Inhibition of adult liver progenitor (oval) cell growth and viability by an agonist of the peroxisome proliferator activated receptor (PPAR) family member gamma, but not alpha or delta

Multifaceted evidence links the development of liver tumours to the activation and proliferation of adult liver progenitor (oval) cells during the early stages of chronic liver injury. The aim of this study was to examine the role of the peroxisome proliferator activated receptors (PPARs): PPARalpha, delta and gamma, in mediating the behaviour of liver progenitor cells during pre-neoplastic disease and to investigate their potential as therapeutic targets for the treatment of chronic liver injury. We observed increased liver expression of PPARalpha and gamma in concert with expanding oval cell numbers during the first 21 days following commencement of the choline deficient, ethionine supplemented (CDE) dietary model of carcinogenic liver injury in mice. Both primary and immortalized liver progenitor cells were found to express PPARalpha, delta and gamma, but not gamma2, the alternate splice form of PPARgamma. WY14643 (PPARalpha agonist), GW501516 (PPARdelta agonist) and ciglitazone (PPARgamma agonist) were tested for their ability to modulate the behaviour of p53-immortalized liver (PIL) progenitor cell lines in vitro. Both PPARdelta and gamma agonists induced dose-dependent growth inhibition and apoptosis of PIL cells. In contrast, the PPARalpha agonist had no effect on PIL cell growth. None of the drugs affected the maturation of PIL cells along either the hepatocytic or biliary lineages, as judged by their patterns of hepatic gene expression prior to and following treatment. Administration of the PPARgamma agonist ciglitazone to mice fed with the CDE diet for 14 days resulted in a significantly diminished oval cell response and decreased fibrosis compared with those receiving placebo. In contrast, GW501516 did not affect oval cell numbers or liver fibrosis, but inhibited CDE-induced hepatic steatosis. In summary, PPARgamma agonists reduce oval cell proliferation and fibrosis during chronic liver injury and may be useful in the prevention of hepatocellular carcinoma.     

Ethanol interactions with a choline-deficient,ethionine-supplemented feeding regime potentiate pre-neoplastic cellular alterations in rat liver

To examine the effect of ethanol on hepatocarcinogenesis induced by a choline-deficient, ethionine-supplemented (CDE) diet, rats were fed either an ethanol-supplemented diet or ethanol-free, isocaloric diet for 2 months, followed by a CDE diet or control diet for up to 8 months. Changes to cellular composition and pattern of gene expression in the liver were determined at 0 and 3 days, and 1, 2 and 3 weeks after commencing the CDE diet, using histological/immunochemical techniques and northern analysis. Oval cells in the liver were identified morphologically and by expression of pi-glutathione S-transferase (pi-GST), alpha-fetoprotein (AFP) and the embryonic isoform of pyruvate kinase (M2-PK). Oval cell numbers and changes in the pattern of gene expression induced by the CDE diet were accelerated by pre-treatment with ethanol. At all stages, the proportion of oval cells in the test group exceeded that in controls. After 1 week, oval cells had spread sufficiently from the periportal region to be observed pericentrally in test animals and by 3 weeks, extensive formation of ductal structures was apparent, which were absent in controls. Additionally, M2-PK and AFP mRNA were detected earlier, and in greater abundance in animals pre-treated with ethanol. After 8 months of CDE treatment, one or two small hepatic foci (<10 hepatocytes), strongly positive for pi-GST, were detected in the liver of ethanol-pre-treated animals. These foci were absent in CDE-treated animals; however, animals pre-treated with ethanol followed by chronic CDE treatment showed increased size (>40 hepatocytes) and numbers of foci, correlating with the extent of liver damage and varying from 5 to 50% of the liver section. Our data suggest that ethanol pre-treatment potentiates the short-term effects of the CDE diet by enhancing oval cell proliferation, while chronic CDE administration enhances the appearance of pre-malignant hepatic foci that are observed with ethanol pre-treatment alone.     

Expression of alpha, mu and pi class glutathione S-transferases in oval and ductal cells in liver of rats placed on a choline-deficient, ethionine-supplemented diet.

Expression of the alpha, mu and pi class glutathione S-transferases (GSTs) in hepatocytes, oval cells and ductal cells derived from the livers of rats placed on a choline-deficient, ethionine-supplemented (CDE) diet for 5 weeks was investigated. An overall decrease in the expression of alpha and mu class GSTs and an over-expression of pi class GST was observed in the liver after CDE treatment as indicated by Northern blotting analysis. Massive disruption of the liver with oval cell infiltration in the sinusoids throughout the lobule occurred after 5 weeks CDE treatment. 'Duct-like' structures consisting of oval-like cells (ductal cells) with rounder nuclei and more cytoplasm than oval cells within the sinusoids were also apparent. Immunocytochemical analysis revealed that the altered expression of GST in the whole liver is attributed to a differential expression of alpha, mu and pi class GSTs in the different cell types in the liver, including hepatocytes, oval cells around the portal region and among the sinusoids, and oval-like cells (ductal cells) in the 'duct-like' structures. In vitro studies using purified oval-ductal cells and hepatocyte populations confirmed the differential expression of GSTs in the varying cell populations in situ. The expression of the alpha and mu class GSTs in hepatocytes does not appear to be altered by the CDE diet. Heterogeneity in distribution of pi class GST was observed in the hepatocyte population, some hepatocytes were stained strongly while no staining was observed in others. Oval and ductal cells represent two distinct populations displaying different expression of GSTs. Pi class GST was detected in the majority of oval and ductal cells. Alpha class GST was detected in < 5% of the oval cell population and was found in > 50% of the ductal cell population. In contrast, mu class GST was absent in ductal cells and was present in 24% of oval cells around the portal region. This supports the view that ductal cells are not of bile ductal origin since mu GST is present in normal bile duct epithelial cells. Furthermore the change in expression of GSTs in the liver after CDE treatment is attributed to the large increase in oval and ductal cell populations.     

Differential susceptibility of mice humanized for peroxisome proliferator-activated receptor alpha to Wy-14,643-induced liver tumorigenesis

Peroxisome proliferators, such as lipid-lowering fibrate drugs, are agonists for the peroxisome proliferator-activated receptor alpha (PPARalpha). Sustained activation of PPARalpha leads to the development of liver tumors in rodents. Paradoxically, humans appear to be resistant to the induction of peroxisome proliferation and development of liver tumors by peroxisome proliferators. To examine the species differences in response to peroxisome proliferators, a PPARalpha humanized mouse (hPPARalpha) was generated, in which the human PPARalpha was expressed in liver under control of the Tet-OFF system. To evaluate the susceptibility of hPPARalpha mice to peroxisome proliferator-induced hepatocarcinogenesis, a long-term feeding study of Wy-14,643 was carried out. hPPARalpha and wild-type (mPPARalpha) mice were fed either a control diet or one containing 0.1% Wy-14,643 for 44 and 38 weeks, respectively. Gene expression analysis for peroxisomal and mitochondrial fatty acid metabolizing enzymes revealed that both hPPARalpha and mPPARalpha were functional. However, the incidence of liver tumors including hepatocellular carcinoma was 71% in Wy-14,643-treated mPPARalpha mice, and 5% in Wy-14,643-treated hPPARalpha mice. Upregulation of cell cycle regulated genes such as cd1 and Cdks were observed in non-tumorous liver tissue of Wy-14,643-treated mPPARalpha mice, whereas p53 gene expression was increased only in the livers of Wy-14,643-treated hPPARalpha mice. These findings suggest that structural differences between human and mouse PPARalpha are responsible for the differential susceptibility to the peroxisome proliferator-induced hepatocarcinogenesis. This mouse model will be useful for human cancer risk assessment of PPARalpha ligands.     

Hepatic expression of acute-phase protein genes during carcinogenesis induced by peroxisome proliferators

Concern exists regarding peroxisome proliferator (PP) xenobiotic exposure because many PPs are potent hepatocarcinogens in rodents. The mechanism of carcinogenicity induced by PPs is atypical compared with those of other hepatocarcinogens in that the former appears to involve alterations in expression of PP-activated receptor (PPAR) target genes rather than direct mutagenicity. To begin to identify some of these genes, we used differential display to compare mRNA expression between hepatic adenomas and adjacent non-tumor liver from rats fed the potent PP Wy-14643 (WY) for 78 wk. Here, we report increased expression of the acute-phase protein (APP) gene alpha-1 antitrypsin (AT) and decreased expression of alpha2-urinary globulin in the tumors. Similar changes were seen in hepatic adenomas induced by a diethylnitrosamine and phenobarbital protocol, indicating a lack of specificity for PP-induced tumors. Additional APP genes, including ceruloplasmin, haptoglobin, beta-fibrinogen, and alpha1-acid glycoprotein were also upregulated in WY-induced tumors but were downregulated in the livers of rats administered a different PP for 13 wk. Mice treated with either WY or di(2-ethylhexyl) phthalate for 3 wk had decreased hepatic AT expression but increased expression of ceruloplasmin and haptoglobin. PPARalpha-null mice showed no hepatic APP gene alteration after PP treatment but had higher basal expression than did wild-type controls. We conclude that PPARalpha activation by several different PPs leads to dysregulation of hepatic APP gene expression in rats and mice. This dysregulation may indicate alterations in cytokine signaling networks regulating both APP gene expression and hepatocellular proliferation.     

Bone marrow-derived cells fuse with hepatic oval cells but are not involved in hepatic tumorigenesis in the choline-deficient ethionine-supplemented diet rat model

Bone marrow cells (BMCs) have been reported to behave as tissue-specific stem cells in some organs and to participate in tumorigenesis. However, the roles of BMCs in hepatic regeneration and carcinogenesis are still unknown. A choline-deficient, ethionine-supplemented (CDE) diet leads to the appearance of oval cells, a type of hepatic progenitor cell, and activates their replication. Furthermore, this type of diet induces preneoplastic nodules and hepatocellular carcinomas (HCCs) derived from oval cell progenitors. The aims of this study were to determine whether oval cells are derived from BMCs and whether preneoplastic nodules or HCCs originate from BMCs in the CDE diet rat model. To clarify the origin of constituent cells in the liver, we transplanted BMCs from green fluorescent protein (GFP) transgenic female rats into male Lewis rats, which were then exposed to a CDE diet to induce hepatocarcinogenesis. Some oval cells showed both donor-derived GFP expression and the recipient-specific Y chromosome, indicating that donor BMCs fused with recipient oval cells. Several preneoplastic nodules (precancerous lesions) identified by their glutathione S-transferase placental (GSTp) positivity were induced by CDE treatment. However, these preneoplastic GSTp-positive nodules were not GFP positive. In conclusion, this study has produced two major findings. First, BMCs fuse with some oval cells. Second, BMC-fused oval cells and BMCs might not have malignant potential in the CDE-treated rat model.     

Tumor necrosis factor alpha is not required for WY14,643-induced cell proliferation

It has been proposed that the cytokine tumor necrosis factor alpha (TNFalpha) stimulates peroxisome proliferator-induced hepatic cell proliferation. To test this hypothesis, induction of peroxisome proliferation and hepatocyte proliferation were compared in wild-type C57Bl/6 and TNFalpha knockout mice. Animals were dosed with either vehicle or 100 mg/kg/day WY14,643 by oral gavage for 4 days. Liver to brain weight ratios increased in both wild-type and TNFalpha knockout animals after WY14,643 administration. In addition, WY14,643-treated wild-type C57Bl/6 and TNFalpha knockout mice displayed marked hepatic induction of fatty acyl-CoA oxidase activity (approximately 8-fold) and mRNA content (approximately 5-fold). Electron microscopic examination confirmed increased numbers of peroxisomes in hepatocytes in both mouse models. Moreover, WY14,643 markedly induced hepatic cell proliferation (approximately 15-fold) in both wild-type C57Bl/6 and TNFalpha knockout mice as measured by bromodeoxyuridine incorporation into hepatocyte nuclei. In addition, a 50% decrease in TNFalpha mRNA was observed in wild-type mice after treatment with WY14,643. These results suggest that the hepatocellular proliferation induced after peroxisome proliferator treatment occurs independently of TNFalpha signaling.     

Evidence that reduction of hepatocyte growth factor (HGF) is not required for peroxisome proliferator-induced hepatocyte proliferation

The mechanisms underlying peroxisome proliferator-induced hepatocarcinogenesis are not understood. Because of the uncertainty of human cancer risk associated with peroxisome proliferators, delineating the mechanisms of carcinogenesis by these agents is of great interest. Alterations in liver growth factors were postulated to contribute to the carcinogenic effect of peroxisome proliferators. Administration of these compounds to rodents results in down-regulation of hepatocyte growth factor (HGF) and supplementing culture medium with HGF is reported to suppress cell proliferation of preneoplastic and neoplastic cells from WY-14,643-treated livers. Combined, these observations suggest that reduced levels of hepatic HGF contribute to the mechanisms underlying peroxisome proliferator-induced hepatocarcinogenesis. To determine if HGF can prevent the effects of peroxisome proliferators in liver, the short-term influence of WY-14,643 in two different lines of HGF transgenic mice was examined. Mice were fed either a control diet or one containing 0.1% WY-14-643 for one week. Hepatomegaly was found in both HGF transgenic mouse lines fed WY-14,643 compared with controls. Additionally, hepatic expression of typical mRNA markers of peroxisome proliferation including those encoding peroxisomal fatty acid metabolizing enzymes and cell cycle control proteins were all significantly elevated in HGF transgenic mice fed WY-14,643 compared with controls. Down-regulation of HGF was found to be dependent on PPARalpha since lower levels of HGF mRNA and protein were observed in wild-type mice fed WY-14,643 for 1 week and not in similarly treated PPARalpha-null mice. These results demonstrate that the early increase in hepatic mRNAs associated with peroxisome and cell proliferation induced by WY-14,643 treatment can not be prevented by overexpression of HGF in vivo.     

Loss of peroxisomal enzyme expression in preneoplastic and neoplastic lesions induced by peroxisome proliferators in rat livers.

Immunohistochemical staining of enoyl CoA hydratase (ECH), a key peroxisomal enzyme, revealed that the putative preneoplastic lesions induced in livers by administration of the peroxisome proliferator (PP) clofibrate (0.3% in diet) to rats for 60 weeks or more, lacked this enzyme so that they could be detected as ECH-negative foci. ECH and other peroxisomal enzymes such as acyl CoA oxidase, catalase and carnitine-dependent acetyltransferase were also either not or only weakly expressed in most hepatic hyperplastic nodules and hepatomas induced by ciprofibrate (0.025% in diet), Wy-14,643 (0.1%) or BR-931 (0.2%), while being strongly induced in surrounding hepatocytes. These results indicate that the expression of ECH and other peroxisomal enzymes is repressed in putative preneoplastic and neoplastic lesions induced by PPs in rat livers and that these peroxisomal enzymes might therefore be used as negative markers.     

Distribution of alpha-fetoprotein- and albumin-containing cells in the livers of Fischer rats fed four cycles of N-2-fluorenylacetamide.

Immunofluorescent localization of alpha-fetoprotein (AFP)- and albumin-containing cells was determined in the livers of Fischer rats fed 0.05% N-2-acetylfluorenamide for four 2-weeks-on, 1-week-off cycles. After this exposure multiple changes in the liver include over 1000 neoplastic nodules/liver, as well as extensive production of so-called oval cells and focal zones of atypical hepatocellular hyperplasia. Approximately 1% of the oval cells contain AFP, and about half of the zones of atypical hyperplasia include cells that contain AFP, but none of the neoplastic nodules or normal hepatocytes have any AFP-containing cells. Since up to 60% of the hepatocellular carcinomas developing from this regimen will predictably produce AFP, it is tentatively concluded that hepatocellular carcinoma may arise not only from "premalignant" neoplastic nodules but also from oval cells or the atpyical differentiation of hepatocytes.     

Isozyme profiles of oval cells, parenchymal cells, and biliary cells isolated by centrifugal elutriation from normal and preneoplastic livers

The fetal liver isozymes aldolase A and pyruvate kinase K increase in livers of adult rats fed a choline deficient-diet containing 0.1% ethionine. Oval cells isolated by centrifugal elutriation from preneoplastic livers of animals receiving the carcinogenic diet contained these fetal forms as well as fetal-adult isozyme hybrids. In contrast, parenchymal cells isolated from the livers of these animals had only aldolase B and pyruvate kinase L, the same isozymes present in parenchymal cells of normal adult rats. Liver homogenates from rats receiving the carcinogenic diet contain lactate dehydrogenase (LDH) 1, LDH 2, and LDH 3 in addition to LDH 4 and LDH 5, which are the forms detected in normal liver homogenates. LDH 1, LDH 2, and LDH 3 are present in oval cells of preneoplastic livers and in biliary epithelial cells of normal livers, but not in parenchymal cells isolated from normal and preneoplastic livers. Cells of biliary epithelium from normal livers also contain aldolase A and pyruvate kinase K, but not the fetal-adult isozymes present in oval cell populations. The results indicate that, in animals receiving this carcinogenic diet, isozyme alterations associated with neoplasia result from the proliferation of a new cell population which contains these enzymes and not from "dedifferentiation" of mature hepatocytes. Furthermore, the data suggest that this new cell population may include a liver stem cell compartment containing cells in transitional states of differentiation.     

Differences between the promoting activities of the peroxisome proliferator WY-14,643 and phenobarbital in rat liver

In order to characterize the promoting activity of the peroxisome proliferator [4-chloro-6-(2,3-xylidino)-2-pyrimidinylthio]acetic acid (WY-14,463), male F344 rats which received a single 150-mg/kg dose of diethylnitrosamine (DEN) were fed 0.1% WY-14,643 or 0.05% phenobarbital in the diet for 11, 22, or 54 wk. WY-14,643 promoted the development of ATPase-deficient foci but not GGTase-positive or G6Pase-deficient foci, in contrast to phenobarbital which promoted development of foci detected by all three markers. The mode of promotion of ATPase-deficient foci by WY-14,643 was distinctly different from that of phenobarbital. WY-14,643 primarily increased mean volume of foci at 11 and 22 wk, while phenobarbital primarily increased the numerical density of foci at these time points. At 54 wk, the yield of hepatic neoplasms per liver was higher in rats fed WY-14,643 than in rats fed phenobarbital. To evaluate the possibility that the promotional activity of WY-14,643 was more effective at a later stage in hepatocarcinogenesis, rats receiving a dose of DEN and then phenobarbital in the diet for 11 wk were changed to a diet containing WY-14,643 for an additional 11 or 43 wk. However, WY-14,643 feeding from wk 11 to 22 caused a reduction in volume density of ATPase-deficient foci relative to the volume density of foci at 11 wk. In addition, feeding WY-14,643 from wk 11 to 54 caused similar yields of hepatic neoplasms whether or not phenobarbital was fed for the initial 11 wk. WY-14,643 induced hepatic peroxisome proliferation as indicated by palmitoyl CoA oxidase activity regardless of prior treatment with DEN and/or phenobarbital. The yield of neoplasms in rats not receiving DEN was greater in rats fed WY-14,643 for wk 11 to 54 than in rats fed WY-14,643 for wk 1 to 54. In summary, the peroxisome proliferator WY-14,643 was a more efficient promoter of hepatocarcinogenesis in DEN-initiated rats than phenobarbital. The promotional activity of WY-14,643, when evaluated by stereological analysis and by changing promoters, is distinct from that of phenobarbital, perhaps suggesting different cellular and/or molecular mechanisms of promotion. Understanding the role of promotion by WY-14,643 and other peroxisome proliferators may be important in understanding the mechanism of their hepatocarcinogenicity.     

Apoptosis, mitosis and cyclophilin-40 expression in regressing peroxisome proliferator-induced adenomas

Chronic exposure to peroxisome proliferators (PP), including certain industrial and pharmaceutical chemicals, causes liver cancer in rodents. Continuous exposure to PP is needed for tumor development since the frequency of hepatocellular neoplasms is decreased in animals returned to control diet. To determine cellular and molecular events responsible for enhanced growth in PP-induced liver tumors, we evaluated the relationships of WY-14,643 levels, apoptosis, mitosis and cyclophilin-40 (Cyp-40) expression in regressing tumors induced by WY-14,643, a potent PP. Male F344 rats were fed WY-14,643 (0.1%) in the diet for 43 weeks and then switched to control diet for 2, 3, 5 or 36 days. Mean serum and hepatic concentrations of WY-14,643 were decreased as early as 2 days following removal of WY-14,643 as compared with rats continuously fed WY-14,643. Adenomas from rats maintained on WY-14,643 markedly compressed surrounding parenchyma. Evidence of adenoma regression was observed by 3 days of WY-14,643 withdrawal and was characterized by loss of compression. Decreased compression corresponded to increases in the apoptotic index and decreases in the mitotic index in regressing adenomas at 2, 3, and 5 days following the switch to control diet. Cyclophilins are multifunctional receptor proteins involved in numerous signal transduction pathways, including those mediated by cyclosporin, a liver tumor promoter in rats. Cyp-40 expression was markedly increased in adenomas from continuously exposed rats, but expression returned to levels similar to surrounding parenchyma in adenomas after 5 days of WY-14,643 withdrawal. Taken together, these results indicate that WY-14, 643-induced adenomas regress rapidly following withdrawal of the PP in association with declining liver WY-14,643 levels, suggesting that peroxisome proliferator-activated receptor alpha may mediate PP-induced alterations in mitogenic and/or apoptotic regulation in growing tumors, in conjunction with alterations in Cyp-40 signal transduction.     

Monoclonal antibodies recognizing oval cells induced in the liver of rats by N-2-fluorenylacetamide or ethionine in a choline-deficient diet

In this paper, we describe the production of a panel of monoclonal antibodies which define antigens which distinguish between hepatocytes and oval cells. These antibodies were obtained from hybridomas constructed from the spleens of mice immunized by a novel protocol designed to suppress response to unwanted or immunodominant epitopes. Of the antibodies obtained, four, 258.7, 270.11, 258.34, and 270.38, were directed to antigens of morphologically defined oval cells, while two, 258.26 and 270.26, defined cytoplasmic antigens of hepatocytes. Examination of frozen sections of normal, regenerating adult and fetal liver and livers from rats fed 2-acetylaminofluorene or ethionine in a choline-deficient diet indicates that morphologically defined oval cells may in fact comprise a phenotypically complex set of cells composed of at least three antigenically distinct subpopulations. The patterns of expression of the antigens defined by these antibodies suggest two possible pathways of liver cell differentiation.     

Autoradiography of "oval cells" appearing rapidly in the livers of rats fed N-2-fluorenylacetamide in a choline devoid diet

Autoradiographic analysis of liver sections from rats fed thehepatocarcinogen N-2-fluorenylacetamide (FAA) in a choline devoid(CD) diet suggests that proliferating small "oval" cells arisefrom a few small portally-situated cells, and spread rapidlyacross the entire liver lobule. Small cells with detectablegrains are first located where liver plates meet the portalareas. This cell type gradually increases in number over a 10–12day period, then proliferates rapidly. After 28 days, microscopicnodules consisting of heavily labeled large eosinophilic cellsappear, whereas residual hepatocytes are not labeled. Combinedimmunofluorescent and autoradiographic labeling studies revealthat many of the small cells contain AFP; approximately halfof the a-fetoprotein-containing cells are labeled with [³H]thymidine(dT). Feeding CD-FAA diets to rats with hepatocytes prelabeledwith [³H]dT after 70% hepa-tectomy 7 weeks earlier providesdata which suggest that small "oval" cells do not arise fromprelabeled hepatocytes but, instead, infiltrate the prelabeledhepatocytes during the diet induced proliferative phase. Weconclude that "oval" cells arise from a small number of portalcells, not from hepatocytes. Exact identification of the ovalcell precursor is not possible, but it could be a "stem" cell.Although hepatocyte-like properties are found in small cells(e.g., albumin staining), there is no evidence that they differentiateinto normally functioning hepatocytes.     

Carcinohistogenesis and expression of alpha fetoprotein in experimental hepatocarcinoma

In1973,WhenAnthony,etal.inUganda[']werestudyinghumanhepatocirrhosisandhepatocarcinomatheyfirstreportedthatatypicalhyperplasia(AH)ofhepatocyteswasprecancerosisofhepatocellularcarcinoma(HCC).Roncalli,etal.[2]dividedAHintolargehepatocyticandsmallhepatocyticAH,andtheyconsideredthelatterwascloselyrelatedtoformingHCC.In1988,Xu,etal.[3]pointedoutthatHCCoccursfromproliferativenodulesofbasophilicandacidophilichepatocytes.Indeed,carcinohistogenesis,inparticularthechangeofalpha-fetoprotein(AFP)i…