In vivo differentiation of rat liver oval cells into hepatocytes

The Solt-Farber protocol, in the absence of an initiating agent, was used to examine the precursor-product relationship between oval cells and hepatocytes in rat liver. The animals were administered 2-acetylaminofluorene (AAF) by gavage for 2 wk combined with partial hepatectomy 1 wk after administering AAF Two dose levels of AAF were used: 9- and 21-mg total dose for animals in Groups I and II, respectively. [3H]Thymidine was administered i.p. to one-half of the animals at Day 6 post-partial hepatectomy. Animals were sacrificed 7, 9, 11, and 13 days after surgery. Only oval cells became labeled on Day 7 in both groups. On Day 9 both labeled oval cells and labeled basophilic hepatocytes were present in Group I, whereas in Group II only oval cells remained labeled. On Days 11 and 13 both oval cells and basophilic hepatocytes were labeled in both groups. The total amount of radioactivity in Group II livers remained the same on Day 9 when only labeled oval cells were present and on Days 11 and 13 when both labeled oval cells and labeled basophilic hepatocytes were present. The calculated half-life for basophilic hepatocytes was about 50 h. The differentiation of oval cells into basophilic hepatocytes was delayed in Group II as compared to Group I, and the higher dose of AAF also induced the formation of both intestinal metaplasia and bile duct formation. In situ hybridization with an alpha-fetoprotein probe showed a strong expression in groups of typical oval cells and in cells arranged in duct-like structures. In addition a transient expression of AFP was also observed in the areas of basophilic hepatocytes 9 to 11 days after partial hepatectomy. Administration of AAF decreased the level of albumin mRNA in preexisting hepatocytes and caused a significant decrease of serum albumin. In contrast, oval cells showed a strong albumin expression, and basophilic hepatocytes formed islands of albumin-expressing cells. Oval cells and the foci of early basophilic hepatocytes lacked glucose-6-phosphatase activity. At Day 13 significant numbers of basophilic hepatocytes were positive for glucose-6-phosphatase. Oval cells were strongly gamma-glutamyltranspeptidase positive, whereas the foci of basophilic hepatocytes were negative for gamma-glutamyltranspeptidase. Only occasionally were transiently gamma-glutamyltranspeptidase-positive hepatocytes observed in basophilic foci. In summary our data indicate that oval cells can differentiate to hepatocytes and may have an important physiological function as a source of major serum proteins when hepatocytes are unable to synthesize these proteins.     

The resistant hepatocyte model of carcinogenesis in the rat: the apparent independent development of oval cell proliferation and early nodules

The early cellular changes in the Solt-Farber resistant hepatocytemodel of carcinogenesis have been studied to clarify the relationshipof oval cell proliferation to the development of early hepatocytenodules. Cellular proliferation, intermediate filament profilesand the expression of specific cytochrome P450 enzymes wereexamined. At 24 h after partial hepatectomy (PH) many of thebile ductular cells were in S phase, but over the next few daysDNA synthesis progressively decreased in the portal bile ductsand was more common in arborizing ductules (oval cells) radiatingfrom the portal areas. These cells strongly expressed cytokeratins8 and 19 and vimentin, and from 1 week after PH they frequentlyunderwent differentiation either into hepatocytes, expressingcytochrome P450 enzymes, or into intestinal-type cells. Fivedays after PH, numerous basophilic foci were discernible, andthese expanded rapidly. The ductular cells swirled around thefoci, but their antigenic profile clearly indicated that thesecells were not involved in the development of these early nodules.In normal hepatocytes, cytokeratin 8 immunoreactivity was distinctlymembranous in location, and could only be readily detected inperiportal hepatocytes. In the basophilic hepatocyte foci, overexpressionof cytokeratin 8 was consistently associated with cells organizinginto acini, with expression reminiscent of authentic bile ducts,possibly indicating a structure-function relationship. In conclusion,early foci and nodules in this model are derived from resistanthepatocytes and not ductular oval cells, the latter being afacultative multipotential stem cell compartment.     

In situ hybridization studies on expression of albumin and alpha-fetoprotein during the early stage of neoplastic transformation in rat liver

The expressions of albumin and alpha-fetoprotein (AFP) genes were studied in early preneoplastic liver lesions produced by the Solt-Farber protocol using "in situ" hybridization with single stranded RNA probes. In normal rat liver, albumin was expressed at a lower level in the centrilobular than in the periportal areas of the liver acinus, whereas the bile duct epithelium did not show any expression. Five weeks after initiation with diethylnitrosamine, islands of hepatocytes were present which showed heterogeneous expression of albumin and were surrounded by cells comprised of albumin negative hepatocytes and oval cells. gamma-Glutamyltranspeptidase positive foci of enzyme altered cells were located in albumin positive areas. Albumin expression gradually decreased in permanent nodules but increased in the hepatocytes outside the nodules during the first five months after initiation with diethylnitrosamine. Remodeling nodules, which were partly gamma-glutamyltranspeptidase and albumin positive, were also present. However, no consistent correlation was found between gamma-glutamyltranspeptidase positive and albumin negative areas during the first 5 months after initiation. Occasionally, cells showing an elevated expression of albumin were found in permanent nodules. These cells were located in the vicinity of oval type cells, which also showed a weak expression of albumin. AFP was expressed at high level in oval cells 5 weeks after the initiation. However, oval cells observed at later time points, either around the neoplastic nodules or inside the nodules showed only low expression of AFP. Hepatocytes in the enzyme-altered foci and in neoplastic nodules were always negative for AFP. The presence of strongly albumin positive cells inside the neoplastic nodules in close proximity to oval type cells suggests that these cells may be derived from primitive "stem-cell"-like oval cells.     

Production of monoclonal antibodies to preneoplastic liver cell populations induced by chemical carcinogens in rats and to transplantable Morris hepatomas

Monoclonal antibodies (moabs) to neoplastic and preneoplastic liver cells in rats have been selected to follow cellular changes in the livers during chemical carcinogenesis. The moabs were induced by immunizations of BALB/c mice with four partially purified liver cell preparations: 1) oval cells induced in male Fischer rats fed 0.05% N-2-acetylaminofluorene in a choline deficient diet: 2) preneoplastic gamma-glutamyltranspeptidase positive hepatocytes induced by i.p. injection of diethylnitrosamine into male Fischer rats followed by 0.02% N-2-acetylaminofluorene and partial hepatectomy (Solt-Farber model): 3) sharply dissected neoplastic nodules induced in male Fischer rats by five 2-week cycles of 0.05% N-2-acetylaminofluorene diet: and 4) Morris hepatomas 7777 and 5123 passaged in male Buffalo rats. The hybridomas were screened by enzyme linked immunosorbent assay or by indirect immunofluorescence on composite cryostat sections of fetal and adult rat liver, liver containing neoplastic nodules, and Morris hepatoma 7777. Positive clones were limit diluted and partially characterized by indirect immunofluorescence on cryostat sections of other preneoplastic and neoplastic rat livers as well as normal rat tissues. Two moabs to oval cells, two moabs to hepatocytes, and one moab to hepatomas have been selected for further study.     

A precursor--product relationship exists between oval cells and hepatocytes in rat liver

Oval cell proliferation was induced in twelve male Fischer ratsby administration of 2-acetylaminofluorene (2-AAF) for 2 weeksand by performing partial hepatectomy one week after the beginningof 2-AAF administration. Albumin expression in liver was studiedby using in situ hybridization of ³H-labeled rat albumin riboprobe.Radiolabeled thy-midine was administered to a group of animalsat day 6 after partial hepatectomy. The animals were killedat 0, 3, 7, 9, 11 and 13 days after partial hepatectomy. Bothoval cells and basophilk hepatocytes showed a prominent expressionof albumin, whereas albumin expression hi acidophilic preexistinghepatocytes was decreased. Oval cell nuclei were exclusivelylabeled one day after administration of [³H]thymidine. At day9, 11 and 13 basophilic hepatocytes became labeled and the areaoccupied by these cells increased. This is the first demonstrationof the transfer of radiolabeled thymidine from oval ceDs tonewly formed hepatocytes in vivo. Thus the precursor—productrelationship between oval cells and basophilic hepatocytes hasbeen established.     

Initiation by bleomycin of hepatocellular foci in the rat.

The antitumor antibiotic, bleomycin, was tested for activity as an initiator of hepatocellular foci and neoplasms in rats. The compound was administered in a single dose via the portal vein 4 h after the proliferative stimulus of a two-thirds partial hepatectomy. Rats were subsequently fed diet containing phenobarbital for up to 41 weeks to promote the development of initiated hepatocytes. Bleomycin-treated livers displayed significantly increased frequencies of basophilic hepatocellular foci and hepatocellular foci which retain glycogen during fasting. Foci that express glutathione-S-transferase (placental form) were not initiated by bleomycin. Hepatocellular neoplasms were infrequently seen in bleomycin-treated livers (5% incidence). The results suggest that oxygen radical-mediated DNA damage may initiate, within populations of proliferating hepatocytes, new lineages of altered hepatocytes that form foci but have low probability of progressing to neoplasms during promotion with phenobarbital.     

Gamma-glutamyltransferase in putative premalignant liver cell populations during hepatocarcinogenesis.

The activity of gamma-glutamyltransferase, as measured quantitatively and by histochemical staining, was studied in different cell populations during the induction of liver cancer with 2-acetylaminofluorene (2-AAF) or diethylnitrosamine and compared with findings in fetal and in intact and regenerating adult liver. The enzyme activity is 20-fold higher in 12-week nodules than in control livers and 30-fold higher in 20-week nodules than in controls. A similar 30-fold increase in activity relative to control is present in hepatomas, induced by either 2-AAF or diethylnitrosamine, and in fetal hepatocytes. The enzyme shows increases in activity in foci of very early putative preneoplastic hepatocytes induced by a single dose of diethylnitrosamine and selected by low doses of 2-AAF plus partial hepatectomy. By 7 days, the foci show a 4-fold increase in enzyme activity, and by 3 weeks they are 40-fold higher than in the control liver. Histochemically, the foci are strongly positive for gamma-glutamyltransferase, especially in the bile canaliculi. By 21 days, the ductular (oval) cells induced by 2-AAF have disappeared. When stained for the enzyme activity, the foci stand out clearly against the negative background of the liver, allowing easy quantitation. It appears that gamma-glutamyltransferase is a useful marker for preneoplastic hepatocytes.     

Ploidy and karyotype of hepatocytes isolated from enzyme-altered foci in two different protocols of multistage hepatocarcinogenesis in the rat

The ploidy and karyotypes of hepatocytes isolated from the livers of rats subjected to the protocols of Peraino et al. and of Solt and Farber were determined by the examination of such cells in primary culture. A study of 100 or more metaphases from each of five rats on each protocol revealed that 75-80% of gamma-glutamyltranspeptidase-positive (GGT+) hepatocytes isolated from livers of rats in either protocol were diploid, whereas only 23-33% of GGT- cells were diploid. Fifty percent or more of the karyotypes of hepatocytes from livers of rats receiving the Solt-Farber protocol exhibited one or more chromosomal breaks, whereas hepatocytes from livers of rats subjected to the Peraino protocol showed no increase in chromosomal breakage over that in normal controls. These studies demonstrate that the majority of GGT+ cells from altered hepatic foci are diploid and that the greater toxicity of the Solt-Farber protocol over that of Peraino is correlated with marked chromosomal breakage of GGT+ hepatocytes.     

Possible tumor development from double positive foci for TGF-alpha and GST-P observed in early stages on rat hepatocarcinogenesis

Expression of TGF-alpha during promotion of neoplastic development from GST-P-positive foci in rat chemical hepatocarcinogenesis was investigated. One-hundred male F344 rats were given a single intraperitoneal injection of DEN (200 mg/kg bodyweight) and subjected to two-thirds partial hepatectomy at week 3. Commencing 2 weeks from the start, PB at doses of 0 or 500 p.p.m. was fed to the rats for 46 weeks. Groups of 10 rats were killed at weeks 4, 8, 16, 32, 48 and their livers were immunohistochemically examined for expression of GST-P and TGF-alpha. TGF-alpha-positive foci and single positive cells were observed from week 4, partially overlapping with GST-P-positive foci but being much fewer. Numbers of TGF-alpha-positive lesions did not increase from weeks 4-48, but their areas showed increment at weeks 32 and 48, especially with PB administration. Almost all of the tumors observed at weeks 16, 32 and 48 were positive for TGF-alpha (98%). In addition, epidermal growth factor receptor overexpression was observed in most TGF-alpha-positive lesions (foci and tumors). The proliferating cell nuclear antigen labeling index in double positive foci for GST-P and TGF-alpha was significantly higher than that in TGF-alpha-negative foci. In conclusion, TGF-alpha may be closely related with promotion from altered foci to neoplasms in rat hepatocarcinogenesis. Our data suggest that double positive foci for GST-P and TGF-alpha in the early stages of rat hepatocarcinogenesis may develop into tumors with promotion.     

Flow cytometric investigation of a possible precursor-product relationship between oval and parenchymal cells in the rat liver

The question of a possible precursor-product relationship betweenoval cells and hepatocytes was examined in rats treated for2 weeks with 2-acetylaminofluorene (2-AAF) with a two-thirdspartial hepatectomy (PH) performed after the first week of 2-AAFtreatment (modified Soft-Farber model). Liver cells were pulse-chaselabelled with bromodeoxynridine (BrdU) on day 6 post PH. Onday 7 post PH the nonparenchymal (NPC) fraction, which containsthe oval cells, exhibited a labelling index (LI) {small tilde}10times higher than that of the hepatocytes as analysed by flowcytometry (FCM), the majority of the proliferating cells beingoval cells. At later time points, there was no significant increasein the LI of diploid hepatocytes, and no detectable shift ofBrdU-labelled cells from the NPC fraction to the hepatocytefraction, suggesting that no extensive conversion of BrdU-labelledoval cells to hepatocytes was taking place. Throughout the experimentalperiod there was a significant increase in the diploid hepatocytecell fraction, from 12% on day 7 to 25% on day 13 post PH. Diploidhepatocytes pulse-labelled on days 7 or 9 post PH had a highU (7–8%), in contrast to the low LI (1%) of tetra- andoctoploid cells. Proliferation of diploid hepatocytes may thusexplain the large increase in the diploid hepatocyte fractionobserved from days 9 through 15 post PH. Our results, therefore,provide no reason to invoke oval cells as precursors of hepatocytesin the modified Solt-Farber carcinogenesis model.     

Sparse Distribution of Hepatocyte Growth Factor-producing Cells inside Hepatocellular Foci in Rats Treated with Hepatocarcinogens

The distribution of hepatocyte growth factor (HGF)–synthesizing cells in rat liver during development of glutathione S –transfcrasc P form (GST–P)–positive nodules after diethylnitrosamine initiation followed hy promotion with 2–acetylaminofluorene plus partial hepatectomy (PH) was investigated using in situ hybridization, HGF–produclng cells were iioii–parenchymal in nature, and were suspected to be mainly of Kupffer type. They were mostly located outside GST–P–positive lesions, in the surrounding parenchyma. In the oval cell proliferation phase 1 week after PH, they increased and they were mainly localized around the portal triads. It is concluded that HGF is directly involved in an endogenous paracrine growth pathway controlling proliferation in oval cells and in normal, hut not GST–P–positive, hepatocytes.     

Low selection of preneoplastic hepatocytes after treatment with the 2-acetylaminofluorene diet-partial hepatectomy regimen in the liver of hepatocarcinogenesis-resistant DRH strain rats

In hepatocarcinogenesis-resistant DRH rats, preneoplastic hepatocytic lesions are smaller than those of usual rats during carcinogenesis. When preneoplastic hepatocytes from DRH and Donryu (original strain of DRH) were reciprocally transplanted into the livers of DRH and Donryu treated with 2-acetylaminofluorene (2-AAF) diet/two-thirds hepatectomy (PH), the Donryu cells formed small colonies within the DRH liver, whereas the DRH cells formed large colonies within the Donryu liver. The DRH liver showed less degree of oval cell proliferation after treatment with 2-AAF and PH, and DRH hepatocytes were more resistant to the growth-inhibitory effect of 2-AAF after PH. Furthermore, DRH hepatocytes were generally resistant to cytotoxicity of hepatotoxins. The tissue environment of the DRH liver, therefore, is less effective for selective growth of preneoplastic hepatocytes during the carcinogen treatment, which is probably a major cause of the hepatocarcinogenesis-resistance in DRH rats.     

Proliferation of diploid hepatocytes and nonparenchymal (oval) cells during rat-liver regeneration in the presence of 2-acetylaminofluorene

Treatment of rats with the carcinogen 2-acetylaminofluorene (2-AAF) during liver regeneration (Solt-Farber protocol) induced a selective outgrowth of diploid, gamma-glutamyltranspeptidase (GGT)-positive hepatocytes (3-4 times increase) as well as of nonparenchymal (oval) liver cells. After cessation of treatment the oval cells rapidly disappeared, while the population of diploid, GGT-positive hepatocytes declined more slowly over the subsequent ten weeks. In animals pretreated with the initiating carcinogen diethylnitrosamine (DEN) a large fraction of the diploid, GGT-positive hepatocytes persisted. The results differ from those obtained with our standard, sequential treatment protocol (2-AAF given after completed regeneration), where there is no hyperproliferation of oval cells and where GGT-positive hepatocytes are found only in DEN-pretreated animals (Saeter et al, Carcinogenesis 9: 581-587, 1988). Different experimental models of liver carcinogenesis may thus present different patterns of liver cell proliferation, which should be taken into account when general hypotheses on the cellular origin of liver cancer are proposed.     

Evolution of hepatocyte population in the process of chemical carcinogenesis]

By means of two different markers of differentiation, using immunofluorescene method, the authors have characterized changes in the population of hepatic cells 1-8 weeks following the start of 3'-methyl-4-dimethyl-aminoazobenzene or 2-acetyl-aminofluorene action. Alpha-fetoprotein served as a marker of embryonic hepatocyte differentiation; while ligandin-as a marker of high-differentiated mature hepatocytes. The toxic effect of the carcinogens on hepatic stem cells was accompanied with a decrease of ligandin content in centrilobular hepatocytes. Among newly proliferating elements "oval" cells, cells of bile tract epithelium and most of basophilic hepatocyte-like cells fail to contain either alpha-fetoprotein or ligandin. Small groups of basophilic hepatocyte-like cells would contein alpha-fetoprotein. In cells of high cylinder-shaped epithelium of intestinal type ligandin was found, but alpha-fetoprotein was not found. The latter was absent in oxyphilous hepatocytes of hyperplastic nodules. In terms of ligandin content three types of morphologically identical nodules were differentiated; a) ones not containing this protein, b) ones containing it in amounts common to normal mature hepatocyte, and c) hyperdifferentiated nodules containing abnormally high concentrations of ligandin. Within one nodule cell all cells were identical in ligandin content. Thus, it is shown that at early stages of chemical carcinogenesis there occure in the liver multiple foci of differentiation of various kind. These intensive processes are assumed to be essential for tumor evolution in the tissue.     

Epidermal growth factor and proliferation in rat hepatocytes in primary culture isolated at different times after partial hepatectomy

Primary hepatocyte cultures have been prepared from normal adult rat liver and from rat liver at 4, 8, 12, 24, and 48 h following partial hepatectomy (removal of 70% of the liver). Cells were maintained in minimal essential medium alone or supplemented with hormones. Comparing DNA synthesis in normal adult rat hepatocytes with DNA synthesis in hepatocytes isolated from regenerating livers, we found with minimal essential medium alone little DNA synthesis in normal adult rat hepatocytes and in hepatocytes isolated 4, 8, or 12 h after 70% hepatectomy. In hepatocytes isolated 24 h after partial hepatectomy, however, the incorporation of [3H]thymidine was 3 times the rate of normal hepatocytes. The addition of insulin to minimal essential medium had minimal effect on DNA synthesis in all hepatocytes. Addition of epidermal growth factor alone or in combination with insulin resulted in a dramatic increase in DNA synthesis in hepatocytes from regenerating rat liver. Increased incorporation was detectable as early as 4 h after partial hepatectomy and reached a maximum at 24 h after the operation. Results obtained with [3H]thymidine incorporation were confirmed by autoradiography and by direct DNA determinations in hepatocyte cultures. Epidermal growth factor binding to the hepatocytes was determined and agreed with previously reported binding studies. Binding of epidermal growth factor in hepatocytes isolated at 4 h after partial hepatectomy was the same as in normal hepatocytes but was undetectable in hepatocytes isolated from rats at 12 and 24 h after partial hepatectomy.     

Induced drug resistance inhibits selection of initiated cells and cancer development

Compounds exerting a mitoinhibitory effect on normal hepatocytes are potent promoters in the resistant hepatocyte model of chemical carcinogenesis in combination with stimulation of regenerative growth by partial hepatectomy or treatment with carbon tetrachloride. 2- Acetylaminofluorene (2-AAF) almost completely inhibits liver cell regeneration after partial hepatectomy, allowing only resistant cells to participate in regenerative growth. After initiation by diethylnitrosamine and promotion with 2-AAF and partial hepatectomy (PH), focal growth of initiated cells generates liver lesions which occupy 40% of the hepatic volume three weeks after PH. In this work the mechanism for the anti promoting effects of phenobarbital and 3- methylcholantrene were investigated as well as their effects on the development of malignant hepatocellular carcinoma in the resistant hepatocyte model. Treatment with phenobarbital or, especially, 3- methylcholanthrene rendered normal rat hepatocytes resistant to the mitoinhibitory effect of 2-AAF. In combination with 2-AAF/PH, 3- methylcholanthrene shortened the regenerative growth period to less than one week. In the Solt-Farber protocol for experimental hepatocarcinogenesis, treatment with phenobarbital or 3- methylcholanthrene during promotion with 2-AAF/PH permitted hepatocytes surrounding the focal lesions to respond with regenerative growth. The foci and surrounding liver grew until the liver/body mass index reached the control value. With phenobarbital treatment the total focal volume was 20% of the liver volume three weeks after PH, whereas the corresponding value in the case of 3-methylcholanthrene was only 1%. Labelling index data supported the conclusion that growth of the liver lesions in the resistant hepatocyte model was dependent on differential inhibition of normal hepatocyte growth by the promoter and that the size of the foci obtained was related to the length of time after PH required to complete liver regeneration. 3-methylcholanthrene induced 2- AAF resistance prevented the development of large persistent nodules and hepatocellular carcinoma while phenobarbital delayed cancer development with several month. The data thus supports the idea that the degree of clonal expansion during promotion determines the size of the population at risk for malignant transformation, as well as the final frequency of carcinomas.