Role of paracellular junction complexes in baculovirus-mediated gene transfer to nondividing rat hepatocytes.

Gene delivery to differentiated hepatocytes is notoriously difficult. Hepatocytes plated on collagen-coated dishes and maintained in dimethyl sulfoxide (DMSO)-supplemented medium acquire paracellular junctions, arrange themselves in multicellular islands and are an excellent in vitro model for studying liver function. Baculovirus-mediated gene delivery to hepatocytes in this culture system is restricted to peripheral cells of the islands. However, this limitation can be overcome by transient calcium depletion of the cells prior to and during baculovirus infection. Examination of the mechanism underlying this process revealed that calcium depletion was accompanied by a transient loss of intercellular contacts and paracellular junction complex integrity, increased distance between adjoining cells, and internalization of the tight junction protein, zona occludens ZO-1. Internalization of ZO-1 was accompanied by baculovirus infection of internal cells of hepatocyte islands. When calcium levels were restored, paracellular junction complex integrity returned to normal by 12 h. No permanent alterations in hepatocyte ultrastructure and albumin mRNA, and protein expression were caused by this gene transfer method. Loss in paracellular junction complex integrity exposes the basolateral (sinusoidal) surface of hepatocytes resulting in homogeneous baculovirus-mediated gene delivery to approximately 75% of the cells in long-term DMSO culture. We conclude that the use of recombinant baculovirus as a vector in combination with transient calcium depletion is a highly efficient method for delivering exogenous genes to hepatocytes without loss of hepatic differentiation.

Transient disruption of intercellular junctions enables baculovirus entry into nondividing hepatocytes.

Baculovirus infection has extended the capabilities for transfection of exogenous genes into a variety of mammalian cell types. Because rat hepatocytes plated on collagen-coated dishes and maintained in dimethyl sulfoxide (DMSO)-supplemented chemically defined medium are an excellent model system for studying liver function in vitro, we investigated the ability of baculoviruses to infect and deliver exogenous genes to cells in this culture system. Efficient delivery to hepatocytes in short-term culture becomes restricted to peripheral cells, or "edge" cells, as the hepatocytes acquire intercellular junctions and form islands with time in culture. This barrier to baculovirus entry can be overcome, and the percentage of internal cells within the hepatocyte islands that are infected with the baculovirus can be increased more than 100-fold, when cells are subjected to transient calcium depletion before and during infection. These findings suggest that at least in some cell types, such as hepatocytes, baculovirus entry may require contact with the basolateral surface. We conclude from this study that recombinant baculovirus infection following transient depletion of extracellular calcium results in delivery of exogenous genes to at least 75% of hepatocytes in long-term DMSO culture, thereby making it possible for the first time to carry out gain-of-function and loss-of-function studies in this cell system.

A selection system for identifying accessible sites in target RNAs.

Although ribozymes offer tremendous potential for posttranscriptionally controlling expression of targeted genes, their utility is often limited by the accessibility of the targeted regions within the RNA transcripts. Here we describe a method that identifies RNA regions that are accessible to oligonucleotides. Based on this selection protocol, we show that construction of hammerhead ribozymes targeted to the identified regions results in catalytic activities that are consistently and substantially greater than those of ribozymes designed on the basis of computer modeling. Identification of accessible sites should also be widely applicable to design of antisense oligonucleotides and DNAzymes.

Cross-resistance testing of antihepadnaviral compounds using novel recombinant baculoviruses which encode drug-resistant strains of hepatitis B virus.

Long-term nucleoside analog therapy for hepatitis B virus (HBV)-related disease frequently results in the selection of mutant HBV strains that are resistant to therapy. Molecular studies of such drug-resistant variants are clearly warranted but have been difficult to do because of the lack of convenient and reliable in vitro culture systems for HBV. We previously developed a novel in vitro system for studying HBV replication that relies on the use of recombinant baculoviruses to deliver greater than unit length copies of the HBV genome to HepG2 cells. High levels of HBV replication can be achieved in this system, which has recently been used to assess the effects of lamivudine on HBV replication and covalently closed circular DNA accumulation. The further development of this novel system and its application to determine the cross-resistance profiles of drug-resistant HBV strains are described here. For these studies, novel recombinant HBV baculoviruses which encoded the L526M, M550I, and L526M M550V drug resistance mutations were generated and used to examine the effects of these substitutions on viral sensitivity to lamivudine, penciclovir (the active form of famciclovir), and adefovir, three compounds of clinical importance. The following observations were made: (i) the L526M mutation confers resistance to penciclovir and partial resistance to lamivudine, (ii) the YMDD mutations M550I and L526M M550V confer high levels of resistance to lamivudine and penciclovir, and (iii) adefovir is active against each of these mutants. These findings are supported by the limited amount of clinical data currently available and confirm the utility of the HBV-baculovirus system as an in vitro tool for the molecular characterization of clinically significant HBV strains.

Regulation of heme metabolism in rat hepatocytes and hepatocyte cell lines: delta-aminolevulinic acid synthase and heme oxygenase are regulated by different heme-dependent mechanisms.

Regulation of delta-aminolevulinic acid (ALA) synthase and heme oxygenase was analyzed in primary rat hepatocytes and in two immortalized cell lines, CWSV16 and CWSV17 cells. ALA synthase was induced by 4,6-dioxohepatnoic acid (4,6-DHA), a specific inhibitor of ALA dehydratase, in all three systems; however, the induction in CWSV17 cells was greater than in either of the other two systems. Therefore, CWSV17 cells were used to explore the regulation of both enzymes by heme and 4,6-DHA. Data obtained from detailed concentration curves demonstrated that 4,6-DHA induced the activity of ALA synthase once ALA dehydratase activity became rate-limiting for heme biosynthesis. Heme induced heme oxygenase activity with increases occurring at concentrations of 10 microM or greater. Heme blocked the 4,6-DHA-dependent induction of ALA synthase with an EC50 of 1.25 microM. Heme-dependent decreases of ALA synthase mRNA levels occurred more quickly and at lower concentrations than heme-dependent increases of heme oxygenase mRNA levels. ALA synthase mRNA remained at reduced levels for extended periods of time, while the increases in heme oxygenase mRNA were much more transient. The drastic differences in concentrations and times at which heme-dependent effects were observed strongly suggest that two-different heme-dependent mechanisms control the ALA synthase and heme oxygenase mRNAs. In CWSV17 cells, heme decreased the stability of ALA synthase mRNA from 2.5 to 1.3 h, while 4,6-DHA increased the stability of the mRNA to 5.2 h. These studies demonstrate that regulation of ALA synthase mRNA levels by heme in a mammalian system is mediated by a change in ALA synthase mRNA stability. The results reported here demonstrate the function of the regulatory heme pool on both ALA synthase and heme oxygenase in a mammalian hepatocyte system.

Use of the hepatitis B virus recombinant baculovirus-HepG2 system to study the effects of (-)-beta-2',3'-dideoxy-3'-thiacytidine on replication of hepatitis B virus and accumulation of covalently closed circular DNA.

(-)-Beta-2',3'-Dideoxy-3'-thiacytidine (lamivudine [3TC]) is a nucleoside analog which effectively interferes with the replication of hepatitis B virus (HBV) DNA in vitro and in vivo. We have investigated the antiviral properties of 3TC in vitro in HepG2 cells infected with recombinant HBV baculovirus. Different types of information can be obtained with the HBV baculovirus-HepG2 system because (i) experiments can be carried out at various levels of HBV replication including levels significantly higher than those that can be obtained from conventional HBV-expressing cell lines, (ii) cultures can be manipulated and/or treated prior to or during the initiation of HBV expression, and (iii) high levels of HBV replication allow the rapid detection of HBV products including covalently closed circular (CCC) HBV DNA from low numbers of HepG2 cells. The treatment of HBV baculovirus-infected HepG2 cells with 3TC resulted in an inhibition of HBV replication, evidenced by reductions in the levels of both extracellular HBV DNA and intracellular replicative intermediates. The effect of 3TC on HBV replication was both dose and time dependent, and the reductions in extracellular HBV DNA that we observed agreed well with the previously reported efficacy of 3TC in vitro. As expected, levels of HBV transcripts and extracellular hepatitis B surface antigen and e antigen were not affected by 3TC. Importantly, the HBV baculovirus-HepG2 system made it possible to observe for the first time that CCC HBV DNA levels are lower in cells treated with 3TC than in control cells. We also observed that the treatment of HepG2 cells prior to HBV baculovirus infection resulted in a slight increase in the efficacy of 3TC compared to treatments starting 24 h postinfection. The treatment of HepG2 cells with the highest concentration of 3TC tested in this study (2 microM) prior to the initiation of HBV replication markedly inhibited the accumulation of CCC DNA, whereas treatment with the same concentration of 3TC at a time when CCC HBV DNA pools were established within the cells was considerably less effective. In addition, our results suggest that in HepG2 cells, non-protein-associated relaxed circular HBV DNA and particularly CCC HBV DNA are considerably more resistant to 3TC treatment than other forms of HBV DNA, including replicative intermediates and extracellular DNA. We conclude from these studies that the HBV baculovirus-HepG2 system has specific advantages for drug studies and can be used to complement other in vitro model systems currently used for testing antiviral compounds.

Metabolism of 3,5,5-trimethylhexanoyl-ferrocene by rat liver: release of iron from 3,5,5-trimethylhexanoyl-ferrocene by a microsomal, phenobarbital-inducible cytochrome P-450.

3,5,5-Trimethylhexanoyl (TMH)-ferrocene has been used to produce iron loading in whole animals and in cultured hepatocytes. Iron loading produced by TMH-ferrocene is highly efficient and, of the compounds used to produce iron loading in experimental systems, most closely mimics the loading patterns observed in the human disease hemochromatosis. Previous work with TMH-ferrocene has shown that TMH-ferrocene is degraded in vivo because the iron is released from the ferrocene nucleus. Because TMH-ferrocene is highly lipophilic and stable chemically, we hypothesize that this molecule indeed could be degraded enzymatically. To measure the breakdown of TMH-ferrocene, iron release from the molecule was analyzed using a Ferrochem II analyzer, which uses constant potential coulometry to measure the amount of ionic iron within a biological sample. In this study, we show that TMH-ferrocene is degraded by a microsomal enzyme that requires NADPH and molecular oxygen. The TMH-ferrocenase activity is heat labile, requires a physiologic temperature, is induced by phenobarbital, and is inhibited by carbon monoxide and piperonyl butoxide but not by dicoumarol. The enzyme follows Michaelis-Menten kinetics and has a Km of 58.5 microM and a Vmax of 57.5 nmol Fe released/mg protein/min. We conclude that TMH-ferrocene is degraded by a phenobarbital-inducible cytochrome P-450.

Hepatitis B virus replication in human HepG2 cells mediated by hepatitis B virus recombinant baculovirus.

A novel transient mechanism for studying hepatitis B virus (HBV) gene expression and replication using recombinant HBV baculovirus to deliver the HBV genome to HepG2 cells was generated. In HBV baculovirus infected HepG2 cells, HBV transcripts, and intracellular and secreted HBV antigens are produced; replication occurs as evidenced by the presence of high levels of intracellular replicative intermediates and protected HBV DNA in the medium. Density-gradient analysis of extracellular HBV DNA indicated that the DNA was contained predominantly in enveloped HBV virions. Covalently closed circular (CCC) DNA is present indicating that, in this system, HBV core particles are capable of delivering newly synthesized HBV genomes back into the nuclei of infected cells. HBV gene expression is driven exclusively from endogenous promoters. Levels of HBV gene expression and replication can be achieved in HBV baculovirus-infected HepG2 cells which far exceed levels found in HepG2 2.2.15 cells. HBV baculovirus infection of HepG2 cells lends itself readily to experimental manipulation as follows: 1) HBV expression can be initiated any time relative to seeding of HepG2 cells; 2) levels of HBV replication can be regulated over a wide range simply by changing the baculovirus multiplicity of infection; 3) HBV replication is readily detectable by one day post infection with HBV baculovirus and persists at least through day eleven post infection; and (4) the transient nature of the infection can be extended and/or enhanced by superinfecting the cultures. We conclude that infection of HepG2 cells by HBV recombinant baculovirus represents a simple to use and highly flexible system for studying the effects of antivirals and/or cytokines on HBV production and for understanding HBV replication and pathogenesis at the molecular level.

Accumulation of iron by primary rat hepatocytes in long-term culture: changes in nuclear shape mediated by non-transferrin-bound forms of iron.

We have previously shown that hepatocytes in long-term dimethylsulfoxide (DMSO) culture, fed a chemically defined medium, are highly differentiated and an excellent in vitro model of adult liver. Hepatocytes in long-term DMSO culture can be iron loaded by exposure to non-transferrin-bound iron (NTBI) in the form of ferrous sulfate (FeSO4), ferric nitrilotriacetate, or trimethylhexanoyl (TMH)-ferrocene. Holotransferrin, at equivalent times and concentrations, was unable to load hepatocytes. Of the iron compounds tested, TMH-ferrocene most accurately simulated the morphological features of iron-loaded hepatocytes in vivo. When exposed to 25 micromol/L TMH-ferrocene, hepatocytes loaded increasing amounts of iron for 2 months before the cells died. When exposed to lower concentrations of TMH-ferrocene (as low as 2.5 micromol/L), hepatocytes continuously loaded iron and remained viable for more than 2 months. The cellular deposition of iron was different in hepatocytes exposed to TMH-ferrocene compared with those exposed to FeSO4; exposure to TMH-ferrocene resulted in the presence of more ferritin cores within lysosomes than were seen with FeSO4. When the concentration of TMH-ferrocene was increased, a greater number of ferritin cores were observed within the lysosome, and total cellular ferritin, as assessed by Western blot, increased. The formation of hemosiderin was also observed. Furthermore, nuclear shape was distorted in iron-loaded hepatocytes. The extent of deviation from circularity in the nucleus correlated with increasing concentrations of TMH-ferrocene and was greater in hepatocytes exposed to FeSO4 than an equivalent concentration of TMH-ferrocene. The deviation from circularity was smallest in hepatocytes that contained well formed ferritin cores and increased in hepatocytes that contained greater amounts of hemosiderin. Furthermore, in hepatocytes treated with FeSO4, a large amount of cell-associated iron was detected but without a significant increase in the total amount of ferritin. The deviation from circularity was the largest in FeSO4-treated hepatocytes, indicating that iron not properly incorporated into ferritin caused more cellular damage. We conclude that iron-loaded hepatocytes in long-term DMSO culture represent a flexible system for studying the effects of chronic iron loading on hepatocytes.

Exposure of primary rat hepatocytes in long-term DMSO culture to selected transition metals induces hepatocyte proliferation and formation of duct-like structures.

We previously showed that primary rat hepatocytes plated on a rat-tail collagen coated dish and fed a chemically-defined medium supplemented with 2% dimethylsulfoxide (DMSO) can be maintained in a well-differentiated, non-replicating state for periods of several months. In this study, we show that the addition of copper, iron, and zinc to the DMSO-containing chemically defined medium induced DNA synthesis and cell replication during the first two months in culture without loss of hepatic differentiation. DNA synthesis occurred throughout the hepatocyte population without regard to cellular size. No changes were observed in properties indicative of well-differentiated hepatocytes, including cellular morphology, ultrastructure, albumin, or cytokeratin-8 expression. During the third month in culture, after the hepatocytes had become confluent, pseudoduct structures became apparent. Examination of cells lining the ducts by immunohistochemistry showed that these cells lost the ability to express albumin and stained more intensely for cytokeratin 19 and laminin. The ultrastructure of the cells lining the ducts was altered and became more characteristic of bile duct cells. Immunoelectron microscopy revealed that connexin 43, a marker of bile-duct proliferation, was expressed in the duct-like cells. We conclude that under these specific nutritive conditions, primary rat hepatocytes proliferate and, with time, begin to form duct-like structures with altered gene expression and ultrastructural properties.

Effect of transforming growth factor beta on rat hepatocyte cell lines depends upon the state of tumorigenic progression.

The effect of transforming growth factor beta (TGF-beta) on morphology, actin cytoskeleton organization, anchorage-dependent proliferation, anchorage-independent proliferation, and alpha-fetoprotein secretion in a series of related hepatocyte-derived cell lines was measured. We reported previously that TGF-beta 1 partially suppressed the transformed phenotype of a ras- transformed SV40-immortalized hepatocyte cell line designated NR4. In the present study, the CWSV14 cell line, which was derived by transfecting rat hepatocytes with SV40 DNA, was analyzed at low (14LP), mid (14MP), and high (14HP) passage. CWSV14 cells are weakly tumorigenic at low passage and spontaneously transform with in vitro passaging. Tumor cell lines (14T1 and 14T2) derived from hepatocellular carcinomas produced by CWSV14 cells were also analyzed. TGF-beta 1 partially suppressed the transformed phenotype of 14HP, 14T1 and 14T2 cells but had no effect on cell proliferation in these cells. In contrast, TGF-beta 1 induced apoptosis in 14LP and 14MP cells. Studies using both the NR4 and 14T1 cells showed that suppression of the transformed phenotype also could be induced by TGF-beta 2 or TGF-beta 3. We conclude that TGF-beta-induced suppression of the transformed phenotype can be observed in ras-and spontaneously transformed hepatocyte cell lines. It is also apparent that the effect of TGF-beta 1 on hepatocyte cell lines differs depending upon the state of progression toward malignancy; that is, while TGF-beta 1 suppressed the transformed phenotype in highly transformed and tumor cell lines representing late stages in progression, it induced apoptosis in weakly or moderately transformed cell lines representing early stages in progression.

Apoptosis induced by tumor necrosis factor-alpha in rat hepatocyte cell lines expressing hepatitis B virus.

Three well differentiated SV40-immortalized rat hepatocyte cell lines, CWSV1, CWSV2, and CWSV14, and Hepatitis B Virus (HBV)-producing cell lines derived from them were examined for sensitivity to tumor necrosis factor (TNF)-alpha. CWSV1, CWSV2, and CWSV14 cells were co-transfected with a DNA construct containing a dimer of the HBV genome and the neo gene and selected in G418 to generate stable cell lines. Characterization of these cell lines indicated that they contain integrated HBV DNA, contain low molecular weight HBV DNA compatible with the presence of HBV replication intermediates, express HBV transcripts, and produce HBV proteins. The viability of CWSV1, CWSV2, and CWSV2 cells was not significantly altered when they were treated with TNF-alpha at concentrations as high as 20,000 U/ml. The HBV-expressing CWSV1 cell line, SV1di36, and the HBV-expressing CWSV14 cell line, SV14di208, were also not killed when treated with TNF-alpha. However, the HBV-expressing CWSV2 cell line, SV2di366, was extensively killed when treated with TNF-alpha at concentrations ranging from 200 to 20,000 U/ml. Analysis of several different HBV-producing CWSV2 cell lines indicated that TNF-alpha killing depended upon the level of HBV expression. The TNF-alpha-induced cell killing in high HBV-producing CWSV2 cell lines was accompanied by the presence of an oligonucleosomal DNA ladder characteristic of apoptosis.

Tumor necrosis factor-alpha-induced apoptosis in hepatocytes in long-term culture.

Apoptosis occurs naturally in the liver and increases in specific pathogenic processes. We previously described the use of a chemically defined medium supplemented with epidermal growth factor and dimethylsulfoxide to maintain rat hepatocytes in a highly differentiated state for more than 30 days (long-term culture). In this study, we showed that hepatocytes in long-term dimethylsulfoxide culture have definite advantages over using cells in short-term culture (cells in culture for 2 to 4 days) to study apoptosis. We demonstrated that treatment with tumor necrosis factor (TNF)-alpha induced apoptosis (detected morphologically and by formation of an oligonucleosomal DNA ladder) only in hepatocytes that had been subjected to dimethylsulfoxide removal. Neither treatment with TNF-alpha alone or dimethylsulfoxide removal alone induced apoptosis. Apoptosis could be induced by concentrations as low as 500 U of TNF-alpha/ml. Although a DNA ladder was not detected by 12 hours after TNF-alpha treatment, it was easily identified by 24 hours. We conclude that this system can be used 1) to examine the underlying mechanism by which TNF-alpha causes apoptosis in hepatocytes and 2) to study induction of apoptosis in hepatocytes by other agents.

Nuclear scaffold-associated protease: in situ nuclear localization and effects of a protease inhibitor on growth and morphology of a ras-transformed hepatocyte cell line.

We have previously identified a multicatalytic protease (MCP) activity associated with the nuclear scaffold (NS) in hepatocytes and fibroblasts. When we used the chloromethylketone protease inhibitor AAPFcmk, which is targeted to chymotrypsinlike protease activity, we observed a dramatic inhibition of transformation of fibroblasts, with effects that were relatively selective for the NS fraction. Here, we undertook experiments to determine the effects of AAPFcmk on Simian Virus 40-immortalized CWSV1 cells compared with a ras-transformed hepatocyte cell line (NR4) derived from CWSV1. We used biotinAAPFcmk and fluorescent reagents to demonstrate a nuclear chymotrypsinlike protease activity, which is most prominent at the nuclear envelope. The ras-transformed NR4 cells were highly susceptible to growth inhibition in a dose-dependent manner, showing 85% growth inhibition at 50 mumol/LAAPFcmk. In contrast, the immortalized CWSV1 cells were not sensitive at the concentrations (10 to 50 mumol/L) of AAPFcmk tested. In subcellular fractionation studies, the inhibitory effects of AAPFcmk were confined to the NS fraction. The AAPFcmk-induced growth inhibition was accompanied by marked morphological changes in ras-transformed cells, without evidence of overt toxicity. No change in DNA content was observed, but a marked increase in organization of actin cytoskeletal elements was seen. These results suggest that a protease activity associated with the nuclear scaffold has important functions in controlling cytoskeletal filament organization and cell replication.

Responsiveness of an SV40-immortalized hepatocyte cell line to growth hormone.

The response of an SV40-immortalized hepatocyte cell line (CWSV-1) derived from adult male rat hepatocytes to human growth hormone (hGH) was investigated. CWSV-1 cells, which have been characterized extensively, retain certain differentiated functions of normal liver (Woodworth and Isom, Mol Cell Biol 7: 3740-3748, 1987). This cell line consists of tightly associated polygonal, mononucleated cells that grow as monolayers. These cells showed no significant morphological changes with the addition of hGH. Northern blot analysis showed that continuous treatment of the CWSV-1 cells with hGH induced the expression of insulin-like growth factor I (IGF-I) and 5 alpha-reductase RNAs. In addition, continuous exposure to hGH resulted in the induction of expression of the growth hormone receptor/growth hormone binding protein (GHR/GHBP) genes. This study indicates that the CWSV-1 cells may serve as a valuable in vitro model system for studying the signaling pathway of GH.

Transforming growth factor beta 1 suppresses transformation in hepatocytes by regulating alpha 1 beta 1 integrin expression.

We previously reported that (a) treatment of the ras-transformed hepatocyte cell line NR4 with transforming growth factor (TGF) beta 1 suppresses many characteristics associated with the transformed phenotype including altered morphology, actin cytoskeleton reorganization, and anchorage-independent growth such that the cells more closely resemble the immortalized CWSV1 parent cell line; (b) transformed NR4 cells expressed significantly less alpha 1 integrin RNA than the immortalized CWSV1 cells; and (c) TGF-beta 1 treatment of NR4 cells stimulated the expression of alpha 1 and beta 1 integrin RNAs. In this report, the role of the alpha 1 beta 1 integrin in TGF-beta 1-mediated suppression of the ras-transformed phenotype was investigated. We determined that (a) the cell surface integrin that increased in response to TGF-beta 1 treatment of NR4 cells was alpha 1 integrin; (b) TGF-beta 1 altered the ability of NR4 cells to attach to collagen and laminin, the extracellular matrix components that interact with the alpha 1 beta 1 integrin receptor; (c) TGF-beta 1 treatment resulted in relocalization of the alpha 1 integrin on the NR4 cell surface; and (d) TGF-beta 1-mediated inhibition of anchorage-independent growth was blocked by the presence of alpha 1 integrin antibody. A cell line that overexpresses alpha 1 integrin was derived from NR4 cells; characterization of these cells indicated that they continued to express H-ras RNA but were less transformed than the parent NR4 cells. Specifically, they had an altered morphology, an organized actin cytoskeleton, and reduced ability to demonstrate anchorage-independent growth.(ABSTRACT TRUNCATED AT 250 WORDS)

Suppression of albumin enhancer activity by H-ras and AP-1 in hepatocyte cell lines.

We demonstrated, using a transient transfection assay, that the albumin enhancer increased the expression of the albumin promoter in a highly differentiated, simian virus 40 (SV40)-immortalized hepatocyte cell line, CWSV1, but was not functional in two ras-transformed cell lines (NR3 and NR4) derived from CWSV1 by stable transfection with the T24ras oncogene. A transient cotransfection assay showed that T24ras and normal c-Ha-ras were each able to inhibit the activity of the albumin enhancer in an immortal hepatocyte cell line. DNase I footprinting and gel mobility shift assays demonstrated that the DNA binding activities specific to the albumin enhancer were not decreased in the ras-transformed cells. ras also did not diminish the expression of HNF1 alpha, C/EBP alpha, HNF3 alpha, HNF3 beta, or HNF3 gamma but did significantly increase AP-1 binding activity. Three AP-1 binding sites were identified within the albumin enhancer, and DNA binding activities specific to these AP-1 sites were induced in the ras-transformed hepatocytes. Subsequent functional assays showed that overexpression of c-jun and c-fos inhibited the activity of the albumin enhancer. Site-directed mutagenesis of the AP-1 binding sites in the albumin enhancer partially abrogated the suppressing effect of ras and c-jun/c-fos on the enhancer. These functional studies therefore supported the results of the structural studies with AP-1. We conclude that the activity of the albumin enhancer is subject to regulation by ras signaling pathways and that the effect of ras on the albumin enhancer activity may be mediated by AP-1.

Stimulation of DNA synthesis and protooncogene expression in primary rat hepatocytes in long-term DMSO culture.

We have previously described the use of a chemically defined medium (CDM) supplemented with epidermal growth factor (EGF) and dimethylsulfoxide (DMSO) to maintain long-term cultures of rat hepatocytes in a highly differentiated state. In this study, conditions necessary to stimulate high levels of DNA synthesis in hepatocytes in long-term DMSO culture were defined. Hepatocytes were maintained in culture for 20 days in CDM containing DMSO and EGF, insulin, and glucagon. EGF, insulin, and glucagon were then removed for 7 days. Readdition of EGF, insulin, and glucagon at day 27 (shiftup) was accompanied by a three- to sixfold increase in labeling index. If DMSO or dexamethasone (dex) + DMSO were removed at time of shiftup, the labeling index increased by 18- to 54-fold. TGF beta inhibited DNA synthesis stimulated by EGF shiftup, TGF alpha shiftup, or EGF shiftup in combination with removal of dex + DMSO. Stimulation of DNA synthesis was accompanied by a specific, sequential induction of protooncogene mRNA levels; c-fos mRNA was induced 23-fold at 0.5 h after readdition of EGF; c-myc mRNA was induced three- to four-fold by 0.5 h; TGF alpha mRNA was induced sevenfold by 8 h; K-ras mRNA was induced fourfold by 26 h. Changes in protooncogene expression paralleled changes seen in regenerating liver. When DMSO was removed for greater than 48 h, the cells flattened and spread out, chords of cells were no longer well defined, albumin mRNA levels decreased, and fibronectin, beta 1 integrin, and TGF beta transcripts increased.

Transforming growth factor beta 1 partially suppresses the transformed phenotype of ras-transformed hepatocytes.

The effect of transforming growth factor beta type 1 (TGF-beta 1) on DNA synthesis, anchorage-dependent and anchorage-independent proliferation, cytoskeletal organization, and gene expression in ras-transformed simian virus 40 (SV40)-immortalized hepatocyte cell lines was measured. An SV40-immortalized cell line (CWSV1), a control neo-transfected and selected cell line (N1), and neo+ras-transfected and selected cell lines (NR3 and NR4) were used for this study. CWSV1 and N1 cells do not grow in soft agarose and are not tumorigenic. The ras-transformed hepatocytes NR3 and NR4 grow in soft agar and are tumorigenic. TGF-beta 1 treatment did not inhibit DNA synthesis or anchorage-dependent growth in the SV40-immortalized hepatocyte cell line CWSV1 or in the ras-transformed hepatocytes. TGF-beta 1 treatment inhibited anchorage-independent growth, increased actin cytoskeleton organization, and altered the morphology of ras-transformed hepatocytes; that is, with regard to all three of these properties, TGF-beta 1-treated ras-transformed hepatocytes more closely resembled the immortalized parent cell line. c-Ha-ras and c-myc RNA levels were not altered in TGF-beta 1-treated NR4 cells. TGF-beta 1 treatment did alter expression of some genes in NR4 cells. The level of expression of alpha 1 integrin RNA was higher in CWSV1 cells than in NR4 cells and increased in NR4 cells when they were treated with TGF-beta 1. Similarly, the levels and profiles of integrins on the cell surface of CWSV1 cells compared to NR4 cells, as determined by cell surface protein iodination, differed and in TGF-beta 1-treated NR4 cells more closely resembled the surface integrin profile for CWSV1 cells.

Functional analyses of albumin expression in a series of hepatocyte cell lines and in primary hepatocytes.

A series of simian virus 40-immortalized hepatocyte cell lines which are heterogeneous with regard to expression of albumin protein and RNA were characterized for their ability to transcribe the albumin gene. Nascent chain extension assay showed that albumin RNA levels in these cells were determined predominantly at the transcription level. The albumin promoter and enhancer sequences were fused to the bacterial chloramphenicol acetyltransferase gene; the ability of the resulting expression constructs to drive chloramphenicol acetyltransferase expression after transfection into these hepatocyte cell lines was measured. The activity of the albumin promoter and enhancer constructs in primary hepatocytes was also measured. The albumin promoter was expressed differentially in these cells; however, no correlation was found between the transcriptional efficiency of the transfected albumin promoter and endogenous albumin transcription. The albumin enhancer was functional in some but not all albumin-positive cells. The minimal albumin enhancer was mapped to a 330-base pair fragment extending from -9.94 kilobases (kb) to -10.27 kb; three elements within this fragment recently shown to be necessary for enhancer function in a murine hepatocyte cell line were also essential for albumin enhancer function in the rat hepatocyte cell line CWSV1. A transcriptional silencer was identified which could suppress the expression of the homologous albumin promoter and the heterologous herpes simplex virus thymidine kinase promoter. Preliminary analysis localized the albumin silencer between -11 and -12 kb. Our results suggest that multiple regulatory sequences may act cooperatively to determine efficient tissue-specific expression of the albumin gene.