Human hepatic stem cell and maturational liver lineage biology


  • Rachael Turner,

    1. Department of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, NC
    2. Department of Biomedical Engineering, University of North Carolina School of Medicine, Chapel Hill, NC
    Search for more papers by this author
    • R. Turner & L.M. Reid did the majority of the writing & editing of the review. O. Lozoya prepared the sections on mechanical effects on cells, & Y. Wang contributed those on regulation by paracrine signaling. Comments on biliary tree biology & cholangiocytes were written & edited by D. Alvaro, E. Gaudio, V. Cardinale, & G. Alpini, & those on biliary tree stem cells were written by L.M. Reid, Y. Wang & V. Cardinale. Schematic figures were drawn by R. Turner & G. Mendel. C. Barbier & E. Wauthier helped edit the review.

    • Senior coauthors.

  • Oswaldo Lozoya,

    1. Department of Biomedical Engineering, University of North Carolina School of Medicine, Chapel Hill, NC
    Search for more papers by this author
  • Yunfang Wang,

    1. Department of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, NC
    Search for more papers by this author
  • Vincenzo Cardinale,

    1. Division of Gastroenterology, Department of Clinical Medicine, University of Rome, Rome, Italy
    Search for more papers by this author
  • Eugenio Gaudio,

    1. Department of Human Anatomy, University of Rome, Rome, Italy
    Search for more papers by this author
  • Gianfranco Alpini,

    1. Division of Research, Central Texas Veterans Health Care System, Department of Medicine, Scott & White Digestive Disease Research Center, Division of Research and Education, Scott & White, and Texas A&M Health Science Center College of Medicine, Temple, TX
    Search for more papers by this author
  • Gemma Mendel,

    1. Department of Biomedical Engineering, University of North Carolina School of Medicine, Chapel Hill, NC
    Search for more papers by this author
  • Eliane Wauthier,

    1. Department of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, NC
    Search for more papers by this author
  • Claire Barbier,

    1. Department of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, NC
    Search for more papers by this author
  • Domenico Alvaro,

    1. Division of Gastroenterology, Department of Clinical Medicine, University of Rome, Rome, Italy
    Search for more papers by this author
    • Senior coauthors.

  • Lola M. Reid

    Corresponding author
    1. Department of Cell and Molecular Physiology, University of North Carolina School of Medicine, Chapel Hill, NC
    2. Department of Biomedical Engineering, University of North Carolina School of Medicine, Chapel Hill, NC
    3. Program in Molecular Biology and Biotechnology, University of North Carolina School of Medicine, Chapel Hill, NC
    • UNC School of Medicine, Campus Box 7038, Glaxo Building Rms. 32-35, Chapel Hill, NC 27599===

    Search for more papers by this author
    • Senior coauthors.

  • Funding for the investigators at UNC derived from NIH grants (DK52851, AA014243, IP30-DK065933), a DOE grant (DE-FG02-02ER-63477), & sponsored research grants from Vesta Therapeutics (Bethesda, MD), Vertex Pharmaceuticals (Cambridge, MA), & GigaCyte (Branford, CT). D. Alvaro is supported by MIUR (Italian Minister of University & Research) grants PRIN #2007, prot.2007HPT7BA-003 & by Federate Atheneaum funds from the University “Sapienza” of Rome. Dr. Gaudio was supported by MIUR grants PRIN#2007, prot. 2007HPT7BA_001 & Federate Atheneaum funds from the University “Sapienza” of Rome. Dr. Alpini is supported by the Dr. Nicholas C. Hightower Centennial Chair of Gastroenterology from Scott & White Hospital, a VA Research Scholar Award, a VA Merit Award, & NIH grants DK062975, K76898, & DK58411.

  • fax: (919) 966-6112


Livers are comprised of maturational lineages of cells beginning extrahepatically in the hepato-pancreatic common duct near the duodenum and intrahepatically in zone 1 by the portal triads. The extrahepatic stem cell niches are the peribiliary glands deep within the walls of the bile ducts; those intrahepatically are the canals of Hering in postnatal livers and that derive from ductal plates in fetal livers. Intrahepatically, there are at least eight maturational lineage stages from the stem cells in zone 1 (periportal), through the midacinar region (zone 2), to the most mature cells and apoptotic cells found pericentrally in zone 3. Those found in the biliary tree are still being defined. Parenchymal cells are closely associated with lineages of mesenchymal cells, and their maturation is coordinated. Each lineage stage consists of parenchymal and mesenchymal cell partners distinguishable by their morphology, ploidy, antigens, biochemical traits, gene expression, and ability to divide. They are governed by changes in chromatin (e.g., methylation), gradients of paracrine signals (soluble factors and insoluble extracellular matrix components), mechanical forces, and feedback loop signals derived from late lineage cells. Feedback loop signals, secreted by late lineage stage cells into bile, flow back to the periportal area and regulate the stem cells and other early lineage stage cells in mechanisms dictating the size of the liver mass. Recognition of maturational lineage biology and its regulation by these multiple mechanisms offers new understandings of liver biology, pathologies, and strategies for regenerative medicine and treatment of liver cancers. (HEPATOLOGY 2011;)


With respect to specific cellular subpopulations, the species of origin is indicated by a small letter in front of the abbreviation (r = rat; m = mouse; h = human); ABAT, apical bile acid transporter; AFP, α-fetoprotein; αSMA, α-smooth muscle actin; bcl2 protein, B-cell lymphoma 2 protein; CCl4, carbon tetrachloride; CFTR, cystic fibrosis transmembrane conductance regulator; CK, cytokeratin; CS-PG, chondroitin sulfate proteoglycan; DS-PG, dermatan sulfate proteoglycan; ENDRA and ENDRB, endothelin receptors type A and type B; EpCAM, epithelial cell adhesion molecules; GABA, γ-aminobutyric acid; HBs, hepatoblasts; HpSTCs, hepatic stellate cells; HpSCs, hepatic stem cells; HP-PG, heparin proteoglycan; HS-PG, heparan sulfate proteoglycan; ICAM, intercellular adhesion molecules; KDR, kinase insert domain receptor; MDR, multidrug transporter; MHC, major histocompatibility complex; MRP, multidrug resistance associated proteins; NCAM, neural cell adhesion molecule; NTPDase2, ectonucleotidase; PG, proteoglycan; SHH and IHH proteins, Sonic and Indian hedgehog proteins; TGF-β, transforming growth factor beta; VCAM, vascular cell adhesion molecule; VEGFr, vascular endothelial growth factor receptor.

The Liver's Maturational Lineages

Hepatic stem cells and their mesenchymal partners, angioblasts, give rise to daughter cells maturing into lineages of parenchymal and mesenchymal cells with stepwise changes in cell size, morphology, ploidy, gene expression, growth potential, and signaling.1-4 Currently, there is evidence for at least eight intrahepatic lineage stages (Figs. 1, 2).5, 6 Continued efforts to characterize the liver's lineage biology should result in recognition of additional stages. This overview focuses on early intrahepatic lineage stages in human livers and includes aspects of their regulation. Information on later lineage stages of cells, and additional background and references are included in the online supplement.

Figure 1.

Schematic image of liver, the biliary tree, and panceas and their connections with the duodenum. The blue stars indicate sites at which there are high numbers of peribiliary glands, the stem cell niches of the biliary tree.

Figure 2.

Schematic image of intrahepatic maturational lineages.

Stage 1.

Hepatic stem cells (hHpSCs) are multipotent stem cells located within the liver's stem cell compartment, the ductal plates of fetal and neonatal livers, and canals of Hering in pediatric and adult livers.5, 7-12 The compartment represents the anatomic and physiological link between the intralobular canalicular system of hepatocytes and the biliary tree and resides along sites that project starlike from the portal tracts. They constitute ≈0.5%-2% of the parenchyma of livers of all age donors. The hHpSCs cells range in size from 7-9 μm in diameter and have a high nucleus-to-cytoplasm ratio. Tolerant of ischemia, they can remain viable in cadaveric livers for up to ≈6 days after asystolic death.11, 12

The hHpSC phenotypic profile includes epithelial cell adhesion molecule (EpCAM), neural cell adhesion molecule (NCAM), CD133, CXCR4, SOX9, SOX17, FOXA2, cytokeratins (CK) 8/18/19, Hedgehog proteins (Sonic and Indian), intranuclear telomerase protein, claudin 3, MDR1, weak or negligible expression of albumin and major histocompatibility complex (MHC) antigens. They do not express α-fetoprotein (AFP), intercellular adhesion molecule (ICAM-1), P450s, or markers for hemopoietic (e.g., CD34/38/45/90, glycophorin), endothelial (e.g., vascular endothelial growth factor receptor [VEGFr], CD31, von Willebrand factor), or mesenchymal cells (e.g., CD146, desmin, vitamin A, CD105).6, 7, 13 It remains unclear whether C-kit (CD117), expressed in the liver's stem cell niches,8, 14, 15 is on hHpSCs or associated angioblasts, as CD117+ flow cytometry selects for angioblasts.7, 13

Some proteins, such as CK19, are synthesized and found in punctuate form, but not converted to filaments, as seen in hepatoblasts.7 Similarly, small amounts of albumin are synthesized and are not packaged as in later lineage stages, implicating lineage-dependent distinctions in posttranscriptional and translational protein processing.

The hHpSCs are isolated by dual immunoselection for EpCAM+/NCAM+ cells from livers of all donor ages. In adult livers, which have scarce hepatoblast populations, EpCAM+ selection alone results in isolation of predominantly hHpSCs.7, 16

In culture, the hHpSCs form colonies capable of self-replication17 and of differentiation to mature cells in culture and in vivo.7, 18 Cells expand ex vivo if cultured in Kubota's medium, a serum-free medium containing only insulin, transferrin/fe, lipids, no copper, and low calcium19, 20 and if cocultured with angioblasts. These feeders are replaceable with purified type III collagen substrates, embedding into weakly crosslinked hyaluronan hydrogels, or a mixture of both.13, 21 If transplanted in vivo, they yield mature liver tissue. If cultured under distinct conditions (see below) they lineage-restrict into hepatoblasts.13

Stage 2.

Hepatoblasts (hHBs) are diploid bipotent cells giving rise to hepatocytic and cholangiocytic lineages, associated with precursors of both endothelia and hepatic stellate cells, and the liver's probable transit amplifying cells.13 They reside throughout parenchyma of fetal and neonatal livers or as single cells and small cell aggregates tethered to the ends of canals of Hering in adult livers.8 With donor age, hHBs decline to <0.01% of the parenchymal cells in postnatal livers.7, 8 They expand during regenerative processes associated with certain diseases such as cirrhosis. Previously, hHBs were referred to as “intermediate hepatobiliary cells of the ductular reactions”22; extensive characterization enabled us to refer to them, as hepatoblasts.8 They can be isolated by dual immunoselection for EpCAM+/ICAM-1+. They have enormous expansion potential cultured in Kubota's medium, especially if supplemented with epidermal growth factor (EGF) and hepatocyte growth factor (HGF), or on feeders of stellate cell precursors replaceable by substrata of type IV collagen, laminin, hyaluronans, or mixtures of these expansion is without proven self-replication.13, 23, 24

The hHBs, larger (10-12 μm) and with higher amounts of cytoplasm than hHpSCs, have an antigenic profile that overlaps with hHpSCs.6, 7, 15 Shared phenotypic traits include CXCR4, CD133, SOX17, MDR1, cytokeratins (CK) 8/18 and 19, Hedgehog proteins (Sonic and Indian), and null expression of late P450s (e.g., P450-3A) or markers for hemopoietic, endothelia, or mesenchymal cells (as for hHpSCs). Protein expression changes include reduction in EpCAM levels with primary localization to plasma membrane surfaces; filamentous CK14 and CK198, 15, 25; elevated albumin levels with discrete cytoplasmic packaging7; switch from NCAM to ICAM-1; expression of early P450s (e.g., P450-A7) and CK7; and strong positive expression of hepatic-specific AFP, distinct from a hemopoietic progenitor variant form with alternative splicing of exon 1, a probable clue of mesendoderm to endoderm differentiation.26 They have ≈5× the telomerase activity found in hHpSCs and with telomerase protein localized both in the nucleus and in the cytoplasm.27 A comparison of the phenotypic profiles of HpSCs and HBs can be found in Table 1 and in Figs. 3, 4.

Figure 3.

Human hepatic stem cell and hepatoblast phenotypes in vivo. (A,B) EpCAM expression in fetal liver: EpCAM expressed in the ductal plate is not only at the cell surface but also in the cytoplasm. EpCAM expressed in the hepatoblasts is restricted to the cell surface. (D,E) EpCAM expression in adult liver: one end of the Canal of Hering connects to the bile duct, the other end connects to hepatoblasts (arrow), indicating that the hepatoblasts are derived from hepatic stem cells harbored in Canals of Hering. (C) Double staining for CK-19/AFP and (F) Ep-CAM/AFP of human fetal liver in the portal triad area and analyzed by confocal microscopy. CK-19 (C, green) is expressed not only by remodeling ductal plates but faintly expressed by some of the hepatoblasts. Ep-CAM (F, green) is detected in all the parenchymal cells and biliary epithelial cells forming bile duct and ductal plate (DP). AFP (red) is expressed by hepatoblasts throughout the fetal liver and is undetectable in the ductal plate. PT: portal triad; DP: ductal plate. Originally published in Hepatology, DOI:10.1002/hep.22516 and J Exp Med, DOI:10.1084/jem.20061603.

Figure 4.

Human hepatic stem cell and hepatoblast phenotypes in culture. (A) Morphology of human hepatic stem cells and (B) human hepatoblasts in culture on plastic. (C) Albumin staining of human hepatic stem cells, transitioning to hepatoblasts. (D) hepatic stem cells expressing NCAM (green) and hepatoblasts expressing ICAM (red). Originally published in J Exp Med DOI:10.1084/jem.20061603.

Table 1. Phenotypic Profiles of Multipotent Cell Populations in Human Livers
PropertyHuman Hepatic Stem Cells (hHpSCs)Human Hepatoblasts (hHBs)
  • *

    Mesenchymal markers found on hepatic stellate cells and myofibroblasts: α-smooth muscle actin (ASMA), desmin, CD146 (MelCAM), vitamin A, VCAM, HGF, and SDF-1α. The precursors to the stellate cells express lower levels of these markers.

  • **

    Angioblast/Endothelial cell markers: CD31, VEGFr, Von Willebrand Factor, VCAM, and on hemangioblast subpopulations: CD34. Note: CD117 was found inconsistently on cells within hHpSC colonies and was never observed on hHB colonies. Flow cytometric sorting for CD117+ cells from fetal liver cell suspensions yielded angioblasts. Therefore, we assume that it is on angioblasts closely associated with the hHpSCs.

  • ***

    Hemopoietic cell markers: CD14, CD34, CD38, CD45, CD90, glycophorin A (red blood cell antigen).

Average diameter (measured by forward scatter in flow cytometric analyses of isolated cells)7-9 μm10-12 μm
Nucleus to cytoplasmic ratioHighest observed of all parenchymal progenitor subpopulations evaluatedIntermediate between that in hHpSCs and mature parenchymal cells
Percentage of parenchymal cells(7)0.5-2.0 % in livers of all donor ages and with minimal ischemia; percentages higher in ischemic livers>80% (fetal livers)
∼50% (neonatal livers) [percentages change rapidly day by day postnatally]
<0.01% (adult livers)
Survival after cardiac arrest (tolerance for ischemia)(7)Viable cells for several days after cardiac arrestViable for more than a day, but not as long as hHpSCs
Morphology of colonies in vitro(7)Uniform; densely packed; look similar to ES cell coloniesCord-like colonies interspersed with clear channels that are presumptive canaliculi
Evidence for self-renewal(17)Clonogenic expansion with stability of phenotype; doubling times of ∼36 hours on plastic; can be passaged repeatedly; fastest doubling times (∼20-24 hours) for hHpSCs on substrata of type III collagenSignificant expansion potential but not yet evidence for self-replication (under the conditions tested to date). Probable transit amplifying cells
Anaerobic metabolism (metabolomic studies)(21)++++++
Conditions for clonogenic expansion(19)Kubota's Medium plus feeders of angioblasts replaceable with type III collagen (monolayers) or uncrosslinked hyaluronans into which is mixed type III collagen (3-D)Kubota's Medium plus feeders of hepatic stellate cells replaceable with type IV collagen/laminin (monolayers) or uncrosslinked hyaluronans into which is mixed type IV collagen/laminin (3-D)
CD44 (hyaluronan receptor)(24)High concentrationsHigh concentrations
Claudin 3(16)+++Negative
Indian Hedgehog(5)+++++
 Highest level in cells in the center of the coloniesLower levels, but pattern of distribution is the same
Sonic Hedgehog(5)+++
 Located at edge of cells; concentrated in cells at edge of colonies at sites of high concentration of angioblastsLower levels, but pattern of distribution is the same
(Hedgehog receptor)(5)Found in all cells and in colonies throughout the colonyLevels lower, but still evident
 mRNA encoding telomerase and the protein found in nucleus. No telomerase protein in the cytoplasmmRNA encoding telomerase and the protein found in nucleus; with differentiation, increasing numbers of the cells have it in the cytoplasm; 5X higher activity than in hHpSCs
P450s(16)Negative for all assayedP450 A7 but not late forms of P450s
CK 8 and 18(16)++++
CK 19(7, 16)++ (not in filament form)++ (filaments evident)
 (throughout the cells)(plasma membrane)
NCAM /ICAM-1(7-8)++ /−−/++
Mesenchymal markers*NegativeNegative
Angioblasts/endothelial cell markers**NegativeNegative
Hemopoietic markers***NegativeNegative

Stage 3.

Committed progenitors are ≈12-15μm diploid, unipotent, immature cells. These precursors give rise to only one adult cell type. They lose most stem cell gene expression (e.g., NCAM, Hedgehog proteins), express either hepatocytic or biliary markers, and abound in fetal and neonatal tissues or chronic liver diseases (viral, alcoholic, and nonalcoholic fatty liver diseases, autoimmune hepatitis, cholangiopathies), unlike normal adult tissues.28

Committed hepatocytic progenitors, also called intermediate hepatocytes, express albumin, enzymes associated with glycogen synthesis (e.g., glucose-6-phosphate), and lack biliary markers (e.g., CK19) and AFP. They are associated with endothelial cell precursors and are located in vivo in the liver plates between the HBs and the diploid adult hepatocytes.

“Small cholangiocytes” are diploid biliary cells, 6-8 μm with cuboidal shape, a high nucleus-to-cytoplasm ratio, small endoplasmic reticulum,29, 30 and are associated with hepatic stellate cell precursors.13 They colocalize with hHpSCs in the stem cell niche, lining the canals of Hering, intrahepatic bile ducts, and bile ductules with internal diameters below 15 μm. Direct links between the canals of Hering and bile ductules, which may traverse the limiting plate and thus may have an intralobular segment (periportal) in addition to their intraportal location, support current hypotheses that point to small cholangiocytes as committed biliary progenitors.31 In human and rodent livers, they express high levels of the antiapoptotic proteins annexin V and bcl2 (B-cell lymphoma 2 protein). At a functional level, they express endothelin receptors type A (EDNRA) and type B (EDNRB), endogenous opioid peptides, insulin, histamine (H1), acetylcholine (M3), and α-1-adrenergic agonists, aquaporin 4. They are negative for the Cl/HCO3 exchanger and receptors for secretin or somatostatin. During chronic feeding with bile salts (taurocholate and taurolithocholate), small cholangiocytes express Na+-dependent apical bile acid transporter (ABAT) de novo, suggesting a role in the cholehepatic recirculation of bile salts in conditions of overload.32 Finally, cystic fibrosis transmembrane conductance regulator (CFTR) is present in human, but not rodent, small cholangiocytes.31

Stages 4-6.

Diploid adult cells are the only parenchymal cells with significant proliferative capacity under all known in vitro or in vivo conditions. Exceptions are in conditions potentially involving genetic reprogramming through chromatin demethylation, the only known mechanism for restoring cytokinensis, as occurs in tyrosinemia33 or with massive loss of mature parenchymal cells (e.g., >80%) due to a suicide transgene.34 Diploid adult hepatocytes (“small hepatocytes”), partnered with endothelia, can undergo six to seven rounds of division within 3 weeks in culture but have limited subcultivation capacity.19 Large cholangiocytes, partnered with stellate cells, are columnar in shape, display a small nucleus and conspicuous cytoplasm, an abundant Golgi apparatus between the apical pole and the nucleus, and rough endoplasmic reticulum more abundant than small cholangiocytes.30, 35, 36 Large cholangiocytes line interlobular ducts located in the portal triads. The connections of hHpSCs in canals of Hering to the septal and segmental bile ducts have not yet been investigated, and markers in septal ducts, segmental ducts, and larger ducts are found also in cells in peribiliary glands, the stem cell niches of the biliary tree.37 Large cholangiocytes express CFTR and Cl/HC03 exchanger, aquaporin 4 and aquaporin 8, secretin and somatostatin receptors other than receptors for hormones and neuropeptides. In addition, they express the Na+-dependent bile acid transporter ABAT (apical bile acid transporter), MDR (multidrug transporter), and MRP (multidrug resistance associated proteins). When large cholangiocytes are damaged by acute carbon tetrachloride (CCl4) or GABA administration, small cholangiocytes proliferate, and acquire phenotypical and functional features of large cholangiocytes,38, 39 suggesting that the population of small cholangiocytes lining the canals of Hering and ductules may represent precursors of large cholangiocytes lining larger ducts. The integrated differential microarray gene expression between small and large normal cholangiocytes demonstrate that the proteins related to cell proliferation tend to be highly expressed by small cholangiocytes, whereas large cholangiocytes express functional and differentiated genes.36 This is consistent with studies showing, either with bile duct injury due to CCl4 and GABA administration or with bile duct regrowth following partial hepatectomy, that small cholangiocyte proliferation is activated presumably to repopulate bile ducts. These findings suggest that small cholangiocytes are less mature, have a high resistance to apoptosis, and have marked proliferative activities, whereas large cholangiocytes are more differentiated contributing mainly to bile secretion and absorption. Therefore, whereas hepatocytic cell lineages proceed from periportal areas towards the central vein, cholangiocytes proceed in the opposite direction from canals of Hering/ductules toward larger ducts. (See the online supplement for further information.)

Regulation of the Parenchymal Cell Lineages

Paracrine Signaling between Epithelial-Mesenchymal Partners.

Paracrine signaling is the primary form of regulation between parenchymal cells and partnering mesenchymal cells and represents classic epithelial-mesenchymal relationships widely described in developmental biology since the 1930s. A new facet is that coordinate maturation of these [parenchymal]:[mesenchymal] cell associations, starting with [hHpSCs]: [angioblasts] and splitting into lineages of [hepatocyte]:[endothelia] and [cholangiocyte]:[stellate cells], gives rise to lineage-dependent gradients of paracrine signals13 that govern the biological responses of cells at each lineage stage. Defined subsets of these lineage-dependent paracrine signals, soluble and insoluble matrix ones, can be used to establish cells at a specific lineage stage in culture (Fig. 5).

Figure 5.

Schematic image indicating the coordinate maturation of the epithelia (parenchymal cells) and their mesenchymal partners and some of the identified extracellular matrix components found at the particular lineage stages. Not shown in the figure are the soluble signals that also are lineage dependent. Some of the lineage dependent soluble signals identified are noted in parentheses beside the lineage stage at which they are found: hepatic stem cells (LIF, IL-6, IL-11, and acetylcholine); hepatoblasts (HGF, EGF, bFGF, IL-6, IL-11, and acetylcholine); hepatocytes (HGF, EGF, bFGF, T3, glucagon, and hydrocortisone); cholangiocytes (VEGF, HGF, bFGF and acetylcholine).

The intrahepatic stem cell niche contains type III collagen, α6β4 integrin-binding form of laminin, hyaluronans, and a minimally sulfated chondroitin sulfate proteoglycan (CS-PG).13 Transition to [hHBs]:[endothelia and stellate cell precursors] results in changes to type IV collagen, αVβ1 integrin-binding laminin, hyaluronans, more sulfated CS-PGs, and forms of heparan sulfate-PGs (HS-PGs). The [hepatocyte]:[endothelia] lineages are associated with network collagens (e.g., type IV and VI) and increasingly sulfated forms of HS-PGs ending, in zone 3, in heparin-PGs (HP-PGs). The [cholangiocyte]:[stellate cell] lineages are associated with fibrillar collagens (e.g., types I and III) and progression from CS-PGs towards highly sulfated PGs, including dermatan sulfate-PGs (DS-PGs).13, 17, 24 Many soluble signals bind to and work synergistically with matrix components to regulate cells, particularly PGs and their glycosaminoglycan chains (GAGs). Matrix-bound soluble signals are biphasic, yielding mitogenic versus differentiative responses depending on the specific matrix chemistry with which they are associated.

Feedback Loop Signaling.

Late lineage stage cells produce positive and negative signaling regulators, including bile salts, various soluble factors, and matrix components.40 Positive regulators include hepatopoietin, released by dying zone 3 cells that stimulate stem/progenitors expansion (M. Roach, J. Hambor, unpubl. observations). Negative regulators include ecto-nucleotidases expressed by portal hepatoblasts like NTPDase2, which inhibits purinergic activation of basolateral P2Y receptors in periportal cholangiocytes under homeostatic conditions. Conversely, loss of NTPDase2 expression after experimental cholestasis in portal hepatoblasts allows activation of periportal P2Y receptors and increases cholangiocyte proliferation.41

Another facet of regulation is mediated by acetylcholine. It stimulates proliferation of stem/progenitor cells and cholangiocytes expressing M3 acetylcholine receptors.42 In normal liver and even after partial hepatectomy, late lineage stage hepatocytes lacking M3 receptors release acetyl cholinesterase into the bile that delivers it to zone 1 where it destroys acetylcholine in the stem cell niche, thus blocking proliferation of stem/progenitor cells and cholangiocytes. In contrast, during conditions of pericentral damage, the acetyl cholinesterase is not released, resulting secondarily in induction by acetylcholine of stem/progenitor cell expansion. Denervated transplanted livers lack acetylcholine modulation of proliferation of cells lining the canals of Hering. Hepatitis-injured transplanted livers also exhibit lower numbers of progenitor and reactive ductular cells than innervated matched controls. Experiments in rats with galactosamine-damaged livers confirm that vagotomy induces impaired regeneration of progenitors and ductal reaction in cholangiocytes.43

Mechanotransduction mechanisms are another major set affecting lineage biology, most involving cytoskeletal rearrangements. The cytoskeleton is a ubiquitous cellular component with characteristics of amplification systems and connections with matrix. Some of these connections allow cells to sense microenvironment rigidity through nonmuscle myosin II, which directs stiffness-dependent differentiation in mesenchymal stem cells.44 Germ layer organization and cell sorting depends on cell adhesion forces and cortex tension relying on actomyosin network activity.45 Integrins connect the cytoskeleton to matrix substrata, recruit focal adhesions that adapt cells to mechanical stresses, bind ligands, and regulate intracellular signaling.46 Mechanical stretch in liver cells induces activation and synthesis of morphogens in the transforming growth factor beta (TGF-β) family of Activin/Nodal signaling.47 SMAD transcription factors regulate TGF-β signaling pathways and regulate gene expression through kinesin-mediated nucleocytoplasmic shuttling along intact microtubules.48, 49

Primary cilia in cells from soft organs also participate in mechanotransduction by probing and amplifying the effects of intraluminal flow above the cells apical surfaces. They mediate polarized signal transduction pathways that use the cytoskeleton to ensure specific and nondiffusable signal trafficking to the nucleus.50 PDGRα and Hedgehog signaling take place in primary cilia51, 52 in livers of all ages5 through dynein-mediated shuttling of Gli transcription factors.53 Some chromatin targets of Gli transcription factors include PTCH, WNT, and BMP genes, all involved in embryonic development and differentiation mechanisms.54-56 Hedgehog expression gradients also demarcate the extension of endodermal organs during development.52, 57 In conjunction, this information suggests primary cilia are relevant participants in endoderm maturation and differentiation.

Bile secretion is an important mechanism for homeostatic control of tissue mass, operating as an inductor in mechanotransduction. Bile is a Newtonian fluid in normal physiological conditions with salt concentration-dependent viscosity.58 As hepatic parenchyma perform secretory functions; bile tonicity increases while flowing in the pericentral-to-periportal direction. Abnormal bile tonicity is characteristic of pathological conditions.59 Shear forces from bile flow, proportional to bile viscosity, can function as long-range mechanical signals communicating states of hepatic function across the entire liver maturational gradient to cholangiocytes in the proximal biliary tree through primary cilia bending. This bending triggers stress-induced Ca2+, cAMP signaling cascades, and receptor-mediated PDGRα and Hedgehog signaling, which makes bile a mechanical probe for liver homeostatic control.42

Liver Regeneration.

Two distinct forms of liver regeneration take place after: (1) partial hepatectomy, and (2) selective loss of pericentral cells. After partial hepatectomy, feedback loop signaling is essentially intact. DNA synthesis occurs in cells across the liver plates but only a portion of the cells undergo cytokinesis, yielding increased numbers of polyploid cells, higher numbers of apoptotic cells, and more rapid turnover of the liver with restoration of the normal ploidy profiles within weeks.60 Feedback loop signaling is the explanation for liver cells in culture in which secreted signals from late lineage stage cells inhibit the growth of any early lineage stage cells.20

Selective loss of pericentral cells with toxic injury to zone 3 cells (and sometimes also to zone 2) results in muting of the feedback loop signaling that activates rapid cell division of early lineage stage cells.12, 61 In response, periportal cells undergo rapid hyperplastic growth (complete cell division) followed by differentiation. These phenomena, the classic “oval cell response” in rodents and the “ductular reactions” seen in humans in massive hepatic necrosis (e.g., acetaminophen toxicity, acute hepatotropic viral infection), have long been recognized to involve extensive expansion of the stem/progenitor cell populations.12 Chronic injury to the liver, as occurs with repeated drug exposures, radiation, or certain viral infections like hepatitis B or C, result in loss of late lineage stage cells, eliciting chronic regenerative responses that can lead to oncogenesis.

Relevance to Clinical Programs.

Hepatic lineage biology and mechanisms of its regulation will have relevance for many clinical programs. Examples include tissue sourcing for clinical programs, strategies for liver cell therapies, immunological issues, and, most profoundly, an understanding of liver tumors and logical strategies by which to treat liver cancers.

Sourcing of tissue for any clinical therapy is dictated by the proportion of cells at the different lineage stages in tissue of a given donor age. Fetal and neonatal tissues with lineages skewed towards early stages will be ideal for stem/progenitor cell therapies, whereas adult livers will be ideal for programs requiring rapid need for late lineage stage functions.

Liver cell therapies for inborn errors of metabolism will be affected by feedback loop signaling, because there will be no selection for the transplanted cells over endogenous cells, necessitating higher numbers of cells to be transplanted. By contrast, patients with liver failure due to virus, drugs, or radiation (involving a loss of feedback loop signals) can be transplanted with smaller numbers of cells given the strong selective pressure for transplanted cells to expand quickly to reconstitute liver mass.

Concerns regarding a need for immunosuppression will be affected by lineage biology. Nonimmunogenic stem/progenitors can acquire immunogenicity with maturation that potentially can be managed by use of stellate cells, known to produce immunomodulatory signals. Liver cell therapies should also use grafting methods that optimize liver engraftment and prevent cell loss to ectopic sites as occurs in vascular route delivery, especially for stem/progenitors.62

Liver Cancer Stem Cells.

The idea that cancers are transformed stem/progenitor cells originated with the pioneering work of Van Potter in the 1960s, who proposed that hepatomas contain cells undergoing “blocked ontogeny.”63 This idea was further elucidated for all types of cancers by Barry Pierce and Stewart Sell,64 who clarified that many functions thought to be related to cancer (e.g., AFP expression) are normal functions of an expanded stem/progenitor cell population and that identification of key distinctions must involve comparison of cancer cells to their normal stem cell counterpart.61, 65 Indeed, normal stem/progenitor cells are strikingly similar to tumor cells in morphology, gene expression, and growth properties, and tumors can be identified as an expanded lineage stage.61, 66 The clinical use of stem cells may come with an increased risk of tumors depending on the donors (e.g., if there are undiagnosed tumor cells among the endogenous stem cells) and on the patient's medical condition (e.g., severe immunosuppression).

Strategies for cancer therapies will be revolutionized if revamped with lineage biology knowledge. Treatments with drugs or radiation are known to affect later lineage stages preferentially. If a specific treatment also targets the lineage stage(s) containing malignantly transformed cells, then the treatment can be curative. If they fail to target that stage, there will be a lethal rebound effect: the treatment kills cells in later lineage stages, mutes feedback loop signaling, and secondarily unhinges early lineage stages where malignant cells reside. Therefore, future cancer therapies should involve strategies identifying the lineage stage of the tumor and whether the treatment targets that stage or, alternatively, uses lineage mechanism regulation, such as feedback loop signals, to control the rate of growth of tumor cells.


The intrahepatic maturational lineages begin within the stem cell compartments, located periportally, and progress through the midacinar region and ending near the central vein. The parenchymal cells, along with their mesenchymal cell partners, are governed by gradients of paracrine signals, including sets of soluble factors and insoluble extracellular matrix components, and by specific mechanical forces. Feedback loop signals regulate the stem/progenitors, controlling liver mass and tissue regeneration. Understanding stem cell and lineage biology in the liver and their regulation offers new considerations for basic and industrial investigations and for more biologically rational clinical program strategies.