Conceived and designed the experiments: SB CB MFB. Performed the experiments: SB NB SN MFB. Analyzed the data: SB DS ALZ AA CB MFB. Contributed reagents/materials/analysis tools: PC. Wrote the paper: SB CB MFB.
The authors have declared that no competing interests exist.
Chronic hepatitis C virus (HCV) infection and associated liver cirrhosis represent a major risk factor for hepatocellular carcinoma (HCC) development. TGF-β is an important driver of liver fibrogenesis and cancer; however, its actual impact in human cancer progression is still poorly known. The aim of this study was to investigate the role of HCC-derived HCV core natural variants on cancer progression through their impact on TGF-β signaling.
We provide evidence that HCC-derived core protein expression in primary human or mouse hepatocyte alleviates TGF-β responses in terms or growth inhibition or apoptosis. Instead, in these hepatocytes TGF-β was still able to induce an epithelial to mesenchymal transition (EMT), a process that contributes to the promotion of cell invasion and metastasis. Moreover, we demonstrate that different thresholds of Smad3 activation dictate the TGF-β responses in hepatic cells and that HCV core protein, by decreasing Smad3 activation, may switch TGF-β growth inhibitory effects to tumor promoting responses.
Our data illustrate the capacity of hepatocytes to develop EMT and plasticity under TGF-β, emphasize the role of HCV core protein in the dynamic of these effects and provide evidence for a paradigm whereby a viral protein implicated in oncogenesis is capable to shift TGF-β responses from cytostatic effects to EMT development.
Epithelial to mesenchymal transition (EMT) is defined as a process in which epithelial cells lose their phenotypic characteristic and acquire mesenchymal cell's features. While EMT is involved in the context of embryonic development it also plays a role in the genesis of fibroblasts during organ fibrosis in adult tissues and might contribute to the metastatic carcinoma development
One of the mechanisms whereby cells undergo neoplastic transformation and escape from normal growth control involves an altered response to the cytostatic effects of TGF-β
TGF-β initiates responses by contacting two types of trans-membrane serine/threonine kinases called receptors type I and type II, promoting activation of the type I by the type II kinase. The activated type I receptor then propagates the signal to the nucleus by phosphorylating Smad2 and Smad3. Once phosphorylated, Smad2 and Smad3 associate with the shared partner Smad4 and the complexes accumulate in the nucleus where they regulate the expression of TGF-β target genes through cooperative interactions with transcriptional partners
Chronic hepatitis C virus (HCV) infection and associated liver cirrhosis represent a major risk factor for hepatocellular carcinoma (HCC) development, and despite epidemiologic evidence connecting HCV infection to HCC, the clinical impact of this virus on hepatocarcinogenesis is still unclear
The structural component of HCV, HCV core protein has attracted particular attention after its characterization and various reports have suggested its potential role in HCV pathogenesis. Indeed, besides its role in viral RNA packaging, HCV core protein has been reported to interact with several cellular proteins such as TNFR
We and others have previously demonstrated an interaction between Smad3 and the HCV core protein
Overexpression of TGF-β and concomitant decrease in hepatocyte growth inhibition is frequently observed in HCC supporting the notion that TGF-β could play a tumor promoting role in liver cancer
In this study, we made use of natural HCV core variants isolated from HCV-related HCC tissues to analyze their impact on the dual function of TGF-β in a pathophysiogically-relevant condition. Thus, we investigated the effects of core protein variants isolated from both tumor or non tumor cirrhotic areas in primary human hepatocytes; indeed, cirrhosis is a well-known preneoplastic condition, associated in at least 90% of cases of HCC. Using these variants we provide evidence for a paradigm in which a viral protein is capable to shift TGF-β responses from cytostatic effects to EMT development.
Recombinant TGF-β1 and recombinant TRAIL/Apo2L were purchased from Abcys, the chemical inhibitor of TGF-β signaling SB-431542 that acts by specifically interfering with the type I receptor
Full length HCV core sequences were amplified from HCV-RNA extracted from tumor (T) or cirrhotic (NT) nodules of a patient (patient B) infected with HCV 1b genotype as previously described
(CAGA)9-Luc was kindly provided by Dr J.M. Gauthier. The expression vectors for HA-TβRI.act, and Flag-TβRImL45.act were a gift from Dr. Y.E. Zhang
To obtain transgenic mice, the HCV core cDNAs isolated from tumor (T) or cirrhotic nodules (NT) were cloned downstream of hepatitis B virus regulatory elements and introduced into C57BL/6 embryos (Institut Clinique de la Souris, Strasbourg, France). Transgenic mice were identified by subjecting 1 µg of tail DNA to amplification by PCR.
The human hepatoma cell line Huh7
Primary mouse hepatocytes were isolated by liver perfusion with a collagenase blend as previously described
Primary human hepatocytes were isolated from the healthy liver tissue of surgical liver biopsy specimens collected after informed consent obtained from patient undergoing therapeutic partial hepatectomy for liver metastasis and benign hepatic tumor. Collagenase (Sigma Aldrich) perfusion (500 µg/ml, 2.4 mg/ml CaCl2 in HEPES buffer, pH 7.4) was preceded by extensive washing of the liver tissue with HEPES/EDTA buffer (pH 7.4) using a catheter inserted into the vessels on the cut surface of the resected fragment. Cells were then washed twice and hepatocytes were separated from nonparenchymatous cells by Percoll fractionation (30% isotonic Percoll solution, centrifuged at 450 g for 4 min) and immediately infected at 37°C for 2 h with lentiviral vectors, washed and plated in Williams medium supplemented as described elsewhere
TRIP-ΔU3-CMV-T, TRIP-ΔU3-CMV-NT and TRIP-ΔU3-CMV-Cinv vectors were obtained by substituting GFP in TRIP-ΔU3-CMV-GFP with cDNA coding for HCV core sequences. An inverted core sequence TRIP-ΔU3-CMV-Cinv was used as a control.
Vector particles were produced by the transient calcium phosphate cotransfection of 293T cells as a previously described
Cells were washed twice with PBS and lysed in RIPA buffer containing 0.5% SDS and Benzon nuclease. Proteins were quantified with the Bio-Rad protein assay (Bio-Rad, France) and 30 µg of extracts were separated on SDS polyacrylamide gel, transferred on nitrocellulose membrane and blotted using different primary antibodies directed against HCV core protein, E-cadherin, Fibronectin (Santa Cruz Biotechnology), Vimentin (Chemicon), phospho-Smad3 (Cell signaling), Smad3 (Abcam), Flag, Myc and HA tags (Sigma). Membranes were revealed using a chemioluminescence detection kit (ECL Plus, GE Healthcare).
Primary mouse hepatocytes were cultured for 48 h with or without TGF-β (2 ng/ml) and routine stain hematoxylin-eosin was performed after fixation of cells with EtOH 70% at 4°C for 15 min.
Cells were washed with PBS and fixed with a 4% PFA solution at 4°C for 20 min followed by methanol permeabilization for 5 min at −20°C. Cells were then incubated with a primary mouse anti-vimentin, rabbit anti-αSMA, or rabbit anti-E-cadherin antibody and then with an Alexa Fluor 488 conjugated anti-mouse antibody and an Alexa Fluor 594 conjugated goat anti-rabbit antibody (Molecular Probes). They were then stained with Hoechst and examined by fluorescence microscopy.
Cell proliferation was assessed by BrdU incorporation (Roche), cell viability and caspase 3 activity were estimated using a Celltiter-Glo luminescent cell viability assay or the CaspaseGlo 3/7 assay respectively (Promega) according to the manufacturer's instructions.
Mitochondrial transmembrane potential (ΔΨm) was evaluated by staining cells (106) with the fluorescent dye DiOC6 at a final concentration of 40 nM for 15 min at 37°C. Cells were immediately dissociated by trypsin and their fluorescence estimated by analysis with a FACScan flow cytometer (Becton-Dickinson) using the FL1 channel
Flow cytometric analysis and sorting were performed using a FacsDiva flow cytometer (Becton Dickinson Immunocytometry Systems). Forward Scatter (FSC) and side scatter (SSC) were collected through a filter. The GFP signal was collected in the FL1 channel. A light gate was drawn in the SSC versus FSC to exclude dead cells/debris. Cells in the gate were displayed in a biparameter histogram (FS versus FL1) and final gating settings determined to collect the labeled cells. GFP positive cells were sorted at 5000 cells/sec.
Cells were cotransfected with vectors coding for the gene of interest together with the CAGA-luc reporter plasmid and the Renilla luciferase plasmid to normalize the results. They were incubated 24 h later in the absence or presence of TGF-β for another 18 h. Luciferase activity was measured with the Dual Luciferase reporter assay (Promega) system according to the manufacturer's instructions.
The significance between the different conditions and their control was determined by paired Student's
We have previously demonstrated that, when transiently expressed in hepatic cells, HCV core proteins isolated from tumor or cirrhotic nodules bind Smad3 differently and that this interaction inhibits Smad3-dependent transcriptional activity
(A,B,C) Mouse hepatocytes obtained from livers of transgenic mice expressing or not HCV core proteins isolated from tumor (T) or cirrhotic (NT) tissues were treated with TGF-β for 48 h before determination of cell proliferation, estimated by BrDU incorporation (A), cell viability (B) or caspase 3 activity (C). (D) Cells were treated with TRAIL (20 ng/ml) for 18 h before determination of caspase3 activity. Results represent the mean+/−SD of triplicates from a representative experiment. * p≤0.05, ** p≤0.005, *** p≤0.0005.
Several lines of evidence support the notion that epithelial cancer cells lose their capacity to respond to TGF-β cytostatic effects but in some cases retain their ability to respond to other TGF-β -mediated functions such as EMT. The observation that HCV core proteins interfere with the ability of TGF-β to execute cell growth inhibition and cell killing prompted us to consider the possibility that these proteins might influence TGF-β mediated EMT. Since recent findings have demonstrated that TGF-β could induce an EMT in mature mouse hepatocytes
(A) Morphologic changes of mouse hepatocytes expressing or not HCV T core protein observed after 48 h of culture with or without TGF-β (2 ng/ml). (B) Hepatocytes isolated from transgenic mice expressing HCV core proteins were treated with TGF-β (2 ng/ml) or SB431542 (1 µM) for 48 h and expression of αSMA was examined by immunofluorescence using a αSMA antibody. Data are representative of three independent experiments. (C) Hepatocytes isolated from transgenic mice expressing HCV core proteins were treated with TGF-β or SB431542 for 48 h and expression of E-cadherin was determined by Western blotting. Anti-p38 western blotting was used as control loading. Data are representative of three independent experiments.
To obtain further evidence that HCV core proteins could modulate the magnitude of the negative growth regulatory effects of TGF-β we also performed experiments in human primary hepatocytes. Freshly isolated hepatocytes were infected with lentiviruses coding for the T or NT core variants or an inverted core sequence as control. Western blot analyses confirmed the expression of the core proteins (
Freshly isolated cells were infected with lentiviruses encoding the HCV core protein variants or an inverted core sequence as control (CTL) (A) Levels of core expression were estimated by Western blot analysis different time points after lentivirus transduction. (B, C) Determination of cell viability (B) or caspase 3 activity (C) was performed after 96 h of treatment with TGF-β (5 ng/ml). Results represent the mean+/−SD of triplicates from a representative experiment. * p≤0.05, *** p≤0.0005.
Although TGF-β -mediated EMT has been described in primary mouse or rat hepatocytes as well as in cancerous human cells, no such study has been yet investigated in primary human hepatocytes in vitro. Interestingly, we observed that human hepatocytes could express stress fibers as spikes mainly located in membrane protrusions under TGF-β treatment (
(A) Expression of αSMA or Vimentin was estimated by immunofluorescence analysis after treatment with TGF-β (5 ng/ml) or SB431542 (1 µM). (B) Expression of Fibronectin or E-Cadherin was estimated by Western blot analysis in the same experimental conditions. Data are representative of three independent experiments.
Western blots analyses evidenced a lower expression of E-cadherin after TGF-β treatment which was totally recovered in the presence of the TbRI inhibitor. On the contrary, expression of the mesenchymal marker fibronectin was greatly increased by TGF-β (
Taken together these data strongly suggest that HCV core interfere with TGF-β responses in terms of cell growth inhibition and apoptosis in hepatocytes isolated from transgenic mice as well as human primary hepatocytes. Remarkably, TGF-β responses, in terms of EMT are increased by expression of T or NT core protein variants in both mouse and human hepatocytes. This might reflect both direct effects of core on TGF-β-induced EMT and reduction of TGF-β induced apoptosis by the core protein, allowing more cells to undergo EMT as compared to control cells.
In order to dissect the molecular mechanisms activated by the HCV core protein, we established Huh7 cell lines stably expressing the T core protein (
(A) Expression of HCV core protein determined by Western blot analysis. (B, C) Cells were treated with TGF-β (5 ng/ml) for 48 h before determination of cell viability (B) or caspase3 activity (C). Results represent the mean+/−SD of triplicates from a representative experiment. *** p≤0.0005 (D) Mitochondrial membrane potential (ΔΨm) was estimated by FACS analysis in cells treated with TGF-β for 48 h. After staining with DiOC6(3), cells with low fluorescence intensity corresponding to low (ΔΨm) were gated and their number expressed as a percentage of the total population. A representative experiment is shown. (E) Cells were treated with TGF-β for 48 h and E-cadherin or αSMA expression was assessed by immunofluorescence analysis. Data are representative of three independent experiments. (F) Comparative expression of HCV core proteins. Extracts from cultured cells expressing the core protein (Huh7, human or mouse primary hepatocytes) or from livers of HCV-related HCC patients were analyzed by Western blot. Anti-p38 western blotting was used as control loading.
We then determined EMT process in Huh7 cell lines expressing this core protein. Immunofluorescence studies showed that αSMA was highly polymerized after TGF-β treatment associated with a strong decrease of E-cadherin from the cell membranes (
All together, these data indicate that the effects of HCV core proteins on TGF-β responses observed in primary hepatocytes were reproduced in a human hepatoma cell line that could thus constitute an useful tool to dissect the mechanisms that are involved in the modulation of TGF-β responses.
We also compared protein core expression in our different cellular models and in extracts from liver of HCV/HCC patients. The strongest expression was obtained in human hepatocytes, which is consistent with an efficient lentiviral transduction. HCV core protein expression could be also detected in different liver extracts although at different levels. Interestingly, core expression in these extracts was comparable to the one observed in mouse hepatocytes (
To analyze in more details the contribution of Smad activation in the effects of HCV core on TGF-β responses, we made use of a mutant of the TGF-β receptor I, TβRImL45Act that retains a constitutively active kinase domain but is unable to induce Smad phosphorylation. Huh7 cells were transfected with this mutant or with the wild type activated form of TβRI, together with a plasmid coding for the HCV core and GFP to detect the transfected cells. Immunofluorescence analysis was performed 48 h later. A marked polymerization of αSMA was observed through expression of the constitutively active TβR1 that was similar or even greater when cells also expressed the HCV core protein (
(A) Huh7 cells were co-transfected GFP together with TbR1act or TbR1l45M act plasmids in the presence or absence of HCV core vector. Immunofluorescence analysis was performed 48 h later with an anti αSMA antibody. (B) Expression of TbR1act or TbR1l45M act and HCV core protein were assessed by Western blotting using anti-HA, anti-Flag or anti-core antibodies respectively.
To confirm this result, we established different independent Huh7 cell clones, stably expressing or not the HCV core protein, in which the expression of endogenous Smad3 was reduced by stable expression of a short-hairpin RNA (shRNA). As expected, Smad3 depletion prevented TGF-β -induced expression of the CAGA-luc reporter plasmid in the four independent clones tested, two of them expressing the core protein (
(A) Smad3 expression determined by Western blot analysis in four independent clones selected after stable transfection with pRetroSuper-shRNA-Smad3 plasmid. Anti-p38 antibody was used as control loading. (B) Different clones were transfected with the CAGA-luc reporter plasmid and treated or not with TGF-β (5 ng/ml) for 18 h before determination of luciferase activity. Results were normalized with renilla luciferase and represent the mean of triplicates+/−SD. (C, D) Different clones were treated with TGF-β for 48 h before determination of cell viability (C) or caspase3 activity (D). Results represent the mean+/−SD of triplicates from a representative experiment. * p≤0.05, *** p≤0.0005, NS : not significant. (E) Different clones were treated with TGF-β for 48 h and αSMA polymerization was estimated by immunofluorescence analysis.
We next investigated the possibility that different threshold levels of Smad3 contribute to the differential effects of TGF-β on apoptosis or EMT. For this, we reintroduced increasing amounts of Smad3 in Huh7-shRNA-Smad3 clones (
(A) Huh7-shRNA-Smad3 cells (Clone 3) were transfected with increasing amounts of Myc-Smad3 expression vector together with pCMV renilla luciferase. Smad3 protein expression was evaluated 24 h later by Western blot analysis using an anti-Myc antibody and loading was normalized with renilla luciferase expression. (B) Huh7-shRNA-Smad3 cells (Clone 3) were cotransfected with the CAGA-luc reporter plasmid and increasing amounts of Myc-Smad3 vector together with pCMV renilla luciferase. 24 h later, they were treated or not with different doses of TGF-β for 18 h before determination of luciferase activity. Results were normalized with renilla luciferase and represent the mean+/−SD of triplicates from a representative experiment. (C, D) Huh7-shRNA-Smad3 cells (Clone 3) were transfected with increasing amounts of Myc-Smad3 vector together with pGFP plasmid and sorted by FACS 24 h later on the basis of GFP expression. Cells were then cultured for 24 h and treated with different doses of TGF-β for 48 h before determination of cell viability (C) or caspase3 activity (D). * p≤0.05, ** p≤0.005, *** p≤0.0005, NS : not significant. (E) Huh7-shRNA-Smad3 cells (Clone 3) were transfected with increasing amounts of Myc-Smad3 vector together with pGFP plasmid and sorted 24 h later on the basis of GFP expression. αSMA expression was estimated by immunofluorescence analysis after treatment with TGF-β (1 ng/ml) for 48 h.
The GFP positive cells were also analyzed for αSMA expression and polymerization after TGF-β treatment. In contrast with apoptotic data, TGF-β-induced EMT could occur in the context of low Smad3 expression (
Our study offers relevant observations regarding both the mechanisms of HCV-related carcinogenesis and the impact of TGF-β in human cancer. Indeed, we provide evidence that HCC-derived HCV core proteins alleviate cell growth inhibition and apoptosis mediated by TGF-β indicating a biological significance of the binding of HCV core protein to Smad3. This effect was not restricted to stably transfected cell lines, since it was also observed in primary mouse hepatocytes isolated from transgenic animals expressing the core proteins as well as in primary human hepatocytes infected in vitro with lentiviruses encoding the same variants. Thus HCV core protein has also the potential to negatively impact the cytostatic actions of TGF-β in systems that may better reflect an
Interestingly, in cells expressing HCV core proteins TGF-β was still able to reduce E-cadherin expression and increase αSMA expression and polymerization that are hallmarks of EMT. These alterations were associated with the ability of these cells to exhibit anchorage independent growth. Importantly, we also observed that core protein expression was sufficient to provoke EMT induction in primary hepatocytes. This effect was reverted by addition of a specific inhibitor of TGF-β I receptor thus demonstrating a TGF-β dependent effect of core on EMT development. . These data emphasize a differential effect on TGF-β actions in terms of apoptosis or EMT.
Different levels of HCV core expression have been observed in HCV-derived HCC at the mRNA level or in immunohistochemistry
It is commonly accepted that TGF-β signaling pathway plays a tumor suppressor role thought to be associated with growth inhibitory and apoptotic responses and a tumor promoter role thought to reflect the positive effects of TGF-β on tumor cell invasion. Taken together, our data suggest that HCV core, by reducing Smad3 signal strength, renders the cells to become less sensible to tumor suppressive effects of TGF-β although they retain the tumor promoting effects, assuming that Smad3 may regulate different targets in function of its level of activation. This is consistent with the notion that critical signal amplitude may be needed to evoke a biological effect. In addition to Smad pathways, non-smad-dependent signal transduction downstream of TGF-β receptors has been proposed
It has been recently reported that hyperactive Ras mediates a decrease in TGF-β -induced Smad3 phosphorylation in the COOH-terminal and an increase in JNK-induced Smad3 phosphorylation in the linker region, shifting the TGF-β pathway from a tumor suppressive to an invasive capacity in human colorectal as well as hepatic carcinogenesis
Although activation and transdifferentiation of hepatic stellate cells are still regarded as key mechanisms of fibrogenesis, recent studies have pointed out that other liver cells, including hepatocytes may contribute to the pool of myofibroblasts in fibrosing liver. Our results showing that TGF-β is able to induce EMT in primary mouse and human hepatocytes add further evidence for this concept. Furthermore, because HCV replicates in hepatocytes, the fact that EMT could develop in HCV core-expressing cells under TGF-β might provide a new notion to explain the fibrotic effect of this virus.
In conclusion, our data ties together TGF-β and HCV which are both known to be keys in the development of fibrosis and HCC, highlight the ability of hepatocytes to develop EMT under TGF-β and emphasize the role of HCV core protein in the dynamic of these effects.
Moreover, one report has suggested EMT as a mechanism of Epstein-Barr virus-related tumor cell dissemination
We would like to thank Dr. J. Massagué for providing the Smad3 shRNA plasmid, D. Clay for cell sorting and C. Desterke for statistical analysis.