Abstract

For growth stimulation of liver cells by hepatocyte growth factor (HGF) or transforming growth factor α (TGFα) via receptor tyrosine kinases, c-fos/c-jun has been considered a point of intersection for cross-talk between the different signal transduction pathways. Recent evidence strongly implicates translocation of pro-TGFα into the nucleus as an important step preceding the initiation of hepatic DNA synthesis. We asked whether an active c-jun is required for the nuclear translocation of pro-TGFα and its stimulatory effect on DNA synthesis. For this purpose we used mice with c-jun inactivated post partum in hepatocytes by the Cre-loxP recombination system (c-jun Δliver ). Nuclear fractions from control and c-jun Δliver mouse livers contained TGFα as pro-form of 17 kDa and the epidermal growth factor receptor (erbb-1) with molecular weights of 170 and 150 kDa (truncated form). Hepatocytes were isolated by collagenase perfusion and cultivated. A lack of c-jun did not alter the apoptotic activity but significantly suppressed DNA synthesis in the cultured hepatocytes. In control and c-jun Δliver cells DNA synthesis was almost always associated with nuclear presence of pro-TGF α. 76.5 ± 6.8% of hepatocytes with pro-TGF α positive nuclei and only 4.52 ± 1.31% of hepatocytes with negative nuclei exhibited DNA replication. About 85% of the pro-TGF α positive nuclei also contained erbb-1. Treatment of cultures with mature TGFα or HGF elevated the frequency of pro-TGF α positive nuclei replicating DNA; HGF and TGFα-induced nuclear pro-TGF α and DNA synthesis significantly more in c-jun Δliver than in control hepatocytes. These results suggest that (i) a lack of c-jun suppresses basal rates of DNA replication in hepatocytes; (ii) c-jun deficient hepatocytes show a pronounced growth response towards HGF or TGFα; (iii) nuclear translocation of pro-TGF α together with erbb-1 and its association with DNA synthesis are independent of c-jun.

Introduction

Primary cultures of hepatocytes, kept under serum-free conditions, are a good in vitro model to dissect the different growth signals and downstream signalling pathways that lead to a mitogenic response within the liver. In cultured rat hepatocytes TGFα stimulates DNA synthesis alone, or synergistically with other factors ( 14 ). The hepatocyte growth factor/scatter factor (HGF/SF) is a further mitogen for rodent and human hepatocytes ( 59 ). TGFα and HGF bind on the cell surface to erbb- or Met-receptors, respectively ( 10 ). Ligand binding to these tyrosine kinase receptors triggers autophosphorylation, upon which the receptors associate with various transducers containing SH 2 -domains, such as phosphoinositide 3-kinase, the phospholipase Cγ, Grb2, ras GAP, p59 FYN , Shc and others ( 3 , 9 , 11 ). This triggers the activation of protein kinase B and C, the RAF-MEK-MAPK or the JNK cascade leading to transcriptional changes by phosphorylation of nuclear proteins, including the jun family of transcription factors. Jun-proteins (c-jun, junB, junD) are key players in the growth-signalling cascade. They form the transcription factor AP-1, along with fos proteins (c-foc, fosB, fra1, fra2) as well as CREB/ATF and maf proteins ( 12 ). The c-jun proto-oncogene itself is one of the earliest genes that is transcribed after stimulation of cells with growth factors and activation of the MAPK and JNK pathways. When cultured rat hepatocytes are incubated with EGF, TGFα or HGF, there is an early peak of c-jun and c-fos synthesis, followed by increased AP-1 transcriptional activity and further potentiation of the activation domain of the c-jun protein ( 9 , 11 ). Likewise, growth stimulatory xenobiotics, such as the tumour promoter phenobarbital (PB), induce c-jun expression in the intact liver ( 13 ). Thus, c-fos/c-jun may be the point of intersection for cross-talk between the different signal transduction pathways in the liver. C-jun deficient cells suffer from growth arrest or at least a delayed cell cycle. Accordingly, c-jun has been considered essential for liver cell growth. The Cre-loxP recombination system has recently been used to bypass the embryonic lethality of c-jun knockout mice and to study c-jun functions in hepatocytes ( 1417 ). When c-jun was inactivated in hepatocytes during postnatal life, mice stayed viable and showed no overt liver abnormalities, but they displayed impaired liver regeneration in response to partial hepatectomy ( 17 ).

We have shown recently that hepatocytes in the intact liver and in primary culture synthesize pro-forms of TGFα, that translocate to the nucleus in G 1 where they appear to serve in a novel function for mitogenic signalling ( 4 ). Nuclear location of pro-TGFα was induced by various different growth stimuli in cultured hepatocytes and in the intact liver. It seems to operate in addition to the well-known signal pathway via erbb-1 receptor binding of mature TGFα on the cell surface. The erbb-1 belongs to the family of four erbb receptors, each of which is able to form heterodimers with other family members and to bind several ligands including TGFα, EGF, amphiregulin, β-cellulin and others ( 18 , 19 ). Like pro-TGFα erbb-1 was found in the nucleus of proliferating cells; it contains a transactivating domain and associates with the promoter region of the cyclin D1 gene ( 20 ). Upon binding of heregulin β1 erbb-3 was found to shuttle between nuclear and non-nuclear compartments of mammary epithelial cells ( 21 ). Erbb-4, a further member of the erbb-receptor family, may be cleaved by secretase upon activation, which releases the cytoplasmic domain of the receptor for translocation into the nucleus. The C-terminus of this erbb-4 fragment has transcriptional activity ( 22 ). Multiple modes have been suggested by which erbb receptors may escape from the cell membrane and travel to the nucleus ( 23 ). Taken together these recent findings indicate that TGFα and/or erbb-receptors may—at least under certain conditions—bypass cytoplasmic phosphorylation cascades, widely thought to be essential in transducing mitogenic stimuli. Because TGFα is up-regulated in different types of human malignancies, such as hepatocellular carcinoma, novel therapeutic approaches concentrate on the possible benefit of blocking TGFα-evoked signal transduction, e.g. by blockade of the receptor or ligand receptor interactions on the cell surface ( 2428 ). However, human hepatocellular adenoma and carcinoma often express pro-TGFα in their nuclei (E.Schausberger et al ., in preparation). To explore effective strategies for tumour therapy, it is of great importance to focus further investigations on the novel pathway of mitogenic signalling by TGFα gene products.

In the present study we investigated whether the high association between nuclear pro-TGF α and DNA replication in hepatocytes depends on an active c-jun. For this purpose, the inducible cre-loxP recombination system was used to delete c-jun in mouse hepatocytes during postnatal life; the growth potential of these hepatocytes was tested in an ex vivo system. We found that c-jun Δliver mouse hepatocytes show a lower basal rate of DNA replication in primary culture but are stimulated to DNA synthesis by HGF and mature TGFα to a much higher extent than control hepatocytes. DNA synthesis was detected almost exclusively in mouse hepatocyte nuclei expressing nuclear pro-TGF α and erbb-1. Thus, the nuclear pro-TGF α system together with nuclear erbb-1 seems to be involved in DNA replication independent of c-jun. This is further evidence for novel intracellular mitogenic signal transduction pathways that act in addition to ‘classical’ phosphorylation cascades.

Materials and methods

Animals and treatment

The mice used in this study were recombinant inbred strains between 129 × C57/B6 genetic background and were bred at the Research Institute of Molecular Pathology (Vienna, Austria). Animals were kept under standardized SPF-conditions. The inducible Mx-cre approach was applied to delete floxed c-jun alleles in the liver, according to published procedures ( 15 , 17 ). For all experiments, sex-matched littermates, 3 months old, were used. Two weeks before collagenase perfusion the group of control animals without the Mx-cre transgene (c-jun f/f ) and the c-jun Δliver animals with the Mx-cre transgene (Mx-cre c-jun f/f ) were treated with a single injection of 400 µg pI–pC to delete the floxed c-jun alleles. All experiments were performed according to the Austrian guidelines for animal care and protection.

Southern blot

Southern blotting was performed according to standard protocols ( 29 ). Ten microgram of genomic DNA was digested with Xba I yielding a 6.9 kb fragment for the floxed c-jun allele and a 3.3 kb fragment for the deleted c-jun allele. For detection of the bands a 600 bp Bam HI fragment from the c-jun promoter was used as probe.

Cell fractionation, one- and two-dimensional gel electrophoresis and immunoblotting

Nuclei were isolated from the cytoplasmic/membrane fraction applying 2.0 M sucrose for purification ( 30 ). This was followed by nuclear matrix preparation, as has been described in detail ( 31 ). Protein was measured by the Bio-Rad Assay using BSA as a standard. Twenty micrograms of protein of isolated nuclei or of the cytoplasmic/membrane fraction were dissolved in 2× SDS buffer and were applied to one-dimensional SDS–polyacrylamide gels (stacking gel: 5%; resolving gel: 6%). Forty micrograms of the matrix preparations were dissolved and subjected to high-resolution two-dimensional gel electrophoresis. The separated proteins were transferred and detected, as described ( 4 , 32 ). The mouse monoclonal antisera against mature TGFα (Oncogene Science, Uniondale, NY; Ab-1, clone 134A-2B3; 1:300), against the C-terminal amino acids 144–160 of pro-form of wild-type TGFα (InnoGenex, San Ramon, CA; 1:300), against the extracellular domain of erbb-1 (1:500, Genosys Biotechnologies, Pampisford, UK), or against residues 985–996 of the intracellular domain of human erbb-1 (1:500; clone F4; Sigma, St Louis, MO) were applied as primary antibodies.

Studies on primary hepatocyte cultures

Hepatocyte preparation

Male mice were anaesthetized with isoflurane in a carrier gas of dinitrogen oxide and oxygen until exsanguination was terminated. Liver perfusion according to Seglen ( 33 ) with modifications as described ( 34 , 35 ) was done through the portal vein with calcium-free buffer at a constant flow rate of 4.5 ml/min. After the liver was freed of blood the processus papilliformis caudatus was ligated, cut off and fixed in Carnoy's solution for histology (see below). After 10 min of pre-perfusion the perfusion buffer was switched to collagenase buffer (0.2 mg/ml collagenase Sigma Type IV) at decreasing flow rates from 4.5 to 2.5 ml/min. Parenchymal cells were isolated from the initial cell suspension by four cycles of low-speed centrifugation (30 g , 5 min). Viability of the cell isolates was 80 ± 7% by Trypan Blue exclusion test with an average cell yield of 32 ± 17 million cells per mouse liver.

Culture and treatment

Conditions and treatments of cultures have been published in detail ( 35 ). If not stated differently treatment commenced 4 h after plating (time point 0 in the experimental protocol) and was renewed with every medium change. Phenobarbital-Na (PB, Fluka, Buchs, Switzerland) was dissolved in 0.9% of NaCl and was added to the medium for a final concentration of 1 mM. HGF (Sigma), dissolved in PBS as a stock of 20 ng/μl, was applied at a final concentration of 10 ng/ml medium. Recombinant human mature TGFα (UBI, Lake Placid, NY) was prepared as a stock of 10 μg/10 μl in 10 mM acetic acid and added as final concentrations of 20 and 50 ng/μl.

Histology

Liver tissue of untreated C57/B6 mice, fixed in Carnoy's solution, was processed as described ( 36 , 37 ); three serial sections, 2 μm thick, were cut; one of the sections was stained with haematoxylin and eosin, the second one for pro-TGF α and the third one for erbb-1 ( 4 ).

Immunostaining for pro-TGFα or erbb-1 of tissue sections and culture plates

For pre-treatment and staining of tissue sections or culture plates see ( 4 , 35 ). Primary antibodies applied were: mouse monoclonal IgGs against either recombinant mature TGFα (originally clone 213-4.4, Oncogene Science) or the C-terminus of human pro-TGF α (Innogenex); rabbit polyclonal IgG against mature TGFα (Santa Cruz, Santa Cruz, CA 1:50); mouse monoclonal IgG against extracellular domain of erbb-1 (Genosys Biotechnologies); a sheep polyclonal IgG against the intracellular domain of human erbb-1 (Fitzgerald Industries International, Concord, MA).

Double-immunostaining for TGFα and erbb-1; confocal laser scanning microscopy

The procedure followed published descriptions with the following modifications ( 4 , 35 ): primary antisera (anti-TGFα from Oncogene Science; anti-erbb-1 from Fitzgerald, both 1:50) were applied together overnight at 4°C; secondary antisera, i.e. Cy2-labelled donkey-anti-sheep (1:100; Dianova, Hamburg, FRG) and biotinylated rabbit-anti-mouse (Dako, Glostrup, Denmark; 1:300), were used together for 60 min at room temperature, which was followed by Cy3-labelled streptavidin (1:1000 in TBS, 60 min room temperature, Dianova). Stained culture plates were examined by confocal laser scanning microscopy (TCSNT/SP series, Leica-Microsystems, Wetzlar, FRG).

Determination of DNA replication and apoptosis in cultures

Newly synthesized DNA was detected with [ 3 H]thymidine (sp. act. 60–80 Ci/mmol; added at 0.5 μCi/ml to the medium 24 h before harvesting; ARC, St Louis, MO) and subsequent autoradiography ( 4 , 35 ). Plates were stained for 5 min with 8 μg/ml Hoechst 33258 (Riedel de Haen, Seelze, FRG) and mounted. To determine DNA synthesis a total of 1000 TGFα and 300–400 pro-TGFα + hepatocyte nuclei were evaluated per plate. For the evaluation of apoptosis at least 1000 hepatocyte nuclei per plate were scored for apoptotic morphology (condensed and fragmented nuclei) as described ( 38 ). The percentage of labelled hepatocyte nuclei (LI%) and the incidences of apoptotic bodies were calculated.

Statistics

For in vitro studies, four hepatocyte cultures per animal, treatment and time point were run and harvested in parallel. If not stated differently, the mean and SD of three equally treated cultures from at least four different unfloxed and floxed donor mice are given. The significance of differences of means was tested by the non-parametric Wilcoxon's test.

Results

c-jun-is deleted by Cre-lox P recombination

The inducible Mx-cre approach was applied to delete the floxed c-jun alleles in the liver ( 15 , 17 ). The level of c-jun deletion in the mice was checked by Southern blot analysis and nearly 100% deletion of the c-jun alleles was found 2 weeks after injection of pI–pC ( Figure 1 ).

Fig. 1.

Detection of Cre-mediated deletion of c-jun in mouse liver. Southern blot analysis of DNA from liver of groups of control (c-jun f/f ) (+pI–pC) transgenic mice (lanes 1 and 2) in comparison to c-jun Δliver (Mx-cre c-jun f/f ) (+pI–pC) (lanes 3 and 4).The positions of the fragments derived from c-jun flox (Flox) and c-jun Δliver (Δ) alleles are indicated.

Fig. 1.

Detection of Cre-mediated deletion of c-jun in mouse liver. Southern blot analysis of DNA from liver of groups of control (c-jun f/f ) (+pI–pC) transgenic mice (lanes 1 and 2) in comparison to c-jun Δliver (Mx-cre c-jun f/f ) (+pI–pC) (lanes 3 and 4).The positions of the fragments derived from c-jun flox (Flox) and c-jun Δliver (Δ) alleles are indicated.

Basal and induced DNA synthesis in cultured hepatocytes from control and c-jun D Δliver mice

Hepatocytes were isolated from control and c-jun Δliver mice by collagenase perfusion and were cultivated. DNA synthesis of c-jun Δliver hepatocytes was reduced to 30% of DNA synthesis of the control cells ( Figure 2A ). The incidence of apoptoses was not significantly different in both groups (control: 0.35 ± 0.18; c-jun Δliver : 0.23 ± 0.11; n = 4). Thus, a lack of c-jun suppresses DNA replication in adult hepatocytes without significant effect on the apoptotic activity.

Fig. 2.

DNA synthesis in untreated control (Co) and c-jun Δliver mouse hepatocytes associated with nuclear occurrence of pro-TGFα. ( A ) Per cent of all hepatocyte nuclei in S-phase; ( B ) total column gives per cent of all pro-TGFα + hepatocyte nuclei; the dark portion of the column gives per cent of pro-TGFα + nuclei in S-phase; ( C ) total column gives per cent of all pro-TGFα-neg hepatocyte nuclei; the dark portion of the column gives per cent of pro-TGFα nuclei in S-phase. Means ± SD are given from four independent experiments after 48 h in culture. Statistics by Wilcoxon's test: aP < 0.05.

Fig. 2.

DNA synthesis in untreated control (Co) and c-jun Δliver mouse hepatocytes associated with nuclear occurrence of pro-TGFα. ( A ) Per cent of all hepatocyte nuclei in S-phase; ( B ) total column gives per cent of all pro-TGFα + hepatocyte nuclei; the dark portion of the column gives per cent of pro-TGFα + nuclei in S-phase; ( C ) total column gives per cent of all pro-TGFα-neg hepatocyte nuclei; the dark portion of the column gives per cent of pro-TGFα nuclei in S-phase. Means ± SD are given from four independent experiments after 48 h in culture. Statistics by Wilcoxon's test: aP < 0.05.

Next we asked whether cultured primary hepatocytes from control and c-jun Δliver mice show a different response towards various growth stimuli. We chose three compounds with presumably different modes of action ( 39 ). A relatively high concentration of PB in the medium for 48 h did not alter DNA replication in the cultured cells ( Figure 3 ). However, the two growth factors, HGF and TGFα raised the labelling index ( Figure 3 ). The relative growth response was more pronounced in hepatocytes with an inactive c-jun than in control cells. This suggests that the response of hepatocytes towards TGFα and HGF is not impaired by an inactive c-jun. We therefore investigated whether the growth response in c-jun deficient hepatocytes is triggered by other signalling mechanisms, such as represented by the nuclear occurrence of pro-TGF α.

Fig. 3.

Effect of PB, TGFα and HGF on the percentage of DNA synthesis [LI(%)] in control and c-jun Δliver mouse hepatocytes. Total column gives per cent of all hepatocyte nuclei in S-phase (total LI; graphic ); the dark portion of the column ( graphic ) gives per cent of hepatocyte nuclei showing pro-TGFα and S-phase. Number on the top of the columns indicates fold increases of LI over control. Abbreviations: C, control hepatocytes, ΔL, c-jun Δliver hepatocytes; TGFα 20, 20 ng of TGFα/ml medium; TGFα 50, 50 ng of TGFα/ml medium; HGF, 10 ng of HGF/ml medium. Means ± SD are given from four independent experiments after 48 h in culture. Statistics by Wilcoxon's test: aP < 0.05.

Fig. 3.

Effect of PB, TGFα and HGF on the percentage of DNA synthesis [LI(%)] in control and c-jun Δliver mouse hepatocytes. Total column gives per cent of all hepatocyte nuclei in S-phase (total LI; graphic ); the dark portion of the column ( graphic ) gives per cent of hepatocyte nuclei showing pro-TGFα and S-phase. Number on the top of the columns indicates fold increases of LI over control. Abbreviations: C, control hepatocytes, ΔL, c-jun Δliver hepatocytes; TGFα 20, 20 ng of TGFα/ml medium; TGFα 50, 50 ng of TGFα/ml medium; HGF, 10 ng of HGF/ml medium. Means ± SD are given from four independent experiments after 48 h in culture. Statistics by Wilcoxon's test: aP < 0.05.

The pro-form of wild-type TGFα and erbb-1 co-localize in hepatocyte nuclei

In immunostained liver sections of control mice pro-TGF α was detectable in the cytoplasm, cell membranes and most prominent in almost all hepatocyte nuclei located in zone III of the acinus ( Figure 4A and B ). This nuclear location of pro-TGF α could also be seen with an antiserum raised against the C-terminus of pro-TGF α ( Figure 4E ). Nuclear matrix preparations from control and c-jun Δliver mouse livers were subjected to separation by two-dimensional gel electrophoresis and to immunoblotting. TGFα was detectable as pro-form at 17 kDa and an isoelectric point of 7.5 ( Figure 5 ). The mature form of TGFα (5.6 kDa) was not found within the nucleus. The detected spots correspond to those found in rat and human liver, in which MALDI-TOF analysis verified spot 1 and 2 to be pro-TGF α wild-type with the characteristic valines at position 159 and 160 ( 4 ). The other two spots could not yet be identified unequivocally. Pro-TGF α protein appeared less abundant in nuclear matrix preparations from c-jun Δliver than from control animals ( Figure 5 ). In silver-stained two-dimensional gels of the nuclear matrix proteins, the pattern and intensity of spots of the wild-type and c-jun Δliver mice did not reveal significant differences (not shown).

Fig. 4.

Pro-TGFα and erbb-1 are localized to hepatocyte nuclei in vivo and in culture. ( A ) Mouse liver with nuclei positive for pro-TGFα. The central vein is indicated with an arrow. ( B ) Detail of (A). In ( C ) and ( D ) sections were stained with an antiserum raised against the intracellular domain of erbb-1; erbb-1 (C) is localized in hepatocyte nuclei of zones II and III of the liver; the central vein is indicated by an arrow; (D) is a detail of (C). Untreated primary culture of control mouse hepatocytes at 48 h: ( E ) plate was stained with antiserum against the C-terminus of pro-TGFα; pro-TGFα occurs in hepatocyte nuclei with incorporated [ 3 H]thymidine; in ( FH ) confocal laser scanning microscopy for erbb-1 (F), pro-TGFα (G), and in (H) overlap of (F) and (G). Magnifications ×25 (A), ×225 (B and D), ×100 (C), ×200 (E–H).

Fig. 4.

Pro-TGFα and erbb-1 are localized to hepatocyte nuclei in vivo and in culture. ( A ) Mouse liver with nuclei positive for pro-TGFα. The central vein is indicated with an arrow. ( B ) Detail of (A). In ( C ) and ( D ) sections were stained with an antiserum raised against the intracellular domain of erbb-1; erbb-1 (C) is localized in hepatocyte nuclei of zones II and III of the liver; the central vein is indicated by an arrow; (D) is a detail of (C). Untreated primary culture of control mouse hepatocytes at 48 h: ( E ) plate was stained with antiserum against the C-terminus of pro-TGFα; pro-TGFα occurs in hepatocyte nuclei with incorporated [ 3 H]thymidine; in ( FH ) confocal laser scanning microscopy for erbb-1 (F), pro-TGFα (G), and in (H) overlap of (F) and (G). Magnifications ×25 (A), ×225 (B and D), ×100 (C), ×200 (E–H).

Fig. 5.

Detection of pro-TGFα as wild-type form by two-dimensional immunoblots of nuclear matrix proteins of control and c-jun Δliver mouse liver. The gel covers a range of pI 3–8 and Mr 6.5–40 kDa. Corresponding spots (encircled) were analysed from human liver by MALDI-TOF; spots 1 and 2 were identified as wild-type pro-TGFα ( 4 ).

Fig. 5.

Detection of pro-TGFα as wild-type form by two-dimensional immunoblots of nuclear matrix proteins of control and c-jun Δliver mouse liver. The gel covers a range of pI 3–8 and Mr 6.5–40 kDa. Corresponding spots (encircled) were analysed from human liver by MALDI-TOF; spots 1 and 2 were identified as wild-type pro-TGFα ( 4 ).

Immunostains with two different antibodies against the extracellular (not shown) or intracellular domain ( Figure 4C and D ) of erbb-1 showed that the distribution of erbb-1 in the liver was similar to that of pro-TGF α, i.e. in the cell membranes and nuclei of hepatocytes in zone III of the acinus. In immunoblots of control and c-jun Δliver mouse livers erbb-1 appeared predominantly in the cytoplasmic/membrane fraction and faintly in the nuclear fraction at 170 and 150 kDa (truncated form). These results were obtained with antisera raised against the extracellular ( Figure 6 ) or the intracellular (not shown) domain of erbb-1. The cytoplasmic/membrane and the nuclear fraction of c-jun Δliver samples appear to contain less erbb-1 protein than controls ( Figure 6 ).

Fig. 6.

Detection of erbb-1 in one-dimensional immunoblots of nuclear proteins from mouse liver. Nuclear fraction (N) and cytoplasmic membrane fraction (C) of control ( 1 ) and c-jun Δliver ( 2 ) mouse liver.

Fig. 6.

Detection of erbb-1 in one-dimensional immunoblots of nuclear proteins from mouse liver. Nuclear fraction (N) and cytoplasmic membrane fraction (C) of control ( 1 ) and c-jun Δliver ( 2 ) mouse liver.

In immunostained primary hepatocyte cultures from control and c-jun Δliver mice, confocal laser scanning microscopy detected pro-TGF α in the nuclei and erbb-1 in the cell membranes and nuclei. Pro-TGF α and erbb-1 often co-localized in the nucleus ( Figure 4F –H); 84.2 ± 7.9% of the pro-TGFα + nuclei were positive for erbb-1 and 62.6 ± 10.6 of erbb-1 positive nuclei also expressed nuclear pro-TGF α.

Basal and induced DNA replication in cultured mouse hepatocytes is highly associated with nuclear pro-TGF α

In untreated primary cultures the incidence of pro-TGFα + nuclei was ∼3-fold higher in controls (23.6 ± 10.7%) than in c-jun Δliver hepatocytes (7.9 ± 4.8%); 83.5 ± 9.8% of the mouse hepatocytes with pro-TGFα + nuclei and only 4.1 ± 3.1% with negative nuclei replicated DNA, which was observed in both control and c-jun Δliver hepatocytes ( Figures 2B and C and 4E ). As 84.2% of the pro-TGFα + nuclei also expressed erbb-1, at least 67.7% of the pro-TGFα + nuclei in DNA synthesis should be positive for erbb-1. In conclusion, DNA replication in pro-TGF α nuclei was a rare event and nuclear pro-TGF α together with nuclear erbb-1 were strongly associated with DNA synthesis. Similar findings have been reported for primary rat hepatocyte cultures ( 4 ).

We asked whether nuclear pro-TGF α is involved in the growth response of control and c-jun Δliver hepatocytes in primary culture. Forty-eight hours after addition of HGF the percentage of cultured hepatocytes with pro-TGFα + nuclei was raised 2-fold, paralleled by a doubled frequency of nuclei in S-phase ( Figure 3 ). The induction of both, DNA synthesis and pro-TGFα + within the same nuclei, correlated highly (Pearson's test: r2 = 0.7127; significant at P < 0.0001). Treatment of cultures with mature TGFα dose-dependently induced a similar effect. Thus, the induction of DNA synthesis by HGF and TGFα was highly associated with de novo occurrence of nuclear pro-TGF α; this mechanism was active in both control and c-jun Δliver hepatocytes.

Discussion

The present study suggests that an active c-jun supports DNA replication in hepatocytes and that the high association between nuclear pro-TGF α and replicative DNA synthesis does not depend on a functional c-jun, which is supported by the following findings.

Suppressed basal DNA replication after inactivation of c-jun

C-jun Δliver hepatocytes showed lower basal DNA synthesis than control cells. Thus, an inactive c-jun suppresses DNA replication under the present experimental conditions. This agrees with previous data showing that c-jun directly regulates the cyclin D1 and PCNA promoters in fibroblasts. Furthermore, c-jun was shown to control p21 protein levels by inhibiting p53 binding to the p21 promoter. Elevated p21 levels could account for the delayed cell cycle progression of hepatocytes lacking c-jun, as shown by impaired liver regeneration in mice with transgenic overexpression of p21 ( 40 ). In hepatocytes lacking c-jun, mRNA and protein levels of p21 accumulate and again impair regenerative liver growth ( 17 ). Therefore, the diminished progression through the cell cycle in c-jun Δliver hepatocytes might be due to inefficient activation of cyclin-dependent kinases, which are essential for a proliferative response.

Mitogen-induced DNA synthesis occurs in the absence of c-jun

The effects of PB or TGFα on DNA replication of cultured mouse hepatocytes have not yet been studied. In the present work, treatment of mouse hepatocytes in vitro with 1 mM PB did not alter DNA synthesis. The findings obtained are consistent with previous ones in cultured primary rat hepatocytes in which slight stimulation of DNA replication occurred at 1.5 mM and pronounced mitoinhibition at 3 mM ( 35 ). Also, in the intact liver of rats and mice PB affects DNA replication only marginally and exerts its effect on growth of liver mostly via suppression of apoptosis ( 39 ). The present study shows that TGFα stimulates DNA synthesis dose-dependently in cultured mouse hepatocytes. HGF is a well-characterized, potent mitogen for hepatocytes in different mammalian species; in vivo infusion of recombinant HGF into untreated mice results in low level of cell proliferation ( 8 ). Recombinant HGF induced DNA synthesis in primary cultures of rat, mouse and human hepatocytes ( 57 , 9 ). Also, under the conditions of the present study HGF raised significantly the replication of DNA in cultured mouse hepatocytes.

Although basal DNA replication was suppressed in c-jun Δliver hepatocytes, stimulation was not compromised at all in the present study. In contrast, the relative induction of DNA replication was more pronounced in c-jun Δliver than in control hepatocytes. It is presently unclear whether the inactivated c-jun is functionally replaced by other jun proteins (jun-B, jun-D) under our experimental conditions. c-Jun and jun-B have antagonistic functions in a variety of biological processes such as cell proliferation, e.g. in fibroblasts c-jun activates and jun-B suppresses cell proliferation ( 41 ). Recent unexpected data show that jun-B, when introduced by a knock-in strategy at a supraphysiological gene dosage, can functionally replace c-jun in jun-null mice, that otherwise would die between day 12.5 and 14.5 due to defects of the heart and the liver ( 42 ). As the mice still die postnatally, c-jun and jun-B may be functionally interchangeable during development but may have different functions after birth. This suggests that in the present study the induction of DNA synthesis in c-jun deficient hepatocytes, that have been isolated postpartally, may not involve jun-B.

Pro-TGFα occurs in the nucleus of mouse hepatocytes and is associated with basal and induced DNA synthesis

Analogous to findings in rat and human liver pro-forms of wild-type TGFα are present in the nuclei of both, control and c-jun Δliver mice. We have shown recently that a subfraction of the isolated primary hepatocytes synthesize pro-TGF α. Immediately after seeding of the cells pro-TGF α is distributed evenly throughout the cytoplasm and soon thereafter starts to translocate to the nucleus; 16 h later hepatocytes show intranuclear accumulation of pro-TGF α without cytoplasmic residues and after a further lag phase of ∼22 h almost all of the positive nuclei replicate DNA ( 4 ). Addition of mature TGFα to the medium increased DNA synthesis exclusively in pro-TGF α negative hepatocytes. This stimulatory effect was abrogated by neutralizing antibodies or by the erbb-1-specific tyrosine kinase inhibitor tyrphostin A25. These interventions did not affect the impact of nuclear pro-TGF α on DNA replication ( 4 ). The blockade of pro-TGF by RNA interference is currently under investigation.

Like in vitro , the start of liver growth in the intact animal induces de-novo synthesis and nuclear import of pro-TGF α in hepatocytes scattered throughout the liver lobule. This again is followed by DNA replication in the fraction of pro-TGF α positive nuclei. In the resting liver, however, the constitutive occurrence of nuclear pro-TGF α in hepatocytes of zone III seems not to be linked to cell replication ( 4 ). Thus, under conditions of induced hepatocellular growth the nuclear occurrence of newly synthesized pro-TGF α is highly associated with replicative DNA synthesis and may be part of a novel mitogenic-signalling pathway, that acts in addition to the pathway initiated by binding of TGFα to erbb-receptors at the cell membrane ( 4 ).

One advantage of a direct action of pro-TGF α in the nucleus becomes evident in culture, where cells often lack intercellular contacts, cytokines or hormones present in the whole body. Then, the novel pathway may circumvent the secretion and possible loss of mature TGFα to the outside of the cell and may confer autonomy and an inherent growth advantage. The present work shows that mouse hepatocytes utilize this pathway in primary culture in a way similar to rat hepatocytes. A further advantage of a direct action of a growth factor in the nucleus may be to guarantee replication of the cell in case of functional loss of components of other signal transduction pathways; this may be used by hepatocytes lacking an active c-jun. A conceivable third advantage is to maintain specificity, which may be compromised by the degeneracy of signalling pathways, shared by many different cell surface receptors.

The growth stimulatory effects of TGFα or HGF were associated with induction of nuclear pro-TGF α in cultured mouse hepatocytes of both controls and and c-jun Δliver mice. In vivo various hepatomitogenic signals led to pronounced induction of nuclear occurrence of pro-TGF α. In vitro the number of pro-TGF α + hepatocyte nuclei was elevated by treatment with hepatomitogenic steroid hormones or prostaglandins ( 4 ). A further interesting finding is, that treatment with mature TGFα induces the nuclear pro-TGF α system. A similar result was recently obtained in cultured primary rat hepatocytes treated with EGF (E.Schausberger, manuscript in preparation). This suggests that two signal pathways, i.e. the ‘classical’ signal transduction via membrane receptor binding of mature TGFα and the pathway involving nuclear pro-TGF α are interdependent.

Possible interaction of pro-TGFα and erbb-1

How pro-TGF α is translocated to the nucleus and its targets therein are still unclear; the present work indicates that c-jun may not be involved. However, we found that ∼85% of the pro-TGFα + hepatocyte nuclei also were positive for erbb-1. It is currently under investigation whether the large TGFα precursor may attach within the binding site of any of the erbb-receptors suited for the small molecule of mature TGFα and whether pro-TGF α may be co-targeted to the nucleus as a receptor-bound ligand. Recently it was demonstrated that the EGF/erbb-1 complex translocates to the nucleus in tissues with a high proliferative status ( 20 ). In the pre-S phase of liver regeneration 125 I-labelled EGF accumulated in the nucleus of hepatocytes ( 43 ). EGF binding sites and second messenger systems associated with erbb-1 signalling were isolated from matrix of rat liver nuclei ( 4447 ). It was also reported that erbb-1 directly activates the signal transducers and activators of transcription (STAT) proteins independent of jak-1 ( 48 ). Moreover, evidence was provided that nuclear erbb-1 acts as a transcription factor or co-activator of cyclin D1 ( 20 ). In a further study erbb-4 was found to be cleaved upon activation; the released cytoplasmic domain of the receptor translocates into the nucleus and exerts transcriptional activity ( 22 ). The recent findings together with the present study imply that growth factors/erbb-receptors may bypass the protein phosphorylation cascades and c-jun for the transduction of mitogenic stimuli ( 23 ).

In conclusion, the present work shows that the nuclear occurrence of pro-TGF α is associated with DNA replication of hepatocytes, which is independent of the function of c-jun. Further research is necessary to elucidate the mechanisms that regulate the different intracellular routes of pro-TGF α and that link nuclear pro-TGF α to DNA replication.

The excellent technical assistance of K.Bukowska, M.Käfer, and H.Koudelka is gratefully acknowledged. This study was also supported by ‘Herzfeldersche Familienstiftung’.

References

1.
Mead,J.E. and Fausto,N. (
1989
) Transforming growth factor α may be a physiological regulator of liver regeneration by means of an autocrine mechanism.
Proc. Natl Acad. Sci. USA
  ,
86
,
1558
–1562.
2.
Michalopoulos,G. and DeFrances,M. (
1997
) Liver regeneration
Science
  ,
276
,
61
–66.
3.
Moriuchi,A., Hirono,S., Ido,A., Ochiai,T., Nakama,T., Uto,H., Hori,T., Hayashi,K. and Tsubouchi,H. (
2001
) Additive and inhibitory effects of simultaneous treatment with growth factors on DNA synthesis through MAPK pathway and G1 cyclins in rat hepatocytes.
Biochem. Biophys. Res. Commun.
  ,
280
,
368
–373.
4.
Grasl-Kraupp,B., Schausberger,E., Hufnagl,K., Gerner,C., Löw-Baselli,A., Rossmanith,W., Parzefall,W. and Schulte-Hermann.,R. (
2002
) A novel mechanism for mitogenic signaling via pro-transforming growth factor α within hepatocyte nuclei.
Hepatology
  ,
35
,
1372
–1380.
5.
Nakamura,T., Teramoto,H. and Ichihara,A. (
1986
) Purification and characterization of a growth factor from rat platelets for mature parenchymal hepatocytes in primary culture.
Proc. Natl Acad. Sci. USA
  ,
83
,
6489
–6493.
6.
Yee,C.J., DeFrances,M.C., Bell,A., Bowen,W., Petersen,B., Michalopoulos,G.K. and Zarnegar,R. (
1993
) Expression and characterization of biologically active human hepatocyte growth factor (HGF) by insect cells infected with HGF-recombinant baculovirus.
Biochemistry
  ,
32
,
7922
–7931.
7.
Petersen,B., Yee,C.J., Bowen,W., Zarnegar,R. and Michalopoulos,G.K. (
1994
) Distinct morphological and mito-inhibitory effects induced by TGF-β1, HGF and EGF on mouse, rat and human hepatocytes.
Cell Biol. Toxicol.
  ,
10
,
219
–230.
8.
Roos,F., Ryan,A.M., Chamow,S.M., Bennett,G.L. and Schwall,R.H. (
1995
) Induction of liver growth in normal mice by infusion of hepatocyte growth factor/scatter factor.
Am. J. Physiol.
  ,
268
,
G380
–G386.
9.
Gomez-Lechon,M.J., Guillen,I., Ponsoda,X., Fabra,R., Trullenque,R., Nakamura,T. and Castell,J.V. (
1996
) Cell cycle progression proteins (cyclins), oncogene expression and signal transduction during the proliferative response of human hepatocytes to hepatocyte growth factor.
Hepatology
  ,
23
,
1012
–1019.
10.
Naldini,L., Vigna,E., Narsimhan,R.P., Gaudino,G., Zarnegar,R., Michalopoulos,G.K. and Comoglio,P.M. (
1991
) Hepatocyte growth factor (HGF) stimulates the tyrosine kinase activity of the receptor encoded by the proto-oncogene c-MET.
Oncogene
  ,
6
,
501
–504.
11.
Brenner,D.A. (
1998
) Signal transduction during liver regeneration.
J. Gastroenterol. Hepatol.
  ,
13
(suppl.),
93
–95.
12.
Wagner,E.F. (
2001
) AP-1.
Oncogene
  ,
20
,
2334
–2335.
13.
Holden,P.R., Odum,J., Soames,A.R., Foster,J.R., Elcombe,C.R. and Tugwood,J.D. (
1998
) Immediate-early gene expression during regenerative and mitogen-induced liver growth in the rat.
J. Biochem. Mol. Toxicol.
  ,
12
,
79
–82.
14.
Hilberg,F., Aguzzi,A., Howells,N. and Wagner,E.F. (
1993
) c-jun is essential for normal mouse development and hepatogenesis.
Nature
  ,
365
,
179
–181.
15.
Kühn,R., Schwenk,F., Aguet,M. and Rajekski,K. (
1995
) Inducible gene targeting in mice.
Science
  ,
269
,
1427
–1429.
16.
Eferl,R., Sibilia,M., Hilberg,F., Fuchsbichler,A., Kufferath,I., Guertl,B., Zenez,R., Wagner,E.F. and Zatloukal,K.J. (
1999
) Functions of c-Jun in liver and heart development.
J. Cell Biol.
  ,
145
,
1049
–1061.
17.
Behrens,A., Sibilia,M., David,J.-P., Möhle-Steinlein,U., Tronche,F., Schütz,G. and Wagner,E.F. (
2002
). Impaired postnatal hepatocyte proliferation and liver regeneration in mice lacking c-jun in the liver.
EMBO J.
  ,
21
,
1782
–1790.
18.
Olayioye,M.A., Neve,R.M., Lane,H.A. and Hynes,N.E. (
2000
) The ErbB signaling network: receptor heterodimerization in development and cancer.
EMBO J.
  ,
19
,
3159
–3167.
19.
Yarden,Y. and Sliwkowski,M.X. (
2001
) Untangling the ErbB signalling network.
Nature Rev. Mol. Cell Biol.
  ,
2
,
127
–137.
20.
Lin,S.Y., Makino,K., Xia,W., Matin,A., Wen,Y., Kwong,K.Y., Bourguignon,L. and Hung,M.C. (
2001
) Nuclear localization of EGF receptor and its potential new role as a transcription factor.
Nature Cell Biol.
  ,
3
,
802
–808.
21.
Offterdinger,M., Schöfer,C., Weipoltshammer,K. and Grunt,T.W. (
2002
) C-erbB-3: a nuclear protein in mammary epithelial cells.
J. Cell Biol.
  ,
157
,
929
–939.
22.
Ni,C.Y., Murphy,M.P., Golde,T.E. and Carpenter,G. (
2001
) Gamma-secretase cleavage and nuclear localization of ErbB-4 receptor tyrosine kinase.
Science
  ,
294
,
2179
–2181.
23.
Wells,A. and Marti,U. (
2002
) Signalling shortcuts: cell-surface receptors in the nucleus?
Nature Rev. Mol. Cell Biol.
  ,
3
,
1
–6.
24.
Yamaguchi,K., Carr,B.I. and Nalesnik,M. A. (
1995
). Concomitant and isolated expression of TGFα and EGF-R in human hepatoma cells supports the hypothesis of autocrine, paracrine and endocrine growth of human hepatoma.
J. Surg. Oncol.
  ,
58
,
240
–245.
25.
Pitot,H.C., Dragan,Y.P., Teeguarden,J., Hsia,S. and Campbell,H. (
1996
) Quantitation of multistage carcinogenesis in rat liver.
Toxicol. Pathol.
  ,
24
,
119
–128.
26.
Kiss,A., Wang,N.J., Xie,J.P. and Thorgeirsson,S.S. (
1997
) Analysis of transforming growth factor (TGF)-alpha/epidermal growth factor receptor, hepatocyte growth factor/c-met, TGF-β receptor type II and p53 expression in human hepatocellular carcinomas.
Clin. Cancer Res.
  ,
3
,
1059
–1066.
27.
Levitzki,A. and Gazit,A. (
1995
) Tyrosine kinase inhibition: an approach to drug development.
Science
  ,
267
,
1782
–1788.
28.
Mendelsohn,J. (
1997
) Epidermal growth factor receptor inhibition by a monoclonal antibody as anticancer therapy.
Clin. Cancer Res.
  ,
276
,
60
–66.
29.
Sambrook,J., Fritsch,E.F. and Maniatis,T. (
1989
) Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
30.
Tata,J.R. (
1974
) Isolation of nuclei from rat liver and other tissues.
Methods Enzymol.
  ,
31
,
253
–262.
31.
Gerner,C., Holzmann,K., Grimm,R. and Sauermann,G. (
1998
) Similarity between nuclear matrix proteins of various cells revealed by an improved isolation method.
J. Cell Biochem.
  ,
71
,
363
–374.
32.
Gerner,C., Fröhwein,U., Gotzmann,J., Bayer,E., Gelbmann,D., Bursch,W. and Schulte-Hermann,R. (
2000
) The fas-induced apoptosis analyzed by high throughput proteome analysis.
J. Biol. Chem.
  ,
275
,
39018
–39026.
33.
Seglen,P.O. (
1976
) Preparation of isolated rat liver cells.
Methods Cell Biol.
  ,
13
,
29
–83.
34.
Parzefall,W., Kainzbauer,E., Qin, H.M., Chabicovsky, M. and Schulte-Hermann, R. (
2002
) Response of isolated hepatocytes from carcinogen sensitive (C3H) and insensitive (C57BL) mice to signals inducing replication or apoptosis.
Arch. Toxicol.
  ,
76
,
699
–706.
35.
Löw-Baselli,A., Hufnagl,K., Parzefall,W., Schulte-Hermann,R. and Grasl-Kraupp,B. (
2000
) Initiated rat hepatocytes in primary culture: a novel tool to study alterations in growth control during the first stage of carcinogenesis.
Carcinogenesis
  ,
21
,
79
–86.
36.
Grasl-Kraupp,B., Bursch,W., Ruttkay-Nedecky,B., Wagner,A., Lauer,B. and Schulte-Hermann,R. (
1994
) Food restriction eliminates preneoplastic cells through apoptosis and antagonizes carcinogenesis in rat liver.
Proc. Natl Acad. Sci. USA
  ,
91
,
9995
–9999.
37.
Grasl-Kraupp,B., Rossmanith,W., Ruttkay-Nedecky,B., Müllauer,L., Kammerer,B., Bursch,W. and Schulte-Hermann,R. (
1998
) Levels of transforming growth factor β and transforming growth factor β receptors in rat liver during growth, regression by apoptosis and neoplasia.
Hepatology
  ,
28
,
717
–726.
38.
Oberhammer,F.A., Pavelka,M., Sharma,S., Tiefenbacher,R., Purchio,A.F., Bursch,W. and Schulte-Hermann,R. (
1992
) Induction of apoptosis in cultured hepatocytes and in regressing liver by transforming growth factor β1.
Proc. Natl Acad. Sci. USA
  ,
89
,
5408
–5412.
39.
Schulte-Hermann,R., Bursch,W. and Grasl-Kraupp,B. (
1995
) Active cell death (apoptosis) in liver biology and disease. In Boyer,J.L. and Ockner,R.K. (eds) Progress in Liver Diseases . Saunders, Philadelphia, pp. 1–35.
40.
Wu,H., Wade,M., Krall,L., Grisham,J., Xiong,Y. and Van Dyke,T. (
1996
) Targeted in vivo expression of the cyclin-dependent kinase inhibitor p21 halts hepatocyte cell-cycle progression, postnatal liver development and regeneration.
Genes Dev.
  ,
10
,
245
–260.
41.
Jochum,W., Passegué,E. and Wagner,E.F. (
2001
) AP-1 in mouse development and tumorigenesis.
Oncogene
  ,
20
,
2401
–2412.
42.
Passegué,E., Jochum,W., Behrens,A., Ricci,R. and Wagner,E.F. (
2002
). JunB can substitute for jun in mouse development and cell proliferation.
Nature Genet.
  ,
30
,
158
–166.
43.
Raper,S.E., Burwen,J.S., Barker,M.E. and Jones,A.L. (
1987
) Translocation of epidermal growth factor to the hepatocyte nucleus during rat liver regeneration.
Gastroenterology
  ,
9
,
1243
–1250.
44.
Ichii,S., Yoshida,A. and Hoshikawa,T. (
1988
) Binding sites of epidermal growth factor in nuclear fraction from rat liver.
Endocrinologica Japonica
  ,
35
,
567
–575.
45.
Marti,U., Burwen,J.S., Wells,A., Barker,M.E., Huling,S., Feren,A.M. and Jones,A.L. (
1991
) Localization of epidermal growth factor receptor in hepatocyte nuclei.
Hepatology
  ,
13
,
15
–20.
46.
Holt,S.J., Alexander,P., Inman,C.B. and Davies,D.E. (
1994
) Epidermal growth factor induced tyrosine phosphorylation of nuclear proteins with translocation of epidermal growth factor receptor into the nucleus.
Biochem. Pharm.
  ,
47
,
117
–126.
47.
Payrastre,B., Nievers,M., Boonstra,J., Breton,M., Verkleij,A.J. and Van Bergen en Henegouwen,P.M.P. (
1992
) A different localization of phosphoinositide kinases, diacylglycerol kinase and phospholipase C in the nuclear matrix.
J. Biol. Chem.
  ,
267
,
5078
–5084.
48.
David,M., Wong,L., Flavell,R., Thompson,S.A., Wells,A., Larner,A.C. and Johnson,G.R. (
1996
) STAT activation by epidermal growth factor (EGF) and amphiregulin. Requirement for the EGF receptor kinase but not for tyrosine phosphorylation sites or JAK1.
J. Biol. Chem.
  ,
271
,
9185
–9188.