Stem cells of the adult liver

Unlike rapidly renewing epithelial tissues (such as the intestinal mucosa or epidermis), in which an active stem cell lineage system continually initiates replacement of differentiated cells that are shed (4,5), the liver is normally a quiescent organ with minimal or slow rates of cell turnover in the adult (51). Nonetheless, the liver possesses an extraordinary capacity for the regeneration of tissue mass following loss of normal hepatocyte numbers because of partial tissue loss (surgical resection) or hepatotoxic injury (necrosis).

A number of different cell types can be activated to repair or regenerate the liver depending on the nature and extent of injury or tissue deficit (15). In an otherwise healthy liver, the replacement of hepatocytes (and tissue mass) lost to surgical resection or toxic injury is achieved through the proliferation of differentiated, normally quiescent hepatocytes contained in the residual (viable) tissue (Fig. 1). However, certain forms of liver injury impair the capacity of the remaining differentiated hepatocytes to proliferate in response to liver tissue deficit. When this occurs, a reserve or facultative stem cell compartment is activated to proliferate and replace the lost hepato-cytes. Evidence suggests that there are at least two distinct cell populations that can be activated to generate new hepatocytes or cholangiocytes (Fig. 1). Rodent livers contain a population of normally quiescent (facultative), undifferentiated stem cells that reside in or around the biliary ductules of the portal tracts, which can be activated under certain pathological conditions to reestablish a proliferating-differentiating lineage (the oval cell reaction), capable of generating hepatocytes and some other cell types (52,53). In addition, the adult liver contains a population of incompletely differentiated small hepatocytelike progenitor cells (SHPC) that can be activated to replace lost hepatocytes in some forms of tissue injury (54).

Fig. 1. Cellular responses to signals for liver regeneration. (A) Liver regeneration after surgical partial hepatectomy involves proliferation of differentiated hepatocytes. However, when differentiated hepatocytes cannot proliferate in response to the regenerative stimulus (such as in retrorsine-exposed rats), other liver progenitor cell populations are activated, such as (B) small hepatocyte progenitor cells or (C) oval cells.

Fig. 1. Cellular responses to signals for liver regeneration. (A) Liver regeneration after surgical partial hepatectomy involves proliferation of differentiated hepatocytes. However, when differentiated hepatocytes cannot proliferate in response to the regenerative stimulus (such as in retrorsine-exposed rats), other liver progenitor cell populations are activated, such as (B) small hepatocyte progenitor cells or (C) oval cells.

These observations combine to suggest that there are at least three distinct populations of cells with stemlike potential in the adult rat liver (Fig. 1). It is conceivable that all these stem cells of the adult liver are derived from the same primordial stem cell population of the developing liver. In the following sections, the evidence for the existence of each of these stem cell populations in liver is reviewed.

4.1. Unipotential Liver Progenitor Cells: Differentiated Hepatocytes and Biliary Epithelial Cells

The liver possesses an enormous capacity to replace cells lost to surgical resection or necrosis (55,56). Activation of undifferentiated stem cells does not occur after cell loss when mature hepatocytes and biliary epithelial cells are capable of proliferating to restore the normal liver mass and structure (57,58). In rats subjected to surgical partial hepatectomy, the residual (viable) hepatocytes undergo a rapid burst of proliferation that ultimately restores the normal hepatocyte number (51,59-62). Likewise, biliary epithelial cells proliferate after partial hepatectomy to form expansions of the intra-hepatic duct system (51,62-64). Irrespective of their location in the parenchyma (periportal to pericentral), virtually all hepatocytes proliferate and divide at least once during restoration of the hepatocyte number (65,66).

The ability of quiescent hepatocytes to reenter the cell cycle and proliferate in response to liver deficit has fascinated investigators throughout history. More recently, the extensive growth potential and enormous proliferative (replicative) capacity of the mature hepatocyte has become evident. In rats, hepatocytes proliferated and divided at least 8 to 12 times during the prolonged process of liver growth following five consecutive partial hepatectomies (67). In transgenic mice that express the urokinase gene in the liver under the direction of the albumin promoter-enhancer, the majority of hepatocytes succumbed to the toxic transgene product (68). In this model, the toxic transgene became inactivated in random hepato-cytes, enabling them to proliferate, undergoing 10-12 cycles of cell division to yield discrete nodular aggregates (clones) that repopulated the liver parenchyma (69).

In a similar experimental system, transplanted normal hepatocytes repopulated the livers of transgenic mice that lack fumarylacetoacetate hydrolase (FAH-/ ) enzyme activity (70) because of the targeted disruption of exon 5 of the Fah gene (71). In this model, the transplanted FAH-expressing hepatocytes exhibited a selective growth advantage over host FAH-deficient hepatocytes, allowing the former to repopulate the livers of mutant mice. In these studies, it was estimated that transplanted hepatocytes proliferated through at least 15 cell divisions during repopulation of mutant livers (70).

In other studies, wild-type male hepatocytes were serially transplanted at limiting dilution through the livers of female FAH/- mice (72). Complete repopulation of the diseased liver was accomplished in each round. The complete replacement of host liver by the progeny of transplanted hepatocytes through seven rounds of transplantation suggests that the transplanted hepatocytes were capable of at least 100 population doublings (72).

Together, these studies demonstrate the incredible capacity for cell proliferation by differentiated hepatocytes, consistent with the suggestion that these cells represent a unipotential progenitor cell population of the adult liver.

4.2. Unipotential Liver Progenitor Cells: Small Hepatocyte Progenitor Cells

In several experimental models, hepatocytes were rendered incapable of proliferation through treatment with mito-inhibitory compounds, facilitating the outgrowth of stem cells in response to liver deficit. We recently described the cellular responses and time course for liver regeneration after surgical partial hepatectomy (PH) in rats with retrorsine-induced hepatocellular injury (54). Similar to other models of chemical liver injury (15,21), systemic exposure to retrorsine, a member of the pyrrolizidine alkaloid (PA) family, resulted in severe inhibition of the replicative capacity of fully differentiated hepatocytes (54,73-75). When confronted with a strong proliferative stimulus such as PH (54,73,74) or hepatocellular necrosis (76), retrorsine-injured hepatocytes synthesized deoxyribonucleic acid (DNA), but were unable to complete mitosis and arrested as nonproliferative giant cells (megalocytes). In this model, neither retrorsine-injured, fully differentiated hepatocytes nor oval cells proliferated sufficiently to contribute significantly to the restoration of liver mass after PH. Instead, the entire liver mass was reconstituted after PH through a novel cellular response that was mediated by the emergence and rapid expansion of a population of SHPCs, which share some phenotypic traits with fetal hepatoblasts, oval cells, and fully differentiated hepatocytes, but are morphologically and phenotypically distinct from each (54). SHPCs emerged in all regions of the liver lobule after PH and were not solely associated with modest oval cell outgrowth in periportal regions, suggesting that SHPCs represent a novel cell population (54).

The SHPCs morphologically most closely resemble differentiated (but small) hepatocytes at early time points after PH, perhaps suggesting that

SHPCs are a subset of retrorsine-resistant hepatocytes and not a novel progenitor cell population. However, the phenotype of SHPCs indicates that they are in fact distinct from fully differentiated hepatocytes because a subset of SHPCs expresses the oval cell/bile duct/fetal liver markers OC.2 and OC.5 through 5 d post-PH (54). Coexpression of hepatocyte markers and oval cell markers by early-appearing SHPCs suggests that these cells are not fully differentiated, and that they display a phenotype similar to that expected for a cell type transitional between the bipotential hepatoblast (E14) and a fetal hepatocyte (E18-E20).

Retrorsine-exposed rats were able to regenerate their liver mass completely after PH, as evidenced by liver weights and liver:body weight ratios (54). At 30 d post-PH, liver weights and liver:body weight ratios do not differ significantly after either retrorsine/PH or control/PH (54). By this time, the progeny of SHPCs occupied virtually the entire (87% by area) parenchyma in retrorsine/PH rats. However, comparison of the time course for liver regeneration in control and retrorsine-exposed rats after partial hepatectomy showed that liver regeneration through activation and expansion of SHPCs is a much more protracted process. Complete regeneration of the liver mass in retrorsine-exposed animals required nearly 30 d, compared to about 10 d in control rats (54).

Using a combined approach involving gene expression analysis of tissues isolated using laser capture microdissection and in situ immunohistochem-istry, the expression patterns of select mRNAs and proteins were examined in the earliest (least-differentiated) SHPCs that emerged after PH in retrorsine-exposed rat livers (77). The results showed that early-appearing SHPCs (at 3-7 d post-PH) expressed messenger RNA (mRNA) or protein for all of the major liver-enriched transcription factors (hepatic nuclear factor 1a [HNFla], HNF1p, HNF3a, HNF3p, HNF3y, HNF4, HNF6, C/EBPa, C/ EBPp, and C/EBPy), WT1, a-fetoprotein, and P-glycoprotein (77).

Compared to surrounding hepatocytes, early-appearing SHPCs lack (or have significantly reduced) expression of mRNA for hepatocyte differentiation markers tyrosine aminotransferase and a1-antitrypsin (77). Likewise, SHPCs that emerge and proliferate during the early phase of liver regeneration lack (or have reduced expression of) several hepatic CYP (cytochrome P450) proteins known to be induced in rat livers after retrorsine exposure (CYP2E1, CYP1A2, and CYP3A1). However, by 30 d post-PH, expression patterns for all markers expressed by SHPCs mirrored that expected for fully differentiated hepatocytes. Both a-fetoprotein and WT1 protein are uniquely expressed by SHPCs during the early phase of liver regeneration, suggesting that these markers may be used to identify the earliest progenitors of these cells (77). These results suggest that SHPCs represent a unique parenchymal (less differentiated than mature hepatocytes) progenitor cell population of adult rodent liver (54,77).

4.3. Multipotential Liver Progenitor Cells: Oval Cells

Oval cells, which proliferate in several hepatocarcinogenesis models (78-80) and in some forms of noncarcinogenic liver damage (81-85), may be related to liver stem cells. A number of different experimental models elicit the proliferation of oval cells (52,81,86-89). All of these models are characterized by concurrent stimulation of liver growth and inhibition of normal mechanisms for liver tissue restoration (i.e., blockade of the proliferation of hepatocytes). The stimulus for liver growth can be satisfied through several different methods, including surgical resection, nutritional stress, or chemically induced necrosis. Blockade of hepatocyte proliferation is frequently achieved using chemicals (such as 2-acetylaminofluorene) that impede or prevent mitotic division of mature hepatocytes (52). The cellular response common to each of these models involves the proliferation of small cells with scant cytoplasm and ovoid nuclei that are morphologically described as oval cells (90). Although most of the models of oval cell proliferation involve rats, similar models have been developed using carcinogen-treated mice (91-94), transgenic mice that express viral oncogenes (95,96), or other transgenes (97).

The timing of cellular events differs, sometimes dramatically, among the various models of oval cell proliferation (15). However, the majority of oval cell proliferation models share a common sequence of events: (1) proliferation of oval cells in or around the portal spaces, (2) invasion of the lobular parenchyma by the proliferating oval cells, (3) appearance of transitional cell types and immature hepatocytes, and (4) maturation of hepatocytes and restoration of normal liver structure. Oval cells are initially seen in the portal zones of the liver lobule in the regions of terminal bile ductules or cholangioles (29,98). Proliferating oval cells are recognized as representing a collection of phenotypically distinct cells that compose a heterogeneous cell population or compartment (32,99,100). Morphologically, the typical oval cell possesses cellular characteristics similar to those of cells of terminal bile ductules (101-103). However, the oval cell compartment also contains transitional cells that display morphologic features intermediate between oval cells and hepatocytes (86,103).

Proliferating oval cells form irregular ductlike structures connected to preexisting bile ducts (103-107). As they proliferate, oval cells migrate from the portal regions into the lobular parenchyma, sometimes occupying a large percentage of the liver mass. Groups of small basophilic hepatocytes appear among oval cells; these immature hepatocytes proliferate and differentiate as the oval cells gradually disappear and the normal liver structure is restored (83,93,103,108-110). The possibility that oval cells might possess stemlike properties and give rise to hepatocytes or biliary epithelial cells has been recognized for some time (26,28,90,111).

Several studies have attempted to document the fate of oval cells that proliferated in various hepatocarcinogenesis models and after noncarcino-genic liver injury, producing evidence that oval cells are precursors of hepatocytes (83,108-110,112,113). Using the modified Solt-Farber model of oval cell proliferation, Evarts et al. (109) demonstrated unequivocally that oval cells radiolabeled with 3H-thymidine could give rise directly to tagged basophilic hepatocytes. More recently, Alison and colleagues examined the proliferation and fate of oval cells in rats using the modified Solt-Farber model with various doses of 2-acetylaminofluorene (114).

Proliferation of oval cells also has been described in the chronic injury produced in mouse liver by transgenic expression of both hepatitis B virus (95) and SV40 T antigen (96). Bennoun and colleagues (96) demonstrated a transition between proliferating oval cells in SV40 T antigen transgenic mice and newly formed hepatocytes. Likewise, in mice treated with diethylnitro-samine, oval cells proliferated and subsequently differentiated into hepatocytes (94).

Employing the d-galactosamine model of oval cell proliferation in rats, Lemire et al. (110) and Dabeva and Shafritz (83) also demonstrated the transfer of radiolabel from oval cells to small hepatocytes. In both of these studies, transition from oval cells to hepatocytes was accompanied by a shift from the biliary epithelial/oval cell phenotype (expression of a-fetoprotein, Y-glutamyltranspeptidase, and biliary epithelial-type cytokeratins) to a cellular phenotype characteristic of hepatocyte differentiation (expression of albumin, glucose-6-phosphatase, and other hepatocyte markers; reduction of a-fetoprotein expression) (83,110).

4.4. Liver Progenitor Cells From Extrahepatic Tissues

In addition to the progress characterizing stem cell responses in liver and the various populations of liver cells with stemlike potential, advances have also been made in the identification of multipotent adult stem cells with broad differentiation potential that includes liver. A number of studies have identified extrahepatic sources of stemlike cells that can colonize the liver or give rise to hepatocytes in vivo or in culture. Most recently, several investigators have reported that progenitor cells in bone marrow or periph-

eral blood can give rise to cells of the liver. In somewhat older studies, the ability of pancreatic cells to give rise to hepatocytes has been described. In addition to these two extrahepatic sources of stem cells for liver epithelial cells, several other sources have also been suggested, including neural stem cells. Evidence that extrahepatic stem cells can give rise to liver is summarized in the sections that follow.

4.4.1. Liver From Progenitor Cells of the Bone Marrow

Bone marrow contains several different cell types with stemlike potential, including hematopoietic (115,116), stromal (117), and mesenchymal stem cells (118-120). In addition to these progenitor cell types, it has been suggested that bone marrow contains a multipotent adult progenitor cell that expresses a broader tissue differentiation potential. However, whether this multipotent progenitor cell compartment of the bone marrow represents a single cell type with broad differentiation capacity or whether it represents an admixture of several tissue stem cell types has not been resolved (3). In fact, the multipotent progenitor cell of bone marrow may be related (or identical) to one of these other cell types (hematopoietic or mesenchymal stem cells) of the bone marrow. Bone marrow transplants generate cell lineages of the blood and have now been suggested to give rise to a number of other cell types, including cardiac muscle (121), skeletal muscle (122,123), neurons (124-126), lung epithelium (127), oval cells (128), hepatocytes (129133), and biliary epithelial cells (127,130). Bone marrow progenitor cell-derived hepatocytes have been demonstrated in rats (128), mice (129,132,133), and humans (130,131).

Transdifferentiation of bone marrow progenitor cells into hepatocytes has yielded variable replacement of liver, possibly related to the animal model employed. Transplantation of unfractionated bone marrow from male donors into lethally irradiated syngeneic female mice (B6D2F1) resulted in efficient reconstitution of the host hematopoietic system and generation of donor-derived hepatocytes in the host livers (129). The bone marrow-derived hepatocytes were identified in the hepatic plates of recipient livers using Y chromosome in situ hybridization (129). Quantitative analysis suggested that 1-2% of hepatocytes were bone marrow derived (129). These results suggested that bone marrow-derived stem cells could engraft and give rise to hepatocytes, albeit at low frequency, in normal liver.

Grompe and colleagues employed the murine model of hereditary tyrosinemia type I (70,71) to investigate the potential for bone marrow-derived stem cells to repopulate diseased liver (132,133). Transplantation of unfractionated bone marrow into FAH-/- mice resulted in replacement of 30-50% of the liver mass after a period of selection (132). Furthermore, when highly purified KTLS (c-kithlghThylowLin-Sca-1+) hematopoietic stem cells from male ROSA26/BA mice were transplanted into lethally irradiated female FAH-/- mice, liver engraftment and hepatocytic differentiation of transplanted cells was observed (132). The hepatocyte phenotype of engrafted cells was confirmed by expression of albumin and bile canalicular dipeptidylpeptidase IV (132). In addition, the engrafted cells expressed the FAH protein and were positive for P-galactosidase and the Y chromosome (132). It was also suggested that the c-kit- and Lin+ fractions of the bone marrow do not contain significant numbers of progenitor cells that can give rise to hepatocytes (132). Although substantial liver repopulation by bone marrow-derived hepatocytes was observed when selective conditions were employed (132,133), negligible hepatocyte replacement was observed in the absence of selective pressure (133).

In a similar study, unfractionated bone marrow from transgenic mice expressing Bcl2 under the control of the liver pyruvate kinase gene promoter was transplanted into normal mice, some of which were subjected to lethal irradiation (134). In this study, the frequency of hepatocyte differentiation from transplanted bone marrow progenitor cells was rare. However, when selection pressure was applied through the administration of anti-Fas antibodies, the small number of bone marrow-derived hepatocytes present in the livers of recipient mice expanded 6-fold to 20-fold, ultimately occupying approx 1% of the liver mass (134).

These studies showed that positive selection pressure can result in a higher degree of replacement of liver by bone marrow-derived hepatocytes. Other investigators have failed to detect donor-derived hepatocytes in undamaged livers after bone marrow transplant despite reconstitution of the hematopoi-etic system and replacement of the liver endothelium (135,136). These observations suggest that generation of hepatocytes from bone marrow stem cells is uncommon in the absence of strong selective pressure, such as that in the FAH-/- mouse model.

A few studies have examined the differentiation potential of bone marrow-derived progenitor cells in vitro. Reyes and colleagues isolated and established cultured multipotent adult progenitor cells from bone marrow of humans, mice, and rats (119,120). These multipotent adult progenitor cells copurified with the mesenchymal stem cell fraction of the bone marrow (119,120). When these cells are propagated on Matrigel in the presence of fibroblast growth factor 4 (FGF-4) and human growth factor (HGF), hepato-cytelike cells expressing albumin, cytokeratin 18 (CK18), and HNF3P resulted after 14 d of culture (137). Furthermore, these multipotent adult progenitor cell-derived hepatocytes expressed several functional character istics of mature hepatocytes, including secretion of urea and albumin, expression of phenobarbital-inducible cytochrome P450, the capacity to store glycogen, and the ability to take up low-density lipoprotein (LDL) (137), suggesting that the bone marrow of adult mammals contains progenitor cells with the potential to give rise to hepatocytes, and that these cells can be propagated in culture without loss of potency (137).

4.4.2. Liver From Stem Cells of the Pancreas

Several experimental models have been developed in which large eosino-philic cells that morphologically and phenotypically resemble hepatocytes are induced in the pancreas of rats and hamsters following severe pancreatic injury (138-141). Rats maintained on a copper-deficient diet for 8-10 wk showed widespread injury to exocrine elements of the pancreas (139,140). When the copper-deficient diet was replaced with a normal diet, hepatocytelike cells developed during the regeneration of the pancreatic tissue (139,140). In this model, cells that resembled hepatic oval cells were thought to represent the progenitor cells for pancreatic hepatocytes (140).

These observations coupled with those of other studies led to the suggestion that liver and pancreas may share a common stem cell (142). To examine the possibility that pancreatic oval cells could serve as liver progenitor cells, proliferating pancreatic oval cells were isolated and introduced into the livers of dipeptidylpeptidase IV-deficient rats via transplantation into the spleen (143). Following transplantation, hepatocytelike cells that express dipeptidylpeptidase IV activity were observed in the livers of recipient dipeptidylpeptidase IV-deficient animals, suggesting that oval cells proliferating in response to pancreatic injury caused by the copper-deficient diet can serve as hepatocyte progenitor cells (143). In a similar study, suspensions of pancreatic cells from normal adult mice were transplanted into FAH-/- mice to examine the possibility that the normal pancreas contains a population of hepatocyte progenitor cells (144). When selection pressure was applied, extensive liver repopulation (>50% replacement of liver) was observed in a subset of recipient mice, and another subset showed nodules of donor-derived hepatocytes (144).

Chen et al. (145) examined the fate of normal rat pancreatic ductal epithelial cells following implantation into the abdominal subcutaneous tissue or intraperitoneal cavity of adult syngeneic rats. In these studies, RP-2 pancreatic duct epithelial cells (146) were embedded in a gel composed of extracellular matrix (collagen I and Matrigel) prior to implantation (145). Eight weeks following subcutaneous implantation, nests of eosinophilic epi-thelioid cells and rare duct structures were observed in the recovered extracellular matrix gel (145). Six weeks following intraperitoneal implantation, trabeculae and clusters of large polygonal epithelioid cells with granular eosinophilic cytoplasms, resembling mature hepatocytes of the adult liver, were observed (145). These hepatocytelike cells expressed high levels of tyrosine aminotransferase, albumin, and transferrin and stained positively with HES6 monoclonal antibodies (145).

These studies demonstrated that normal pancreatic duct epithelial cells can differentiate into functional hepatocytes following implantation into an appropriate host microenvironment. They provide additional support for the suggestion that liver and pancreas share a similar stem cell with differentiation options that are determined by the tissue microenvironment.

4.4.3. Liver From Neural Stem Cell Cultures

Cultured neural stem cells isolated from mice (147) have been shown to give rise to neurons and glia following transplantation into brain tissue of host animals (148-150). Such neural stem cells have been suggested to have a differentiative plasticity when transplanted into various tissue microenvironments other than the brain. For instance, neural stem cells derived from adult donor tissue differentiate into hematopoietic lineages when engrafted into the bone marrow (151). When introduced into developing mouse blastocysts, neural stem cells contribute to cells of various germ layers and tissues of chimeric embryos, including the liver (152). Additional studies will be required to demonstrate whether neural stem cells can give rise to differentiated hepatocytes in the adult liver.

4.5. Human Liver Stem Cells

4.5.1. Evidence for Liver Stem Cells in Humans

In recent years, numerous investigators have attempted to identify and isolate human liver stem cells or have made observations in pathological human livers that suggest the existence of these cells. In many instances, investigators have attempted to determine if activation and proliferation of oval cells occurs in humans in a fashion similar to the oval cell reaction observed in rodents (153). Using morphologic criteria or immunohis-tochemical staining, several reports suggested the presence of cells resembling oval cells in several different human liver diseases (154-169). In some studies, cells exhibiting a phenotype consistent with oval cells were observed in normal human liver (169,170).

4.5.2. Liver From Bone Marrow in Humans

In rodents, bone marrow has been suggested to contain stem cells with hepatocytic differentiation potential (129,132). Likewise, several studies have suggested that human bone marrow contains stemlike progenitor cells that can differentiate into hepatocyte progeny under various conditions (130,131). The first evidence that bone marrow-derived progenitor cells give rise to hepatocytes in humans emerged from a study of archival autopsy or biopsy liver specimens from gender-mismatched transplant patients (130).

Liver tissue from two female patients who received therapeutic bone marrow transplants from male donors and liver tissue from four male patients who received orthotopic liver transplants from female donors was studied (130). Variable numbers of Y chromosome-positive hepatocytes and biliary epithelial cells were detected using in situ hybridization (130). In this study, the interval from transplant to liver sampling was 1 mo to 2 yr, and the numbers of Y chromosome-positive hepatocytes observed varied from 1 to 8% among the patients (130). However, when the investigators adjusted the data to account for their assessment of the insensitivity of Y chromosome in situ hybridization, the corrected results suggested that 5-40% of hepatocytes observed were derived from circulating progenitor cells, probably of bone marrow origin (130).

In a similar study, the livers of 9 female patients who received bone marrow transplant from male donors and the livers of 11 male patients who received orthotopic liver transplants from female donors were evaluated for Y chromosome-positive hepatocytes (131). Among these two groups of patients, 0.5-2% of hepatocytes were derived from extrahepatic progenitor cells, and clusters of Y chromosome-positive hepatocytes were observed in several instances, suggesting that hepatocyte progeny had clonally expanded after colonization of the liver by circulating progenitor cells (131).

In a third study, patients with hematologic or breast malignancies received high-dose chemotherapy and transplants of allogeneic peripheral blood stem cells from donors pretreated with granulocyte colony-stimulating factor (171). Y chromosome-positive hepatocytes were detected as early as 13 d posttransplant, and 4-7% of hepatocytes examined were donor derived (171).

In all of these studies, few hepatocytes derived from bone marrow (or other circulating progenitor cells) were detected, and the interval from transplant to sampling of the liver was short (less than a year in most cases). It has been suggested that the engraftment of bone marrow-derived stem cells in the liver may represent an early feature in liver transplants, but that hepa-tocytes derived from bone marrow progenitor cells may not persist as a long-term feature of grafted livers (172). In a study of gender-mismatched liver transplant patients with long interval between liver transplant and biopsy (1.2-12 yr), no host-derived hepatocytes were detected when the Y chromosome was used in in situ hybridization experiments (172).

4.6. Isolation and Culture of Adult Liver Stem Cells

Several types of epithelial cells can be isolated from rodent livers and established in primary or propagable cultures (173). Differentiated hepatocytes and bile duct cells can be maintained in primary culture for short periods, but generally these cell types exhibit limited propagability and lifespan in culture (174-176). In contrast, simple (undifferentiated) liver epithelial cells can be readily established and propagated in culture (177). These simple liver epithelial cell types possess some stemlike properties, suggesting that they may represent the cultured counterpart of epithelial stem cells in the adult liver (178).

4.6.1. Early Studies of Propagable Liver Epithelial Cells

Early long-term rat liver epithelial cell cultures were established from cell outgrowths in liver tissue explant cultures (179). Development of enzymatic techniques for the preparation of viable single-cell suspensions of liver cells made possible the selective culture of several liver epithelial cell types (180,181). Brief digestion of liver tissue with collagenase produces liver cell suspensions enriched for hepatocytes (181,182), whereas enrichment of nonparenchymal epithelial cell types can be accomplished by the selective removal of hepatocytes from collagenase-dispersed liver using various strong proteases, such as Pronase or trypsin (183-185). The nonparenchymal cells remaining following protease treatment of liver cell suspensions include macrophages (Kupffer cells), endothelial cells, bile ductular cells, Ito cells, and various hematopoietic cells (183,186). Also present in dispersed liver cell suspensions are simple epithelial cells (177,187).

4.6.2. Rat Liver Epithelial Stem Cells: Oval Cell Lines

Oval cells have been isolated from diseased liver and established in culture by several laboratories. Morphologically, cultured oval cells are cuboi-dal and grow in a monolayer (188,189). Some established lines of oval cells are stably diploid or pseudodiploid, are nontumorigenic, and do not proliferate in soft agar (188-191). Ultrastructurally, cultured oval cells exhibit cata-lase-positive peroxisomes that proliferate in response to treatment with clofibrate (189,190). Cultured oval cells generally express glucose-6-phos-phatase activity and lactate dehydrogenase isozymes 2-5, and they are variably positive for albumin and a-fetoprotein (189-192). A few oval cell lines have been characterized for cytokeratin expression and express CK8 and CK18 and variably express or not express CK7 and CK19 (189,191). As with some other liver epithelial cell lines, cultured oval cells tend to be anti-genically simple. Cultured LE/6 oval cells do not express antigens for mono-

Polygonal Cells Liver
Fig. 2. Morphology of cultured WB-F344 rat liver epithelial stem cells. Low-passage WB-F344 cells viewed by phase contrast microscopy.

clonal antibodies OC.1, OC.2, BD.1, H.1, or H.2 (reviewed in ref. 21). The presence of peroxisomes and glucose-6-phosphatase activity in cultured oval cells suggests that these cells are part of the hepatocyte lineage.

4.6.3. Rat Liver Epithelial Stem Cells: The WB-F344 Line

Several lines of rat liver epithelial cells have been established from the livers of normal adult rats (21). The WB-F344 rat liver epithelial cell line represents one such propagable rat liver epithelial cell line clonally derived from a single epithelial cell (20). WB-F344 cells are phenotypically similar to other established rat liver epithelial cell lines. A detailed comparison of the phenotypic properties of several rat liver epithelial cell lines was reviewed in ref. 21. WB-F344 cells are small (9- to 15-^m diameter), polygonal cells that grow in a monolayer (Fig. 2).

Ultrastructurally, WB-F344 cells exhibit a relatively simple cytoplasm with few organelles. Adjacent cells in confluent monolayers are joined by numerous desmosomes (20) and nexus junctions containing connexins 26 and 43 (193-196) and are dye coupled (194,196,197). Cells are polarized, surfaces directed to the growth medium interface contain microvilluslike projections, and a basement membranelike material containing fibronectin is deposited at the substrate interface (20). WB-F344 cells possess a stable diploid or quasidiploid karyotype (20). They do not proliferate in soft agar culture and are nontumorigenic following transplantation into neonatal syn-

geneic rats (20). WB-F344 cells share some phenotypic traits with both hepatocytes and biliary epithelial cells, but their overall phenotype differs distinctively from either differentiated cell type (20). Most notably, WB-F344 cells are null for the major antigens that typify and distinguish hepatocytes or biliary epithelial cells (198; A. E. Wennerberg and J. W. Grisham, 1993, unpublished observations).

4.6.4. Progenitor Cells From the Human Liver

Very few studies have appeared that reported the isolation and culture of human liver cells. Nussler and colleagues isolated a cell population from human liver and established it in propagable culture (199). The resulting cell line, AKN-1, has been characterized in culture and shows many characteristics of biliary epithelial cells (199). It is tempting to speculate that AKN-1 cells represent a cultured counterpart to a putative undifferentiated human liver stem cell. However, these cells contain chromosomal abnormalities, display an aneuploid DNA content, and are tumorigenic following transplantation into nude mice (199), indicating that the AKN-1 cell line may not represent propagable normal human liver stem cells.

In a similar study, cells expressing c-Kit and CD34 were isolated and cultured from diseased human liver (170). These cells were localized to portal tracts close to bile ducts in cirrhotic livers (170). In cell culture, these cells expressed markers (such as CK19) that suggested differentiation toward the biliary epithelial cell lineage (170). Of great significance, cells positive for c-Kit and CD34 were also isolated from normal human liver, albeit in smaller numbers, and these cells also acquired biliary epithelial differentiation in vitro (170).

0 0

Post a comment