The diversity and plasticity of adult hepatic progenitor cells and their niche

Abstract

The liver is a unique organ for homeostasis with regenerative capacities. Hepatocytes possess a remarkable capacity to proliferate upon injury; however, in more severe scenarios liver regeneration is believed to arise from at least one, if not several facultative hepatic progenitor cell (HPC) compartments. Newly-identified pericentral stem/progenitor cells residing around the central vein is responsible for maintaining hepatocyte homeostasis in the uninjured liver. In addition, HPCs have been reported to contribute to liver fibrosis and cancers. What drives liver homeostasis, regeneration and diseases is determined by the physiological and pathological conditions, and especially the HPC niches which influence the fate of HPCs. The HPC niches are special microenvironments consisting of different cell types, releasing growth factors and cytokines and receiving signals, as well as the extracellular matrix (ECM) scaffold. The HPC niches maintain and regulate stem cells to ensure organ homeostasis and regeneration. In recent studies, more evidence has been shown that hepatic cells such as hepatocytes, cholangiocytes or myofibroblasts can be induced to be oval cell-like state through transitions under some circumstance, those transitional cell types as potential liver-resident progenitor cells play important roles in liver pathophysiology. In this review, we describe and update recent advances in the diversity and plasticity of HPC and their niches and discuss evidence supporting their roles in liver homeostasis, regeneration, fibrosis and cancers.

INTRODUCTION

The liver has an impressive regenerative capacity, and regeneration occurs through division of mature hepatocytes and cholangiocytes within the liver, which leave their normal mitotically quiescent state, termed G0, and enter cell cycle and mitosis.¹ During persistent or severe liver injury, such as submassive necrosis, chronic viral hepatitis and non alcoholic fatty liver disease,^2,3 this normally efficient renewal from mature epithelial cells is overwhelmed. In this scenario, adult hepatic stem/progenitor cells, termed hepatic progenitor cells (HPCs) in humans⁴ and oval cells in rodents,⁵ emerge and become activated.^6,7 There cells, with a large nuclear-to–cytoplasm ratio and an oval-shaped nucleus, express markers of both biliary and hepatocyte lineages.⁸ Three-dimensional reconstructions in human liver suggest that HPCs arise from the interface between the hepatocyte canalicular system and the biliary tree, known as the canals of Hering (CoH).⁹ The CoH are partially lined by small hepatocytes and partially by bile duct epithelial cells. They are the physiologic link between the hepatocyte canalicular system and the biliary tree.¹⁰ This anatomical position of HPC accommodates their physiological role following activation, differentiating towards both hepatocellular and biliary type cells to restore function to the damaged liver.⁶ Recently new HPC and their niches have been identified, which are responsible for hepatocyte homeostasis in the uninjured liver.¹¹ In addition, human and mouse hepatocytes can be induced to de-differentiate into ductular cells by ductular metaplasia in chronic injury; these hepatocyte-derived ducts display bipotent liver progenitors which can be differentiated into hepatocytes or cholangiocytes. This appears to involve transition through an “oval cell” like state,¹² and these cell types are considered potential liver-resident progenitors. During liver injury some epithelial cells contribute to fibrogenesis by undergoing epithelial-to-mesenchymal transition (EMT); moreover, it has been shown that certain mesenchymal cells can be reverted into epithelial cells by undergoing mesenchymal-to-epithelial transition (MET), and ultimately differentiate into either hepatocytes or cholangiocytes. Thus, it suggests that multiple cell types modulate the outcome of liver injury.¹³

On the other hand, it has also been reported that HPC may be transformed into cancer stem cells or tumor-initiating cells that drive tumor initiation and thus play a role in the tumorigenesis of hepatocellular carcinomas.^14–16 In addition, some researchers have found co-expression of epithelial and mesenchymal markers in hepatic progenitor cells, demonstrating that HPC have an ability to differentiate towards hepatic stellate cells or myofibroblasts.^17,18 Thus, these in vitro and in vivo results indicate that the fate of HPC depends not only upon which signaling cascades are activated within the HPC, but also on the disease context in which HPC evolve. HPC niches or microenvironments are critical in theses fate selections.

The concept of a stem cell niche was first developed by Raymond Schofield in 1978,¹⁹ who defined it as the microenvironment which regulates stem cell behavior including stem cell maintenance, self-renewal and differentiation. This paper seeks to provide an in-depth review of the literature regarding the role of the diversity and plasticity of HPC and their niche in maintaining the characteristics of HPCs and determining the fate choice of the HPC’s differentiation towards a specific cell type after activation.

The location and composition of HPC niche

In adult livers, the probable progenitor cell niche has been shown to reside in the most proximal biliary structures, the CoH.^9,20 Support for the existence of the HPC niche in the CoH was provided by a lineage-tracing study in an HPC-mediated liver injury model, which showed a continuous supply of HPCs from the Sox9-expressing progenitor zone within the portal area.²¹ In addition to the CoH, other areas of the liver can also transiently provide a niche for HPCs, such as the space of Disse.²² General features of stem cell niches can be found in both the CoH and the space of Disse, but there are some differences between these two sites: One difference is that HPCs are directly exposed to high concentrations of bile acids in the CoH.²³ In addition, the CoH appear to contain slow-cycling precursor cells in the normal liver,^20,24 whereas the space of Disse normally contains quiescent stellate cells and hematopoietic stem cells that only rarely divide.²² Moreover, the type of cell which can release CXCL12 to attract chemokine cysteine-X-cysteine receptor 4 (CXCR4)-expressing cells in these two niches is different; they are liver sinusoidal endothelial cells (LESC) in the space of Disse ²⁵ and cholangiocytes in the CoH.²⁶

However, it has been reported recently that newly-identified hepatic stem/progenitor cells resides around the central vein in normal liver,¹¹ and the endothelial cells at the central vein produce Wnt2 and Wnt9b as short-range signals for these stem/progenitor cells, and constitute a portion of their niche. These pericentral stem/progenitor cells are distinct from aforementioned HSCs, as pericentral stem/progenitor cells maintain hepatocyte homeostasis in the uninjured liver while HSCs have only been reported after injury. Another difference is that pericentral stem/progenitor cells contribute only to the hepatocyte lineage rather than other liver cell types including biliary epithelial cells.¹¹

The HPC niche is defined not only by the site where it is located, but also by the composition of the niche (Figure 1). It is a special microenvironment consisted of different cell types, extracellular matrix (ECM) scaffold, growth factors and cytokines released by the niche cells and signaling pathways that help maintain the characteristics of HPCs and the balance among their activation, transition, proliferation and differentiation^27–27 (Figure 2).

Figure 1. Model of the adult hepatic stem/progenitor cell niche.

(a): In the uninjured liver the Wnt-responsive stem cell population resides around the central vein. These cells self-renew and differentiate into hepatocytes and replace other hepatocytes for liver homeostasis. (b and c): In the injured liver, the canal of Herring (b) and the space of Disse (c) can provide a niches for adult HPCs. These niches are composed of different cell types, ECM components, growth factors and cytokines released by the niche cells and signaling pathways that help to maintain the characteristics of adult HPCs and control activation and proliferation, and govern differentiation fate decisions.

Figure 2. A schematic illustration shows that the activation and transition, proliferation and differentiation of the adult HPCs are determined by the factors and signals received.

Resident adult hepatic/stem cells are activated from their niche after receiving certain factors and signals or transitioned from hepatocytes or myofibroblasts by ductal metaplasia or MET (Left column); and proliferated (Central column); then differentiated towards hepatocytes by the activation of Wnt/β-catenin signaling and the up-regulation of Wnt2 and Wnt9b, differentiated towards cholangiocytes by the activation of Notch signaling pathway, trans-differentiated into hepatic stellate cells by the up-regulation of TGF-β and CTGF, trans-differentiated into myofibroblasts by the up-regulation of Fox1 and TGF-β and the activation of non-canonial Wnt signaling pathway, and transformed into tumor-initiating cells by the up-regulation of TGF-β and Bmil and the activation of Wnt/β-catenin signaling pathway (Right column). Abbreviations: MET, mesenchymal-epithelial transitions; HGF: hepatocyte growth factor; TGF-β: transforming growth factor-β; EGF, epidermal growth factor; aFGF, acidic fibroblast growth factor; TNF, tumor necrosis factor; IL-6: interleukin-6; LT-β: lymphotoxin β; LT-α: lymphotoxin α: IFN-α: interferon α; IFN-γ: interferon γ; TWEAK: TNF-like weak inducer of apoptosis; CTGF: connective tissue growth factor.

Markers of HPC

Accurately identifying HPC in vivo remains a major challenge in understanding HPC biology. To understand the complicated cross-talk between the different cell types and analyze their biological properties in detail, markers of HPC are required to track them at the single-cell level. In the many models of HPC activation it has been shown that numerous markers have been identified and employed (Table 1). However, their unequivocal identification and characterization as a pure fraction is a major obstacle in HPC research; some reported markers are not specific for HPCs and frequently these cells are confused with other cell types expressing the same genes. Therefore, further studies are required to explore unique and more versatile HPC-specific markers and genetic tools to achieve to clonal analyses of HPC.

Table 1.

Hepatic progenitor cell markers and references

Hepatic progenitor cell markers	References
Rodents
Cytokeratin 19 (CK19)	(105, 106)
CK7	(107)
CK14	(106)
γ-Glutamyl transpeptidase (γGT)	(108)
Glutathione-S-transferase P(GST-P)	(109)
OV6	(110)
OV1	(111)
A6	(112)
OC.2	(113)
OC.3	(114)
Connexin 43	(115)
CX3Cl1	(116)
CD24	(116)
MUC1	(116)
Deleted in malignant brain tumour 1	(117)
CD34	(118)
c-kit	(119)
SerpinB3	(120)
Foxl1	(121)
Stromal-derived factor 1 (SDF1)	(122)
BDS7	(123)
Albumin	(123)
CK8	(124)
CK18	(125)
Hepatocyte nuclear factor 4 (HNF4)	(126)
HBD.1	(127)
α-Fetoprotein (AFP)	(78)
Delta-like protein (dlk)	(128)
Aldolase A and C	(129, 130)
c-Met	(131)
Muscle pyruvate kinase (MPK)	(48)
Cadherin 22(PB-cadherin)	(116)
Cadherin 3(P-cadherin)	(116),129
CD44	(132)
Glutathione-S-transferase P (GST-P)	(109)
EpCAM	(133)
Glypican-3	(134)
CXCR4	(134)
Sca-1	(111)
Thy-1(CD90)	(106)
Stromal-derived factor 1 (SDF1)	(135)
CD133	(136)
chromogranin-A	(116)
parathyroid hormone-related peptide	(116)
Prox1	(137)
Liv2	(138)
Dlk	(139)
SEK1	(140)
SMAD 5	(141)
T-Box3(Tbx3)	(27)
Lgr5	(142)
LIF	(143)
claudin-7	(116)
claudin-22	(116)
ros-1	(116)
Gabrp	(116)
Podoplanin	(144)
Humans
OV6	(145)
EpCAM	(102)
CD133	(131)
ALB	(146)
AFP	(146)
CK19	(9, 102)
ICAM1	(9)
CD44h	(102)
CD44	(146)
CK8	(102)
CK18	(102)
c-kit	(102)
Dlk-1	(9)
Prox1	(145)
CD34	(102, 137)
CD90	(102)
c-met	(102)
SerpinB3	(120)
GCTM-5	(147)
α1-Antitrypsin	(123)
NCAM	(148)

The neighboring cells with HPCs in the niche

Potentially all cell types in the liver can interact with HPCs including cholangiocytes, hepatocytes, stellate cells, myofibroblasts, Kupffer cells, LSECs, and cells from the immune system such as macrophages and lymphocytes. These cells are closely associated with HPCs and have a potential ability to influence the survival, proliferation, migration and differentiation of HPCs not only through secreting growth factors, chemokines, and cytokines,^22,30,31 but also by direct cell-cell contracts.³²

The epithelial cell types in liver, which incur the most damage, can influence the activation and the fate of HPCs to differentiate into hepatocytes and cholangiocytes. It has been demonstrated that in humans a threshold of 50% loss of hepatocytes is required for extensive HPC activation.³³ In severe viral hepatitis and acute liver failure, the hepatocyte is the most damaged cell type, consequently HPCs differentiate into hepatocytes.³⁴ In primary biliary cirrhosis (PBC) and primary sclerosing cholangitis, the cholangiocyte is the primary damaged cell type, and HPCs differentiate towards biliary cells,³⁵ though prior to cholangiocyte injury in the ducts, the CoH actually disappear, due to unknown causes. Thus, the loss of the CoH niche may be a primary event in the development of PBC.

In the uninjured liver, adjacent central vein endothelial-cell-derived Wnt signals (Wnt2 and Wnt9b) are necessary for maintaining the precursor state and the high proliferative state of pericentral stem/progenitor cells, and it appears that glutamine synthetase as a molecule important to niche maintenance in rodents and in humans.¹¹

It has been shown that HPC expansion requires a close cooperation with accompanying hepatic stellate cells (HSCs) and myofibroblasts, which can produce hepatocyte growth factor (HGF), TGF-β, TGF-α, epidermal growth factor (EGF) and acidic fibroblast growth factor (aFGF).^{25,28,29,36,37} HSCs express signaling pathways required for HPC function such as hedgehog,³⁸ Wnt/β-catenin³⁹ and Notch signaling.²⁵ Moreover, HSCs express the HPC marker CD133 and exhibit the capacity to differentiate into hepatocyte-like cells in vitro.⁴⁰ Pintilie et al. have demonstrated that the inhibition of HSC activation after 2-AAF/partial hepatectomy (PHx) not only reduces HPC proliferation, but also affects HPC differentiation.⁴¹

Further invesitgations suggested that inflammatory cells played a central role in the HPC response, as multiple inflammatory cytokines were responsible for activating the HPC response.⁴² Lim, et al. found that IFN-α based treatment can reduce the number of HPCs in the choline-deficient ethionine-supplemented (CDE) diet model of chronic liver injury by modulating apoptosis, proliferation, and differentiation of HPCs.⁴³ TNF-like weak inducer of apoptosis (TWEAK), a member of the pro-inflammatory TNF family, can be secreted by NK cells and macrophages, which stimulates the proliferation of HPCs through the activation of fibroblast growth factor-inducible 14 (Fn14) and nuclear factor kappa B (NF-κB).⁴⁴ Nagyd et al. demonstrated that dexamethasone, a well-known inhibitor of TNF and IL-6, effectively inhibited HPC proliferation and liver regeneration after PHx.⁴⁵ Knight et al. demonstrated that TNF signaling participated in the proliferation of HPCs during the preneoplastic phase of liver carcinogenesis, and inhibition of this signaling reduced the incidence of tumor formation.⁴⁶ It has been reported that T lymphocytes and NK cells produced pro-inflammatory cytokines including IFN-γ and TNF-α, thereby promoting the expansion of HPCs in the CDE treatment model.⁴⁷ LT-β, LT-β receptor, and IFN-γ are involved in HPC-mediated liver regeneration, and the absence of these cytokines impairs the HPC-dependent regenerative response.⁴⁸ It has also been shown that HPC proliferation was reduced in TNF and LT-α double knockout mice.⁴⁹ Yeoh et al, demonstrated that IL-6 administration resulted in increased migration and proliferation of HPC in vivo, and hyperactive STAT-3 signaling enhanced HPC numbers whereas ERK1/2 activation suppressed HPC proliferation.⁵⁰ Inflammatory cells can also remodel the ECM through the production of metalloproteinases, a process that may be necessary for the extension of the HPC response.⁵¹ Thus there are a series of cell types which can interact and cross-talk with HPCs, thereby influencing the proliferative and differentiation capacities of HPC within the niche.

The extracellular matrix (ECM) in the HPC niche

The ECM is a structural network that consists of different types of collagens (types I, III, IV, V and XVIII),⁵² proteoglycans (heparan, dermatan, chondroitin sulphate, perlecan, hyaluronic acid, biglycan, and decorin)⁵³ and glycoproteins (laminin, fibronectin, tenascin, nidogen, and SPARC).⁵² The ECM plays a vital role in the determination, differentiation, proliferation, and survival of HPCs by binding, integrating, and presenting growth factor signals to cells.⁵⁴ A study using a CDE-induced murine model revealed that HPCs required ECM deposition and activation of matrix-producing cells, HSCs or myfibroblasts, for migration and differentiation for the repopulation of the damaged liver.⁵⁵ They also demonstrated that new ECM deposition was an initial phase, prior to HPC expansion, along the porto-veinous gradient of lobular invasion, and that the ECM deposition played a fundamental role in establishing HPC niches. An in vitro study has demonstrated that HPCs can proliferate and maintain an undifferentiated phenotype when grown on laminin, but their proliferation was inhibited on collagen I matrix.⁵⁶ It is intriguing to note that unlike any other epithelial lined structure in the body, the CoH-parenchymal interface has a discontinuous basement membrane: the biliary cells of the CoH having laminin-rich basement membrane while the adjoining hepatocytes only have loosely aggregated type 4 collagen fibers in their space of Dissse which may play a role in niche functionality.⁵⁷ Moreover, an in vivo study has shown that a failure of ECM remodeling, i.e., fibrosis resolution and laminin deposition, hinders the ability of the liver to activate HPCs. In additions, laminin-HPC interactions within the HPC niche are critical for HPC activation and expansion.⁵⁸ Future studies are necessary to better define the interactive cross-talk among HPCs, HSCs, myofibroblasts, and the ECM.

Signaling pathways in HPC and its niche

Determining which signals are responsible for such processes as induction of differentiation, self-renewal and/or maintenance of stemness will lead to a better understanding of the biology of HPCs and their role in liver regeneration. In the HPC niches, signaling pathways such as Wnt, Notch, and Hedgehog are key regulators of HPCs. β-catenin is an essential component of the canonical Wnt signaling, which is clearly involved in the HPC response observed in rodents and human. A noteworthy increase in total and active β-catenin was observed at the time of extensive oval cell activation and proliferation in the 2-AAF/PHx model. And a dramatic decrease of the A6-positive HPC numbers in the β-catenin conditional knockout mouse confirmed this role.⁵⁹ The data from Yang et al. not only supported such a role for Wnt/β-catenin signaling in activation and proliferation of the HPCs in the liver, but also demonstrated that a considerable proportion of human HCCs appear to arise from HPCs, and activation of the Wnt/β-catenin can be markedly enriched in the HPC positive HCC cells.⁶⁰ 32 children with biopsy-proven non-alcoholic fatty liver disease (NAFLD) were studied to evaluate the role of Wnt3a production by macrophages in HPC response in the progression of pediatric non-alcoholic fatty liver disease by Eugenio Gaudio’s team, they found that after pediatric patients were treated with docosahexaenoic acid, the macrophage polarization was determined to an anti-inflammatory phenotype with the upregulation of macrophage Wnt3a expression, resulting in reduction the degree of NAFLD and portal fibrosis through β-catenin phosphorylation in HPC and commitment towards hepatocyte fate.⁶¹

Notch signaling has also been shown to play a role in the differentiation and proliferation of oval cells in the rat model of 2AAF/PHx.⁶² Wang et al. found that Notch 3 is not only important in the generation of hepatocytes from fetal hepatic progenitor cells (FHPCs), but it also appears to be a potential marker of FHPCs.⁶³ Also, Hedgehog (Hh) signaling is essential for survival of HPCs, as evidenced by an inhibitor of Hedgehog signaling, cyclopamine, which can induce HPC death.⁶⁴ Another researches investigated the regulation of adult liver repair and regeneration, and liver diseases by Hh pathway,^65–67 Hh pathway activation was accompanied by increased HPCs after PHx, and after treatment of mice with cyclopamine, a specific Smo antagonist, Hh signaling was inhibited and the expressions of various HPC markers were also reduced, indicating that Hh signaling is critical for normal liver regeneration.⁶⁵ Leptin-deficient ob/ob mice with fatty livers and their healthy lean liver littermates were studied to determined if Hh signaling regulated HPCs and epithelial-mesenchymal transitions (EMT) during tissue remodeling before and after exposure to ethionine,⁶⁶ the results showed that hepatic accumulation of HPCs which produced and responded to Hh ligands were increased in the ob/ob mice after treatment with ethionine, which resulted in increased Hh activity and liver fibrosis through EMT. Michelotti G et al, also demonstrated that Hh signaling was disrupted after Smoothened (SMO), an obligate intermediate in the Hh pathway, was conditionally deleted in α-SMA expressing cells in BDL mice, then inhibited fibrosis and caused liver atrophy by accumulation of myofibroblasts and HPCs, these all identify SMO was a master regulator of adult liver repair.⁶⁷ In addition, it has been shown that the proliferation and self-renewal of human HPCs was promoted by the Bmi 1 pathway.⁶⁸ And the clonogenicity of HPCs was significantly improved in culture medium containing the Rho-associated kinase inhibitor Y-27632.⁶⁹ There is as yet no consensus as to which of these pathways is most critical in the function of the HPCs as well as their specific contributions to the HPC niche.

During liver development, epithelial cell adhesion molecule (EpCAM), a transmembrane glycoprotein, demonstrates a dynamic expression in fetal liver, including cells of the parenchyma, whereas mature hepatocytes are devoid of EpCAM. It has been demonstrated that liver regeneration is associated with a population of EpCAM-positive cells within ductular reactions, which gradually lose the expression of EpCAM along with maturation into hepatocytes.^70,71 EpCAM can be switched on and off through a diverse panel of strategies to fine-tune EpCAM-dependent functional and differentiative traits. HPCs highly express EpCAM, and EpCAM is switched off in the derivatives of HPC. Thus, the molecular switches involved in the processes of HPC maintenance and differentiation into mature liver cells is of great consequence, but currently understudied. Based on the strong expression in HPCs and the dynamic expression pattern of EpCAM during liver cell differentiation and maturation, this molecule might represent a central target in the field of hepatocellular differentiation and liver regeneration.⁷²

Experimental models of HPC activation

Since the first observation of HPCs as small, ovoid-shape cells reported to appear in the liver of rats, a series of experimental models of HPC-mediated liver regeneration in rodents have been employed.⁷³ In rats, a precursor-product relationship exists between oval cells and hepatocytes with N-2-acetylaminofluorene (2-AAF) treatment. Oval cell activation, proliferation and differentiation resulting in the liver regeneration has been studied in a series of models, including the combination of 2-AAF with 70% hepatectomy (2-AAF/PHx),⁷⁴ 2-AAF treatment combined with exposure to carbon tetrachloride (2-AAF/CCl4),⁷⁵ 2-AAF treatment combined with allyl alcohol injury (2-AAF/AA),⁷⁵ galactosamine (GalN) treatment,⁷⁶ a choline-deficient, ethionine-supplemented (CDE) diet,⁷⁷ and long-term exposure to ethanol.⁷⁸ Recently, several models of HPC activation have been developed for mice, that are more genetically tractable than rats.⁷⁹ For example, 1,4-bis[N,No-di-(ethylene)-phosphamide] piperazine (Dipin) treatment combined with 70% hepatectomy (Dipin/PHx),⁸⁰ the CDE diet,⁷⁹ galactosamine injury (GalN),⁸¹ allyl alcohol injury (AA),⁸² carbon tetrachloride (CCl₄) injury,⁸³ acetaminophen (APAP) injury,⁸⁴ a 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)-containing diet,⁸⁵ and long-term exposure to ethanol have been employed.³ These approaches have been shown to be highly effective for the induction of hepatic oval cells capable of differentiating into hepatocytes and/or biliary epithelial cells, without causing high mortality. Of these, only APAP is a significant toxin in humans; indeed it is the leading cause for transplant for acute liver failure suggesting it should be more widely used as correlate to human disease. In humans, HPC activation is not observed after acute and complete extra-hepatic biliary obstruction, but HPC activation has been described in various conditions such as severe acute hepatitis due to viral infection, toxin (e.g. acetaminophen) or autoimmune hepatitis,¹⁰ chronic hepatitis B and C, primary biliary cirrhosis,⁸⁶ alcoholic liver disease, chronic hepatitis B,⁸⁷ fulminant hepatic failure, focal nodular hyperplasia, primary sclerosing cholangitis,⁸⁸ and pediatric nonalcoholic fatty liver disease.⁸⁹

The HPCs niche influences the fate of HPC differentiation

A handful of studies in recent years have shed light on factors and signals from the HPC niche that determine the fate choice of HPC differentiation to a specific cell type depending on the type and severity of the liver disease and resultant microenvironment created by it.⁹⁰ HPCs are very different from tissue-resident progenitor cells in other epithelial tissues such as intestine and skin. They are not required to replenish liver tissue under normal, healthy conditions. Under the condition of liver parenchymal cells loss and impaired liver parenchymal cell proliferation, HPCs can differentiate into hepatocytes and cholangiocytes to restore parenchymal architecture and liver function.⁹¹ It has been shown that there can be robust hepatocyte repopulation by HPC in late stage human liver disease.^70,92 Boulter et al. showed that Notch and Wnt signaling direct HPC specification within the activated myofibroblast and macrophage HPC niche in liver disease patients and murine models of the ductular reaction with biliary and hepatocyte regeneration. During adult biliary regeneration there is a requirement for an activated Notch signaling pathway in order to specify biliary epithelium from HPCs, while ectopic activation of the canonical Wnt pathway commits HPCs to the hepatocyte fate. Tanimizu et al. demonstrated that HPCs could differentiate into either hepatocytes or cholangiocytes by overlaying an Engelbreth-Holm-Swarm sarcoma (EHS) gel or by embedding them in a type I collagen gel, respectively.⁹³ Inhibition of ECM accumulation resulted in an increased number of HPC-derived hepatocytes, suggesting that the ECM is also involved in the fate determination of HPCs.⁶ Wang et al. revealed that the fundamental mechanisms regulating liver renewal are similar to other organs in which homeostatic renewal involves small populations of stem/progenitor cells that maintain the tissue,¹¹ and that these stem/progenitor cells only contribute to the hepatocyte lineage. In a recent study, Theise, ND et al. showed that the niche activities are sometimes simply responsible for the re-establishment of the hepato-biliary link. ⁵⁷

Some studies have also suggested that deregulated HPCs might be a potential source of HCCs as well as cholangiocarcinomas (CC).^94,95 As early as 2002, Dumble et al. demonstrated direct evidence for the involvement of HPCs in the pathogenesis of HCC. When HPCs, isolated from p53 null mice, were transplanted into athymic nude mice, they produced HCCs.⁹⁶ Chiba et al. showed that the transplantation of Bmi1- or β-catenin-transduced cells clonally expanded from a single HPC produced tumors, which exhibited the histologic features of combined hepatocellular and cholangiocarcinoma.⁶⁸ Wu et al revealed that TGF-β induced the neoplastic transformation of adult HSCs to hepatic TICs and facilitated HCCs.⁹⁷ These observations imply that the dysregulated self-renewal of HPCs serves as an early event in hepatocarcinogenesis, and they also highlight the important roles of Bmi1 and the Wnt/β-catenin pathway in regulating the self-renewal of normal or cancer stem cells in the liver. Our recent studies indicated that HPCs were the origin of a liver cancer stem cell formed by the fusion with hematopoietic precursor-derived myeloid intermediates,¹⁶ and these liver cancer stem cells could initiate three types of human liver carcinomas (HCC, CC and the combined HCC and CC (CHC)).⁹⁸

Recent studies have clearly demonstrated that there is an association between the severity of fibrosis and progenitor cell activation.⁹⁹ And some investigations have suggested that HPCs may not only make up for loss of hepatocytes but may also contribute to or promote periportal fibrogensis. For example, in the 2-AAF/PHx rat model, connective tissue growth factor (CTGF) was up-regulated, and co-localized with HPCs and hepatic stellate cells.¹⁰⁰ An HPC line, HPPL, derived from long-term culture of mouse Dlk+ hepatoblasts on laminin, can produce ECM proteins, including laminin-α5,-β1,-β2,-γ1, and -γ2 subunits.⁹³ Some investigators found co-expression of epithelial and mesenchymal markers in hepatic progenitor cells, demonstrating that HPCs have an ability to differentiate towards hepatic stellate cells or myofibroblasts.^17,101 Noemi et al. showed that expression profiling studies have identified a subpopulation of HPCs that express markers consistent with a mixed epithelial/ mesenchymal phenotype including members of the forkhead winged helix transcription factors family, Foxl 1.¹⁰² Yovchev et al. found that HPCs express a number of mesenchymal markers including vimentin, mesothelin, CD44, bone morphogenetic protein 7, and Tweak receptor (tumor necrosis factor receptor superfamily, member 12A).¹⁷ Wang et a.l demonstrated that freshly cultured HPCs can be induced by TGF-β1 to differentiate into mesenchymal cells, resembling hepatic stellate cells, through an EMT process.¹⁰³ Dan et al. reported that multipotent progenitor cells isolated from human fetal liver were capable of differentiating into liver and mesenchymal lineages.¹⁰² Our recent study has shown that HPCs differentiated into α-SMA positive myofibroblasts via activation of the non-canonical Wnt pathway and exhibited a profibrotic effect in in the 2-AAF/CCl4 rat model.¹⁸ It also has been shown that HPCs can differentiate into pancreatic cells in the culture medium which was depleted with dexamethasome and supplemented with retinoic acid.⁹³

Taken together, these reports suggest that HPCs are highly plastic cells which are capable of differentiating into hepatocytes, cholangiocytes, hepatic stellate cells, myofibroblasts and pancreatic cells, and can contribute to liver cancers depending on the HPC niche and the signals received.

CONCLUSION

Throughout life, the liver is bombarded with chemical, physiological and pathological insults which require the activation and appropriate differentiation of HPCs. To maintain liver homeostasis or restore the function of the damaged liver, the HPCs must differentiate into hepatocytes, the key metabolic cells of the liver, or/ and cholangiocytes, which line the biliary tree and transport bile into the intestine. In addition to resident hepatic progenitor cells, those transitional cell types as potential resident hepatic progenitor cells play important roles in liver regeneration after liver injury. However, on the other hand, HPCs contribute to liver fibrosis and cancers. The HPC niche regulates the maintenance of HPC quiescence, controls the activation and proliferation of HPCs, and governs fate decisions of HPC differentiation. Thus, the identification and elucidation of HPC niche composition, how it changes with liver injury and regeneration, as well as the interplay of its components in the normal and regenerating liver are pivotal for developing therapeutic strategies to enhance liver regeneration for a wide variety of human liver diseases (Table 2). In summary, the diversity and plasticity of HPCs and their niche interpret liver homeostasis, regeneration, fibrosis and cancers under physiological and pathological conditions (Figure 3).

Table 2.

Potential approaches to treat human liver diseases through modulation of adult hepatic progenitor cells and their niche

Inhibit the transdifferentiation of HPC into myofibroblasts and/or hepatic stellate cells Small molecules with anti-TGF-β activity (102,103) The non-canonical Wnt pathway inhibitors (18,59,60) Small molecules with anti-Foxl 1 activity (102)

Inhibit the transformation of HPCs into tumor-initiating cells Small molecules with anti-TGF-β activity (97) Small molecules with anti-Bmi1 activity (68) Wnt/β-catenin pathway inhibitors (60,68)

Stimulate the degradation of accumulated extracellular matrix in liver fibrosis* Tissue inhibitor of metalloproteinase I (TIMP1) inhibitors (149) Matrix metalloproteinase gene therapy (150) Relaxin (150)

HPC-based approaches promote hepatic regeneration Adjustment of the HPC microenvironment and optimization of the differentiation protocols to promote restoration of liver function (70,92,104,105)

Affect HPCs by modifying macrophage phenotype in vivo Promotion of hepatic regeneration by inducing an anti-inflammatory phenotype of macrophages and inhibiting the release of inflammatory factors (61)

^*: Inhibition of extracellular matrix accumulation could increase the number of HPC-derived hepatocytes.

Figure 3. A schematic illustration of the involvement of adult hepatic progenitor cells in liver disease.

Resident or transitional hepatic progenitor cells (HPC) will be differentiated towards hepatocytes or cholangiocytes under liver injury conditions such as in acute/chronic viral hepatitis, non-alcoholic and alcoholic steatohepatitis, or in primary biliary cholangitis, primary sclerosing cholangitis. Under some circumstance, HPCs can be trans-differentiated into myofibroblasts and/or hepatic stellate cells to contribute to the development and progression of liver fibrosis, and can be transformed into tumor initiating cells to initiate and sustain hepatocellular carcinoma, or cholangiocarcinoma, or both.

Since some investigations have shown that transplanted HPCs have the ability to differentiate into hepatocytes in vivo,^104,105 HPC-based transplantation represents a therapeutic option which provides significant advantages compared to organ or primary hepatocyte transplantation. Thus, by adjusting the HPC microenvironment and optimizing the differentiation protocols, there is the potential to theoretically restore liver function as a clinical application in the future (Table 2).

Key Points.

• HPC are responsible for liver homeostasis and regeneration. Under some circumstance, hepatic cells can be induced to function as potential liver-resident progenitors.

• The canals of Hering, the space of Disse are and the central vein and their endothelial cells are defined with features of stem/progenitor cell niches.

• The fate of HPC is determined by the interactions among extrinsic and intrinsic factors within HPC niche.

• Extrinsic factors include different cell types, releasing growth factors, cytokines and receiving signals, and ECM scaffold in HPC niche.

• Intrinsic factors mainly consist of activated signaling pathways in HPC, including Wnt, Notch, Hedgehog, and EpCAM pathways.

Abbreviations

HPC

hepatic progenitor cells

ECM

extracellular matrix

EMT

epithelial to mesenchymal transition

MET

mesenchymal to epithelial transition

LSEC

liver sinusoidal endothelial cells

PBC

primary biliary cholangitis

CoH

canals of Hering

HSC

hepatic stellate cells

HGF

hepatocyte growth factor

EGF

epidermal growth factor

aFGF

acidic fibroblast growth factor

2-AAF

N-2-acetylaminofluorene

PHx

partial hepatectomy

EpCAM

epithelial cell adhesion molecule

HCC

hepatocellular carcinomas

CC

cholangiocarcinomas

TIC

tumor-initiating cells