ln vited Revie w Ductular hepatocytes - Digitum

Histol Histopathol(l995) 10: 433-456 Histology and Histopathology

ln vited Revie w

Ductular hepatocytes A.E. Sirica Medical College of Virginia, Virginia Cornmonwealth University, Richrnond, USA

Summary. Ductular hepatocytes are observed in the livers of both experimental animals and man under various conditions of severe toxin-, carcinogen- or viral- induced hepatic injury with prominent loss of parenchymal hepatocytes. These unique hepatic epithelial cells are characterized by phenotypic traits that are intermediate between those of hepatocytes and intrahepatic biliary epithelium. The origin of ductular hepatocytes is controversial, but it has been hypothesized that they may represent a transitional cell stage associated with either (1) a ductular metaplasia of parenchymal hepatocytes into intrahepatic biliary epithelium, (2) a metaplastic conversion of intrahepatic bile duct or ductular epithelium into hepatocytes, or (3) differentiation of a putative liver stem cell along the hepatocyte lineage. Depending on the liver disease state being investigated, evidence is presented to support al1 three of these possibilities. Of particular interest is the increasing evidence supporting the existence of a facultative pluripotent stem-like cell associated with the intrahepatic biliary tract, which appears capable of differentiating into various gut endoderm-derived cell types, including hepatocytes, small intestinal mucosa1 cells, and pancreatic acinar cells. Ductular cells of pancreas have also been demonstrated to alter their differentiation commitment under various induced conditions of pancreatic injury and regeneration, so as to give rise to pancreatic hepatocytes. The presence of a putative stem-like cell in liver together with the plasticity exhibited by some hepatocytes and biliary epithelial cells in various forms of severe hepatic and biliary tract injury can have important implications for carcinogenesis and aberrant regenerative responses in liver. In addition, novel in vivo and cell culture models have been developed, which are serving as potentially powerful tools for investigating the effects of specific growth factors, extracellular matrix components, hormones and other agents on the ability of nonparenchymal epithelial liver cell types to differentiate into hepatocyte-like cells.

Offprint requests to: Dr. Alphonse E. Sirica, Department of Pathology, Medical College of Virginia, Virginia Commonwealth University, Box 980297, Richmond, VA 23298-0297, USA

Key words: Ductular hepatocyte, Neocholangiole, Ductular metaplasia, Oval cells, Bile ductules

lntroduction

~Ductular hepatocytes» is the term given to epithelial cells having phenotypic characteristics intermediate between those of hepatocytes and intrahepatic bile duct epithelium, which have been observed to occur in a variety of liver diseases characterized by severe hepatic injury and liver parenchymal cell loss (Gerber et al., 1983; Thung, 1990). This review will detail what is currently known about ductular hepatocytes in various clinical and experimental forms of liver disease, with an emphasis on critically evaluating hypotheses that have been developed to explain their cellular origin and potential role in liver in which the normal regenerative ability of preexisting hepatocytes has become greatly impaired as the result of severe toxin-, carcinogen-, or viral-induced injury. In addition, the saiient features of two novel experimental rat models of ductular hepatocyte formation recently developed in the author's laboratory will be described (Elmore and Sirica, 1991; Sirica, 1992; Sirica and Williams, 1992; Sirica et al., 1994a). Recent studies concerning the effects of specific growth factors, extracellular matrix components, hormones and other agents on the developmental potential of various nonparenchymai epithelial liver cell types will also be reviewed.

Ductular structures in normal liver

The intrahepatic biliary tract of normal liver consists of the bile canaliculi, which are formed as specialized regions of the plasma membranes of adjacent hepatocytes, together with the canals of Hering, bile ductules, interlobular bile ducts, septal bile ducts, and the large intrahepatic bile ducts near the hepatic hilum (Yamada et al., 1992; Terada and Nakanuma, 1993). The canals of Hering are minute channels located at the edge of the portal tracts in the region of the limiting plate of hepatocytes and serve to connect the bile canalicular network of the hepatocytes to the bile ductules (Steiner

Ductular hepatocytes

et al., 1965; Tanikawa, 1979; Phillips et al., 1987). They are formed in part by hepatocytes from the limiting plate and in part by simple flattened or low cuboidal biliary epithelial cells (Steiner et al., 1965; Popper and Schaffner, 1970; Tanikawa, 1979; Phillips et al., 1987; Sirica et al., 1992). Hepatocytes and biliary epithelial cells of the canals of Hering are attached by junctional complexes, but only the biliary epithelial cells rest on a basement membrane (Philips et al., 1987). In contrast, bile ductules do not contain hepatocytes in their lining epithelium (Steiner et al., 1965; Tanikawa, 1979; Phillips et al., 1987). Rather, they are short fine tubules lined by a rosette of two to I six cuboidal epithelial cells surrounded by a continuous basement membrane (Steiner et al., 1965; Popper and Schaffner, 1970; Tanikawa, 1979; Kono and Nakanuma, 1992). Yamamoto and Phillips (1984) have further demonstrated by the use of the scanning electron rnicroscopy of biliary tract casts, the presence of a richly anastomosing bile ductular plexus located at the borders of the portal tracts of the normal liver of rats, a species which lacks a gallbladder. Interestingly, the existence of such a periportal bile ductular plexus was not demonstrated by these investigators in biliary tract casts prepared from species with a gallbladder, including guinea pig, rabbit, dog, pig, rhesus monkey and man (Yamarnoto and Phillips, 1984; Yamamoto et al., 1985). The functional significance of the bile ductular plexus in rat liver needs to be elucidated, but Yamamoto and Phillips (1984) have proposed that since the rat does not possess a gallbladder, it may be acting as a reservoir for storing and modifying original canalicular bile secreted by the hepatocytes. In addition, the role played by this bile ductular plexus in proliferative bile ductular and oval cell responses to carcinogen and toxin-induced liver injury in rats still remains to be defined.

The bile ductules cornmunicate with the interlobular bile ducts within the portal tracts. In cross section, the lining epithelium of the interlobular bile ducts is composed of a single cell layer of cuboidal biliary cells, present in a larger number and taller than those lining the bile ductular lumens. These ducts are also enveloped by a well developed continuous basement membrane (Tanikawa, 1979; Phillips et al., 1987). However, unlike bile ductules, the interlobular bile ducts are accompanied by terminal branches of the hepatic artery and portal vein (Popper and Schaffner, 1970; Yamamoto and Phillips, 1984), which together fonn the main structures of the portal triad (Phillips et al., 1987).

As the size of the intrahepatic bile ducts continues to enlarge, so does the number and height of their lining epithelium, with the larger ducts becoming lined by a simple columnar epithelium. Yamamoto et al. (1985) initially had shown that the large hilar intrahepatic bile ducts of human liver contained numerous irregular side branches and pouches, forrning complex biliary plexus that interconnects with the main bile ducts. Sub- sequently, Terada and his associates (1987, 1993) demonstrated that the hilar biliary plexus reported by

Yamamoto et al. (1985) was actually composed of intrahepatic peribiliary glands, consisting of both sparse intramural glands and abundant extramural glands. The intramural glands were further shown to be simple tubular mucous glands, while the extramural glands were found to be branched seromucous tubuloalveolar glands (Terada et al., 1987).

Various terms have been used synonymously, particularly when referring to the terminal portions of the normal intrahepatic biliary tract. For example, other terms used to denote the canal of Hering have included the intermediate piece, canalicular-ductular junction, ampulla, and cholangiole (Tanikawa, 1979; Phillips et al., 1987). The use of the term cholangiole has also made for an added confusion in nomenclature, since it has been applied by different authors to describe either the canals of Hering, bile ductules, or even bile ducts (Grisham and Hartroft, 1961; Popper and Schaffner, 1970; Phillips and Poucell, 198 1; Uchida and Peters, 1983; Phillips et al., 1987). In this review, the term cholangiole will be used as a synonym only for the canal of Hering. In addition, the terms canal of Hering-like or cholangiolar-like will be employed to describe structures which are not found as part of the normal intrahepatic biliary tract, but which can occur in liver under select conditions of severe hepatic injury and parenchymal cell loss.

Ductular hepatocytes In clinical and experimental liver disease

Ductular structures become conspicuously increased in liver in a variety of clinical and experimental forms of liver disease with hepatic injury (Popper et al., 1957; Steiner et al., 1962; Grisham and Porta, 1964; Desmet, 1985; Tavoloni, 1987; Sirica et al., 1990; Thung, 1990; Van Eyken and Desmet, 1992; Phillips et al., 1993). This phenomenon has also been referred to as ductular «proliferation» (Thung, 1990; Van Eyken and Desmet, 1992; Phillips et al., 1993), with four histological types having been described: «typical» ductular proliferation, «atypical» ductular proliferation, «oval cell» proliferation, and «neocholangiolar» proliferation (Farber, 1956; Phillips and Poucell, 1981; Desmet, 1986; Nakanuma and Ohta, 1986; Tavoloni, 1987; Sirica et al., 1990; Thung, 1990; Van Eyken and Desmet, 1992; Phillips et al., 1993). In general, the latter three fonns of ductular proliferation have been found to be associated with the appearance of ductular hepatocytes.

The origin of proliferated ductules and of ductular hepatocytes forrning in response to particular types of liver injury remains a matter of significant controversy and debate (Burt and MacSween, 1993). Depending upon the liver disease state being investigated, it has been postulated that proliferating ductules are derived from either multiplication of preexisting bile ductules (Desmet, 1986; Nakanuma and Ohta, 1986; Tavoloni, 1987; Gall and Bhathal, 1990; Slott et al., 1990; Sirica et al., 1992; Van Eyken and Desmet, 1992). ductular


metaplasia of hepatocytes (Uchida and Peters, 1983; Desmet, 1986; Nakanuma and Ohta, 1986; Van Eyken and Desmet, 1992), or possibly from activation and proliferation of a putative liver stem cell population (Sell and Salman, 1984; Zajicek, 1991). Moreover, ductular hepatocytes have been hypothesized to represent an intermediate cell stage associated with either the transdifferentiation or metaplastic conversion of hepatocytes into bile duct-like cells (Thung, 1990; Van Eyken and Desmet, 1992), with differentiation of a liver stem-like cell along the hepatocyte lineage (Sell, 1990; Lemire et al., 1991; Dabeva and Shafritz, 1993; Fausto et al., 1993; Thorgeirsson et al., 1993), or with transdifferentiation of bile duct or ductular cells into hepatocytes (Nomoto et al., 1992; Sirica, 1992; Sirica and Williams, 1992; Sirica et al., 1994a). Each of the different histological forms of ductular proliferation listed above will now be discussed in the context of these different hypotheses.

Typical ductular proliferation in acute cholestatic liver injury

In both experimental animals and in man, typical ductular proliferation is a prorninent histological feature of acute cholestatic liver injury. The experimental prototype for this condition in rats is bile duct ligation, and its clinical counterpart in humans is acute extrahepatic biliary obstruction (Tavoloni, 1987; Sirica et al., 1990; Van Eyken and Desmet, 1992). A similar pattern of ductular proliferation is also seen in the livers of rats during their chronic dietary intoxication with chemical agents such as a-naphthylisothiocyanate (McLean and Rees, 1958; Steiner and Carruthers, 1963; Chou and Gibson, 1972; Sirica and Cihla, 1984) and 4,4'-diaminodiphenylmethane (Fukushima et al., 1979; Sell, 1983).

Typical ductular proliferation is characterized by an increase in liver of well formed bile ductular-like structures, usually cut in cross section, and which exhibit definite lumina (Nakanuma and Ohta, 1986; Slott et al., 1990; Van Eyken and Desmet, 1992; Sansonno and Damrnacco, 1993). In the case of the bile duct-ligated rat, the evidence is persuasive that this type of ductular proliferation is due almost entirely to the multiplication of preexisting bile ductules (Grisham and Porta, 1964; Wu et al., 1981; Gall and Bhathal, 1990; Slott et al., 1990; Sirica et al., 1992). In this context, Slott et al. (1990) have shown that after bile duct ligation in rats, replicative DNA synthesis occurs in the lining cells of hyperplastic bile ductal and ductular-like structures of al1 sizes and not just in those lining the smallest ductules, strongly implying that the proliferative compartment in this model is not segregated in the terminal bile ductules or canals of Hering. Gall and Bhathal (1990) have also reported that the contribution to the hyperplastic bile ductular epithelial cell pool in the livers of bile duct- ligated rats from an intermediate cell type within the canals of Hering is very small. It has been further

demonstrated by Slott et al. (1990) that the typical bile ductular proliferation associated with bile duct ligation in rats is manifested as an elongation, as opposed to circumferential enlargement or sprouting of new side branches, of the preexisting bile ductular tree, which after severa1 weeks of bile duct ligation has been generally found to be expanded largely into zone 1 of the liver acini (Gall and Bhathal, 1990; Sirica et al., 1991; Sirica and Williams, 1992). The increase in typical ductules which occurs preferentially at the margins of the portal tracts of human liver in acute extrahepatic biliary obstruction also appears to be due to the actual proliferation of preexisting bile ductules, since mitotic figures may be observed in their lining cells (Desmet, 1985, 1986; Van Eyken and Desmet, 1992).

Transitional cells having morphological and functional characteristics intermediate between those of intrahepatic biliary epithelial cells and hepatocytes either have not been observed (Grisham and Porta, 1964; Sirica et al., 1985; Yokoyama et al., 1986; Sirica et al., 1991; Sirica and Williams, 1992) or have only very rarely been seen (Sirica and Cihla, 1984; Gall and Bhathal, 1990) in the livers of rats that have been subjected just to bile duct ligation. Moreover, the transplantation of a 295% pure population of hyperplastic bile ductular epithelial cells freshly isolated from the livers of bile duct-ligated rats into the interscapular fat pads of syngeneic recipient rats was found to give rise to only well-differentiated bile ductular-like structures within a dense connective tissue stroma, without morphological evidence of transitional cells or ductular hepatocytes forming at the cell transplantation site (Sirica et al., 1991).

Apoptotic cell death has been demonstrated to be the major mechanism for the involution of hyperplastic bile ductular epithelial cells in the livers of bile duct-ligated rats, being more prominent by 6-7 weeks after ligation of the common bile duct than at earlier time points (Aronson et al., 1988; Gall and Bhathal, 1990). Deletion of hyperplastic intrahepatic biliary epithelial cells by apoptosis has also been demonstrated in rats previously subjected to bile duct ligation or to a chronic feeding of a-napthylisothiocyanate, following removal of the proliferative stimulus (Bhathal and Gall, 1985). Interestingly, the bcl-2 protooncogene protein, which prolongs cell survival by blocking apoptosis, has recently been reported to be detected by immunohistochemistry in intrahepatic bile ducts and ductules of normal human liver, as well as in proliferated ductules in a variety of human hepatobiliary disorders, including bile duct obstruction, and in 8 of 11 cholangiocarcinomas that were analyzed in this particular study (Charlotte et al., 1994).

Atypical ductular proliferation and ductular metaplasia of hepatocytes

Atypical ductular proliferation has been largely investigated in various chronic forms of human liver disease with injury and in particular, has been reported


to occur in those conditions characterized by longstanding incomplete biliary obstmction, such as primary biliary cirrhosis, primary sclerosing cholangitis, chronic alcoholic liver disease, and foca1 nodular hyperplasia (Uchida and Peters, 1983; Butron Vila et al., 1984; Desmet, 1985,1986; Nakanuma and Ohta, 1986; Burt et al., 1987; Yamada et al., 1987; Van Eyken et al., 1989; Ayres et al., 1991; Van Eyken and Desmet, 1992; Sansonno and Dammacco, 1993). In contrast to typical hyperplastic bile ductules, proliferated atypical ductules display an anastomosing, elongated and tortuous configuration with poorly defined lumens lined by irregular flattened or cuboidai epithelial cells. They are usually cut longitudinally and are generally formed adjacent to the hepatic parenchyma (Uchida and Peters, 1983; Nakanuma and Ohta, 1986; Yamada et al., 1987; Van Eyken and Desmet, 1992; Sansonno and Dammacco, 1993). Nakanuma and Ohta (1986) have further reported that in a spectrum of human hepatobiliary diseases, proliferated atypical ductules lacked immunohistochemically-detectable epithelial membrane antigen, whereas this antigen was usually reveaied along the luminal surfaces of proliferated typical ductules.

There is an increasing body of evidence suggesting that proliferated atypical ductules do not arise by the proliferation of preexisting bile ductules, but rather, are derived from a metaplastic conversion of hepatocytes into bile ductules (Van Eyken and Desmet, 1992). Support for the concept of cductular metaplasia» of hepatocytes in various human hepatobiliary diseases is based in large part on the following descriptive findings: 1) the presence in proliferated atypical ductules of intermediate or transitional cells exhibiting ultrastmctural andlor select phenotypic characteristics of both hepatocytes and intrahepatic biliary epithelium (Uchida and Peters, 1983; Butron Vila et al., 1984; Van Eyken and Desmet, 1992; Hiley et al., 1993); 2) the demonstration by computerized three-dimensional reconstruction of biliary pathways in primary biliary cirrhosis that proliferated atypical ductules are in communication with liver cell plates, but not with preexisting bile ductules and ducts (Yamada et al., 1987); and (3), the demonstration of positive immunohistochemical staining of hepatocytes either near or continuous with areas of proliferated atypical ductules with antibodies reactive for the bile duct-type cytokeratins 7 and 19 (Van Eyken et al., 1989; Van Eyken and Desmet, 1992), as well as with antibodies against a variety of different non-cytokeratin antigens, which in normal liver has been shown to be restricted to the bile duct epithelium. These included positive immunohistochemical reactivity for S-100 protein (Vanstapel et al., 1984, 1986; Desmet, 1986), tissue polypeptide antigen (Burt et d. , 1987), a 41-KD antigen temed 17-1A Ag that is also expressed in human colon carcinoma cells (Sansonno and Dammacco, 1993), an undefined biliary antigen identified by its reactivity with a monoclonal antibod designated clone 5 (Ayres et al., 4: 1991), H, Lea, and Le blood group antigens (Nakanuma

and Sasaki, 1989), and VLA adhesion molecules a2, 3 and 6 (Volpes et al., 1991). Interistingly, Van Eyken et al. (1989) reported that in both alcoholic and cholestatic liver diseases, cytokeratin 7 seems to appear earlier than cytokeratin 19 during apparent ductular metaplasia of hepatocytes, whereas during the embryonic and fetal development of the intrahepatic bile ducts, immunoreactivity for cytokeratin 19 precedes that of cytokeratin 7 in cells of the ductal plates.

Hepatocytes forming cholestatic liver rosettes in human livers under conditions of chronic cholestasis have also been considered to be a manifestation of ductular metaplasia of hepatocytes (Butron Vila et al., 1984; Nagore, 1989; Slott et al., 1990; Van Eyken and Desmet, 1992). In cases of primary biliary cirrhosis, cholestatic liver rosettes were demonstrated by electron rnicroscopy to be connected to newly formed atypical ductules by means of a transitional zone containing some cells with ultrastmctural features characteristic of both hepatocytes and intrahepatic biliary epithelium (Nagore et al., 1989). Basement membrane was identified surrounding these intermediate cells. In comparison, Uchida and Peters (1983) reported that in chronic alcoholic liver disease, the ductular hepatocytic cells which intermingled with biliary-appearing cells of proliferated atypical ductules were often surrounded by a poorly developed or incomplete basa1 lamina.

Evidence has been presented suggesting that the cholangiocellular component of human combined hepatocellular-cholangiocarcinomas may also have a hepatocellular lineage (Yano et al., 1986; Fisher et al., 1988). Here, the findings of Yano et al. (1986) are particularly noteworthy. These investigators demonstrated the emergence of adenocarcinoma following subcutaneous inoculation of nude mice with cells of a human hepatocellular carcinoma cell line (KYN-1). The developed tumor, while retaining features of the original primary hepatocellular carcinoma used to generate the cell line, including positive irnrnunohisto- chemical staining for albumin, a-fetoprotein, ferritin and a-1 antitrypsin, also exhibited ultrastructural characteristics of adenocarcinoma with gland formations containing mucicarmine-positive material.

Modulation of hepatocytes to duct cells has been suggested to occur in the livers of adult mice (Leduc, 1959) and guinea pigs (Bathal and Christie, 1969) fed diets containing a-naphthylisothiocyanate. On the other hand, it appears that in rats, bile ductular proliferation associated with the chronic feeding of a-naphthylisothiocyanate is due to multiplication of preexisting bile ductules (Goldfarb et al., 1962; Steiner and Camthers, 1963). However, Richards et al. (1982) have found that 34 to 42 percent of the cells in proliferated ductules induced in rat liver after continuous feeding of 0.03% a- naphthylisothiocyanate for 13 days were irnmunohisto- ~ h ~ m i ~ a l l ~ - ~ o s i ~ v e for the fetal hepatocyte protein, a- fetoprotein. In contrast, typical proliferated ductules appéaring in the livers of Gts at iarious time intervals after bile duct ligation have been shown to be immuno-


histochemically-negative for a-fetoprotein (Yokoyama et al., 1986; Sirica et al., 1991, 1994b; Sirica, 1992). Hyperplastic bile ductular structures occuring in rat liver in response to chronic dietary exposure to 4,4'- diaminodiphenylmethane were also found not to contain immunohistochernically-detectable a-fetoprotein (Sell, 1983). Moreover, Sell, Dunsford, and their associates were unable to repeat the results of Richards et al. and did not find a-fetoprotein in rats fed a-naphthylisothiocyanate (Dunsford et al., 1985; Sell, 1993).

More direct experimental support for ductular metaplasia of hepatocytes comes from the morphological study of Hillan et al. (1989), who assessed the fate of freshly isolated rat hepatocytes following their transplantation into the spleens of syngeneic rats, which at the time of cell transplantation, were also subjected to a common bile duct ligation. Each recipient received an intrasplenic inoculation of 107 cells, having initial viabilities of >80% and a demonstrated hepatocyte purity that was reported to be consistently greater than 98%. Thirty days after the hepatocyte transplantation, between 15% and 50% of the splenic parenchyma of the bile duct-ligated rats was observed to be largely replaced by duct-like structures, with many of the cells forming these structures exhibiting the ultrastructural features of bile duct epithelium. Some of the transplanted cells also had the ultrastructural appearance of hepatocytes, while others possessed cytological features intermediate between hepatocytes and bile duct cells. In contrast, when the hepatocytes were transplanted into the spleens of recipient rats whose livers had been previously subjected to carbon tetrachloride injury with or without an end-to-side portal cava1 shunt, the transplanted cells retained ultrastructural characteristics of normal adult rat '

hepatocytes arranged in the form of irregular sheets surrounded by a fibrous tissue frarnework.

As was the case for combined hepatocellular- cholangiocellular carcinomas in man (see above), it has been inferred that ductular metaplasia of hepatocytic cells takes place in primary adenoid neoplasms induced in the livers of rats treated with N-nitrosomorpholine (Ruan et al., 1989). At least three types of ducts were distinguished at the ultrastructural leve1 in these neoplasms, being characterized as of hepatocellular (type 1), transitional (type 11) and cholangiocellular (type 111) phenotypes, respectively. Cells of the transitional phenotype predominated in the adenoid component of the tumors analyzed in this study. In addition, al1 of the neoplastic ducts with a cholangiocellular phenotype and parts of those ducts composed of transitional cells were observed to give positive immunohistochemical staining when reacted with monoclonal antibody KA-4, which in normal liver was found to selectively react with bile duct-type cytokeratins.

While it has been suggested that ductular metaplasia of hepatocytes contributes largely to the increase in ductular structures appearing in human liver in various longstanding hepatobiliary diseases (Desmet, 1986; Van Eyken and Desmet, 1992). it is also apparent that typical

ductular proliferation can also occur, particularly in the early stages of these diseases (Nakanuma and Ohta, 1986). Moreover, the studies described above in support of the ductular metaplasia concept do not fully rule out the possibility that intermediate or transitional cells observed in such severely injured livers, and in combined hepatocellular-cholangiocarcinomas, might have been derived from the proliferation and differentiation of putative liver stem cells (see below). Even the compelling results of Hillan et al. (1989) are limited by the fact that they are based solely on morphology and do not exclude the possibility that the donor hepatocyte fraction might have been contarninated by rare biliary epithelial cells or putative liver stem-like cells capable of selective proliferation and differentiation along both the biliary and hepatocyte lineages in the spleens of bile duct-ligated recipients. Yano et al. (1986) have also not ruled out the possibility that their KYN-1 hepatocellular carcinoma cell line might have been contaminated with latent adenocarcinoma cells, since carcinoembryonic antigen was detected in the serum and tumors cells of the patient from which the KYN-1 cell line was established. Lastly, it should be cautioned that the expression of bile duct-type proteins or antigens by malignant hepatocytic cells might also be a reflection of aberrant differentiation rather than a «ductular metaplastic» process (Van Eyken and Desmet, 1992).

Oval cell proliferation and the putative hepatlc stem cell

One of the most provocative and potentially important, but still controversia1 issues in hepatic pathobiology concerns the hypothesized existence in adult liver of a putative nonparenchymal stem cell population, which when activated has the potential to differentiate along both the hepatocyte and biliary epithelial cell lineages. In recent years, there has been a considerable amount written about this subject and for a comprehensive and balanced account of the evidence supporting the existence of liver stem cells, the reader is referred to the recent reviews by Se11 and Dunsford (1989), Se11 (1990, 1993, 1994), Sirica et al. (1990), Aterman (1992), Siga1 et al. (1992), Se11 and Pierce (1994), the published proceedings of a 1993 symposium entitled FASEB Meeting in Search of a Hepatic Stem Cell, sponsored by the Society for Experimental Biology and Medicine (papers by Fausto et al., Dabeva et al., Thorgeirson et al., Brill et al., Grisham et al., and Yang et al., 1993, and chapters 5-9 in the Role of Cell Types in Hepatocarcinogenesis (Sirica A.E., ed.) 1992, CRC Press, Boca Raton).

Much of the data favoring the existence of a stem cell population in adult liver has come from studies on the developmental potential of proliferating oval cells induced in rodent livers by various hepatocarcinogenic regimens (Tatematsu et al., 1985; Germain et al., 1988; Evarts et al., 1989; Fausto et al., 1992; Factor and Radaeva, 1993; Sell, 1993) and following certain forms


of noncarcinogenic hepatotoxic injury, such as that produced by the administration of D-galactosamine to rats (Lemire et al., 1991; Dabeva and Shafritz, 1993). The remainder of this section will focus on the relationship of the putative liver stem cell to oval cell proliferation, as well as on the more recent evidence supporting its postulated role as a progenitor cell for hepatocytes and for hepatocellular carcinoma.

The term oval cell was originally coined by Farber (1956) to describe small basophilic epithelial cells with scant cytoplasm and ovoid nuclei, first noted in the livers of carcinogen-treated rats (Sirica et al., 1990; Sell, 1993). While oval cell proliferation has most cornrnonly been observed in rat liver as an early cellular response to many chemical hepatocarcinogenic regimens and to D- galactosamine hepatotoxicity, it has also been demonstrated to be a prominent feature of mouse liver during Dipin-induced hepatocarcinogenesis (Engelhardt et al., 1990; Radaeva and Factor, 1990; Factor and Radaeva, 1993) and to occur in the early stages of hepatocarcinogenesis in the livers of Simian virus 40 large T transgenic mice (Bennoun et al., 1993). In addition, the proliferation of oval cells has been reported to be associated with hepatocarcinogenesis in woodchuck liver infected with woodchuck hepatitis virus (Fu et al.. 1988) and in the livers of estuarine sheep&ead rninnows exPosed to the hepatocarcinogen, dimethvlnitrosamine (Couch and Courtnev. 1987). To date, hówever, there'have been very fe; studies to suggest the presence of oval cells in human liver, but oval-type cells have been described in various cases of nonneoplastic human liver diseases with hepatic injury (Sakamoto et al., 1975; Gerber et al., 1983; Vandersteenhoven et al., 1990; De Vos and Desmet, 1992) and in nontumorous portions of human livers with hepatocellular carcinoma (Sakamoto et al., 1975; Hsia et al., 1992, 1994).

It has been generally proposed that in adult liver, proliferating oval cells are derived from cells of the terminal bile ductules or canals of Hering (Grisham and Hartroft, 1961; Germain et al., 1988; Sirica et al., 1990; Aterman, 1992; Farber, 1992; Lenzi et al., 1992; Yang et al., 1993b; Fausto et al., 1993). This view is strongly supported by the fact that oval cell proliferation originates in the portal zones in the regions of the terminal bile ductules, that proliferating oval cells in large part exhibit ultrastructural features closely resembling those of cells comprising terminal bile duetules (Grisham and Hartroft, 1961; Lenzi et al., 1992; Factor and Radaeva, 1993) as well as express markers of normal intrahepatic bile ductlductular epithelium (Sirica et al., 1990; Lemire et al., 1991; Lenzi et al., 1992), and that they form into irregular ductular-like structures, which in the rat have been demonstrated to connect with preexisting portal bile ducts (Dunsford et al., 1985; Makino et al., 1988). In addition, isolated rat oval cells transplanted into the fat pads of syngeneic rats have been observed to form into bile ductular-like structures by two to three weeks following cell transplantation (Germain et

al., 1988; Fausto et al., 1992). However, oval cells also possess unique phenotypic properties which clearly distinguish them from normal intrahepatic biliary epithelial cells as well as from typical hyperplastic bile ductular epithelial cells (Sirica et al., 1990).

Proliferating oval cells appearing in the livers of rats during the early stages of hepatocarcinogenesis or following D-galactosamine intoxication in actuality represent an expanding heterogeneous population of nonparenchymal liver cells made up of subsets of cells with different developmental potentials. Because of such heterogeneity, Fausto et al. (1992) have used the term oval cell compartment when referring to this type of hepatic cell proliferation. This compartment is comprised of varying proportions of different cell types, including (1) cells that ultrastructurally resemble bile ductular epithelium, (2) transitional cells typically characterized by morphological features that are intermediate between those of intrahepatic biliary epithelial cells and hepatocytes, (3) nondescript cells, and (4) associated stromal cell types, such as desmin- positive Ito cells (Grisham and Porta, 1964; Iwasaki et al., 1972; Se11 and Salman, 1984; Evarts et al., 1990; Sirica et al., 1990; Lemire et al., 1991; Fausto et al., 1993).

One of the more notable phenotypic differences distinguishing the proliferating oval cell compartment of hepatocarcinogen- or D-galactosamine-treated rats from typical bile ductular hyperplasia is the presence in a relatively high proportion of oval cells of the 2.1 -kb fetal liver form of a-fetoprotein rnRNA encoding the 68-to- 70-kDa protein (Lemire and Fausto, 1991; Lemire et al., 1991; Fausto et al., 1992). A subset of oval cells induced in such rat livers have also been demonstrated to give positive immunohistochemical staining when reacted with antibodies against a-fetoprotein (Shinozuka et al., 1978, 1979; Kuhlmann and Wurster, 1980; Sell, 1983; Se11 and Salman, 1984; Yokoyama et al., 1986; Lemire et al., 1991). Variable proportions of rat oval cells have also been shown to contain albumin mRNA (Evarts et al., 1987b; Scoazec et al., 1989; Alpini et al., 1992; Fausto et al., 1992; Dabeva et al., 1993), as well as to irnmunostain positively for albumin (Shinozuka et al., 1978, 1979; Sell, 1983; Se11 and Salman, 1984; Sirica and Cihla, 1984; Yokoyama et al., 1986; Fausto et al., 1992). Moreover, both a-fetoprotein and albumin mRNA transcripts have been detected by in situ hybridization in cells of the atypical ductular-like structures that form in association with hepatocarcinogen- and D- galactosamine-induced oval cell proliferation in rat liver (Tournier et al., 1988; Evarts et al., 1989; Scoazec et al., 1989; Lemire et al., 1991; Alpini et al., 1992). It has been further observed that such ductular-like structures also contain immunohistochemically detected a- fetoprotein and alburnin (Kuhlmann and Wurster, 1980; Sell, 1983; Se11 and Salman, 1984; Dunsford et al., 1985). In addition, double label in situ hybridization for the fetal liver form of a-fetoprotein and for albumin mRNAs have demonstrated that some rat oval cells


simultaneously express both of these mRNA transcripts (Dabeva and Shafritz, 1993). It was also noted by Dabeva and Shafritz (1993) that transitional cells in the rat oval cell compartment contain the fetal liver form of a-fetoprotein mRNA. In comparison, Lemire and Fausto (1991) demonstrated by in situ hybridization that-only a very small number of cells apparently localized in the canals of Hering of normal adult rat liver express 2.1 kb a-fetoprotein mRNA. Alpini et al. (1992) subsequently reported that albumin and a-fetoprotein mRNAs could also be detected by in situ hybridization in rare nonparenchymal cells located in the portal tracts near the limiting plate of normal adult rat liver. Moreover, these investigators reported that after bile duct ligation or a- naphthylisothiocyanate feeding in rats, the expression of a-fetoprotein mRNA was not induced in either parenchymal hepatocytes nor in nonparenchymal liver cells, but like normal liver, occasional a-fetoprotein rnRNA-positive nonparenchymal cells were observed in the portal tracts of these treated animals.

Other «hepatocytic» traits of known function which have been demonstrated to be expressed by subsets of nonparenchymal epithelial cells within the proliferating oval cell compartment of hepatocarcinogen-treated rats include glucose-6-phosphatase (Ogawa et al., 1974; Sirica and Cihla, 1984; Yokoyama et al., 1986; Plenat et al., 1988; Dabeva and Shafritz, 1993), peroxisomal catalase activity (Plenat et al., 1988), a 1 acid glycoprotein (Onda, 1976), and both the adult and fetal liver isozyme forms of pyruvate kinase and aldolase (Hayner et al., 1984; Fausto et al., 1992). Rat epithelial oval cells also exhibit functional traits of normal adult intrahepatic biliary epithelial cells, such as constitutive expressions of bile duct-type cytokeratins 7 and 19 (Lemire et al., 1991; Lenzi et al., 1992; Marceau et al., 1992) and of y-glutamyl transpeptidase (Sirica and Cihla, 1984; Yokoyama et al., 1986; Sirica et al., 1990; Lernire et al., 1991; Fausto et al., 1992; Marceau et al., 1992). A number of antigenic «markers», most of whose functions have as yet to be defined, but which have been selectively demonstrated either in the hepatocytes or intrahepatic biliary epithelium of adult rat liver, have also been shown to be expressed in non parenchymal epithelial cells subsets within the proliferating oval cell compartment (Hixson et al., 1992; Marceau et al., 1992; Yang et al., 1993b; Marceau, 1994). One such antigen, however, designated BD1 which was recently reported by Yang et al. (1993a,b) to be expressed on 60-70% of the intrahepatic bile duct epithelial cells of normal adult rat liver, was further observed by these investigators to be completely absent from proliferated oval cells in the livers of hepatocarcinogen-treated rats. Yang et al. (1993b) also observed the apparent absence of imrnuno- histochemicaiiy-detected BD 1 -positive proliferated bile ductular epithelial cells in adult rat liver by 28 days after bile duct ligation.

Like both the bile ducts of normal rat liver and the typical hyperplastic bile ductules induced in rat liver by various noncarcinogenic cholestatic treatments (Sell,

1983; Sirica et al., 1990, 1992; Sirica, 1992), the atypical bile ductular-like structures formed in rat liver in association with hepatocarcinogen- or D-galactosamine- induced oval cell proliferation have been shown to be surrounded by basement membrane (Sell, 1983; Lemire et al., 1991). However, in contrast to what have been observed for normal rat intrahepatic bile ducts or typical hyperplastic bile ductules (Sell, 1983; Sirica et al., 1990, 1992), the atypical bile ductular-like structures associated with proliferated oval cells in the livers of hepatocarcinogen-treated rats have been reported by Se11 (1983) to have basement membranes that were markedly deficient in their immunohistochemical staining for laminin. On the other hand, the extracellular matrix surrounding zones of oval cell proliferation in these rat livers exhibited strong immunohistochemical staining for fibronectin, whereas that associated with a more typical bile ductular hyperplasia immunostained much less intensely for fibronectin (Sell, 1983). In acute galactosamine hepatitis in rats, fibronectin was also demonstrated to be increased in areas of inflarnmation, along the sinusoids, and in the portal tracts of liver, while immunohistochemical staining for the fibronectin receptor 05 B1 was observed to be decreased in liver at 19 to 48 hours after D-galactosamine administration (Jonker et al., 1992). In a more recent immunohistochemical study of hepatic fibrosis and biliary cirrhosis induced in rats by multiple D-galactosamine injections, Jonker et al. (1994) noted a pattern of injury which resembled the hepatic changes seen after bile duct ligation, namely a prominent bile ductular hyperplasia with an increased immunohistochemical staining for basement membrane laminin and type IV collagen in the areas of proliferated bile ductules. Immunohistochemical staining for fibronectin was also increased along the formed fibrous septa and in contrast to what was observed in the acute D-galactosamine model, the fibronectin receptor a 5 B1 increased in its staining intensity, especially around the proliferated bile ductules, and in parallel with the fibronectin accumulation. While studies such as those just described have focused attention on potentially important differences in basement membrane and extracellular matrix content in relation to oval cell proliferation versus typical bile ductular hyperplasia, it is also evident that additional comparative studies are now required to actually begin to define the role such differences might be playing in regulating the differentiation of nonparenchymal epithelial cell populations within the proliferating oval cell compartment as well as those cells associated with more typical bile ductular hyperplasia. Additional findings relating extracellular matrix content to liver cell differentiation are presented in an upcoming section of this review.

The possibility that some of the nonparenchymal cells of the oval cell compartment have the ability to differentiate into hepatocytes is not a new concept, but one which dates back to the early studies on azo dye hepatocarcinogenesis in rats (Kinosita, 1937; Price et al.,


1952; Farber, 1956). Furthermore, it has been almost four decades since Wilson and Leduc (1958) first postulated the existence in mouse liver of hepatic stem- like cells, which they termed indifferent cholangiole cells, that under conditions of prolonged and severe injury to the liver become activated to proliferate (in the form of oval cell proliferation) and differentiate along both the hepatocyte and biliary cell lineages. Some of the more convincing evidence supporting a precursor- product relationship between certain nonparenchymal cells within the proliferating oval cell compartment and hepatocytes has been derived from a limited number of autoradiographic studies on the fate of [3~]-thymidine- labeled oval cells in rat liver during the early stages of hepatocarcinogenesis (Inaoka, 1967; Evarts et al., 1987a, 1989) and after hepatic injury induced by D- galactosamine (Lemire et al., 1991). However, it should also be noted that the results of such studies have not been universal and that a number of earlier studies (Rubin, 19ó4; Grisharn and Porta, 1964; Se11 et al., 1981; Tatematsu et al., 1984), as well as a very recent study (Gerlyng et al., 1994) on the fate of prelabeled oval cells in rat liver have failed to demonstrate a precursor- product relationship between these cells and hepatocytes. No doubt, such conflicting results are due in part to specific differences in the experimental conditions and time points in which the observations were made. On the other hand, some legitimate concerns have also been raised, particularly with regards to the cited findings of Evarts et al. (1987a, 1989). These investigators provided data to support an apparent conversion by days 11 and 13 of [3~]-thymidine-labeled oval cells into basophilic hepatocytes in the livers of rats that were administered 2-acetylaminofluorene (2AAF) by gavage for two weeks combined with a two-thirds hepatectomy performed at the midpoint of the 2-AAF treatment (modified Solt- Farber hepatocarcinogenesis regimen). However, Farber (1992) has argued that such a finding cannot be considered as being conclusive, since it was not shown that the proliferation of the original parenchymal hepatocytes were completely inhibited at the time apparent conversion of the oval cells to basophilic hepatocytes was first seen. It was further stated by Farber (1992) that zone 1 hepatocytes of rat liver can quickly recover from their mitogenic inhibition with 2- AAF. Gerlying et al. (1994) have also pointed out that one problem with studies based on the transition of [ 3 ~ ] - thymidine-labeled oval cells to hepatocytes in the livers of rats subjected to hepatocarcinogenic, or hepatotoxic treatments is that the proliferation of the original parenchymal hepatocytes is usually not completely inhibited by the growth-suppressive effects of such treatments. In this context, Dabeva and Shafritz (1993) have demonstrated that the restoration of hepatocyte mass following D-galactosamipe injury in rat liver is due in part to the proliferation of preexisting hepatocytes as well as to the conversion of some oval cells into hepatocytes, although it is not possible at present to provide an assessment for this model of how much of the

liver is eregeneratedn from preexisting hepatocytes versus from oval cells differentiating along the hepatocyte lineage. In addition, hepatocarcinogenic or hepatotoxic treatments can be lethal to some of the hepatic cells prelabeled with [3~]-thyrnidine, thereby permitting the possibility of reutilization of label by previously unlabeled hepatocytes that may have recovered from their initial mitogenic block. On the other hand, Lemire et al. (1991) have provided data to rule out the possibility of label reutilization in their carefully conducted analyses of the fate of [ 3 ~ ] - thymidine-labeled oval cells as precursors of small hepatocytes in the livers of D-galactosamine-treated rats. However, an equally well conducted study by Gerlyng et al. (1994) using flow cytometry to assess the fate of bromodeoxyuridine (BrdU) - prelabeled oval cells induced in the livers of rats subjected to the modified Solt-Farber hepatocarcinogenesis regimen, failed over time to demonstrate a significant increase in the labeling index of diploid hepatocytes. Gerlyng et al. observed no detectable shift of BrdU-labeled cells from the proliferating oval cell fraction to the hepatocyte fraction, suggesting that no extensive conversion of BrdU-labeled oval cells to hepatocytes was taking place. Again, a precise analysis of extent to which oval cells might have contributed to the restoration of the hepatocyte mass in rats subjected to the modified Solt-Farber hepatocarcinogenic regimen still needs to be performed.

Further support for a precursor-product relationship between oval cells and hepatocytes is derived from observations of mixed hepatic cell ductular-like structures formed in association with oval cell proliferation in rat liver during azo dye hepatocarcinogenesis (Inaoka, 1967) and D-galactosamine- induced hepatic injury (Lemire et al., 1991). as well as in mouse liver during Dipin-induced hepatocarcinogenesis (Factor and Radaeva, 1993). Ultrastructural analysis of these structures has revealed them to be composed of transitional cells in various stages of maturation along the hepatocyte lineage, together with small hepatocyte- like cells as well as cells with morphological features similar to bile ductular epithelium (Inaoka, 1967; Lernire et al., 1991; Factor and Radaeva, 1993). Tight junctions and desmosomes were seen along the apposing plasma membranes of the small hepatocyte-like cells and neighboring transitional or biliary-like cells, and the biliary-like cells and the majority of transitional cells rested on basement membrane (Inaoka, 1967; Lernire et al., 1991; Factor and Radaeva, 1993). It was further noted by Factor and Radaeva (1993) that mixed hepatic cell ductular-like structures formed in the livers of Dipin-treated mice were either completely surrounded or focally surrounded by basement membrane. Moreover, Lemire et al. (1991) have indicated that in the case of D-galactosamine-induced liver injury in rats, the proliferating ductular-like structures lined by ductular cells, many of which exhibited transitional characteristics, and small hepatocyte-like cells appeared similar to the canals of Hering in that only the ductular


cells were lined by a basement membrane. Novikoff et al. (1991) have demonstrated that a subset of proliferated oval cells with bile ductular features induced in the livers of rats fed DL-ethionine in a choline-deficient diet are frequently encountered in close proximity to hepatocytes. Tight junctions and desmosomes were observed along the contiguous plasma membrane domains of these biliary oval cells and hepatocytes but gap junctions were not seen in the apposed plasma membranes between the two cell types. Interestingly, the biliary oval cells that came in contact with the hepatocytes exhibited an altered polarity, were no longer completely surrounded by basement membrane, and acquired a bile canalicular ATPase activity along their apical cell surface. These findings suggest that interactions with hepatocytes may elicit hepatocytic features in biliary oval cells at the leve1 of the plasma membrane. However, it was not determined if such cells can fully differentiate along the hepatocyte lineage.

While it seems evident that a proportion of nonparenchymal cells associated with oval cell proliferation have the capacity to both proliferate and differentiate into hepatocytes, and thus appear to be acting as hepatic stem-like cells, their actual identity and the exact roles that they may be playing in (1) the replenishment of hepatocytes in the normal versus injured adult liver and (2) in hepatocarcinogenesis remains unresolved and controversial. Severa1 potential hepatic stem cell candidates have been postulated for rodent liver. These include «indifferent» cholangiolar cells (Wilson and Leduc, 1958), nondescript periductular cells (Sell and Salman, 1984), transitional ductule cells (Sell, 1990); rare ductular cells localized in the canals of Hering of normal adult rat liver, which express the 2.1 kb fetal liver form of a-fetoprotein mRNA (Fausto et al., 1993) and embryonic bipotential progenitor cells apparently persisting in very small numbers in the adult liver (Hixson et al., 1992; Marceau et al., 1992; Marceau, 1994). DeVos and Desmet (1992) also presented electron microscopic data demonstrating the presence of small oval cells classified as type 1 cells in human pathological liver specimens with chronic ductular reaction that were associated with cells exhibiting ultrastructural features of bile duct cell differentiation (type 11 cells) together with cells showing morphological evidence of hepatocyte differentiation (type 111 cells). These findings were used to support the possible existence of bipotential progenitor epithelial cells in adult human liver. However, as is the case with al1 of the putative hepatic stem cell candidates listed above, only until such small nonparenchymal cells are isolated in sufficient number and cloned, their phenotypes more completely defined, and their developmental potential experimentally tested, can their possible role as stem-like cells be verified.

That a rare stem cell population may exist in normal adult livers has added credence to the ~strearning liver» hypothesis first proposed by Zajicek et al. (1985). This hypothesis predicts that hepatocytes and intrahepatic

biliary epithelial cells of normal adult liver are continuously being renewed, albeit at a slow rate, through a stem cell pool located in the canals of Hering. The newly formed hepatocytes then mature as they migrate or «stream» from the region of the limiting plate to the central vein, where they are eventually elirninated (Arber et al., 1988). Differentiation along the biliary epithelial cell lineage proceeds in the opposite direction as the cells «stream» toward the interlobular bile ducts (Arber and Zajicek, 1990). A number of recent reports, however, have refuted the streaming liver hypothesis (Slott et al., 1990; Bralet et al., 1994; Grisham, 1994), with much of the evidence against hepatocytes streaming in normal liver having recently been reviewed by Grisham (1994). Of particular significance are the findings of Bralet et al. (1994) who followed the fate of hepatocytes in adult rat liver after they were genetically labeled with the Escherichia coli B-galactosidase gene coupled to a nuclear localization signal. The B- galactosidase gene was introducted by direct in vivo retroviral-mediated gene transfer into hepatocytes 24 hours after partial hepatectomy. It was then observed that after more than one year, the pattern of distribution of B-galactosidase-positive hepatocytes within the hepatic acini did not change from that seen at days 3 and 15 after genetic labeling, strongly suggesting that hepatocytes do not migrate from the portal space to the central vein.

Differentiated rodent hepatocytes have been shown to be capable of cycling multiple times without the involvement of stem cells (Grisharn, 1994; Rihrn et al., 1994) and replicative DNA synthesis in hyperplastic bile ductules appearing in rat liver after bile duct ligation has been demonstrated not to be confined to a specific biliary duct cell (i.e., stem cell) compartment (Slott et al., 1990). Hence, in contrast to the viewpoint held by some (Zajicek, 1991; Siga1 et al., 1992; Brill et al., 1993), it is highly unlikely that activation of a putative liver stem cell is playing much of a role in the renewal of hepatocytes and biliary epithelial cells in normal adult liver or in liver after partial hepatectomy or bile duct ligation. On the other hand, based on the available evidence, it seems much more plausible that in adult liver, the proliferation and activation of a nonparenchymal stem-like population to differentiate along the hepatocyte lineage occurs only when the liver is subjected to severe injury, in which the normal regenerative capacity of preexisting differentiated hepatocytes is greatly impaired. Thus, the notion of a facultative liver stem-like cells as first proposed by Grisham (1980), seems more appropriate when considering the role such precursor cells may be playing in the adult liver.

nontumorigenic epithelial cell lines designated LE12 and LE16 have been established in Fausto's laboratory with oval cells isolated from the livers of rats fed DL-ethionine in a choiine-deficient diet for two and six weeks, respectively (Braun et al., 1987). Neoplastic transformation of cultured LE16 cells was accomplished


by transfection of the H-ras oncogene (Braun et al., 1987; Goyette et al., 1990), as well as occurred spontaneously with infrequent subculture (Braun et al., 1989). In both cases, the neoplastically-transformed LE16 cells were found to give rise to hepatocellular carcinomas when inoculated subcutaneously into nude mice (Braun et al., 1987, 1989; Goyette et al., 1990) or transplanted into the liver or spleen of syngeneic rats (Fausto et al., 1992). In addition, ten separate cell clones were obtained from the untransfected, nontumorigenic LE16 cell line that when transfected with H-ras gave rise to neoplastic cell transformants which when inoculated subcutaneously into nude mice produced poor to well differentiated hepatocellular carcinomas and in the case of one clonal line, a glandular carcinoma with goblet cells (Goyette et al., 1990). The hepatocellular liver carcinomas that developed from the H-ras transfected cells were further shown to express both hepatocyte and oval cell antigens, with trabecular hepatocellular carcinomas having a higher reactivity toward monoclonal antibodies recognizing hepatocyte antigens, while tumors with a glandular architecture reacted predominantly with monoclonal antibodies against antigens expressed in oval cells. Furthermore, a higher proportion of tumor cells from one of the H-ras- transfected clones was found to react in culture with antibodies against hepatocyte antigens when these cells were maintained in a culture medium in which hydrocortisone had been removed, whereas under the same conditions, a concornitant drastic decrease in the proportion of cells reacting with monoclonal antibodies against oval cell antigens could be observed (Goyette et al., 1990).

Neoplastic transformation in cell culture of some rat liver epitheliai cell lines of unknown cellular origin have also been demonstrated to give rise to a range of tumors that can occur in live, including hepatocellular carcinomas, when inoculated into suitable rodent recipients (Tsao and Grisham, 1987; Garfield et al., 1988; Lee et al., 1989b; Tsao and Zhang, 1992; Williams et al., 1992; Grisham et al., 1993). Most notable and best studied among these cell lines is the diploid rat liver epithelial cell line, designated WB-F344, which was first isolated in Grisharn's laboratory from a normal adult rat by primary cloning. The origin of these cells in liver remains unknown, although it appears that they are not derived from adult hepatocytes (Grisham et al., 1993). Moreover, WB-F344 cells have been demonstrated to express phenotypic characteristics that closely resemble ihose of cultured oval cells (Tsao et al., 1984), and which are compatible with their being ~embryonic or undifferentiated» variants of either hepatocytes or intrahepatic biliary epithelial cells (Grisham et al., 1993). Tumors produced in syngeneic rats by chemically or spontaneously transformed WB-F344 cells express a wide range of differentiated states, including hepatocellular carcinomas, adenocarcinomas of biliary or intestinal type, hepatoblastomas, squamous cell carcinomas, poorly differentiated carcinomas, and rnixed

epithelial-mesenchymal tumors (Tsao and Grisham, 1987; Grisham et al., 1993). More recently, Coleman et al. (1993) have reported that following intrahepatic transplantation in adult syngeneic Fischer 344 rats, diploid cultured WB-F344 cells retrovirally transfected with the Escherichia coli B-galactosidase gene became integrated into the hepatic plates and acquired the cell size and nuclear features of mature hepatocytes. These B-galactosidase-positive hepatocyte-like cells were further observed in liver tissue sections to give positive immunohistochemical reactions form albumin, a,-antitrypsin, and tyrosine aminotransferase, as do normal adult rat hepatocytes (Grisham et al., 1993). Furthermore, one of two aneuploid, neoplastically- transformed cell lines derived from WB-F344 cells and subsequently transfected with the bacteria1 B- galactosidase gene (BAG2-GNóTF), when transplanted into adult rat liver lost its tumorigenicity, and yielded instead B-galactosidase-positive cells that become integrated into the hepatic plates in a manner analogous to that observed for the genetically marked untransformed WB-F344 cells described above.

The other neoplastically-transformed cell line (BAG2-GWTB) retained its tumorigenicity in liver, but the cells of the intrahepatic tumors were found to have a more differentiated epithelial morphology. In contrast, both the BAG2-GN6TF and BAG2-GP7TB cells lines produced aggressively growing spindle cell tumors when transplanted subcutaneously into 1-day-old syngeneic Fischer 344 rats.

Results such as those just described for cultured neoplastically transformed LE16 and WB-F344 rat liver epithelial cell lines clearly demonstrate that nonparenchymal liver stem-like cells can serve as the progenitors of hepatocellular carcinoma, as well as to other malignant neoplasms of liver. The pluripotential nature of cells within the proliferating oval cell compartment is further exemplified by the findings of Tatematsu et al. (1985) who demonstrated that oval cells can serve as precursors of metaplastic smali intestinal glands in rat liver. Also, as will be detailed in a later section of this review, proliferating oval cells have been associated with the appearance of metaplastic pancreatic acinar tissue in rat liver.

The actual extent to which liver stem-like cells within the proliferating oval cell compartment of rats are involved in the histogenesis of hepatocellular carcinoma in vivo remains an unresolved issue that still is hotly debated (Sell and Dunsford, 1989; Farber, 1992; Sell, 1993; Tarsetti et al., 1993; Se11 and Pierce, 1994). At one extreme is the stem cell model for liver carcinogenesis (Sell, 1993), which states that al1 liver tumors arise from transformation events that accumulate during the differentiation of hepatocytes or intrahepatic biliary epithelial cells from a common stem cell. The other extreme is that oval cells (or liver stem cells) play no essentiai role in hepatocarcinogenesis and that with most hepatocarcinogens, the cellular origin of hepatocellular carcinoma is the hepatocyte (Farber, 1992; Tarsetti et al.,


1993). Perhaps more accurate is the viewpoint held by Fausto et al. (1992) and others (Faris et al., 1991; Sirica et al., 1991) that while transformed oval cells can give rise to some hepatocellular carcinomas, this does not in any way imply that these cells are the exclusive progenitors of such tumors. Rather, depending on the nature of the hepatocarcinogenic regimen and the type liver injury it produces hepatocytes and nonparenchymal oval cells may both serve as progenitors of hepatocellular carcinomas (Se11 and Dunsford, 1989; Fausto et al., 1992). However, it is not yet known if liver stem-like cells are playing a role in the histogenesis of hepatocellular carcinomas in human liver. In addition, and as will soon become apparent, aberrant differentiation in the liver and biliary tract is not limited to cells derived from the canals of Hering, but has been observed to occur in the larger intrahepatic bile ducts, as well as in extrahepatic bile ducts.

Neocholangiolar proliferation

Neocholangiole is a term first used by Phillips and Poucell (1981) to describe a canal of Hering-like structure observed in human liver in various severe forms of viral- or chemically-induced hepatic injury characterized by a prominent loss of hepatocytes. These unique structures are only observed in pathologic liver states and are seen to best advantage in cases of massive

and submassive necrosis, active cirrhosis, chronic active hepatitis, and alcoholic liver disease (Phillips and Poucell, 1981; Gerber et al., 1983; Phillips et al., 1993; Fig. 1B).

Neocholangioles are lined by a combination of biliary epithelial cells, intermediate or transitional cells (ductular hepatocytes) and more mature hepatocytes, al1 of which are enclosed by a well developed basement membrane (Phillips and Poucell, 1981; Phillips et al., 1987, 1993). It is the presence of a complete basement membrane surrounding both biliary epithelial cells and hepatocyte-like cells that readily distinguishes neocholangioles from the cholangioles or canals of Hering of normal liver, which as already noted have basement membrane only around their biliary epithelial cell component.

Like the atypical ductules found in rodent livers in association with oval cell proliferation, the intermediate cells of neocholangioles have also been demonstrated to exhibit specific phenotypic features of both intrahepatic biliary epithelium and hepatocytes. In cases of massive hepatic necrosis, Gerber et al. (1983) have shown that cells of proliferated neocholangioles occurring at the periphery of necrotic liver lobules gave positive immunohistochemical reactions for both al-antitrypsin, a normal hepatocyte export protein, and carcinoembryonic antigen, a family of glycoproteins expressed by normal intrahepatic bile duct epithelium.

Fig. 1. Light photomicrographs of cholangiolar-like structures composed of both bile ductular epithelial-like cells and single ductular hepatocytic cells (small arrowhead) obsewed in histological sections from (A) rat and (B) human liver under conditions of severe hepatic injury with excessive parenchymal cell loss. The liver tissue section shown in (A) is from the right liver lobe of a Fischer 344 young adult male rat following a short-term period of daily administration of a severely hepatotoxic concentration of furan. Large arrowheads point to a basement membrane completely surrounding the cholangiolar-like structure, which is delineated by its strongly-positive immunohistochemical staining for type IV collagen. Hernatoxylin and eosin counterstain, x 330. In (B) the tissue section is from a 55 year old male with chronic alcoholic liver disease without evidence of viral markers. The ductular hepatocytic cell in this section is stained selectively with periodic acid - Schifi (PAS) stain. Note the PAS-positive stained basement membrane (large arrowheads) supporting both the bile ductular-like epithelial cells and the ductular hepatocytic cell of the cholangiolar-like structure. PD= periductular cells. x 320. The photomicrograph in (A) is from Sirica et al., 1994a and is reproduced with pemission. That in (E) was kindly provided by Dr. Anne-Mane JBzBquel of the University of Ancona, Ancona, Italy.


Interestingly, a-fetoprotein and albumin were not detected by immunohistochemistry in any of the cases tested (Gerber et al., 1983). However, the intermediate cells of neocholangioles were further shown to have ultrastructural features in common with both biliary epithelial cells and hepatocytes (Phillips and Poucell, 1981; Phillips et al., 1987).

The origin of neocholangioles in human liver disease continues to be debated. Thung (1990) has suggested that neocholangioles in massive hepatic necrosis are derived through a metaplasia of damaged hepatocytes. This conclusion was based on her observation that neocholangioles connect to injured hepatocytes and that hepatocytes adjacent to these ductular structures show weak immunohistochemical staining for bile duct-type cytokeratins. In an earlier electron microscopic study of liver biopsies from patients with posthepatitic cirrhosis, Phillips and Steiner (1966) observed tubularization of the hepatic parenchyma with the formation of ductular hepatocytes, also suggesting that neocholangioles developed by ductular metaplasia of hepatocytes. More recently, Meybehm et al. (1993) demonstrated the expression of hepatitis B surface antigen (HBsAg) and hepatic B core antigen (HBcAg) in neocholangiolar epithelium in 15 and 20% of liver biopsies, respectively, from 61 patients with chronic active hepatitis B. In addition, these investigators reported that the neocholangiolar cells did not give positive immunohistochemical staining for proliferating cell nuclear antigen (PCNA), a marker of cell proliferative activity. Citing the strong hepatocytotropism of hepatitis B virus and the lack of detectable PCNA immunostaining in neocholangiolar epithelium as support, Meybehm et al. (1993) also concluded that neocholangioles were derived from hepatocytes undergoing ductular metaplasia.

In contrast, Vandersteenhoven et al. (1990) have provided interesting data to suggest a proliferated bile ductular cell origin of ductular hepatocytes and neocholangiolar structures observed in postrnortem liver of a 66-year-old man who died with liver failure as a consequence of end-stage cirrhosis resulting from chronic hepatitis B infection combined with secondary biliary cirrhosis related to a marked cholelithiasis. Cholangiolar-like structures lined by bile ductular epithelium and ductular hepatocytes were observed to be formed exclusively within areas of intense bile ductular cell proliferation. Moreover, these investigators reported the presence of numerous mitoses in the ductular hepatocytes, but not in parenchymal hepatocytes. It was further demonstrated that the ductular hepatocytes, but not the proliferated bile ductular epithelium, exhibited immunohistochemically-positive staining for HBsAg. On the other hand, the ductular hepatocytes did not exhibit immunoreactivity for albumin, but did show immunohistochernically-positive staining for bile duct- type cytokeratins.

While there have been conflicting reports conceming hepatitis B virus infection of human intrahepatic biliary epithelium (Blum et al., 1983; Niedobitek, 1989), Wang

et al. (1991) have detected the presence of hepatitis Bx antigen in intrahepatic bile duct epithelium of chronic hepatitis B carrier patients with liver cancer and in human cholangiocarcinoma tissues. HBcAg and HBsAg have been recently shown to be present by imrnunohistochemistry in occasional typical bile ductules in cases of acute hepatitis B (Delladetsima et al., 1994). These latter findings add credence to the view that some intrahepatic biliary epithelial cells can be directly infected with hepatitis B virus. Oval-type cells in human livers with hepatocellular carcinoma have also been recently reported to express HBcAg and HBsAg (Hsia et al., 1994).

Rare hepatocyte-like cells have been detected in the interlobular bile ducts of human liver in various liver diseases states, particularly in cases of chronic active hepatitis and chronic persistent hepatitis (Nomoto et al., 1992), as well as in alcoholic liver disease (Nomoto et al., 1992; Ray et al., 1993). Phenotypically, these hepatocyte-like cells were demonstrated by Nomoto et al. (1992) to contain glycogen and to give immunohistochemically-positive staining for albumin and a-1 antitrypsin, whereas the biliary epithelial cells of the interlobular bile ducts were immunohistochemically- negative for these characteristics. However, like intrahepatic biliary epithelium, the hepatocyte-like cells occumng within the interlobular bile ducts were found to give a positive immunohistochemical staining when reacted with antibodies against bile duct-type cytokeratins (Nomoto et al., 1992; Ray et al., 1993), including monoclonal cytokeratin 19 (Nomoto et al., 1992). The appearance of hepatocyte-like cells in the interlobular bile ducts in the portal tracts of liver strongly suggests that bile duct cells are capable of differentiating (or transdifferentiating) into hepatocytes under specific pathologic liver conditions. Experimental evidence supporting the possibility of a biliary cell origin of neocholangioles is presented below.

Novel experimental animal models of ductular hepatocyte development

Two novel rat models of severe hepatic injury were recently described by the author and his associates (Sirica et al., 1992, 1994a; Sirica and Williarns, 1992), which provide strong support for a hyperplastic bile ductular cell origin of ductular hepatocytes. Both animal models were developed in young adult Fischer 344 male rats, and each was characterized by a massive loss of parenchymal hepatocytes and by a prominent bile ductular cell proliferation. Not surprisingly, both rat models also exhibited high degrees of mortality with 10- 30% of the rats surviving long enough in each case for ductular hepatocytes to appear in their livers.

The first of these models combined a prior typical bile ductular cell hyperplasia that followed four to six weeks of ligation of the common bile duct with a subsequent zone 3 hepatonecrosis induced by a single necrogenic treatment with CC14. The most notable


consequence of the bile duct ligation/CC14, treatment was the distinct, albeit rare appearance in the livers of moribund rats of cholangiolar-like structures comprised of biliary epithelial cells together with one or more hepatocyte-like cells in various stages of maturation. These cholangiolar-like structures were obsewed to form only in areas of preexisting hyperplastic bile ductular tissue of the bile duct ligated/CC14-treated rats and were not detected in histological sections from the livers of control rats that were subjected to either bile duct ligation alone or to CC14-induced hepatonecrosis without prior bile duct ligation.

The hepatocyte-like cells of the cholangiolar-like structures exhibited cell and nuclear diameters that were intermediate in size between those of hyperplastic bile ductular epithelium and parenchymal hepatocytes. In addition, such hepatocyte-like cells could be readily distinguished from the biliary epithelial cell components of the same cholangiolar-like structure, as well as from epithelial cells lining typical hyperplastic bile ductules present in the same liver section by their strongly eosinophilic cytoplasm, their vesicular nucleus with prominent nucleolus, and their selectively-positive immunohistochemical staining for the hepatocyte cytoplasmic antigen H.4. (Sirica et al., 1992; Sirica and Williams, 1992). However, it is also noteworthy that, in contrast to hepatocarcinogen- or D-galactosamine- induced rat oval cell subpopulations, but like neocholangiolar cells analyzed by Gerber et al. (1983) in massive hepatic necrosis of human liver, a-fetoprotein was not detected by imrnunohistochemistry in any of the cells of the cholangiolar-like structures nor in typical hyperplastic bile ductules observed in histological liver sections from the bile duct-ligation/CC14-treated rats.

In the second model, rare, but distinct cholangiolar- like structures composed of biliary epithelial cells and typically a single hepatocyte-like cell in various stages of maturation (Fig. 1A) appeared in association with an extensive proliferative bile ductular reaction, which was particularly prominent within the atrophied right liver lobe of Fischer 344 rats that has been subjected to a severely hepatotoxic treatment with furan (Sirica et al., 1994a). The morphology of the cholangiolar-like structures that formed in the livers of the furan-treated rats closely resembled that of neocholangioles described for human liver in various hepatic disease states, including alcoholic liver disease (Fig. l), and were very similar in their appearance to the cholangiolar-like structures seen in the livers of the bile duct-ligated/CClg- treated rats. The cholangiolar-like structures occumng in the livers of the furan-treated rats also did not appear to express a-fetoprotein (Elmore and Sirica, 199 1; Sirica et al., 1994a). Moreover, these cholangiolar-like structures were shown to be completely surrounded by a well developed basement membrane that gave strong immunohistochemically-positive staining reactions for larninin and for type IV collagen (Fig. lA, Elmore and Sirica, 1991 ; Sirica, 1992; Sirica et al., 1994a). Al1 of the biliary epithelial cells and ductular hepatocyte-like cells

formed in the livers of the furan-treated rats exhibited a strongly-positive immunohistochemical staining for y- glutamyl transpeptidase and for cytokeratin 8. Bile duct- type cytokeratin 19 was further found to be characteristically expressed in al1 of the biliary epithelial cells in this model, but was also detected by immunohistochemistry in some of the ductular hepatocyte-like cells within cholangiolar-like stnictures (Sirica et al., 1994a). However, only the ductular hepatocyte-like cells were observed to be immunohistochemically-positive for albumin and to contain peroxisomes in their cytoplasm. Cell junctions were demonstrated by electron microscopy to be prominent between individual ductular hepatocyte-like cells and adjacent biliary epithelial cells of the same cholangiolar- like structures. Finally, the range of cell and nuclear diameters determined for the ductular hepatocyte-like cells was consistent with a progressive maturation process, with the smallest ductular cells having transitional features being somewhat larger in size (12- 13 pm) than the largest biliary epithelial cells; the majority of ductular hepatocyte-like cells having diameters (14-17 pn) of small hepatocytes, and very few of the ductular hepatocyte-like cells having cell diameters (18-20 pn) closely approximating those of normal adult rat parenchymal hepatocytes (Sirica et al., 1994a).

The results obtained from both the bile duct ligated/CC14-treated and furan hepatotoxicity rat models of severe hepatic injury and massive hepatocyte loss demonstrate the origin of hepatocyte-like cells within cholangiolar like structures and strongly suggest that they have arisen as a result of a differentiation (or transdifferentiation) of rare proliferated bile ductular-like epithelial cells to transitional ductular cells to more mature ductular hepatocytes under conditions in which the normal regenerative capacity of parenchymal hepatocytes had become greatly impaired. These experimental findings on the origin of hepatocyte-like cells within intrahepatic biliary structures also closely relate to those of Vandersteenhoven et al. (1990) and of Nomoto et al. (1992) for human liver, as described in the preceding section.

The ductular cells of rat and hamster pancreas are likewise capable of altering their normal differentiation commitment so as to serve as precursor cells of pancreatic hepatocytes is well documented (Rao et al., 1989, 1991; Scarpelli et al., 1989; Makino et al., 1990). Particularly relevant to the observations described above for liver are those of Rao et al. (1989, 1991) demonstrating in Fischer 344 male rats the differentiation of pancreatic ductular (and periductular cells) into pancreatic hepatocytes under conditions of massive acinar cells loss and subsequent ductular proliferation induced by a copper depletiodrepletion protocol. It is noteworthy that the proliferated pancreatic ductular cells in the Rao et al. model of massive acinar cell loss appeared to selectively differentiate along the hepatocyte lineage rather than pancreatic acinar cell


lineage. In furan-treated rats, the appearance of hepatocyte-like cells in cholangiolar-like stmctures also seemed to be linked to the severity of the liver injury and degree of parenchymal cell loss (Sirica et al., 1994a). When less toxic, nonlethal doses of furan were given, a different pattern of aberrant intrahepatic biliary cell differentiation was observed, namely the preferential development of small intestinal-like glands whose cellular composition closely resembled that of the crypts of Lieberkühn of normal adult rat small intestine (Elmore and Sirica, 1991, 1992, 1993; Sirica, 1992; Sirica et d. , 1992). Specifically, these metaplastic small intestinal glands were demonstrated (1) to form almost exclusively within the right and caudate liver lobes of the furan treated rats (Elmore and Sirica, 1991, 1992; Sirica et al., 1992), (2) to be composed of well differentiated tal1 colurnnar enterocytic-like absorptive cells, goblet cells, Paneth cells and neuroendocrine cells, which differentiated in a sequential manner from hyperplastic bile ductular precursor cells (Elmore and Sirica, 1992; Sirica, 1992), and (3) to represent an early change related to the histogenesis of primary hepatic «intestinal-type» adenocarcinomas, which subsequently developed in the right and caudate liver lobes following longer term exposures of the rats to furan (Elmore and Sirica, 1993). Cholangiolar-like structures containing ductular hepatocytes were not apparent in the livers of rats subjected to nonlethal exposures to furan. Conversely, metaplastic small intestinal glands were not seen in histological liver sections from rats exhibiting greater parenchymal hepatocyte loss in response to severely hepatotoxic furan treatment, although it should be noted here that such animals did not survive long enough for such metaplastic intestinal structures to develop. Interestingly, it was also observed that when rats were administered furan by gavage at a dose of 45 mgkg body weight, once a day, five times a week for six weeks beginning one week after a bile duct ligation, almost al1 of their livers became replaced with typical hyperplastic bile ductular tissue without evidence of small intestinal mucosa1 cell differentiation or ductular hepatocyte development (Sirica et al., 1994b).

Intestinal metaplasia has been demonstrated in both the intra- and extrahepatic biliary tract of humans in various non-neoplastic and neoplastic conditions (Kozuka et al., 1984; Kurumaya et al., 1990; Sirica, 1992; Duarte et al., 1993), as well as in hamsters following different cholangiocarcinogenic treatments (Moore et d., 1986; Tharnavit et al., 1987; Kinami et al., 1990). In addition, exocrine pancreatic tissue has been reported to develop in human liver in association with small proliferating bile ductules in a case of severe posthepatitic cirrhosis (Wolf et al., 1990). «Metaplastic» pancreatic acinar tissue has also been observed in association with oval cell proliferation in the livers of rats exposed to polychlorinated biphenyls (Kimbrough, 1973; Rao et al., 1986; Scarpelli et al., 1989) and in association with neoplastic biliary cells andlor hepatocellular carcinoma that had been induced in trout

livers by dimethylnitrosarnine (Lee et al., 1989a). More recently, Terada and Nakanuma (1993) have observed in some human livers that the intrahepatic peribiliary glands located around the large intrahepatic bile ducts near the hepatic hilum are associated with the development of ectopic exocrine pancreatic tissue. Apparent differentiation of peribiliary glands into pancreatic acini was first observed at three months after birth and found to persist during childhood and in adult life.

Growth factors, hormones, and stromal-epithelial interactions in oval cell and intrahepatic biliary epithelial cell proliferation and developmental potential

Growth factors, hormones, and stromal-epithelial interactions are al1 known to play key roles in the regulation of cell proliferation and morphogenesis. In the case of carcinogen-induced oval cell proliferation in rat liver, mRNA transcripts for hepatocyte growth factorlscatter factor (HGFISF) have recently been demonstrated by in situ hybridization to be restricted to increasing numbers of desmin-positive Ito cells located in the imrnediate vicinity of proliferating oval cells, but were not detected in oval cells themselves (Alison et al., 1993b; Hu et al., 1993). On the other hand, mRNA transcripts for the c-met proto-oncogene encoding the HGFISF receptor were localized by in situ hybridization to oval cells, but were not detected in Ito cells distributed around the oval cells (Hu et al., 1993). In situ hybridization has further revealed the presence of mRNA transcripts for transforming growth factor-a (TGF-a), acidic fibroblast growth factor (aFGF) and stem cell factor (SCF), respectively, in both oval cells and Ito cells proliferating in the livers of carcinogen- treated rats (Evarts et al., 1992, 1993; Thorgeirsson et al., 1993; Fujio et al., 1994) whereas c-kit transcripts encoding the SCF receptor were reported to be only observed in oval cells and not seen in Ito cells (Fujio et al., 1994). Proliferating rat oval cells have also been shown to exhibit strong imrnunohistochernically-positive staining for TGF-a (Alison et al., 1993a) and more recently, human oval-type cells have been found to express TGF-a (Hsia et al., 1994). Transforrning growth factor-Bl (TGF-BI) mRNA was localized by in situ hybridization mainly to Ito cells that had proliferated along with oval cells in rat liver during the early stages of hepatocarcinogenesis, while most oval cells did not contain detectable TGF-Bl transcripts (Evarts et al., 1990). However, rat oval cells in culture have been reported to express high levels of TGF-B, mRNA (Fausto et al., 1992).

HGFISF was not detected by immunohistochemistry in either the intrahepatic biliary epithelium of normal rat liver (Wolf et al., 1991) nor in typical hyperplastic bile ductular epithelium induced in rat liver following bile duct ligation or bile duct ligation in combination with daily exposures to furan (Sirica et al., 1994b). In


contrast, bile duct epithelium in normal human liver sections were observed to exhibit a moderate immunoreactivity for HGFISF, and normal human gallbladder epithelium were shown to be strongly- positive for HGFISF (Wolf et al., 1991). The c-met encoded HGFISF receptor protein was also detected by immunohistochemistry in the major biliary ducts of normal human liver (Prat et al., 1991).

TGF-a immunoreactivity has recently been demonstrated in the bile duct epithelium of both normal rat liver (Burr et al., 1993; Burt and Mac Sween, 1993) and in models of regenerating rat liver (Burr et al., 1993; Alison et al., 1993a), as well as in normal human intrahepatic bile duct epithelium and within proliferating bile ductules of human cirrhotic livers (Collier et al., 1993). Epidermal growth factor receptors (receptors also for TGF-a) have been detected on the plasma membrane of normal adult rat intrahepatic biliary epithelial cells (Ishii et al., 1990). Moreover, the neoplastic epithelium of some human intrahepatic cholangiocarcinomas have been shown to be immunohistochemically-positive for epidermal growth factor (EGF) and its receptor (EGF- R), although it was further stated that an autocrine model of intrahepatic biliary tumor growth as suggested by the combination of EGF and EGF-R may only be applicable to a very limited number of cases of cholangiocarcinomas (Nonomura et al., 1988).

TGF-B mRNA was not detected by Northern blot analysis in biliary epithelial cells isolated from the livers of bile duct-ligated rats (Braun et al., 1988). Similarly, in normal and regenerating rat liver, as well as in the Solt- Farber rat model of hepatocarcinogenesis, TGF/Bl mRNA transcripts were shown by in situ hybridization to be absent in the intrahepatic bile ducts, but were present in periductal cells of the surrounding connective tissue (Evarts et al., 1990; Nakatsukasa et al., 1990; Milani et al., 1991). In situ hybridization has also revealed that TGF-Bl mRNA transcripts were at best, only weakly expressed in occasional hyperplastic bile ductular epithelial cells appearing in the livers of bile duct-ligated rats and in human fibrotic livers. In contract, TGF-B2 mRNA transcripts were detected at high levels in the proliferating bile ductules in both human and rat liver (Milani et al., 1991).

Proliferating bile ductules induced in rat liver following bile duct ligation were further observed not to immunoreact with a rabbit polyclonal antibody generated from a synthetic peptide containing arninoacid residues 78-109 of human mature TGF-BI (Sirica et al., unpublished data). However, a rabbit polyclonal antibody specific for TGF-Bl-latent-associated protein was very recently shown by Saperstein et al. (1994) to weakly react with the intrahepatic biliary epithelium of sham-operated control rats and to strongly react with proliferating bile ductules in the livers of bile duct- ligated rats. These hyperplastic bile ductules were also found to exhibit a strongly-positive immunohisto- chernical staining for the mannose 6-phosphatelinsulin- like growth factor-11 receptor, which has been shown to

facilitate the proteolytic activation of TGF-BI (Saperstein et al., 1994).

The metaplastic small intestinal-like glands derived from biliary cell precursors in the livers of furan-treated rats were recently demonstrated to selectively immunoreact with antibody against HGFISF (Sirica et al., 1994b). Of potentially greater interest are the results of more recent preliminary immunochemical studies performed in the author's laboratory, which have revealed that the neoplastic glands of primary hepatic «intestinal-type» adenocarcinomas induced in rats following long-term furan exposure are strongly immunoreactive for HGFISF, c-met receptor protein, TGF-Bl and mannose 6-phosphatelinsulin-like growth factor-11 receptor (Sirica et al., unpublished data). TGF-a was not detected by immunohistochemistry in these biliary cell-derived tumors, but was observed in two primary hepatocellular carcinomas that developed in the furan-treated rats (Elmore and Sirica, 1993).

Roskams e t al. (1990) have demonstrated that reactive human bile ductules express neuroendocrine features, including immunoreactivity for chromogranin A and ultrastructural evidence of neuroendocrine granules. Subsequently, this group of investigators have reported that the majority of reactive human bile ductular cells expressing neuroendocrine features also were immunoreactive for parathyroid hormone-related protein (Roskams et al., 1993b). In addition, it was demonstrated that al1 human primary cholangiocarcinomas exarnined showed immunohistochemically- positive staining for both chromogranin A and for parathyroid hormone-related protein (Roskams et al., 1993a). In this context, it is noteworthy that parathyroid hormone-related protein has been described as being an incomplete mitogen requiring the presence of EGF or TGF-B to stimulate cell proliferation (Burt and Mac Sween, 1993; Roskams et al., 1993b).

Although mostly descriptive in nature, the studies just described provide some clues as to what growth factors might be important in the regulation of oval and bile ductular epithelial cell growth and differentiation. Functional studies in this area are limited, but those that have been performed have yielded results that are consistent with the above findings. For example, the addition of EGF in medium containing 1% fetal calf semm to primary cultures of bile ductular epithelial cells from bile duct-ligated rats was observed after a six day lag period to greatly enhance the rate of replicative DNA synthesis in these cells over that measured in solvent control cultures (Mathis and Sirica, 1990). Proliferative growth of a human cholangiocarcinoma cell line (CC- LP-1) in culture was also reported to be significantly augmented by EGF, whereas in the same study, EGF was found to be without effect in stimulating the in vitro growth of a second human cholangiocarcinoma cell line designated CC-SW-1 (Shimizu et al., 1992). EGF was further demonstrated to markedly stimulate growth of the LE16 rat oval cell line in soft agar, but only after this cell line reached its 50th passage. Early passages of


these cells formed no true colonies in agar containing EGF. Moreover, while the late passage LE16 oval cells exhibited an altered sensitivity to the mitogenic action of EGF in soft agar, these cells were not tumorigenic when injected into nude rnice (Braun et al., 1987). In addition, during the process of neoplastic transformation in cell culture, rat oval cells were shown to progressively lose their sensitivity to TGF-Bl-mediated inhibition of cell growth. In this regard, replicative DNA synthesis is fully transformed cultured oval cells was not inhibited by TGF-Bl and may have even been stimulated by it (Fausto et al., 1992).

Untransformed WB-F344 rat liver epithelial cells have also been demonstrated to respond positively to EGF and TGF-a and to be growth inhibited by TGF-B (Tsao et al., 1986; Lin et al., 1987; Earp et al., 1988; Grisham et al., 1993). These cells were further demonstrated to express numerous high affinity receptors for EGF (Tsao et al., 1986; Grisham et al., 1993). In addition, TGF-a was found to be overexpressed in chemicaliy-transformed WB-F344 cells in culture, which correlated with tumorigenicity among those clones in which c-myc was also overexpressed (Lee et al., 1991). EGFR rnRNA was further determined to be slightly decreased in most of the transformed clones compared to the parental cells. However, it was suggested that EGFR levels in the transformed clones were likely sufficient to support a TGF-aIEGFR autocrine growth cycle (Lee et al., 1991). In contrast, spontaneously transformed rat liver epithelial cell lines were shown to express variable amounts of TGF-a mRNA, which was not functionally coordinated with its receptor, EGFR (Tsao et al., 1990). More recently, Tsao and Zhang (1992) observed that EGF enhanced the tumorigenicity of spontaneously transformed rat liver epithelial cells and that the tumors derived from the EGF-treated transformed cells were highly metastatic mucus-secreting ductal adenocarcinomas. In contrast, cells that spontaneously transformed in culture in the absence of chronic EGF exposure formed nonmetastatic tumors with hepatocellular differentiation.

TGF-a was not detected in conditioned media from cultures of untransformed WB-F344 rat liver epithelial cells, but transformed clonal cell strains derived from WB-F344 secreted TGF-a-like activity into their culture media (Liu et al., 1988). In comparison, both untransformed and transformed WB-F344 cells typically produced the inactive form of TGF-B (Liu et al., 1988). However, only tumorigenic cell lines tended to produce activated TGF-B de novo and anchorage-independent growth of those cell lines which elaborated active TGF-B was either stimulated, inhibited or unaffected by TGF-B. Furthermore, cell lines that were inhibited by TGF-B concurrently produced TGF-a-like activity, which in most cases was able to overcome the growth inhibition mediated by TGF-B.

In another study that utilized a different rat liver epithelial cell line, Gupta et al. (1992) reported that hepatic stimulator substance was also able to abrogate

the inhibitory effect of TGF-Bl in cell culture. Thus, as suggested by Liu et al. (1988), the main physiological role of TGF-B when produced in its active form by tumorigenic cells may be to act either as an autocrine growth stimulator or more likely, to function in a paracrine manner. The presence of a strong immunoreactivity for TGF-Bl and the mannose 6-phosphatel insulin-like growth factor 11 receptor in proliferating neoplastic glands of primary ((intestinal-type» adenocarcinomas formed in the livers of furan-treated rats (see above), together with the abundant connective tissue stroma associated with this type of biliary cell- derived malignant neoplasm, is consistent with such a paracrine function for TGF-BI in this tumor. Here, it is also noteworthy that a human cholangiocellular cell line (Hu CC-TI) has been demonstrated to produce a basic fibroblastic growth factor-like factor, which was highly mitogenic for murine BALBIC 3T3 fibroblasts, but was without effects as an autocrine growth factor for Hu CC- TI cells in culture (Matsuzaki et al., 1990).

HGFISF has recently been reported to be a potent mitogen for cultured human intrahepatic biliary epithelial cells (Joplin et al., 1992). However, to the author's knowledge, there are to date no published studies indicating if HGFISF can also act as a mitogen for cultured biliary epithelial cells isolated from the livers of either normal or bile duct-ligated rats.

In addition to its mitogenic function, HGFISF has been demonstrated to induce tubulogenesis in vitro in a number of untransformed and malignat-transformed cell lines (Tsarfaty et al., 1992; Santos et al., 1993; Santos and Nigam, 1993; Morimoto et al., 1994). For Madin- Darby canine kidney cells grown in Type 1 collagen gels, it was further demonstrated that the addition of laminin, entactin and fibronectin facilitated the HGFISF-mediated formation of tubular structues, while type IV collagen, heparin sulfate proteoglycan, vitronectin and TGF-B inhibited HGFISF induced tubulogenesis (Santos and Nigam, 1993).

Of particular relevance are the results of a recent study performed by Johnson et al. (1993) on the differential effects of HGFISF on the morphogenesis of two murine nonparenchymal epithelial liver cell lines cultured in type 1 collagen gels. Histological sections of colonies of the BNL CL.2 cell line, originally derived from embryonic mouse liver, growing in collagen gels in the presence of HGFISF were reported to morphologically resemble neocholangioles. In comparison, sections of HGFISF-treated colonies of the NMuLi cell line, initially generated from postnatal mouse liver, were obsewed to closely resemble in their appearance ductules proliferating in conditions of biliary tract obstruction. Sakata et al. (1994) have more recently reported that FK506, a potent immunosuppressive agent with HGF-like activity appeared to induce the proliferation and apparent development of biliary epithelial cells from hepatocytes transplanted into the spleen of syngeneic rats. However, as was the case for the morphological study of Hillan et al. (1989)


previously described in this review, it is not known if the biliary epithelial cells observed following hepatocyte transplantation into the spleens of the FK506-treated rats were due to a selective proliferation of a small amount of biliary epithelial or putative liver stem-like cell contaminants within the isolated hepatocyte suspension or were the result of ductular metaplasia of the transplanted hepatocytes.

Cultured oval cells derived from the livers of 3'- methyl-4-dimethylaminoazobenzene-treated rats have been demonstrated to produce large amounts of albumin, to cease a-fetoprotein production, and to exhibit a progressive induction of the hepatocyte enzyme tyrosine aminotransferase when maintained in a medium containing sodium butyrate and dexamethasone (Germain e t al., 1988). However, the induction of HESó, a rat hepatocyte lineage marker antigen, was not detected in these cells, nor did they show morphological features of differentiated hepatocytes (Germain et al., 1988; Marceau et al., 1992). Albumin mRNA transcripts were detected by in situ hybridization in rat pancreatic epithelial «oval» cell line cultured on or inside type 1 collagen gels, but such cells when grown in collagen gels in the presence or absence of fibroblasts did not show morphological differentiation into hepatocytes, acinar cells, or islet cells (Ide et al., 1993). In contrast, Lázaro and Fausto (1994) recently presented preliminary findings indicating that both the LE16 and LE12 rat oval cell lines can serve as precursors for hepatocyte-like cells forming in vitro when they have been cocultured with fibroblasts in a tridimensional collagen gel system. Rat embryonic E15 liver cells have also been demonstrated to differentiate in primary culture along the hepatocyte lineage when exposed to TGF-B and to insulin-like growth factor, and along the biliary epithelial cell lineage when cultured in the presence of tryptose phosphate broth (Marceau et al., 1992). E12 rat liver cells exposed to sodium butyrate in primary culture gave rise to cells expressing biliary epithelial features, whereas those cultured in serum exhibited hepatocytic properties (Germain et al., 1988; Marceau, 1994). Shiojiri and Mizuno (1993) have further demonstrated in primary organ cultures of fetal mouse liver fragments without evidence of bile ducts, that the addition of dexamethasone to the medium stimulated the development of both mature hepatocytes and bile duct epithelium. Interestingly, when the immature liver fragments were cultured on substrata such as collagen gel, Millipore filter or Spongel, the presence of dexamethasone dramatically stimulated glycogen storage but did not induce bile duct differentiation. On the other hand, when the fragments were cultured on basement membrane Matrigel, dexamethasone stimulated the expression of bile duct markers. Here, it is also noteworthy that laminin has been demonstrated to accompany the development of intrahepatic bile ducts during successive stages of human liver development (Shah and Gerber, 1990).

Conclusion and perspective

It is now apparent that parenchymal hepatocytes and nonparenchymal epithelial cells of the intrahepatic biliary tract are both capable of demonstrating remarkable plasticity in response to certain forms of severe hepatic injury and cell loss. Depending on the condition, parenchymal hepatocytes appear to be capable of altering their differentiation state so as to undergo a metaplastic conversion into biliary epithelial cells. Conversely, some biliary epithelial cells appear capable of altering their differentiation commitment so as to give rise to either hepatocyte-like cells, small intestinal mucosa cell types, or pancreatic acinar cells. In addition, a substantial body of evidence now exists, derived largely from studies of proliferating rat oval cells and of certain rat liver epithelial cell lines with oval cell properties, to support the existence in liver of a facultative pluripotent stem-like cell compartment that becomes activated under certain conditions of carcinogen- and toxin-induced hepatic injury in which the normal regenerative capacity of liver has become greatly impaired. However, the actual identity of such a facultative liver stem-like cell and i ts specific localization in normal adult liver still remains as a matter of conjecture and controversy. Moreover, that hepatocytes can form in the interlobular bile ducts of diseased human liver, that the peribiliary glands around the large hilar bile ducts of human liver can apparently give rise to metaplastic pancreatic acinar tissue, and that under conditions of chronic persistent injury and proliferation, epithelial cells of both the larger intrahepatic as well as extrahepatic bile ducts of humans can undergo a «metaplastic» conversion to small intestinal mucosa1 cell types al1 point to the fact that such dramatic alterations in biliary cell differentiation are not compartmentalized solely to cells associated with the terminal bile ductules or canals of Henng of liver.

Implicit in these unique patterns of aberrant cell differentiation is the likelihood that such changes, when associated with carcinogenic conditions, are important in the histogenesis of different types of pnmary hepatic and biliary tumors, such as hepatoblastoma, hepatocellular- cholangiocarcinoma, «intestinal type» cholangiocarcinoma or adenocarcinoma, composite or mixed glandular-neuroendocrine carcinoma of the extrahepatic biliary tract (Hsu et al., 1991; Nishihara et al., 1993), and possibly primary carcinoid or carcinoid-like tumors (Karaki et al., 1991) of liver and of bile ducts. Furthermore, that a nonparenchymal cell population located at the leve1 of the terminal bile ductules of adult liver is capable of being activated to proliferate and differentiate along both the biliary and hepatocyte lineages suggest the possibility of using such cells to develop novel strategies for hepatic gene therapy as well as to potentially rescue liver in some cases of fulminant hepatic failure.

Much of our current understanding of the development of ductular hepatocytes in both


experimental animal and human livers has been based in large part on data derived from phenomenological studies. New experimental approaches are now needed employing various clonally expanded liver cell populations of known origin and with identifying phenotypic features in order to more clearly establish the nature of the putative stem-like cell of liver, as well as to define mechanisms governing the differentiation of select biliary epithelial (and oval cell) types along either the hepatocyte, small intestinal mucosa1 cell or pancreatic acinar cell lineages. Further development of novel in vitro and cell transplantation models, such as those described in this review, is also likely to provide potentially powerful tools for investigating how specific growth factors, stromal elements, hormones and other differentiation factors may be interacting to change the differentiation comrnitment of various hepatic epithelial cell types so as to give rise to ductular hepatocytes and other nonliver epithelial ceel types of gut endodermal derivation.

- -

Acknowledgements. The author wishes to thank Ms. Teresa W. Gainey and Ms. Sara L. Cole for their superb technical assistance in carrying out the experimental studies with furan described in this review and to Ms. Rhonda C. Jackson for typing the manuscript. This publication was supported by grant 5 R o l CA 39225 to A.E.S. from the National Cancer Institute, NIH.

References

Alison M.R., Nasim M.M., Anilkumar T.V. and Sarraf C.E. (1993a). Transforming growth factor-a immunoreactivity in a variety of epithelial tissues. Cell Prolif. 26, 449-460.

Alison M.R., Poulsom R., Jeffery R., Anilkumar T.V., Jagoe R. and Sarraf C.E. (1993b). Expression of hepatocyte growth factor mRNA during oval cell activation in the rat liver. J. Pathol. 171, 291-299.'

Alpini G., Aragona E., Dabeva M., Salvi R., Shafritz D.A. and Tavoloni N. (1992). Distribution of albumin and alpha-fetoprotein mRNAs in normal. hyperplastic, and preneoplastic rat liver. Am. J. Pathol. 141, 623-632.

Arber N. and Zajicek G. (1990). Streaming liver VI: Streaming intrahepatic bile ducts. Liver 10,205-208.

Arber N., Zajicek G. and Ariel 1. (1988). The streaming liver II. Hepatocyte life history. Liver 8, 80-87.

Aronson D.C., De Han J., James J., Bosch K.S., Ketel A.G., Houtkooper J.M. and Heijmans H.S.A. (1988). Quantitative aspects of the parenchyma-stroma relationship in experimentally induced cholestasis. Liver 8, 1 16-1 26.

Aterman K. (1992). The stem cells of the liver - a selective review. J. Cancer Res. Clin. Oncol. 1 18, 87-1 15.

Ayres R.C.S., Hübscher S.G., Shaw J., Gamer C., Joplin R., Williams A. and Neuberger J.M. (1991). New monoclonal antibodies reacting with bile ducts: further insights into the pathogenesis of bile ductular proliferation in biliary diseases. J. Pathol. 165, 153-161.

Bennoun M., Rissel M., Engelhardt N., Guillouzo A., Briand P. and Weber-Benarous A. (1993). Oval cell proliferation in early stages of hepatocarcinogenesis in Simian Virus 40 large T transgenic mice. Am. J. Pathol. 143, 1326-1336.

Bhathal P.S. and Christie G.S. (1969). A fluorescente microscopic study of bile duct proliferation induced in guinea pig by a-naphtyl isothiocyanate. Lab. Invest. 20, 480-487.

Bathal P.S. and Gall J.A.M. (1985). Deletion of hyperplastic biliary epithelial cells by apoptosis following removal of the proliferative stimulus. Liver 5, 31 1-325.

Blum H.E., Stowring L., Figus A., Montgomery C.K., Haase A.T. and Vyas G.N. (1983). Detection of hepatitis B virus DNA in hepatocyies, bile duct epitheliurn, and vascular elements by in situ hybridization. Proc. Natl. Acad. Sci. USA 80,6685-6688.

Bralet M-P., Branchereau S., Brechot C. and Ferry N. (1994). Cell lineage study in the liver using retroviral mediated gene transfer- Evidence against the streaming of hepatocytes in normal liver. Am. J. Pathol. 144, 896-905.

Braun L., Goyette M., Yasween P., Thompson N.L. and Fausto N. (1987). Growih in culture and tumorigenicity after transfection with the ras oncogene of liver epithelial cells from carcinogen-treated rats. Cancer Res. 47,4116-4124.

Braun L., Mead J.E., Panzica M., Mikumo R., Bell G.I. and Fausto N. (1988). Transforming growth factor B mRNA increases during liver regeneration: A possible paracrine mechanism of growth regulation. Proc. Natl. Acad. Sci. USA 85, 1539-1543.

Braun L., Mikumo R. and Fausto N. (1989). Production of hepatocellular carcinoma by oval cells: cell cycle expression of c-myc and p53 at different stages of oval cell transformation. Cancer Res. 49, 1554- 1561.

Brill S., Holst P., Sigal S., Zvibel l., Fiorino A., Ochs A., Somasundaran U. and Reid L.M. (1993). Hepatic progenitor populations in embryonic, neonatal, and adult liver. Proc. Soc. Exp. Biol. Med. 204, 261 -269.

Burr A.W., Carpenter M.R., Hines J.E., Gullick W.J. and Burt A.D. (1993). lntrahepatic distribution of transforming growth factor -alpha (TGF-a) during liver regeneration following carbon tetrachloride- induced necrosis. J. Pathol. 170, 95-100.

Burt A.D. and MacSween R.N.M. (1993). Bile duct proliferation - its true significance. Histopathology 23,599-602.

Burt A.D., Stewart J.A., Aitchison M. and MacSween R.N.M. (1987). Expression of tissue polypeptide antigen (TPA) in fetal and adult liver: changes in liver disease. J. Clin. Pathol. 40, 719-724.

Butron Vila M.M., Haot J. and Desmet V.J. (1984). Cholestatic features in focal nodular hyperplasia of the liver. Liver 4, 387-395.

Charlotte F., L'hermine A., Martin N., Geleyn Y., Nollet M., Gaulard P. and Zafrani E.S. (1994). lmmunohistochemical detection of bcl-2 protein in normal and pathological human liver. Am. J. Pathol. 144, 460-465.

Chou S.T. and Gibson J.B. (1972). A comparative histochemical study of rat livers in alpha-naphtyl-iso-thiocyanate (ANIT) and dl-ethionine intoxication. J. Pathol. 108, 73-83.

Coleman W.B., Wennerberg A.E., Smith G.J. and Grisham J.W. (1993). Regulation of the differentiation of diploid and sorne aneuploid rat liver epithelial (stem-like) cells by lhe hepatic microenvironment. Am. J. Pathol. 142, 1373-1382.

Collier J.D., Gullick W.J., Bassendine M.F. and Burt A.D. (1993). Expression of transforming growth factor alpha in human hepatocellular carcinoma. Liver 13. 151-155.

Couch J.A. and Courtney L.A. (1987). N-nitrosodiethylamine-induced hepatocarcinogenesis in estuarine sheepshead minnow (Cyprinodon variegatus): neoplasms and related lesions compared with mammalian lesions. J. Natl. Cancer Inst. 79,297-321.


Dabeva M.D. and Shafritz D.A. (1993). Activation, proliferation, and differentiation of progenitor cells into hepatocytes in the D- galactosamine model of liver regeneration. Am. J. Pathol. 143, 1606-1620.

Dabeva M.D., Alpini G., Hurston E. and Shafrik D.A. (1993). Models for hepatic progenitor cell activation. Proc. Soc. Exp. Biol. Med. 204, 242-252.

DeVos R. and Desmet V. (1992). Ultrastnictural characteristics of novel epithelial cell types identified in human pathologic liver specimens wiih chronic ductular reaction. Am. J. Pathol. 140, 1441-1450.

Delladetsima J.K., Vafiadis l., Tassopoulos N.C., Kyriakou V., Apostolaki A. and Smyrnoff T. (1994). HBcAg and HBsAg expression in ductular cells in chronic hepatitis B. Liver 14, 71-75.

Desmet V.J. (1985). lntrahepatic bile ducts under the lens. J. Hepatol. 1, 545-559.

Desmet V.J. (1986). Current problems in diagnosis of biliaiy disease and cholestasis. Semin. Liver Dis. 6, 233-245.

Duarte l., Llanos O., Domke H., Harz C. and Valdivieso V. (1993). Metaplasia and precursor lesions of gallbladder carcinoma. Cancer 72, 1878-1 884.

Dunsford H.A., Maset R., Saalman J. and Se11 S. (1985). Connection of duct-like structures induced by a chemical hepatocarcinogen to portal bile ducts in the rat liver detected by injection of bile ducts with a pigmented barium gelatin medium. Am. J. Pathol. 118,218-224.

Earp H.S., Hepler J.R., Petch L.A., Miller A., Berry A.R., Harris J., Raymond V.W., McCune B.K., Lee L.W., Grisham J.W. and Harden T.K. (1988). Epidermal growth factor (EGF) and hormones stimulate phosphoinositide hydrolysis and increase EGF receptor protein synthesis and mRNA levels in rat liver epithelial cells. J. Biol. Chem. 263, 13868-1 3874.

Elmore L.W. and Sirica A.E. (1991). Phenotypic characterization of metaplastic intestinal glands and ductular hepatocytes in cholangiofibrotic lesions rapidly induced in the caudate liver lobe of rats treated with furan. Cancer Res. 51, 5752-5759.

Elmore L.W. and Sirica A.E. (1992). Sequential appearance of intestinal mucosa1 cell types in the right and caudate liver lobes of furan- treated rats. Hepatology 16, 1220-1226.

Elmore L.W. and Sirica A.E. (1993). <(Intestinal-typem of adenocarcinoma preferentially induced in right-ckudate liver lobes of rats treated with furan. Cancer Res. 53,254-259.

Engelhardt N.V., Factor V.M., Yasova A.K., Poltoranina V.S., Baranov V.N. and Lasareva M.N. (1990). Common antigens of mouse oval and biliary epithelial cells. Expression on newly formed hepatocytes. Differentiation 45,29-37.

Evarts R.P., Nagy P., Marsden E. and Thorgeirsson S.S. (1987a). A precursor-product relationship exists between oval cells and hepatocytes in rat liver. Carcinogenesis 8, 1737-1740.

Evarts R.P., Nagy P., Marsden E. and Thorgeirsson S.S. (1987b). ln situ hybridization studies on expression of albumin and a-fetoprotein during the early stage of neoplastic transformation in rat liver. Cancer Res. 47,5469-5475.

Evarts R.P., Nagy P., Nakatsukasa H., Marsden E. and Thorgeirsson S.S. (1989). In vivo differentiation of rat liver oval cells into hepatocytes. Cancer Res. 49,1541-1547.

Evarts R.P., Nakatsukasa H.,, Marsden E.R., Hsia C-C., Dunsford H.A. and Thorgeirsson S.S. (1990). Cellular and molecular changes in the early stages of chemical hepatocarcinogenesis in the rat. Cancer Res. 50,3439-3444.

Evarts R.P., Nakatsukasa H., Marsden E.R., Hu Z. and Thorgeirsson

S.S. (1992). Expression of transforming growth factor-a in regenerating liver during hepatic differentiation. Mol. Carcinogen 5, 25-31.

Evarts R.P., Hu Z., Fujio K., Marsden E.R. and Thorgeirsson S.S. (1993). Activation of hepatic stem cell compartment in the rat: Role of transforming growth factor a, hepatocyte growth factor, and acidic fibroblast groMh factor in early proliferation. Cell Growth Differ. 4, 555-561.

Factor V.M. and Radaeva S.A. (1993). Oval cell-hepatocytes relationships in Dipin-induced hepatocarcinogenesis. Exp. Toxicol. Pathol. 465,239-244.

Farber E. (1956). Similarities in the sequence of early histological changes induced in the liver of the rat by ethionine, P-acetylaminofluorene, and 3'-methyl-4-dimethylamianozobenzene. Cancer Res. 16, 142-148.

Farber E. (1992). On cells of origin of liver cell cancer. In: The role of cell types in hepatocarcinogenesis. Sirica A.E. (ed). CRC Press. Boca Raton. pp 1-28.

Faris R.A., Monfils B.A., Dunsford H.A. and Hixson D.C. (1991). Antigenic relationship between oval cells and a subpopulation of hepatic foci, nodules and carcinomas induced by the ~resistant hepatocyte* model system. Cancer Res. 51,1308-1317.

Fausto N., Lemire J.M. and Shiojlri N. (1992). Oval cells in liver carcinogenesis: Cell lineages in hepatic development and the identification of facultative stem cells in normal liver. In: The role of cell types in hepatocarcinogenesis. Sirica A.E. (ed). CRC Press. Boca Raton. pp 89-108.

Fausto N., Lemire J.M. and Shiojiri N. (1993). Cell lineages in hepatic development and the identification of progenitor cells in normal and injured liver. Proc. Soc. Exp. Biol. Med. 204,237-241.

Fisher H.-P., Doppl W., Osborn M. and Altmannsberger M. (1988). Evidence for a hepatocellular lineage in a combined hepatocellular- cholangiocarcinoma of transitional type. Virchows Arch. (8) 56, 71- 76.

Fu X.X., Su C.Y., Lee Y., Hintz R., Biempica L., Snyder R. and Rogler C.E. (1988). Insulin-like growth factor II expression and oval cell proliferation associated with hepatocarcinogenesis in woodchuck hepatitis virus camers. J. Virol. 62, 3422-3430.

Fujio K., Evarts R.P., Hu Z., Marsden E.R. and Thorgeirsson S.S. (1994). Expression of stem cell factor and its receptor, c-kit, during liver regeneration from putative stem cells in the adult rat. Lab. Invest. 70,511-516.

Fukushima S., Shibata M.. Hibino T., Yoshimura T., Hirose M. and Ito N. (1979). lntrahepatic bile duct proliferation induced by 4,4'-diaminodiphenylmethane in rats. Toxicd. Appl. Phamiacol. 48, 145-155.

Gall J.AM. and Bhathal P.S. (1990). Origin and involution of hyperplastic bile ductules following total biliary obstmction. Liver 10, 106-1 15.

Garfield S., Huber B.E., Nagy P., Cordingley M.G. and Thorgeirsson S.S. (1988). Neoplastic transformation and lineage switching of rat liver epithelial cells by retrovirus-associated oncogenes. Mol. Carcinogen. 1, 189-1 95.

Gerber M.A., Thung S.N., Shen S., Stromeyer F.W. and lshak K. (1983). Phenotypic characterization of hepatic proliferation - Antigenic expression by proliferating epithelial cells in fetal liver, massive hepatic necrosis, and nodular transformation of the liver. Am. J. Pathol. 1 10,70-74.

Gerlyng P., Grotmal T., Stokke T., Erikstein B. and Seglen P.O. (1994). Flow cytometric investigation of a possible precursor-product relationship between oval cells and parenchymal cells in the rat liver.


Carcinogenesis 15, 53-59. Germain L., Blouin M.J. and Marceau N. (1988). Biliary epithelial and

hepatocytic cell lineage relationships in embryonic rat liver as determined by the differential expression of cytokeratins, alpha- fetoprotein, albumin, and cell surface-exposed components. ~anber Res. 48,4909-4918.

Germain L., No61 M., Gourdeau H. and Marceau N. (1988). Prornotion of growth and differentiation of rat ductular oval cells in primary culture. Cancer Res. 48, 368-378.

Goldfarb S., Singer E.J. and Popper H. (1962). Experimental cholangitis due to alpha-naphthyl-isothiocyanate (ANIT). Am. J. Pathol. 40, 685- 698.

Goyette M., Faris R., Braun L., Hixson D. and Fausto N. (1990). Expression of hepatocyte and oval cell antigens in hepatocellular carcinomas produced by oncogene-transfected liver epithelial cells. Cancer Res. 50,4809-4817.

Grisham J.W. (1980). Cell types in long-term propagable cultures of rat liver. Ann. NY Acad. Sci. 349, 128-137.

Grisham J.W. (1994). Migration of hepatocytes along hepatic plates and stem cell-fed hepatocyte lineages. Am. J. Pathol. 144, 849-854.

Grisham J.W. and Hartroft W.S. (1961). Morphological identificaüon by electron microscopy of aovaln cells in experimental hepatic degeneration. Lab. Invest. 10,317-332.

Grisham J.W. and Porta E.A. (1964). Origin and fate of proliferated hepatic ductal cells in the rat: electron rnicroscopic and autoradiographic studies. Exp. Mol. Pathol. 3,242-261.

Grisham J.W., Coleman W.B. and Smith G.J. (1993). Isolation, culture, and transplantation of rat hepatocytic precursor (stem-like) cells. Proc. Soc. Exp. Biol. Med. 204,270-279.

Gupta S., Labrecque D.R. and Shafritz D.A. (1992). Mitogenic effects of hepatic stimulator substance on cultured nonparenchymal liver epithelial cells. Hepatology 15, 485-491.

Hayner N.T., Braun L., Yaswen P., Brooks M. and Fausto N. (1984). lsozyrne profiles of oval cells, parenchymal cells, and biliaiy cells isolated by centrifugal elutriation from normal and preneoplastic livers. Cancer Res. 44, 332-338.

Hiley C.G., Strange R.C., Davies M.H., Elias E. and Hubscher S.G. (1993). Glutathione S-transferase expression in prirnary biliary cirrhosis supports concepts of ~ductular metaplasian of hepatocytes. J. Clin. Pathol. 46, 381.

Hillan K.J., Burt A.D., George W.D., MacSween R.N.M., Griffiths M.R. and Bradley J.A. (1989). lntrasplenic hepatocyte transplantation in rats with experimental liver injury: morphological and morphometric studies. J. Pathol. 159, 67-73.

Hixson D.C., Fark R.A., Yang L. and Novikoff P. (1992). Antigenic clues to liver development, renewal, and carcinogenesis: an integrated model. In: The role of cell types in hepatocarcinogenesis. Sirica A.E. (ed). CRC Press. Boca Raton. pp 151-182.

Hsia C.C., Evarts R.P., Nakatsukasa H., Marsden E.R. and Thorgeirrson S.S. (1 992). Occurrence of oval-type cells in hepatitis B virus-associated human hepatocarcinogenesis. Hepatology 16, 1 327- 1 333.

Hsia C.C., Thorgeirrsson S.S. and Tabor E. (1994). Expression of hepatitis B surface and core antigens and transforming growth factor-a in -oval cells* of the liver in patients with hepatocellular carcinoma. J. Med. Virol. 43,216-221.

Hsu W., Deizel D.J., Gould V.E., Warren W.H., Gooch G.T. and Staren E.D. (1991). Neuroendocrine differentiation and prognosis of extrahepatic biliary tract carcinomas. Surgery 1 1 0,604-61 1.

Hu Z., Evarts R.P., Fujio K., Marsden E.R. and Thorgeirsson S.S. (1993). Expression of hepatocyte growth factor and c-met genes during hepatic differentiation and liver development in the rat. Am. J. Pathol. 142, 1823-1 830.

Ide l., Subbarao V., Reddy J.K. and Rao S. (1993). Formation of ductular structures in virro by rat pancreatic epithelial oval cells. Exp. Cell Res. 209,3&44.

lnaoka Y. (1967). Significance of the so-called oval cell proliferation during azo-dye hepatocarcinogenesis. Gann 58,355-366.

lshii M., Vroman B. and La Russo N.F. (1990). Morphologic demonstration of receptor-rnediated endocytosis of epidermal growth factor by isolated bile duct epithelial cells. Gastroenterology 98, 1284-1291.

lwasaki T., Dempo K., Kaneko A. and Ono6 T. (1972). Fluctuation of various cell populations and their characteristics during azo-dye carcinogenesis. Gann 63,21-30.

Johnson M., Koukoulis G., Matsumoto K., Nakamuta T. and lyer A. (1993). Hepatocyte growth factor induces proliferation and morphogenesis in nonparenchymal epithelial liver cells. Hepatology 17, 1052-1 061.

Jonker A.M., Dujkhuis F.W.J., Boes A., Hardonk M.J. and Grond J. (1992). lmmunohistochernical study of extracellular matrix in acute galactosamine hepatitis in rats. Hepatology 15, 423-431.

Jonker A.M., Dijkhuis F.W.J., Hardonk M.J., Moerkerk P., Kate J.T. and Grond J. (1994). lmmunohistochemical study of hepatic fibrosis induced in rats by multiple galactosamine injectlons. Hepatology 19, 775-781.

Joplin R.. Hishida T., Tsubouchi H., Daikuhara Y., Ayres R., Neuberger J.M. and Strain A.J. (1992). Human intrahepatic biliary epithelial cells proliferate in vitro in response to human hepatocyte growth factor. J. Clin. Invest. 90, 1284-1289.

Karaki Y., Munakata S., Saeki T., Hirota S. and Fujimaki M. (1991). Appearance of a carcinoid-like pattem in rat hepatic tumors induced by 3'-methyl-4-dimethyl-arninoazobenzene. Jpn. J. Cancer Res. 82, 397-402.

Kimbrough R.D. (1973). Pancreatic-type tissue in livers of rats fed polychlorinated biphenyls. J. Natl. Cancer Inst. 51,679-681.

Kinami Y., Ashida Y., Seto K., Takashima S. and Kita 1. (1990). lnfluence of incornplete bile duct obstruction on the occurrence of cholangiocarcinoma induced by diisopropanolnitrosamine in hamsters. Oncology 47, 170-1 76.

Kinosita R. (1937). Studies on the cancerogenic chemical substances. Trans. Soc. Pathol. Jpn. 27,665-727.

Kono N. and Nakanuma Y. (1992). Ultrastructural and immunohistochemical studies of the intrahepatic peribiliary capillary plexus in normal livers and extrahepatic biliaiy obstruction in human beings. Hepatology 15,411-418.

Kozuka S., Kurashina M., Tsubone M., Hachisuka K. and Yasui A. (1984). Significance of intestinal metaplasia for the evolution of cancer in the biliary tract. Cancer 54, 2277-2285.

Kuhlmann W.D. and Wurster K. (1980). Correlation of histology and alpha,-fetoprotein resurgence in rat liver regeneration after experimental injury by galactosamine. Virchows Arch. (A) 387, 47- 57.

Kurumaya H., Terada T. and Nakanuma Y. (1990). ~Metaplastic lesions- in intrahepatic bile ducts in hepatolithiasis: a histochemical and immunohistochemical study. J. Gastroenterol. Hepatol. 5, 530- 536.

Lázaro C. and Fausto N. (1994). The differentiation of liver progenitor


cells in culture. FASEB J. 8, A535. Leduc E.H. (1959). Cell modulation in liver pathology. J. Histochem.

Cytochem. 7,253-255. Lee B.C., Hendricks J.D. and Bailey G.S. (1989a). Metaplastic

pancreatic cells in liver tumors induced by diethylnitrosamine. Exp. Mol. Pathol. 50, 104-1 13.

Lee L.W., Tsao M-S., Grisham J.W. and Smith G.J. (1989b). Emergence of neoplastic transformants spontaneously or after exposure to N-methyl-N'-nitro-N-nitrosoguanidine in populations of rat liver epithelial cells cultured under selective and nonselective conditions. Am. J. Pathol. 135, 63-71.

Lee L.W., Raymond V.W., Tsao M-S., Lee D.C., Earp H.S. and Grisham J.W. (1991). Clonal cosegregation of tumorigenicity with overexpression of c-myc and transforming growth factor a genes in chemically transformed rat liver epithelial cells. Cancer Res. 51, 5238-5244.

Lemire J.M. and Fausto N. (1991). Multiple a-fetoprotein RNAs in adult rat liver: Cell type-specific expression and difierential regulation. Cancer Res. 51,4656-4664.

Lemire J.M., Shiojiri N. and Fausto N. (1991). Oval cell proliferation and the origin of small hepatocytes in liver injury induced by D- galactosamine. Am. J. Pathol. 139,535552.

Lenzi R., Liu M.H., Tarsetü F., Slott P.A., Alpini G., Zhai W-R., Paronetto F., Lenzen R. and Tavoloni N. (1992). Histogenesis of bile duct-like cells proliferating during ethionine hepatocarcinogenesis - Evidence for a biliary epithelial nature of oval cells. Lab. Invest. 66, 390-402.

Lin P., Liu C., Tsao M-T. and Grisham J.W. (1987). lnhibition of proliferation of cultured rat liver epithelial cells at specific cell cycle stages by transforming growth factor-@. Biochem. Biophys. Res. Commun. 143,2640.

Liu C., Tsao M-S. and Grisham J.W. (1988). Transforming growth factors produced by normal and neoplastically transformed rat liver epithelial cells in culture. Cancer Res. 48, 850-855.

Makino Y., Yamamoto K. and Tsuji T. (1988). Three-dimensional arrangement of ductular structures formed by oval cells during hepatocarcinogenesis. Acta Med. Okayama 3, 143-150.

Makino T., Usuda N., Rao S., Reddy J.K. and Scarpelli D.G. (1 990). Transdifferentiation of ductular cells into hepatocytes in regenerating hamster pancreas. Lab. Invest. 62,552-561.

Marceau N. (1994). Epithelial cell lineages in developing, restoring, and transforming liver: Evidence for the existence of a wdlfferentiation window*. Gut 35,294-296.

Marceau N,, Blouin M-J., Nbel M., T6rbk N. and Loranger A. (1992). The role of bipotential progenitor cells in liver ontogenesis and neoplasia. In: The role of cell types in hepatocarcinogenesis. Sirica A.E. (ed). CRC Press Boca Raton. pp 121-149.

Mathis G.A. and Sirica A.E. (1990). Effects of medium and substratum conditions on the rates of DNA synthesis in primary cultures of bile ductular epithelial cells. In Vitro Cell Dev. Biol. 26, 113-1 18.

Matsuzaki K., Yoshitake Y., Miyagiwa M., Minemura M., Tanaka M., Sasaki H. and Nishikawa K. (1990). Production of basic fibroblast growth factor-like factor by cultured human cholangiocellular carcinoma calls. Jpn. J. Cancer Res. 81,345-354.

McLean M.R. and Rees K.R. (1958). Hyperplasia of bile ducts included .by alpha-naphthyl-iso-thiocyanate: Experimental biliary cirrhosis free from biliary obstniction. J. Path. Bact. LXXVI, 175-188.

Meybehm M., Fischer H.-P. and Pfeifer U. (1993). Expression of HBs- and HBc-antigen in neoductular epithelium in chronic active hepatitis B. Virchows Arch. (B) 63, 167-172.

Milani S., Herbst H., Schuppan D., Stein H. and Surrenti C. (1991). Transforming growth factors B1 and B2 are differentially expressed in fibrotic liver disease. Am. J. Pathol. 139, 1222-1229.

Moore M.A., Fukushima S., lchihara A., Sato K. and Ito N. (1986). Intestinal metaplasia and altered enzyme expression in propyl- nitrosamine-induced Syrian hamster cholangiocellular and gallbladder lesions. Virchows Arch. (B) 51,29-38.

Morimoto A., Tada K., Nakayama Y., Kohno K., Naito S., Ono M. and Kuwano M. (1994). Cooperative roles of hepatocyte growtti factor and plasminogen activator in tubular morphogenesis by human microvascular endothelial cells. Jpn. J. Cancer Res. 85, 53-62.

Nagore N., Howe S., Boxer L. and Scheuer P.-J. (1989). Liver cell rosettes: structural dierences in cholestasis and hepatitis. Liver 9, 43-51.

Nakanuma Y. and Ohta G. (1986). lmmunohistochemical study on bile ductular proliferation in various hepatobiliary diseases. Liver 6, 205- 211.

Nakanuma Y. and Sasaki M. (1989). Expression of blood group-related antigens in the intrahepatic biliary tree and hepatocytes in normal livers and various hepatobiliary diseases. Hepatology 10, 174-178.

Nakatsukasa H., Evarts R.P., Hsia C.C. and Thorgeirsson S.S. (1990). Transforming growth factor-B, and type 1 procollagen transcripts during regeneration and early fibrosis of rat liver. Lab. Invest. 63, 171-180.

Niedobitek G., Finn T., Herbst H. and Stein H. (1989). Detedon of viral genomes in the liver by in situ hybridization uslng 3%-bromodeoxyuridine-, and biotin-labeled probes. Am. J. Pathol. 134, 633-639.

Nishihara K., Tsuneyoshi M., Niiyama H. and lchimiya H. (1993). Case Report - Composite glandular-endocrine cell carcinoma of the extrahepatic bile duct: immunohistochemical study. Pathology 25, 90-94.

Nomoto M., Uchikosi Y., Kajikazawa N., Tanaka Y. and Asakura H. (1992). Appearance of hepatocyte-like cells in the interlobular bile ducts of human liver in various liver disease states. Hepatology 16, 11 99-1 205.

Nonomura A., Ohta G., Nakanuma Y., lzumi R., Mizukami Y., Matsubara F., Hayashi M., Watanabe K. and Takayanagl N. (1988). Simultaneous detection of epidermal growth factor receptor (EGF- R), epidermal growth factor (EGF) and ras p21 in cholangiocarcinoma by an immunocytochemical method. Liver 8, 157-166.

Novikoff P.M., lkeda T., Hixson D.C. and Yam A. (1991). Characterizations and interactions between bile ductule cells and hepatocytes in ea@ stages of rat hepatocarcinogenesis induced by ethionine. Am. J. Pathol. 139, 1351-1368.

Ogawa K., Minase T. and Onoe T. (1974). Demonstration of glucose-6- phosphatase activity in the oval cells of rat liver and the significance of the oval cells in azo dye carcinogenesis. Cancer Res. 34, 3379- 3386.

Onda H. (1976). lmmunohistochemical studies on al-fetoprotein and al- acid glywprotein during azo dye hepatocarcinogenesis in rats. Gann 67, 253-262.

Phillips M.J. and Poucell S. (1981). Modem aspects of the morphology of viral hepatitis. Hum. Pathol. 12, 1060-1084.

Phillips M.J. and Steiner J.W. (1966). Electron microscopy of cirrhotic nodules-tubularization of the parenchyma by biliary hepatocytes. Lab. Invest. 15,801-817.

Phillips M.J., Poucell S., Patterson J. and Valencia P. (1987). The Liver- An Atlas and Text of Ultrastructural Pathology. Raven Press. New York.

Phillips M.J., Latham P.S. and Poucell-Hatton S. (1993). Electron


microscopy of human liver diseases. In: Diseases of the Liver. 7th ed. Vol. 1. Schiff L. and Schiff E.R. (eds). J.B. Lippincott Co. Philadelphia. pp 189-215.

Plenat F., Braun L. and Fausto N. (1988). Demonstration of glucose-6- phosphatase and peroxisomal catalase activiíy by ultrastructural cytochemistry in oval cells from livers of carcinogen-treated rats. Am. J. Pathol. 130,91-102.

Popper H. and Schaffner F. (1970). Pathophysiology of cholestasis. Hurnan Pathol. 1, 1-24.

Popper H., Kent G. and Stein R. (1957). Ductular cell reaction in the liver in hepatic injury. J. Mount Sinai Hospital XXIV, 551-556.

Pral M., Narsimhan R.P., Crepaldi T., Nicotra M.R., Natal1 P.G. and Comoglio P.M. (1991). The receptor encoded by the human c-met oncogene is expressed in hepatocytes, epithelial cells and solid tumors. Int. J. Cancer 49,323-328.

Price J.M., Harman J.W., Miller E.C. and Miller J.M. (1952). Progressive microscopic alterations in the livers of rats fed the hepatic carcinogens 3'-methyl-4-dimethylaminoazobenzene and 4'-fluoro-4- dimethylaminoazobenzene. Cancer Res. 12, 192-200.

Radaeva S.A. and Factor V.M. (1990). Dynamics of oval cells deveiopment induced in mice liver by Dipin and pariial hepatectomy. Bull. Exp. Biol. Med. (Russ.) 109,514-517.

Rao M.S., Bendayan M., Kimbrough R.D. and Reddy J.K. (1986). Characterization of pancreatic-type tissue in the liver of rats induced by polychlorinated biphenyls. J. Histochem. Cytochem. 34, 197-201.

Rao M.S., Dwivedi R.S., Yeldandi A.V., Subbarao V., Tan X., Usman M.I., Thangada S., Nemali M.R., Kumar S., Scarpelli D.G. and Reddy J.K. (1989). Role of periductal and ductular eptihelial cells of the adult rat pancreas in pancreatic hepatocyte lineage - A change in the difierentiation commitment. Am. J. Pathol. 134. 1069-1 086.

Rao M.S., Subbarao V., Sato K. and Reddy J.K. (1991). Alterations of pancreatic hepatocytes in rats exposed to carcinogens. Am. J. Pathol. 139, 1111-1117.

Ray M.B., Mendenhall C.L., French S.W., Gartside P.S. and the Veterans Administration Cooperative Study Group (1 993). Bile duct changes in alcoholic liver disease. Liver 13, 36-45.

Rhim J., Sandgren E.P., Degen J.L., Palmiter R.D. and Brinster R.L. (1994). Replacement of diseased mouse liver by hepatic cell transplantation. Science 263, 1 149-1 152.

Richards W.L., Tsukada Y. and Potter V.F. (1982). y-Glutamyl transpeptidase and a-fetoprotein expression during a-naphthylisothiocyanate-induced hepatotoxiciíy in rats. Cancer Res. 42, 5133- 5138.

Roskams T., van den Oord J.J., De Vos R. and Desmet V.J. (1990). Neuroendocrine features of reactve bile ductules in cholestatic liver disease. Am. J. Pathol. 137, 1019-1025.

Roskams T., Willems M., Campos R.V., Drucker D.J., Yap S.H. and Desmet V.J. (1993a). Parathyroid hormone-related peptide expression in primary and metastatic liver tumours. Histopathology 23,519-525.

Roskams T., Campos R.V., Drucker D.J. and Desmet V.J. (1993b). Reactive human bile ductules express parathyroid hormone-related peptide. Histopathology 23, 11-19.

Ruan Y., Hacker H.J., Zerban H. and Bannasch P. (1989). Ultrastructural and immunocytochemical characterization of the cellular phenotype in primary adenoid liver tumors of the rat. Pathol. Res. Pract. 184,223-233.

Rubin E. (1964). The origin and fate of proliferated bile ductular cells. Exp. Mol. Pathol. 3,279-286.

Sakamoto S., Yachi A,, Anzai T. and Wada T. (1975). AFP-producing cells in hepatitis and in liver cirrhosis. Ann. NY Acad. Sci. 259, 253- 258.

Sakata H., Jiang B., Kusano M. and Mito M. (1994). Proliferation and morphogenesis of biliary epithelial cells in syngeneic hepatocytes transplantation into the spleen under FK506. Int. Hepatol. Cornmun. 2,67-69.

Sansonno D. and Dammacco F. (1 993). Expression and distribution of a human colon-carcinoma-associated antigen in normal and diseased liver tissue. Pathobiology 61, 193-196.

Santos O.F.P. and Nigam S.K. (1993). HGF-induced tubulogenesis and branching of epithelial cells is rnodulated by extracellular matrix and TGF-O. Dev. Biol. 160,293-302.

Santos O.F.P., Moura L.A., Rosen E.M. and Nigam S.K. (1993). Modulation of HGF-induced tubulogenesis and branching by rnultiple phosphorylation mechanisms. Dev. Biol. 159,535-548.

Saperstein L.A., Jirile R.L., Farouk M., Thompson H.J., Chung K.S. and Meyers W.C. (1994). Transforming growth factor-0, and mannose 6- phosphatelinsulin-like growth factor-ll receptor expression during intrahepatic bile duct hyperplasia and biliary fibrosis in the rat. Hepatology 19,412-41 7.

Scarpelli D.G., Reddy J.K. and Rao S.M. (1989). Metaplastic transformation of pancreatic cells to hepatocytes. In: The pathobiology of neoplasia. Sirica A.E. (ed). Plenum Press. New York. pp 447-495.

Scoazec J-Y., Moreau A., Feldmann G. and Bemau D. (1989). Cellular expression of a-fetoprotein gene and its relation to albumin gene expression during rat azo-dye hepatocarcinogenesis. Cancer Res. 49, 1790-1 796.

Sell S. (1983). Comparsion of oval cells induced in rat liver by feeding N-2-fluorenylacetamide in a choline-devoid diet and bile duct cells induced by feeding 4,4'-diaminodiphenylmethane. Cancer Res. 43, 1761-1767.

Sell S. (1990). Is there a liver stem cell? Cancer Res. 50, 381 1-3815. Sell S. (1993). The role of determined stem-cells in the cellular lineage

of hepatocellular carcinoma. Int. J. Dev. Biol. 37, 189-201. Sell S. (1994). Liver stem cells. Modem Pathol. 7, 105-1 12. Sell S. and Dunsford H.A. (1989). Evidence for the stem cell origin of

hepatocellular carcinoma and cholangiocarcinoma. Am. J. Pathol. 134, 1347-1 363.

Sell S. and Pierce G.B. (1994). Biology of disease-rnaturation arrest of stem cell differentiation is a common pathway for the cellular origin of teratocarcinomas and epithelial cancers. Lab. Invest. 70,6-22.

Sell S. and Salman J. (1984). Light- and electron-microscopic autoradiographic analysis of proliferating cells during the early stages of chemical hepatocarcinogenesis in the rat induced by feedlng N-2-fluorenylacetamide in a choline-deficient diet. Am. J. Pathol. 114,287-300.

Sell S., Osbom K. and Leffert H.L. (1981). Autoradiography of .oval cells* appearing rapidly in the livers of rats fed N-2-fluorenylacetamide in a choline devoid diet. Carcinogenesis 2, 7-14.

Shah K.D. and Gerber M.A. (1990). Development of intrahepatic bile ducts in humans - possible role of laminin. Arch. Pathol. Lab. Med. 1 14,597-600.

Shimizu Y., Demetris A.J., Gollin S., Storio P.D., Bedford H.M., Altarac S., lwatsuki S., Herberman R.B. and Whiteside T.L. (1992). Two new human cholangiocarcinorna cell lines and their cytogenetics and responses to growth factors, hormones, cytokines or immunologic effector cells. Int. J. Cancer 52,252-260.


Shinozuka H., Lombardi B., Sell S. and lammarino R.M. (1978). Early histological and functional alterations of ethionine liver carcinogenesis in rats fed a cholinedeficient diet. Cancer Res. 38, 1092-1098.

Shinozuka H., Sells M.A., Katyal S.L., Sell S. and Lombardi B. (1979). Effects of a choline-devoid diet on the emergence of y- glutamyltranspeptidase-positive foci in the liver of carcinogen- treated rats. Cancer Res. 39,2515-2521.

Shiojiri N. and Mizuno T. (1993). Differentiation of functional hepatocytes and biliary epithelial cells from immature hepatocytes of the fetal mouse in vitro. Anal. Embryol. 187, 221-229.

Sigal S.H., Brill S., Fiorino A.S. and Reid L.M. (1992). The liver as a stem cell and lineage system. Am. J. Physiol. 263, G139-G148.

Sirica A.E. (1992). Biology of biliary epithelial cells. In: Progress in liver diseases. Vol. X. Boyer J.L. and Ockner R.K. (eds). W.B. Saunders Co. Philadelphia. pp 63-87.

Sirica A.E. and Cihla H.P. (1984). lsolation and partial characterizations of oval and hyperplastic bile ductular cell-enriched populations from the livers of carcinogen and noncarcinogen-treated rats. Cancer Res. 44,3454-3466.

Sirica A.E. and Williams T.W. (1992). Appearance of ductular hepatocytes in rat liver after bile duct ligation and subsequent zone 3 necrosis by carbon tetrachloride. Am. J. Pathol: 140, 129-136.

Sirica A.E., Sattler C.A. and Cihla H.P. (1985). Characterization of a primary bile ductular cell culture from the livers of rats during extrahepatic cholestasis. Am. J. Pathol. 120, 67-78.

Sirica A.E., Mathis G.A., Sano N. and Elmore L.W. (1990). Isolation, culture, and tranplantation of intrahepatic biliary epithelial cells and oval cells. Pathobiology 58, 44-64.

Sirica A.E., Elmore L.W. and Sano N. (1991). Characterization of rat hyperplastic bile ductular epithelial cells in culture and in vivo. Digest. Dis. Sci. 36, 494-501.

Sirica A.E., Elmore L.W., Williams T.W. and Cole S.L. (1992). Differentiation potential of hyperplastic bile ductular epithelial cells in rat modeis of hepatic injury and cholangiocarcinogenesis. In: The role of cell types in hepatocarcinogenesis. Sirica A.E. (ed). CRC Press. Boca Raton. pp 183-208

Sirica A.E., Gainey T.W. and Mumaw V.R. (1994a). Ductular hepatocytes: Evidence for a bile ductular cell origin in furan-treated rats. Am. J. Pathol. 145,375-383.

Sirica A.E., Cole S.L. and Williams T. (1994b). A unique rat model of bile ductular hyperplasia in which liver is almost totally replaced with well-differentiated bile ductules. Am. J. Pathol. 144, 1257-1268.

Slott P.A., Liu M.H. and Tavoloni N. (1990). Origin, pattern, and mechanism of bile duct proliferation following biliary obstruction in the rat. Gastroenterology 99, 466-477.

Steiner J.W. and Carruthers J.S. (1963). Electron microscopy of hyperplastic ductular cells in a-naphthylisothiocyanate-induced cirrhosis. Lab. Invest. 12, 471-498.

Steiner J.W., Camthers J.S. and Kalifat S.R. (1962). The ductular cell reaction of rat liver in extrahepatic cholestasis. l. Proliferated biliary epithelial cells. Exp. Mol. Pathol. 1, 162-185.

Steiner J.W., JBzBquel A-M., Phillips M.J., Miyai K. and Arakawa K. (1 965). Some aspects of the ultrastnictural pathology of the liver. In: Progress in liver diseases. Vol. II. Popper H. and Schaffner F. (eds). Grune & Strattop. New York. pp 303-372.

Tanikawa K. (1979). Ultrastructural aspects of the liver and its disorders. 2nd ed. Igaku-Shoin. Tokyo. pp 33-41.

Tarsetti F., Lenzi R., Salvi R., Schuler E., Rijhsinghani K.. Lenzen R.

and Tavoloni T. (1993). Liver carcinogenesis associated with feeding of ethionine in a choline-free diet: evidence against a role of oval cells in the emergence of hepatocellualr carcinoma. Hepatology 18,596403.

Tatematsu M., Kaku T., Ekem J.K. and Farber E. (1984). Studies on the proliferation and fate of oval cells in the liver of rats treated with 2- acetylaminofluorene and partial hepatectomy. Am. J. Pathol. 114, 41 8-430.

Tatematsu M., Kaku T., Medline A. and Farber E. (1985). Intestinal metaplasia as a common option of oval cells in relation to cholangio- fibrosis in liver of rats exposed to 2-acetylaminofluorene. Lab. Invest. 52, 354-362.

Tavoloni N. (1987). The intrahepatic biliary epithelium: An area of growing interest in hepatology. Semin. Liver Dis. 7, 280-292.

Terada T. and Nakanuma Y. (1993). Development of human intrahepatic peribiliary glands-histological, keratin immunohistochemical, and mucus histochemical analyses. Lab. Invest. 68, 261-269.

Terada T., Nakanuma Y. and Goroku 0. (1987). Glandular elements around the intrahepatic bile ducts in man; their morphology and distribution in normal livers. Liver 7, 1-8.

Thamavit W., Kongkanuntn R., Tiwawech D. and Moore M.A. (1987). Level of Opisthorchis infestation and carcinogen dose-dependence of cholangiocarcinoma induction in Syrian golden hamsters. Virchows Arch. (B) 54,52-58.

Thorgeirsson S.S., Evarts R.B., Bisgaard C., Fujio K. and Hu 2. (1993). Hepatic stem cell compartment: activation and lineage commitment. Proc. Soc. Exp. Biol. Med. 204, 253-260.

Thung S.N. (1990). The development of proliferating ductular structures in liver disease - An immunohistochemical study. Arch. Pathol. Lab. Med. 1 14,407-41 1.

Tournier l., Legrbs L., Schoevaert D., Feldmann G. and Bemuau D. (1988). Cellular analysis of a-fetoprotein gene activation during carbon tetrachloride and D-galactosamine-induced acute liver injury in rats. Lab. Invest. 59,657-665.

Tsafarty l., Resau J.H., Rulong S., Keydar l., Faletto D.L. and Vande Woude G.F. (1992). The met proto-oncogene receptor and lumen formation. Science 257, 1258-1261.

Tsao M-S. and Grisham J.W. (1987). Hepatocarcinomas, cholangiocarcinomas, and hepatoblastomas produced by chemically transformed cuitured rat liver epithelial ceils - A light- and electron microscopic analysis. Am. J. Pathol. 127, 168-181.

Tsao M-S. and Zhang X-Y. (1992). The effects of continuous exposure to epidermal growth factor on the spontaneous transformation of cultured rat liver epithelial cells. Am. J. Pathol. 140, 85-94.

Tsa M.-S., Smith J.D., Nelson K.G. and Grisham J.W. (1984). A diploid epithelial cell line from normal adult rat liver with phenotypic properties of oval^ cells. Exp. Cell Res. 154, 38-52.

Tsao M-S., Earp H.S. and Grisham J.W. (19&6). The effects of epidermal growth factor and the state of confluente on enzymatic activities of cultured rat liver epithelial cells. J. Cell. Physiol. 126, 167-1 73.

Tsao M-S., Shepherd J. and Batist G. (1990). Phenotypic expression in spontaneously transformed cultured rat liver epithelial cells. Cancer Res. 50, 1941 -1 947.

Uchida T. and Peters R.L. (1983). The nature and origin of proliferated bile ductules in alcoholic liver disease. Am. J. Clin. Pathol. 79, 326- 333.

Van Eyken P. and Desmet V.J. (1992). Development of intrahepatic bile ducts, ductular metaplasia of hepatocytes, and cytokeratin pattems

ln vited Revie w Ductular hepatocytes - Digitum

Documents