Volume 137, Issue 1, Pages 62-79 (July 2009)

33 of 80

ABSTRACT

FULL TEXT

FULL-TEXT PDF (1691 KB)

Mechanisms of Liver Development: Concepts for Understanding Liver Disorders and Design of Novel Therapies

Received 10 February 2009; accepted 18 March 2009. published online 31 March 2009.

The study of liver development has significantly contributed to developmental concepts about morphogenesis and differentiation of other organs. Knowledge of the mechanisms that regulate hepatic epithelial cell differentiation has been essential in creating efficient cell culture protocols for programmed differentiation of stem cells to hepatocytes as well as developing cell transplantation therapies. Such knowledge also provides a basis for the understanding of human congenital diseases. Importantly, much of our understanding of organ development has arisen from analyses of patients with liver deficiencies. We review how the liver develops in the embryo and discuss the concepts that operate during this process. We focus on the mechanisms that control the differentiation and organization of the hepatocytes and cholangiocytes and refer to other reviews for the development of nonepithelial tissue in the liver. Much progress in the characterization of liver development has been the result of genetic studies of human diseases; gaining a better understanding of these mechanisms could lead to new therapeutic approaches for patients with liver disorders.

John P. Lynch and David C. Metz, Section Editors

Abbreviations used in this paper: APC, adenomatous polyposis coli, ARF6, adenosine diphosphate–ribosylation factor 6, ATF, activating transcription factor, BMP, bone morphogenetic protein, C/EBP, CCAAT-enhancer binding protein, E, embryonic day, ER, endoplasmic reticulum, FGF, fibroblast growth factor, Fox, Forkhead box factor, HDGF, hepatoma-derived growth factor, Hes-1, homolog of hairy/enhancer of split 1, Hex, hematopoietically expressed homeobox factor, HGF, hepatocyte growth factor, HNF, hepatocyte nuclear factor, Id3, inhibitor of differentiation 3, LRH-1, liver receptor homolog 1, NF-κB, nuclear factor κB, OC, onecut factor, OSM, oncostatin M, Prox-1, Prospero-related homeobox 1, SOX9, SRY-related HMG box transcription factor 9, Tbx3, T-box transcription factor 3, TGF, transforming growth factor, TNF, tumor necrosis factor, WT1, Wilms' tumor suppressor gene, XBP-1, X-box binding protein 1, Zhx2, Zinc finger and homeoboxes factor 2

Article Outline

• Abstract

• Formation of the Hepatic Bud From the Endoderm: Coordination of Proliferation, Migration, Intercellular Adhesion, and Differentiation

• Expansion of the Liver: Integrated Signaling and Transcriptional Networks That Control Proliferation and Survival of Hepatoblasts

• Segregation of Hepatobiliary Lineages: Cell Fate Decision Controlled by Cell-Extrinsic and Cell-Intrinsic Cues

• Hepatocyte Maturation: Development of Function and Morphology

• Dynamic Transcriptional Network and Synergistic Interdependence

• Gene Repression During Hepatocyte Maturation

• Extracellular Signaling and Hepatocyte Maturation

• Hepatocyte Heterogeneity: The Concept of Metabolic Zonation

• Hepatocyte Maturation and Beyond: Programmed Differentiation of Stem Cells to Hepatocytes for Cell Therapy

• Morphogenesis of the Biliary Tract

• Extrahepatic Biliary Tract Development

• Transient Asymmetry During Intrahepatic Bile Duct Development

• Human Biliary Diseases and Biliary Developmental Mechanisms

• Conclusions

• References

• Copyright

The liver derives from the endoderm, one of the 3 germ layers formed during gastrulation. The endoderm delineates the primitive gut and gives rise to the epithelial compartment of the gastrointestinal tract and of a number of organs such as the thyroid, the liver, and the pancreas. A fate mapping study in which mouse endoderm cells were labeled with a dye to follow their location and migration revealed that the progenitor cells are located in 3 endodermal domains.1, 2 Two domains are paired and located laterally, whereas the third domain is found along the ventral midline (Figure 1A). The 2 lateral domains migrate toward the midline to fuse with the ventral midline domain and to generate a single prehepatic domain adjacent to the cardiogenic mesoderm. It is not clear whether the descendants of the 3 progenitor domains contribute differentially to adult liver cells.

View Large Image
Download to PowerPoint

Figure 1. Initiation of liver development in the mouse embryo. (A) The left panel shows a mouse embryo at E8.25 (10 somites). The dashed line indicates the foregut delineated by the endoderm. The gut is still open at that stage; the arrow points to the anterior ventral opening (anterior intestinal portal). The right panel schematizes a slightly younger embryo with a ventral view on the anterior intestinal portal. The blue areas correspond to the hepatic progenitor domains in the ventral endoderm. (B) Schematic representation (profile view) of the respective position of liver, heart, and septum transversum at 2 stages of liver development. When FGF production by the heart increases, the liver moves away from the heart and becomes adjacent to the septum transversum to ensure that the liver cells are exposed to the appropriate FGF concentrations.

The first molecular evidence of liver development is the expression of albumin, transthyretin, and α-fetoprotein in a region of the ventral endoderm located near the developing heart.3, 4 The concept of “specification” applies to this initial event, which occurs before organogenesis is detected: cells of the endoderm become committed to a specific differentiation fate. Hepatic specification occurs in mouse embryos at approximately the 7-somite stage, which corresponds to embryonic day (E) 8.25.

The principles that guide the molecular mechanisms of hepatic specification are fairly well understood. In vivo DNA/protein analyses have shown that the albumin gene promoter is bound by Forkhead box (Fox) A and GATA-4 transcription factors at a stage that precedes transcription of the gene.3, 5 Given the capacity of FoxA and GATA factors to decompact chromatin,6 it has been proposed that the binding of such factors opens chromatin to provide access to additional transcription factors, such as nuclear factor 1 and CCAAT-enhancer binding protein (C/EBP) β, eventually resulting in the transcriptional activation of albumin. The mode of action of FoxA factors illustrates the concept of developmental “competence,” which characterizes a cell that is not yet specified but has acquired the capacity to respond to specification-inducing signals. This function of FoxA factors was supported by the analysis of mouse embryos deficient in FoxA1 and FoxA2; these embryos failed to express α-fetoprotein, and their endoderm did not respond to extracellular inducers of specification.7 Hepatocyte nuclear factor (HNF)-1β (also called vHNF-1 or TCF2) stimulates expression of FoxA1 and FoxA2 in the prehepatic endoderm; in its absence, the endoderm cells lack competence and hepatic specification is impaired.8 Taken together, these data indicate that HNF-1β, FoxA1, FoxA2, and GATA-4 are critical initiators of liver development.

Competent endoderm cells respond to extracellular signals that induce their specification to a hepatic fate. Fibroblast growth factors (FGFs) secreted by the cardiogenic mesoderm were the first to be identified.2 Hepatic gene induction by FGF is mediated by a mitogen-activated protein kinase signaling cascade9 and is dependent on low concentrations of FGF. Liver gene induction in mouse endoderm starts when FGF production by the cardiogenic mesoderm is initiated. At this point, the developing heart is adjacent to the prehepatic endoderm and produces low amounts of FGF. Slightly later, the heart produces more FGF but becomes separated from the prehepatic endoderm by the septum transversum, a mesodermal tissue located caudally to the heart. These morphogenetic events, which move the endoderm away from the source of FGF, ensure that hepatic cells are not exposed to excessive concentrations of FGF10 (Figure 1B). This is necessary because higher concentrations of FGF induce differentiation of endoderm cells toward a more anterior fate, such as lung. Importantly, FGF signaling from the cardiogenic mesoderm is not sufficient for hepatic specification. In mouse embryos, the septum transversum produces bone morphogenetic protein (BMP)-2 and BMP-4, which cooperate with FGF to induce hepatic gene expression. The activity of BMPs is mediated at least in part by stimulation of the expression of the GATA-4 transcription factor, which, as indicated previously, contributes to competence and specification of the prehepatic endoderm. The study of BMP signaling, like that of FGF signaling, provides another illustration of the need for coordinated movements of endoderm and mesoderm. In zebrafish, knockdown of myosine phosphatase targeting subunit 1 caused abnormal bundling of actin filaments and disorganization of mesoderm and endoderm cells. This led to misalignment of BMP-producing mesoderm with respect to the liver primordium and eventually to apoptosis of the hepatoblasts.11

The mechanisms that control the initiation of liver development are well conserved among various organisms. For instance, studies in chick and zebrafish models have provided genetic evidence for the roles of FGF and BMP signaling in hepatic induction.12, 13, 14 Wnt signaling was also shown to have a role in liver induction. When the primitive gut becomes patterned along its anteroposterior axis, Wnt signaling must be repressed anteriorly to maintain foregut identity and to allow subsequent liver development. This repression depends on the appropriate secretion by the endoderm of Wnt inhibitors.15, 16 In particular, it was shown in Xenopus that the Wnt antagonist secreted frizzled-related protein 5 inhibits the activity of Wnt11, which would otherwise repress foregut identity and foregut epithelium morphogenesis.17 Importantly, beyond the stage of patterning, Wnt signaling becomes necessary for liver development; the Zebrafish mutant Prometheus is deficient in Wnt2bb and exhibits a profound but transient defect in liver specification.18

Formation of the Hepatic Bud From the Endoderm: Coordination of Proliferation, Migration, Intercellular Adhesion, and Differentiation

When endoderm cells have been specified, a diverticulum forms from the primitive gut. This occurs at E9 in the mouse and at day 22 in humans. The liver diverticulum is lined by endoderm cells, which at this stage are called hepatoblasts, that become columnar and undergo a transition to a pseudostratified epithelium19 (Figure 2). This transition results from interkinetic nuclear migration, a process characterized by the migration of the nucleus to the apical part of the cell during mitosis. The hepatoblasts proliferate and form a tissue bud delineated by a basement membrane that contains laminin, collagen IV, nidogen, fibronectin, and heparan sulfate proteoglycan.20 The hepatoblasts then leave the endoderm epithelium, migrate through the basement membrane, and invade the septum transversum. This is reminiscent of an epithelial-mesenchymal transition, because the hepatoblasts transiently lose their epithelial morphology and reduce expression of E-cadherin when they move away from the endoderm.

View Large Image
Download to PowerPoint

Figure 2. Budding of the liver out of the endoderm. The upper panel schematizes the changes in morphology of the liver when it buds out of the endoderm. The lower panel summarizes the control of key regulatory events by transcription factors and extracellular signaling.

A network of transcription factors controls the onset of liver development. Studies of knockout mice have shown that the hematopoietically expressed homeobox factor (Hex, also known as Prh) promotes interkinetic nuclear migration and hepatoblast proliferation19, 21, 22; GATA-6 is required to maintain the differentiation state of the hepatoblasts.23 Consequently, in the absence of either factor, liver development is arrested soon after budding. In prospero-related homeobox 1 (Prox-1)–deficient mice, hepatoblasts remain abnormally clustered in the liver bud.24 It has been proposed that the lack of Prox-1 leads to a failure of the hepatoblasts to delaminate from the liver bud and to migrate in the septum transversum, resulting from overexpression of E-cadherin, which holds the cells together by cell-cell contacts. In this context, it is interesting to consider the role of the T-box transcription factor 3 (Tbx3), which, in addition to mediating proliferation and cell-fate decisions (see the following text), is also required for hepatoblast migration, most likely via stimulation of Prox-1 expression.25 A phenotype similar to that of Prox-1–deficient mice was observed in embryos that have combined inactivation of the Hnf6 and Oc2 genes.26 Taken together, the analysis of these mutant mice suggests that Hex, GATA-6, HNF-6, Onecut (OC)-2, Tbx3, and Prox-1 control a network of genes involved in cell migration and adhesion (Figure 2).

The early phase of liver expansion depends on continuous interactions between hepatoblasts and adjacent mesodermal tissues; this has been observed in studies of all tissue compartments of the liver bud. When the activity of metalloproteases produced by the hepatoblasts and mesenchymal cells is blocked, migration of the hepatoblasts through the basement membrane is inhibited.27 Also, the endothelial cells that surround the liver bud promote hepatoblast proliferation, as illustrated by the lack of liver expansion in mice with deficiencies in endothelial cell development.28 Finally, the septum transversum, a tissue whose integrity requires the transcription factor GATA-4 and is a source of BMPs, is necessary for early liver expansion.29, 30

Expansion of the Liver: Integrated Signaling and Transcriptional Networks That Control Proliferation and Survival of Hepatoblasts

Once the hepatoblasts have invaded the septum transversum, they continue proliferating and the liver further expands. Several growth factors are involved in this process31 (Figure 3). Hepatocyte growth factor (HGF) is expressed by the septum transversum, the endothelial cells, and the hepatoblasts. Its receptor, c-met, is found at the surface of hepatoblasts, where it initiates a signaling cascade mediated by SEK1/MKK4 and possibly by c-jun. This cascade promotes hepatoblast proliferation, as indicated by the reduced proliferation of cells in mice with disruptions in genes that encode HGF, c-Met, SEK1/MMK4, or c-jun.32, 33, 34, 35, 36 It is not clear whether c-jun actually lies downstream of SEK1/MMK4, because liver defects appear 2 days later in c-jun–deficient mice than in SEK1/MMK4-deficient mice. Nevertheless, other known targets of SEK1/MMK4, namely activating transcription factor (ATF)-2 and ATF-7, were found to be required for liver development at the same stage as SEK1/MMK4. These factors are activated by JNK and p38, kinases that are downstream of SEK1/MMK4. ATF-2 was shown to have negative feedback control on the signaling cascade by stimulating the expression of phosphatases that repress and modulate the activity of p38.37

View Large Image
Download to PowerPoint

Figure 3. Regulation of liver growth during development. The signaling factors and downstream cascades that control proliferation and apoptosis of hepatoblasts are illustrated. Liver growth depends on the coordinated control of these biological processes.

The transforming growth factor (TGF) β Smad2/Smad3 pathway also stimulates proliferation, because mice with heterozygous mutations in the Smad2 and Smad3 genes showed liver hypoplasia.38 Interestingly, the HGF and TGF-β pathways functionally interact. Indeed, cultured explants from livers deficient in Smad2 and Smad3 had reduced expression of β1-integrin, but this reduction was rescued by the addition of HGF to the medium. Because cells deficient in β1-integrin cannot colonize the liver,39 it is likely that the HGF and TGF-β pathways converge on β1-integrin expression, which is necessary for proliferation.

Hepatoma-derived growth factor (HDGF) is produced by fetal hepatoblasts and stimulates their proliferation when they are cultured in vitro. The expression of HDGF remains high at the hepatoblast stage and is extinguished between E15.5 and E18.5, when the hepatoblasts mature to hepatocytes.40 However, HDGF is dispensable in vivo for normal liver development, suggesting the existence of unidentified compensatory pathways.41

β-catenin is best known as a mediator of the canonical Wnt signaling pathway. Wnt binding to its receptor, Frizzled, inactivates the β-catenin degradation complex, which contains adenomatous polyposis coli (APC). This allows β-catenin to dissociate from the complex, to translocate to the nucleus, and to bind to Tcf/Lef transcription factors to control transcription of Wnt target genes. Wnt signaling is particularly complex in the liver, because a total of 11 Wnt ligands and 8 Frizzled receptors were found in this organ in mice42 and at least 4 Frizzled receptors in chicken.43 Despite this complexity, analyses of β-catenin–deficient mouse livers indicate a role for β-catenin in stimulation of hepatoblast proliferation; these livers have hypoplasia and differentiation defects.44 The chicken liver provides an excellent model system for investigation of Wnt signaling. Indeed, the liver mesothelium contributes to sinusoidal and stellate cells, cell types that produce Wnt9a. Because the liver mesothelium in chicken is accessible to retroviral-mediated gene transfer, an elegant set of gain- and loss-of-function experiments was designed to address the role of Wnt9a. Overexpression of Wnt9a or small interfering RNA–mediated repression of Wnt9a showed that Wnt9 stimulates hepatoblast proliferation.43

Moreover, accurate analysis of chicken liver development indicated the existence of localized growth zones in the liver. These zones are located at the periphery of the developing liver lobes, where Wnt3a and β-catenin are abundant. By exerting a localized control on cell proliferation, Wnt signaling not only promotes growth of the liver, but it also controls the acquisition of global liver morphology.45 In addition, β-catenin seems to represent a key node at the intersection of multiple signaling cascades. It binds to the HGF receptor c-met in hepatocytes and is translocated to the nucleus upon HGF stimulation.46 Although this HGF/β-catenin cascade was not found to function in hepatoblasts, all components of this cascade are present in hepatoblasts. Moreover, FGF-10, which is secreted by myofibroblastic cells, stimulates proliferation of hepatoblasts and controls β-catenin activation,47 suggesting that β-catenin contributes to the growth stimulatory effects of FGF-10. The latter is produced predominantly from E12.5 to E13.5 and might replace or complement the role of other FGFs such as FGF-8, which stimulates liver development at earlier stages.4

Diffusible signals that stimulate proliferation of hepatoblasts do not necessarily act directly on hepatoblasts; this process is believed to be regulated by retinoic acid.48 The Wilms' tumor suppressor gene (WT1) encodes a transcription factor that is expressed in coelomic cells covering the liver as well as in endothelial and subendothelial cells that line the sinusoids. Mice with genetic disruption of WT1 showed abnormal liver lobulation and reduced proliferation essentially at the tips of the lobes. Because the WT1-deficient coelomic cells in the mutant livers had decreased expression of retinaldehyde dehydrogenase 2, an enzyme that catalyzes the synthesis of retinoic acid, the data indicated that mesenchymal cells positively control proliferation mediated by retinoic acid. This conclusion has been supported by experiments in chick embryos that were exposed to a retinaldehyde dehydrogenase 2 inhibitor and in quail liver explants that were incubated in the presence of retinoic acid; inhibition of retinoic acid synthesis was associated with reduced hepatoblast proliferation, whereas retinoic acid stimulated proliferation in explants. The retinoic acid receptor RXRα is expressed in mesodermal cells scattered between the hepatoblasts and often in contact with sinusoids. This suggests that retinoic acid stimulates proliferation of hepatoblasts by inducing the production of trophic factors by mesodermal cells rather than a direct effect on the hepatoblasts.

A different perspective is taken when considering the role of tumor necrosis factor (TNF). TNF stimulates a signaling cascade that activates the transcription factor nuclear factor κB (NF-κB). NF-κB is composed of the subunits p50 and p65/RelA and is activated by IκB kinases. Mice deficient in IκB kinase β or IκB kinase γ, or in the NF-κB subunit RelA, showed massive apoptosis in the liver. Interestingly, the dramatic phenotypes of IκB kinase β- and RelA-null mice were rescued by the additional inactivation of the TNF receptor, indicating that NF-κB protects hepatoblasts against TNF-induced apoptosis.49 Similar observations were made in studies of c-Raf-1; in its absence, hepatoblasts were apoptotic and displayed increased sensitivity to FasL, and this is another example of a mechanism that protects cells from apoptosis.50

In addition to those mentioned previously, several other transcription factors regulate hepatoblast proliferation. These include Prox-1,51 which promotes proliferation according to a mechanism that is antagonized by liver receptor homolog 1 (LRH-1; also called NR5A2 or fetoprotein transcription factor), and FoxM1B, which activates expression of regulators of the G2/M phase of the cycle.52 The X-box binding protein 1 (XBP-1), a transcription factor of the CREB/ATF family, is activated by endoplasmic reticulum (ER) stress and controls the expansion of the ER surface.53 It was proposed that in xbp1−/− mice, the lack of XBP-1 leads to insufficient ER production, resulting in activation of the ER stress-induced apoptotic pathway.31 Inhibitor of differentiation 3 (Id3), a basic-helix loop helix transcription factor, is transiently expressed during the earliest stages of chicken liver development, and small interfering RNA–mediated depletion of Id3 inhibited hepatoblast proliferation.54 The stimulatory activity of Id3 appears to depend on its abilty to inhibit the E2A-related protein E47. In addition, there is some evidence that Id3 acts downstream of FGF and/or BMP signals. Finally, the transcription factors N-Myc and Hlx are, unlike the previously discussed transcription factors, expressed in mesenchymal cells. Both factors promote liver growth, most likely by regulating the expression of paracrine factors.55, 56

Segregation of Hepatobiliary Lineages: Cell Fate Decision Controlled by Cell-Extrinsic and Cell-Intrinsic Cues

In the liver, the concept of the cell fate decision applies, in theory, to hepatoblasts that have reached a developmental stage at which they become committed either to the hepatocyte lineage or the cholangiocyte lineage (Figure 4). In contrast to concepts discussed previously, there are several reasons that the notion of cell fate determination is difficult to explain, in clear molecular terms, in the context of liver development. First, the time point at which the segregation of hepatobiliary lineages occurs has not been accurately determined; second, the determination of the cholangiocytic fate does not take place at a single developmental stage; third, the determination of the cholangiocytic lineage could be reversible; and fourth, the mechanisms are only partly understood.

View Large Image
Download to PowerPoint

Figure 4. Mechanisms of cell fate determination in the liver. Hepatoblasts give rise to either hepatocytes or cholangiocytes. Transcription factors and extracellular regulators that operate in the 3 cell types and during the transition from one cell type to the other are illustrated. The orange dashed oval integrates biliary differentiation mechanisms, and the green dashed oval delineates hepatocyte-inducing mechanisms. In cholangiocytes, HNF-6 controls a gradient of TGF-β signaling that is active at higher levels in the periportal cholangiocytes than in the rest of the parenchyma. The question mark refers to in vitro data not confirmed by in vivo experiments (Notch signaling).

Hepatocytes have a number of specific markers that are already present at the endodermal stage, so the onset of hepatocyte differentiation cannot be accurately defined. Between E9.5 and E12.5 in mice, the hepatoblasts start to express genes that are active in mature hepatocytes. Electron micrographs show that livers at E12 contain a subpopulation of hepatoblasts that have an abundant rough ER as well as lipid vesicles, indicating that they have entered the hepatocyte differentiation process. The analysis of mice with a disruption in the gene encoding the transcription factor HNF-4α showed deficient expression at E12 of a number of hepatocyte-specific proteins such as apolipoproteins or albumin, suggesting that HNF-4α is critical for hepatocyte fate determination.57

Another way to determine the timing of lineage segregation is by analysis of expression of biliary markers. For example, the expression of the intermediate filament proteins cytokeratin 19 and 7 is considered a sign of biliary differentiation. However, because cytokeratin 19 is expressed at low levels in hepatoblasts and at increasing levels in differentiating cholangiocytes,58 cytokeratin 19 expression is not a good indicator of cell fate commitment. Moreover, cytokeratin 7 becomes detectable when cholangiocytes are already committed to a cholangiocyte lineage. Recent data on the expression of the transcription factor SRY-related HMG box transcription factor 9 (SOX9) could resolve this issue59; this factor is present in endodermal cells that line the hepatic diverticulum, but its expression disappears when the cells start to invade the septum transversum. SOX9 becomes reexpressed at E11.5 in liver epithelial cells that are located a short distance from the branches of the portal vein, where the biliary cells differentiate; at later developmental stages, it is restricted to biliary cells. The expression of SOX9 at E11.5 can therefore be considered the earliest indication of hepatoblast differentiation toward the biliary lineage. However, a functional analysis of the transcription factor Tbx3 revealed that the control of hepatoblast cell fate determination extends over a period ranging at least from the liver primordium stage (E9.5) to E13.5.25, 60 Tbx3 expression in liver peaks from E9.5 to E13.5; in Tbx3−/− embryos, biliary differentiation is promoted at the expense of hepatocyte differentiation. In these embryos, at E9.5 the expression of HNF-6 and HNF-1β increases whereas that of HNF-4α decreases; this also occurs in differentiating biliary cells. Because Tbx3−/− embryos showed premature increase of HNF-6 and HNF-1β, it is possible that Tbx3 controls the timing of hepatoblast lineage decision. Interestingly, the functions of Tbx3 quickly evolve at the early stages of liver development. At E12.5, Tbx3 controls hepatoblast fate by repressing p19Arf /Cdkn2a. Elegant experiments showed that hepatic epithelial cells isolated from E12.5 Tbx3−/− embryos displayed growth arrest as well as increased levels of p19Arf and the biliary markers cytokeratin 7 and 19. A near-normal phenotype was restored to these cells by inhibition of p19Arf expression using a short hairpin RNA strategy. Also, overexpression of p19Arf in wild-type cells induced an increase in cytokeratin 7 and 19, confirming the link between p19Arf and hepatoblast lineage segregation.60 Such a mode of action of Tbx3 was not apparent at earlier stages; at E9.5, Tbx3 prevented a premature increase in HNF-6 and HNF-1β without affecting p19Arf expression.25

The link between biliary differentiation and proliferation is not entirely clear. Whereas Tbx3−/− livers associated reduced hepatoblast proliferation with increased biliary differentiation, the lack of FoxM1B was characterized by reduced proliferation and an absence of biliary development.52

Transcription factors could influence hepatoblast cell fate determination by modulating extracellular signaling. TGF-β promotes differentiation of hepatoblasts to biliary cells and represses hepatocyte differentiation. TGF-β signaling is highest near the portal vein, most likely as a result of the high expression of TGF-β2 and TGF-β3 in the periportal mesenchyme.59, 61 The transcription factors HNF-6 and OC-2 are expressed in hepatoblasts, cholangiocytes, and hepatocytes, but the highest levels are found in cholangiocytes. Livers deficient in HNF-6 and OC-2 had increased TGF-β signaling, which resulted in the generation of hepatic cells that coexpressed hepatocyte and cholangiocyte markers. Therefore, HNF-6 and OC-2 control hepatoblast differentiation by modulating a gradient of TGF-β signaling activity.

The livers deficient in HNF-6 also showed premature expression of biliary differentiation markers,62 indicating that HNF-6 controls the timing of hepatoblast fate decision and that its regulators participate in this control. As discussed previously, loss of Tbx3 increased the levels of HNF-6 above that of the normal levels found in hepatoblasts; this increase was associated with premature biliary differentiation,25 suggesting that high expression levels of HNF6 promote biliary differentiation of hepatoblasts. Moreover, blocking C/EBPα activity in hepatic stem cells isolated from E13.5 liver repressed HNF-6 expression but allowed induction of biliary markers, suggesting in this case that low expression levels of HNF-6 in hepatoblasts promote biliary differentiation.63 Also, in the absence of Hex, HNF-6 expression was reduced and hybrid “hepatobiliary cells” developed as they do in the livers of HNF-6 knockout mice,64 suggesting that in hepatoblasts, levels of HNF-6 that are too low induce a hybrid phenotype. Taken together, the seemingly conflicting data on the effects of Tbx3, Hex, and C/EBPα on HNF-6 and on biliary differentiation can be reconciled if one considers that timely and normal hepatoblast differentiation only occurs when HNF-6 expression is maintained at appropriate levels in hepatoblasts.

Another issue that relates to the timing of hepatoblast differentiation pertains to the mechanisms of bile duct formation. Based on their SOX9 expression profile, biliary cells first differentiate at approximately E11.5 and then line the branches of the portal vein to form a ductal plate (see the following text). Then, starting around E15.5, ducts start to form by apposition of hepatoblasts to the cholangiocytes of the ductal plate. These hepatoblasts delineate the lumen of the future ducts and later differentiate to cholangiocytes, so the cell fate decision must be made at approximately E15.5, a time point much later than that of the appearance of the first signs of biliary differentiation. This indicates that at least a subset of hepatoblasts must remain bipotent to allow for this second wave of biliary differentiation. Interestingly, cells of the ductal plate, which express cholangiocyte markers, can be purified and induced to differentiate into hepatocytes in vitro.65 Also, cholangiocyte marker–positive fetal liver cells can give rise to hepatocytes in a regenerating liver when transplanted into animals that underwent partial hepatectomy.66 Lineage tracing analyses are required to determine the reversibility of cholangiocyte differentiation under physiologic conditions in vivo.

Beyond the questions of timing and reversibility, the mechanisms of cell fate decision must be addressed (Figure 4). The involvement of key transcription factors and the TGF-β signaling pathway has been described previously, but other signaling mechanisms have been investigated. HGF stimulates expression of C/EBPα in isolated hepatoblasts, promoting differentiation toward the hepatocyte lineage.63 The Jagged-Notch pathway controls biliary development, but arguments in favor of its involvement at the stage of cell fate decision are based on in vitro experiments.67 In chicken liver, bFGF (FGF-2) and FGF-7 induce differentiation of hepatoblasts toward the biliary lineage68; this occurs in cooperation with BMP-4 and extracellular matrix components. Further evidence for a role of BMP (and/or TGF-β) signaling includes the high expression levels of Smad5 in early differentiating cholangiocytes that form the ductal plate.69 Finally, the Wnt/β-catenin pathway was investigated using explant cultures and in vivo mouse models. Exposure of early liver explants to Wnt3a or inhibition of expression of the Wnt signaling mediator β-catenin by an antisense approach led to the conclusion that Wnt signaling stimulates differentiation toward the biliary lineage.70, 71 This was supported by in vivo experiments in which Cre-mediated ablation of β-catenin in early liver strongly reduced biliary differentiation.44 In addition, a gain-of-function approach, in which β-catenin was stabilized by knockout of APC, also indicated that Wnt signaling represses hepatocyte differentiation and stimulates biliary development.72 Still, the mechanisms by which Wnt signaling controls hepatoblast fate decision are not clear. It is likely that the various Wnt ligands have specific differentiation-inducing properties, and studies that have targeted β-catenin have not been able to specifically address the effects of these ligands on cell fate decisions. Moreover, Wnt also signals via β-catenin–independent pathways that have not been investigated in liver development.

Finally, many extracellular matrix components have a role in cell fate determination. Laminin and collagen types I and IV facilitated the hepatocytic differentiation of a subpopulation of hepatic stem cells that were isolated by flow cytometry.63 In contrast, chicken liver epithelial cells grown in a collagen type I gel alone or in the presence of fibronectin, laminin, or collagen type IV did not induce expression of hepatocyte-specific genes but did induce biliary-specific genes.68 These contradictory data could result from differences in culture conditions but also reveal the difficulty in making conclusions about the role of the extracellular matrix. In vivo, stellate cells are a source of extracellular matrix and their activation, via ablation of the stellate cell–specific gene Lhx2, induced excessive extracellular matrix deposition.73 Extracellular matrix deposition was associated with the development of hybrid hepatobiliary cells, so alterations in this process might impair normal cell fate decisions.

Hepatocyte Maturation: Development of Function and Morphology

Dynamic Transcriptional Network and Synergistic Interdependence

Following the lineage segregation stage, cells that have differentiated to the hepatocyte lineage undergo a process of maturation that consists of the progressive acquisition of morphology and physiologic functions. Several microarray analyses of gene expression at various stages of liver development support the view that maturation is a process that extends throughout development, even after birth.74, 75, 76 The transcription factors that control hepatocyte maturation nicely illustrate the concept of a “dynamic transcriptional network.” A set of transcription factors, commonly referred to as liver-enriched factors, organize to form a network of autoregulatory and cross-regulatory loops. In elegant studies, the binding of each individual liver-enriched factor to the gene regulatory regions of all other liver-enriched factors was tested by in vivo analyses of protein-DNA interactions.77 This revealed that the number and complexity of interactions increases when hepatocyte maturation proceeds (Figure 5A). This increase correlates with the progressive rise in concentrations of most liver-enriched factors and with the increased stability of the network. In a study of transcription factor binding to target promoters, the removal of a single factor had much less of an impact when it was deleted after birth than when it was deleted in the prenatal period.77 In this study, 12 liver-enriched factors were analyzed and a “core” of 6 factors (HNF-1α, HNF-1β, FoxA2, HNF-4α1, HNF-6, and LRH-1) was found to occupy the gene regulatory regions of each other and the other factors. These core factors not only exerted a mutual control, but they also cooperated with cofactors to stimulate transcription of common targets, illustrating the concept of “synergistic interdependence.” This was shown for the regulation of glucose-6-phosphatase (g6pc) expression, in which the stimulatory activity of HNF-6 required the presence of HNF-4α and vice versa. This interdependence went pairwise with synergistic activity, which relied, at least in part, on the corecruitment by HNF-6 and HNF-4α of the coactivator PGC1α to the g6pc promoter.78 Interestingly, the synergistic interdependence only occurs when the concentrations of HNF-6, HNF-4α, and PGC-1α reach threshold levels. Similar observations were made with other combinations of liver-enriched factors; increased levels of C/EBPα and HNF-6 proteins are required to stimulate association of these factors with the CBP coactivator protein to bind to the FoxA2 promoter.79 In integrating the concepts of a dynamic transcriptional network and synergistic interdependence, we propose a model in which hepatocyte maturation depends on a network of factors that increase in concentration during maturation, thereby allowing an increase in the number of autoregulatory and cross-regulatory loops that stabilize the network. The increasing concentration of factors reaches threshold levels at specific developmental stages, thereby permitting corecruitment of coactivators and time-specific activation of target genes (Figure 5B).

View Large Image
Download to PowerPoint

Figure 5. Dynamic organization of a hepatic transcriptional network. (A) Changes in interactions between the core hepatocyte-enriched transcription factors at 2 stages of development and in the adult. Green arrows, unidirectional regulations; blue double arrows, reciprocal interactions. Adapted with permission from Kyrmizi et al.77 (B) Model for time-specific gene activation during development. The concentration of 3 transcription factors (TF1, TF2, and TF3) and of 2 coactivators (CoAct1 and CoAct2) rises during development. When threshold concentration levels are reached, these proteins can interact interdependently and synergistically to induce specific genes at specific time points during development.

The liver-enriched transcription factors have been studied using gene knockout strategies in mice; the functions of most of these have been compiled in excellent reviews.80, 81, 82 Transcription factors have developmental stage–specific effects that can be studied in detail by phenotype analysis of mice that have liver-specific and inducible (ie, temporal-specific) gene inactivations. The metabolic roles of liver-enriched factors in developing and adult hepatocytes depend on the identity of their target genes, which can now be studied at the genomic level using chromatin immunoprecipitation combined with DNA microarray or sequence analysis of the immunoprecipitated gene fragments.83, 84 HNF-1α and HNF-4α control glucose metabolism as well as several other hepatic functions such as lipid and amino acid metabolism. HNF-1β is essential for bile acid sensing and fatty acid oxidation.85 The 3 FoxA factors (FoxA-1, -2, and -3) have overlapping DNA-binding properties and, like the other liver-enriched factors, regulate numerous hepatic functions.86 LRH-1 belongs to the nuclear hormone receptor family and controls bile acid and cholesterol metabolism.87, 88 HNF-6 mediates some effects of growth hormone,89 inhibits glucocorticoid activity,90 and stimulates expression of genes in the gluconeogenic, glycolytic, and bile acid synthesis pathways78, 91, 92 as well as hepatocyte proliferation.93 Finally, members of the C/EBP family are not among the factors of the core transcriptional network but were among the first to attract the attention of hepatologists.94 C/EBPα, the founding member of the C/EBP family, regulates glucose and glycogen metabolism as well as lipid homeostasis and hepatocyte proliferation. Another member of the family, C/EBPβ, is a regulator of gluconeogenesis and a potent stimulator of phosphoenolpyruvate carboxykinase.

Beyond their role in the control of metabolic functions, liver-enriched factors are also important determinants of hepatocyte morphology. Livers that are deficient in HNF-4α not only exhibit metabolic defects but also have abnormal architecture. Hepatocytes from these livers were small and failed to epithelialize, and this impaired normal sinusoidal organization. The role of HNF-4α in determination of cell morphology depends at least in part on cell-autonomous mechanisms. Overexpression of HNF-4α in cultured fibroblasts converted the cells to an epithelial phenotype,95 whereas the HNF-4α–null phenotype was similar to that of fetal mice with livers deficient in the expression of cited2, a cofactor of HNF-4α.96 Moreover, HNF-4α is required for normal expression of an impressive set of genes whose products are involved in cell junction assembly and adhesion97 as well as in regulating the endoplasmic stress response.98 However, the effects induced by the absence of HNF-4α might also rely on noncell autonomous mechanisms. Indeed, when HNF-4α–deficient cells were grown in vitro as primary cultures, cell morphology and polarity were restored.99 Because the absence of HNF-4α was associated with signs of stress response, such response may contribute to the deficiencies observed in HNF-4α–null livers.

Gene Repression During Hepatocyte Maturation

Hepatocyte maturation also requires repression of a number of genes during the prenatal and postnatal periods. In addition to the proliferation factor HDGF mentioned earlier,40 the down-regulation of the gene that encodes α-fetoprotein serves as a classic example of liver-specific, transcriptional repression. Cis-acting sequences and trans-acting repressors have been identified. The zinc finger and homeoboxes factor 2 (Zhx2) is one of these repressors. It is more strongly expressed in adult liver than in prenatal liver. Mice with a retroposon inserted in the Zhx2 locus have persistent postnatal expression of α-fetoprotein, and compensatory expression of a transgene coding for Zhx2 in these mice restored α-fetoprotein repression.100 The Zinc finger factor ZBTB20 is another α-fetoprotein repressor. It is expressed in the liver only after birth and it binds to the α-fetoprotein gene promoter, where it exerts transcriptional inhibition.101 Interestingly, ZBTB20, like Zhx2, is not a liver-specific protein, indicating that non–tissue-specific proteins could be responsible for liver-specific functions.

Components of the basal transcription machinery or chromatin remodeling factors, which are all expressed ubiquitously, can exert hepatocyte maturation-specific effects. Liver-specific inactivation of TAF-10 (a component of the TFIID complex) in the embryo led to widespread loss of hepatocyte-specific gene expression and deficient hepatocyte differentiation and morphology, in agreement with the transcription-activation function of TFIID.102 However, inactivation of TAF-10 in the liver after birth was associated with normal liver architecture and differentiation but with persistent expression of genes, such as α-fetoprotein, that are normally repressed postnatally. The link between differentiation and chromatin remodeling is well illustrated by the phenotype of mice that are deficient in the expression of the subunit SNF5 of the switching defective/sucrose nonfermenting (SWI/SNF) complex. The livers of these mice are impaired in morphogenesis as a result of defective expression of proteins required for cell-cell junction formation and proliferation; these mice also have profound anomalies in glucose metabolism resulting from down-regulation of genes required for gluconeogenesis and glycogen synthesis.103

Extracellular Signaling and Hepatocyte Maturation

Starting around E12 in the mouse embryo, the differentiating hepatocytes are closely associated with hematopoietic precursor cells. The latter colonize the embryonic liver, and by the end of gestation and the early postnatal period they leave the organ and migrate to the bone marrow. These hematopoietic cells are essential for hepatocyte maturation, because they secrete oncostatin M (OSM), an interleukin-6–related cytokine. OSM binds to the gp130 receptor at the hepatocyte membrane and induces a signaling cascade mediated by STAT3. This results in stimulated expression of terminal hepatocyte differentiation markers such as the gluconeogenic enzymes glucose-6-phosphatase and phosphoenolpyruvate carboxykinase.104, 105 Jumonji, a transcription factor expressed in several cell types including hepatocytes, is also necessary for the activity of OSM.106 Other cytokines are expected to have roles that are similar or compensatory to that of OSM, because OSM receptor-null mice do not display obvious hepatocyte anomalies.107 Moreover, the inducing effects of OSM seem to be tempered by TNF-α signaling in the prenatal period. The production of TNF-α decreases after birth, and this relieves a repressive effect that TNF-α exerts on OSM-induced maturation.108

Hepatocyte maturation is also manifested by their organization as cord-like structures. The small guanosine triphosphatase adenosine diphosphate–ribosylation factor 6 (ARF6) has a crucial role in this context because ARF6-deficient livers failed to develop cords and instead showed clumped hepatic cells.109 In addition, ARF6 is activated in response to HGF, and Arf6−/− hepatocytes were unable to organize as cords upon stimulation by HGF when grown in collagen gels, thereby uncovering one more role of HGF in liver development.

Wnt signaling controls hepatoblast proliferation but is also involved in the maturation of hepatocytes. In chicken livers, Wnt9a is produced by sinusoidal cells and stimulates glycogen accumulation in hepatocytes; gene expression analysis suggested that this resulted from negative regulation of glucose-6-phosphatase and positive regulation of glycogen synthase.43

Other types of intercellular signaling mediate sinusoid-hepatocyte interactions. This was shown by recent work on the function of sinusoids in the zebrafish.110 In the zebrafish mutant valentine, a cytoplasmic scaffolding protein is disrupted, resulting in hepatocytes with defective polarity. The protein heart of glass is a transmembrane protein that genetically interacts with valentine, and heart of glass mutants also displayed hepatocyte polarity defects. Both valentine and heart of glass are expressed in endothelial sinusoidal cells; the transmembrane structure of heart of glass suggested that it is responsible for the observed effects on cell polarity. This conclusion was supported by the observation that overexpression of heart of glass perturbs hepatocyte polarity. In particular, such overexpression induced down-regulation of an apical polarity marker and up-regulation of a basolateral hepatocyte marker, leading to the conclusion that the endothelial sinusoidal cells can influence the apico-basal polarity of hepatocytes.

Hepatocyte maturation is not complete at birth; the neonatal period of liver development is an often-overlooked phase of development, despite its importance, which is illustrated by the ontogenic gene expression patterns of cytochrome P450 genes.111 Production of HGF is greater in the postnatal than in the prenatal period and HGF further stimulates maturation of cultured fetal hepatocytes, suggesting that it has a similar role in vivo.112

Hepatocyte Heterogeneity: The Concept of Metabolic Zonation

The hepatocytes constitute a heterogeneous cell population that expresses different gene sets and exerts different metabolic functions, depending on their location in the hepatic lobule. This concept, known as metabolic zonation, begins in the first weeks after birth.113, 114 Within a hepatic lobule, it distinguishes a periportal zone from a pericentral zone on the basis of the expression of a number of genes, mainly metabolism-regulating genes.115 HNF-4α and other regulators (described in the following text) contribute to the establishment of lobular zonation. Wild-type liver expresses glutamine synthase, ornithine aminotransferase, and thyroid hormone receptor β1 exclusively in pericentral hepatocytes, whereas HNF-4α knockout livers have the same pericentral expression pattern but these factors are also expressed in the periportal zone. It is likely that the periportal repression of glutamine synthase depends on the recruitment by HNF-4α of a repressor, namely histone deacetylase type I. Indeed, histone deacetylase type I recruitment to the glutamine synthase regulatory regions depends on the presence of HNF-4α.116 Importantly, HNF-4α cannot be considered a master regulator of metabolic zonation, because the zonation-keeping property of HNF-4α is restricted to only a subset of its gene targets.

Several experiments have indicated the importance of Wnt signaling in regulation of metabolic zonation.117 In the liver, β-catenin and its negative regulator APC show complementary expression; the former is found in the perivenous area and the latter in the periportal hepatocytes. Moreover, activation of the β-catenin pathway by APC inactivation or expression of a dominant active β-catenin mutant imposed a perivenous gene program on the entire liver lobule. The reverse approach, inactivation of the Wnt pathway by overexpression of the Wnt antagonist dickkopf-1, imposed a periportal pattern to the perivenous hepatocytes. Because APC is the key regulator of β-catenin levels along the lobular axis, APC is proposed to be the master regulator of zonation, but the identity and the source of the Wnt ligand(s) that regulate metabolic zonation are unknown.

Hepatocyte Maturation and Beyond: Programmed Differentiation of Stem Cells to Hepatocytes for Cell Therapy

Deciphering the developmental pathways that lead to mature hepatocytes has implications for regenerative therapy of liver disease. A survey of such therapy is beyond the scope of the present review, and we refer to a recent editorial in Gastroenterology.118 Studies on embryonic development established the concept that stem cells could give rise in vitro to hepatocytes by recapitulation of in vivo developmental pathways. This concept is illustrated in protocols for embryonic stem cell differentiation,119 which initially included growth factors such as FGF-2, HGF, and OSM to mimic hepatic specification and maturation. Such protocols were recently refined by the addition of Activin A and Wnt3a to recapitulate early stages of liver development (endoderm formation)120, 121 and BMP-4 to mimic hepatic specification.120 Because clinically relevant hepatocytes have not yet been obtained in vitro, it is expected that more fine-tuning of differentiation is necessary and that developmental biologists will further complement the list of markers and growth factors (Table 1) and characterize their function.

Table 1.

Transcriptional Regulators and Secreted Factors That Control Liver Development


Transcriptional regulator	Function	Secreted factor	Function
ATF-2	Negative feedback on HGF signaling; represses hepatoblast proliferation	BMP-2	Hepatic specification of prehepatic endoderm; early liver expansion
β-catenin	Mediator of Wnt signaling; stimulates hepatoblast proliferation and biliary differentiation, zonation keeper	BMP-4	Hepatic specification of prehepatic endoderm; early liver expansion; cooperates with FGF to stimulate biliary differentiation
C/EBPα	Represses biliary differentiation, controls hepatocyte metabolism	FGF-1	Promotes hepatic specification of prehepatic endoderm
c-Jun	Mediator of HGF signaling; stimulates hepatoblast proliferation	FGF-2	Promotes hepatic specification of prehepatic endoderm; stimulates biliary differentiation
FoxAfactors	Promote competence of prehepatic endoderm	FGF-7	Stimulates biliary differentiation
FoxA2	‘Core' hepatocyte transcription factor; stimulates hepatocyte differentiation	FGF-8	Promotes early liver morphogenesis
FoxM1B	Stimulates hepatoblast proliferation	FGF-10	Stimulates hepatoblast proliferation
GATA-4	Promotes competence of prehepatic endoderm; stimulates early liver expansion	HDGF	Stimulates hepatoblast proliferation
GATA-6	Stimulates hepatoblast differentiation	HGF	Stimulates hepatoblast proliferation, hepatocyte maturation and cord formation
Hex	Stimulates hepatoblast proliferation; controls duct morphogenesis	OSM	Promotes hepatocyte maturation
Hes-1	Mediator of Jagged1/Notch2 signaling; stimulates duct morphogenesis	TGF-β	Stimulates hepatoblast proliferation
Hlx	Stimulates liver growth	TGF-β2,-β3	Stimulate biliary differentiation
HNF-1α	‘Core' hepatocyte transcription factor; stimulates hepatocyte differentiation	TNF	Induces liver apoptosis; represses OSM activity
HNF-1β	Promotes competence of prehepatic endoderm; ‘core' hepatocyte transcription factor, controls duct morphogenesis	Wnt	Represses foregut identity, promotes liver specification; zonation keeper
HNF-4α	‘Core' hepatocyte transcription factor; stimulates hepatocyte differentiation and morphogenesis; zonation keeper	Wnt 3a	Stimulates hepatoblast proliferation and biliary differentiation
HNF-6	Stimulates hepatoblast migration; controls hepatocyte and biliary differentiation; ‘core' hepatocyte transcription factor; controls hepatocyte differentiation and duct morphogenesis	Wnt 9a	Stimulates hepatoblast proliferation; controls hepatocyte metabolism
Id3	Stimulates hepatoblast proliferation; potential mediator of FGF and BMP
Jumonji	Regulator of OSM signaling
Lhx2	Maintains stellate cells quiescent
LRH-1	‘Core' hepatocyte transcription factor; represses hepatoblast proliferation
NF-κB	Protects against TNF-induced apoptosis
N-Myc	Stimulates liver growth
OC-2	Stimulates hepatoblast migration; promotes biliary differentiation; controls duct morphogenesis
Prox-1	Stimulates hepatoblast migration
Sox9	Controls timing of duct morphogenesis
Tbx3	Stimulates hepatoblast migration; promotes hepatocyte differentiation
WT1	Promotes liver lobulation; stimulates hepatoblast proliferation
XBP-1	Promotes ER production; protects against ER stress
Zhx2	Repressor of α-fetoprotein

Morphogenesis of the Biliary Tract

Extrahepatic Biliary Tract Development

Cholangiocytes line the extrahepatic and intrahepatic bile ducts but have a dual origin. The biliary tract consists of an extrahepatic portion that comprises the hepatic ducts, the cystic duct, the gallbladder, and the common bile duct and an intrahepatic portion that consists of the intrahepatic bile duct network. The extrahepatic biliary tract develops from an outpocket of the endoderm, that is located caudally to the liver, and is closely associated with the ventral pancreatic bud. Anatomists used to distinguish a cranial and caudal portion in the liver diverticulum; the caudal portion is the part of the diverticulum that participates in extrahepatic biliary tract development.122 The 2 portions are difficult to distinguish anatomically, but a gene expression map of the ventral endoderm identified a region that coexpresses the pancreatic and duodenal homeobox factor 1, Prox-1, and HNF-6 and that could be the gallbladder anlage, because this combination of factors is found in extrahepatic biliary epithelial cells.123 The location of the extrahepatic biliary tract in the vicinity of the ventral pancreas indicates the relationship between the 2 organ systems that is reflected in the phenotype of some knockout mice. In mice deficient in the transcription factor homolog of hairy/enhancer of split-1 (Hes-1), the cholangiocytes of the extrahepatic biliary tract differentiate toward a pancreatic phenotype,124, 125 and mice deficient in pancreatic and duodenal homeobox factor 1 showed absence of peribiliary glands and of mucin-producing cells in the common bile duct in addition to deficient pancreatic morphogenesis.126 The morphogenesis of the extrahepatic biliary tract depends on the appropriate development of both the cholangiocyte epithelium and the surrounding mesenchyme. Loss of expression of epithelial transcription factors such as HNF-6 or HNF-1β or haploinsufficiency of the mesenchymal transcription factor FoxF1 was associated with dysmorphogenesis of the gallbladder and common bile duct.62, 85, 127

Morphologic analysis of the developing liver suggests that the limit between the extrahepatic biliary tract and intrahepatic bile duct resides at the level of the hepatic ducts.128, 129 This is supported by the observation that disruption of the Hex gene led to replacement of the extrahepatic biliary tract with duodenal-like tissue, with persistence of the extremities of the hepatic ducts.64 The extrahepatic biliary tract develops before the intrahepatic bile duct, but the mechanism by which they anastomose is not known.

Transient Asymmetry During Intrahepatic Bile Duct Development

Extrahepatic biliary tract cholangiocytes derive from the endoderm and intrahepatic bile duct cholangiocytes from hepatoblasts. Intrahepatic bile duct morphogenesis starts with the alignment of the cholangiocytes around the periportal mesenchyme to form a single-layer ring of cells called the ductal plate. The formation of ducts relies on a unique tubulogenic process in which hepatoblasts become apposed to specific areas of the ductal plate to generate ducts that have a transiently asymmetrical structure.59 These ductal structures are considered to be asymmetrical because their lumen is delineated on the portal side by ductal plate cholangiocytes and on the parenchymal side by hepatoblasts (Figure 6). The latter differentiate to cholangiocytes between E15.5 and E18.5, thereby allowing the development of ducts entirely lined by cholangiocytes. During this period, the expression of cell adhesion molecules such as E-cadherin59 and spermatogenic immunoglobulin superfamily (nectin-like molecule-2)130 increases and that of neural cell adhesion molecule decreases.131 During duct formation, the ductal plate areas not involved in tubulogenesis regress, whereas the ducts grow in length following a hilum-periphery axis and become surrounded by periportal mesenchyme (Figure 6).

View Large Image
Download to PowerPoint

Figure 6. Morphogenesis of the intrahepatic bile ducts. During bile duct development, the cholangiocytes first form a ring of cells (ductal plate) around the branches of the portal vein. Hepatoblasts then come into contact with the ductal plate and delineate the lumen of transiently asymmetrical ducts. When the hepatoblasts on the parenchymal side of the asymmetrical ducts differentiate to cholangiocytes, the ducts become symmetrical (totally delineated by cholangiocytes) and surrounded by portal mesenchyme. The ducts grow from the hilum toward the periphery of the liver lobes, which is reflected by the observation that sections at different levels along the hilum-periphery axis reveal different levels of duct maturation. Adapted with permission from Antoniou et al.59

Several transcription factors involved in biliary determination of hepatoblasts seem to be involved in intrahepatic bile duct morphogenesis, but it is not clear whether the aberrant morphogenesis observed in mice with disruptions in the genes that encode these transcription factors results from a role of the factors in duct formation or is a secondary consequence of abnormal lineage segregation, or both. Despite this uncertainty, there has been speculation about the organization of the biliary transcriptional network. A few factors were shown to impact tubulogenesis without being required for cell fate determination. SOX9 is downstream of HNF-6 and upstream of C/EBPα in this network, and SOX9-deficient livers are delayed in duct development.59 C/EBPα is normally absent from biliary cells; this absence is induced by SOX9 and might be required for normal expression of HNF-6 and HNF-1β.59, 132 In contrast, the biliary expression of HNF-6, which stimulates that of HNF-1β, is necessary, because in the absence of HNF-6 or HNF-1β the ducts give rise to ductal plate malformations and biliary cysts.62, 85

Human Biliary Diseases and Biliary Developmental Mechanisms

Hes-1–deficient mice have normal duct plates but fail to initiate tubulogenesis.133 Hes-1 is an effector of Notch signaling; clues about the function of this pathway came from the analysis of patients affected with Alagille syndrome. This syndrome is associated with mutations in the genes that encode Jagged1 and its receptor Notch2. It has been a challenge to characterize the expression pattern of Notch signaling ligands and receptors.134 A plausible model, based on expression data and on functional analysis of zebrafish and mouse models,135, 136, 137, 138 is that Jagged-1 is expressed in the periportal mesenchyme, including portal fibroblasts and stellate cells,139 and interacts with Notch-2 in cholangiocytes to stimulate expression of Hes-1 to induce ductal morphogenesis.140

Further information on the mechanisms of bile duct morphogenesis came from the study of human diseases characterized by polycystic bile ducts that were or were not associated with cysts in other organs, such as the kidney. A detailed discussion of these diseases goes beyond the scope of the present review (see Everson et al,141 Johnson et al,142 and Kamath and Piccoli143 for reviews). Several genes associated with polycystic liver in humans encode proteins that are associated with primary cilia functions.144 The primary cilia are nonmotile organelles located at the apical pole of the cholangiocytes that project into the duct lumen. They are considered to be osmosensors, mechanosensors, and chemosensors.145 Intracellular levels of calcium and adenosine 3′,5′-cyclic monophosphate are modulated by the bending of cholangiocyte cilia or by exposure of the cilia to hypotonicity or to biliary nucleotides, which has been proposed to control bile secretion, cholangiocyte-matrix interactions, and cell proliferation.146, 147, 148 Normal fluid transport might control bile duct lumen development; this occurs during development of the gut lumen in zebrafish.149 The importance of cell-matrix interactions for lumen development was illustrated by in vitro experiments in which liver progenitor cells developed into luminal structures lined by polarized biliary cells when cultured in laminin 1–containing gels.150 In this context, the involvement of laminin is supported by the observation that cholangiocytes express a specific set of laminin receptors, namely integrins α6β1, α2β1, α3β1, and α6β4.151 Proliferation of cholangiocytes is an expected mechanism for regulation of duct diameter. Angiogenic factors stimulate cholangiocyte proliferation, and their expression is up-regulated in cystic diseases.152 Along the same lines, the analysis of a rat model of Caroli's disease identified the epidermal growth factor/mitogen-activated protein kinase pathway as a positive regulator of cholangiocyte proliferation; repressing this pathway with the EGF receptor inhibitor gefitinib inhibited cystic dilations.153, 154 A final remark on cholangiocyte proliferation pertains to its regulation by microRNAs. The latter are 20–22 nucleotide RNAs that modulate translation and stability of messenger RNAs. In cholangiocytes, microRNA-15a down-regulates the expression of the cell cycle regulator cell division stimulator cdc25a; patients with various polycystic diseases (linked to cilia-related genes) have decreased levels of microRNA-15a and a subsequent increase in cdc25a.155 These studies have established an interesting link between cilia function and microRNA expression.

Biliary dysgenesis in humans is not restricted to cystic diseases. Bile duct paucity is observed in Alagille syndrome as well as in other genetic diseases. The arthrogryposis–renal dysfunction–cholestasis syndrome is associated with mutations in VPS33B, a regulator of vesicle trafficking.156 The involvement of vesicle trafficking in biliary morphogenesis was confirmed by the analysis of mutant zebrafish that had a reduced number of ducts, which led to the identification of VPS18 as a trafficking regulator157 and of HNF-6 as a regulator of VPS33B expression.158

Biliary dysgenesis has been associated with abnormal biliary differentiation. SOX9-deficient livers have delayed duct development that results from delayed differentiation of the hepatoblasts that line the asymmetrical ductal structures, and there is evidence that these structures have an abnormal response to TGF-β.59 Moreover, in some patients affected with Meckel syndrome, a cilia-related deficiency, the hepatic epithelial cells displayed hybrid characteristics as they coexpressed cholangiocyte and hepatocyte markers.159 Finally, as for hepatocytes, biliary differentiation continues after birth; cholangiocytes continue to acquire a significant degree of morphologic and functional heterogeneity, depending on the position of the cells in the ramified ductal network.160, 161 It is not known how this process is controlled.

Conclusions

Most efforts in the study of liver development have focused on hepatocytes and biliary cells. A wide range of signaling factors and differentiation markers have been identified. They can be used to translate the developmental processes in cell culture protocols for the in vitro programmed differentiation of cells for regenerative therapies of liver diseases. Alternatively, developmental biologists have learned much from the analysis of human diseases, illustrated by our understanding of bile duct morphogenesis via genetic analyses of patients with biliary dysgenesis. Moreover, the role of nonepithelial cells in epithelial development emerged through investigations of the intercellular mechanisms that promote differentiation and morphogenesis of hepatocyte cords and bile ducts. Little is known about the origin, differentiation, and morphogenesis of the nonepithelial cells,162, 163, 164 but increased understanding of these mechanisms could challenge current concepts about epithelial cell origins. The observation that stellate cells can give rise to epithelial cells during liver regeneration is a good example.165 There are many important questions remaining about all liver cell types; answering these will require constant dialogue between basic scientists and clinicians, along with, whenever possible, the much-appreciated help of patients.

References

1. 1Tremblay KD, Zaret KS. Distinct populations of endoderm cells converge to generate the embryonic liver bud and ventral foregut tissues. Dev Biol. 2005;280:87–99. MEDLINE | CrossRef

2. 2Zaret KS. Genetic programming of liver and pancreas progenitors: lessons for stem-cell differentiation. Nat Rev Genet. 2008;9:329–340. CrossRef

3. 3Gualdi R, Bossard P, Zheng M, et al. Hepatic specification of the gut endoderm in vitro: cell signaling and transcriptional control. Genes Dev. 1996;10:1670–1682. MEDLINE | CrossRef

4. 4Jung J, Zheng M, Goldfarb M, et al. Initiation of mammalian liver development from endoderm by fibroblast growth factors. Science. 1999;284:1998–2003. MEDLINE | CrossRef

5. 5Bossard P, Zaret KS. GATA transcription factors as potentiators of gut endoderm differentiation. Development. 1998;125:4909–4917. MEDLINE

6. 6Cirillo LA, Lin FR, Cuesta I, et al. Opening of compacted chromatin by early developmental transcription factors HNF3 (FoxA) and GATA-4. Mol Cell. 2002;9:279–289. MEDLINE | CrossRef

7. 7Lee CS, Friedman JR, Fulmer JT, et al. The initiation of liver development is dependent on Foxa transcription factors. Nature. 2005;435:944–947. CrossRef

8. 8Lokmane L, Haumaitre C, Garcia-Villalba P, et al. Crucial role of vHNF1 in vertebrate hepatic specification. Development. 2008;135:2777–2786. CrossRef

9. 9Calmont A, Wandzioch E, Tremblay KD, et al. An FGF response pathway that mediates hepatic gene induction in embryonic endoderm cells. Dev Cell. 2006;11:339–348. MEDLINE | CrossRef

10. 10Serls AE, Doherty S, Parvatiyar P, et al. Different thresholds of fibroblast growth factors pattern the ventral foregut into liver and lung. Development. 2005;132:35–47. MEDLINE | CrossRef

11. 11Huang H, Ruan H, Aw MY, et al. Mypt1-mediated spatial positioning of Bmp2-producing cells is essential for liver organogenesis. Development. 2008;135:3209–3218. CrossRef

12. 12Shin D, Shin CH, Tucker J, et al. Bmp and Fgf signaling are essential for liver specification in zebrafish. Development. 2007;134:2041–2050. MEDLINE | CrossRef

13. 13Chung WS, Shin CH, Stainier DY. Bmp2 signaling regulates the hepatic versus pancreatic fate decision. Dev Cell. 2008;15:738–748. CrossRef

14. 14Zhang W, Yatskievych TA, Baker RK, et al. Regulation of Hex gene expression and initial stages of avian hepatogenesis by Bmp and Fgf signaling. Dev Biol. 2004;268:312–326. MEDLINE | CrossRef

15. 15McLin VA, Rankin SA, Zorn AM. Repression of Wnt/beta-catenin signaling in the anterior endoderm is essential for liver and pancreas development. Development. 2007;134:2207–2217. MEDLINE | CrossRef

16. 16Finley KR, Tennessen J, Shawlot W. The mouse secreted frizzled-related protein 5 gene is expressed in the anterior visceral endoderm and foregut endoderm during early post-implantation development. Gene Expr Patterns. 2003;3:681–684. CrossRef

17. 17Li Y, Rankin SA, Sinner D, et al. Sfrp5 coordinates foregut specification and morphogenesis by antagonizing both canonical and noncanonical Wnt11 signaling. Genes Dev. 2008;22:3050–3063. CrossRef

18. 18Ober EA, Verkade H, Field HA, et al. Mesodermal Wnt2b signalling positively regulates liver specification. Nature. 2006;442:688–691. CrossRef

19. 19Bort R, Signore M, Tremblay K, et al. Hex homeobox gene controls the transition of the endoderm to a pseudostratified, cell emergent epithelium for liver bud development. Dev Biol. 2006;290:44–56. MEDLINE | CrossRef

20. 20Shiojiri N, Sugiyama Y. Immunolocalization of extracellular matrix components and integrins during mouse liver development. Hepatology. 2004;40:346–355. MEDLINE | CrossRef

21. 21Martinez-Barbera JP, Clements M, Thomas P, et al. The homeobox gene Hex is required in definitive endodermal tissues for normal forebrain, liver and thyroid formation. Development. 2000;127:2433–2445. MEDLINE

22. 22Keng VW, Yagi H, Ikawa M, et al. Homeobox gene Hex is essential for onset of mouse. Biochem Biophys Res Commun. 2000;276:1155–1161. CrossRef

23. 23Zhao R, Watt AJ, Li J, et al. GATA6 is essential for embryonic development of the liver but dispensable for early heart formation. Mol Cell Biol. 2005;25:2622–2631. MEDLINE | CrossRef

24. 24Sosa-Pineda B, Wigle JT, Oliver G. Hepatocyte migration during liver development requires Prox1. Nat Genet. 2000;25:254–255. MEDLINE | CrossRef

25. 25Lüdtke TH, Christoffels VM, Petry M, et al. Tbx3 promotes liver bud expansion during mouse development by suppression of cholangiocyte differentiation. Hepatology. 2009;49:969–978. CrossRef

26. 26Margagliotti S, Clotman F, Pierreux CE, et al. The Onecut transcription factors HNF-6/OC-1 and OC-2 regulate early liver expansion by controlling hepatoblast migration. Dev Biol. 2007;311:579–589. CrossRef

27. 27Margagliotti S, Clotman F, Pierreux CE, et al. Role of metalloproteinases at the onset of liver development. Dev Growth Differ. 2008;50:331–338.

28. 28Matsumoto K, Yoshitomi H, Rossant J, et al. Liver organogenesis promoted by endothelial cells prior to vascular function. Science. 2001;294:559–563. MEDLINE | CrossRef

29. 29Watt AJ, Zhao R, Li J, et al. Development of the mammalian liver and ventral pancreas is dependent on GATA4. BMC Dev Biol. 2007;7:37. CrossRef

30. 30Rossi JM, Dunn NR, Hogan BL, et al. Distinct mesodermal signals, including BMPs from the septum transversum mesenchyme, are required in combination for hepatogenesis from the endoderm. Genes Dev. 2001;15:1998–2009. MEDLINE | CrossRef

31. 31Tanimizu N, Miyajima A. Molecular mechanism of liver development and regeneration. Int Rev Cytol. 2007;259:1–48. MEDLINE | CrossRef

32. 32Hilberg F, Aguzzi A, Howells N, et al. c-jun is essential for normal mouse development and hepatogenesis. Nature. 1993;365:179–181. MEDLINE | CrossRef

33. 33Nishina H, Vaz C, Billia P, et al. Defective liver formation and liver cell apoptosis in mice lacking the stress signaling kinase SEK1/MKK4. Development. 1999;126:505–516. MEDLINE

34. 34Schmidt C, Bladt F, Goedecke S, et al. Scatter factor/hepatocyte growth factor is essential for liver development. Nature. 1995;373:699–702. MEDLINE | CrossRef

35. 35Uehara Y, Minowa O, Mori C, et al. Placental defect and embryonic lethality in mice lacking hepatocyte growth factor/scatter factor. Nature. 1995;373:702–705. MEDLINE | CrossRef

36. 36Bladt F, Riethmacher D, Isenmann S, et al. Essential role for the c-met receptor in the migration of myogenic precursor cells into the limb bud. Nature. 1995;376:768–771. MEDLINE | CrossRef

37. 37Breitwieser W, Lyons S, Flenniken AM, et al. Feedback regulation of p38 activity via ATF2 is essential for survival of embryonic liver cells. Genes Dev. 2007;21:2069–2082. CrossRef

38. 38Weinstein M, Monga SP, Liu Y, et al. Smad proteins and hepatocyte growth factor control parallel regulatory pathways that converge on beta1-integrin to promote normal liver development. Mol Cell Biol. 2001;21:5122–5131. MEDLINE | CrossRef

39. 39Fässler R, Meyer M. Consequences of lack of beta 1 integrin gene expression in mice. Genes Dev. 1995;9:1896–1908. MEDLINE | CrossRef

40. 40Enomoto H, Yoshida K, Kishima Y, et al. Hepatoma-derived growth factor is highly expressed in developing liver and promotes fetal hepatocyte proliferation. Hepatology. 2002;36:1519–1527. MEDLINE | CrossRef

41. 41Gallitzendoerfer R, Abouzied MM, Hartmann D, et al. Hepatoma-derived growth factor (HDGF) is dispensable for normal mouse development. Dev Dyn. 2008;237:1875–1885. CrossRef

42. 42Zeng G, Awan F, Otruba W, et al. Wnt'er in liver: expression of Wnt and frizzled genes in mouse. Hepatology. 2007;45:195–204. MEDLINE | CrossRef

43. 43Matsumoto K, Miki R, Nakayama M, et al. Wnt9a secreted from the walls of hepatic sinusoids is essential for morphogenesis, proliferation, and glycogen accumulation of chick hepatic epithelium. Dev Biol. 2008;319:234–247. CrossRef

44. 44Tan X, Yuan Y, Zeng G, et al. Beta-catenin deletion in hepatoblasts disrupts hepatic morphogenesis and survival during mouse development. Hepatology. 2008;47:1667–1679. CrossRef

45. 45Suksaweang S, Lin CM, Jiang TX, et al. Morphogenesis of chicken liver: identification of localized growth zones and the role of beta-catenin/Wnt in size regulation. Dev Biol. 2004;266:109–122. MEDLINE | CrossRef

46. 46Monga SP, Mars WM, Pediaditakis P, et al. Hepatocyte growth factor induces Wnt-independent nuclear translocation of beta-catenin after Met-beta-catenin dissociation in hepatocytes. Cancer Res. 2002;62:2064–2071. MEDLINE

47. 47Berg T, Rountree CB, Lee L, et al. Fibroblast growth factor 10 is critical for liver growth during embryogenesis and controls hepatoblast survival via beta-catenin activation. Hepatology. 2007;46:1187–1197. CrossRef

48. 48Ijpenberg A, Pérez-Pomares JM, Guadix JA, et al. Wt1 and retinoic acid signaling are essential for stellate cell development and liver morphogenesis. Dev Biol. 2007;312:157–170. CrossRef

49. 49Doi TS, Marino MW, Takahashi T, et al. Absence of tumor necrosis factor rescues RelA-deficient mice from embryonic lethality. Proc Natl Acad Sci U S A. 1999;96:2994–2999. MEDLINE | CrossRef

50. 50Piazzolla D, Meissl K, Kucerova L, et al. Raf-1 sets the threshold of Fas sensitivity by modulating Rok-alpha signaling. J Cell Biol. 2005;171:1013–1022. MEDLINE | CrossRef

51. 51Kamiya A, Kakinuma S, Onodera M, et al. Prospero-related homeobox 1 and liver receptor homolog 1 coordinately regulate long-term proliferation of murine fetal hepatoblasts. Hepatology. 2008;48:252–264. CrossRef

52. 52Krupczak-Hollis K, Wang X, Kalinichenko VV, et al. The mouse Forkhead Box m1 transcription factor is essential for hepatoblast mitosis and development of intrahepatic bile ducts and vessels during liver morphogenesis. Dev Biol. 2004;276:74–88. MEDLINE | CrossRef

53. 53Reimold AM, Etkin A, Clauss I, et al. An essential role in liver development for transcription factor XBP-1. Genes Dev. 2000;14:152–157. MEDLINE

54. 54Nakayama M, Matsumoto K, Tatsumi N, et al. Id3 is important for proliferation and differentiation of the hepatoblasts during the chick liver development. Mech Dev. 2006;123:580–590. MEDLINE | CrossRef

55. 55Hentsch B, Lyons I, Li R, et al. Hlx homeo box gene is essential for an inductive tissue interaction that drives expansion of embryonic liver and gut. Genes Dev. 1996;10:70–79. MEDLINE | CrossRef

56. 56Giroux S, Charron J. Defective development of the embryonic liver in N-myc-deficient mice. Dev Biol. 1998;195:16–28. MEDLINE | CrossRef

57. 57Li J, Ning G, Duncan SA. Mammalian hepatocyte differentiation requires the transcription factor HNF-4alpha. Genes Dev. 2000;14:464–474. MEDLINE

58. 58Van Eyken P, Sciot R, Callea F, et al. The development of the intrahepatic bile ducts in man: a keratin-immunohistochemical study. Hepatology. 1988;8:1586–1595. MEDLINE | CrossRef

59. 59Antoniou A, Raynaud P, Cordi S, et al. Intrahepatic bile ducts develop according to a new mode of tubulogenesis regulated by the transcription factor SOX9. Gastroenterology (in press).

60. 60Suzuki A, Sekiya S, Büscher D, et al. Tbx3 controls the fate of hepatic progenitor cells in liver development by suppressing p19ARF expression. Development. 2008;135:1589–1595. CrossRef

61. 61Clotman F, Jacquemin P, Plumb-Rudewiez N, et al. Control of liver cell fate decision by a gradient of TGFβ signaling modulated by Onecut transcription factors. Genes Dev. 2005;19:1849–1854. MEDLINE | CrossRef

62. 62Clotman F, Lannoy VJ, Reber M, et al. The onecut transcription factor Hnf6 is required for normal development of the biliary tract. Development. 2002;129:1819–1828. MEDLINE

63. 63Suzuki A, Iwama A, Miyashita H, et al. Role for growth factors and extracellular matrix in controlling differentiation of prospectively isolated hepatic stem cells. Development. 2003;130:2513–2524. MEDLINE | CrossRef

64. 64Hunter MP, Wilson CM, Jiang X, et al. The homeobox gene Hhex is essential for proper hepatoblast differentiation and bile duct morphogenesis. Dev Biol. 2007;308:355–367. CrossRef

65. 65Zhang L, Theise N, Chua M, et al. The stem cell niche of human livers: symmetry between development and regeneration. Hepatology. 2008;48:1598–1607. CrossRef

66. 66Simper-Ronan R, Brilliant K, Flanagan D, et al. Cholangiocyte marker-positive and -negative fetal liver cells differ significantly in their ability to regenerate the livers of adult rats exposed to retrorsine. Development. 2006;133:4269–4279. MEDLINE | CrossRef

67. 67Tanimizu N, Miyajima A. Notch signaling controls hepatoblast differentiation by altering the expression of liver-enriched transcription factors. J Cell Sci. 2004;117:3165–3174. MEDLINE | CrossRef

68. 68Yanai M, Tatsumi N, Hasunuma N, et al. FGF signaling segregates biliary cell-lineage from chick hepatoblasts cooperatively with BMP4 and ECM components in vitro. Dev Dyn. 2008;237:1268–1283. CrossRef

69. 69Ader T, Norel R, Levoci L, et al. Transcriptional profiling implicates TGFbeta/BMP and Notch signaling pathways in ductular differentiation of fetal murine hepatoblasts. Mech Dev. 2006;123:177–194. MEDLINE | CrossRef

70. 70Hussain SZ, Sneddon T, Tan X, et al. Wnt impacts growth and differentiation in ex vivo liver development. Exp Cell Res. 2004;292:157–169. MEDLINE | CrossRef

71. 71Monga SP, Monga HK, Tan X, et al. Beta-catenin antisense studies in embryonic liver cultures: role in proliferation, apoptosis, and lineage specification. Gastroenterology. 2003;124:202–216. Abstract | Full Text | Full-Text PDF (10834 KB) | CrossRef

72. 72Decaens T, Godard C, de Reyniès A, et al. Stabilization of beta-catenin affects mouse embryonic liver growth and hepatoblast fate. Hepatology. 2008;47:247–258. CrossRef

73. 73Wandzioch E, Kolterud A, Jacobsson M, et al. Lhx2-/- mice develop liver fibrosis. Proc Natl Acad Sci U S A. 2004;101:16549–16554. MEDLINE | CrossRef

74. 74Jochheim A, Cieslak A, Hillemann T, et al. Multi-stage analysis of differential gene expression in BALB/C mouse liver development by high-density microarrays. Differentiation. 2003;71:62–72. MEDLINE | CrossRef

75. 75Kelley-Loughnane N, Sabla GE, Ley-Ebert C, et al. Independent and overlapping transcriptional activation during liver development and regeneration in mice. Hepatology. 2002;35:525–534. MEDLINE | CrossRef

76. 76Petkov PM, Zavadil J, Goetz D, et al. Gene expression pattern in hepatic stem/progenitor cells during rat fetal development using complementary DNA microarrays. Hepatology. 2004;39:617–627. MEDLINE | CrossRef

77. 77Kyrmizi I, Hatzis P, Katrakili N, et al. Plasticity and expanding complexity of the hepatic transcription factor network during liver development. Genes Dev. 2006;20:2293–2305. MEDLINE | CrossRef

78. 78Beaudry JB, Pierreux CE, Hayhurst GP, et al. Threshold levels of hepatocyte nuclear factor 6 (HNF-6) acting in synergy with HNF-4 and PGC-1alpha are required for time-specific gene expression during liver development. Mol Cell Biol. 2006;26:6037–6046. MEDLINE | CrossRef

79. 79Yoshida Y, Hughes DE, Rausa FM, et al. C/EBPalpha and HNF6 protein complex formation stimulates HNF6-dependent transcription by CBP coactivator recruitment in HepG2 cells. Hepatology. 2006;43:276–286. MEDLINE | CrossRef

80. 80Schrem H, Klempnauer J, Borlak J. Liver-enriched transcription factors in liver function and development (Part I: the hepatocyte nuclear factor network and liver-specific gene expression). Pharmacol Rev. 2002;54:129–158. MEDLINE | CrossRef

81. 81Schrem H, Klempnauer J, Borlak J. Liver-enriched transcription factors in liver function and development (Part II: the C/EBPs and D site-binding protein in cell cycle control, carcinogenesis, circadian gene regulation, liver regeneration, apoptosis, and liver-specific gene regulation). Pharmacol Rev. 2004;56:291–330. MEDLINE | CrossRef

82. 82Costa RH, Kalinichenko VV, Holterman AX, et al. Transcription factors in liver development, differentiation, and regeneration. Hepatology. 2003;38:1331–1347. MEDLINE | CrossRef

83. 83Odom DT, Zizlsperger N, Gordon DB, et al. Control of pancreas and liver gene expression by HNF transcription factors. Science. 2004;303:1378–1381. CrossRef

84. 84Odom DT, Dowell RD, Jacobsen ES, et al. Core transcriptional regulatory circuitry in human hepatocytes. Mol Syst Biol. 2006;2:2006–0017.

85. 85Coffinier C, Gresh L, Fiette L, et al. Bile system morphogenesis defects and liver dysfunction upon targeted deletion of HNF1beta. Development. 2002;129:1829–1838. MEDLINE

86. 86Friedman JR, Kaestner KH. The Foxa family of transcription factors in development and metabolism. Cell Mol Life Sci. 2006;63:2317–2328. CrossRef

87. 87Fayard E, Auwerx J, Schoonjans K. LRH-1: an orphan nuclear receptor involved in development, metabolism and steroidogenesis. Trends Cell Biol. 2004;14:250–260. MEDLINE | CrossRef

88. 88Lee YK, Moore DD. Liver receptor homolog-1, an emerging metabolic modulator. Front Biosci. 2008;13:5950–5958.

89. 89Lahuna O, Fernandez L, Karlsson H, et al. Expression of hepatocyte nuclear factor 6 in rat liver is sex-dependent and regulated by growth hormone. Proc Natl Acad Sci U S A. 1997;94:12309–12313. MEDLINE | CrossRef

90. 90Pierreux CE, Stafford J, Demonte D, et al. Antiglucocorticoid activity of hepatocyte nuclear factor-6. Proc Natl Acad Sci U S A. 1999;96:8961–8966. MEDLINE | CrossRef

91. 91Lannoy VJ, Decaux JF, Pierreux CE, et al. Liver glucokinase gene expression is controlled by the onecut transcription factor hepatocyte nuclear factor-6. Diabetologia. 2002;45:1136–1141. CrossRef

92. 92Wang M, Tan Y, Costa RH, et al. In vivo regulation of murine CYP7A1 by HNF-6: a novel mechanism for diminished CYP7A1 expression in biliary obstruction. Hepatology. 2004;40:600–608. MEDLINE | CrossRef

93. 93Tan Y, Yoshida Y, Hughes DE, et al. Increased expression of hepatocyte nuclear factor 6 stimulates hepatocyte proliferation during mouse liver regeneration. Gastroenterology. 2006;130:1283–1300. Abstract | Full Text | Full-Text PDF (1104 KB) | CrossRef

94. 94Lekstrom-Himes J, Xanthopoulos KG. Biological role of the CCAAT/enhancer-binding protein family of transcription factors. J Biol Chem. 1998;273:28545–28548. MEDLINE | CrossRef

95. 95Parviz F, Matullo C, Garrison WD, et al. Hepatocyte nuclear factor 4alpha controls the development of a hepatic epithelium and liver morphogenesis. Nat Genet. 2003;34:292–296. MEDLINE | CrossRef

96. 96Qu X, Lam E, Doughman YQ, et al. Cited2, a coactivator of HNF4alpha, is essential for liver development. EMBO J. 2007;26:4445–4456. CrossRef

97. 97Battle MA, Konopka G, Parviz F, et al. Hepatocyte nuclear factor 4alpha orchestrates expression of cell adhesion proteins during the epithelial transformation of the developing liver. Proc Natl Acad Sci U S A. 2006;103:8419–8424. MEDLINE | CrossRef

98. 98Luebke-Wheeler J, Zhang K, Battle M, et al. Hepatocyte nuclear factor 4alpha is implicated in endoplasmic reticulum stress-induced acute phase response by regulating expression of cyclic adenosine monophosphate responsive element binding protein H. Hepatology. 2008;48:1242–1250. CrossRef

99. 99Hayhurst GP, Strick-Marchand H, Mulet C, et al. Morphogenetic competence of HNF4 alpha-deficient mouse hepatic cells. J Hepatol. 2008;49:384–395. Abstract | Full Text | Full-Text PDF (2249 KB) | CrossRef

100. 100Perincheri S, Dingle RW, Peterson ML, et al. Hereditary persistence of alpha-fetoprotein and H19 expression in liver of BALB/cJ mice is due to a retrovirus insertion in the Zhx2 gene. Proc Natl Acad Sci U S A. 2005;102:396–401. MEDLINE | CrossRef

101. 101Xie Z, Zhang H, Tsai W, et al. Zinc finger protein ZBTB20 is a key repressor of alpha-fetoprotein gene transcription in liver. Proc Natl Acad Sci U S A. 2008;105:10859–10864. CrossRef

102. 102Tatarakis A, Margaritis T, Martinez-Jimenez CP, et al. Dominant and redundant functions of TFIID involved in the regulation of hepatic genes. Mol Cell. 2008;31:531–543. CrossRef

103. 103Gresh L, Bourachot B, Reimann A, et al. The SWI/SNF chromatin-remodeling complex subunit SNF5 is essential for hepatocyte differentiation. EMBO J. 2005;24:3313–3324. MEDLINE | CrossRef

104. 104Kamiya A, Kinoshita T, Ito Y, et al. Fetal liver development requires a paracrine action of oncostatin M through the gp130 signal transducer. EMBO J. 1999;18:2127–2136. MEDLINE | CrossRef

105. 105Ito Y, Matsui T, Kamiya A, et al. Retroviral gene transfer of signaling molecules into murine fetal hepatocytes defines distinct roles for the STAT3 and ras pathways during hepatic development. Hepatology. 2000;32:1370–1376. MEDLINE | CrossRef

106. 106Anzai H, Kamiya A, Shirato H, et al. Impaired differentiation of fetal hepatocytes in homozygous jumonji mice. Mech Dev. 2003;120:791–800. MEDLINE | CrossRef

107. 107Tanaka M, Hirabayashi Y, Sekiguchi T, et al. Targeted disruption of oncostatin M receptor results in altered hematopoiesis. Blood. 2003;102:3154–3162. MEDLINE | CrossRef

108. 108Kamiya A, Gonzalez FJ. TNF-alpha regulates mouse fetal hepatic maturation induced by oncostatin M and extracellular matrices. Hepatology. 2004;40:527–536. MEDLINE | CrossRef

109. 109Suzuki T, Kanai Y, Hara T, et al. Crucial role of the small GTPase ARF6 in hepatic cord formation during liver development. Mol Cell Biol. 2006;26:6149–6156. MEDLINE | CrossRef

110. 110Sakaguchi TF, Sadler KC, Crosnier C, et al. Endothelial signals modulate hepatocyte apicobasal polarization in zebrafish. Curr Biol. 2008;18:1565–1571. CrossRef

111. 111Hart SN, Cui Y, Klaassen CD, et al. Three patterns of cytochrome P450 gene expression during liver maturation in mice. Drug Metab Dispos. 2009;37:116–121. CrossRef

112. 112Kamiya A, Kinoshita T, Miyajima A. Oncostatin M and hepatocyte growth factor induce hepatic maturation via distinct signaling pathways. FEBS Lett. 2001;492:90–94. Abstract | Full Text | Full-Text PDF (390 KB) | CrossRef

113. 113Katz N, Jungermann K. Autoregulatory shift from fructolysis to lactate gluconeogenisis in rat hepatocyte suspensions (The problem of metabolic zonation of liver parenchyma). Hoppe Seylers Z Physiol Chem. 1976;357:359–375. MEDLINE

114. 114Jungermann K, Katz N. Functional specialization of different hepatocyte populations. Physiol Rev. 1989;69:708–764. MEDLINE

115. 115Gebhardt R, Baldysiak-Figiel A, Krügel V, et al. Hepatocellular expression of glutamine synthetase: an indicator of morphogen actions as master regulators of zonation in adult liver. Prog Histochem Cytochem. 2007;41:201–266. MEDLINE | CrossRef

116. 116Stanulović VS, Kyrmizi I, Kruithof-de Julio M, et al. Hepatic HNF4alpha deficiency induces periportal expression of glutamine synthetase and other pericentral enzymes. Hepatology. 2007;45:433–444. MEDLINE | CrossRef

117. 117Benhamouche S, Decaens T, Godard C, et al. Apc tumor suppressor gene is the ”zonation-keeper” of mouse liver. Dev Cell. 2006;10:759–770. MEDLINE | CrossRef

118. 118Zern MA. Cell transplantation to replace whole liver transplantation. Gastroenterology. 2009;136:767–769. Full Text | Full-Text PDF (415 KB) | CrossRef

119. 119Lavon N, Benvenisty N. Study of hepatocyte differentiation using embryonic stem cells. J Cell Biochem. 2005;96:1193–1202. MEDLINE | CrossRef

120. 120Gouon-Evans V, Boussemart L, Gadue P, et al. BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm. Nat Biotechnol. 2006;24:1402–1411. MEDLINE | CrossRef

121. 121Hay DC, Fletcher J, Payne C, et al. Highly efficient differentiation of hESCs to functional hepatic endoderm requires ActivinA and Wnt3a signaling. Proc Natl Acad Sci U S A. 2008;105:12301–12306. CrossRef

122. 122Roskams T, Desmet V. Embryology of extra- and intrahepatic bile ducts, the ductal plate. Anat Rec. 2008;291:628–635.

123. 123Sherwood RI, Chen TY, Melton DA. Transcriptional dynamics of endodermal organ formation. Dev Dyn. 2009;238:29–42. CrossRef

124. 124Sumazaki R, Shiojiri N, Isoyama S, et al. Conversion of biliary system to pancreatic tissue in Hes1-deficient mice. Nat Genet. 2004;36:83–87. MEDLINE | CrossRef

125. 125Fukuda A, Kawaguchi Y, Furuyama K, et al. Ectopic pancreas formation in Hes1 -knockout mice reveals plasticity of endodermal progenitors of the gut, bile duct, and pancreas. J Clin Invest. 2006;116:1484–1493. MEDLINE | CrossRef

126. 126Fukuda A, Kawaguchi Y, Furuyama K, et al. Loss of the major duodenal papilla results in brown pigment biliary stone formation in pdx1 null mice. Gastroenterology. 2006;130:855–867. Abstract | Full Text | Full-Text PDF (1193 KB) | CrossRef

127. 127Kalinichenko VV, Zhou Y, Bhattacharyya D, et al. Haploinsufficiency of the mouse Forkhead Box f1 gene causes defects in gall bladder development. J Biol Chem. 2002;277:12369–12374. MEDLINE | CrossRef

128. 128Tan CE, Moscoso GJ. The developing human biliary system at the porta hepatis level between 29 days and 8 weeks of gestation: a way to understanding biliary atresia. Part 1. Pathol Int. 1994;44:587–599. MEDLINE | CrossRef

129. 129Tan CE, Moscoso GJ. The developing human biliary system at the porta hepatis level between 11 and 25 weeks of gestation: a way to understanding biliary atresia. Part 2. Pathol Int. 1994;44:600–610. MEDLINE | CrossRef

130. 130Ito A, Nishikawa Y, Ohnuma K, et al. SgIGSF is a novel biliary-epithelial cell adhesion molecule mediating duct/ductule development. Hepatology. 2007;45:684–694. MEDLINE | CrossRef

131. 131Fabris L, Strazzabosco M, Crosby HA, et al. Characterization and isolation of ductular cells coexpressing neural cell adhesion molecule and Bcl-2 from primary cholangiopathies and ductal plate malformations. Am J Pathol. 2000;156:1599–1612. MEDLINE

132. 132Yamasaki H, Sada A, Iwata T, et al. Suppression of C/EBPα expression in periportal hepatoblasts may stimulate biliary cell differentiation through increased Hnf6 and Hnf1β expression. Development. 2006;133:4233–4243. MEDLINE | CrossRef

133. 133Kodama Y, Hijikata M, Kageyama R, et al. The role of notch signaling in the development of intrahepatic bile ducts. Gastroenterology. 2004;127:1775–1786. Abstract | Full Text | Full-Text PDF (824 KB) | CrossRef

134. 134Loomes KM, Russo P, Ryan M, et al. Bile duct proliferation in liver-specific Jag1 conditional knockout mice: effects of gene dosage. Hepatology. 2007;45:323–330. MEDLINE | CrossRef

135. 135Lorent K, Yeo SY, Oda T, et al. Inhibition of Jagged-mediated Notch signaling disrupts zebrafish biliary development and generates multi-organ defects compatible with an Alagille syndrome phenocopy. Development. 2004;131:5753–5766. MEDLINE | CrossRef

136. 136McCright B, Lozier J, Gridley T. A mouse model of Alagille syndrome: Notch2 as a genetic modifier of Jag1 haploinsufficiency. Development. 2002;129:1075–1082. MEDLINE

137. 137Lozier J, McCright B, Gridley T. Notch signaling regulates bile duct morphogenesis in mice. PLoS One. 2008;3:e1851.

138. 138Geisler F, Nagl F, Mazur PK, et al. Liver-specific inactivation of Notch2, but not Notch1, compromises intrahepatic bile duct development in mice. Hepatology. 2008;48:607–616. CrossRef

139. 139Suzuki K, Tanaka M, Watanabe N, et al. p75 Neurotrophin receptor is a marker for precursors of stellate cells and portal fibroblasts in mouse fetal liver. Gastroenterology. 2008;135:270–281. Abstract | Full Text | Full-Text PDF (7623 KB) | CrossRef

140. 140Lemaigre FP. Notch signaling in bile duct development: new insights raise new questions. Hepatology. 2008;48:358–360. CrossRef

141. 141Everson GT, Taylor MR, Doctor RB. Polycystic disease of the liver. Hepatology. 2004;40:774–782. MEDLINE | CrossRef

142. 142Johnson CA, Gissen P, Sergi C. Molecular pathology and genetics of congenital hepatorenal fibrocystic syndromes. J Med Genet. 2003;40:311–319.

143. 143Kamath BM, Piccoli DA. Heritable disorders of the bile ducts. Gastroenterol Clin North Am. 2003;32:857–875. Full Text | Full-Text PDF (148 KB) | CrossRef

144. 144Adams M, Smith UM, Logan CV, et al. Recent advances in the molecular pathology, cell biology and genetics of ciliopathies. J Med Genet. 2008;45:257–267.

145. 145Masyuk AI, Masyuk TV, LaRusso NF. Cholangiocyte primary cilia in liver health and disease. Dev Dyn. 2008;237:2007–2012. CrossRef

146. 146Masyuk AI, Masyuk TV, Splinter PL, et al. Cholangiocyte cilia detect changes in luminal fluid flow and transmit them into intracellular Ca2+ and cAMP signaling. Gastroenterology. 2006;131:911–920. Abstract | Full Text | Full-Text PDF (1494 KB) | CrossRef

147. 147Gradilone SA, Masyuk AI, Splinter PL, et al. Cholangiocyte cilia express TRPV4 and detect changes in luminal tonicity inducing bicarbonate secretion. Proc Natl Acad Sci U S A. 2007;104:19138–19143. CrossRef

148. 148Masyuk AI, Gradilone SA, Banales JM, et al. Cholangiocyte primary cilia are chemosensory organelles that detect biliary nucleotides via P2Y12 purinergic receptors. Am J Physiol Gastrointest Liver Physiol. 2008;295:G725–G734. CrossRef

149. 149Bagnat M, Cheung ID, Mostov KE, et al. Genetic control of single lumen formation in the zebrafish gut. Nat Cell Biol. 2007;9:954–960. CrossRef

150. 150Tanimizu N, Miyajima A, Mostov KE. Liver progenitor cells develop cholangiocyte-type epithelial polarity in three-dimensional culture. Mol Biol Cell. 2007;18:1472–1479. MEDLINE | CrossRef

151. 151Couvelard A, Bringuier AF, Dauge MC, et al. Expression of integrins during liver organogenesis in humans. Hepatology. 1998;27:839–847. MEDLINE | CrossRef

152. 152Fabris L, Cadamuro M, Fiorotto R, et al. Effects of angiogenic factor overexpression by human and rodent cholangiocytes in polycystic liver diseases. Hepatology. 2006;43:1001–1012. MEDLINE | CrossRef

153. 153Sato Y, Harada K, Kizawa K, et al. Activation of the MEK5/ERK5 cascade is responsible for biliary dysgenesis in a rat model of Caroli's disease. Am J Pathol. 2005;166:49–60. MEDLINE

154. 154Sato Y, Harada K, Furubo S, et al. Inhibition of intrahepatic bile duct dilation of the polycystic kidney rat with a novel tyrosine kinase inhibitor gefitinib. Am J Pathol. 2006;169:1238–1250. MEDLINE | CrossRef

155. 155Lee SO, Masyuk T, Splinter P, et al. MicroRNA15a modulates expression of the cell-cycle regulator Cdc25A and affects hepatic cystogenesis in a rat model of polycystic kidney disease. J Clin Invest. 2008;118:3714–3724.

156. 156Gissen P, Johnson CA, Morgan NV, et al. Mutations in VPS33B, encoding a regulator of SNARE-dependent membrane fusion, cause arthrogryposis-renal dysfunction-cholestasis (ARC) syndrome. Nat Genet. 2004;36:400–404. MEDLINE | CrossRef

157. 157Sadler KC, Amsterdam A, Soroka C, et al. A genetic screen in zebrafish identifies the mutants vps18, nf2 and foie gras as models of liver disease. Development. 2005;132:3561–3572. MEDLINE | CrossRef

158. 158Matthews RP, Plumb-Rudewiez N, Lorent K, et al. Zebrafish vps33b, an ortholog of the gene responsible for human arthrogryposis-renal dysfunction-cholestasis syndrome, regulates biliary development downstream of the onecut transcription factor hnf6. Development. 2005;132:5295–5306. MEDLINE | CrossRef

159. 159Clotman F, Libbrecht L, Killingsworth MC, et al. Lack of cilia and differentiation defects in the liver of human foetuses with the Meckel syndrome. Liver Int. 2008;28:377–384.

160. 160Strazzabosco M, Fabris L. Functional anatomy of normal bile ducts. Anat Rec. 2008;291:653–660.

161. 161Glaser S, Francis H, Demorrow S, et al. Heterogeneity of the intrahepatic biliary epithelium. World J Gastroenterol. 2006;12:3523–3536. MEDLINE

162. 162Collardeau-Frachon S, Scoazec JY. Vascular development and differentiation during human liver organogenesis. Anat Rec. 2008;291:614–627.

163. 163Geerts A. On the origin of stellate cells: mesodermal, endodermal or neuro-ectodermal?. J Hepatol. 2004;40:331–334. Full Text | Full-Text PDF (82 KB) | CrossRef

164. 164Friedman SL. Hepatic stellate cells: protean, multifunctional, and enigmatic cells of the liver. Physiol Rev. 2008;88:125–172. CrossRef

165. 165Yang L, Jung Y, Omenetti A, et al. Fate-mapping evidence that hepatic stellate cells are epithelial progenitors in adult mouse livers. Stem Cells. 2008;26:2104–2413. CrossRef

de Duve Institute, Université Catholique de Louvain, Brussels, Belgium

Reprint requests Address requests for reprints to: Frédéric P. Lemaigre, MD, PhD, de Duve Institute, Université Catholique de Louvain, Avenue Hippocrate 75/7529, 1200 Brussels, Belgium. fax (32) 2 764 7507

Conflicts of interest The author discloses no conflicts.

Funding Supported by the Interuniversity Attraction Poles Program (Belgian Science Policy), the D.G. Higher Education and Scientific Research of the French Community of Belgium, the Fund for Scientific Medical Research (Belgium), and the Alphonse & Jean Forton Fund.

PII: S0016-5085(09)00463-6

doi:10.1053/j.gastro.2009.03.035

33 of 80