**1. Introduction**

16 Liver Regeneration

Strick-Marchand, H. and M. C. Weiss (2002). "Inducible differentiation and morphogenesis

Talbot, N. C., V. G. Pursel, et al. (1994). "Colony isolation and secondary culture of fetal

Taub, R. (2004). "Liver regeneration: from myth to mechanism." Nat Rev Mol Cell Biol 5(10):

Theise, N. D., R. Saxena, et al. (1999). "The canals of Hering and hepatic stem cells in

Thiery, J. P., H. Acloque, et al. (2009). "Epithelial-mesenchymal transitions in development

Thomson, J. A., J. Itskovitz-Eldor, et al. (1998). "Embryonic stem cell lines derived from

Thorgeirsson, S. S. and J. W. Grisham (2006). "Hematopoietic cells as hepatocyte stem cells: a

Tomiya, T., I. Ogata, et al. (2000). "The mitogenic activity of hepatocyte growth factor on rat

Turner, R., O. Lozoya, et al. (2011). "Human hepatic stem cell and maturational liver lineage

Vassilopoulos, G., P. R. Wang, et al. (2003). "Transplanted bone marrow regenerates liver by

Wang, X., M. Foster, et al. (2003). "The origin and liver repopulating capacity of murine oval

Wilson J.W., Leduc E.H., (1950). "Abnormal mitosis in mouse liver". *Am J Anat*. Jan;8 6(1):51-

Xu, H. et al. (2010) "Liver-enriched transcription factors regulate microRNA-122 that targets

Yasui, O., N. Miura, et al. (1997). "Isolation of oval cells from Long-Evans Cinnamon rats

Yin, L., D. Lynch, et al. (1999). "Participation of different cell types in the restitutive response

Yovchev, M. I., P. N. Grozdanov, et al. (2008). "Identification of adult hepatic progenitor cells capable of repopulating injured rat liver." *Hepatology* 47(2): 636-647. Zhou, H., L.E. Rogler, et al. (2007). "Identification of hepatocytic and bile ductular cell

CUTL1 during liver development". *Hepatology* 52, 1431-1442,

hepatocytes is dependent upon endogenous transforming growth factor-alpha."

biology". Hepatology 53(3):1035-45.Ujike, K., T. Shinji, et al. (2000). "Kinetics of expression of connective tissue growth factor gene during liver regeneration after partial hepatectomy and D-galactosamine-induced liver injury in rats." Biochem

cells." Proceedings of the National Academy of Sciences of the United States of

and their transformation into hepatocytes in vivo in the rat liver." *Hepatology* 25(2):

of the rat liver to periportal injury induced by allyl alcohol." *Journal of hepatology*

lineages and candidate stem cells in bipolar ductular reactions in cirrhotic human

1): 794-804.

836-847.

73.

doi:10.1002/hep.23818.

31(3): 497-507.

liver" *Hepatology* 45: 716-724.

329-334.

biology. Animal 30A(12): 851-858.

and disease." Cell 139(5): 871-890.

Am J Pathol 157(5): 1693-1701.

Biophys Res Commun 277(2): 448-454.

cell fusion." Nature 422(6934): 901-904.

America 100 Suppl 1: 11881-11888.

humans." Hepatology 30(6): 1425-1433.

human blastocysts." Science 282(5391): 1145-1147.

critical review of the evidence." Hepatology 43(1): 2-8.

of bipotential liver cell lines from wild-type mouse embryos." Hepatology 36(4 Pt

porcine hepatocytes on STO feeder cells." In vitro cellular & developmental

There is great interest in the biology of liver progenitor cells (LPCs) because of their stem cell-like ability to regenerate the liver when the hepatocyte pool is exhausted. Barely detectable in healthy tissue, they emerge upon chronic insult in periportal regions, proliferate and migrate to injury sites in the parenchyma and eventually differentiate into hepatocytes and cholangiocytes to restore liver mass, morphology and function. The increasing worldwide shortage of livers for orthotopic transplantation means LPCs have assumed more prominence as candidates for cell therapy as an alternative therapeutic approach for the treatment of various liver diseases. However, an LPC response is usually seen in pre-cancerous liver pathologies and their high proliferation potential makes them possible transformation targets; associations that overshadow their restorative capability. This mandates that we continue to investigate the factors that govern their activation, proliferation and especially their differentiation into mature, functional cells to effectively direct transplanted cells towards regeneration and not tumorigenicity.

### **2. Normal liver tissue turnover**

Tissue regeneration and maintenance in healthy intestine and skin is achieved within days and weeks respectively. In contrast healthy liver has a very slow cell turnover rate and the vast majority of hepatocytes is considered to be in the quiescent, non-proliferative G0 phase of the cell cycle. It has been estimated that at any one time only 1 in 20,000 to 40,000 hepatocytes is undergoing mitotic cell division with an average life span of 200 to 300 days (Bucher & Malt, 1971).

The mechanisms by which hepatic cells are replaced in healthy liver are controversial. An early model, the "streaming liver" hypothesis is based on the metabolic zonation and differential gene expression patterns of periportal compared to pericentral hepatocytes. Periportal cells were proposed to proliferate and migrate ("stream") towards the central area with maturation during the journey and terminal differentiation achieved when the cells reached the central zone (Zajicek *et al.*, 1985; Arber *et al.*, 1988; Sigal *et al.*, 1992). However there is no convincing evidence for a periportal to pericentral differentiation gradient and while hepatocytes in opposing lobular areas are responsible for different

Liver Progenitor Cells, Cancer Stem Cells and Hepatocellular Carcinoma 19

hepatocytes has previously been underestimated. Rhim *et al.* showed that newborn uPA overexpressing mice with continuous hepatocytic necrosis could be rescued by transplantation of a small number of hepatocytes that required between 10 to 15 rounds of replication to generate sufficient liver mass (Rhim *et al.*, 1994; Rhim *et al.*, 1995). In addition, serial transplantation experiments performed in tyrosinemic mice caused by a deficiency for fumarylacetoacetate hydrolase (FAH) revealed that hepatocytes are capable of undergoing more than 70 cell doublings without loss of functionality (Overturf *et al.*, 1997). Conversely there is also recent evidence that hepatocytes might reach a state of "replicative senescence" under certain chronic conditions such as advanced cirrhosis, perhaps due to telomere

Repeated replication of healthy hepatocytes is the most efficient way to restore liver mass and function during normal tissue renewal and repair. If this process is inhibited or blocked during chronic chemical or carcinogenic hepatocyte insult, the liver relies on stem cell-like LPCs for its restoration. These cells are also referred to as "oval cells" in rodents (Fausto & Campbell, 2003) and the "Ductular Reaction" in humans due to their rather ductular phenotype in most human chronic liver diseases (Roskams & Desmet, 1998; Theise *et al*.,

The appearance of oval-like cells in the livers of rats treated with the azo dye "Butter Yellow" was originally reported in 1937 (Kinosita, 1937). Two decades later, Farber introduced the term "oval cell" for this population after observing small ovoid cells with a scant basophilic cytoplasm and a high nuclear to cytoplasmic ratio following treatment of rats with carcinogenic agents (Farber, 1956a, 1956b). Shortly after, Wilson and Leduc documented the proliferation of ductular cells that gave rise to hepatocytes and possibly new interlobular bile ducts in mice fed a methionine-rich, bentonite-supplemented diet and they were the first to suggest the existence of a bipotential liver progenitor or stem cell (Wilson & Leduc, 1958). Many experimental models involving toxins and carcinogens, alone or in combination with other surgical or dietary regimes, have since been developed and these facilitated the study of these progenitor cells, which are now widely accepted to

The precise origin of LPCs remains uncertain, even though many researchers have addressed this question. The lack of definite evidence regarding the cellular source of LPCs may reflect differences in the models used to induce them and has also been hampered by a lack of specific LPC markers. Lenzi *et al.* suggested bile ducts as the structure of origin and argued that LPCs express biliary markers such as cytokeratin (CK) 7 and CK19 and lack expression of the mesenchymal cell markers vimentin and desmin. Additionally, the degree of LPC proliferation during early ethionine-induced carcinogenesis was found to be proportional to the increase in biliary tree volume and the authors claimed that LPCs are

Other investigators have proposed an extrahepatic origin for LPCs. After it became apparent that some LPCs share c-kit, CD34 and Thy-1 expression with haematopoietic stem cells

shortening (Paradis *et al.*, 2001; Wiemann *et al.*, 2002).

**3.2 Liver progenitor cell-mediated regeneration** 

**3.2.1 History, origin and features of liver progenitor cells** 

represent adult LPCs; the progeny of hepatic stem-like cells.

simply part of spatially expanded cholangioles (Lenzi *et al.*, 1992).

1999).

metabolic functions, cells in either location are considered to be fully differentiated. By reversing the blood flow in the liver, Thurman and Kauffman demonstrated that this lobular zonation is not dependent on hepatocyte lineage progression but rather due to metaboliteinduced gene regulation (Thurman & Kauffman, 1985). Retroviral marking studies provided additional evidence against the "streaming liver" model since transplanted cells, traceable by β-galactosidase expression, remained in the original location for 15 months (Bralet *et al.*, 1994). Furthermore, experiments performed with mosaic livers of chimeric rats (Ng & Iannaccone, 1992) as well as approaches using transgenic hAAT/β-gal mice (Kennedy *et al.*, 1995) demonstrated that hepatocytes proliferate clonally during normal tissue renewal throughout the whole liver lobule. Collectively, these findings led to the conclusion that normal liver cell plates lack the existence of a main proliferative compartment and instead randomly distributed hepatocytes mediate normal liver turnover by slow clonal expansion without involvement of a liver stem cell (Ponder, 1996).

#### **3. Liver regeneration**

The liver has an enormous capacity to regenerate by (1) replication of remaining, healthy hepatocytes, (2) activation, expansion and differentiation of a stem cell compartment, or (3) by a combination of these processes. Which pathway is employed depends on the nature of the injury, its severity and duration. This is discussed in greater detail in the sections to follow.

#### **3.1 Hepatocyte-mediated regeneration**

The hepatic regenerative capacity is most clearly seen after surgical removal of liver mass. This model, referred to as partial hepatectomy (PHx), was introduced by Higgins and Anderson (Higgins & Anderson, 1931) and it is unquestionably the best studied liver regeneration model due to its simplicity of design and reproducibility. In the rat two-thirds PHx is performed, whereas in the mouse usually only the left lobe is removed due to technical difficulties in the performance of two-thirds PHx surgery in mice, with resultant high mortality (Fausto *et al.*, 2006). The removed lobes do not re-grow. Instead there is compensatory, hyperplastic growth of all residual cellular populations until the size of the organ achieves proportionality to the body size, as determined by metabolic demands of the organism (Kawasaki *et al.*, 1992; Starzl *et al.*, 1993). The different liver cell types do not divide simultaneously but show different kinetics in DNA synthesis. Periportal hepatocytes, with a presumably shorter G1 phase than pericentral cells (Rabes, 1976), are the first to undergo a wave of mitosis but DNA synthesis progresses to eventually involve the whole lobule with the exception of a few glutamine synthetase-positive, pericentral cells (Gebhardt, 1988). The proliferating hepatocytes are thought to provide mitogenic stimuli for the other hepatic cell populations. Biliary ductular cells, Kupffer and hepatic stellate cells (HSCs) and finally sinusoidal endothelial cells enter DNA synthesis about 24 hours later (Michalopoulos & DeFrances, 1997) with synchronised proliferation of each cell type for at least the first wave of replication. The greatest increase in liver mass can be seen by 72 hours with complete mass restoration after about one week (Grisham, 1962).

Although it was known from early experiments that repeated PHx does not exhaust hepatocyte growth (Simpson & Finck, 1963), the enormous proliferative capacity of adult

metabolic functions, cells in either location are considered to be fully differentiated. By reversing the blood flow in the liver, Thurman and Kauffman demonstrated that this lobular zonation is not dependent on hepatocyte lineage progression but rather due to metaboliteinduced gene regulation (Thurman & Kauffman, 1985). Retroviral marking studies provided additional evidence against the "streaming liver" model since transplanted cells, traceable by β-galactosidase expression, remained in the original location for 15 months (Bralet *et al.*, 1994). Furthermore, experiments performed with mosaic livers of chimeric rats (Ng & Iannaccone, 1992) as well as approaches using transgenic hAAT/β-gal mice (Kennedy *et al.*, 1995) demonstrated that hepatocytes proliferate clonally during normal tissue renewal throughout the whole liver lobule. Collectively, these findings led to the conclusion that normal liver cell plates lack the existence of a main proliferative compartment and instead randomly distributed hepatocytes mediate normal liver turnover by slow clonal expansion

The liver has an enormous capacity to regenerate by (1) replication of remaining, healthy hepatocytes, (2) activation, expansion and differentiation of a stem cell compartment, or (3) by a combination of these processes. Which pathway is employed depends on the nature of the injury, its severity and duration. This is discussed in greater detail in the sections to

The hepatic regenerative capacity is most clearly seen after surgical removal of liver mass. This model, referred to as partial hepatectomy (PHx), was introduced by Higgins and Anderson (Higgins & Anderson, 1931) and it is unquestionably the best studied liver regeneration model due to its simplicity of design and reproducibility. In the rat two-thirds PHx is performed, whereas in the mouse usually only the left lobe is removed due to technical difficulties in the performance of two-thirds PHx surgery in mice, with resultant high mortality (Fausto *et al.*, 2006). The removed lobes do not re-grow. Instead there is compensatory, hyperplastic growth of all residual cellular populations until the size of the organ achieves proportionality to the body size, as determined by metabolic demands of the organism (Kawasaki *et al.*, 1992; Starzl *et al.*, 1993). The different liver cell types do not divide simultaneously but show different kinetics in DNA synthesis. Periportal hepatocytes, with a presumably shorter G1 phase than pericentral cells (Rabes, 1976), are the first to undergo a wave of mitosis but DNA synthesis progresses to eventually involve the whole lobule with the exception of a few glutamine synthetase-positive, pericentral cells (Gebhardt, 1988). The proliferating hepatocytes are thought to provide mitogenic stimuli for the other hepatic cell populations. Biliary ductular cells, Kupffer and hepatic stellate cells (HSCs) and finally sinusoidal endothelial cells enter DNA synthesis about 24 hours later (Michalopoulos & DeFrances, 1997) with synchronised proliferation of each cell type for at least the first wave of replication. The greatest increase in liver mass can be seen by 72 hours

Although it was known from early experiments that repeated PHx does not exhaust hepatocyte growth (Simpson & Finck, 1963), the enormous proliferative capacity of adult

with complete mass restoration after about one week (Grisham, 1962).

without involvement of a liver stem cell (Ponder, 1996).

**3. Liver regeneration** 

**3.1 Hepatocyte-mediated regeneration** 

follow.

hepatocytes has previously been underestimated. Rhim *et al.* showed that newborn uPA overexpressing mice with continuous hepatocytic necrosis could be rescued by transplantation of a small number of hepatocytes that required between 10 to 15 rounds of replication to generate sufficient liver mass (Rhim *et al.*, 1994; Rhim *et al.*, 1995). In addition, serial transplantation experiments performed in tyrosinemic mice caused by a deficiency for fumarylacetoacetate hydrolase (FAH) revealed that hepatocytes are capable of undergoing more than 70 cell doublings without loss of functionality (Overturf *et al.*, 1997). Conversely there is also recent evidence that hepatocytes might reach a state of "replicative senescence" under certain chronic conditions such as advanced cirrhosis, perhaps due to telomere shortening (Paradis *et al.*, 2001; Wiemann *et al.*, 2002).

#### **3.2 Liver progenitor cell-mediated regeneration**

Repeated replication of healthy hepatocytes is the most efficient way to restore liver mass and function during normal tissue renewal and repair. If this process is inhibited or blocked during chronic chemical or carcinogenic hepatocyte insult, the liver relies on stem cell-like LPCs for its restoration. These cells are also referred to as "oval cells" in rodents (Fausto & Campbell, 2003) and the "Ductular Reaction" in humans due to their rather ductular phenotype in most human chronic liver diseases (Roskams & Desmet, 1998; Theise *et al*., 1999).

#### **3.2.1 History, origin and features of liver progenitor cells**

The appearance of oval-like cells in the livers of rats treated with the azo dye "Butter Yellow" was originally reported in 1937 (Kinosita, 1937). Two decades later, Farber introduced the term "oval cell" for this population after observing small ovoid cells with a scant basophilic cytoplasm and a high nuclear to cytoplasmic ratio following treatment of rats with carcinogenic agents (Farber, 1956a, 1956b). Shortly after, Wilson and Leduc documented the proliferation of ductular cells that gave rise to hepatocytes and possibly new interlobular bile ducts in mice fed a methionine-rich, bentonite-supplemented diet and they were the first to suggest the existence of a bipotential liver progenitor or stem cell (Wilson & Leduc, 1958). Many experimental models involving toxins and carcinogens, alone or in combination with other surgical or dietary regimes, have since been developed and these facilitated the study of these progenitor cells, which are now widely accepted to represent adult LPCs; the progeny of hepatic stem-like cells.

The precise origin of LPCs remains uncertain, even though many researchers have addressed this question. The lack of definite evidence regarding the cellular source of LPCs may reflect differences in the models used to induce them and has also been hampered by a lack of specific LPC markers. Lenzi *et al.* suggested bile ducts as the structure of origin and argued that LPCs express biliary markers such as cytokeratin (CK) 7 and CK19 and lack expression of the mesenchymal cell markers vimentin and desmin. Additionally, the degree of LPC proliferation during early ethionine-induced carcinogenesis was found to be proportional to the increase in biliary tree volume and the authors claimed that LPCs are simply part of spatially expanded cholangioles (Lenzi *et al.*, 1992).

Other investigators have proposed an extrahepatic origin for LPCs. After it became apparent that some LPCs share c-kit, CD34 and Thy-1 expression with haematopoietic stem cells

Liver Progenitor Cells, Cancer Stem Cells and Hepatocellular Carcinoma 21

hepatic plate (Grisham and Porta, 1964). Microscopic studies of early histological changes in rats following 2-acetylaminofluorene (2-AAF)/partial hepatectomy (PHx) treatment also show elongated ductular branches that are formed by proliferating LPCs, which originate from a stem cell compartment located in these canalicular-ductular junctions. The newly formed biliary structures represent cellular extensions of the Canals of Hering and remain connected to the terminal biliary ductules by a continuous basement membrane (Paku *et al.*, 2001). Reid and colleagues suggested epithelial cell adhesion molecule (EpCAM) as a suitable marker for isolation and study of these Canals of Hering-derived LPCs (Schmelzer *et al*., 2007). Lineage tracing of Sry (sex determining region Y)-box 9 (Sox9)-expressing cells supports the hypothesis that LPCs derive from the epithelial lining of bile ducts (Furuyama *et al*., 2011). Theise *et al*. conducted studies comparing normal with acetaminophen-induced necrotic liver and identified the human equivalent to the rodent Canals of Hering, a niche which is similarly thought to harbour stem-like cells that give rise to LPCs or the Ductular

Fig. 1. LPC ontogeny. During liver development hepatoblasts (Hb) differentiate into cholangiocytes (C) and hepatocytes (Hep) and might be incorporated into the Canals of Hering to serve as a stem cell compartment during chronic liver injury. Activated liver progenitor cells (LPC) proliferate after appropriate stimuli, are capable of self-renewal and later commit towards either the cholangiocytic or hepatocytic lineage to regenerate the liver. If kept in a proliferative state, LPCs are likely candidates for transformation and might

LPCs are a heterogeneous cell population and immature as well as intermediate phenotypes are observed before cells that express a differentiated phenotype are identified. Importantly, from activation to differentiation or transformation, they continuously change their morphology, phenotype and accordingly marker expression. LPCs express different combinations of phenotypic markers from both the hepatocytic and biliary lineage (Fig. 2) and also share epitopes with haematopoietic cells and cancer stem cells (CSCs; see table 1).

Reaction (Theise *et al.*, 1999).

represent cancer stem cells (CSCs).

(Fujio *et al.*, 1994; Omori *et al.*, 1997; Petersen *et al.*, 1998a), Petersen *et al.* were the first to suggest that LPCs could be derived from epithelial precursors in the bone marrow (Petersen *et al.*, 1999). Bone marrow-derived cells that potentially contribute to liver regeneration would enter via the portal vasculature and locate adjacent to the ducts in the periportal region, which is why Sell extended the preceding proposition by suggesting the periductular LPC as the candidate cell for an extrahepatic, bone marrow-derived stem cell in the liver (Sell, 2001). To test the hypothesis that cells from the bone marrow contribute to the formation of LPCs and hepatocytes, several investigators performed cell transplantation studies. They generally followed the fate of male bone marrow cells or purified haematopoietic stem cells transplanted into lethally irradiated female recipients that were in most cases subjected to liver injury. It was demonstrated that very minor fractions of LPCs or hepatocytes were donor-derived in both healthy and diseased livers (Petersen *et al.*, 1999; Theise *et al.*, 2000a, 2000b; Krause *et al.*, 2001; Wang *et al.*, 2002). The responsible population in the bone marrow capable of repopulating the liver was thought to be of ckithighThylowLinnegSca-1pos phenotype (Lagasse *et al.*, 2000). Soon after, the bone marrow was found to contain another stem cell subpopulation, the multipotent adult progenitor cell (MAPC), which can be induced to express hepatocyte phenotype and functions *in vitro* (Schwartz *et al.*, 2002) and is capable of differentiating into hepatocyte-like cells when transplanted into the liver (Jiang *et al.*, 2002). When donor-derived hepatocytes were examined genotypically, it was noted that they contained both donor and host genetic markers, indicating cell fusion as the likely mechanism by which hepatocytes are generated from bone marrow and not by transdifferentiation of haematopoietic stem cells (Vassilopoulos *et al.*, 2003; Wang *et al.*, 2003b). On the other hand, haematopoietic stem cells co-cultured with injured liver tissue separated by a trans-well membrane were shown to convert to a hepatocyte phenotype without fusion due to humoral factors released from the liver tissue. When engrafted into injured liver the haematopoietic stem cells differentiated into functional hepatocytes and their plasticity was proposed to facilitate the conversion, rather than the rare cell fusion event that was only seen at later stages of the experiment (Jang *et al.*, 2004). Recently it was demonstrated by transplantation of lacZ-transgenic bone marrow into virally or steatotically challenged mice that the contribution of extrahepatic cells to LPC-generated hepatocytes is minimal (Tonkin *et al*., 2008). Collectively, these experiments show that some bone marrow cells are capable of producing hepatocytes (with or without fusion, depending on the model and cell population used) to restore injured liver. However, it occurs at a low frequency and efficiency unless a strong selective pressure is applied (Thorgeirsson & Grisham, 2006). It is likely that the more significant role of bone marrow cells is to generate non-parenchymal cells during liver regeneration (Forbes *et al.*, 2004). The usual regeneration processes after acute and chronic liver injuries appear to rely predominantly on intrahepatic cells.

The most widely accepted view is that LPCs originate from liver-resident precursor or stem cells, which lie dormant and present in such low numbers as to be undetectable in normal liver. However, they can be activated to proliferate under certain pathological conditions (Fig. 1). Evidence from experiments showing that LPCs always emerge from periportal liver zones and the fact that selective periportal damage inhibits the LPC response (Petersen *et al.*, 1998b) have led to the conclusion that the precursor cell likely resides somewhere in the vicinity of the portal triad. Grisham and Porta found ductular proliferation in carcinogentreated rats that they attributed to activated stem-like cells from the Canals of Hering, the anatomical boundary between terminal bile ducts and the most distal hepatocytes of the

(Fujio *et al.*, 1994; Omori *et al.*, 1997; Petersen *et al.*, 1998a), Petersen *et al.* were the first to suggest that LPCs could be derived from epithelial precursors in the bone marrow (Petersen *et al.*, 1999). Bone marrow-derived cells that potentially contribute to liver regeneration would enter via the portal vasculature and locate adjacent to the ducts in the periportal region, which is why Sell extended the preceding proposition by suggesting the periductular LPC as the candidate cell for an extrahepatic, bone marrow-derived stem cell in the liver (Sell, 2001). To test the hypothesis that cells from the bone marrow contribute to the formation of LPCs and hepatocytes, several investigators performed cell transplantation studies. They generally followed the fate of male bone marrow cells or purified haematopoietic stem cells transplanted into lethally irradiated female recipients that were in most cases subjected to liver injury. It was demonstrated that very minor fractions of LPCs or hepatocytes were donor-derived in both healthy and diseased livers (Petersen *et al.*, 1999; Theise *et al.*, 2000a, 2000b; Krause *et al.*, 2001; Wang *et al.*, 2002). The responsible population in the bone marrow capable of repopulating the liver was thought to be of ckithighThylowLinnegSca-1pos phenotype (Lagasse *et al.*, 2000). Soon after, the bone marrow was found to contain another stem cell subpopulation, the multipotent adult progenitor cell (MAPC), which can be induced to express hepatocyte phenotype and functions *in vitro* (Schwartz *et al.*, 2002) and is capable of differentiating into hepatocyte-like cells when transplanted into the liver (Jiang *et al.*, 2002). When donor-derived hepatocytes were examined genotypically, it was noted that they contained both donor and host genetic markers, indicating cell fusion as the likely mechanism by which hepatocytes are generated from bone marrow and not by transdifferentiation of haematopoietic stem cells (Vassilopoulos *et al.*, 2003; Wang *et al.*, 2003b). On the other hand, haematopoietic stem cells co-cultured with injured liver tissue separated by a trans-well membrane were shown to convert to a hepatocyte phenotype without fusion due to humoral factors released from the liver tissue. When engrafted into injured liver the haematopoietic stem cells differentiated into functional hepatocytes and their plasticity was proposed to facilitate the conversion, rather than the rare cell fusion event that was only seen at later stages of the experiment (Jang *et al.*, 2004). Recently it was demonstrated by transplantation of lacZ-transgenic bone marrow into virally or steatotically challenged mice that the contribution of extrahepatic cells to LPC-generated hepatocytes is minimal (Tonkin *et al*., 2008). Collectively, these experiments show that some bone marrow cells are capable of producing hepatocytes (with or without fusion, depending on the model and cell population used) to restore injured liver. However, it occurs at a low frequency and efficiency unless a strong selective pressure is applied (Thorgeirsson & Grisham, 2006). It is likely that the more significant role of bone marrow cells is to generate non-parenchymal cells during liver regeneration (Forbes *et al.*, 2004). The usual regeneration processes after acute and chronic liver

injuries appear to rely predominantly on intrahepatic cells.

The most widely accepted view is that LPCs originate from liver-resident precursor or stem cells, which lie dormant and present in such low numbers as to be undetectable in normal liver. However, they can be activated to proliferate under certain pathological conditions (Fig. 1). Evidence from experiments showing that LPCs always emerge from periportal liver zones and the fact that selective periportal damage inhibits the LPC response (Petersen *et al.*, 1998b) have led to the conclusion that the precursor cell likely resides somewhere in the vicinity of the portal triad. Grisham and Porta found ductular proliferation in carcinogentreated rats that they attributed to activated stem-like cells from the Canals of Hering, the anatomical boundary between terminal bile ducts and the most distal hepatocytes of the hepatic plate (Grisham and Porta, 1964). Microscopic studies of early histological changes in rats following 2-acetylaminofluorene (2-AAF)/partial hepatectomy (PHx) treatment also show elongated ductular branches that are formed by proliferating LPCs, which originate from a stem cell compartment located in these canalicular-ductular junctions. The newly formed biliary structures represent cellular extensions of the Canals of Hering and remain connected to the terminal biliary ductules by a continuous basement membrane (Paku *et al.*, 2001). Reid and colleagues suggested epithelial cell adhesion molecule (EpCAM) as a suitable marker for isolation and study of these Canals of Hering-derived LPCs (Schmelzer *et al*., 2007). Lineage tracing of Sry (sex determining region Y)-box 9 (Sox9)-expressing cells supports the hypothesis that LPCs derive from the epithelial lining of bile ducts (Furuyama *et al*., 2011). Theise *et al*. conducted studies comparing normal with acetaminophen-induced necrotic liver and identified the human equivalent to the rodent Canals of Hering, a niche which is similarly thought to harbour stem-like cells that give rise to LPCs or the Ductular Reaction (Theise *et al.*, 1999).

Fig. 1. LPC ontogeny. During liver development hepatoblasts (Hb) differentiate into cholangiocytes (C) and hepatocytes (Hep) and might be incorporated into the Canals of Hering to serve as a stem cell compartment during chronic liver injury. Activated liver progenitor cells (LPC) proliferate after appropriate stimuli, are capable of self-renewal and later commit towards either the cholangiocytic or hepatocytic lineage to regenerate the liver. If kept in a proliferative state, LPCs are likely candidates for transformation and might represent cancer stem cells (CSCs).

LPCs are a heterogeneous cell population and immature as well as intermediate phenotypes are observed before cells that express a differentiated phenotype are identified. Importantly, from activation to differentiation or transformation, they continuously change their morphology, phenotype and accordingly marker expression. LPCs express different combinations of phenotypic markers from both the hepatocytic and biliary lineage (Fig. 2) and also share epitopes with haematopoietic cells and cancer stem cells (CSCs; see table 1).

Liver Progenitor Cells, Cancer Stem Cells and Hepatocellular Carcinoma 23

They have been shown to differentiate at least bipotentially into hepatocytes and cholangiocytes (Tirnitz-Parker *et al*., 2007), and in some models display multipotentiality, also producing intestinal and pancreatic lineages (Tatematsu *et al.*, 1985; Yang *et al.*, 2002; Leite *et al.*, 2007). Hence it is not surprising that there is still not a single LPC-specific marker available and a combination of phenotypic markers is required for their identification or

LPCs infiltrate the parenchyma in close spatial and temporal association with hepatic stellate cells (HSCs). Following activation, HSCs differentiate from quiescent, vitamin A-rich cells into α-smooth muscle actin-positive myofibroblastic cells, which are capable of matrix degradation to generate space for cell migration as well as fibrogenesis and collagen deposition to provide chronically injured liver with architectural support. The activation, proliferation, migration and differentiation status of LPCs and HSCs, as well as their beneficial as opposed to pathological contributions, are controlled by key cytokines. LPCs and HSCs have been reported to influence each other's behaviour through paracrine signalling. LPCs produce a range of cytokines, including lymphotoxin β (LTβ), which signals via the LTβ receptor on HSCs to activate the NFkB pathway. This results in production of intercellular adhesion molecule 1 (ICAM-1) and regulated upon activation, normal T-cell expressed and secreted (RANTES), which then act as chemotactic agents for LPCs and inflammatory cells involved in the wound healing response to chronic liver injury (Ruddell *et al.*, 2009). Several other factors mediating the LPC response have been identified, including tumour necrosis factor (TNF), TNF-like weak inducer of apoptosis (TWEAK), interferon gamma (IFNγ), and transforming growth factor beta (TGFβ) among others (Knight *et al.*, 2000; Akhurst *et al.*, 2005; Knight *et al.*, 2005; Knight & Yeoh, 2005; Knight *et al.*, 2007; Tirnitz-Parker *et al.*, 2010). Abrogation of these key signalling pathways inhibits the LPC response to injury and prevents or diminishes liver fibrosis in animal models (Davies *et al.*, 2006; Lim *et al.*, 2006; Knight *et al.*, 2008). In the setting of impaired wound healing combined with chronic inflammation, the regenerative fibrotic response turns into pathological fibrogenesis, which can progress to cirrhosis and eventually hepatocellular carcinoma (HCC).

The majority of commonly used LPC induction models was originally developed to study the process of hepatocarcinogenesis. They generally combine an injuring mitotic stimulus, usually in the form of functional liver mass loss (chemically or otherwise-induced), with a manipulation that chronically damages hepatocytes or blocks their ability to divide and prevents them from contributing to the liver regeneration process. Described below are four

This model is mainly used to induce liver injury in the rat. Administration of Dgalactosamine inhibits RNA and protein synthesis in centrilobular hepatocytes by trapping and depleting uridine-nucleotides and UDP-glucose (Decker & Keppler, 1972), leading to acute necrosis. Hepatocyte replication is not fully blocked in this model; the response is only delayed. LPCs are resistant to the chemical as they do not metabolise D-galactosamine and are induced to proliferate within 48 hours after injury. They migrate into the parenchyma, where they generate both ductular cells and small hepatocytes (Lemire *et al.*, 1991; Dabeva &

**3.2.2 Rodent liver progenitor cell induction models** 

examples of the most commonly used regimens.

**3.2.2.1 D-galactosamine** 

Shafritz, 1993).

isolation.

Fig. 2. Bipotentiality of LPCs. Immunofluorescent characterisation of the clonally established LPC line BMOL (Tirnitz-Parker *et al*., 2007) demonstrates the cells' bipotentiality. Immature BMOL cells co-express the hepatocytic markers muscle 2-pyruvate kinase (A, green) and transferrin (B, green) with the biliary markers A6 (A, red) and CK19 (B, red).


Table 1. Marker expression by adult liver cells. A6, murine marker, epitope unknown; AFP, α-fetoprotein; Alb, albumin; CD, cluster of differentiation; CK, cytokeratin; c-kit, CD117, stem cell factor receptor; Cx, connexin; Dlk, delta-like protein; E-cad, E-cadherin; EpCAM, epithelial cell adhesion molecule; GGT IV, γ-glutamyl transpeptidase IV; M2PK, muscle 2-pyruvate kinase ; OV-6, rat and human marker, epitope shared by CK14 and 19; π-GST, pi-glutatione-Stransferase; Sca-1, stem cell antigen 1; Thy-1, thymocyte differentiation antigen 1.

Fig. 2. Bipotentiality of LPCs. Immunofluorescent characterisation of the clonally established LPC line BMOL (Tirnitz-Parker *et al*., 2007) demonstrates the cells' bipotentiality. Immature BMOL cells co-express the hepatocytic markers muscle 2-pyruvate kinase (A, green) and

Table 1. Marker expression by adult liver cells. A6, murine marker, epitope unknown; AFP, α-fetoprotein; Alb, albumin; CD, cluster of differentiation; CK, cytokeratin; c-kit, CD117, stem cell factor receptor; Cx, connexin; Dlk, delta-like protein; E-cad, E-cadherin; EpCAM, epithelial cell adhesion molecule; GGT IV, γ-glutamyl transpeptidase IV; M2PK, muscle 2-pyruvate kinase ; OV-6, rat and human marker, epitope shared by CK14 and 19; π-GST, pi-glutatione-S-

transferase; Sca-1, stem cell antigen 1; Thy-1, thymocyte differentiation antigen 1.

transferrin (B, green) with the biliary markers A6 (A, red) and CK19 (B, red).

They have been shown to differentiate at least bipotentially into hepatocytes and cholangiocytes (Tirnitz-Parker *et al*., 2007), and in some models display multipotentiality, also producing intestinal and pancreatic lineages (Tatematsu *et al.*, 1985; Yang *et al.*, 2002; Leite *et al.*, 2007). Hence it is not surprising that there is still not a single LPC-specific marker available and a combination of phenotypic markers is required for their identification or isolation.

LPCs infiltrate the parenchyma in close spatial and temporal association with hepatic stellate cells (HSCs). Following activation, HSCs differentiate from quiescent, vitamin A-rich cells into α-smooth muscle actin-positive myofibroblastic cells, which are capable of matrix degradation to generate space for cell migration as well as fibrogenesis and collagen deposition to provide chronically injured liver with architectural support. The activation, proliferation, migration and differentiation status of LPCs and HSCs, as well as their beneficial as opposed to pathological contributions, are controlled by key cytokines. LPCs and HSCs have been reported to influence each other's behaviour through paracrine signalling. LPCs produce a range of cytokines, including lymphotoxin β (LTβ), which signals via the LTβ receptor on HSCs to activate the NFkB pathway. This results in production of intercellular adhesion molecule 1 (ICAM-1) and regulated upon activation, normal T-cell expressed and secreted (RANTES), which then act as chemotactic agents for LPCs and inflammatory cells involved in the wound healing response to chronic liver injury (Ruddell *et al.*, 2009). Several other factors mediating the LPC response have been identified, including tumour necrosis factor (TNF), TNF-like weak inducer of apoptosis (TWEAK), interferon gamma (IFNγ), and transforming growth factor beta (TGFβ) among others (Knight *et al.*, 2000; Akhurst *et al.*, 2005; Knight *et al.*, 2005; Knight & Yeoh, 2005; Knight *et al.*, 2007; Tirnitz-Parker *et al.*, 2010). Abrogation of these key signalling pathways inhibits the LPC response to injury and prevents or diminishes liver fibrosis in animal models (Davies *et al.*, 2006; Lim *et al.*, 2006; Knight *et al.*, 2008). In the setting of impaired wound healing combined with chronic inflammation, the regenerative fibrotic response turns into pathological fibrogenesis, which can progress to cirrhosis and eventually hepatocellular carcinoma (HCC).
