Potential conflict of interest: Nothing to report.
Financial support: National Institutes of Health (NIH): K08DK080980 (to S.C.); K01DK082613 (to S.B.); Duke Center for AIDS Research: K08DK076598, R03DK084120 (to R.J.).
Studies of the hepatitis C virus (HCV) life-cycle rely heavily on Huh7.5 cells, but the reasons why these cells are exceptionally permissive for HCV replication are not clear. Based on recent clinical observations, we hypothesized that the Hedgehog (Hh) pathway, which has not been previously associated with HCV replication, may be involved in the Huh7.5 phenotype of increased permissiveness. We tested this hypothesis by comparing levels of a variety of Hh-related cellular markers in Huh7.5 cells with the parental Huh7 cells, which are far less permissive. Here we demonstrate that Huh7.5 cells, when compared with Huh7 cells, have substantially decreased expression of epithelial markers, increased levels of mesenchymal markers, and markedly up-regulated Hh pathway activity: Shh, >100-fold, Gli1, >30-fold, Ptc, 2-fold. In Huh7.5 cells, we found that cyclopamine, an Hh pathway antagonist, reduced HCV RNA levels by 50% compared with vehicle and inactive isomer controls. Moreover, in Huh7 cells treatment with recombinant Shh ligand and SAG, both Hh pathway agonists, stimulated HCV replication by 2-fold and 4-fold, respectively. These effects were observed with both viral infections and a subgenomic replicon. Finally, we demonstrated that GDC-0449 decreased HCV RNA levels in a dose-response manner. Conclusion: We have identified a relationship between HCV and Hh signaling where up-regulated pathway activity during infection promotes an environment conducive to replication. Given that Hh activity is very low in most hepatocytes, these findings may serve to further shift the model of HCV liver infection from modest widespread replication in hepatocytes to one where a subset of cells support high-level replication. These findings also introduce Hh pathway inhibitors as potential anti-HCV therapeutics. (HEPATOLOGY 2011;)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hepatitis C virus (HCV) is an important cause of chronic liver disease, with the severe consequences of hepatocellular carcinoma (HCC) and cirrhosis occurring in some patients.1, 2 When considering determinants of HCV persistence and propagation of infection, little consideration has been given to differences between cells within the liver. Recent studies have demonstrated HCV Core protein localized to discrete foci within HCC sections from patients, and laser captured microdissection samples indicated that HCV genomes exist at unexpectedly low average copy numbers per cell.3, 4 These observations suggest that HCV infection is not widespread throughout the liver, but rather selective or restrained in its target cells.
In vitro studies of HCV rely heavily on the Huh7.5 cell line. This cell line was generated after Huh7 cells selected for harboring an HCV subgenomic replicon were cured of replicating viral RNA with interferon-α.5 The resulting cells were highly permissive for HCV replication when retransfected with replicon constructs. As they support replication of the JFH1 viral isolate and produce infectious virus in tissue culture, Huh7.5 cells have propelled studies of the HCV life-cycle forward.6 Similar cell lines with increased HCV permissivity, like LH86 cells, have been directly isolated from patient samples, although HCV RNA levels are 1-2 log lower compared with Huh7.5 cells.7 The reasons for Huh7.5 cells being exceptionally permissive for HCV replication were attributed to a defect in RIG-I, a pattern recognition receptor that activates type I interferon expression during viral infection.8 However, recent studies found no RIG-I defects in novel cell lines also generated from Huh7 cells with increased permissiveness to HCV.9, 10 Thus, RIG-I alone may not explain this phenomenon.
The Hedgehog (Hh) pathway plays an important role during embryogenesis, normal tissue growth, regeneration after injury, and carcinogenesis.11-15 Most hepatocytes in healthy adult livers do not express detectable Hh ligands Sonic hedgehog (Shh) or Indian hedgehog (Ihh), whereas some Hh ligand expression can be demonstrated in ductular-type cells.16 After injury, Shh expression increases, causing Gli family transcription factors to accumulate in Hh-responsive cells as part of the regenerative and fibrotic responses.14, 15 Hh pathway activation has also been observed in some HCC cell lines, although significant heterogeneity exists within actual tumors.16 A vigorous debate exists as to whether liver epithelial cells, such as cholangiocytes and hepatocytes, undergo epithelial-to-mesenchymal transitions (EMT) in injured livers, but some evidence supports this concept and suggests Hh-mediated regulation.17-21 In viral hepatitis patients, recent data suggests EMT may occur in response to infection.22 HCV infection of hepatoma cell lines in vitro alters cell polarity to expose gap junction complex proteins key to viral entry.23 To our knowledge, such studies have not addressed the effects of altered cell polarity on HCV replication, or mechanisms by which viral infection might promote EMT.
We hypothesized that Huh7.5 cells are highly permissive for HCV because they possess a “transitional” phenotype skewed toward mesenchymal characteristics due to increased Hh pathway activity. We subsequently asked whether Hh pathway activation may create an environment conducive to viral replication and whether Hh pathway inhibition would inhibit HCV replication.
EMT, epithelial-to-mesenchymal transition; Hh, Hedgehog; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; Ihh, Indian hedgehog; LDH, lactate dehydrogenase; qRT-PCR, quantitative reverse-transcription polymerase chain reaction; Shh, Sonic hedgehog.
Materials and Methods
Cells, Constructs, and Reagents.
Huh7 cells, Huh7.5 cells (a gift from C. Rice, Rockefeller University), LH86 cells (a gift from C. Liu, University of Florida), and HepG2 cells were used for these studies. Primary human hepatic stellate cells were isolated and cultured as described17 and primary human hepatocytes were commercially prepared (a gift from R. Witek, Invitrogen, Durham, NC). JFH1 cDNA (kindly provided by T. Wakita, National Institute of Infectious Diseases, Tokyo, Japan) was used to generate cell culture HCV, and HCV Con1 replicon (a gift from C. Rice, Rockefeller University) to study HCV replication. Primer sequences for quantitative reverse-transcription polymerase chain reaction (qRT-PCR) reactions are listed in Supporting Table 1. Purchased reagents were: Interferon-α A/D (Sigma-Aldrich, St. Louis, MO), cyclopamine and tomatidine (Toronto Research Chemicals, Toronto, Canada), recombinant N-terminal mouse Shh (StemCell Technologies, Vancouver, Canada), SAG (Enzo Life Sciences, Plymouth Meeting, PA), GDC-0449 (Selleck Chemicals, Houston, TX), and mouse immunoglobulin G (IgG)1 isotype control antibody (R&D Systems, Minneapolis, MN). Antibodies to Shh and Gli1 (Santa Cruz Biotechnology, Santa Cruz, CA), α-SMA (Dako, Carpinteria, CA), and α-tubulin (Sigma-Aldrich) were used for immunoblotting and α-HCV Core (C7-50, Abcam) was used for immunofluorescence/immunoblotting.
Messenger RNA Quantification by RT-PCR.
Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA), reverse-transcribed using Superscript reverse transcriptase (Invitrogen). cDNA samples were used in triplicate for qRT-PCR using iQ-SYBR Green Supermix (Bio-Rad Laboratories).17 Target gene levels are presented as ratio of levels detected in treated cells over levels detected in control cells, according to the ΔΔCt method.17
HCV Replicon, JFH1 Infections, and Focus-Forming Assay (FFU).
Huh7.5 cells were grown in supplemented Dulbecco's modified Eagle medium (DMEM).5 HCV replicon-harboring Huh7.5 cells were established by transfecting in vitro transcribed Con1 replicon RNA followed by selection with 1 mg/mL G418.24 After selection, cells were propagated in DMEM with 0.75 mg/mL G418. Infectious JFH1 virus was obtained by transfection of Huh7.5 cells with in vitro transcribed RNA and harvesting of cell supernatant as described.6, 25 To generate viral stocks, clarified supernatant was used to infect naïve Huh7.5 cells, supernatants were recovered 7 days postinfection, and concentrated using an Amicon 100k device. Virus-containing supernatants were titered by FFU as described.25 Briefly, serial 10-fold dilutions of samples were plated in triplicate on 96-well plates containing subconfluent Huh7.5 cells. After 72 hours of incubation, cells were washed with phosphate-buffered saline (PBS) and fixed with ice-cold methanol. Infected cells were subsequently identified by immunofluorescence using mouse α-Core antibody (C7-50; Abcam) and FFU/mL values were calculated using the mean of three separate wells per sample.
Time course infections-protein analysis: Huh7.5 cells (3 × 105) were seeded onto four T-25 flasks and two of the flasks were infected the following day with HCV at a multiplicity of infection of 0.5-1.0. The virus was removed 16 hours postinfection and replaced with normal growth medium. Uninfected cells were split and grown in culture for the same time as their respective infected culture counterparts. Cells were harvested and lysed in RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 0.5% sodium deoxycholate, 1% NP40, 0.1% SDS) on day 2 and day 5 postinfection. The cell lysate was collected in 1.5-mL microcentrifuge tubes and allowed to sit on ice for 15 minutes prior to centrifugation at 10,000g. The supernatant was collected as total cell lysate and amount of protein was estimated using the Bradford method.
Drugs were incubated with cells prior to RNA and protein isolation. The following concentrations and incubation times were used: cyclopamine and tomatidine (5 μM for 48 hours for initial experiment; 24, 48, and 72 hours for time course), Shh (100 ng/mL for 48 hours), SAG (0.3 μM for 24 hours), interferon-α (500 U/mL for 48 hours), GDC-0449 (range 0-25 μM for 24 hours), 5E1 (Shh neutralizing antibody, 10 μg/mL, for 48 hours), mouse IgG1 isotype control (10 μg/mL, for 48 hours). If the experiment was done with the JFH1 HCV, drug was added at the time of infection.
Lactate Dehydrogenase (LDH) Release Assay.
Supernatant media was collected from Huh7.5 cells infected and uninfected cells after 72 hours. LDH assay was performed according to the manufacturer's protocol (LDH Cytotoxicity Detection Kit, Takara Bio, Japan). Uninfected cells lysed with Triton-X were used as positive control.
Huh7.5 Cells Are More Permissive for JFH1 HCV Infection, Have Decreased Epithelial Markers, Increased Mesenchymal Markers, and Markedly Higher Hh Pathway Activity.
We first compared Huh7.5 cells to Huh7 cells grown under standard conditions and infected with JFH1 HCV to confirm that the cell lines in our laboratory follow the observation made by other laboratories. Indeed, we found that Huh7.5 cells were more permissive for HCV infection than Huh7 cells infected under the same conditions (Fig. 1A). We next compared uninfected cells grown under standard conditions to assess differences in baseline gene expression. Analysis focused on markers that are differentially expressed primarily in epithelial and mesenchymal cells, as well as Hh pathway markers involved in regulating the “transitional” state. We found that Huh7.5 cells expressed significantly reduced transcript levels of the epithelial markers E-cadherin and keratin 19 (Krt19) (Fig. 1B), and significantly increased transcript levels of mesenchymal markers α-smooth muscle actin (αSMA), collagen, type I, alpha 1 (Col1α1), and S100 calcium-binding protein A4 (S100A4) (Fig. 1C). We also noted highly significant increases in Hh pathway component transcript levels in Huh7.5 cells compared with Huh7 cells: Shh >150-fold, Gli1 >30-fold, and Patched >2-fold (Fig. 1D).
We next asked if these findings were unique to Huh7.5 cells or also observed in LH86 cells, another cell line permissive for HCV replication.7 We found LH 86 cells had: 1) increased transcript levels of mesenchymal markers, significantly higher than Huh7 cells but lower than Huh7.5 cells (Supporting Fig. 1A); 2) markedly increased Shh transcript levels comparable to Huh7.5 cells; 3) significantly higher levels of Gli1 and Ptc transcripts compared with Huh7 cells, but significantly lower than Huh7.5 cells (Supporting Fig. 1B). We confirmed the relevance of observed changes in messenger RNA (mRNA) levels by analyzing protein expression. αSMA and Shh proteins were undetectable in Huh7 cells, but robustly expressed in both Huh7.5 and LH86 cells (Supporting Fig. 2). To further confirm our results, we profiled HepG2 cells, which are nonpermissive for HCV infection, and determined that these cells exhibit markers consistent with an epithelial phenotype and express Shh and Gli1 at levels even lower than parental Huh7 cells and far lower than Huh7.5 cells (data not shown).
Interferon Treatment of Huh7 Cells Causes Increased Mesenchymal Marker Transcripts and Higher Transcripts of Upstream and Downstream Hh Pathway Components.
We evaluated the possible role of interferon treatment in the genesis of the Huh7.5 cell phenotype by asking if Huh7 cells treated with interferon-α demonstrated similar changes in gene expression. We observed that interferon-α caused the expected changes in Huh7 cells: increased αSMA and Col1α1 transcripts; significantly increased Shh and Gli1 transcripts (Supporting Fig. 3).
Hh Pathway Inhibition and Activation Modulate HCV Replication.
Given the significantly up-regulated Hh pathway activity in Huh7.5 cells, we examined if manipulation of the Hh pathway would inhibit or promote HCV viral RNA replication. Huh7.5 cells were infected with JFH1 HCV and treated with cyclopamine (Smoothened antagonist), tomatidine (biologically inactive isomer), or vehicle control.26 Compared with control treatments, cyclopamine reduced cell-associated HCV RNA by 70%, mirroring the observed decreases in Shh and Gli1 expression (Fig. 2A). Cyclopamine treatment also noticeably reduced the extent of infection as ascertained by immunofluorescence using antibody to the viral Core protein (Fig. 2B). Reductions in HCV Core content correlated strongly with the drop in Shh expression that followed cyclopamine treatment.
To further characterize the effect of cyclopamine on infected cells, we performed a time course experiment in which we isolated RNA from JFH1 infected Huh7.5 cells treated with cyclopamine and controls at 24, 48, and 72 hours postinfection (Supporting Fig. 4). HCV RNA mirrored reductions in Gli1 RNA beginning at 24 hours and maximizing by 48-72 hours. In order to ascertain whether cyclopamine treatment was associated with changes in cell viability, we performed an analysis of LDH levels in supernatant media under different conditions. LDH levels were comparable between uninfected and infected cells regardless of treatment (Supporting Fig. 5), indicating that reduced HCV replication was not due to potential toxic effects of cyclopamine treatment. Finally, we performed a FFU assay to quantify infectious virus from supernatant media at 72 hours from infected cells after cyclopamine and control treatment (Supporting Fig. 6). Cyclopamine treatment led to a one log reduction in focus forming units/mL compared with control treated cells.
To further verify these results obtained with pharmacologic inhibition, we also used a neutralizing antibody to Shh (5E1) to inhibit pathway activity in Huh7.5 cells and observed similar reductions in HCV RNA, Shh, and Gli1 observed with chemical inhibition (Fig. 2C).
We next examined if recombinant N-terminal fragments of Shh, an agonist of the Hh pathway, would promote HCV viral titers in Huh7 cells. Incubation with exogenous Shh for 48 hours produced increased Shh and Gli1 transcripts and caused a 2-fold increase in HCV RNA levels (Fig. 3). It should be noted that JFH1 infection alone produced increased Shh and Gli1 transcripts and protein, and the increase in Shh expression paralleled the increase in core protein over time postinfection (Supporting Fig. 7).22 We replicated the increase in Huh7 permissiveness results using SAG, an Hh agonist that acts downstream by directly binding to Smoothened. SAG treatment resulted in a 3-fold increase in Hh pathway transcripts, and a corresponding 3-fold increase in HCV RNA levels (Fig. 4A). Corresponding increases in protein expression levels were observed (Fig. 4B). To confirm that this increase in HCV RNA correlated with functional virus, we used supernatants collected from the above experiment to infect naïve Huh7 cells. We observed foci of HCV Core-staining cells in Huh7 cells infected with supernatants from SAG-treated cells but no foci in cells infected with supernatants from untreated or vehicle-treated cells (data not shown). SAG also potentiated HCV RNA accumulation in Huh7.5 cells, with a 3-fold increase in Shh and Gli1 transcripts accompanied by a 4-fold increase in HCV RNA levels (Fig. 5A). Corresponding increases in protein expression levels were observed (Fig. 5B). We subsequently treated commercially prepared primary human hepatocytes with SAG to assess if they would support HCV replication. Given that mature hepatocytes do not express Hh pathway intermediaries and have little detectable Hh activity, these cells were not SAG-responsive and did not support HCV replication (data not shown).
Hh Pathway Activity Modulates Replication of a Subgenomic HCV Replicon.
We next attempted to determine if up-regulated Hh pathway activity promotes HCV RNA replication versus some other event such as assembly or entry. We treated Huh7.5 cells harboring the Con1 HCV subgenomic replicon with cyclopamine, tomatidine, or vehicle. Cyclopamine-treated cells exhibited a 50% reduction in HCV RNA compared with cells treated with either tomatidine or vehicle, corresponding to the reduction in Shh and Gli1 transcript levels (Fig. 6).
We also performed an experiment to assess the effect of Hh pathway on HCV cell entry. Huh7.5 cells were infected after 24 hours incubation with cyclopamine, tomatidine, or vehicle plus or minus the presence of antibody to CD81. HCV RNA levels at 4 hours postinfection were equivalent in infected cells in the presence of absence of cyclopamine (data not shown). Antibody to CD81 inhibited association of HCV with target cells under all conditions. This suggests that the Hh pathway promotes HCV replication, at least partially through enhancing HCV RNA synthesis and/or translation, and does not alter viral attachment.
GDC-0449 Has Anti-HCV Activity.
Although cyclopamine treatment was able to reduce HCV viral titers, this agent has well-described toxicity, and thus, no potential as a future anti-HCV pharmaceutical. In contrast, GDC-0449 is a Smoothened antagonist currently in phase 1 and 2 clinical studies as a chemotherapeutic agent for various malignancies, with an excellent safety profile and thus much greater potential as an HCV treatment.27-30 Therefore, we tested GDC-0449 in Huh7.5 cells infected with the JFH1 virus. GDC-0449 at 5 μM concentration inhibited HCV RNA by >50% when compared with untreated or vehicle-treated cells (Fig. 7A). Reductions in HCV RNA mirrored the decreases in Shh and Gli1 transcripts, and similar reductions in protein levels were observed (Fig. 7B). We subsequently determined that GDC-0449 resulted in Hh pathway and HCV RNA inhibition in a dose-response fashion beginning at concentrations as low as 0.05 μM with a plateau at 5 μM (Fig. 7C). Based on this curve, the IC50 is estimated to be 0.16 μM.
We have demonstrated a significant association between HCV infection and Hh pathway activity in liver-derived cells. Cells with dramatically increased Hh pathway activity such as Huh7.5 and LH86 cells support higher levels of HCV replication compared with Huh7 cells with virtually undetectable Hh pathway activity. Pharmacological and antibody treatment to either antagonize or promote Hh signaling in liver cells at different points in the pathway had the same direct effects on HCV replication. Included in the current study was GDC-0449, an Hh pathway antagonist already in preclinical development for other indications, which we have shown can inhibit HCV replication.
The observation that HCV replication is closely associated with Hh pathway activity has not been previously appreciated. Certainly it has never been explored as an explanation for why Huh7.5 cells are exceptionally permissive for HCV replication. Hh pathway components were previously identified in a functional genomic primary screen, but were not identified in the secondary screen.31 This may be due to the broad changes in the intracellular environment induced by Hh pathway activation that may not be detected in single gene small interfering RNA (siRNA) knockdowns. It may be that the intracellular changes that occur upon Hh pathway activation and its possible association with EMT are part of the same continuum when considering an environment conducive to HCV infection. The concept of a continuum is supported by the mixed phenotype we detected in LH86 cells compared with Huh7.5 cells and how that correlates with the original observation that LH86 cells were less permissive than Huh7.5 cells.7 One of the key remaining questions is at what level does Hh pathway activation exerts its effects on HCV replication. Although we used multiple agonists and antagonists, these agents all act close to the “top” of the pathway. We suspect the changes may occur further downstream and are likely to involve other pathways known to be associated with Hh in promoting an environment conducive for HCV replication.
Our results may be further cause to reevaluate the model of HCV infection in the liver. Liang et al.3 demonstrated in their analysis of explanted livers from HCV patients with HCC that Core protein was detected in a minority of cells despite this advanced-stage of disease. Perhaps Huh7.5 cells are representative of a minor population of liver cells that retain an active Hh pathway.18 HCV may preferentially infect and replicate in this minority population of cells so hospitable to efficient replication. Infection of these cells combined with a possible contribution of the resulting interferon release may induce further Hh pathway activation, including increased Shh ligand production. Moreover, increased Shh ligand may serve as a paracrine factor that allows neighboring cells to become more easily infected by promoting a “transitional” phenotype, exposing the gap junction complexes to facilitate viral entry and/or altering the intracellular environment.
The association between HCV and the Hh pathway will not alone settle the vigorous debate regarding the occurrence of EMT within the liver.17, 19-21, 32 The results we present should be examined when considering the cellular environment that is most conducive for replication. Given that HCV has evolved to infect humans over a long period of time, like other viruses, it has developed patterns of infection that most benefit its propagation and/or persistence. Recent data suggests that dedifferentiated cells within hepatoma cell lines had increased Hh activity and demonstrated increased proliferation and invasion in a mouse xenograft model.33 Given the increased proliferation of these dedifferentiated cells with increased Hh activity, one could hypothesize that HCV prefers this type of cell as a target because proliferation allows persistence of chronic infection.
The results of our studies have certain limitations that must be acknowledged. All of these observations were made in vitro, and should be interpreted with caution when considering in vivo pathogenesis. Our in vitro evidence linking Hh signaling and viral infection are supported by observations that have been made about Hh pathway activity in liver samples from patients with chronic HCV.22 Another pertinent caveat is that most of our data relied on real-time PCR analysis to quantify RNA expression. We included selected analyses of protein levels and infectious virus levels to support these findings and all have correlated with the RNA results. Finally, our findings do not indicate whether the association between HCV replication and Hh pathway activity is direct or indirect. Further exploration of changes in the microenvironment of Hh-responsive cells will be required to determine the specific proteins that mediate changes in cellular content of HCV RNA and the contribution of viral entry and replication in supporting this microenvironment.
In conclusion, we have discovered a significant association between HCV replication and Hh pathway activity that has not been previously described. The implications of this work are that subpopulations of hepatocytes may support disproportionately high levels of HCV replication and Hh-mediated effects on HCV replication may represent an important new therapeutic target or reveal other intracellular targets against HCV.
The authors thank Thiago Pereira, Rafal Witek, and all members of the Diehl laboratory for their support. We also thank Michael Cohen-Wolkowiez for assistance with IC50 estimation, Carlos Goller for assistance with the LDH release assays, and Stan Lemon, Patrick Seed, Raphael Valdivia, Bryan Cullen, and Robert Mook for their constructive comments on this project and article.