyzedNumber ChimericNumber in Set ANumber in Set B
DPPIV null6913NANA

These data showed a varying, but widespread, distribution of HBC-3 cells in embryonic and extraembryonic tissues of chimeras, as would be expected. It also showed that the HBC-3 cells, which become migratory when exposed to Matrigel in culture,20 did not specifically home to the liver in the chimeras, because some of the embryonic livers were negative.

Quantitative Analysis of Male and Female Cells in Male Chimeric Mice.

To quantitatively estimate the contribution of HBC-3 cells in the chimeras, we analyzed the X chromosome marker, hypoxanthine-guanine phosphoribosyltransferase, and the Y chromosome marker, Sry, by quantitative real time PCR in DNA preparations that were positive or negative for EGFP. The standard curve generated using artificial mixes of male and female control animals is shown in Supplementary Fig. 2. HBC-3 cells have a normal female chromosome complement, and this allowed us to estimate the contribution of HBC-3 cells within tissues of male chimeric mice, by calculating the relative increase in the X chromosome marker, normalized to the Y chromosome marker, in DNA from candidate chimeric organs (Table 3). This approach showed a significant degree of chimerism, generally between 5% and 10% (Table 3). The highest degree of chimerism was found in the liver and colon.

Table 3. Percent Chimerism of Isolated Tissues from Chimeric Fetuses, Pups, and Adult Chimeric Mice
Description of Tissue*Approximate % Chimera
  • *

    All tissues were positive for HBC-3 cells as determined by detection of EGFP DNA by PCR. The percent chimerism of tissues isolated from male chimeras was determined using quantitative PCR to measure the relative abundance of the genes for Syr (male) and HPRT (female) in the samples as described in Materials and Methods.

E14.5 #8 liver; ♂5%
pup #8 brain; 3–4w ♂11%
pup # 2 colon; 1w21%
pup #1 Liv 4; 1w ♂25%
pup #1 Liv 1; 1w ♂8%
pup #1 small intestine; 1w ♂10%
pup #2 lung; 1w ♂5%
pup #2 heart; 1w ♂5%
pup #2 spleen; 1w ♂15%
5lr Liv 2; 3–4w ♂3%
11-6 Liver lobe 2; 10 months6%
11-6 Liver lobe 3; 10 months<5%
11-6 Liver lobe; 10 month5%

HBC-3 Cells Maintain Liver Specification in Chimeric Mice.

We initially attempted to use fluorescence from the EGFP transgene, driven by an SV40 promoter, to identify HBC-3 cells in the chimeric mice. However, expression of this transgene was extinguished in the cells. Therefore, we prepared a second set of chimeric mice in which the blastocyst donors were DPPIV null animals. Because HBC-3 cells are DPPIV positive (wild-type), this provided a marker that is easily detected by histochemistry as a bright red staining pattern. In particular, in the adult liver of wild-type animals, hepatocytes display a characteristic pattern of canalicular staining (Fig. 4A).

Figure 4.

DPPIV-positive HBC-3 cells detected in DPPIV knockout e16.5 chimeric embryos. Arrows mark clusters of red DPPIV-positive cells. (A) DPPIV staining of wild-type liver (red) (original magnification ×400). (B) HBC-3 in brain (original magnification ×80). (C) Higher magnification of cluster in B showing diploid nucleus and canalicular DPPIV staining, arrow (original magnification ×360). (D) HBC-3 in mesenchyme (original magnification ×80). (E) HBC-3 in liver, hepatocyte cluster (original magnification ×40). (F) Higher magnification of cluster in E showing canalicular staining, arrow (original magnification ×200). (G) HBC-3 in bile duct clusters (original magnification ×80). (H) HBC-3 in rib (original magnification ×80). (I) Higher magnification of positive area in H (original magnification ×360).

As expected, we identified DPPIV-positive HBC-3 cells in the livers of chimeric mice. HBC-3 cells formed hepatocytes with clearly delineated bile canalicular staining of DPPIV (Fig. 4B–F and H–I). Their bipotential differentiation was confirmed by the observation of DPPIV-positive bile ducts in the chimeric livers (Fig. 4G). Interestingly, HBC-3–derived hepatocytes and bile ducts were not highly prevalent in the livers, suggesting that they did not have a selective growth advantage and might in fact have a growth disadvantage, because of their stem cell phenotype. We also observed that as the chimeric mice aged, the frequency of HBC-3 cells in the chimeric organs decreased.

We also identified HBC-3 cells in nonliver locations using DPPIV staining as an assay. In all cases, the DPPIV-positive staining occurred as a distinctly bile canalicular pattern. A dramatic example was found in the brain, where a group of HBC-3 cells were present and a complete bile canalicular network was observed (Fig. 4B,C). The region of the brain in which these cells were located did not have any DPPIV staining in wild-type mice. The only DPPIV-positive cells in wild-type brain are the epithelial cells of the choroid plexus, and they have a much different phenotype (Supplementary Fig. 3).

Therefore, these data strongly suggest that HBC-3 cells had formed a small node of liver in the brain.

We also observed HBC-3 cells with bile canalicular DPPIV staining patterns in mesenchyme and developing bone (Fig. 4D, H, I) where there was no staining in wild-type counterparts. Again the HBC-3 cells were present in small clusters. In an HBC-3 cluster in the mesenchyme, we were able to carry out double staining for DPPIV and albumin (Fig. 5A–C). This analysis demonstrated that the DPPIV-positive cells were also albumin positive, supporting our conclusion that the HBC-3 cells maintain their liver specification in nonliver locations, in the chimeric mice.

Figure 5.

Detection of DPPIV and albumin in the same cell cluster. (A) DPPIV-positive small HBC-3 cell cluster in the mesenchyme of an e16.5-day-old fetus (original magnification ×80). (B) Albumin staining of the same cluster using immunofluorescent staining with Alexa 488–labeled rabbit anti-mouse albumin antibody. (C) DAPI staining.

HBC-3 Cells Do Not Assume the Phenotype of the Organ in Which They Are Present.

Additional HBC-3 cells may have been present in nonliver tissues, and their DPPIV expression may have been silenced when they differentiated in that tissue. To test this, we developed an independent test for HBC-3 cells based on the unique H2 haplotype of HBC-3 cells. C57Bl/6 mice express H2Kb and CBA mice express H2Kk. All of the host blastocysts that were used for chimera production were from C57Bl/6 mice, and HBC-3 cells were derived from a hybrid of C57Bl/6/CBA embryo. We determined that HBC-3 cells expressed H2Kk. This made it possible to detect HBC-3 cells as H2Kk-expressing cells using an anti-mouse H2Kk antibody (Fig. 6A). These cells should be positive cells against the negative background of the H2Kb cells of the C57Bl/6 chimeric embryos.

Figure 6.

H2Kk-positive HBC-3 cells identified in small clusters in chimeric mice. Brownish/red staining marks H2Kk-positive cells. (A) H2Kk-positive cell cluster in the liver of 1–week-old chimeric mouse (original magnification ×40). Inset: High magnification (×200) of H2Kk-positive cells. (B) H2Kk-positive small cluster in liver of 2-week-old chimeric mouse (original magnification ×40). Inset: High magnification of isolated positive cells (original magnification ×200). (C) Cluster of H2Kk-positive cells from the brain of 4-week-old chimeric mouse (original magnification ×40). Arrow marks H2Kk-positive cell mass.

H2Kk staining revealed positive cells sparsely distributed within tissues, including liver, and the brain (Fig. 6A–C) in accordance with the DPPIV data. Similar to the DPPIV staining, no broad areas of positive reacting cells were observed in any of the chimeric tissues examined. Within the liver, the positive cells resembled hepatocytes (Fig. 6A, B). The positive staining from the brain was from an animal in which brain was the only green fluorescing protein–positive tissue. As shown in Fig. 6C, the H2Kk positive cells were found adjacent to the cerebellum, in a mass that appeared to consist of undifferentiated cells. Thus, it appears that in this animal the HBC-3 cells were sequestered as a mass of undifferentiated cells adjacent to brain tissue.

In Situ Hybridization to X and Y Chromosome Markers.

The histological data strongly suggested that HBC-3 cells were distributed throughout some embryos, raising the possibility that they could be the product of somatic cell fusion. To test this possibility, we isolated cells from the liver of male chimeric embryos that were determined to contain female HBC-3 cells by PCR for EGFP. Cells derived from fusion should contain 3 X chromosomes and 1 Y chromosome or possibly 2Y, by FISH analysis. However, detection of both X chromosome signals in control female liver cells was not possible with our detection system. We were only able to detect 1, or in many cases, no X chromosome markers in control female cells. In contrast, using control male liver cells, we were able to easily detect the Y chromosome marker in 93% of the cells with only 7% not yielding a convincing Y chromosome signal.

Therefore, we could determine whether a liver contained normal unfused female cells by measuring the percentage of Y chromosome–negative cells in a random sample of cells from the liver. FISH analysis from e14.5 male chimeric embryos showed that 27% of the cells did not have a Y chromosome signal compared with only 7.14% of the cells from control male liver. Multiple X chromosomes were not observed in any of these cells. In addition, we were unable to identify any cells that had 1 Y chromosome and 3 or more X chromosomes in any of the samples examined. Therefore, the FISH data demonstrated that fusion was not a major contributor to the transplanted HBC-3 cells in the chimeric embryos.


Concepts surrounding the question of plasticity of stem cells have undergone significant changes since the development of cell and nuclear transplantation and transfection technologies. An important factor in stem cell plasticity is the source of the stem cells, and the most comprehensive test for pluripotency is the ability to form multiple tissues in chimeric mice derived from blastocysts injected with the cells. Embryonal stem cells (ES cells) that are derived from blastocysts are the most pluripotent stem cells so far identified, because they reproducibly form all mammalian tissues in chimeric mice derived from blastocyst injection.

The second major source of stem cells is tissue-specific stem cells. General understanding in the past has been that tissue-specific stem cells can differentiate into multiple lineages but are not pluripotent and generally show less self-renewal capability. However, transplantation studies with some tissue-specific stem cells, primarily those derived from bone marrow or enriched hematopoietic stem cells, have suggested a broader pluripotentiality.16, 23–27 Transplantation experiments have suggested that hematopoietic stem cells are capable of forming vastly different tissues of mesenchymal, endothelium, and endoderm origin.16, 23–27 Specialized cell cultures of bone marrow mesenchymal stem cells have led to the identification of rare multipotent adult progenitor cells, in the cultures. Multipotent adult progenitor cells transplanted into blastocysts have shown the ability to contribute to most, if not all, somatic cells types in the resultant chimeric mice. These include cells of mesenchymal origin, endothelial origin, and endoderm origin, including liver, lung, and gut.17 Furthermore, one report showed that neural stem cells when injected into blastocysts contributed to many tissues of the chimeric mouse embryos.15

A third potential source of pluripotent stem cells has recently been reported using differentiated fibroblasts (induced pluripotent stem cells).28–30 Using a specialized cocktail of transfected genes, fully differentiated mouse embryo fibroblasts or tail-tip fibroblasts were converted, at a low frequency, into stem cells capable of forming all the tissues of the adult, including the germline.

Several experimental approaches have focused on the isolation of stem/progenitor cells from the liver and tests for their transplantation capability.31–33 Cells have been isolated from both adult and embryonic livers and directly transplanted into appropriately prepared recipient livers.33, 34 These approaches have demonstrated that fetal hepatocytes have extensive capacity to grow and integrate into host liver.33 The transplanted fetal stem/progenitor cells can form both hepatocytes and bile ducts in the host livers, suggesting that they are committed bipotential stem cells.33 However, until now, the broader developmental potential of such liver specified stem cells has not been determined.

In light of work described, determining whether cultured tissue-derived embryonic liver stem cells could exhibit a broader pluripotentiality when placed in the blastocyst environment was of interest to us. This would address the question of whether embryonic hepatoblasts were irreversibly committed to the hepatic lineage or could be “reprogrammed” by the environment of the developing embryo. We pursued this goal by injecting cultured cells into wild-type and DPPIV-null C57Bl/6 blastocysts.

Initial analysis of injected blastocysts showed that the genetically marked HBC-3 cells distributed in an apparently random manner and formed chimeras in which they were present in both embryonic and extraembryonic (yolk sac) tissues. Genetic analysis further demonstrated that HBC-3 cells distributed widely throughout the major organs of many chimeric mice. Phenotypic analysis of HBC-3–derived cells present in the chimeric mice showed that they maintained their hepatocytic specification in diverse cellular environments. This was most striking in the chimeric brain, where HBC-3 cells formed a DPPIV-positive bile canalicular network. A bile canalicular staining pattern also occurred in the mesenchyme, and HBC-3 cells in the cluster were positive for albumin, further establishing their liver identity.

In general, the number of HBC-3 cells that could be detected was the highest in chimeric embryos, and they became more difficult to detect as the animals increased in age. This could be because of their stem cell nature and their presence in organs in which they do not normally exist. Thus, when HBC-3 cells are restricted in growth, they would be expected to decrease according to their normal half-life. Another possibility is that there was a weak negative selection against HBC-3 cells in the chimeric mice, as has been observed in other transplantation systems in which the donor and recipient genomes are not identical.35

The results of this study are different from the reported work with other tissue-derived stem cells such as hematopoietic stem cells, bone marrow–derived mesenchymal cells, and neural stem cells discussed previously. Rather, they are consistent with the classical understanding of lineage-specified stem cell biology. That is, once a cell population becomes committed to a lineage, that commitment is maintained even when the cells are present in diverse environments in the developing embryo.2 Therefore, use of cultured bipotent embryonic hepatoblasts will most likely be limited to uses for transplantation into the liver.


The authors thank Dr. Teresa DiLorenzo for her assistance in developing the MHC assay to detect HBC-3 cells.